Compare commits

...
Author SHA1 Message Date
bquarkz d7c2ec880f implements PLN-0126
Intrepid/Prometeu/Studio/pipeline/pr-master Build started...
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.95% (17766/28676) * Branch Coverage: 52.69% (6860/13019) * Lines of Code: 28676 * Cyclomatic Complexity: 11490 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 645
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Publish the language-versus-platform validation matrix in compiler-general
specs, retitle spec 19, and recut PBS diagnostics so common structural
failures are not PBS coverage.

Housekeep DSC-0060 with LSN-0063.
2026-09-19 01:11:50 +01:00
bquarkz c248b91147 Merge pull request 'implements PLN-0125' (#19) from dev/multi-frontend-common-lifecycle into master
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.95% (17766/28676) * Branch Coverage: 52.69% (6860/13019) * Lines of Code: 28676 * Cyclomatic Complexity: 11490 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 643
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Reviewed-on: #19
2026-09-18 23:40:50 +00:00
bquarkz 46bd0c42d3 implements PLN-0125
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.95% (17766/28676) * Branch Coverage: 52.69% (6860/13019) * Lines of Code: 28676 * Cyclomatic Complexity: 11490 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 643
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
Move lifecycle assembly out of the PBS frontend into a common
AssembleLifecyclePipelineStage. PBS emits a typed IRLifecycleDeclaration;
the assembler owns wrappers, boot guard, and single project-init execution.

Housekeep DSC-0059 with LSN-0062, and include the DSC-0062 PVM/PBX
neutrality spec and lesson already present on this branch.
2026-09-19 00:19:45 +01:00
bquarkz c6ef56114f Merge pull request 'dev/multi-frontend-serializable-ir' (#18) from dev/multi-frontend-serializable-ir into master
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.73% (17590/28493) * Branch Coverage: 52.45% (6778/12924) * Lines of Code: 28493 * Cyclomatic Complexity: 11423 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 624
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Reviewed-on: #18
2026-07-15 12:47:12 +00:00
bquarkz 701f5b0339 housekeep DSC-0058
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.73% (17590/28493) * Branch Coverage: 52.45% (6778/12924) * Lines of Code: 28493 * Cyclomatic Complexity: 11423 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 624
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
2026-07-15 13:12:36 +01:00
bquarkz cc4cb23b76 implements PLN-0123 2026-07-15 13:03:57 +01:00
bquarkz f42c83bdca implements PLN-0122 2026-07-15 13:02:09 +01:00
bquarkz cf14039077 implements PLN-0121 2026-07-15 12:59:35 +01:00
bquarkz 1f1a92d43f implements PLN-0120 2026-07-15 12:57:36 +01:00
bquarkz 38790c0b45 implements PLN-0119 2026-07-15 12:55:50 +01:00
bquarkz 5f6bd408a4 Unique and Serializable IR 2026-07-15 12:54:39 +01:00
bquarkz a77dce1c56 Merge pull request 'dev/multi-frontend-frontend-backend-contract' (#17) from dev/multi-frontend-frontend-backend-contract into master
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.73% (17588/28493) * Branch Coverage: 52.44% (6777/12924) * Lines of Code: 28493 * Cyclomatic Complexity: 11423 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 619
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Reviewed-on: #17
2026-07-15 11:45:42 +00:00
bquarkz b9a32e460a housekeep DSC-0057
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.73% (17588/28493) * Branch Coverage: 52.44% (6777/12924) * Lines of Code: 28493 * Cyclomatic Complexity: 11423 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 619
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
2026-07-15 12:18:44 +01:00
bquarkz e8079b67fb implements PLN-0118 2026-07-15 12:13:40 +01:00
bquarkz bde999f1fe implements PLN-0117 2026-07-15 12:10:10 +01:00
bquarkz 26c69d5660 Stable contract FE/BE 2026-07-15 12:07:07 +01:00
bquarkz ffd2975484 Merge pull request 'dev/multi-frontend-remove-pbs-branches' (#16) from dev/multi-frontend-remove-pbs-branches into master
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.73% (17588/28493) * Branch Coverage: 52.44% (6777/12924) * Lines of Code: 28493 * Cyclomatic Complexity: 11423 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 616
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Reviewed-on: #16
2026-07-15 10:47:09 +00:00
bquarkz a62ddb0f74 LSP contract generalization
Intrepid/Prometeu/Studio/pipeline/head Build started...
Intrepid/Prometeu/Studio/pipeline/pr-master Build started...
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.73% (17588/28493) * Branch Coverage: 52.44% (6777/12924) * Lines of Code: 28493 * Cyclomatic Complexity: 11423 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 616
2026-07-15 11:40:02 +01:00
bquarkz 79e4a64ff2 implements PLN-0116 2026-07-15 11:33:43 +01:00
bquarkz 3ca58dcc24 implements PLN-0115 2026-07-15 11:31:02 +01:00
bquarkz bb8669bf46 implements PLN-0114 2026-07-15 11:29:02 +01:00
bquarkz 6205b82e0b implements PLN-0113 2026-07-15 11:26:53 +01:00
bquarkz 088101449a implements PLN-0112 2026-07-15 11:24:22 +01:00
bquarkz b1e75973f9 implements PLN-0111 2026-07-15 11:22:38 +01:00
bquarkz 3ee0da3c0a cleanup
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.54% (17430/28323) * Branch Coverage: 52.41% (6739/12858) * Lines of Code: 28323 * Cyclomatic Complexity: 11353 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 611
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
2026-07-15 10:40:50 +01:00
bquarkz 29f46c6606 Merge pull request 'dev/multi-frontend-compiler-vs-language-services' (#15) from dev/multi-frontend-compiler-vs-language-services into master
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.54% (17430/28323) * Branch Coverage: 52.41% (6739/12858) * Lines of Code: 28323 * Cyclomatic Complexity: 11353 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 611
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Reviewed-on: #15
2026-07-15 09:32:07 +00:00
bquarkz 30e3e4145a Merge branch 'master' into dev/multi-frontend-compiler-vs-language-services
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.54% (17430/28323) * Branch Coverage: 52.41% (6739/12858) * Lines of Code: 28323 * Cyclomatic Complexity: 11353 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 611
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
2026-07-15 10:01:21 +01:00
bquarkz 4233b90b82 housekeep DSC-0055
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.54% (17430/28323) * Branch Coverage: 52.41% (6739/12858) * Lines of Code: 28323 * Cyclomatic Complexity: 11353 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 611
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
2026-07-15 09:55:08 +01:00
bquarkz da00e9c50c implements PLN-0110 2026-07-15 09:45:49 +01:00
bquarkz fee89d7522 implements PLN-0109 2026-07-15 09:43:30 +01:00
bquarkz bc95b33bd9 implements PLN-0108 2026-07-15 09:41:54 +01:00
bquarkz 6cd24bf14f implements PLN-0107 2026-07-15 09:40:13 +01:00
bquarkz b86cf0117d Merge pull request 'dev/multi-frontend-provider-contract' (#14) from dev/multi-frontend-provider-contract into master
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.49% (17397/28294) * Branch Coverage: 52.32% (6724/12852) * Lines of Code: 28294 * Cyclomatic Complexity: 11339 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 608
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Reviewed-on: #14
2026-07-15 07:42:35 +00:00
bquarkz 6fdb4828c3 housekeep DSC-0054
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.49% (17397/28294) * Branch Coverage: 52.32% (6724/12852) * Lines of Code: 28294 * Cyclomatic Complexity: 11339 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 608
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
2026-07-15 08:37:22 +01:00
bquarkz 44f03f771b implements PLN-0106
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.49% (17397/28294) * Branch Coverage: 52.32% (6724/12852) * Lines of Code: 28294 * Cyclomatic Complexity: 11339 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 608
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
2026-07-15 08:32:19 +01:00
bquarkz ba3af3c01b implements PLN-0105 2026-07-15 08:29:51 +01:00
bquarkz 66c1b1a033 implements PLN-0104 2026-07-15 08:28:30 +01:00
bquarkz f8b9c5b27d implements PLN-0103 2026-07-15 08:26:37 +01:00
bquarkz e99bb58453 Multi Provider FE content 2026-07-15 08:24:56 +01:00
bquarkz db0b71c9ad fix issues with wrong magic number on PBX
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.37% (17351/28272) * Branch Coverage: 52.31% (6722/12850) * Lines of Code: 28272 * Cyclomatic Complexity: 11325 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 601
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
2026-07-15 08:15:09 +01:00
bquarkz 35b65bb524 multi FE adjustment agendas
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.37% (17351/28272) * Branch Coverage: 52.31% (6722/12850) * Lines of Code: 28272 * Cyclomatic Complexity: 11325 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 601
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
2026-07-15 07:58:23 +01:00
bquarkz 6e594c15d7 add LSP agendas
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.37% (17351/28272) * Branch Coverage: 52.31% (6722/12850) * Lines of Code: 28272 * Cyclomatic Complexity: 11325 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 601
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
2026-07-15 07:40:45 +01:00
bquarkz d58a96a4c4 Merge pull request 'dev/pbs-symbol-documentation-and-hover-markdown' (#13) from dev/pbs-symbol-documentation-and-hover-markdown into master
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.37% (17351/28272) * Branch Coverage: 52.31% (6722/12850) * Lines of Code: 28272 * Cyclomatic Complexity: 11325 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 601
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
Reviewed-on: #13
2026-07-15 06:36:29 +00:00
bquarkz 725c28b3f5 implements PLN-0102
JaCoCo Coverage #### Project Overview No changes detected, that affect the code coverage. * Line Coverage: 61.37% (17351/28272) * Branch Coverage: 52.31% (6722/12850) * Lines of Code: 28272 * Cyclomatic Complexity: 11325 #### Quality Gates Summary Output truncated.
Test / Build skipped: 15, passed: 601
Intrepid/Prometeu/Studio/pipeline/pr-master This commit looks good
Intrepid/Prometeu/Studio/pipeline/head This commit looks good
2026-07-15 07:22:58 +01:00
bquarkz d8db99eb81 implements PLN-0101 2026-07-15 07:20:42 +01:00
bquarkz 63215ac02f implements PLN-0100 2026-07-15 07:18:19 +01:00
bquarkz e7647165bc implements PLN-0099 2026-07-15 07:12:03 +01:00
bquarkz f32a25eb59 implements PLN-0098 2026-07-15 07:10:15 +01:00
bquarkz d680504c31 implements PLN-0097 2026-07-15 07:08:11 +01:00
bquarkz aeea8ca900 implements PLN-0096 2026-07-15 07:05:45 +01:00
bquarkz 05dd37a30a implements PLN-0095 2026-07-15 07:00:43 +01:00
bquarkz f700e400d5 implements PLN-0094 2026-07-15 06:59:16 +01:00
bquarkz 2b25364252 implements PLN-0093 2026-07-15 06:57:19 +01:00
137 changed files with 7717 additions and 1794 deletions
+29 -2
View File
@@ -1,7 +1,34 @@
{"type":"meta","next_id":{"DSC":39,"AGD":42,"DEC":40,"PLN":103,"LSN":55,"CLSN":1}}
{"type":"meta","next_id":{"DSC":66,"AGD":69,"DEC":48,"PLN":127,"LSN":64,"CLSN":1}}
{"type":"discussion","id":"DSC-0065","status":"open","ticket":"multi-frontend-avoid-premature-abstractions","title":"Evitar abstracoes prematuras na preparacao multi-frontend","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","studio","frontend","architecture","multi-frontend","simplicity"],"agendas":[{"id":"AGD-0068","file":"AGD-0068-multi-frontend-avoid-premature-abstractions.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0064","status":"open","ticket":"multi-frontend-architectural-tests","title":"Testes arquiteturais para fronteiras multi-frontend","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","studio","frontend","architecture","tests","multi-frontend"],"agendas":[{"id":"AGD-0067","file":"AGD-0067-multi-frontend-architectural-tests.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0063","status":"open","ticket":"multi-frontend-synthetic-test-frontend","title":"Frontend sintetico de teste para provar neutralidade do pipeline","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","frontend","tests","backend","multi-frontend"],"agendas":[{"id":"AGD-0066","file":"AGD-0066-multi-frontend-synthetic-test-frontend.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0062","status":"done","ticket":"multi-frontend-pvm-neutrality","title":"Neutralidade da PVM e identificacao PBX independente de PBS","created_at":"2026-07-15","updated_at":"2026-09-18","tags":["vm-arch","runtime","pvm","pbx","compiler","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0061","file":"discussion/lessons/DSC-0062-multi-frontend-pvm-neutrality/LSN-0061-pbx-runtime-boundary.md","status":"done","created_at":"2026-09-18","updated_at":"2026-09-18"}]}
{"type":"discussion","id":"DSC-0061","status":"open","ticket":"multi-frontend-sdk-canonical-definition","title":"Auditoria e centralizacao gradual da definicao canonica do SDK","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","sdk","stdlib","hostcalls","intrinsics","multi-frontend"],"agendas":[{"id":"AGD-0064","file":"AGD-0064-multi-frontend-sdk-canonical-definition.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0060","status":"done","ticket":"multi-frontend-validation-boundaries","title":"Separar validacoes de linguagem e validacoes de plataforma","created_at":"2026-07-15","updated_at":"2026-09-19","tags":["compiler","compiler-general","compiler-pbs","backend","validation","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0063","file":"discussion/lessons/DSC-0060-multi-frontend-validation-boundaries/LSN-0063-language-vs-platform-validation-ownership.md","status":"done","created_at":"2026-09-19","updated_at":"2026-09-19"}]}
{"type":"discussion","id":"DSC-0059","status":"done","ticket":"multi-frontend-common-lifecycle","title":"Extrair lifecycle comum das responsabilidades do frontend PBS","created_at":"2026-07-15","updated_at":"2026-09-19","tags":["compiler","compiler-general","compiler-pbs","lifecycle","backend","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0062","file":"discussion/lessons/DSC-0059-multi-frontend-common-lifecycle/LSN-0062-common-lifecycle-assembly-stage.md","status":"done","created_at":"2026-09-19","updated_at":"2026-09-19"}]}
{"type":"discussion","id":"DSC-0058","status":"done","ticket":"multi-frontend-serializable-ir","title":"Manter a IR comum serializavel por design","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","ir","backend","serialization","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0060","file":"discussion/lessons/DSC-0058-multi-frontend-serializable-ir/LSN-0060-data-contract-first-for-serializable-irbackend.md","status":"done","created_at":"2026-07-15","updated_at":"2026-07-15"}]}
{"type":"discussion","id":"DSC-0057","status":"done","ticket":"multi-frontend-frontend-backend-contract","title":"Estabilizar contrato entre frontend e backend comum","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","compiler-pbs","ir","backend","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0059","file":"discussion/lessons/DSC-0057-multi-frontend-frontend-backend-contract/LSN-0059-common-irbackend-handoff-and-backend-guardrails.md","status":"done","created_at":"2026-07-15","updated_at":"2026-07-15"}]}
{"type":"discussion","id":"DSC-0056","status":"done","ticket":"multi-frontend-remove-pbs-branches","title":"Generalizar o contrato LSP/editorial para frontends","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","compiler-pbs","studio","frontend","coupling","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0058","file":"discussion/lessons/DSC-0056-multi-frontend-remove-pbs-branches/LSN-0058-generic-frontend-editorial-contract-for-lsp.md","status":"done","created_at":"2026-07-15","updated_at":"2026-07-15"}]}
{"type":"discussion","id":"DSC-0055","status":"done","ticket":"multi-frontend-compiler-vs-language-services","title":"Separar compilacao de servicos editoriais de frontend","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","studio","lsp","editor","frontend","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0056","file":"discussion/lessons/DSC-0055-multi-frontend-compiler-vs-language-services/LSN-0056-compile-first-frontends-with-optional-editorial-capabilities.md","status":"done","created_at":"2026-07-15","updated_at":"2026-07-15"}]}
{"type":"discussion","id":"DSC-0054","status":"done","ticket":"multi-frontend-provider-contract","title":"Introduzir provider completo de frontend","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["compiler","compiler-general","studio","frontend","registry","multi-frontend"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0055","file":"discussion/lessons/DSC-0054-multi-frontend-provider-contract/LSN-0055-static-frontend-providers-before-plugin-architecture.md","status":"done","created_at":"2026-07-15","updated_at":"2026-07-15"}]}
{"type":"discussion","id":"DSC-0053","status":"open","ticket":"pbs-lsp-call-and-type-hierarchy","title":"PBS LSP Call and Type Hierarchy","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","call-hierarchy","type-hierarchy"],"agendas":[{"id":"AGD-0056","file":"AGD-0056-pbs-lsp-call-and-type-hierarchy.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0052","status":"open","ticket":"pbs-lsp-document-links","title":"PBS LSP Document Links","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","document-links","imports"],"agendas":[{"id":"AGD-0055","file":"AGD-0055-pbs-lsp-document-links.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0051","status":"open","ticket":"pbs-lsp-diagnostics-ux","title":"PBS LSP Diagnostics UX","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","diagnostics","ux"],"agendas":[{"id":"AGD-0054","file":"AGD-0054-pbs-lsp-diagnostics-ux.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0050","status":"open","ticket":"pbs-lsp-folding-and-selection-ranges","title":"PBS LSP Folding and Selection Ranges","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","folding","selection-range"],"agendas":[{"id":"AGD-0053","file":"AGD-0053-pbs-lsp-folding-and-selection-ranges.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0049","status":"open","ticket":"pbs-lsp-completion-depth","title":"PBS LSP Completion Depth","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","completion"],"agendas":[{"id":"AGD-0052","file":"AGD-0052-pbs-lsp-completion-depth.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0048","status":"open","ticket":"pbs-lsp-performance-cache-and-snapshots","title":"PBS LSP Performance Cache and Snapshots","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","performance","caching"],"agendas":[{"id":"AGD-0051","file":"AGD-0051-pbs-lsp-performance-cache-and-snapshots.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0047","status":"open","ticket":"pbs-lsp-semantic-tokens-semantic-classification","title":"PBS LSP Semantic Token Classification","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","semantic-tokens"],"agendas":[{"id":"AGD-0050","file":"AGD-0050-pbs-lsp-semantic-token-classification.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0046","status":"open","ticket":"pbs-lsp-import-assistance","title":"PBS LSP Import Assistance","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","imports","completion"],"agendas":[{"id":"AGD-0049","file":"AGD-0049-pbs-lsp-import-assistance.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0045","status":"open","ticket":"pbs-lsp-formatting","title":"PBS LSP Formatting","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","formatting"],"agendas":[{"id":"AGD-0048","file":"AGD-0048-pbs-lsp-formatting.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0044","status":"open","ticket":"pbs-lsp-code-actions","title":"PBS LSP Code Actions and Quick Fixes","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","code-actions","quick-fix"],"agendas":[{"id":"AGD-0047","file":"AGD-0047-pbs-lsp-code-actions-and-quick-fixes.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0043","status":"open","ticket":"pbs-lsp-rename-symbol","title":"PBS LSP Rename Symbol","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","rename"],"agendas":[{"id":"AGD-0046","file":"AGD-0046-pbs-lsp-rename-symbol.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0042","status":"open","ticket":"pbs-lsp-workspace-symbols","title":"PBS LSP Workspace Symbols","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","workspace-symbols"],"agendas":[{"id":"AGD-0045","file":"AGD-0045-pbs-lsp-workspace-symbols.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0041","status":"open","ticket":"pbs-lsp-document-symbols-outline","title":"PBS LSP Document Symbols and Outline","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","outline","document-symbols"],"agendas":[{"id":"AGD-0044","file":"AGD-0044-pbs-lsp-document-symbols-and-outline.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0040","status":"open","ticket":"pbs-lsp-find-references","title":"PBS LSP Find References","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","references"],"agendas":[{"id":"AGD-0043","file":"AGD-0043-pbs-lsp-find-references.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0039","status":"open","ticket":"pbs-lsp-go-to-definition","title":"PBS LSP Go to Definition","created_at":"2026-07-15","updated_at":"2026-07-15","tags":["studio","lsp","vscode","compiler-pbs","editor","definition"],"agendas":[{"id":"AGD-0042","file":"AGD-0042-pbs-lsp-go-to-definition.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15"}],"decisions":[],"plans":[],"lessons":[]}
{"type":"discussion","id":"DSC-0038","status":"done","ticket":"studio-packer-rgba8888-asset-pipeline","title":"Studio and Packer RGBA8888 Asset Pipeline Alignment","created_at":"2026-05-23","updated_at":"2026-07-14","tags":["studio","packer","assets","glyph-bank","palette","rgba8888","runtime-alignment"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0053","file":"discussion/lessons/DSC-0038-studio-packer-rgba8888-asset-pipeline/LSN-0053-rgba8888-is-the-canonical-studio-packer-palette-contract.md","status":"done","created_at":"2026-07-14","updated_at":"2026-07-14"}]}
{"type":"discussion","id":"DSC-0037","status":"done","ticket":"pbs-autocomplete-parameter-names","title":"PBS autocomplete parameter names for stdlib and method calls","created_at":"2026-05-08","updated_at":"2026-05-14","tags":["compiler-pbs","studio","lsp","autocomplete","signature-help","stdlib"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0052","file":"discussion/lessons/DSC-0037-pbs-autocomplete-parameter-names/LSN-0052-canonical-callable-parameter-names-through-pbs-editor-assistance.md","status":"done","created_at":"2026-05-14","updated_at":"2026-05-14"}]}
{"type":"discussion","id":"DSC-0036","status":"in_progress","ticket":"pbs-symbol-documentation-and-hover-markdown","title":"Modelo de documentacao de simbolos em PBS e consumo markdown no hover","created_at":"2026-05-08","updated_at":"2026-07-15","tags":["compiler","compiler-pbs","studio","lsp","vscode","editor","hover","documentation","markdown"],"agendas":[{"id":"AGD-0039","file":"AGD-0039-pbs-symbol-documentation-and-hover-markdown.md","status":"accepted","created_at":"2026-05-08","updated_at":"2026-07-15"}],"decisions":[{"id":"DEC-0039","file":"DEC-0039-pbs-symbol-documentation-with-doc-markdown-text-blocks.md","status":"accepted","created_at":"2026-07-15","updated_at":"2026-07-15","ref_agenda":"AGD-0039"}],"plans":[{"id":"PLN-0092","file":"PLN-0092-specify-pbs-doc-attribute-and-markdown-text-block-syntax.md","status":"done","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0093","file":"PLN-0093-implement-pbs-lexer-support-for-documentation-text-blocks.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0094","file":"PLN-0094-extend-pbs-attribute-parser-and-ast-for-doc-markdown-payloads.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0095","file":"PLN-0095-implement-documentation-text-block-normalization.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0096","file":"PLN-0096-validate-doc-attribute-semantics-and-diagnostics.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0097","file":"PLN-0097-attach-doc-metadata-to-pbs-semantic-symbols.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0098","file":"PLN-0098-expose-doc-metadata-through-pbs-editorial-and-lsp-surfaces.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0099","file":"PLN-0099-render-doc-markdown-in-hover-composition.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0100","file":"PLN-0100-author-doc-metadata-for-stdlib-sdk-and-interface-declarations.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0101","file":"PLN-0101-protect-runtime-artifacts-from-doc-metadata-lowering.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]},{"id":"PLN-0102","file":"PLN-0102-add-end-to-end-doc-documentation-conformance-coverage.md","status":"open","created_at":"2026-07-15","updated_at":"2026-07-15","ref_decisions":["DEC-0039"]}],"lessons":[]}
{"type":"discussion","id":"DSC-0036","status":"done","ticket":"pbs-symbol-documentation-and-hover-markdown","title":"Modelo de documentacao de simbolos em PBS e consumo markdown no hover","created_at":"2026-05-08","updated_at":"2026-07-15","tags":["compiler","compiler-pbs","studio","lsp","vscode","editor","hover","documentation","markdown"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0057","file":"discussion/lessons/DSC-0036-pbs-symbol-documentation-and-hover-markdown/LSN-0057-compiler-owned-pbs-doc-markdown-for-editor-assistance.md","status":"done","created_at":"2026-07-15","updated_at":"2026-07-15"}]}
{"type":"discussion","id":"DSC-0035","status":"done","ticket":"pbs-lsp-editor-assistance-wave-1","title":"Wave 1 de assistencia editorial via LSP para PBS no VS Code","created_at":"2026-05-08","updated_at":"2026-05-08","tags":["studio","lsp","vscode","compiler","compiler-pbs","editor","completion","hover","signature-help"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0051","file":"discussion/lessons/DSC-0035-pbs-lsp-editor-assistance-wave-1/LSN-0051-compiler-backed-editor-assistance-for-pbs.md","status":"done","created_at":"2026-05-08","updated_at":"2026-05-08"}]}
{"type":"discussion","id":"DSC-0034","status":"done","ticket":"frontend-semantic-host-projection-flexibility","title":"Frontend semantic vocabulary flexibility and declarative host projection","created_at":"2026-05-06","updated_at":"2026-05-07","tags":["compiler","compiler-general","frontend","semantics","vscode","host-projection","lsp"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0048","file":"discussion/lessons/DSC-0034-frontend-semantic-host-projection-flexibility/LSN-0048-frontend-owned-semantic-vocabularies-with-declarative-host-projection.md","status":"done","created_at":"2026-05-07","updated_at":"2026-05-07"}]}
{"type":"discussion","id":"DSC-0033","status":"done","ticket":"frontend-visual-theme-spec-and-css-retirement","title":"Frontend visual theme spec and retirement of host-consumed semantic CSS","created_at":"2026-05-06","updated_at":"2026-05-08","tags":["compiler","compiler-general","frontend","presentation","theming","studio","vscode","lsp","pbs"],"agendas":[],"decisions":[],"plans":[],"lessons":[{"id":"LSN-0050","file":"discussion/lessons/DSC-0033-frontend-visual-theme-spec-and-css-retirement/LSN-0050-frontend-owned-visual-themes-with-structured-contract-and-host-adapters.md","status":"done","created_at":"2026-05-08","updated_at":"2026-05-08"}]}
@@ -0,0 +1,104 @@
---
id: LSN-0057
ticket: pbs-symbol-documentation-and-hover-markdown
title: Compiler-Owned PBS Doc Markdown for Editor Assistance
created: 2026-07-15
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Original Problem
PBS editor assistance had compiler-backed completion, hover, and signature-help surfaces, but it did not yet have a canonical authored documentation model for named API symbols.
Without that model, hover could show mechanical facts such as kind, signature, origin, and type shape, but it could not reliably explain intent, parameter meaning, semantic constraints, or usage notes. Any editor-side solution would have risked reparsing PBS source or inventing a second documentation binding model outside the compiler.
## Consolidated Decision
PBS symbol documentation is authored with a reserved compiler-recognized `Doc` attribute:
```pbs
[Doc(markdown = """
Draws a sprite.
- `x`: screen x coordinate
- `y`: screen y coordinate
""")]
```
The v1 shape is intentionally narrow:
1. `Doc` has exactly one required named argument, `markdown`.
2. The `markdown` argument uses a triple-quoted documentation text block.
3. The text block is documentation-only metadata, not a general runtime multiline string literal.
4. `[Doc]` attaches to the following declaration using the normal PBS attribute association rule.
5. Attribute order is semantically irrelevant.
6. Parameters do not receive `[Doc]` in v1; parameter notes live inside the owning declaration's Markdown.
7. Duplicate `Doc`, missing `markdown`, extra arguments, aliases, positional arguments, empty normalized payloads, and invalid targets are semantic errors.
The compiler owns documentation resolution. LSP and editor consumers receive normalized compiler metadata; they must not reparse PBS source to discover documentation.
## Final Implementation Shape
The work landed as a staged compiler-to-editor pipeline:
1. PBS specs define the syntax, static semantics, AST shape, diagnostics, and stdlib documentation policy for `[Doc(markdown = """...""")]`.
2. The PBS lexer recognizes triple-quoted documentation text blocks without changing ordinary string literal behavior.
3. The parser and AST preserve documentation text block payloads and spans as structured attribute values.
4. A shared normalizer removes incidental surrounding blank lines and common indentation while preserving Markdown content and internal line breaks.
5. Semantic validation reserves `Doc` and enforces the single canonical v1 shape.
6. Normalized Markdown is attached to resolved PBS semantic symbols as structured documentation metadata.
7. PBS editorial and LSP surfaces transport that metadata without source reparsing.
8. Hover composition includes the normalized Markdown while preserving existing fallback behavior for undocumented symbols.
9. Stdlib, SDK, and interface declarations use the same authored documentation surface.
10. Runtime-facing artifacts remain unchanged by documentation metadata.
This preserves a clean boundary: PBS source authors write documentation in the language, the compiler resolves it, LSP transports it, and editors render it.
## Examples
Valid documentation belongs on named API declarations that introduce semantic symbols and can carry attributes, such as functions, methods, services, hosts, interface declarations, and repository-owned stdlib or SDK declarations.
Parameter documentation is written inside the callable's Markdown:
```pbs
[Doc(markdown = """
Fills a rectangle.
- `x`: left edge in screen coordinates
- `y`: top edge in screen coordinates
- `width`: rectangle width in pixels
- `height`: rectangle height in pixels
""")]
```
Invalid v1 forms include:
```pbs
[Doc("Draws a sprite.")]
[Doc(text = """Draws a sprite.""")]
[Doc(markdown = """Draws a sprite.""", format = "markdown")]
[Documentation(markdown = """Draws a sprite.""")]
```
These are rejected because accepting aliases or positional shortcuts would immediately expand the language contract and make future compatibility harder.
## Pitfalls
Do not treat `"""..."""` as a general PBS runtime string literal. Runtime multiline strings raise separate questions about type behavior, lowering, constant pools, equality, escaping, and artifact representation.
Do not let LSP or editor code rediscover documentation by scanning source. That duplicates language semantics outside the compiler and breaks as soon as imports, stdlib declarations, interface modules, or generated sources participate.
Do not attach `[Doc]` to parameters in v1. Parameter docs are authored inside the declaration Markdown so the compiler can keep one documentation payload per API symbol.
Do not reflow authored Markdown in a formatter or hover composer. The normalizer may remove incidental indentation, but it must preserve the author's Markdown content.
Do not lower documentation into runtime artifacts by default. Documentation is compile/editor metadata unless a future runtime or packaging decision explicitly changes that boundary.
## References
- `DEC-0039` captured the normative decision for PBS `Doc` markdown text blocks.
- `PLN-0092` specified the PBS docs and diagnostics contract.
- `PLN-0093` through `PLN-0099` implemented lexer, parser, normalization, semantics, symbol metadata, LSP exposure, and hover composition.
- `PLN-0100` authored stdlib, SDK, and interface documentation.
- `PLN-0101` protected runtime artifacts from documentation lowering.
- `PLN-0102` closed end-to-end conformance coverage.
@@ -0,0 +1,74 @@
---
id: LSN-0055
ticket: multi-frontend-provider-contract
title: Static frontend providers before plugin architecture
created: 2026-07-15
tags: [compiler, compiler-general, studio, frontend, registry, multi-frontend]
---
## Context
Prometeu needed a multi-frontend boundary without turning the compiler or Studio into a dynamic plugin platform. Before this work, shared code could ask the registry for frontend specs or phase services, and PBS was registered directly as the default language. That worked for a single frontend, but it made the next frontend harder because common code could still depend on PBS-specific construction details.
The durable ownership rule is now: common compiler and Studio code select a frontend by `languageId`, then consume a provider contract. PBS remains the first and only real provider, but it is no longer the shape of the shared abstraction.
## Key Decisions
### `FrontendProvider` is the common registration unit
**What:** Each registered frontend is represented by a `FrontendProvider` with:
- `specification()`, for the static frontend contract and `languageId`;
- `compiler()`, for compiler-facing frontend execution;
- `languageService()`, as an optional editor-facing capability.
**Why:** This groups the frontend identity, compiler entrypoint, and optional tooling capability without forcing every frontend to implement Studio/editor services.
**Trade-offs:** The provider is intentionally small. It does not solve dynamic discovery, plugin installation, external loading, or all future language-service APIs. Those are separate architectural decisions.
### Registration is explicit
**What:** Frontend registration is static and composition-root-owned. The application composition path bootstraps the default provider set, and the registry resolves providers by `languageId`.
**Why:** Explicit registration keeps dependencies visible and testable. It lets composition know about PBS while preventing common compiler and Studio consumers from directly constructing PBS services.
**Trade-offs:** Adding a new in-repository frontend still requires a code change in composition. That is acceptable because this decision optimizes for repository-internal multi-frontend readiness, not external plugin extensibility.
### PBS is a provider, not the shared model
**What:** PBS is wrapped by `PBSFrontendProvider`. Shared build and Studio consumers obtain frontend behavior through `FrontendRegistryService.require(languageId)` or provider-derived listings.
**Why:** PBS remains the correctness baseline, but it must not become the exclusive owner of compiler pipeline semantics.
**Trade-offs:** PBS-specific tests may still instantiate `PBSFrontendPhaseService` directly because they test PBS internals. Common modules should not.
## Patterns and Algorithms
Use this path for common compiler execution:
1. Resolve the project language from the manifest or default provider.
2. Resolve `FrontendProvider` by `languageId`.
3. Use `provider.specification()` for static metadata.
4. Use `provider.compiler()` for frontend compilation.
5. Use `provider.languageService()` only when a tooling consumer needs optional editor services.
Use this path for Studio language catalogs:
1. List providers from the registry.
2. Derive templates from `provider.specification()`.
3. Do not hardcode PBS templates in Studio catalog code.
## Pitfalls
- Do not add reflection, classpath scanning, external JAR loading, or runtime plugin installation to solve repository-internal frontend registration.
- Do not make `languageService()` mandatory. Compilation and editor assistance are separate capabilities.
- Do not instantiate `PBSFrontendPhaseService` from common build, dependency, Studio, or app code.
- Do not use PBS as a fallback for unknown `languageId`. Unknown languages must fail explicitly.
- Do not move frontend semantic ownership into Studio just because Studio consumes provider metadata.
## Takeaways
- Static providers are enough to prepare for multiple in-repository frontends.
- The registry boundary is provider-based, not PBS-based.
- Composition may know which frontends exist; common consumers should only know `FrontendProvider`.
- Optional editor capabilities keep compiler readiness independent from Studio feature depth.
@@ -0,0 +1,88 @@
# Compile-first frontends with optional editorial capabilities
## Original Problem
Prometeu is moving from a PBS-centered compiler/editor stack toward a multi-frontend compiler. PBS already provides compilation plus rich editor assistance, so it was easy for shared code, LSP, or Studio consumers to assume that every frontend can provide diagnostics, semantic tokens, completion, hover, and signature help as one package.
That assumption makes future frontend adoption too expensive. A new frontend must be able to compile before it has a complete IDE experience.
## Consolidated Decision
The frontend contract is compile-first.
Every frontend must provide a compiler capability through `FrontendProvider.compiler()`. Editorial services are optional and live behind `FrontendProvider.languageService()`.
A frontend that can compile but has no language service is valid. Missing editor assistance must not block `analyze`, `compile`, `build`, backend lowering, bytecode emission, verification, or artifact writing.
LSP and Studio consumers must query editorial capability before invoking editor-specific behavior. Absence of a capability is a normal state, not a compiler failure.
## Final Implementation
The compiler pipeline spec now defines the provider boundary explicitly:
- `compiler()` is the only required executable frontend capability.
- `languageService()` is optional.
- editor capabilities such as completion, hover, definition, signature help, semantic tokens, formatting, rename, and code actions are optional unless a future accepted decision narrows the requirement.
- absent editor capabilities have deterministic fallback behavior.
- compiler diagnostics and editor diagnostics are separate ownership surfaces.
PBS keeps its rich editor assistance, but now exposes it through `PBSFrontendLanguageService`, returned by `PBSFrontendProvider.languageService()`.
The LSP bridge resolves the active `FrontendProvider` and checks for `languageService()` before using editor behavior. For now, PBS remains an adapter-specific implementation because the common language-service DTOs are not yet generalized. That is acceptable because the important boundary is already enforced: the LSP no longer assumes that compiler support implies editor support.
Conformance coverage now includes:
- a compile-only provider fixture with no language service;
- LSP neutral responses when editor capability is absent;
- PBS regression coverage for existing completion, hover, signature help, semantic tokens, and documentation behavior;
- architecture tests preventing common modules from directly instantiating PBS editor support.
## Examples
A compile-only frontend should look like this at the provider boundary:
```java
public final class ExampleFrontendProvider implements FrontendProvider {
@Override
public FrontendSpec specification() {
return EXAMPLE_SPEC;
}
@Override
public FrontendPhaseService compiler() {
return exampleCompiler;
}
}
```
It does not need to override `languageService()`. The default empty optional is a valid frontend state.
An editor consumer should treat missing services as a normal fallback:
```java
return provider.languageService()
.filter(ExpectedLanguageService.class::isInstance)
.map(ExpectedLanguageService.class::cast)
.map(service -> service.completion(...))
.orElseGet(List::of);
```
## Pitfalls
Do not use PBS as the minimum frontend contract. PBS is the first rich frontend, not the definition of frontend validity.
Do not make editor diagnostics a prerequisite for compilation. Build/analyze diagnostics are compiler output; live editor diagnostics may reuse compiler results but have different latency, cancellation, overlay, and UX concerns.
Do not split the language service into many top-level interfaces before implementation pressure proves a real lifecycle or ownership boundary. The current contract intentionally keeps an aggregated optional language-service surface.
Do not call PBS editorial classes directly from common compiler, LSP, Studio, or app code. Go through `FrontendProvider.languageService()` and adapter-local capability checks.
## References
- Spec: `docs/specs/compiler/23. Compiler Pipeline Entry Points Specification.md`
- Provider API: `prometeu-compiler/prometeu-frontend-api/src/main/java/p/studio/compiler/services/FrontendProvider.java`
- PBS language service: `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/PBSFrontendLanguageService.java`
- LSP bridge: `prometeu-lsp/prometeu-lsp-v1/src/main/java/p/studio/lsp/services/compiler/CompilerLanguageServiceBridge.java`
- Conformance tests:
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/specs/FrontendProviderBoundaryTest.java`
- `prometeu-lsp/prometeu-lsp-v1/src/test/java/p/studio/lsp/services/compiler/CompilerLanguageServiceBridgeTest.java`
@@ -0,0 +1,92 @@
---
id: LSN-0058
ticket: multi-frontend-remove-pbs-branches
title: Generic Frontend Editorial Contract for LSP
created: 2026-07-15
tags: [compiler, compiler-general, compiler-pbs, studio, frontend, lsp, coupling, multi-frontend]
---
## Original Problem
Prometeu was preparing for multiple compiler frontends while the LSP bridge still knew too much about PBS internals. The first framing was broad: remove explicit PBS checks from common code. After inspection, that was too blunt.
PBS was already mostly centered in `prometeu-frontend-pbs`. Defaults, templates, `.pbs` fixtures, VS Code language registration, semantic keys, and composition-root bootstrap were legitimate PBS references. The real architectural problem was narrower: common LSP code imported and called PBS-specific editorial types and services for semantic tokens, completion, hover, signature help, and semantic read preparation.
## Consolidated Decision
The common LSP/editorial path must consume frontend-provided editorial capabilities through generic contracts exposed by `prometeu-frontend-api`.
The provider contract remains stable:
1. `FrontendProvider.compiler()` is mandatory.
2. `FrontendProvider.languageService()` is optional.
3. Compile-only frontends are valid.
4. Unknown `languageId` values fail explicitly.
5. PBS may be the product default only when the user has not chosen a language.
PBS is not the shared editor model. PBS implements the generic contract from inside its frontend boundary.
## Final Implementation Shape
The implementation introduced generic frontend editorial DTOs and default methods in `FrontendLanguageService`:
1. `FrontendDocumentRequest`
2. `FrontendEditorialContext`
3. `FrontendCompletionCandidate`
4. `FrontendHover`
5. `FrontendSignatureHelp`
6. `FrontendSignature`
7. `FrontendSymbolKind`
8. `FrontendDocumentation`
`CompilerLanguageServiceBridge` now resolves the selected provider and calls generic `FrontendLanguageService` methods for semantic tokens, completion, hover, and signature help.
PBS-specific analysis setup moved behind `PBSFrontendLanguageService.prepareEditorialContext(...)`. The bridge no longer imports `p.studio.compiler.pbs.*`, no longer knows `PbsAst.File`, and no longer calls `PBSFrontendPhaseService.semanticReadSurface(...)` directly.
An architectural test now protects the common LSP code from reintroducing PBS implementation imports.
## Examples
A frontend with no language service remains valid:
```java
public final class CompileOnlyFrontendProvider implements FrontendProvider {
@Override
public FrontendSpec specification() {
return SPEC;
}
@Override
public FrontendPhaseService compiler() {
return compiler;
}
}
```
An LSP consumer should resolve the selected provider, ask for `languageService()`, and fall back deterministically when editorial capability is absent.
PBS can still keep rich internal models such as `PbsEditorialResolvedSymbol`, but those models must be mapped to generic `FrontendHover`, `FrontendCompletionCandidate`, or `FrontendSignatureHelp` before common LSP consumes them.
## Pitfalls
Do not remove legitimate PBS defaults just because they mention PBS. PBS remains the product default when the user has not chosen another language.
Do not use PBS as a fallback for unknown `languageId`. Unknown language identity is a configuration error and must fail explicitly.
Do not make editor assistance mandatory. Compilation and editor support are separate frontend capabilities.
Do not move frontend semantic ownership into LSP or Studio. The LSP bridge transports frontend-owned editorial results; it does not author language semantics.
Do not reintroduce PBS imports in common LSP code for convenience. If a capability needs PBS-specific state, create or extend a frontend-owned generic boundary and keep the PBS mapping inside `prometeu-frontend-pbs`.
## References
- `DEC-0042` locked the generic frontend editorial language-service contract.
- `PLN-0111` introduced generic frontend editorial DTOs and defaults.
- `PLN-0112` migrated semantic tokens.
- `PLN-0113` migrated completion.
- `PLN-0114` migrated hover and documentation.
- `PLN-0115` migrated signature help.
- `PLN-0116` removed PBS imports from common LSP and added boundary coverage.
- Related lesson: `LSN-0055` Static frontend providers before plugin architecture.
- Related lesson: `LSN-0056` Compile-first frontends with optional editorial capabilities.
@@ -0,0 +1,98 @@
---
id: LSN-0059
ticket: multi-frontend-frontend-backend-contract
title: Common IRBackend handoff and backend guardrails
created: 2026-07-15
tags: [compiler, compiler-general, compiler-pbs, ir, backend, multi-frontend]
---
# Common IRBackend handoff and backend guardrails
## Context
The multi-frontend work needed a clear boundary between language-specific frontends and common backend stages.
The existing code was already close to the desired architecture: PBS parsing, AST construction, and semantic analysis lowered into `IRBackend`, while the backend consumed `IRBackend` and lowered it into `IRVM`. The risk was not a known broken dependency. The risk was that common backend behavior still looked PBS-owned in specs, names, and missing guardrails, which could invite future coupling when another frontend is introduced.
## Key Decisions
### `IRBackend` is the common executable handoff
**What:** `IRBackend` remains the normative frontend-to-backend executable handoff. Frontends emit it; common backend stages consume it.
**Why:** The existing model already provides a language-neutral executable contract with modules, callable identities, executable functions, instruction metadata, globals, synthetic functions, host calls, intrinsics, spans, and capabilities. Renaming or redesigning it without a concrete leak would add churn without improving the boundary.
**Trade-offs:** Keeping the name `IRBackend` means the architecture relies on documentation and tests to communicate its common role. That is acceptable because the name is already established and the guardrails now enforce the boundary.
### PBS owns lowering into `IRBackend`, not backend consumption of it
**What:** PBS specs describe PBS admission and translation into `IRBackend`. Compiler-general specs describe the common `IRBackend -> IRVM` contract and backend obligations.
**Why:** PBS is currently the only rich frontend producing executable `IRBackend`, so many examples naturally come from PBS. Specs must still separate source-language obligations from backend obligations so future frontends do not inherit PBS-specific rules by accident.
**Trade-offs:** Some PBS-facing examples remain useful, especially around lifecycle wrappers and asset metadata. They must be phrased as current frontend sources of a common backend obligation, not as backend ownership by PBS.
## Implementation Result
The completed work:
1. rewrote the general backend spec so `IRBackend -> IRVM` is explicitly compiler-general;
2. rewrote the PBS lowering spec so PBS emits the common handoff instead of owning backend behavior;
3. updated the conformance matrix with DEC-0043 guardrail rows;
4. added a direct backend test that manually constructs common `IRBackend` and lowers it without PBS parser or frontend services;
5. added an architectural test preventing common backend imports from `p.studio.compiler.pbs`;
6. added a reflection-based contract test preventing public `IRBackend` model types from exposing PBS packages;
7. renamed the common lifecycle helper from `validatePbsLifecycleStructure` to `validateCommonLifecycleStructure`.
## Patterns and Algorithms
### Boundary consolidation before abstraction
When an existing boundary is already technically sound, prefer consolidation over redesign:
1. document the current contract as normative;
2. identify only concrete leaks;
3. add tests that would fail if the boundary regresses;
4. avoid renames or model extraction until there is evidence of real coupling.
This keeps multi-frontend preparation focused on preventing regressions rather than inventing abstractions prematurely.
### Guardrails for common backend neutrality
A useful backend neutrality guardrail has two layers:
1. source/package guard: common backend source must not import language frontend packages;
2. public-contract guard: common handoff types must not expose language-owned types through fields, constructors, methods, or record components.
The package guard catches implementation coupling. The reflection guard catches API coupling.
### Direct handoff construction test
A backend test should be able to construct executable `IRBackend` manually using only common model types and lower it through `LowerToIRVMService`.
That test proves the backend consumes the contract, not a specific frontend pipeline.
## Pitfalls
- Do not treat the presence of PBS examples as proof that the backend is PBS-owned.
- Do not move PBS source rules into compiler-general specs.
- Do not let lifecycle terms such as published wrapper, boot guard, or frame root hide whether the obligation is common or PBS-specific.
- Do not rename `IRBackend` for aesthetics when the real problem is missing ownership language and missing guardrails.
- Do not use a future synthetic test frontend as a substitute for simple direct `IRBackend` construction tests.
## Takeaways
- `IRBackend` is the common executable handoff.
- PBS owns PBS-to-`IRBackend` lowering; compiler-general owns `IRBackend -> IRVM`.
- Common backend code must not depend on `p.studio.compiler.pbs`.
- Public `IRBackend` contract types must not expose PBS AST, token, parser, semantic, or editorial types.
- Multi-frontend preparation should first lock existing good boundaries before introducing new abstractions.
## References
- `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`
- `docs/specs/compiler/22. Backend Spec-to-Test Conformance Matrix.md`
- `docs/specs/compiler-languages/pbs/13. Lowering IRBackend Specification.md`
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/backend/irvm/LowerToIRVMServiceTest.java`
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/specs/CommonBackendArchitectureTest.java`
- `prometeu-compiler/prometeu-frontend-api/src/test/java/p/studio/compiler/models/IRBackendExecutableContractTest.java`
@@ -0,0 +1,82 @@
---
id: LSN-0060
ticket: multi-frontend-serializable-ir
title: Data contract first for serializable IRBackend
created: 2026-07-15
tags: [compiler, compiler-general, ir, backend, serialization, multi-frontend]
---
# Data contract first for serializable IRBackend
## Original Problem
The compiler already had a common executable handoff: `IRBackend`. PBS emits it, and common backend stages consume it before lowering to `IRVM`.
The remaining risk was subtler than direct PBS coupling. A Java-only handoff can accidentally grow around process-local assumptions: callbacks, services, object identity, mutable global state, unordered collections, lazy values, or direct references between live objects. Those shapes can work inside one JVM and still make a future external frontend or out-of-process boundary expensive to add.
The work needed to keep the handoff serializable by design without choosing JSON, Protobuf, RPC, plugin loading, or a separate process too early.
## Consolidated Decision
`IRBackend` is a data contract first.
Public `IRBackend` contract types must remain modelable as an acyclic, deterministic data graph made from primitives, strings, enums, immutable values, ordered lists, explicit ids, stable symbolic keys, and source attribution values such as file ids and spans.
The public handoff must not expose callbacks, lambdas, service objects, registries, visitors, live compiler services, frontend AST/parser/token/semantic/editorial objects, mutable global state, object identity semantics, unordered output-affecting collections, cyclic public references, process-dependent lazy values, or a selected wire format.
The important boundary is the data shape, not a codec. A future serializer should be able to consume the handoff because the public model is already disciplined, but this decision does not require a serializer to exist now.
## Final Implementation
The implementation updated the compiler-general backend spec with a new `IRBackend` serializability-by-design subsection. The spec now states the allowed and forbidden public shapes and keeps wire format selection explicitly deferred.
The conformance matrix added `G20-4.2.x` rows for the new requirements:
- public `IRBackend` modelability as an acyclic deterministic data graph;
- rejection of callbacks, services, mutable or unordered collections, frontend-owned objects, process-dependent lazy values, cyclic public references, and wire-format commitments;
- use of explicit ids or stable symbolic keys for cross-object references.
The executable guardrails live in `IRBackendExecutableContractTest`:
- the public contract audit covers all current `p.studio.compiler.models` handoff types and nested value types;
- support types such as `FileId`, `ModuleId`, `CallableId`, `IntrinsicId`, `Span`, table reference records, and `ReadOnlyList` are explicitly classified;
- reflection checks inspect public fields, record components, public constructors, public methods, and generic type arguments;
- guardrails reject PBS/frontend exposure, callback shapes, service-like public shapes, mutable/unordered collection contracts, and other non-serializable public exposures;
- deterministic aggregation tests prove stable ordering for functions, synthetic functions, globals, executable functions, module pools, callable signatures, intrinsic pools, reserved metadata, and required capabilities;
- known references are locked to explicit ids or stable keys rather than direct object references.
The audit found no concrete public serialization leak that required changing the production model. That result is meaningful: the current `IRBackend` shape was already close to the target, so the correct implementation was spec clarification plus tests, not a redesign.
## Examples
Good public handoff shapes:
- `ModuleId`, `CallableId`, `IntrinsicId`, and `FileId` for table-scoped references;
- `ReadOnlyList<T>` for ordered collections;
- `ModuleReference`, `CallableSignatureRef`, and `IntrinsicReference` as stable table values;
- `Span` for source attribution;
- enums and records for fixed value surfaces.
Bad public handoff shapes:
- `Map`, `Set`, or mutable collection interfaces where output order matters;
- `Runnable`, `Callable`, `Function`, visitors, or callbacks;
- compiler services or registries exposed as public fields or constructor parameters;
- frontend AST, parser, token, semantic, or editor types;
- direct references to another handoff object where an id or ordered table entry is the stable contract.
## Pitfalls
- Do not confuse "serializable by design" with "serialize it now." A codec is a future implementation detail.
- Do not add typed ids everywhere by taste. Add them when a real public field would otherwise depend on textual identity, object identity, or direct object references.
- Do not treat `ReadOnlyList` as cosmetic. Ordered collections are part of the deterministic contract.
- Do not let a private assembler or aggregator implementation shape leak into public handoff semantics.
- Do not move the rule into PBS specs. PBS owns how it populates the handoff; compiler-general owns the common handoff contract.
## References
- `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`
- `docs/specs/compiler/22. Backend Spec-to-Test Conformance Matrix.md`
- `prometeu-compiler/prometeu-frontend-api/src/test/java/p/studio/compiler/models/IRBackendExecutableContractTest.java`
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/specs/BackendConformanceMatrixSpecTest.java`
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/backend/irvm/LowerToIRVMServiceTest.java`
@@ -0,0 +1,143 @@
---
id: LSN-0062
ticket: multi-frontend-common-lifecycle
title: Common lifecycle assembly after frontend declaration
created: 2026-09-19
tags: [compiler, compiler-general, compiler-pbs, lifecycle, backend, multi-frontend]
---
# Common lifecycle assembly after frontend declaration
## Original Problem
`IRBackend` was already the common executable handoff, but the PBS frontend still
owned two different jobs. It admitted language markers `[Init]`, `[Frame]`, and
`InitAllowed`, and it also assembled the platform lifecycle: module init,
project-init wrapper, published frame wrapper, physical entrypoint, and boot
guard.
That second job discovered roles through `__pbs.*` names and lived inside
`PBSFrontendPhaseService`. Another frontend could not reuse the boot sequence
without copying PBS. Common backend code also risked treating PBS naming as the
integration contract.
The old path had a concrete defect. The `[Init]` colocated with `[Frame]` is
project init, but it was also included in the file-init fragment. The published
wrapper then called module init and project init separately, so the same
callable could run twice. Specs required project init exactly once, after all
module initializers and before the frame root.
## Consolidated Decision
Declaration and assembly are separate.
An executable frontend emits an immutable, ordered, typed
`IRLifecycleDeclaration` on `IRBackend`. The declaration carries:
- modules in explicit dependency order;
- each module's ordered file-init fragment references;
- an optional project-init callable;
- exactly one frame-root callable.
References use `ModuleId`, `FileId`, `CallableId`, a typed role, and source
attribution. Absence of a declaration means a deliberately non-executable
handoff. It must not synthesize a default `main`.
`FrontendPhasePipelineStage` stops at that declarative backend.
`AssembleLifecyclePipelineStage` runs immediately after it and is the terminal
stage of `analyze()`. `LifecycleAssemblerService` is the only owner of derived
lifecycle artifacts. `compile()` and `build()` reuse that single assembly pass.
`ResolveDepsPipelineStage` remains the ordering authority. Frontends propagate
that order; the assembler validates and applies it. Neither side may rebuild
order from synthetic names, `FileId`, textual module names, or incidental map
order.
Project init is excluded from file-init fragments and runs exactly once. Common
derived identities are neutral and deterministic. Common code must not import
PBS packages or inspect `__pbs.*` prefixes. Typed roles, ids, and origins are
the stable inspection surface; synthetic names are internal.
## Final Implementation
PBS now emits per-file lifecycle contributions and a project-level declaration.
A colocated `[Init]` is the project-init reference only. File-init fragments
remain frontend-lowered executables, consumed by id and role rather than by
name prefix. `synthesizePublishedLifecycle` was removed.
The common assembler validates the declaration and materializes:
1. module init callables in declared module/file order;
2. an optional project-init wrapper, invoked once;
3. the published frame wrapper, which is the physical entrypoint;
4. a hidden boot guard that prevents a second bootstrap after success.
The declaration remains on the assembled `IRBackend` as provenance. IRVM
lowering keeps only assembled-graph preconditions such as wrapper identity,
synthetic origin, and boot-guard presence. It does not reconstruct lifecycle
from names.
`analyze()` still returns a fully assembled executable backend, without IRVM,
bytecode, or file output. Neutral names use the `__prometeu.lifecycle.*` prefix.
Diagnostics stay split: PBS reports marker, signature, placement, cardinality,
and colocation errors; the common layer reports unresolved ids, incompatible
roles, missing frame root, invalid origins, and frontend-preassembled derived
artifacts.
## Examples
Good frontend output:
- a file with globals or a non-project `[Init]` contributes a
`FILE_INIT_FRAGMENT`;
- the `[Init]` next to `[Frame]` is `PROJECT_INIT` only;
- `[Frame]` is `FRAME_ROOT`;
- module order follows the resolved dependency graph even when names or file
ids would sort the other way.
Good assembled output:
- wrapper call order is module inits, then project init once, then frame root;
- boot guard is a hidden global, not a user global;
- `getLifecycleDeclaration()` still exposes the original typed references.
Bad output:
- PBS emitting `MODULE_INIT`, `PROJECT_INIT` wrappers, published wrappers, or
boot guards;
- common code matching `__pbs.frame_wrapper$m0` to discover the entrypoint;
- sorting modules by `FileId` or textual path because the declaration order was
inconvenient;
- leaving both the PBS assembler and the common stage active.
## Pitfalls
- Do not treat `analyze()` as "frontend only." It now includes common lifecycle
assembly so callers still receive an executable `IRBackend`.
- Do not keep a temporary dual path. Two assemblers duplicate wrappers or init.
- Do not infer missing frame roots or repair invalid declarations in the common
layer. Fail deterministically.
- Do not move PBS marker rules into the assembler. Language admission stays in
PBS; structural handoff checks stay common.
- Do not preserve `__pbs.*` names for common-derived artifacts as compatibility
behavior. They were never a public ABI.
- Do not confuse a frontend file-init fragment name with a common-derived
identity. PBS may still name its own fragments; common code must not parse
those names.
## References
- Decision: `DEC-0046`
- Plan: `PLN-0125`
- Related lessons: `LSN-0059`, `LSN-0060`
- `docs/specs/compiler/23. Compiler Pipeline Entry Points Specification.md`
- `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`
- `docs/specs/compiler/22. Backend Spec-to-Test Conformance Matrix.md`
- `docs/specs/compiler-languages/pbs/4. Static Semantics Specification.md`
- `docs/specs/compiler-languages/pbs/9. Dynamic Semantics Specification.md`
- `docs/specs/compiler-languages/pbs/12. Diagnostics Specification.md`
- `docs/specs/compiler-languages/pbs/13. Lowering IRBackend Specification.md`
- `prometeu-compiler/prometeu-frontend-api/src/main/java/p/studio/compiler/models/IRLifecycleDeclaration.java`
- `prometeu-compiler/prometeu-build-pipeline/src/main/java/p/studio/compiler/lifecycle/LifecycleAssemblerService.java`
- `prometeu-compiler/prometeu-build-pipeline/src/main/java/p/studio/compiler/workspaces/stages/AssembleLifecyclePipelineStage.java`
@@ -0,0 +1,129 @@
---
id: LSN-0063
ticket: multi-frontend-validation-boundaries
title: Language vs platform validation ownership
created: 2026-09-19
tags: [compiler, compiler-general, compiler-pbs, backend, validation, multi-frontend]
---
# Language vs platform validation ownership
## Original Problem
PBS mixed two kinds of rejection. Some failures only make sense with the
language: syntax, types, markers, barrel files, imports. Others belong to any
executable frontend: malformed `IRBackend`, invalid lifecycle declarations,
missing assembled wrappers, host-capability identity, bytecode/preload.
Without an ownership rule, the next frontend would copy PBS validators, or the
common layer would start knowing `[Init]`, barrel, and `E_SEM_*`. Spec 19 was
still titled as PBS verification, and the PBS diagnostics catalog listed
wrapper and boot-guard failures as PBS coverage.
Lifecycle assembly had already moved (`LSN-0062`). That proved the pattern. It
did not inventory the rest.
## Consolidated Decision
A validation lives at the lowest layer that understands the failure without
knowing the source language.
- Language: needs AST, tokens, a language marker, a barrel, or another source
rule.
- Platform: can be checked on `IRBackend`, `IRLifecycleDeclaration`, IRVM, PBX,
host/capability tables, or ids.
The discussion published that matrix. It did not extract a generic
validator-core and did not move validators whose common contract is not closed.
Host admission is split. PBS owns the source shape of `[Host]` and
`[Capability]`. Platform owns canonical `(module, name, version)` identity and
capability consistency once those facts exist on `IRReservedMetadata`. The
current validator may stay in the PBS package until a later plan. `E_HOST_*`
is not eternal PBS semantics.
The same fact may exist on both sides with distinct identities. PBS can report
`E_SEM_MISSING_PROJECT_FRAME`; the common layer still rejects an executable
declaration with no frame root.
Platform codes are `COMMON_*`, `LOWER_IRVM_*`, or `MARSHAL_*`. PBS keeps
`E_PARSE_*`, `E_SEM_*`, `E_LINK_*`, and `E_HOST_*`. Copy origin/span when
present; a missing origin is a structural error, not a reason to invent a span.
Linking and barrel stay language-owned. `Addressable` is platform-owned and was
classified only. SDK canonical source and the broad architectural harness stay
on adjacent discussions.
## Final Implementation
The work was editorial and test-facing.
Spec 19 is now `Compiler Verification and Safety Checks Specification`. Gates
S-U and S-I are unchanged. Section 11 states the classification rule, the
English ownership matrix, diagnostic identity, dual reporting, and the
host-admission split. It explicitly refuses validator-core extraction and
package moves.
PBS diagnostics spec 12 keeps marker, syntax, semantics, linking, and
host-attribute coverage. Common structural failures now point at spec 19
section 11 and spec 20 instead of appearing as PBS required coverage.
Spec 20 cross-references that same compiler-general ownership. The conformance
matrix added `G19-11.1` through `G19-11.6` and `PBS12-11`.
`CommonBackendArchitectureTest` now guards `backend/`, `lifecycle/`, and all
common pipeline stages. `ValidationOwnershipSpecTest` locks the title, matrix,
and PBS recut.
`PbsHostAdmissionValidator` remains under `p.studio.compiler.pbs`. No linking
or asset production code moved.
## Examples
Language-owned:
- parse errors;
- `[Init]` / `[Frame]` signature and colocation;
- barrel and import failures;
- `host call not allowed during init`.
Platform-owned and already implemented:
- invalid `IRLifecycleDeclaration`;
- assembled wrapper, boot guard, and synthetic origin preconditions;
- IRVM and marshal rejection families.
Classified as platform, not moved:
- canonical host-binding identity on `IRReservedMetadata`;
- `Addressable` / asset path.
Dual reporting:
- missing `[Frame]` in PBS source is `E_SEM_MISSING_PROJECT_FRAME`;
- an executable declaration without frame root is a common structural error.
## Pitfalls
- Do not treat “classified as platform” as “move the class now.” Host admission
still waits on the SDK source discussion.
- Do not merge language and platform failures into one hybrid code. Two
identities are cheaper than a lost message.
- Do not put PBS governance or PBS diagnostics in charge of compiler-general
safety gates.
- Do not scan PBS frontend sources with the platform architecture guard. The
guard is for common validators, not for the language package that must remain.
- Do not invent a validator-core because several common checks exist. They
already live on their stages.
## References
- Decision: `DEC-0047`
- Plan: `PLN-0126`
- Related lesson: `LSN-0062` - Common lifecycle assembly after frontend declaration
- Related lesson: `LSN-0059` - Common IRBackend handoff and backend guardrails
- `docs/specs/compiler/19. Verification and Safety Checks Specification.md`
- `docs/specs/compiler-languages/pbs/12. Diagnostics Specification.md`
- `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`
- `docs/specs/compiler/22. Backend Spec-to-Test Conformance Matrix.md`
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/specs/ValidationOwnershipSpecTest.java`
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/specs/CommonBackendArchitectureTest.java`
@@ -0,0 +1,136 @@
---
id: LSN-0061
ticket: multi-frontend-pvm-neutrality
title: PBX is the language-neutral runtime boundary
created: 2026-09-18
tags: [vm-arch, runtime, pvm, pbx, compiler, multi-frontend]
---
# PBX is the language-neutral runtime boundary
## Original Problem
PBS was the first implemented frontend, so its name still appeared in a few
common compiler specification titles. That editorial residue made it easier to
misread the architecture as if PBS owned bytecode, IRVM, or the PVM execution
model.
The implementation was already mostly neutral: the compiler emitted PBX, the
runtime loaded PBX, and the common backend did not import PBS frontend packages.
The real risk was future drift. In particular, IRVM could be mistaken for a
runtime input, `PBS\0` could be confused with the executable magic, and duplicated
VM-architecture documents could silently become competing authorities.
## Consolidated Decision
The executable pipeline is:
```text
language frontend -> IRBackend -> IRVM -> PBX -> loader/verifier -> PVM
```
`IRBackend`, `IRVM`, optimization, and PBX emission are compiler-owned.
`IRVM` is an internal compiler representation and never crosses the runtime
boundary. PBX is the only executable artifact consumed by the runtime.
The runtime repository owns PBX binary layout, Core ISA, loader behavior,
verification, and PVM execution. Studio owns the common compiler transformations
that produce conforming PBX. Studio copies of VM-architecture documents are
mirrors, not independent authorities.
`PBS` names a language, frontend, stdlib, and explicitly language-specific
integration. It does not name the executable format, magic, ISA, ABI, program
image, or runtime type. The canonical persisted artifact is `program.pbx`, and
its binary magic is `PBX\0`.
## Final Implementation
The common compiler specifications now identify bytecode/PBX and IRVM
optimization as language-neutral surfaces. They explicitly state that IRVM is
internal to the compiler and that PBX is the compiler/runtime boundary. Existing
PBS references were retained where they describe legitimate frontend-specific
publication or metadata behavior.
Studio's copies of `ARCHITECTURE.md`, `ISA_CORE.md`, and `INTRINSICS.csv` were
synchronized with the runtime repository. A mirror policy now identifies the
runtime repository as canonical and provides byte-for-byte comparison commands.
Existing Studio tests already proved two important invariants:
- `BytecodeModule` serializes the exact `PBX\0` header;
- common backend contracts neither import nor publicly expose PBS frontend
types.
Runtime conformance was strengthened in three focused places:
- the bytecode invalid-magic regression now explicitly proves that legacy
`PBS\0` is rejected;
- the PVM initialization regression now creates a `BytecodeModule` directly,
serializes it, loads and verifies it, prepares the boot call, and executes it
through `HALT` without any frontend;
- a quality check inspects the `prometeu-bytecode` and `prometeu-vm` manifests
and rejects dependencies whose identities belong to compiler or frontend
layers.
No production VM behavior, ISA, ABI, or binary format changed.
## Examples
Language-specific work stops before the common backend:
```text
PBS AST and semantic model
-> PBS lowering
-> neutral IRBackend data
-> common IRVM lowering and optimization
-> PBX\0
```
A future frontend follows the same boundary:
```text
FutureLanguage AST and semantic model
-> FutureLanguage lowering
-> neutral IRBackend data
-> the same common IRVM and PBX stages
```
The runtime sees the same PBX contract in both cases. It does not receive a
`languageId`, source extension, AST object, IRVM object, or frontend service.
## Pitfalls and Anti-Patterns
- Do not expose IRVM as a convenient runtime protocol. It is optimized for
compiler transformations, not binary compatibility or runtime loading.
- Do not rename PBX concepts after the first frontend. `PBS\0` is invalid even
when PBS produced the program.
- Do not globally ban the word `PBS` from runtime documentation. PBS is a valid
guest-language example; guards should inspect dependencies, public types, and
format identities instead of raw text.
- Do not duplicate tests when an existing guard proves the entire invariant.
Extend coverage only for the missing distinction, such as explicit `PBS\0`
rejection.
- Do not let a documentation mirror become a second authority. Synchronize it
against runtime or replace it with an explicit reference.
- Loading a PBX image is not the same as beginning execution. The PVM boot test
must prepare the published boot call before running the instruction budget.
## References
- `docs/specs/compiler/15. Bytecode and PBX Mapping Specification.md`
- `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`
- `docs/specs/compiler/21. IRVM Optimization Pipeline Specification.md`
- `docs/vm-arch/README.md`
- `prometeu-compiler/prometeu-build-pipeline/src/main/java/p/studio/compiler/backend/bytecode/BytecodeModule.java`
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/p/studio/compiler/specs/CommonBackendArchitectureTest.java`
- Runtime: `crates/console/prometeu-bytecode/src/model.rs`
- Runtime: `crates/console/prometeu-vm/src/virtual_machine.rs`
- Runtime: `crates/dev/prometeu-quality-checks/tests/pvm_frontend_neutrality.rs`
## Takeaways
- PBX is the stable, language-neutral executable boundary.
- IRVM stays private to the compiler pipeline.
- Runtime authority and compiler authority are complementary, not competing.
- Small dependency and artifact tests are enough to protect this boundary
without introducing a plugin platform or a second language.
@@ -1,187 +0,0 @@
---
id: AGD-0039
ticket: pbs-symbol-documentation-and-hover-markdown
title: Modelo de documentacao de simbolos em PBS e consumo markdown no hover
status: accepted
created: 2026-05-08
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Pain
A wave atual de assistencia editorial quer `hover` util, mas ainda nao existe um modelo claro para documentacao authored dentro de PBS.
Sem esse modelo, o hover pode até mostrar:
- kind;
- assinatura;
- origem;
- tipo/shape.
Mas continua sem uma superficie limpa para explicar intencao de uso, contrato semantico, detalhes de parametros ou observacoes de API.
Hoje a direcao consolidada eh algo como:
- `[Doc(markdown = """qualquer markdown aqui dentro...""")]`
como anotacao consumida diretamente pelo hover.
Isso parece simples no uso, mas toca em decisoes de linguagem e contrato editorial que merecem discussao propria.
## Context
Domain owner: `compiler/pbs`
Subdomains touched:
- `compiler/general`
- `studio/lsp`
- `tools/vscode-extension`
Esta agenda eh deliberadamente separada de `AGD-0038`.
Razao:
- `AGD-0038` trata da wave 1 de `completion`, `hover` e `signature help`;
- esta nova agenda trata da autoria, armazenamento, propagacao e consumo de documentacao de simbolos.
O objetivo eh evitar que a primeira wave de assistencia editorial fique bloqueada por uma decisao de linguagem ainda nao fechada.
## Open Questions
- A autoria da documentacao deve acontecer por anotacao explicita como `[Doc(...)]`, por comentario estruturado, ou por outro mecanismo?
- O payload authored deve aceitar um subset Markdown como fonte canonica, ou deve haver outra normalizacao/restricao?
- Quais declaracoes nomeadas de API podem carregar documentacao na wave inicial:
funcoes, methods, classes/structs, services, hosts, builtin/stdlib declarations e interface declarations?
- A documentacao deve viver no AST/semantics como superficie compiler-backed explicita?
- Como a documentacao authored em PBS se relaciona com documentacao de stdlib gerada ou importada de outras fontes?
- O hover deve consumir markdown pronto sem transformacao, ou deve haver um pipeline intermediario de sanitizacao/normalizacao?
- Como tratar multilinhas, escaping e ergonomia de escrita se a surface for `[Doc(...)]`?
- A agenda deve definir apenas o modelo de autoria, ou tambem o transporte LSP e a composicao final do hover?
## Options
### Option A - Anotacao `[Doc(...)]` com Markdown subset como payload canonico
- **Approach:** introduzir uma anotacao explicita na linguagem PBS cujo conteudo ja eh markdown pronto para o hover.
- **Pro:** autoria direta, intencao clara, pipeline simples de consumo no LSP e no hover.
- **Con:** precisa desenhar sintaxe, escaping, multiline e regras de onde a anotacao eh permitida.
- **Maintainability:** forte se a surface de linguagem ficar limpa.
### Option B - Comentarios estruturados como fonte de documentacao
- **Approach:** usar comentarios especiais ou doc-comments, com parser especifico para gerar documentacao de simbolo.
- **Pro:** mais familiar para quem vem de outras linguagens e menos intrusivo no corpo semantico da linguagem.
- **Con:** exige novo contrato de binding entre comentarios e simbolos, e pode aumentar ambiguidade editorial.
- **Maintainability:** media.
### Option C - Nao definir autoria em PBS agora e suportar apenas docs de stdlib por metadata externa
- **Approach:** deixar a linguagem sem surface authored por enquanto e alimentar hover apenas com docs externas/geradas.
- **Pro:** menor custo imediato de linguagem.
- **Con:** resolve mal a autoria de codigo de usuario e deixa o modelo incompleto.
- **Maintainability:** fraca.
## Tradeoffs
Option C nao parece suficiente como direcao de produto se a intencao eh ter hover realmente util para APIs authored em PBS.
Option B eh plausivel, mas move o problema para um binding entre comentario e simbolo que tambem precisa ser projetado com cuidado. Isso costuma parecer mais barato do que realmente eh.
Option A tem um custo claro de sintaxe e modelo, mas pelo menos deixa explicito que documentacao eh metadado semantico e nao apenas texto solto ao redor do codigo.
Se a equipe realmente quer markdown authored pronto para consumo no hover, Option A parece a discussao mais direta e honesta.
## Recommendation
Discutir esta agenda com viés inicial para **Option A** e fechar a decision com um shape canonico unico:
```pbs
[Doc(markdown = """
Draws a sprite.
- `x`: screen x
- `y`: screen y
""")]
```
O consenso da discussao eh que `Doc` eh curto e pratico, e que o nome do argumento `markdown` ja carrega a semantica do payload. Um prefixo lexical como `md"""..."""` nao deve entrar na v1 porque adiciona uma nova familia de literais antes de haver necessidade concreta para multiplos payloads especializados.
A recomendacao preliminar eh separar o problema em duas perguntas:
1. qual eh a surface canonical de autoria da documentacao em PBS?
2. como essa documentacao vira um payload editorial resolvido para hover e futuros consumers?
Mesmo que `[Doc(...)]` seja a direcao vencedora, a decision deve deixar claro:
- onde ela pode ser usada;
- qual formato textual aceita;
- como ela eh preservada no compiler;
- como chega ao LSP;
- e como o hover deve compor isso com assinatura e metadados do simbolo.
## Proposed Resolution
Fechar a agenda com a seguinte direcao:
- PBS deve suportar documentacao explicita de simbolos por atributo `[Doc(markdown = """...""")]`.
- `Doc` deve ter um unico shape canonico na v1: argumento nomeado `markdown` com payload em triple-quoted text block.
- O atributo segue a regra normal de atributos PBS: associa-se a declaracao imediatamente seguinte.
- `[Doc]` deve ser valido para declaracoes nomeadas de API que introduzem simbolo semantico nomeado e que possam carregar atributos, incluindo funcoes, methods, classes/structs, services, hosts, builtin/stdlib declarations e interface declarations.
- Parametros nao devem receber `[Doc]` na v1 em nenhuma situacao; a documentacao de parametros deve viver no Markdown da declaracao que os possui.
- A intencao normativa eh nao limitar artificialmente declaracoes documentaveis por origem; se algum ponto da gramatica de declaracao ainda nao aceita atributos, o plano deve tratar isso como trabalho sintatico explicito.
- A ordem entre `[Doc]` e outros atributos deve ser irrelevante.
- Mais de um `[Doc]` no mesmo simbolo deve ser erro semantico, nao erro de parse.
- `[Doc]` sem `markdown`, com argumentos extras, com payload vazio/whitespace-only, ou aplicado fora da superficie documentavel deve produzir diagnostico claro.
- O payload `"""..."""` deve ser tratado inicialmente como text block documental, nao como string runtime geral de PBS.
- O text block deve aplicar normalizacao de indentacao equivalente ao modelo de text blocks de Java/Javadoc: remover linhas externas vazias incidentais, calcular indentacao comum das linhas nao vazias, remover essa indentacao comum, preservar quebras internas e preservar o conteudo Markdown.
- O compiler deve preservar o Markdown normalizado como metadata semantica do simbolo.
- O LSP e demais consumers devem consumir documentacao resolvida pelo compiler; nao devem reparsear source para descobrir docs.
- Um subset Markdown deve ser o protocolo de documentacao authored em PBS.
- O Markdown deve ser definido como payload editorial preservado, nao como garantia de renderizacao identica em todos os clientes. Renderizacao, sanitizacao e suporte a constructs especificos sao responsabilidades do consumer.
- Formatters devem preservar o conteudo interno do text block e nao devem reflowar Markdown authored.
- Rename/refactor nao precisa atualizar referencias textuais dentro do Markdown nesta fase do projeto.
- Specs e documentacao da stdlib devem ser escritas em ingles.
- Testes devem validar o payload Markdown normalizado no compiler/LSP; validacao visual no VSCode pode ser manual.
- Stdlib, SDK e interface modules devem usar o mesmo modelo `[Doc(markdown = """...""")]` para documentacao authored.
- Se no futuro houver documentacao importada ou gerada, documentacao authored no source deve ter precedencia, e conflitos devem ter regra deterministica ou diagnostico explicito.
## Maintenance Friction Points
Pontos que a decision e o plano devem proteger para evitar que uma feature simples de editor vire divida de linguagem:
- "Todas as declaracoes nomeadas de API documentaveis" precisa ser especificado como "toda declaracao de API que introduz simbolo semantico nomeado", evitando ambiguidade sobre parametros, fields, enum cases, tuple fields, expressoes e variaveis locais.
- A gramatica de atributos pode precisar ser expandida para declaracoes que hoje ainda nao aceitam atributos; isso deve ser tratado como parte real do trabalho, nao como excecao informal.
- A v1 nao deve aceitar aliases como `text`, argumento posicional, `html`, `format`, `Document`, `Documentation`, ou `md"""..."""`.
- `"""..."""` nao deve virar literal runtime geral sem uma decision propria.
- A normalizacao do text block deve ser especificada no texto normativo, nao apenas referenciada informalmente como "igual Java".
- Parametros nao recebem `[Doc]` na v1; fields, enum cases e tuple fields so devem receber `[Doc]` se a decision os classificar explicitamente como declaracoes nomeadas de API documentaveis.
- O formatter futuro eh um ponto sensivel: ele pode alinhar a declaracao, mas nao pode alterar o Markdown authored dentro do text block.
- A composicao visual do hover pode ficar para plano/implementacao, mas a ownership deve ficar clara: compiler resolve documentacao; LSP transporta; editor renderiza.
## Discussion
Esta agenda nao deve ser reduzida a "mostrar markdown no hover".
O problema real eh definir ownership e boundary da documentacao:
- o author escreve onde?
- o compiler preserva como?
- o LSP transporta em qual shape?
- o hover monta o resultado com qual ordem e responsabilidade?
Tambem vale observar que a escolha aqui pode influenciar mais do que hover:
- completions futuras podem querer `detail` e `documentation`;
- signature help pode querer descricoes de parametros;
- docs de stdlib podem querer surface unificada com docs authored.
Por isso, mesmo sendo uma agenda separada, ela conversa diretamente com a futura qualidade do editor.
## Resolution
Consenso de agenda formado em favor de `[Doc(markdown = """...""")]` como anotacao explicita, compiler-owned, com payload em subset Markdown e text block documental normalizado.
## Next Step
Aceitar a agenda e fechar uma `decision` propria para o modelo de documentacao de simbolos em PBS.
@@ -0,0 +1,62 @@
---
id: AGD-0042
ticket: pbs-lsp-go-to-definition
title: PBS LSP Go to Definition
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, definition]
---
## Pain
Domain owner: `studio/lsp`
PBS users cannot jump from a symbol use to its declaration, which makes navigation across source files, stdlib imports, services, and generated/editorial surfaces slow and error-prone.
## Context
The current LSP exposes diagnostics, hover, completion, signature help, and semantic tokens. The compiler already builds semantic read surfaces and editorial symbol resolution for hover/completion, but there is no LSP definition capability or stable location mapping exposed to VS Code.
## Open Questions
- [ ] Which symbol categories must be supported in the first definition wave?
- [ ] Should stdlib and generated/supplemental declarations navigate to source files, virtual documents, or remain hover-only?
- [ ] What identity model should connect an editor token to the canonical declaration without relying on text search?
## Options
### Option A - Syntactic location lookup
- **Approach:** Use parser spans and local AST traversal to map the token under cursor to a declaration with matching text.
- **Pro:** Fast to implement and useful for same-file declarations.
- **Con:** Breaks on overloads, imports, stdlib, aliases, and same-name symbols.
- **Maintainability:** Weak; it creates a second navigation model separate from semantic resolution.
### Option B - Semantic identity lookup
- **Approach:** Extend the compiler/editorial surface with resolved declaration locations keyed by symbol identity, then expose `textDocument/definition`.
- **Pro:** Correct for imports, overloads, methods, stdlib-backed declarations, and future references/rename.
- **Con:** Requires explicit location metadata in semantic/editorial surfaces.
- **Maintainability:** Strong; it becomes shared infrastructure for references, rename, hierarchy, and code actions.
## Tradeoffs
The syntactic option is tempting for speed, but it would likely be thrown away once references and rename arrive. The main tradeoff is whether stdlib/supplemental declarations must navigate to physical files immediately or can initially return no location while still supporting project symbols.
## Recommendation
Prefer Option B. Build a reusable semantic location index and implement the first wave for project source declarations, then define a policy for stdlib/supplemental targets.
## Discussion
This should probably be the first navigation agenda to resolve because references, rename, workspace symbols, and hierarchy depend on the same symbol identity/location model.
## Resolution
Ainda em aberto.
## Next Step
Decide the first-wave symbol categories and whether stdlib/supplemental declarations require virtual document support.
@@ -0,0 +1,62 @@
---
id: AGD-0043
ticket: pbs-lsp-find-references
title: PBS LSP Find References
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, references]
---
## Pain
Domain owner: `studio/lsp`
PBS users cannot ask where a symbol is used, so changes to functions, services, structs, constants, and stdlib-facing APIs require manual search that confuses overloads and unrelated text matches.
## Context
Completion and hover already resolve symbols in local editor context, but the LSP does not expose references and the compiler/editorial layer does not yet provide project-wide usage locations keyed by semantic identity.
## Open Questions
- [ ] Which declarations need reference tracking first: local functions, methods, types, constants, imports, stdlib symbols, or all of them?
- [ ] How should overloads and same-name symbols in different scopes be disambiguated?
- [ ] Should references include declarations, write/read classification, or only usage sites in the first wave?
## Options
### Option A - Text search with filters
- **Approach:** Search workspace text for the selected identifier and filter obvious false positives by token kind.
- **Pro:** Simple and fast for early demos.
- **Con:** Incorrect for overloads, scope, shadowing, fields, methods, imports, and generated/stdlib surfaces.
- **Maintainability:** Poor; later semantic references would duplicate and replace it.
### Option B - Semantic usage index
- **Approach:** During semantic analysis, emit definition and usage sites tied to stable symbol identities and expose `textDocument/references`.
- **Pro:** Correct foundation for rename, call hierarchy, diagnostics UX, and code actions.
- **Con:** Requires compiler-facing usage modeling and careful invalidation.
- **Maintainability:** Strong; one source of truth for all symbol usage features.
## Tradeoffs
Reference accuracy matters more than early breadth. False positives are worse than missing unsupported categories because users will trust rename and references for safe edits.
## Recommendation
Prefer Option B, but scope the first wave to project-owned declarations whose identity is already resolved reliably.
## Discussion
References should follow go-to-definition because both need the same identity and location model.
## Resolution
Ainda em aberto.
## Next Step
Define first-wave symbol categories and whether declarations are included in reference results.
@@ -0,0 +1,62 @@
---
id: AGD-0044
ticket: pbs-lsp-document-symbols-outline
title: PBS LSP Document Symbols and Outline
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, outline, document-symbols]
---
## Pain
Domain owner: `studio/lsp`
PBS files do not populate a reliable VS Code outline, forcing users to navigate long source files manually instead of scanning declarations and members structurally.
## Context
The parser and semantic surfaces know top-level declarations and many member declarations. The LSP currently does not announce documentSymbolProvider and does not map PBS declarations to DocumentSymbol or SymbolInformation payloads.
## Open Questions
- [ ] Should the first outline be syntactic, semantic, or hybrid?
- [ ] Which hierarchy should be shown for structs, services, hosts, contracts, builtin types, enums, and methods?
- [ ] How should invalid or partially parsed files contribute to outline during editing?
## Options
### Option A - AST-only outline
- **Approach:** Map parsed declarations and member spans directly to LSP `DocumentSymbol`.
- **Pro:** Works even when semantic analysis is incomplete and gives immediate editor value.
- **Con:** Cannot classify every symbol with semantic precision.
- **Maintainability:** Good if kept as structural outline, not overloaded with semantic behavior.
### Option B - Semantic outline
- **Approach:** Build outline from semantic symbols after analysis.
- **Pro:** More accurate kinds and can hide invalid/unresolved surfaces.
- **Con:** More fragile during active editing and depends on full semantic success.
- **Maintainability:** Good for final precision, but heavier than needed for outline.
## Tradeoffs
Outline should remain available during broken intermediate edits. Semantic precision is useful, but not at the cost of disappearing structure while typing.
## Recommendation
Prefer Option A with semantic enrichment where available. Use AST spans as the stable backbone.
## Discussion
Document symbols are relatively independent and can be implemented before definition/references.
## Resolution
Ainda em aberto.
## Next Step
Decide hierarchy and symbol kinds for PBS declarations and members.
@@ -0,0 +1,62 @@
---
id: AGD-0045
ticket: pbs-lsp-workspace-symbols
title: PBS LSP Workspace Symbols
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, workspace-symbols]
---
## Pain
Domain owner: `studio/lsp`
VS Code cannot search PBS symbols across the workspace, so users cannot quickly jump to functions, services, structs, constants, or stdlib-exposed APIs by name.
## Context
Document-level editorial features exist, but the LSP does not announce workspaceSymbolProvider and the compiler pipeline does not expose a project-wide symbol index designed for editor queries.
## Open Questions
- [ ] Should workspace symbols include only project files or also imported stdlib and SDK surfaces?
- [ ] What symbol kinds and ranking should be returned for duplicate names or overloads?
- [ ] Should the symbol index be cached from analysis snapshots or rebuilt per request in the first implementation?
## Options
### Option A - Project source index only
- **Approach:** Index project-owned declarations from analysis snapshots and expose `workspace/symbol`.
- **Pro:** Clear ownership and physical locations.
- **Con:** Does not help users discover stdlib/SDK symbols.
- **Maintainability:** Strong as a first wave.
### Option B - Project plus stdlib index
- **Approach:** Include project declarations and imported stdlib/SDK surfaces in the workspace symbol result set.
- **Pro:** Better discovery, especially for new users.
- **Con:** Requires a navigation policy for stdlib source or virtual targets.
- **Maintainability:** Good if stdlib target handling is solved; awkward otherwise.
## Tradeoffs
Workspace symbols are partly navigation and partly discovery. Including stdlib too early can create broken jumps if target locations are not defined.
## Recommendation
Prefer Option A first, then add stdlib once definition has a target policy.
## Discussion
This should follow document symbols and the semantic location/index work.
## Resolution
Ainda em aberto.
## Next Step
Decide whether first-wave workspace search includes stdlib or project-only symbols.
@@ -0,0 +1,62 @@
---
id: AGD-0046
ticket: pbs-lsp-rename-symbol
title: PBS LSP Rename Symbol
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, rename]
---
## Pain
Domain owner: `studio/lsp`
PBS users cannot safely rename a symbol through the editor, so they must edit text manually and risk missing references or changing unrelated overloads and same-name declarations.
## Context
Rename requires the same semantic identity and reference machinery needed by definition/references, plus workspace edits and conflict detection. The current LSP does not announce renameProvider.
## Open Questions
- [ ] Which symbols should be renameable in the first wave, and which should be protected such as stdlib, generated, host, or builtin declarations?
- [ ] How should rename handle overloads, member methods, imports, and labels with the same text?
- [ ] What validation should run before returning a WorkspaceEdit?
## Options
### Option A - Local file rename
- **Approach:** Rename only references in the current file using local semantic resolution.
- **Pro:** Smaller blast radius and easier conflict handling.
- **Con:** Incomplete for imported symbols and project-wide APIs.
- **Maintainability:** Acceptable only if explicitly presented as limited.
### Option B - Workspace semantic rename
- **Approach:** Use a project-wide reference index and return a `WorkspaceEdit` for all usage and declaration sites.
- **Pro:** Matches user expectations for editor rename.
- **Con:** Requires reliable references, collision checks, and write-safety across files.
- **Maintainability:** Strong once references exist.
## Tradeoffs
Rename is high trust. A partial or text-based rename can corrupt code silently, so scope must be explicit and conservative.
## Recommendation
Prefer Option B after semantic references are implemented. If delivered earlier, ship only a clearly limited local rename.
## Discussion
Do not implement rename before definition/references are resolved.
## Resolution
Ainda em aberto.
## Next Step
Decide whether rename waits for workspace references or ships a limited local-only wave.
@@ -0,0 +1,62 @@
---
id: AGD-0047
ticket: pbs-lsp-code-actions
title: PBS LSP Code Actions and Quick Fixes
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, code-actions, quick-fix]
---
## Pain
Domain owner: `studio/lsp`
PBS diagnostics are visible, but users do not get actionable quick fixes for common compiler/editor problems, making simple repairs slower than necessary.
## Context
The LSP maps diagnostics but does not announce codeActionProvider. Compiler diagnostics already carry codes and ranges, but not every diagnostic has enough structured repair data to generate safe edits.
## Open Questions
- [ ] Which diagnostics should receive quick fixes first?
- [ ] Should code actions be produced from diagnostic codes only, or from structured compiler repair hints?
- [ ] How should imports, reserved attributes, missing methods, and invalid Doc shapes be prioritized?
## Options
### Option A - Diagnostic-code quick fixes
- **Approach:** Map known diagnostic codes to simple edits in the LSP layer.
- **Pro:** Fast for a small set of obvious repairs.
- **Con:** Couples fixes to diagnostic text/ranges and can become brittle.
- **Maintainability:** Moderate if limited to trivial edits.
### Option B - Compiler repair hints
- **Approach:** Extend diagnostics with structured repair metadata and let the LSP map hints to code actions.
- **Pro:** Keeps product logic close to compiler knowledge.
- **Con:** Requires a new diagnostic/repair contract.
- **Maintainability:** Strong for non-trivial fixes and future frontends.
## Tradeoffs
Quick fixes are only useful if safe. Simple code-based fixes can ship first, but imports and generated stubs need stronger compiler context.
## Recommendation
Prefer Option B as the target design, with Option A allowed only for a small bootstrap set of unambiguous repairs.
## Discussion
This should follow diagnostics UX and import assistance for the first high-value fixes.
## Resolution
Ainda em aberto.
## Next Step
Pick the first three diagnostic codes eligible for safe quick fixes.
@@ -0,0 +1,62 @@
---
id: AGD-0048
ticket: pbs-lsp-formatting
title: PBS LSP Formatting
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, formatting]
---
## Pain
Domain owner: `studio/lsp`
PBS formatting is not automated, so style drifts across files, especially around attributes, services, blocks, imports, and multi-line Doc text blocks.
## Context
The LSP does not announce documentFormattingProvider. PBS has a parser and AST, but formatting must preserve documentation text block content while normalizing surrounding syntax and indentation.
## Open Questions
- [ ] Should formatting be AST-based, token-based, or hybrid to preserve comments and text blocks?
- [ ] What canonical style should PBS use for attributes, braces, imports, params, and text blocks?
- [ ] Should the first wave support full document formatting only, or also range/on-type formatting?
## Options
### Option A - Token-preserving formatter
- **Approach:** Format from tokens and trivia, preserving comments and documentation text block contents.
- **Pro:** Safer for active editing and text blocks.
- **Con:** Harder to enforce deep structural layout.
- **Maintainability:** Good if PBS formatting remains mostly syntactic.
### Option B - AST pretty-printer
- **Approach:** Reprint code from the AST using canonical style rules.
- **Pro:** Produces consistent output.
- **Con:** Risks losing comments/trivia and cannot handle malformed files well.
- **Maintainability:** Strong only after AST/trivia ownership is explicit.
## Tradeoffs
Formatting must not damage `Doc` text blocks. That makes trivia preservation a first-class requirement rather than an implementation detail.
## Recommendation
Prefer Option A or a hybrid token/AST formatter. Avoid a pure AST pretty-printer until comments/trivia are modeled.
## Discussion
Formatter decision should define canonical style before implementation.
## Resolution
Ainda em aberto.
## Next Step
Decide PBS canonical formatting rules for attributes, braces, imports, params, and text blocks.
@@ -0,0 +1,62 @@
---
id: AGD-0049
ticket: pbs-lsp-import-assistance
title: PBS LSP Import Assistance
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, imports, completion]
---
## Pain
Domain owner: `studio/lsp`
PBS users do not get import-aware assistance, so discovering exported symbols and adding the right import from barrels, stdlib, or project modules remains manual.
## Context
Completion already exists for local/editorial contexts, and stdlib imports are understood by the compiler pipeline. The LSP does not yet provide auto-import completion, organize imports, or import quick fixes.
## Open Questions
- [ ] Should auto-import target project modules, stdlib modules, or both in the first wave?
- [ ] How should barrels and explicit module imports be ranked and rendered in completion?
- [ ] Should organize imports be a code action, source action, formatter responsibility, or separate feature?
## Options
### Option A - Import completion only
- **Approach:** Suggest symbols from known modules and include import detail, but do not edit imports automatically.
- **Pro:** Low risk and improves discovery.
- **Con:** Still leaves manual import edits.
- **Maintainability:** Good as a stepping stone.
### Option B - Auto-import and organize imports
- **Approach:** Completion/code actions insert missing imports and source actions sort/remove unused imports.
- **Pro:** Matches modern editor expectations.
- **Con:** Requires robust module export indexing and edit placement rules.
- **Maintainability:** Strong if built on a proper import graph.
## Tradeoffs
Auto-import depends on knowing canonical exports and choosing an edit location. Doing that ad hoc in completion would create formatting and duplicate import issues.
## Recommendation
Prefer Option B as the target, but start with import-aware suggestions plus a limited missing-import quick fix.
## Discussion
This should be coordinated with code actions and formatter.
## Resolution
Ainda em aberto.
## Next Step
Decide first-wave import sources: stdlib only, project modules only, or both.
@@ -0,0 +1,62 @@
---
id: AGD-0050
ticket: pbs-lsp-semantic-tokens-semantic-classification
title: PBS LSP Semantic Token Classification
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, semantic-tokens]
---
## Pain
Domain owner: `studio/lsp`
PBS semantic highlighting exists, but parts of the token classification can remain lexical or superficial, which limits visual accuracy for same-looking identifiers with different semantic roles.
## Context
The current LSP announces full semantic tokens and maps frontend semantic presentation to VS Code. The compiler has semantic read surfaces that can classify symbols more accurately than raw tokenization.
## Open Questions
- [ ] Which token categories must become semantic identity-based first?
- [ ] Should semantic tokens be produced from compiler semantic surfaces, parser AST, lexer tokens, or layered output?
- [ ] How should tokens behave in files with syntax or semantic errors during active editing?
## Options
### Option A - Lexical tokens with semantic overlays
- **Approach:** Keep lexer-based tokenization as baseline and overlay semantic categories where resolution succeeds.
- **Pro:** Stable during broken edits and incremental to improve.
- **Con:** Some unresolved identifiers remain visually generic.
- **Maintainability:** Strong because fallback behavior is explicit.
### Option B - Fully semantic token stream
- **Approach:** Generate all semantic tokens from compiler semantic surfaces.
- **Pro:** Maximum accuracy when analysis succeeds.
- **Con:** Highlighting may disappear or degrade sharply on incomplete code.
- **Maintainability:** Risky unless snapshot/error recovery is excellent.
## Tradeoffs
Highlighting must be resilient while the user is typing. A perfect semantic pass that fails on partial code is worse than a layered model.
## Recommendation
Prefer Option A: lexical baseline plus semantic overlays from resolved symbols.
## Discussion
This can proceed independently but benefits from LSP snapshot/cache work.
## Resolution
Ainda em aberto.
## Next Step
Choose semantic overlay categories for the first pass.
@@ -0,0 +1,62 @@
---
id: AGD-0051
ticket: pbs-lsp-performance-cache-and-snapshots
title: PBS LSP Performance Cache and Snapshots
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, performance, caching]
---
## Pain
Domain owner: `studio/lsp`
Several PBS LSP requests rebuild or reanalyze substantial project state, which can become slow or jittery as projects grow and as richer features are added.
## Context
The current bridge favors correctness and simplicity by invoking compiler pipeline analysis for diagnostics and editorial documents. A complete LSP needs snapshot caching, invalidation, cancellation, and predictable latency.
## Open Questions
- [ ] What is the canonical project/document snapshot boundary for LSP requests?
- [ ] Which requests can reuse semantic read surfaces safely across document changes?
- [ ] How should cancellation, stale responses, and background reanalysis be handled?
## Options
### Option A - Request-local analysis
- **Approach:** Keep each LSP request self-contained and rebuild analysis as needed.
- **Pro:** Simple, deterministic, and less stateful.
- **Con:** Latency grows with project size and feature richness.
- **Maintainability:** Good for correctness, weak for interactive performance.
### Option B - Project snapshot cache
- **Approach:** Maintain per-project analysis snapshots with document overlays, invalidation, and cancellation-aware request handling.
- **Pro:** Required for responsive definition, references, rename, and semantic tokens.
- **Con:** Introduces lifecycle complexity and stale-result risks.
- **Maintainability:** Strong if snapshot ownership and invalidation are explicit.
## Tradeoffs
A complete LSP needs shared state, but shared state must not leak stale analysis into edits. The decision is about the boundary, not just caching.
## Recommendation
Prefer Option B before implementing workspace-wide features like references and rename.
## Discussion
Define snapshot lifecycle, invalidation triggers, and cancellation policy.
## Resolution
Ainda em aberto.
## Next Step
Decide whether snapshot caching is a prerequisite for references/rename or can be staged later.
@@ -0,0 +1,62 @@
---
id: AGD-0052
ticket: pbs-lsp-completion-depth
title: PBS LSP Completion Depth
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, completion]
---
## Pain
Domain owner: `studio/lsp`
PBS completion works for core cases, but a complete daily-use experience needs ranking, resolve details, snippets, overload handling, context-aware suggestions, and richer documentation.
## Context
The current LSP announces completion with trigger character '.' and returns detail/documentation for resolved compiler candidates. It does not use resolveProvider, snippets, broad trigger contexts, or advanced ranking.
## Open Questions
- [ ] Which completion contexts should be supported first beyond member access?
- [ ] Should completion use resolveProvider for expensive documentation/details or return everything eagerly?
- [ ] How should overloads, snippets, imports, keywords, and stdlib candidates be ranked?
## Options
### Option A - Eager rich completion
- **Approach:** Return labels, details, docs, and snippets directly in every completion response.
- **Pro:** Simple client behavior and immediate docs.
- **Con:** Can become expensive for broad/global completion.
- **Maintainability:** Good while candidate sets remain small.
### Option B - Completion resolve provider
- **Approach:** Return lightweight candidates first and populate expensive docs/details through `completionItem/resolve`.
- **Pro:** Scales better for large candidate sets and import-aware suggestions.
- **Con:** Requires client/server resolve plumbing and stable candidate identities.
- **Maintainability:** Strong for a full LSP.
## Tradeoffs
The current eager model is fine for scoped completions, but global/import-aware completion may need resolve to avoid latency.
## Recommendation
Prefer a staged approach: keep eager for current contexts, design resolve before broad workspace/import completion.
## Discussion
Define completion contexts and ranking before adding snippets or resolve.
## Resolution
Ainda em aberto.
## Next Step
List first-wave completion contexts beyond member access.
@@ -0,0 +1,62 @@
---
id: AGD-0053
ticket: pbs-lsp-folding-and-selection-ranges
title: PBS LSP Folding and Selection Ranges
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, folding, selection-range]
---
## Pain
Domain owner: `studio/lsp`
PBS users do not get structural folding or smart selection ranges, so navigating and selecting nested functions, blocks, services, attributes, and Doc text blocks is more manual than it should be.
## Context
The parser knows syntactic spans for declarations and blocks. The LSP currently does not announce foldingRangeProvider or selectionRangeProvider.
## Open Questions
- [ ] Should folding ranges be computed from AST spans, token pairs, or a hybrid that handles malformed files?
- [ ] Which constructs should fold in the first wave: imports, declarations, blocks, attributes, text blocks, or all of them?
- [ ] What selection range hierarchy should VS Code expose for identifiers, calls, params, blocks, and declarations?
## Options
### Option A - Parser span ranges
- **Approach:** Use AST spans for declarations, blocks, parameter lists, and text blocks.
- **Pro:** Accurate for well-formed code and easy to map to LSP ranges.
- **Con:** Can degrade when parsing fails mid-file.
- **Maintainability:** Good because parser owns structure.
### Option B - Token pair scanning
- **Approach:** Compute ranges from brace/paren/text-block tokens without relying on full AST validity.
- **Pro:** More resilient during malformed edits.
- **Con:** Less semantically meaningful and can produce noisy ranges.
- **Maintainability:** Good only as fallback.
## Tradeoffs
Folding and selection should keep working while typing. AST ranges are higher quality, but token fallback may be needed for broken files.
## Recommendation
Prefer a hybrid: AST ranges first, token fallback for unmatched/incomplete constructs.
## Discussion
This is an editor UX feature with low semantic dependency and can be planned independently.
## Resolution
Ainda em aberto.
## Next Step
Decide first-wave foldable constructs and selection hierarchy.
@@ -0,0 +1,62 @@
---
id: AGD-0054
ticket: pbs-lsp-diagnostics-ux
title: PBS LSP Diagnostics UX
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, diagnostics, ux]
---
## Pain
Domain owner: `studio/lsp`
PBS diagnostics reach VS Code, but the editor experience can still be hard to act on when ranges, related information, severity, and phase context are too generic.
## Context
The LSP maps compiler diagnostics to LSP diagnostics. Compiler diagnostics include phase/code/message/range in several paths, but not all editor-friendly fields are consistently exposed.
## Open Questions
- [ ] Which diagnostics need better ranges or related information first?
- [ ] Should diagnostic source include compiler phase, frontend, or both?
- [ ] What structured diagnostic metadata is needed for future quick fixes without coupling UI to text messages?
## Options
### Option A - Mapper-only UX improvements
- **Approach:** Improve LSP diagnostic mapping using existing diagnostic fields.
- **Pro:** Quick improvements to source, severity, and display consistency.
- **Con:** Cannot add missing related spans or repair metadata that compiler diagnostics do not expose.
- **Maintainability:** Good for presentation-only changes.
### Option B - Structured diagnostic contract
- **Approach:** Extend compiler diagnostics with editor-facing metadata such as phase, related locations, stable quick-fix hints, and refined ranges.
- **Pro:** Enables better diagnostics and future code actions without parsing message text.
- **Con:** Requires compiler/core diagnostic API changes.
- **Maintainability:** Strong as the foundation for quick fixes and richer UX.
## Tradeoffs
Some improvements belong in the LSP mapper, but the most valuable changes require structured information at diagnostic creation time.
## Recommendation
Prefer Option B as the target, with mapper-only cleanup allowed for low-risk display fixes.
## Discussion
This agenda should inform code actions before quick fixes become broad.
## Resolution
Ainda em aberto.
## Next Step
Inventory high-frequency diagnostics and identify which need better ranges, related information, or repair hints.
@@ -0,0 +1,62 @@
---
id: AGD-0055
ticket: pbs-lsp-document-links
title: PBS LSP Document Links
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, document-links, imports]
---
## Pain
Domain owner: `studio/lsp`
PBS import/module references and asset-like surfaces are plain text in VS Code, so users cannot click through to imported modules or related project resources.
## Context
The compiler understands module references and project files. The LSP does not announce documentLinkProvider, and the VS Code extension delegates all language behavior to the server.
## Open Questions
- [ ] Which references should become document links first: imports, barrels, stdlib modules, assets/addressables, or docs?
- [ ] How should virtual or stdlib resources be represented as link targets?
- [ ] Should document links overlap with go-to-definition or remain a separate lightweight navigation feature?
## Options
### Option A - Import/module links only
- **Approach:** Expose `documentLink` for module references such as `@sdk:gfx` and barrel paths.
- **Pro:** Narrow, useful, and mostly aligned with existing module resolution.
- **Con:** Does not cover assets/addressables or documentation links.
- **Maintainability:** Strong first wave.
### Option B - General resource links
- **Approach:** Link imports, stdlib modules, assets, addressables, and possibly documentation/resource surfaces.
- **Pro:** Richer navigation across project resources.
- **Con:** Requires resource-specific target policy and may overlap with definition.
- **Maintainability:** Good only after resource ownership is defined.
## Tradeoffs
Document links should remain lightweight. If a link requires semantic identity and symbol resolution, it may belong in go-to-definition instead.
## Recommendation
Prefer Option A first: imports and module references only.
## Discussion
Coordinate stdlib target handling with go-to-definition.
## Resolution
Ainda em aberto.
## Next Step
Decide first-wave link targets and how stdlib/module URIs are represented.
@@ -0,0 +1,62 @@
---
id: AGD-0056
ticket: pbs-lsp-call-and-type-hierarchy
title: PBS LSP Call and Type Hierarchy
status: open
created: 2026-07-15
resolved:
decision:
tags: [studio, lsp, vscode, compiler-pbs, editor, call-hierarchy, type-hierarchy]
---
## Pain
Domain owner: `studio/lsp`
PBS users cannot inspect call relationships or type/member relationships from the editor, which limits understanding of larger projects after basic navigation exists.
## Context
Call hierarchy and type hierarchy are advanced LSP features. They depend on reliable semantic identity, references, callsite classification, and type ownership surfaces, so they should probably follow definition/references rather than precede them.
## Open Questions
- [ ] Should call hierarchy and type hierarchy be one discussion or split after dependencies are clearer?
- [ ] Which relationships are valuable for PBS: function call graph, service method calls, struct methods, contract implementations, host/intrinsic calls, or all of them?
- [ ] What prerequisite symbol/reference infrastructure must be completed before this feature is safe to implement?
## Options
### Option A - Call hierarchy only
- **Approach:** Implement incoming/outgoing calls for functions and methods after callsite references are stable.
- **Pro:** Directly useful and builds on executable call resolution.
- **Con:** Does not cover type ownership, contracts, or implementations.
- **Maintainability:** Strong if based on the reference/callsite index.
### Option B - Split call hierarchy and type hierarchy
- **Approach:** Treat call graph and type/member/contract hierarchy as separate decisions and implementations.
- **Pro:** Avoids mixing two different semantic models.
- **Con:** Creates more workflow artifacts and sequencing work.
- **Maintainability:** Strong; each feature can evolve on its own contract.
## Tradeoffs
Call hierarchy and type hierarchy share LSP vocabulary but not the same compiler data. Coupling them too early would make the first implementation harder to reason about.
## Recommendation
Prefer Option B: split after this agenda confirms dependencies, likely implementing call hierarchy first.
## Discussion
This should wait until definition and references are complete.
## Resolution
Ainda em aberto.
## Next Step
Decide whether to split into two discussions before moving to decisions.
@@ -0,0 +1,78 @@
---
id: AGD-0064
ticket: multi-frontend-sdk-canonical-definition
title: Auditoria e centralizacao gradual da definicao canonica do SDK
status: open
created: 2026-07-15
resolved:
decision:
tags: [compiler, compiler-general, sdk, stdlib, hostcalls, intrinsics, multi-frontend]
---
# Agenda - Centralizar a definicao do SDK
## Objetivo
Domain owner: `compiler/general`, com referencias a `compiler/pbs` e `vm-arch`.
Auditar onde hostcalls, intrinsics, tipos fundamentais e APIs do console sao definidos hoje antes de propor uma fonte canonica do SDK.
## Contexto atual
Existem specs em `docs/vm-arch/INTRINSICS.csv`, docs de stdlib em `docs/specs/compiler/18. Standard Library Surface Specification.md`, specs PBS de stdlib e host ABI, alem de possiveis registries em codigo e fixtures.
## Escopo
- Mapear fontes atuais da verdade.
- Identificar definicoes geradas, duplicadas e manuais.
- Propor caminho gradual para um schema canonico.
## Fora de escopo
- Migracao grande do SDK.
- Gerar bindings para C#, Kotlin, TypeScript ou outras linguagens.
- Escolher formato definitivo antes da auditoria.
## Arquivos e componentes a inspecionar
- `docs/vm-arch/INTRINSICS.csv`
- `docs/specs/compiler/18. Standard Library Surface Specification.md`
- `docs/specs/compiler-languages/pbs/5. Manifest, Stdlib, and SDK Resolution Specification.md`
- `docs/specs/compiler-languages/pbs/6.1. Intrinsics and Builtin Types Specification.md`
- `prometeu-compiler/...` registries de hostcalls/intrinsics/stdlib
- runtime/PVM quando presente no workspace ou submodulos relacionados
## Alteracoes propostas
Opcao A: agenda de auditoria pura, produzindo mapa de fontes e duplicacoes.
Opcao B: criar schema canonico minimo depois da auditoria, sem migrar todos os consumidores.
Recomendacao inicial: primeiro fechar auditoria; so uma decision posterior deve escolher formato e migracao.
## Estrategia de implementacao
Listar cada hostcall/intrinsic/tipo fundamental em suas fontes atuais, comparar assinatura, capability, versao e docs, e classificar divergencias.
## Testes necessarios
- Testes de paridade entre registry runtime/backend e specs existentes, se ja houver fonte estruturada.
- Testes de nao divergencia para IDs e assinaturas apos qualquer centralizacao.
## Criterios de aceitacao
- Existe mapa claro de fonte real da verdade.
- Duplicacoes e divergencias ficam registradas.
- Nenhuma migracao ampla e iniciada sem nova decisao.
## Riscos
- Tratar docs como fonte real quando o codigo ja diverge.
- Iniciar geracao antes de saber quem consome cada definicao.
## Decisoes que devem ser registradas
- Fonte canonica proposta.
- Campos obrigatorios do SDK.
- Sequencia de migracao segura.
@@ -0,0 +1,77 @@
---
id: AGD-0066
ticket: multi-frontend-synthetic-test-frontend
title: Frontend sintetico de teste para provar neutralidade do pipeline
status: open
created: 2026-07-15
resolved:
decision:
tags: [compiler, compiler-general, frontend, tests, backend, multi-frontend]
---
# Agenda - Criar um frontend sintetico de teste
## Objetivo
Domain owner: `compiler/general`.
Discutir um frontend exclusivo de testes que prove que o pipeline comum nao depende implicitamente do PBS.
## Contexto atual
O criterio arquitetural final exige que um frontend sintetico implemente contratos equivalentes a `FrontendProvider`, `FrontendSpec` e `FrontendCompiler`, produza IR comum minima e passe pelo backend ate PBX/verifier.
## Escopo
- Escolher entre frontend totalmente sintetico ou extensao ficticia.
- Definir fixture de IR minima com `frame` vazio.
- Garantir isolamento em testes.
## Fora de escopo
- Expor frontend sintetico no Studio.
- Criar segunda linguagem real.
- Criar parser completo ou grammar dummy complexa.
## Arquivos e componentes a inspecionar
- `prometeu-compiler/prometeu-frontend-registry/...`
- `prometeu-compiler/prometeu-build-pipeline/...`
- testes do build pipeline e backend
- `prometeu-studio/src/main/java/p/studio/projects/...` apenas se a selecao por extensao for exercitada
## Alteracoes propostas
Opcao A: provider sintetico que ignora arquivos e retorna IR minima diretamente.
Opcao B: extensao `.dummy` reconhecida em fixture, com conteudo ignorado ou trivial.
Recomendacao inicial: preferir Opcao A para provar neutralidade do pipeline sem criar outra superficie de linguagem.
## Estrategia de implementacao
Esperar provider/IR/lifecycle estarem suficientemente definidos, entao adicionar fixture de teste que registra provider sintetico e executa pipeline comum sem PBS.
## Testes necessarios
- Provider registrado e resolvido por `languageId`.
- Extensao/source handling testado se a opcao B for escolhida.
- IR comum produzida, lifecycle montado, IRVM gerado, PBX emitido e verificado.
## Criterios de aceitacao
- Teste falha se pipeline comum instanciar PBS implicitamente.
- Synthetic frontend existe apenas em test fixtures.
- PBS nao e afetado.
## Riscos
- Fixture sintetica exercitar caminho feliz pequeno demais e nao proteger acoplamentos reais.
- Virar uma segunda linguagem informal.
## Decisoes que devem ser registradas
- Forma do frontend sintetico.
- Escopo minimo da IR fixture.
- Onde o provider de teste pode ser registrado.
@@ -0,0 +1,79 @@
---
id: AGD-0067
ticket: multi-frontend-architectural-tests
title: Testes arquiteturais para fronteiras multi-frontend
status: open
created: 2026-07-15
resolved:
decision:
tags: [compiler, compiler-general, studio, frontend, architecture, tests, multi-frontend]
---
# Agenda - Criar testes arquiteturais
## Objetivo
Domain owner: `compiler/general`, com impacto em `studio` e `vm-arch`.
Definir testes que protejam fronteiras: backend comum sem PBS, IR comum sem AST PBS, Studio comum sem instanciacao PBS fora do composition root e runtime neutro.
## Contexto atual
O repo ja tem testes Java para backend, bytecode e IRVM. ArchUnit pode ser util, mas o alinhamento proibe adicionar biblioteca arquitetural quando testes simples bastarem.
## Escopo
- Escolher ferramenta minima.
- Definir regras por pacote.
- Adicionar testes que sejam baratos e claros.
## Fora de escopo
- Cobrir toda arquitetura do monorepo.
- Adicionar ArchUnit sem justificativa concreta.
- Bloquear referencias PBS em testes PBS ou no frontend PBS.
## Arquivos e componentes a inspecionar
- `build.gradle.kts` e version catalog se houver.
- `prometeu-compiler/prometeu-build-pipeline/src/test/java/...`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/test/java/...`
- `prometeu-studio/src/test/java/...`
- pacotes runtime/PVM quando presentes
## Alteracoes propostas
Opcao A: testes Java/reflection simples para imports, nomes de tipos e packages publicos.
Opcao B: ArchUnit caso a quantidade de regras e modulos justifique.
Recomendacao inicial: comecar com testes simples; reavaliar ArchUnit se as regras virarem frageis.
## Estrategia de implementacao
Definir allowlist de packages PBS, composition root e testes PBS. Escrever verificacoes independentes por dominio para falhas legiveis.
## Testes necessarios
- Backend/common nao importa `frontend.pbs`, `parser.pbs` ou `semantic.pbs`.
- IR comum nao expoe tipos PBS.
- Studio comum nao instancia servicos PBS fora da allowlist.
- Runtime nao usa nomenclatura ou semantica PBS.
## Criterios de aceitacao
- Regras rodam no build local.
- Falhas apontam arquivo/pacote responsavel.
- Excecoes autorizadas estao documentadas.
## Riscos
- Teste textual produzir falso positivo por docs ou fixtures.
- Criar barreira rigida antes de definir provider e lifecycle.
## Decisoes que devem ser registradas
- Ferramenta escolhida.
- Allowlist de packages.
- Ordem de introducao das regras.
@@ -0,0 +1,77 @@
---
id: AGD-0068
ticket: multi-frontend-avoid-premature-abstractions
title: Evitar abstracoes prematuras na preparacao multi-frontend
status: open
created: 2026-07-15
resolved:
decision:
tags: [compiler, compiler-general, studio, frontend, architecture, multi-frontend, simplicity]
---
# Agenda - Evitar abstracoes prematuras
## Objetivo
Domain owner: `compiler/general`, com impacto em `studio`.
Definir limites para a preparacao multi-frontend: remover acoplamento com baixo custo sem construir marketplace, hot reload, RPC, sandbox ou infraestrutura generica antes da necessidade real.
## Contexto atual
O PBS continuara sendo o unico frontend implementado no curto prazo. O objetivo e preparar contratos, fronteiras e testes de neutralidade, nao criar uma plataforma completa de plugins.
## Escopo
- Estabelecer principios para as agendas derivadas.
- Definir o que e explicitamente proibido nesta fase.
- Criar criterio para rejeitar abstracoes sem uso real.
## Fora de escopo
- Plugin marketplace.
- Carregamento dinamico de JARs.
- RPC, sockets, process supervisor ou version negotiation complexa.
- Service locator global.
## Arquivos e componentes a inspecionar
- `prometeu-compiler/prometeu-frontend-registry/...`
- `prometeu-app/src/main/java/p/studio/AppContainer.java`
- `prometeu-studio/src/main/java/p/studio/...`
- planos futuros derivados das discussions multi-frontend
## Alteracoes propostas
Opcao A: registrar esta decisao como guardrail transversal antes das demais.
Opcao B: deixar cada agenda repetir suas restricoes locais.
Recomendacao inicial: criar uma decisao curta e transversal apos discussao, para guiar provider, language services, synthetic frontend e testes arquiteturais.
## Estrategia de implementacao
Transformar o alinhamento em criterios de review: interface pequena, registro explicito, DI/composicao, records imutaveis e testes; rejeitar infra que nao seja exigida pelo PBS ou pelo frontend sintetico de teste.
## Testes necessarios
- Nao ha teste direto para simplicidade, mas plans devem incluir criterio de aceite contra reflection/scanning/RPC quando aplicavel.
- Testes arquiteturais podem proteger contra service locator global ou acoplamentos indevidos.
## Criterios de aceitacao
- Plans derivados nao introduzem infraestrutura de plugin antes da necessidade.
- Registro manual do PBS e suficiente.
- O frontend sintetico prova neutralidade sem virar plataforma externa.
## Riscos
- Usar "evitar abstracao" como desculpa para manter acoplamento PBS.
- Criar contratos tao pequenos que precisem ser quebrados imediatamente.
## Decisoes que devem ser registradas
- Guardrails obrigatorios para esta fase.
- Lista de mecanismos proibidos.
- Criterio para aceitar uma nova abstracao.
@@ -1,137 +0,0 @@
---
id: DEC-0039
ticket: pbs-symbol-documentation-and-hover-markdown
title: PBS symbol documentation with Doc markdown text blocks
status: accepted
created: 2026-07-15
ref_agenda: AGD-0039
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Context
PBS editor assistance needs hover content that can explain author intent, API contracts, parameter meaning, semantic constraints, and usage notes. Existing symbol surfaces can expose mechanical information such as kind, signature, origin, and type/shape, but PBS does not yet define a canonical authored documentation model.
This decision belongs to domain owner `compiler/pbs`.
Touched subdomains:
- `compiler/general`
- `studio/lsp`
- `tools/vscode-extension`
This decision resolves `AGD-0039`. It does not replace the existing wave-1 editor-assistance work; it defines the source-language and compiler-backed documentation model that hover, completion, signature help, stdlib documentation, and future editor consumers can reuse.
## Decision
PBS MUST support explicit authored documentation for named API declarations through a reserved `[Doc]` attribute.
The only canonical v1 shape is:
```pbs
[Doc(markdown = """
Draws a sprite.
- `x`: screen x
- `y`: screen y
""")]
```
`Doc` MUST be a reserved compiler-recognized attribute name. It MUST NOT be treated as an ordinary user-defined attribute with language-specific reinterpretation.
`Doc` MUST use a single required argument named `markdown`. The `markdown` argument MUST use a triple-quoted documentation text block as its payload. The v1 surface MUST NOT accept aliases or alternate forms such as `text`, positional arguments, `html`, `format`, `Document`, `Documentation`, or `md"""..."""`.
`[Doc]` MUST attach to the declaration immediately following it, following the normal PBS attribute association rule. Attribute order MUST be semantically irrelevant; `[Doc]` before or after other attributes on the same declaration MUST produce the same documentation result.
`[Doc]` MUST be valid only on named API declarations that introduce named semantic symbols and can carry attributes. The v1 documentable surface includes:
- functions;
- methods;
- classes/structs, when present in the PBS declaration surface;
- services;
- hosts;
- builtin and stdlib declarations;
- interface declarations.
Parameters MUST NOT receive `[Doc]` in v1 under any circumstance. Parameter documentation, when needed, MUST be authored inside the Markdown payload of the owning declaration.
Fields, enum cases, tuple fields, local variables, expressions, and other non-API declaration surfaces MUST NOT receive `[Doc]` unless a future decision explicitly classifies them as named API declarations documentable by `[Doc]`.
The `markdown` payload MUST be a PBS documentation text block, not a general runtime string literal. `"""..."""` MUST NOT become a general PBS runtime string literal as part of this decision.
The documentation text block MUST be normalized before semantic validation and before exposure to editor consumers. Normalization MUST:
- remove an incidental leading empty line after the opening delimiter;
- remove an incidental trailing empty line before the closing delimiter;
- compute the common indentation of non-empty content lines;
- remove that common indentation;
- preserve internal line breaks;
- preserve Markdown content after indentation normalization.
After normalization, an empty or whitespace-only documentation payload MUST be rejected.
The authored documentation format for PBS v1 MUST be a supported Markdown subset. PBS defines Markdown as the documentation protocol, not as a guarantee that every client renders every Markdown construct identically.
The compiler MUST preserve normalized Markdown documentation as semantic metadata of the documented symbol. LSP and editor consumers MUST consume compiler-resolved documentation metadata; they MUST NOT reparse PBS source to discover documentation.
`[Doc]` MUST be compile/editor metadata. It MUST NOT be lowered into runtime or cartridge artifacts unless a future runtime or packaging decision explicitly requires that.
Stdlib, SDK, and interface modules MUST use the same `[Doc(markdown = """...""")]` model for authored documentation. If imported or generated documentation exists in the future, authored source documentation MUST take precedence. Conflicts MUST be deterministic or produce explicit diagnostics.
## Rationale
An explicit attribute keeps documentation attached to language-owned semantic symbols instead of relying on proximity between comments and declarations. This avoids a separate comment-binding model and gives the compiler one canonical source of truth for editor assistance.
The short attribute name `Doc` is practical at the authoring site and remains clear when paired with the required `markdown` argument. The argument name carries the payload format, so a prefixed literal such as `md"""..."""` would add lexer and grammar surface without enough v1 benefit.
Restricting the v1 shape to `[Doc(markdown = """...""")]` protects maintainability. It avoids aliases, ambiguous forms, and compatibility pressure around multiple documentation payload formats.
Restricting triple-quoted text blocks to documentation avoids accidentally introducing general multiline runtime strings. Runtime strings raise separate questions about type behavior, lowering, constant pools, equality, escapes, and artifact representation; those questions do not belong to this documentation decision.
Compiler ownership keeps hover, completion, signature help, stdlib docs, and future editor consumers aligned. The LSP layer should transport resolved documentation, not infer language semantics from raw source text.
## Implications
- PBS syntax and parsing MUST recognize triple-quoted documentation text blocks in `[Doc(markdown = ...)]`.
- PBS semantic validation MUST enforce the canonical `Doc` shape, allowed declaration sites, duplicate detection, empty-payload rejection, and invalid-argument diagnostics.
- A declaration MUST NOT have more than one `[Doc]`; duplicates MUST be semantic errors rather than parse errors.
- Formatter implementations MUST preserve the internal Markdown text of documentation blocks and MUST NOT reflow authored Markdown.
- Rename/refactor tools are not required in this phase to update textual references inside Markdown documentation.
- Specs and stdlib documentation authored for repository-owned surfaces MUST be written in English.
- Tests SHOULD validate normalized Markdown payloads at compiler and LSP boundaries. VSCode visual validation may remain manual for this wave.
- Hover composition may be implemented later, but the ownership boundary is fixed: compiler resolves documentation, LSP transports it, and the editor renders it.
## Propagation Targets
- specs:
- `docs/specs/compiler-languages/pbs/3. Core Syntax Specification.md`
- `docs/specs/compiler-languages/pbs/4. Static Semantics Specification.md`
- `docs/specs/compiler-languages/pbs/11. AST Specification.md`
- `docs/specs/compiler-languages/pbs/12. Diagnostics Specification.md`
- `docs/specs/compiler-languages/pbs/5. Manifest, Stdlib, and SDK Resolution Specification.md`
- LSP/editor-facing specs if/when a dedicated Studio/LSP spec surface exists.
- plans:
- Create an implementation plan for syntax, AST, semantics, documentation metadata, LSP exposure, tests, and stdlib authoring.
- code:
- PBS lexer/parser/AST.
- PBS semantic validation and symbol/editorial surfaces.
- LSP hover/editorial transport.
- Stdlib/interface PBS sources.
- tests:
- Lexer/parser coverage for documentation text blocks.
- Semantic diagnostics for invalid `Doc` usage.
- Normalization tests for indentation and empty payloads.
- Compiler/editorial surface tests proving documentation is attached to resolved symbols.
- LSP hover or symbol-documentation payload tests.
- docs:
- PBS language specs.
- Any lessons generated after implementation and housekeeping.
## References
- Agenda: AGD-0039
- Source discussion: DSC-0036
## Revision Log
- 2026-07-15: Initial decision drafted from accepted `AGD-0039`.
@@ -1,63 +0,0 @@
---
id: PLN-0092
ticket: pbs-symbol-documentation-and-hover-markdown
title: Specify PBS Doc attribute and markdown text block syntax
status: done
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Update the PBS normative specs so `DEC-0039` is reflected before code changes begin. This plan is the specification gate for the `[Doc(markdown = """...""")]` feature.
## Objective
Define the syntax, static semantics, AST shape, diagnostics, and stdlib documentation policy for compiler-owned PBS symbol documentation.
## Dependencies
- Requires accepted decision `DEC-0039`.
- Must be completed before parser, semantic, LSP, and stdlib implementation plans are accepted as complete.
## Scope
- Add `[Doc(markdown = """...""")]` to the PBS attribute surface.
- Define triple-quoted documentation text blocks as documentation-only payloads, not runtime strings.
- Define documentable named API declarations and explicitly exclude parameters.
- Define normalization, duplicate handling, invalid shape diagnostics, and non-runtime lowering rules.
## Non-Goals
- Implement lexer, parser, semantics, or LSP code.
- Define general multiline runtime strings.
- Define VSCode rendering details beyond the compiler/LSP ownership boundary.
## Execution Method
1. Update `docs/specs/compiler-languages/pbs/3. Core Syntax Specification.md` with the canonical `[Doc(markdown = """...""")]` syntax, allowed declaration sites, attribute ordering rule, and text block delimiter rules.
2. Update `docs/specs/compiler-languages/pbs/4. Static Semantics Specification.md` with reserved `Doc` validation, duplicate rejection, parameter exclusion, empty normalized payload rejection, and non-documentable target diagnostics.
3. Update `docs/specs/compiler-languages/pbs/11. AST Specification.md` with an AST representation for documentation attributes and documentation text block payloads.
4. Update `docs/specs/compiler-languages/pbs/12. Diagnostics Specification.md` with required diagnostic categories for unterminated text block, invalid `Doc` shape, duplicate `Doc`, empty payload, and invalid target.
5. Update `docs/specs/compiler-languages/pbs/5. Manifest, Stdlib, and SDK Resolution Specification.md` with the rule that stdlib, SDK, and interface modules use the same authored `Doc` surface.
## Acceptance Criteria
- Specs define all normative rules from `DEC-0039` without deferring syntax or semantics to implementation judgment.
- The specs explicitly state that `"""..."""` is documentation-only for this feature.
- The specs explicitly state that parameters never receive `[Doc]` in v1.
- The specs identify propagation boundaries: compiler resolves, LSP transports, editor renders.
## Tests
- Run any existing spec conformance tests that validate PBS spec references.
- No code behavior tests are required in this plan.
## Affected Artifacts
- `docs/specs/compiler-languages/pbs/3. Core Syntax Specification.md`
- `docs/specs/compiler-languages/pbs/4. Static Semantics Specification.md`
- `docs/specs/compiler-languages/pbs/11. AST Specification.md`
- `docs/specs/compiler-languages/pbs/12. Diagnostics Specification.md`
- `docs/specs/compiler-languages/pbs/5. Manifest, Stdlib, and SDK Resolution Specification.md`
@@ -1,62 +0,0 @@
---
id: PLN-0093
ticket: pbs-symbol-documentation-and-hover-markdown
title: Implement PBS lexer support for documentation text blocks
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Add lexical support for triple-quoted documentation text blocks required by `[Doc(markdown = """...""")]`.
## Objective
Teach the PBS lexer to tokenize documentation text blocks without turning triple quotes into a general runtime string feature.
## Dependencies
- Depends on `PLN-0092` for the normative syntax.
## Scope
- Add a dedicated token kind for documentation text block payloads or an equivalent lexer representation.
- Recognize `"""` open and close delimiters.
- Preserve source text and spans for parser and diagnostics.
- Report unterminated documentation text blocks with useful span information.
## Non-Goals
- Parse or validate `[Doc]` attribute shapes.
- Normalize indentation.
- Lower documentation to runtime artifacts.
- Add general multiline string expressions.
## Execution Method
1. Update `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/lexer/PbsTokenKind.java` with a documentation text block token if a distinct token is the cleanest local pattern.
2. Update `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/lexer/PbsLexer.java` to recognize `"""..."""` and preserve its raw lexeme and span.
3. Ensure ordinary string scanning remains unchanged for `"..."`.
4. Add or update lexer diagnostics in `LexErrors.java` only if existing unterminated-string diagnostics cannot accurately represent unterminated documentation text blocks.
5. Add focused tests in `PbsLexerTest` for valid text blocks, multiline payloads, embedded single quotes/double quotes, and unterminated blocks.
## Acceptance Criteria
- The lexer can tokenize a multiline documentation text block.
- Unterminated documentation text blocks produce a deterministic diagnostic.
- Ordinary string literals keep current behavior.
- `"""..."""` is not accepted as an expression-level runtime string by this plan.
## Tests
- `./gradlew :prometeu-compiler:frontends:prometeu-frontend-pbs:test` or the repository-equivalent targeted compiler frontend test task.
- Focused `PbsLexerTest` cases for documentation text blocks.
## Affected Artifacts
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/lexer/PbsLexer.java`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/lexer/PbsTokenKind.java`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/lexer/LexErrors.java`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/test/java/p/studio/compiler/pbs/lexer/PbsLexerTest.java`
@@ -1,59 +0,0 @@
---
id: PLN-0094
ticket: pbs-symbol-documentation-and-hover-markdown
title: Extend PBS attribute parser and AST for Doc markdown payloads
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Extend parsing and AST representation so `[Doc(markdown = """...""")]` can be represented structurally before semantic validation.
## Objective
Parse documentation text block attribute values and expose them in AST without accepting alternate `Doc` shapes as valid semantics.
## Dependencies
- Depends on `PLN-0092`.
- Depends on `PLN-0093` for lexer support.
## Scope
- Extend attribute values to carry documentation text block payloads.
- Preserve raw payload and spans needed for normalization and diagnostics.
- Keep parser recovery consistent with existing attribute parsing.
## Non-Goals
- Validate whether `Doc` is allowed on a declaration.
- Normalize the text block.
- Attach documentation to semantic symbols.
## Execution Method
1. Update `PbsAst.AttributeValue` variants in `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/ast/PbsAst.java` to include a documentation text block value or equivalent typed payload.
2. Update `PbsAttributeParser` so attribute arguments can parse `markdown = """..."""`.
3. Keep existing string, int, and bool attribute argument parsing unchanged.
4. Ensure parser diagnostics remain syntax-level: malformed attribute structure is parse error; invalid `Doc` shape is left to semantic validation where possible.
5. Add parser tests proving `Doc` attributes parse on top-level and nested declaration forms that already support attributes.
## Acceptance Criteria
- `[Doc(markdown = """...""")]` parses into a distinct AST value or unambiguous attribute value representation.
- Existing attributes such as `[Init]`, `[Frame]`, `[Host(...)]`, and `[InitAllowed]` continue to parse.
- Invalid generic attribute syntax still recovers at the same boundaries as before.
## Tests
- Targeted parser tests under `prometeu-frontend-pbs/src/test/java/p/studio/compiler/pbs/parser`.
- Existing PBS parser test suite.
## Affected Artifacts
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/ast/PbsAst.java`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/parser/PbsAttributeParser.java`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/test/java/p/studio/compiler/pbs/parser/PbsParserTest.java`
@@ -1,59 +0,0 @@
---
id: PLN-0095
ticket: pbs-symbol-documentation-and-hover-markdown
title: Implement documentation text block normalization
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Implement the normalization algorithm required before semantic validation and editor exposure.
## Objective
Produce normalized Markdown text from a raw documentation text block using the rules locked by `DEC-0039`.
## Dependencies
- Depends on `PLN-0092`.
- Depends on `PLN-0094` for AST access to raw text block payloads.
## Scope
- Strip incidental leading and trailing empty lines.
- Compute and remove common indentation from non-empty lines.
- Preserve internal line breaks and Markdown content after normalization.
- Provide a single reusable normalization implementation for parser/semantics/editorial surfaces.
## Non-Goals
- Render Markdown.
- Sanitize Markdown for a specific editor.
- Reflow authored text.
## Execution Method
1. Add a small normalization helper in the PBS frontend package, colocated with parser or semantic support according to local patterns.
2. Feed the helper raw documentation text block content without delimiters.
3. Normalize line endings deterministically before indentation calculation if existing source handling requires it.
4. Validate empty/whitespace-only content after normalization.
5. Add unit tests for no indentation, common indentation, blank first/last lines, mixed blank internal lines, Markdown lists, code spans, and whitespace-only payloads.
## Acceptance Criteria
- Normalized output matches the spec for all required indentation cases.
- Whitespace-only content is detectable after normalization.
- The normalizer does not interpret Markdown and does not alter non-indentation content.
## Tests
- New focused unit tests for the normalizer.
- Parser/semantic tests that consume normalized output through `[Doc]`.
## Affected Artifacts
- New or existing PBS frontend helper under `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs`
- Corresponding tests under `prometeu-compiler/frontends/prometeu-frontend-pbs/src/test/java/p/studio/compiler/pbs`
@@ -1,64 +0,0 @@
---
id: PLN-0096
ticket: pbs-symbol-documentation-and-hover-markdown
title: Validate Doc attribute semantics and diagnostics
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Enforce the semantic contract of `[Doc]` after syntax can represent it.
## Objective
Reject invalid `Doc` usage and produce clear diagnostics for all invalid shapes and targets required by `DEC-0039`.
## Dependencies
- Depends on `PLN-0092`.
- Depends on `PLN-0094`.
- Depends on `PLN-0095`.
## Scope
- Reserve `Doc` as a compiler-recognized attribute.
- Accept only `Doc(markdown = """...""")`.
- Reject aliases, positional arguments, extra arguments, duplicate `Doc`, empty normalized payload, parameter targets, and non-documentable targets.
- Keep duplicates as semantic errors rather than parse errors.
## Non-Goals
- Attach documentation to final symbol metadata.
- Expose documentation through LSP.
- Author stdlib documentation.
## Execution Method
1. Locate existing declaration and reserved attribute validators, including `PbsDeclarationSemanticsValidator`, `PbsDeclarationRuleValidator`, and interface-module reserved attribute checks.
2. Add `Doc` validation in the semantic phase that already validates declaration attributes.
3. Define allowed declaration categories according to the spec: named API declarations that can carry attributes.
4. Add or extend `PbsSemanticsErrors` diagnostics for invalid `Doc` target, invalid shape, duplicate attribute, missing `markdown`, extra arguments, and empty payload.
5. Add tests covering valid declaration targets and every invalid case.
## Acceptance Criteria
- Every invalid shape named by `DEC-0039` produces a semantic diagnostic.
- Parameters never accept `[Doc]`.
- Duplicate `[Doc]` on one declaration is rejected semantically.
- Existing reserved attributes keep their current behavior.
## Tests
- `PbsSemanticsDeclarationsTest`
- `PbsInterfaceModuleSemanticsTest`
- `PbsDiagnosticsContractTest`
## Affected Artifacts
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/semantics/PbsDeclarationSemanticsValidator.java`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/semantics/PbsDeclarationRuleValidator.java`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/semantics/PbsSemanticsErrors.java`
- PBS semantics tests
@@ -1,60 +0,0 @@
---
id: PLN-0097
ticket: pbs-symbol-documentation-and-hover-markdown
title: Attach Doc metadata to PBS semantic symbols
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Once `[Doc]` is syntactically and semantically valid, attach normalized documentation to compiler-owned semantic symbol data.
## Objective
Make documentation available from resolved PBS symbols without requiring LSP or editor code to reparse source.
## Dependencies
- Depends on `PLN-0095`.
- Depends on `PLN-0096`.
## Scope
- Introduce an internal documentation metadata model with format and text.
- Attach normalized Markdown to documented named API symbols.
- Preserve the distinction between documentation metadata and runtime semantics.
## Non-Goals
- Render hover markdown.
- Write stdlib documentation content.
- Lower documentation into runtime artifacts.
## Execution Method
1. Add a compiler-side documentation metadata model, for example `format = markdown` plus normalized `text`.
2. Extend symbol or editorial-resolution structures so documented declarations expose this metadata.
3. Update symbol collection/binding paths for functions, methods, structs/classes, services, hosts, builtin/stdlib declarations, and interface declarations.
4. Ensure imported/stdlib declarations can carry the same metadata once parsed.
5. Add compiler tests proving documentation is attached to the intended declaration and not to parameters or unrelated symbols.
## Acceptance Criteria
- Resolved symbols can expose documentation as structured metadata, not an untyped hover-only string.
- Documentation metadata is absent for undocumented symbols.
- Documentation does not change type checking, name resolution, or runtime behavior.
## Tests
- Focused PBS semantic/editorial support tests.
- Existing name resolution and declaration tests to catch regressions.
## Affected Artifacts
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/semantics`
- `PbsEditorialResolvedSymbol`
- `PbsEditorialSupportService`
- Symbol/declaration binding code touched by documented declaration categories
@@ -1,60 +0,0 @@
---
id: PLN-0098
ticket: pbs-symbol-documentation-and-hover-markdown
title: Expose Doc metadata through PBS editorial and LSP surfaces
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Transport compiler-resolved documentation through the existing PBS editorial and LSP message boundaries.
## Objective
Expose normalized Markdown documentation to LSP consumers without source reparsing.
## Dependencies
- Depends on `PLN-0097`.
## Scope
- Extend compiler editorial result models to include documentation metadata.
- Extend LSP baseline messages where required.
- Keep documentation as Markdown-oriented structured data at the boundary.
## Non-Goals
- Final hover layout composition.
- VSCode automated rendering tests.
- Runtime artifact changes.
## Execution Method
1. Extend PBS editorial models such as `PbsEditorialResolvedSymbol`, completion candidates, or hover source models with optional documentation metadata.
2. Extend LSP API records under `prometeu-lsp/prometeu-lsp-api/src/main/java/p/studio/lsp/messages` only where the baseline protocol needs to carry documentation.
3. Update bridge code under `prometeu-lsp/prometeu-lsp-v1/src/main/java/p/studio/lsp/services/compiler` to map compiler documentation to LSP baseline messages.
4. Preserve backward-compatible behavior for symbols without documentation.
5. Add tests proving the LSP service receives compiler-resolved documentation.
## Acceptance Criteria
- LSP-facing code consumes compiler-resolved documentation metadata.
- No LSP code reparses PBS source to discover `[Doc]`.
- Existing completion, hover, and signature help behavior remains stable for undocumented code.
## Tests
- PBS editorial support tests.
- LSP bridge tests.
- Existing `prometeu-lsp` protocol mapping tests.
## Affected Artifacts
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/semantics/PbsEditorialResolvedSymbol.java`
- `prometeu-lsp/prometeu-lsp-api/src/main/java/p/studio/lsp/messages`
- `prometeu-lsp/prometeu-lsp-v1/src/main/java/p/studio/lsp/services/compiler`
- Related LSP tests
@@ -1,60 +0,0 @@
---
id: PLN-0099
ticket: pbs-symbol-documentation-and-hover-markdown
title: Render Doc markdown in hover composition
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Use transported documentation metadata in hover output while preserving the compiler/LSP/editor ownership boundary.
## Objective
Compose hover Markdown from signature/kind metadata and normalized `[Doc]` documentation.
## Dependencies
- Depends on `PLN-0098`.
## Scope
- Add documentation content to PBS hover composition.
- Keep Markdown rendering responsibility in the editor/client.
- Preserve hover behavior for undocumented symbols.
## Non-Goals
- Automate VSCode visual tests.
- Define a full documentation website generator.
- Change signature help parameter documentation behavior.
## Execution Method
1. Locate the PBS hover composition path in `PbsEditorialSupportService` and LSP bridge/mapping code.
2. Compose hover Markdown in deterministic order: signature/kind surface, documentation Markdown, then origin/type/shape metadata as currently appropriate.
3. Avoid escaping or reflowing normalized Markdown beyond safe composition separators.
4. Ensure undocumented symbols keep existing hover output.
5. Add hover tests with paragraphs, lists, and code spans from `[Doc]`.
## Acceptance Criteria
- Hover includes normalized authored Markdown for documented symbols.
- Hover does not include empty documentation sections.
- Markdown lists do not become accidental code blocks due to indentation.
- VSCode validation remains manual for this wave.
## Tests
- PBS editorial hover tests.
- LSP hover mapping tests that assert `MarkupKind.MARKDOWN` payloads.
## Affected Artifacts
- `PbsEditorialSupportService`
- `prometeu-lsp/prometeu-lsp-v1/src/main/java/p/studio/lsp/services/compiler`
- `prometeu-lsp/prometeu-lsp-v1/src/main/java/p/studio/lsp/services/protocol/mapping`
- Hover-related tests
@@ -1,60 +0,0 @@
---
id: PLN-0100
ticket: pbs-symbol-documentation-and-hover-markdown
title: Author Doc metadata for stdlib SDK and interface declarations
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Apply the new documentation model to repository-owned PBS stdlib, SDK, and interface declarations.
## Objective
Author English `[Doc(markdown = """...""")]` documentation for stdlib/API surfaces so editor assistance has useful content.
## Dependencies
- Depends on `PLN-0092`.
- Should run after `PLN-0096` so authored docs are validated by compiler tests.
- Benefits from `PLN-0099` for hover verification but does not require it to author source docs.
## Scope
- Add English Markdown documentation to stdlib, SDK, and interface declarations.
- Document parameters inside the owning declaration Markdown.
- Keep documentation concise and API-focused.
## Non-Goals
- Add `[Doc]` to parameters.
- Generate external documentation sites.
- Change stdlib semantics or host binding behavior.
## Execution Method
1. Inventory stdlib and SDK PBS files under `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/resources/game/stdlib`.
2. Add `[Doc(markdown = """...""")]` to public named API declarations that users see in completion/hover.
3. Document parameter meaning inside callable Markdown when needed.
4. Ensure host/interface declarations use the same `Doc` shape.
5. Run stdlib compile and interface conformance tests.
## Acceptance Criteria
- Public stdlib/SDK declarations that are user-facing have concise English documentation.
- No parameters receive `[Doc]`.
- Existing stdlib behavior and host ABI bindings are unchanged.
## Tests
- `PbsGateUStdlibCompileTest`
- `PbsGateUSdkInterfaceConformanceTest`
- Relevant PBS frontend tests.
## Affected Artifacts
- `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/resources/game/stdlib/1/**`
- Stdlib/interface compile tests
@@ -1,58 +0,0 @@
---
id: PLN-0101
ticket: pbs-symbol-documentation-and-hover-markdown
title: Protect runtime artifacts from Doc metadata lowering
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Guard the decision that `[Doc]` is compile/editor metadata and must not be lowered into runtime artifacts by default.
## Objective
Ensure documentation metadata does not affect IR, bytecode, cartridge output, stack behavior, or runtime validation.
## Dependencies
- Depends on `PLN-0097`.
- Should be validated before end-to-end conformance closure.
## Scope
- Confirm lowering ignores `Doc` metadata.
- Add regression tests comparing runtime-facing output with and without docs where practical.
- Protect constant pool and cartridge artifacts from documentation payload bloat.
## Non-Goals
- Define a future packaging format for documentation.
- Remove compiler/editor metadata before LSP consumption.
## Execution Method
1. Inspect PBS lowering paths under `prometeu-compiler/frontends/prometeu-frontend-pbs/src/main/java/p/studio/compiler/pbs/lowering`.
2. Ensure documented declarations lower identically to undocumented declarations for runtime-facing IR/backend artifacts.
3. Add golden or structural tests comparing compiled output for equivalent documented and undocumented programs.
4. Ensure large documentation payloads do not appear in runtime constant pools unless a future decision explicitly changes that.
## Acceptance Criteria
- Runtime-facing artifacts are unchanged by adding `[Doc]`.
- Compiler/editor metadata remains available before lowering for LSP/editor consumers.
- Golden artifact tests or equivalent regression tests cover this boundary.
## Tests
- `PbsGoldenArtifactsTest`
- Backend/lowering tests relevant to PBS IR output.
- Any artifact comparison test already used by the compiler frontend.
## Affected Artifacts
- PBS lowering package
- Golden artifact test resources if needed
- Compiler backend artifact tests
@@ -1,62 +0,0 @@
---
id: PLN-0102
ticket: pbs-symbol-documentation-and-hover-markdown
title: Add end-to-end Doc documentation conformance coverage
status: open
created: 2026-07-15
ref_decisions: [DEC-0039]
tags: [compiler, compiler-pbs, studio, lsp, vscode, editor, hover, documentation, markdown]
---
## Briefing
Close the feature with end-to-end tests that prove the whole `[Doc]` contract works across compiler and LSP boundaries.
## Objective
Validate that syntax, normalization, semantics, metadata attachment, LSP exposure, hover composition, stdlib docs, and runtime exclusion all satisfy `DEC-0039`.
## Dependencies
- Depends on `PLN-0092` through `PLN-0101`.
## Scope
- Add integrated compiler tests for valid and invalid `[Doc]`.
- Add LSP tests for hover documentation payloads.
- Add conformance coverage for normalization and runtime exclusion.
- Define a manual VSCode checklist for final visual validation.
## Non-Goals
- Automate VSCode UI rendering.
- Add new documentation syntax beyond `[Doc(markdown = """...""")]`.
- Reopen decision scope.
## Execution Method
1. Create one or more fixture PBS files using `[Doc]` on functions, methods, structs/classes when available, services, hosts, and interface declarations.
2. Add negative fixtures for parameters, duplicate `Doc`, empty payload, missing `markdown`, extra arguments, alias names, and invalid targets.
3. Add LSP tests proving hover transports Markdown and does not reparse source.
4. Add runtime artifact tests proving docs do not lower.
5. Record a short manual VSCode validation checklist covering paragraph rendering, list rendering, code spans, indentation behavior, and undocumented-symbol fallback.
## Acceptance Criteria
- All decision rules have at least one compiler or LSP test.
- Manual VSCode validation scope is documented but not automated.
- The feature can be implemented without returning to the agenda or decision for missing requirements.
## Tests
- Full relevant PBS frontend test suite.
- Relevant LSP module tests.
- Runtime/lowering golden tests.
- Manual VSCode checklist execution before final housekeeping.
## Affected Artifacts
- PBS frontend tests and fixtures
- LSP tests and fixtures
- Golden/runtime artifact tests
- Optional manual validation note in the implementation PR or plan completion record
@@ -82,7 +82,8 @@ The following inputs are already fixed elsewhere and must not be contradicted he
- Stable source span metadata is required at the token level and must remain useful to later diagnostics.
- Traps are fatal runtime outcomes rather than a recoverable userland diagnostic flow.
- Qualitative cost facts such as retention-bearing operations and host-boundary crossings are normative, but quantitative performance metrics are not.
- Topic `19` requires deterministic diagnostics for globals, lifecycle markers, init admission, and published-wrapper structural failures.
- PBS requires deterministic diagnostics for source-form globals, lifecycle markers, and init admission.
- Published-wrapper, boot-guard, synthetic-origin, and invalid-lifecycle-declaration failures are compiler-general. See `docs/specs/compiler/19. Verification and Safety Checks Specification.md` section 11 and `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`.
## 6. Required Diagnostic Phases
@@ -97,6 +98,8 @@ PBS-facing tooling must distinguish at least the following external diagnostic p
`linking` is a first-class phase in the external diagnostics contract and must not be collapsed away into a generic resolution bucket.
The `host-binding/capability admission` phase covers the source shape of `[Host]` and `[Capability]`. Canonical host-binding identity and capability consistency over `IRReservedMetadata` are platform-owned and are not moved by this specification.
This document does not require one finer-grained public phase split for verifier internals, artifact offsets, or transport-specific backend stages beyond what is stated here.
## 7. Required Diagnostic Identity, Payload, and Attribution
@@ -180,7 +183,7 @@ At minimum, the PBS diagnostics baseline must cover:
5. malformed, unauthorized, or capability-rejected host usage required by `6.2. Host ABI Binding and Loader Resolution Specification.md` and `7. Cartridge Manifest and Runtime Capabilities Specification.md`,
6. source-attributable backend-originated failures that remain user-actionable under normative lowering or load-facing rules.
7. dependency-scoped fail-fast admission at frontend boundary: when one module is rejected, modules that import it (directly or transitively) must not be emitted, while diagnostics collection for independent modules must continue.
8. lifecycle and globals diagnostics required by topic `19`, including:
8. lifecycle-marker and source-form global diagnostics, including:
- `global initializer uses unsupported form`,
- `global dependency cycle detected`,
- `global import must resolve through a global barrel entry`,
@@ -194,16 +197,13 @@ At minimum, the PBS diagnostics baseline must cover:
- `project init must be colocated with frame`,
- `Init attribute target invalid`,
- `InitAllowed is valid only on SDK host methods`,
- `host call not allowed during init`,
- `synthetic wrapper entrypoint missing`,
- `published entrypoint is not function index 0`,
- `hidden boot guard is missing`,
- `synthetic callable origin missing`.
9. symbolic asset-reference diagnostics required by backend-owned `Addressable` resolution, including:
- `host call not allowed during init`.
9. compiler-general lifecycle-assembly and IRVM-precondition failures, which this PBS catalog MUST NOT claim as PBS required coverage; see `docs/specs/compiler/19. Verification and Safety Checks Specification.md` section 11 and `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`.
10. symbolic asset-reference diagnostics required by backend-owned `Addressable` resolution, including:
- unresolved terminal asset reference,
- namespace-versus-terminal asset misuse,
- structurally invalid symbolic asset path admitted by source syntax but rejected by semantic surface rules.
10. ignored-value diagnostics for expression statements, including:
11. ignored-value diagnostics for expression statements, including:
- a stable warning when a materialized value is produced and then ignored,
- and no warning for unit-like expression statements.
@@ -1,13 +1,13 @@
# PBS Lowering IRBackend Specification
Status: Draft v1 (Frontend Scope + Backend Handoff Addendum)
Applies to: first lowering boundary from bound PBS frontend model into `IRBackend`/`IRBackendFile`, plus executable-backend handoff obligations at this boundary
Status: Draft v1 (PBS Frontend Scope + Common Handoff Emission)
Applies to: first lowering boundary from bound PBS frontend model into the common `IRBackend`/`IRBackendFile` handoff
## 1. Purpose
This document defines the normative frontend lowering contract from PBS source semantics to `IRBackend`.
This document defines the normative frontend lowering contract from PBS source semantics to the common `IRBackend` handoff.
Its purpose is to keep the first lowering deterministic and shared across implementations while PBS and backend design evolve.
Its purpose is to keep PBS lowering deterministic while preserving the common frontend-to-backend contract owned by compiler-general backend specs.
## 2. Scope
@@ -18,7 +18,7 @@ This document defines:
- semantic obligations preserved in `IRBackend`,
- deterministic rejection behavior for unsupported frontend-lowering forms,
- diagnostics attribution obligations for lowering failures in frontend scope,
- executable-backend handoff obligations at the `IRBackend` boundary,
- PBS obligations for emitting the common executable `IRBackend` handoff,
- and backend-owned symbolic lowering obligations for asset-facing references.
This document does not define:
@@ -26,7 +26,8 @@ This document does not define:
- VM lowering (`IRVM`),
- bytecode/PBX mapping,
- runtime execution behavior,
- verifier/loader internals.
- verifier/loader internals,
- or common backend ownership of the `IRBackend -> IRVM` contract.
Those concerns belong to shared acceptance specs under `docs/specs/compiler`.
@@ -75,9 +76,9 @@ When the active source slice depends on a backend-owned symbolic surface such as
- the frontend must not query domain services directly to discover operational assets;
- and lowering must preserve enough identity for the backend to complete final operational resolution.
## 6. IRBackend Preserved Obligations
## 6. PBS IRBackend Emission Obligations
For each admitted source unit and callable in the current lowering slice, `IRBackend` must preserve at minimum:
For each admitted PBS source unit and callable in the current lowering slice, PBS lowering must emit `IRBackend` that preserves at minimum:
1. callable identity (name/category as applicable),
2. callable arity,
@@ -85,14 +86,18 @@ For each admitted source unit and callable in the current lowering slice, `IRBac
4. source attribution anchor (`file + span`) for diagnostics and traceability,
5. source-observable parse intent for statement/expression structure (including precedence/associativity outcome already fixed by AST shape).
6. deterministic `requiredCapabilities` derived from admitted host-binding metadata for packer/runtime-manifest assistance.
7. compiler-selected published-wrapper entrypoint identity for backend handoff,
8. explicit global and synthetic lifecycle structure required by executable lowering.
7. a typed `IRLifecycleDeclaration` for executable programs, with file-init fragment, optional project-init, and required frame-root references,
8. explicit user globals and frontend-lowered file-init fragments required by later common lifecycle assembly,
9. backend-owned symbolic identities that must survive to later executable lowering, including symbolic asset references when admitted by the source slice.
PBS lowering MUST NOT emit module init, project-init wrapper, published frame wrapper, physical entrypoint, or boot guard. Those artifacts are produced exclusively by common lifecycle assembly.
Lowering must not collapse source categories in a way that erases required declaration/callable identity needed by downstream diagnostics or conformance assertions.
The normative contract is obligation-based, not tied to one mandatory in-memory class graph.
`IRBackend` itself is the common frontend-to-backend executable handoff. This PBS spec owns only the PBS-specific admission and translation path into that handoff. Common backend rules for consuming `IRBackend` are owned by `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`.
## 7. Deterministic Rejection Policy
If a source form is outside current frontend-lowering support:
@@ -111,7 +116,7 @@ For multi-module builds, lowering admission must apply dependency-scoped fail-fa
## 8. Conformance Boundary
`IRBackend` is the first lowering boundary (frontend responsibility).
PBS emission into `IRBackend` is the first PBS lowering boundary (frontend responsibility).
Conformance-valid claims at this boundary require Gate U evidence from `docs/specs/compiler/13. Conformance Test Specification.md`.
@@ -148,9 +153,9 @@ This document is healthy when:
3. deterministic rejection policy is explicit and test-backed,
4. and scope boundaries with general/backend acceptance specs are explicit.
## 12. Executable Backend Handoff Addendum (v1)
## 12. PBS Executable Handoff Emission Obligations (v1)
For executable backends, the `IRBackend` output from this frontend boundary must satisfy the additional handoff obligations below.
For executable backends, PBS lowering must emit `IRBackend` satisfying the additional handoff obligations below before common backend lowering starts.
### 12.1 Callable obligations
@@ -170,7 +175,7 @@ Each executable callsite in `IRBackend` must be classified into exactly one cate
2. `CALL_HOST`,
3. `CALL_INTRINSIC`.
Backend lowering must not infer this category by textual heuristics.
PBS lowering must not leave this category for backend inference by textual heuristics.
### 12.3 Host-backed metadata obligations
@@ -187,11 +192,11 @@ When a host-backed callsite also depends on backend-owned symbolic asset lowerin
- the frontend boundary MUST preserve which callsite argument is asset-facing;
- for PBS v1 this asset-facing designation is carried by reserved host metadata `[AssetLowering(param = N)]`;
- the preserved symbolic operand MUST remain attributable to a backend-provided `Addressable` identity;
- backend stages MUST be able to rewrite that symbolic operand into runtime-facing `asset_id` before the final low-level host path is emitted.
- common backend stages MUST be able to rewrite that symbolic operand into runtime-facing `asset_id` before the final low-level host path is emitted.
When an expression statement leaves a materialized runtime value on the stack:
- `IRBackend` lowering MUST emit explicit discard behavior rather than relying on implicit backend stack cleanup;
- PBS lowering into `IRBackend` MUST emit explicit discard behavior rather than relying on implicit backend stack cleanup;
- for single-slot materialized values this discard is represented by `POP`;
- and the discard path MUST preserve stack-accounting validity for subsequent backend lowering stages.
@@ -218,20 +223,18 @@ It does not replace:
2. `docs/specs/compiler/21. IRVM Optimization Pipeline Specification.md`,
3. or `docs/specs/compiler/15. Bytecode and PBX Mapping Specification.md`.
### 12.7 Executable Lifecycle and Published Wrapper Obligation
### 12.7 Executable Lifecycle Declaration Obligation
For executable PBS frontends, backend handoff must preserve the compiler-selected published wrapper rather than a frontend-declared nominal entrypoint.
For executable PBS frontends, PBS lowering MUST emit a typed `IRLifecycleDeclaration` rather than assembling the published lifecycle.
At `IRBackend` emission time:
At frontend `IRBackend` emission time:
1. the lowered graph must contain explicit lifecycle structure for:
- user-authored globals,
- file init fragments,
- module init,
- project init when present,
- and the published frame wrapper,
2. the published frame wrapper must be the effective entrypoint identity handed to backend stages,
3. the userland callable marked with `[Frame]` must remain distinguishable as the logical frame root,
4. the wrapper must own final `FRAME_RET`,
5. backend stages must not reintroduce manifest-owned or `FrontendSpec`-owned nominal entrypoint authority,
6. and hidden compiler-owned lifecycle state such as the boot guard must remain structurally distinguishable from user globals.
1. PBS MUST retain marker admission for `[Init]`, `[Frame]`, and `InitAllowed`,
2. PBS MUST emit typed references for:
- file-init fragments that include global materialization and non-project `[Init]`,
- the optional project-init callable, which is the `[Init]` colocated with `[Frame]`,
- and the required frame-root callable marked with `[Frame]`,
3. the colocated project init MUST be excluded from file-init fragments,
4. module order in the declaration MUST propagate the explicit dependency order already derived from resolved project/module data, without sorting by `FileId`, textual module name, or synthetic callable name,
5. PBS MUST NOT emit module init, project-init wrapper, published frame wrapper, physical entrypoint, or boot guard,
6. and common lifecycle assembly remains the owner of derived artifacts, including the published wrapper that becomes the physical entrypoint and owns the final frame return.
@@ -250,6 +250,8 @@ Rules:
- A file may declare at most one `[Init]`.
- The `[Init]` colocated with `[Frame]` is the project init.
- Project init must be colocated with `[Frame]`.
- PBS static semantics and lowering admit those markers and translate them into a typed `IRLifecycleDeclaration`.
- PBS MUST NOT materialize module init, project-init wrapper, published frame wrapper, physical entrypoint, or boot guard; those artifacts belong to common lifecycle assembly.
- `InitAllowed` is valid only on SDK host methods and is invalid on userland `fn`, `global`, `service`, `const`, `struct`, `contract`, and callback surfaces.
- A host method signature in a reserved stdlib/interface module MUST carry exactly one `Host` attribute.
- A host method signature in a reserved stdlib/interface module MUST carry exactly one `Capability` attribute.
@@ -102,17 +102,21 @@ Executable PBS programs have a logical bootstrap sequence before ordinary frame
Normative order:
1. materialize globals for each file according to dependency order,
2. execute that file's `[Init]`, if present,
2. execute that file's non-project `[Init]`, if present,
3. compose those steps into one synthetic module init per module,
4. execute project init, if present,
4. execute project init exactly once, if present, after all module initializers,
5. then invoke the userland callable marked with `[Frame]`.
The `[Init]` colocated with `[Frame]` is project init. It MUST NOT also run as a file-init fragment.
Lifecycle bootstrap is one-shot:
- no automatic retry occurs after boot failure,
- failure during module init aborts boot,
- failure during project init aborts boot.
Common lifecycle assembly materializes this order from the typed declaration. PBS remains the language source of marker meaning; it is not the owner of the published wrapper.
## 6.2 Logical frame root and physical wrapper
PBS distinguishes the logical frame root from the physical published entrypoint:
@@ -1,4 +1,4 @@
# PBS Bytecode and PBX Mapping Specification
# Bytecode and PBX Mapping Specification
Status: Draft v1 (Backend Baseline)
Applies to: mapping from lowered/optimized backend programs into PBX sections, bytecode-facing artifacts, and source-to-artifact invariants required by loader/verifier/runtime
@@ -9,6 +9,10 @@ This document defines the normative mapping between backend-lowered semantics an
Its purpose is to keep artifact emission deterministic and compatible with runtime loader/verifier contracts.
PBX is the only executable boundary between the compiler and runtime. `IRVM`
is a compiler-internal representation consumed by bytecode emission; it MUST
NOT be exposed to or consumed by the runtime loader, verifier, or PVM.
## 2. Scope
This document defines:
@@ -1,18 +1,21 @@
# PBS Verification and Safety Checks Specification
# Compiler Verification and Safety Checks Specification
Status: Draft v1 (Quality-Gate Baseline)
Applies to: test-backed safety gates for backend compiler stages and integration regression safety checks against runtime
Applies to: test-backed safety gates for common compiler stages, validation ownership across executable frontends, and integration regression safety checks against runtime
## 1. Purpose
This document defines the minimum safety/verification quality gates that must be enforced for executable backend pipelines.
This document defines the minimum safety/verification quality gates that must be enforced for executable compiler pipelines, and the ownership rule that separates language validation from platform validation.
Its purpose is to keep safety measurable and deterministic across:
- lowering,
- optimization,
- emission,
- and runtime integration boundaries.
- runtime integration boundaries,
- and the classification of diagnostics that every executable frontend must obey.
PBS is the current executable frontend. It is an example, not the owner of this specification.
## 2. Scope
@@ -20,37 +23,41 @@ This document defines:
- mandatory compiler-stage safety checks for executable pipeline stages,
- conditional integration safety regression checks against runtime line,
- and evidence requirements for safety-related support claims.
- evidence requirements for safety-related support claims,
- and validation-ownership rules that apply to all executable frontends.
This document does not define:
- full runtime verifier algorithm internals,
- formal proof of all runtime safety properties,
- or one mandatory verifier/optimizer architecture.
- one mandatory verifier/optimizer architecture,
- PBS syntax or static-semantics catalogs,
- or a generic validator-core package.
## 3. Authority and Precedence
Normative precedence:
1. `docs/pbs/specs/2. Governance and Versioning.md`
2. `docs/pbs/specs/12. Diagnostics Specification.md`
1. Runtime authority (`docs/specs/hardware/topics/chapter-2.md`, `chapter-3.md`, `chapter-9.md`, `chapter-12.md`, `chapter-16.md`)
2. Bytecode authority (`docs/specs/bytecode/ISA_CORE.md`)
3. `20. IRBackend to IRVM Lowering Specification.md`
4. `21. IRVM Optimization Pipeline Specification.md`
5. `15. Bytecode and PBX Mapping Specification.md`
6. `13. Conformance Test Specification.md`
7. This document
8. Frontend diagnostics specifications, including `docs/specs/compiler-languages/pbs/12. Diagnostics Specification.md`, for language-owned coverage only
If a rule here conflicts with higher-precedence authorities, it is invalid.
If a rule here conflicts with a higher-precedence authority, it is invalid. Frontend diagnostics documents MUST NOT redefine compiler-general safety gates or platform validation ownership.
## 4. Normative Inputs
This document depends on:
- `docs/pbs/specs/12. Diagnostics Specification.md`
- `13. Conformance Test Specification.md`
- `15. Bytecode and PBX Mapping Specification.md`
- `20. IRBackend to IRVM Lowering Specification.md`
- `21. IRVM Optimization Pipeline Specification.md`
- `docs/specs/compiler-languages/pbs/12. Diagnostics Specification.md` as the current frontend diagnostics catalog, not as the owner of this specification
## 5. Safety Gate Model
@@ -133,3 +140,74 @@ This document is healthy when:
2. executable backends have explicit S-I status (`pass`/`fail`/`deferred`),
3. integration safety regressions are preserved in tests,
4. and safety-related support claims remain evidence-backed.
## 11. Validation Ownership
A validation MUST live at the lowest layer that understands the failure without knowing the source language.
- A check is language-owned if it needs AST, tokens, a language marker, a barrel, or another source-language rule.
- A check is platform-owned if it can be performed on `IRBackend`, `IRLifecycleDeclaration`, IRVM, PBX, host/capability tables, or ids.
Adding a validation MUST record:
1. the input artifact the check reads;
2. whether that artifact is language source/AST/marker or a common contract;
3. the matrix owner;
4. the code family (`E_PARSE_*` / `E_SEM_*` / `E_LINK_*` / `E_HOST_*` versus `COMMON_*` / `LOWER_IRVM_*` / `MARSHAL_*`);
5. whether the span comes from a common-contract origin or from language tokens/AST.
If item 2 is both, the implementation MUST emit two diagnostics with distinct identities, not one hybrid.
### 11.1 Diagnostic identity and span
Platform diagnostics MUST use `COMMON_*`, `LOWER_IRVM_*`, or `MARSHAL_*` codes and the phase of the pipeline stage that reports them.
PBS MUST keep its current `E_PARSE_*`, `E_SEM_*`, `E_LINK_*`, and `E_HOST_*` codes.
This specification MUST NOT rewrite diagnostic messages or renumber existing codes.
Common code MUST copy origin/span when the declaration, instruction, or metadata carries one. A missing origin MUST be a structural error. Implementations MUST NOT invent a span.
The same fact MAY exist on both sides. PBS explains the language rule; the common layer rejects an invalid handoff even when the frontend has no PBS markers. The two errors MUST keep distinct identities. Example: `E_SEM_MISSING_PROJECT_FRAME` and a missing frame root on an executable `IRLifecycleDeclaration`.
### 11.2 Initial ownership matrix
| Family | Owner | Status | Notes |
| --- | --- | --- | --- |
| Parse / syntax (`ParseErrors`) | `compiler/pbs` | language | remains |
| Type, flow, declaration, const, and source-form global semantics | `compiler/pbs` | language | remains |
| Markers `[Init]`, `[Frame]`, `InitAllowed`, signature, colocation | `compiler/pbs` | language | already split by lifecycle assembly |
| `host call not allowed during init` | `compiler/pbs` | language | marker/source rule |
| PBS linking (`PbsLinkErrors`, barrel, import, visibility) | `compiler/pbs` | language | do not extract |
| Source shape of `[Host]` / `[Capability]` | `compiler/pbs` | language | host-admission split |
| Canonical host-binding identity and capability over `IRReservedMetadata` | `compiler/general` | platform | classified; validator remains in the PBS package until a later plan |
| Lifecycle assembler | `compiler/general` | platform | already implemented |
| IRVM preconditions for wrapper, boot guard, origin | `compiler/general` | platform | already implemented |
| `IRVMValidator`, lowering, marshal/bytecode | `compiler/general` | platform | already implemented |
| `Addressable` / asset path | `compiler/general` | platform | classified; no extra migration in this specification |
| Gates S-U/S-I | `compiler/general` | platform | this document |
### 11.3 Host admission, assets, and linking
Host admission is split:
- language owns the source shape of `[Host]` and `[Capability]`;
- platform owns canonical binding identity `(module, name, version)`, unknown capability, undeclared capability, and inconsistent binding once those facts exist on `IRReservedMetadata`.
`PbsHostAdmissionValidator` MAY remain in the PBS package until a later plan. `E_HOST_*` MUST NOT be treated as eternal PBS semantics.
`Addressable` / asset-path resolution is platform-owned. This specification classifies it and MUST NOT move additional asset code.
PBS linking, barrel files, and `PbsLinkErrors` remain language-owned. `ResolveDeps` does not authorize extracting them.
### 11.4 What this section does not authorize
This specification MUST NOT be read as authorizing:
- a generic validator-core package;
- moving `PbsHostAdmissionValidator` into a common package;
- moving asset-resolution production code;
- extracting PBS linking;
- or reopening lifecycle-assembly ownership.
Platform validators MUST NOT import `p.studio.compiler.pbs` and MUST NOT discover roles from `__pbs.` prefixes.
@@ -1,7 +1,7 @@
# PBS IRBackend to IRVM Lowering Specification
# IRBackend to IRVM Lowering Specification
Status: Draft v1 (Backend Baseline)
Applies to: executable-backend lowering from `IRBackend` handoff into `IRVM`
Applies to: executable-backend lowering from the common `IRBackend` handoff into `IRVM`
## 1. Purpose
@@ -9,11 +9,20 @@ This document defines the normative lowering contract from executable `IRBackend
Its purpose is to make backend lowering deterministic, reviewable, and compatible with runtime ISA/verification authority.
`IRBackend` is the common frontend-to-backend executable handoff. It is emitted by frontends and consumed by common backend stages. Backend lowering MUST NOT require PBS AST nodes, PBS tokens, PBS parser structures, PBS semantic objects, PBS editorial objects, or equivalent language-owned objects from future frontends.
`IRVM` is an internal compiler representation. It MUST be consumed by common
compiler optimization and PBX emission stages, and it MUST NOT cross the
compiler/runtime boundary or become an input, dependency, public type, or
operational concept of the runtime loader, verifier, or PVM. PBX is the only
executable artifact boundary between compiler and runtime.
## 2. Scope
This document defines:
- required backend preconditions before `IRBackend -> IRVM` starts,
- allowed language-neutral `IRBackend` contract concepts at the backend boundary,
- `IRVM` structural obligations needed before optimization/emission,
- deterministic function-id assignment and control-flow lowering obligations,
- callsite lowering obligations across function, host-backed, and VM-owned intrinsic paths,
@@ -25,7 +34,8 @@ This document does not define:
- binary PBX layout details,
- loader patching internals,
- runtime execution internals,
- or one mandatory optimizer implementation.
- one mandatory optimizer implementation,
- or frontend-specific lowering rules before `IRBackend` emission.
## 3. Authority and Precedence
@@ -33,8 +43,8 @@ Normative precedence:
1. Runtime authority (`docs/specs/hardware/topics/chapter-2.md`, `chapter-3.md`, `chapter-9.md`, `chapter-12.md`, `chapter-16.md`)
2. Bytecode authority (`docs/specs/bytecode/ISA_CORE.md`)
3. `docs/specs/compiler-languages/pbs/13. Lowering IRBackend Specification.md`
4. This document
3. This document
4. Frontend-specific lowering specifications, including `docs/specs/compiler-languages/pbs/13. Lowering IRBackend Specification.md`
If a rule here conflicts with higher-precedence authorities, it is invalid.
@@ -42,22 +52,106 @@ If a rule here conflicts with higher-precedence authorities, it is invalid.
This document depends on:
- `docs/specs/compiler-languages/pbs/13. Lowering IRBackend Specification.md`
- `docs/specs/compiler-languages/pbs/6.1. Intrinsics and Builtin Types Specification.md`
- `docs/specs/compiler-languages/pbs/6.2. Host ABI Binding and Loader Resolution Specification.md`
- frontend-specific lowering specifications that emit the common `IRBackend` handoff
- `15. Bytecode and PBX Mapping Specification.md`
- `21. IRVM Optimization Pipeline Specification.md`
## 4.1 Common IRBackend Contract
At this boundary, `IRBackend` may contain only language-neutral compiler concepts required by executable backend lowering:
1. source identity and spans,
2. modules and module identifiers,
3. callable identities, signatures, names, arity, and type-shape surfaces,
4. executable functions,
5. instruction kinds and instruction metadata,
6. globals and global origins,
7. synthetic executable functions and synthetic origins,
8. host-call metadata,
9. intrinsic metadata,
10. reserved metadata required for backend lowering,
11. required runtime or platform capabilities,
12. and an optional typed `IRLifecycleDeclaration` used as lifecycle provenance after common assembly.
The common backend contract MUST express these concepts as backend obligations, not as PBS syntax or PBS semantic rules. When a concept currently originates from PBS, the backend contract still owns only the language-neutral obligation carried by `IRBackend`.
### 4.3 IRLifecycleDeclaration
`IRLifecycleDeclaration` is the typed, immutable, serializable-by-design lifecycle provenance carried by `IRBackend`.
An executable handoff MUST contain exactly one declaration before common assembly. A deliberately non-executable handoff MAY omit it; absence MUST NOT synthesize a default `main` or published wrapper.
The declaration MUST represent, using ordered collections and typed ids:
1. modules in explicit dependency order,
2. each module's ordered file-init fragment references,
3. an optional project-init callable reference,
4. and exactly one frame-root callable reference.
References MUST use `ModuleId`, `FileId`, `CallableId`, a typed role, and source attribution. The public shape MUST NOT expose maps, sets, callbacks, compiler services, or frontend-owned types.
`AssembleLifecyclePipelineStage` MUST run after `FrontendPhasePipelineStage` and before `LowerToIRVMPipelineStage`. `LifecycleAssemblerService` is the only owner of derived lifecycle artifacts: module init, project-init wrapper, published frame wrapper, physical entrypoint, and hidden boot guard. Those artifacts MUST use deterministic neutral internal identities. Common code MUST NOT discover roles from `__pbs.*` prefixes or depend on PBS packages.
The declaration MUST remain available on the assembled `IRBackend` as typed provenance. Downstream inspection MUST use roles, ids, and origins rather than synthetic names.
Project init MUST be absent from file-init fragments and MUST execute exactly once, after all module initializers and before the frame root. Bootstrap failure MUST abort without retry. The boot guard MUST prevent a second bootstrap after successful completion. The published wrapper owns the final frame return.
The common backend contract MUST NOT expose or require:
1. `PbsExpression`, `PbsStatement`, `PbsToken`, or equivalent PBS-owned syntax objects,
2. PBS parser cursors, parse contexts, parse nodes, or token kinds,
3. PBS semantic validator internals,
4. PBS editorial or language-service objects,
5. or any type under a language frontend package as part of the public backend handoff.
## 4.2 Serializable-by-Design Contract Shape
The public `IRBackend` handoff MUST remain serializable by design. This requirement does not select a wire format, schema language, RPC mechanism, process boundary, plugin runtime, or external frontend transport. It only constrains the public data shape so those choices remain possible later without redesigning the handoff.
Public `IRBackend` contract types MUST be modelable as an acyclic, deterministic data graph made of:
1. primitive values and strings,
2. enums,
3. immutable value objects,
4. ordered lists,
5. nullable or optional scalar fields only where absence is part of the contract,
6. explicit identifier values for cross-references,
7. and source attribution values such as file ids and spans.
The public handoff MUST NOT expose, require, or rely on:
1. callbacks, lambdas, service objects, registries, visitors, or live compiler services,
2. frontend AST, parser, token, semantic, or editorial objects,
3. mutable global state as part of interpretation,
4. object identity equality as a semantic key,
5. unordered collections where iteration order affects output,
6. cyclic references between public handoff objects,
7. lazy values whose result depends on current process state,
8. or a selected wire format.
Cross-object references MUST use explicit ids or stable symbolic keys. Existing table-scoped identifiers such as `FileId`, `ModuleId`, `CallableId`, and `IntrinsicId` are valid handoff values when their table scope is explicit and deterministic. New typed ids SHOULD be introduced only when a concrete public handoff field would otherwise depend on textual identity, object identity, or a direct object reference.
Ordered collections MUST preserve deterministic order. Any map-like data that becomes part of the handoff MUST either be represented as an ordered list of entries or define deterministic key ordering before it reaches the public contract.
Initial enforcement for this contract SHOULD be audit and guardrail based: specs, reflection tests over public handoff surfaces, deterministic ordering tests, and targeted fixes for concrete leaks. A serializer or schema implementation is not required for conformance to this subsection.
## 5. Backend Entry Preconditions
Lowering from `IRBackend` to `IRVM` may start only when:
1. frontend admission to `IRBackend` is complete for the active slice,
2. required callsite categories (`CALL_FUNC`/`CALL_HOST`/`CALL_INTRINSIC`) are available,
3. required canonical host/intrinsic metadata is available for admitted callsites,
4. dependency-scoped fail-fast exclusions have already been applied at the `IRBackend` boundary,
5. compiler-selected published-wrapper entrypoint identity is present and unambiguous,
6. and target `vm_profile` is selected deterministically.
2. common lifecycle assembly has already materialized derived artifacts for an executable handoff,
3. required callsite categories (`CALL_FUNC`/`CALL_HOST`/`CALL_INTRINSIC`) are available,
4. required canonical host/intrinsic metadata is available for admitted callsites,
5. dependency-scoped fail-fast exclusions have already been applied at the `IRBackend` boundary,
6. compiler-selected published-wrapper entrypoint identity is present and unambiguous,
7. and target `vm_profile` is selected deterministically.
Declaration validation, role uniqueness, explicit order, and rejection of frontend-preassembled derived artifacts belong to `LifecycleAssemblerService`. `IRBackend -> IRVM` MUST keep only backend preconditions on the assembled graph, such as published-wrapper identity, synthetic origin presence, and hidden boot-guard presence. It MUST NOT reconstruct lifecycle order from callable names.
Those common structural diagnostics are compiler-general. Classification, code families, and dual language/platform reporting are owned by `docs/specs/compiler/19. Verification and Safety Checks Specification.md` section 11. They MUST NOT be catalogued as PBS-facing required coverage.
## 6. IRVM Model Obligations
@@ -98,12 +192,15 @@ Function indexing must be deterministic:
4. manifest-owned or name-only entrypoint fallback is forbidden,
5. and identical admitted input graphs must produce identical function-id assignments.
For PBS executable lowering:
For executable lowering:
- the userland callable marked with `[Frame]` remains the logical frame root,
- the wrapper path must contain final `FRAME_RET`,
- the declared frame-root callable remains the logical frame root,
- the assembled published wrapper is the physical entrypoint,
- the wrapper path must contain the final frame return,
- and lowering must preserve that distinction through `IRVM`.
PBS `[Init]`/`[Frame]` markers are a frontend source of the declaration. The backend-owned obligation is the neutral requirement that the compiler-selected published wrapper be the physical entrypoint and that lifecycle provenance remain explicit in `IRBackend` through typed roles and ids. Neutral synthetic names are internal and are not a public ABI.
## 9. Callsite Lowering Obligations
### 9.1 Function calls
@@ -150,7 +247,7 @@ Before bytecode emission, backend must run structural pre-verification on lowere
7. and structural validity of host/intrinsic call forms.
8. and structural validity of backend-owned symbolic-to-operational rewrites such as `Addressable -> asset_id`.
For PBS host-backed asset calls in v1, this validation includes the rewrite selected by preserved host metadata such as `[AssetLowering(param = N)]`.
For host-backed asset calls in v1, this validation includes backend-owned symbolic-to-operational rewrites selected by preserved host metadata. PBS currently emits this designation from metadata such as `[AssetLowering(param = N)]`, but final operational resolution is a common backend obligation.
This pre-verification does not replace runtime verifier authority.
@@ -184,4 +281,6 @@ This document is healthy when:
1. `IRBackend -> IRVM` obligations are explicit and testable,
2. deterministic function/call/control-flow lowering rules are explicit,
3. mandatory pre-verification boundary is explicit,
4. and scope boundaries with optimization/emission specs are explicit.
4. common backend obligations are not described as owned by PBS,
5. frontend-specific lowering specs only own emission into `IRBackend`,
6. and scope boundaries with optimization/emission specs are explicit.
@@ -1,4 +1,4 @@
# PBS IRVM Optimization Pipeline Specification
# IRVM Optimization Pipeline Specification
Status: Draft v1 (Pipeline Baseline)
Applies to: placement and invariants of the `OptimizeIRVM` stage between lowering and bytecode emission
@@ -9,6 +9,10 @@ This document defines the normative placement and safety invariants of the `Opti
Its purpose is to allow optimization in the backend pipeline without transferring semantic optimization responsibility to runtime.
`IRVM` is internal to the compiler. Optimization output MUST flow to PBX
emission and MUST NOT be exposed as a runtime input or protocol. PBX remains the
only executable boundary consumed by the runtime loader, verifier, and PVM.
## 2. Scope
This document defines:
@@ -3,7 +3,7 @@
Status: Draft v1 (Traceability Baseline)
Applies to: compiler/backend conformance traceability for canonical stage order and entrypoint-specific contracts
Last Updated: 2026-03-30
Last Updated: 2026-09-19
## 1. Purpose
@@ -13,7 +13,7 @@ This matrix maps each normative backend MUST from:
2. `docs/specs/compiler/20. IRBackend to IRVM Lowering Specification.md`
3. `docs/specs/compiler/21. IRVM Optimization Pipeline Specification.md`
4. `docs/specs/compiler/23. Compiler Pipeline Entry Points Specification.md`
5. `docs/specs/compiler-languages/pbs/13. Lowering IRBackend Specification.md` (Section 12 addendum only)
5. `docs/specs/compiler-languages/pbs/13. Lowering IRBackend Specification.md` (PBS lowering obligations only)
to concrete positive/negative test evidence and current status.
@@ -40,6 +40,22 @@ to concrete positive/negative test evidence and current status.
| G19-5.2.6 | Gate S-I MUST reject missing capability at load-time. | N/A | `BackendGateIIntegrationTest#gateI_rejectMissingCapability` | pass | |
| G19-5.2.7 | Gate S-I MUST cover valid VM-owned intrinsic path. | `BackendGateIIntegrationTest#gateI_validIntrinsicPath` | N/A | pass | |
| G19-5.2.8 | Gate S-I MUST cover repeatability across runtime line. | `RuntimeBackedCompatibilityAdapterTest#checkMustPassInStrictModeWhenRuntimeCommandIsValid`; `RuntimeBackedCompatibilityAdapterTest#checkMustBeRepeatableAcrossDeclaredRuntimeLinesInStrictMode` | `RuntimeBackedCompatibilityAdapterTest#checkMustFailInStrictModeWhenRuntimeCommandIsUnavailable` | pass | Multi-line repeatability is asserted with strict runtime-backed checks over distinct declared runtime lines. |
| G19-11.1 | A validation MUST live at the lowest layer that understands the failure without knowing the source language. | `ValidationOwnershipSpecTest#spec19MustBeCompilerGeneralAndPublishValidationOwnership` | N/A | pass | Spec 19 section 11 states the classification rule. |
| G19-11.2 | Spec 19 MUST publish the validation ownership matrix. | `ValidationOwnershipSpecTest#spec19MustBeCompilerGeneralAndPublishValidationOwnership` | N/A | pass | Matrix families include parse/syntax, PBS linking, and lifecycle assembler. |
| G19-11.3 | Platform diagnostics MUST use `COMMON_*` / `LOWER_IRVM_*` / `MARSHAL_*` and MUST copy origin/span when present. | `ValidationOwnershipSpecTest#spec19MustBeCompilerGeneralAndPublishValidationOwnership`; `LifecycleAssemblerServiceTest#assembleMustRejectMissingFrameRootWithoutRepair` | N/A | pass | Code families and span policy are normative in spec 19 section 11.1. |
| G19-11.4 | The same fact MAY exist as a language diagnostic and a platform diagnostic with distinct identities. | `ValidationOwnershipSpecTest#spec19MustBeCompilerGeneralAndPublishValidationOwnership` | N/A | pass | Dual reporting is required when a check reads both source and common contract. |
| G19-11.5 | Host admission MUST split source attribute shape from canonical `IRReservedMetadata` identity without moving the PBS validator in this plan. | `ValidationOwnershipSpecTest#spec19MustBeCompilerGeneralAndPublishValidationOwnership` | N/A | pass | Classified as platform; validator remains in the PBS package. |
| G19-11.6 | Platform validators MUST NOT import PBS packages or match `__pbs.` prefixes. | `CommonBackendArchitectureTest#commonBackendMustNotImportPbsFrontendPackages`; `CommonBackendArchitectureTest#commonLifecycleAndBackendMustNotDiscoverRolesFromPbsNamePrefixes` | N/A | pass | Guard covers backend, lifecycle, and all common pipeline stages. |
| PBS12-11 | PBS diagnostics MUST NOT claim common structural lifecycle/IRVM failures as PBS required coverage. | `ValidationOwnershipSpecTest#pbsDiagnosticsMustNotClaimCommonStructuralFailures` | N/A | pass | PBS spec 12 points those failures at spec 19 section 11 and spec 20. |
| G20-4.1.1 | `IRBackend` MUST be the common frontend-to-backend executable handoff. | `LowerToIRVMServiceTest#lowerMustAcceptManuallyConstructedCommonIRBackend` | N/A | pass | Direct backend lowering from manually constructed common IRBackend proves PBS parser/frontend services are not required. |
| G20-4.1.2 | Common backend code MUST NOT depend on `p.studio.compiler.pbs`. | `CommonBackendArchitectureTest#commonBackendMustNotImportPbsFrontendPackages` | N/A | pass | Source-level architectural guard rejects PBS frontend imports in common backend packages. |
| G20-4.1.3 | Public `IRBackend` contract MUST NOT expose PBS AST, token, parser, semantic, or editorial types. | `IRBackendExecutableContractTest#publicIRBackendContractMustNotExposePbsTypes` | N/A | pass | Reflection guard covers public contract fields, constructors, methods, and record components. |
| G20-4.2.1 | Public `IRBackend` contract types MUST remain modelable as an acyclic, deterministic data graph. | `IRBackendExecutableContractTest#publicIRBackendContractMustExposeOnlySerializableShapes`; `IRBackendExecutableContractTest#aggregatorMustEmitDeterministicOrderedPublicContractShape` | N/A | pass | Covers public shape guardrails, generic type arguments, ordered aggregate emission, table remapping, and capability de-duplication order. |
| G20-4.2.2 | Public `IRBackend` contract types MUST NOT expose callbacks, services, mutable/unordered collection contracts, frontend-owned objects, process-dependent lazy values, cyclic public references, or wire-format commitments. | `IRBackendExecutableContractTest#publicIRBackendContractMustExposeOnlySerializableShapes` | N/A | pass | Reflection guard covers public contract fields, constructors, methods, record components, and generic type arguments. |
| G20-4.2.3 | Public `IRBackend` cross-object references MUST use explicit ids or stable symbolic keys. | `IRBackendExecutableContractTest#publicIRBackendContractAuditMustClassifyAllowedSupportTypes`; `IRBackendExecutableContractTest#publicIRBackendContractMustUseExplicitIdsOrStableKeysForKnownReferences` | N/A | pass | Current audit found no concrete public serialization leak requiring model correction. Existing `FileId`, `ModuleId`, `CallableId`, and `IntrinsicId` remain explicit table-scoped identifiers. |
| G20-4.3.1 | Executable `IRBackend` MUST carry a typed, ordered, id-based `IRLifecycleDeclaration` as provenance. | `IRBackendExecutableContractTest#aggregatorMustRemapTypedLifecycleReferencesAndPreserveDeclaredModuleOrder`; `LifecycleAssemblerServiceTest#assembleMustMaterializeNeutralLifecycleArtifactsInDeclaredOrder` | `IRBackendExecutableContractTest#aggregatorMustOmitLifecycleDeclarationWhenNoFileContributesOne` | pass | Declaration uses `ReadOnlyList`, typed ids, roles, and source attribution; aggregator remaps ids with explicit module order. |
| G20-4.3.2 | Missing lifecycle declaration MUST mean a deliberately non-executable handoff and MUST NOT synthesize entrypoints. | `LifecycleAssemblerServiceTest#assembleMustNoOpWhenLifecycleDeclarationIsAbsent`; `AssembleLifecyclePipelineStageTest#runMustNoOpForDeliberatelyNonExecutableHandoff` | N/A | pass | Assembler and stage leave the backend unchanged when the declaration is absent. |
| G20-4.3.3 | Common lifecycle assembly MUST be the exclusive owner of derived artifacts and MUST use neutral identities. | `LifecycleAssemblerServiceTest#assembleMustMaterializeNeutralLifecycleArtifactsInDeclaredOrder`; `CommonBackendArchitectureTest#commonLifecycleAndBackendMustNotDiscoverRolesFromPbsNamePrefixes` | `LifecycleAssemblerServiceTest#assembleMustRejectPreMaterializedDerivedArtifacts` | pass | Derived wrappers and boot guard are common-owned; architecture guard rejects PBS imports and `__pbs.` prefix matching. |
| G20-6.2 | `IRVM_EXT` MUST declare structural metadata (`pops/pushes/is_branch/is_terminator`). | `IRVMValidatorTest#validateMustApplyStructuralMetadataForCustomInternalExtension`; `IRVMValidatorTest#validateMustRejectCustomInternalExtensionWhenStructuralMetadataUnderflowsStack`; `IRVMOp` record contract (`pops/pushes/branch/terminator/internal`) | N/A | pass | Dedicated extension fixtures now assert structural metadata is consumed by validation behavior. |
| G20-6.3 | `IRVM_EXT` MUST be eliminable before bytecode emission. | `OptimizeIRVMServiceTest#optimizeDefaultPassesMustEliminateUnreachableInternalExtensionBeforeEmission` | `EmitBytecodePipelineStageTest#runMustFailWhenInternalOpcodesRemain`; `EmitBytecodePipelineStageTest#runMustFailWhenInternalOpcodesRemainEvenWithNonEmptyEmissionPlan`; `IRVMValidatorTest#validateMustRejectInternalOpcodeWhenConfigured` | pass | Optimizer elimination path and emit-stage hard rejection path are both covered. |
| G20-6.4 | IRVM MUST preserve per-function slot and identity headers. | `IRVMProgramTest#constructorMustRejectModuleAndEmissionPlanMismatch` | `IRVMProgramTest#constructorMustRejectModuleAndEmissionPlanMismatch` | pass | Header mismatch is rejected deterministically. |
@@ -67,14 +83,18 @@ to concrete positive/negative test evidence and current status.
| G21-9.1 | Validation MUST include optimized-vs-non-optimized equivalence fixtures. | `OptimizeIRVMEquivalenceHarnessTest#optimizeOnOffMustPreserveObservableTraceForLoweredHostIntrinsicFixture`; `OptimizeIRVMEquivalenceHarnessTest#optimizeOnOffMustPreserveObservableTraceForConditionalJoinFixture`; `OptimizeIRVMEquivalenceHarnessTest#optimizeOnOffMustPreserveObservableTraceForSimpleLoopFixture`; `OptimizeIRVMEquivalenceHarnessTest#optimizeOnOffMustPreserveObservableTraceForLinearCallFixture` | N/A | pass | Dedicated opt on/off harness with reusable interpreter and deterministic trace assertions is in place. |
| G21-9.2 | Validation MUST preserve known negative loader/verifier behavior. | `BackendSafetyGateSUTest#emitStageMustExposeMarshalingLinkageFailureDeterministically`; `BackendGateIIntegrationTest` rejection suite | N/A | pass | |
| G21-9.3 | Validation MUST preserve deterministic artifact-level invariants. | `BackendSafetyGateSUTest#fullPipelineMustProduceDeterministicBytecodeForSameInput`; `BytecodeEmitterTest#emitMustRemainDeterministicAfterInterning` | N/A | pass | |
| G23-5.1 | Canonical compiler pipeline MUST preserve the shared stage order through `WriteBytecodeArtifact`. | `MainProjectPipelineIntegrationTest#analyzeShouldNotWriteProgramBytecode`; `MainProjectPipelineIntegrationTest#compileShouldProduceInMemoryBytecodeWithoutWritingProgramBytecode`; `MainProjectPipelineIntegrationTest#buildShouldWriteProgramBytecode` | N/A | partial | Entrypoint execution now covers the full analyze -> compile -> build path, but a dedicated order-only assertion for every terminal boundary is still missing. |
| G23-6.1 | `analyze` MUST terminate after `FrontendPhase` and MUST NOT execute backend artifact stages. | `MainProjectPipelineIntegrationTest#analyzeShouldNotWriteProgramBytecode` | N/A | pass | The integration fixture asserts the analysis snapshot contract and proves `program.pbx` is not written on the analysis path. |
| G23-5.1 | Canonical compiler pipeline MUST preserve the shared stage order through `WriteBytecodeArtifact`. | `BuilderPipelineServiceOrderTest#canonicalOrderMustContainOptimizeBetweenLowerAndEmit`; `MainProjectPipelineIntegrationTest#analyzeShouldNotWriteProgramBytecode`; `MainProjectPipelineIntegrationTest#compileShouldProduceInMemoryBytecodeWithoutWritingProgramBytecode`; `MainProjectPipelineIntegrationTest#buildShouldWriteProgramBytecode` | N/A | pass | Dedicated order assertion now covers analyses, compile, and build stage lists, including `AssembleLifecyclePipelineStage`. |
| G23-5.2 | `AssembleLifecyclePipelineStage` MUST run immediately after `FrontendPhasePipelineStage` and before IRVM lowering. | `BuilderPipelineServiceOrderTest#canonicalOrderMustContainOptimizeBetweenLowerAndEmit`; `BuilderPipelineServiceOrderTest#analyzeMustTerminateAfterLifecycleAssemblyWithoutBackendArtifacts` | N/A | pass | Analyses terminate at lifecycle assembly; compile stages begin at `LowerToIRVMPipelineStage`. |
| G23-6.1 | `analyze` MUST terminate after `AssembleLifecycle` and MUST NOT execute backend artifact stages. | `BuilderPipelineServiceOrderTest#analyzeMustTerminateAfterLifecycleAssemblyWithoutBackendArtifacts`; `MainProjectPipelineIntegrationTest#analyzeShouldNotWriteProgramBytecode` | N/A | pass | Analyze returns an assembled executable `IRBackend` with no IRVM, bytecode, or file output. |
| G23-6.2 | `compile` MUST terminate after `VerifyBytecode` and MUST return a validated in-memory executable result without disk write. | `MainProjectPipelineIntegrationTest#compileShouldProduceInMemoryBytecodeWithoutWritingProgramBytecode`; `BuilderPipelinePublicSurfaceTest#compileResultMustExposeValidatedInMemoryArtifacts` | N/A | pass | Coverage asserts validated in-memory bytecode artifacts and the no-write boundary for `compile`. |
| G23-6.3 | `build` MUST extend `compile` only with terminal artifact persistence. | `MainProjectPipelineIntegrationTest#buildShouldWriteProgramBytecode`; `BuilderPipelinePublicSurfaceTest#buildResultMustExposeArtifactPathAndCompilePayload` | N/A | pass | Coverage asserts that `build` retains the compile payload and writes the filesystem artifact path expected by callers. |
| G23-7.1 | `AnalysisSnapshot` MUST expose the minimum shared analysis contract. | `MainProjectPipelineIntegrationTest#analyzeShouldNotWriteProgramBytecode`; `BuilderPipelinePublicSurfaceTest#analysisSnapshotMustExposeMinimumSharedContract` | N/A | pass | Both executable and structural contract checks now cover the shared minimum analysis payload. |
| G23-8.1 | Caller-specific configs/contexts MUST NOT redefine canonical stage semantics. | N/A | N/A | missing | Requires explicit multi-entrypoint composition tests once non-filesystem contexts are implemented. |
| G23-9.1 | Legacy public `run` MUST be removed as the normative entrypoint and filesystem-default behavior MUST be expressed through `build`. | `BuilderPipelinePublicSurfaceTest#publicServiceSurfaceMustExposeExplicitEntrypointsAndNoPublicRun`; `MainProjectPipelineIntegrationTest#buildShouldWriteProgramBytecode` | N/A | partial | Public service coverage now proves `run` is no longer public and `build` is the filesystem artifact path, but the CLI composition path is not yet exercised by a dedicated automated test. |
| PBS13-12.0 | Executable backend handoff MUST satisfy addendum obligations. | `IRBackendExecutableContractTest` suite; `PBSFrontendPhaseServiceTest#shouldSynthesizePositiveTopic19FixtureAcrossFileInitProjectInitAndFrame` | `LowerToIRVMServiceTest#lowerMustRejectWhenSyntheticWrapperEntrypointIsMissing`; `LowerToIRVMServiceTest#lowerMustRejectWhenHiddenBootGuardIsMissing`; `LowerToIRVMServiceTest#lowerMustRejectWhenSyntheticCallableOriginIsMissing` | pass | Row group `PBS13-12.x` details each obligation; topic 19 closure now includes lifecycle wrapper/guard/origin evidence. |
| PBS13-12.0 | PBS executable lowering MUST emit `IRBackend` satisfying frontend handoff obligations before common backend lowering. | `IRBackendExecutableContractTest` suite; `PBSFrontendPhaseServiceTest#shouldSynthesizePositiveTopic19FixtureAcrossFileInitProjectInitAndFrame` | `LowerToIRVMServiceTest#lowerMustRejectWhenSyntheticWrapperEntrypointIsMissing`; `LowerToIRVMServiceTest#lowerMustRejectWhenHiddenBootGuardIsMissing`; `LowerToIRVMServiceTest#lowerMustRejectWhenSyntheticCallableOriginIsMissing` | pass | Row group `PBS13-12.x` verifies PBS emission into the common handoff; common backend ownership is covered by G20 rows. |
| PBS13-12.7 | PBS executable lowering MUST emit a typed lifecycle declaration and MUST NOT assemble derived wrappers. | `PBSFrontendPhaseServiceTest#shouldSynthesizePositiveTopic19FixtureAcrossFileInitProjectInitAndFrame`; `PbsFrontendCompilerTest#shouldExposeGlobalsAndTypedLifecycleDeclarationInBackendFile` | N/A | pass | Frontend tests assert declaration roles and the absence of module init, published wrapper, and boot guard. |
| PBS13-12.8 | Project init MUST be excluded from file-init fragments and MUST execute exactly once after module initializers. | `PbsFrontendCompilerTest#shouldExposeGlobalsAndTypedLifecycleDeclarationInBackendFile`; `LifecycleAssemblerServiceTest#assembleMustInvokeProjectInitExactlyOnceAndExcludeItFromFileInit` | N/A | pass | Colocated `[Init]` is declared as project init and invoked once by the assembled wrapper. |
| PBS13-12.9 | Lifecycle module order MUST propagate resolved dependency order without name or `FileId` inference. | `PBSFrontendPhaseServiceTest#shouldPublishDependencyOrderedLifecycleModulesInsteadOfNameOrFileIdOrder`; `IRBackendExecutableContractTest#aggregatorMustRemapTypedLifecycleReferencesAndPreserveDeclaredModuleOrder` | N/A | pass | Adversarial module names and file registration order do not change declared dependency order. |
| PBS13-12.1.1 | Callable identity MUST be preserved at handoff. | `IRBackendExecutableContractTest#aggregatorMustPreserveExecutableFunctionOrderDeterministically` | N/A | pass | |
| PBS13-12.1.2 | Observable callable signature MUST be preserved at handoff. | `IRBackendExecutableContractTest#functionContractMustRejectInvalidSlotAndSpanBounds` | N/A | pass | |
| PBS13-12.1.3 | Callable category MUST be preserved at handoff. | `PbsFrontendCompilerTest#shouldLowerExecutableFunctionsWithCallsiteCategories` | N/A | pass | |
@@ -103,7 +123,7 @@ to concrete positive/negative test evidence and current status.
1. Any backend PR that changes conformance-relevant behavior in FE/BE pipeline MUST update this matrix.
2. If requirement coverage changes, update: mapped tests, status, and `Last Updated` date.
3. New normative MUSTs in specs `19/20/21` or PBS `13` addendum MUST add a new matrix row before merge.
3. New normative MUSTs in specs `19/20/21` or PBS `13` lowering obligations MUST add a new matrix row before merge.
4. Conformance-relevant FE/BE code paths for hard-gate are:
- `prometeu-compiler/prometeu-build-pipeline/`
- `prometeu-compiler/frontends/prometeu-frontend-pbs/`
@@ -14,6 +14,7 @@ Its purpose is to make the compiler usable through explicit entrypoints instead
This document defines:
- the canonical stage order shared by compiler entrypoints,
- the frontend provider and registry boundary used by shared compiler composition,
- the terminal stage and mandatory behavior of `analyze`,
- the terminal stage and mandatory behavior of `compile`,
- the terminal stage and mandatory behavior of `build`,
@@ -25,6 +26,7 @@ This document does not define:
- one mandatory Java class or package layout,
- editor-specific source providers or document-session providers,
- frontend-specific semantic facts beyond the minimum shared `AnalysisSnapshot` contract,
- dynamic frontend plugin discovery,
- or one mandatory CLI architecture.
## 3. Authority and Precedence
@@ -63,12 +65,13 @@ The canonical stage order is:
1. `ResolveDepsPipelineStage`
2. `LoadSourcesPipelineStage`
3. `FrontendPhasePipelineStage`
4. `LowerToIRVMPipelineStage`
5. `OptimizeIRVMPipelineStage`
6. `EmitBytecodePipelineStage`
7. `LinkBytecodePipelineStage`
8. `VerifyBytecodePipelineStage`
9. `WriteBytecodeArtifactPipelineStage`
4. `AssembleLifecyclePipelineStage`
5. `LowerToIRVMPipelineStage`
6. `OptimizeIRVMPipelineStage`
7. `EmitBytecodePipelineStage`
8. `LinkBytecodePipelineStage`
9. `VerifyBytecodePipelineStage`
10. `WriteBytecodeArtifactPipelineStage`
Public entrypoints MAY terminate early according to this document, but they MUST NOT:
@@ -80,21 +83,23 @@ Public entrypoints MAY terminate early according to this document, but they MUST
### 6.1 `analyze`
`analyze` MUST terminate at `FrontendPhasePipelineStage`.
`analyze` MUST terminate at `AssembleLifecyclePipelineStage`.
`analyze` is defined as:
1. `ResolveDeps`
2. `LoadSources`
3. `FrontendPhase`
4. `AssembleLifecycle`
`analyze` MUST:
1. resolve workspace/dependency inputs needed by the frontend,
2. load source surfaces admitted by the selected compiler configuration,
3. run frontend semantic analysis,
4. return an `AnalysisSnapshot` or equivalent result contract,
5. and remain free of backend artifact side effects.
3. run frontend semantic analysis and emit a declarative `IRBackend`,
4. assemble the common lifecycle from `IRLifecycleDeclaration` when present,
5. return an `AnalysisSnapshot` or equivalent result contract whose `IRBackend` is fully assembled when the handoff is executable,
6. and remain free of backend artifact side effects.
`analyze` MUST NOT:
@@ -229,6 +234,56 @@ This authority model MUST NOT:
2. move semantic inline hint policy into LSP,
3. or imply that the host editor may synthesize semantic hints when the frontend did not publish them.
## 8.3 Frontend Provider and Registry Contract
The shared compiler composition surface MUST represent each registered frontend through a `FrontendProvider` or equivalent provider contract.
At minimum, each provider MUST expose:
1. `specification()`, returning the frontend-owned static contract, including its stable `languageId`;
2. `compiler()`, returning the compiler-facing frontend service used by the shared build pipeline;
3. `languageService()`, returning an optional editor-facing language-service capability for tooling consumers.
`compiler()` is the only required executable frontend capability. It MUST be sufficient for `analyze`, `compile`, `build`, and backend-facing pipeline flows.
`languageService()` MUST be optional. A frontend that supports compilation but does not provide editor services MUST remain a valid provider. Absence of `languageService()` MUST NOT prevent project analysis, compilation, backend lowering, bytecode emission, verification, or artifact writing.
The first shared editor-facing contract SHOULD remain an aggregated language-service capability surface. The common compiler contract MUST NOT require separate top-level service interfaces for each editor feature until a future accepted decision and plan identify a real lifecycle, ownership, or testing boundary that requires that split.
When exposed, editor-facing capabilities MAY include:
1. diagnostics intended for live editor sessions,
2. completion,
3. hover,
4. definition and navigation,
5. signature help,
6. semantic tokens for a live document,
7. formatting,
8. rename,
9. code actions,
10. and other host-facing editor assistance.
These capabilities are optional unless a future frontend- or host-specific decision makes one of them mandatory for a narrower surface. Tooling consumers MUST query capability availability before invoking editor-specific behavior.
Absent editor capabilities MUST have deterministic fallback behavior at tooling boundaries:
1. collection-shaped responses SHOULD be empty;
2. scalar optional responses SHOULD be absent or use a documented neutral response;
3. host protocols that can represent unsupported operations SHOULD do so explicitly;
4. and absent editor capability MUST NOT be reported as a compiler failure.
Compiler diagnostics and editor diagnostics are distinct ownership surfaces. Diagnostics returned by `analyze`, `compile`, and `build` are compiler contract output. Editor diagnostics MAY reuse compiler analysis results, live overlays, caches, or cancellation-aware tooling state, but a frontend MUST NOT be required to provide editor diagnostics in order to compile.
`FrontendSpec` remains the source of static frontend-owned presentation metadata such as semantic vocabularies, host projections, and visual themes. Producing semantic tokens for a live document is an optional editor-facing capability; the existence of static presentation metadata MUST NOT imply that every frontend can provide live semantic-token results.
The frontend registry MUST resolve providers by `languageId`. A lookup for an unknown `languageId` MUST fail explicitly with a diagnostic-friendly error. Unknown languages MUST NOT silently fall back to PBS or to any other frontend.
Common compiler, build pipeline, and Studio-facing consumers MUST obtain frontend behavior through provider lookup. They MUST NOT directly instantiate frontend-specific compiler services, except inside the frontend provider itself or inside an approved application composition root.
Provider registration MUST be explicit in v1. The registry contract MUST NOT use dynamic plugin discovery, reflection-based discovery, classpath scanning, external JAR loading, or runtime plugin installation to discover frontends.
PBS is the only initial real provider for this contract. PBS MAY be registered by an application composition root, but PBS MUST NOT become an exclusive semantic owner of the shared compiler pipeline.
## 9. Public Surface Rule
The shared compiler public surface MUST expose explicit entrypoints equivalent in meaning to:
+2 -1
View File
@@ -9,7 +9,7 @@ Scope boundary:
- PROMETEU itself is a fantasy handheld / fantasy console with a broader machine model, firmware model, cartridge model, and virtual hardware surface.
- This document does not define the whole PROMETEU machine.
- This document defines the VM/runtime subsystem that executes bytecode inside that machine.
- For broader machine-level framing, see [`../specs/README.md`](../specs/README.md).
- For broader machine-level framing, see [`../specs/README.md`](../specs/runtime/README.md).
Document roles:
@@ -149,6 +149,7 @@ The verifier statically checks bytecode for structural safety and stack‑shape
- `SyscallMeta` defines expected arity and return slot counts. The loader resolves `HOSTCALL` against this metadata and rejects raw `SYSCALL` in PBX pre-load artifacts; the verifier checks final IDs/arity/return‑slot counts against the same metadata.
- Arguments and returns
- Arguments are taken from the operand stack in the order defined by the ABI. Returns use multi‑slot results via a host‑side return buffer (`HostReturn`) which the VM copies back onto the stack, or zero slots for “void”. A mismatch in result counts is a fault/panic per current hardening logic.
- Example: the canonical asset runtime load surface is `asset.load(asset_id, slot) -> (status, handle)`. The caller does not supply `asset_name` or `asset_type`; bank kind is derived from `asset_table` using `asset_id`.
- Capabilities
- Cartridge capability flags are applied before load-time host resolution. Missing required capability aborts load; invoking a syscall without the required capability also traps defensively at runtime.
+11
View File
@@ -78,6 +78,17 @@ Authority rule:
- `INTRINSIC u32` — final numeric VM-owned intrinsic call.
- `FRAME_SYNC` — yield until the next frame boundary (e.g., vblank); explicit safepoint.
Host service arity is not encoded in the opcode itself. It is defined by resolved syscall metadata.
Example:
- `asset.load` currently resolves with `arg_slots = 2` and `ret_slots = 2`.
- The canonical stack contract is `asset_id, slot -> status, handle`.
- Callers do not provide an explicit asset kind; the runtime derives it from `asset_table`.
- `composer.bind_scene` resolves with `arg_slots = 1` and `ret_slots = 1`.
- The canonical stack contract is `bank_id -> status`.
- `composer.emit_sprite` resolves with `arg_slots = 9` and `ret_slots = 1`.
#### Canonical Intrinsic Registry Artifact
- Final intrinsic IDs and intrinsic stack metadata are published in [`INTRINSICS.csv`](INTRINSICS.csv).
+32
View File
@@ -0,0 +1,32 @@
# VM Architecture Mirror Policy
The Prometeu runtime repository is the normative authority for PBX binary
layout, Core ISA, loader behavior, verification, and PVM execution semantics.
The files in this directory are convenience mirrors for Studio-side compiler
and integration work; they are not an independent architectural authority.
Canonical runtime sources:
- `../runtime/docs/vm-arch/ARCHITECTURE.md`
- `../runtime/docs/vm-arch/ISA_CORE.md`
- `../runtime/docs/vm-arch/INTRINSICS.csv`
When these mirrors are retained, an integration change that updates a canonical
runtime source MUST update the corresponding Studio mirror in the same logical
change. Verify synchronization from the Studio repository root with:
```bash
diff -u docs/vm-arch/ARCHITECTURE.md ../runtime/docs/vm-arch/ARCHITECTURE.md
diff -u docs/vm-arch/ISA_CORE.md ../runtime/docs/vm-arch/ISA_CORE.md
diff -u docs/vm-arch/INTRINSICS.csv ../runtime/docs/vm-arch/INTRINSICS.csv
```
All three commands MUST exit successfully with no output before a change claims
that the mirrors are synchronized. If the runtime repository is not available
in the same checkout, synchronization evidence must be produced against the
canonical runtime revision before merging.
Compiler-owned transformation contracts remain under `docs/specs/compiler`.
Those specifications define `IRBackend -> IRVM -> PBX` emission, but they MUST
conform to the runtime-owned PBX and ISA contracts and MUST NOT redefine runtime
execution semantics.
+2
View File
@@ -7,6 +7,8 @@ dependencies {
implementation(project(":prometeu-infra"))
implementation(project(":prometeu-lsp:prometeu-lsp-api"))
implementation(project(":prometeu-lsp:prometeu-lsp-v1"))
implementation(project(":prometeu-compiler:frontends:prometeu-frontend-pbs"))
implementation(project(":prometeu-compiler:prometeu-frontend-registry"))
implementation(project(":prometeu-packer:prometeu-packer-api"))
implementation(project(":prometeu-packer:prometeu-packer-v1"))
implementation(project(":prometeu-studio"))
@@ -2,6 +2,7 @@ package p.studio;
import com.fasterxml.jackson.databind.ObjectMapper;
import p.packer.Packer;
import p.studio.compiler.FrontendRegistryService;
import p.studio.lsp.api.LspServerLifecycle;
import p.studio.lsp.events.StudioEventBus;
import p.studio.lsp.events.StudioPackerEventAdapter;
@@ -24,6 +25,7 @@ public final class AppContainer implements Container {
private final LspServerLifecycle lspServerLifecycle;
public AppContainer() {
FrontendRegistryService.bootstrapDefaults();
this.i18nService = new I18nService();
this.themeService = new ThemeService();
this.studioEventBus = new StudioEventBus();
@@ -0,0 +1,216 @@
package p.studio.compiler;
import p.studio.compiler.messages.FESurfaceContext;
import p.studio.compiler.messages.BuildingIssueSink;
import p.studio.compiler.messages.FrontendPhaseContext;
import p.studio.compiler.models.FrontendSemanticToken;
import p.studio.compiler.pbs.ast.PbsAst;
import p.studio.compiler.pbs.semantics.PbsEditorialCompletionCandidate;
import p.studio.compiler.pbs.semantics.PbsEditorialResolvedSymbol;
import p.studio.compiler.pbs.semantics.PbsEditorialSignatureHelp;
import p.studio.compiler.pbs.semantics.PbsEditorialSymbolKind;
import p.studio.compiler.pbs.semantics.PbsEditorialSupportService;
import p.studio.compiler.services.FrontendCompletionCandidate;
import p.studio.compiler.services.FrontendDocumentRequest;
import p.studio.compiler.services.FrontendDocumentation;
import p.studio.compiler.services.FrontendEditorialContext;
import p.studio.compiler.services.FrontendHover;
import p.studio.compiler.services.FrontendLanguageService;
import p.studio.compiler.services.FrontendSignature;
import p.studio.compiler.services.FrontendSignatureHelp;
import p.studio.compiler.services.FrontendSymbolKind;
import p.studio.compiler.services.PBSFrontendPhaseService;
import p.studio.compiler.source.diagnostics.DiagnosticSink;
import p.studio.compiler.source.identifiers.FileId;
import p.studio.utilities.structures.ReadOnlyList;
import java.util.List;
import java.util.Objects;
import java.util.Optional;
public final class PBSFrontendLanguageService implements FrontendLanguageService {
private final PbsEditorialSupportService editorialSupportService;
private final PBSSemanticTokenProvider semanticTokenProvider;
public PBSFrontendLanguageService() {
this(new PbsEditorialSupportService(), new PBSSemanticTokenProvider());
}
PBSFrontendLanguageService(
final PbsEditorialSupportService editorialSupportService,
final PBSSemanticTokenProvider semanticTokenProvider) {
this.editorialSupportService = Objects.requireNonNull(editorialSupportService, "editorialSupportService");
this.semanticTokenProvider = Objects.requireNonNull(semanticTokenProvider, "semanticTokenProvider");
}
@Override
public Optional<FrontendEditorialContext> prepareEditorialContext(
final FrontendPhaseContext phaseContext,
final FileId fileId) {
if (phaseContext == null || fileId == null) {
return Optional.empty();
}
final var semanticReadSurface = PBSFrontendPhaseService.semanticReadSurface(
phaseContext,
DiagnosticSink.empty(),
BuildingIssueSink.empty());
final var ast = semanticReadSurface.astByFile().get(fileId);
if (ast == null) {
return Optional.empty();
}
return Optional.of(new FrontendEditorialContext(
ast,
semanticReadSurface.supplementalTopDeclsByFile().getOrDefault(fileId, ReadOnlyList.empty()),
phaseContext.feSurfaceContext()));
}
public List<PbsEditorialCompletionCandidate> completion(
final String sourceText,
final PbsAst.File ast,
final ReadOnlyList<PbsAst.TopDecl> supplementalTopDecls,
final FESurfaceContext feSurfaceContext,
final int offset) {
return editorialSupportService.completion(sourceText, ast, supplementalTopDecls, feSurfaceContext, offset);
}
@Override
public List<FrontendCompletionCandidate> completion(
final FrontendDocumentRequest request,
final FrontendEditorialContext editorialContext,
final int offset) {
if (request == null
|| editorialContext == null
|| !(editorialContext.syntaxTree() instanceof PbsAst.File ast)
|| !(editorialContext.supplementalDeclarations() instanceof ReadOnlyList<?> supplementalTopDecls)
|| !(editorialContext.semanticContext() instanceof FESurfaceContext feSurfaceContext)) {
return List.of();
}
@SuppressWarnings("unchecked")
final ReadOnlyList<PbsAst.TopDecl> typedSupplementalTopDecls =
(ReadOnlyList<PbsAst.TopDecl>) supplementalTopDecls;
return completion(
request.documentText(),
ast,
typedSupplementalTopDecls,
feSurfaceContext,
offset).stream()
.map(this::toFrontendCompletionCandidate)
.toList();
}
public Optional<PbsEditorialResolvedSymbol> hover(
final String sourceText,
final PbsAst.File ast,
final ReadOnlyList<PbsAst.TopDecl> supplementalTopDecls,
final FESurfaceContext feSurfaceContext,
final int offset) {
return editorialSupportService.hover(sourceText, ast, supplementalTopDecls, feSurfaceContext, offset);
}
@Override
public Optional<FrontendHover> hover(
final FrontendDocumentRequest request,
final FrontendEditorialContext editorialContext,
final int offset) {
if (request == null
|| editorialContext == null
|| !(editorialContext.syntaxTree() instanceof PbsAst.File ast)
|| !(editorialContext.supplementalDeclarations() instanceof ReadOnlyList<?> supplementalTopDecls)
|| !(editorialContext.semanticContext() instanceof FESurfaceContext feSurfaceContext)) {
return Optional.empty();
}
@SuppressWarnings("unchecked")
final ReadOnlyList<PbsAst.TopDecl> typedSupplementalTopDecls =
(ReadOnlyList<PbsAst.TopDecl>) supplementalTopDecls;
return hover(
request.documentText(),
ast,
typedSupplementalTopDecls,
feSurfaceContext,
offset).map(this::toFrontendHover);
}
public Optional<PbsEditorialSignatureHelp> signatureHelp(
final String sourceText,
final PbsAst.File ast,
final ReadOnlyList<PbsAst.TopDecl> supplementalTopDecls,
final FESurfaceContext feSurfaceContext,
final int offset) {
return editorialSupportService.signatureHelp(sourceText, ast, supplementalTopDecls, feSurfaceContext, offset);
}
@Override
public FrontendSignatureHelp signatureHelp(
final FrontendDocumentRequest request,
final FrontendEditorialContext editorialContext,
final int offset) {
if (request == null
|| editorialContext == null
|| !(editorialContext.syntaxTree() instanceof PbsAst.File ast)
|| !(editorialContext.supplementalDeclarations() instanceof ReadOnlyList<?> supplementalTopDecls)
|| !(editorialContext.semanticContext() instanceof FESurfaceContext feSurfaceContext)) {
return FrontendSignatureHelp.empty();
}
@SuppressWarnings("unchecked")
final ReadOnlyList<PbsAst.TopDecl> typedSupplementalTopDecls =
(ReadOnlyList<PbsAst.TopDecl>) supplementalTopDecls;
return signatureHelp(
request.documentText(),
ast,
typedSupplementalTopDecls,
feSurfaceContext,
offset)
.map(this::toFrontendSignatureHelp)
.orElseGet(FrontendSignatureHelp::empty);
}
@Override
public List<FrontendSemanticToken> semanticTokens(final FrontendDocumentRequest request) {
return semanticTokens(request == null ? "" : request.documentText());
}
public List<FrontendSemanticToken> semanticTokens(final String documentText) {
return semanticTokenProvider.tokenize(documentText == null ? "" : documentText);
}
private FrontendCompletionCandidate toFrontendCompletionCandidate(final PbsEditorialCompletionCandidate candidate) {
return new FrontendCompletionCandidate(
candidate.label(),
toFrontendSymbolKind(candidate.kind()),
candidate.detail(),
candidate.documentation(),
candidate.origin());
}
private FrontendHover toFrontendHover(final PbsEditorialResolvedSymbol symbol) {
return new FrontendHover(
symbol.displayName(),
toFrontendSymbolKind(symbol.kind()),
symbol.detail(),
symbol.origin(),
symbol.signatures().stream()
.map(this::toFrontendSignature)
.toList(),
FrontendDocumentation.markdown(symbol.documentation()));
}
private FrontendSignatureHelp toFrontendSignatureHelp(final PbsEditorialSignatureHelp signatureHelp) {
return new FrontendSignatureHelp(
signatureHelp.signatures().stream()
.map(this::toFrontendSignature)
.toList(),
signatureHelp.activeSignature(),
signatureHelp.activeParameter());
}
private FrontendSignature toFrontendSignature(final p.studio.compiler.pbs.semantics.PbsEditorialSignature signature) {
return new FrontendSignature(signature.label(), signature.parameterLabels(), "");
}
private FrontendSymbolKind toFrontendSymbolKind(final PbsEditorialSymbolKind kind) {
return FrontendSymbolKind.valueOf(kind.name());
}
}
@@ -0,0 +1,44 @@
package p.studio.compiler;
import p.studio.compiler.models.FrontendSpec;
import p.studio.compiler.services.FrontendLanguageService;
import p.studio.compiler.services.FrontendPhaseService;
import p.studio.compiler.services.FrontendProvider;
import p.studio.compiler.services.PBSFrontendPhaseService;
import java.util.Optional;
public final class PBSFrontendProvider implements FrontendProvider {
private final FrontendPhaseService compiler;
private final PBSFrontendLanguageService languageService;
public PBSFrontendProvider() {
this(new PBSFrontendPhaseService(), new PBSFrontendLanguageService());
}
PBSFrontendProvider(final FrontendPhaseService compiler) {
this(compiler, new PBSFrontendLanguageService());
}
PBSFrontendProvider(
final FrontendPhaseService compiler,
final PBSFrontendLanguageService languageService) {
this.compiler = compiler;
this.languageService = languageService;
}
@Override
public FrontendSpec specification() {
return PBSDefinitions.PBS;
}
@Override
public FrontendPhaseService compiler() {
return compiler;
}
@Override
public Optional<FrontendLanguageService> languageService() {
return Optional.of(languageService);
}
}
@@ -0,0 +1,72 @@
package p.studio.compiler.pbs;
import java.util.ArrayList;
import java.util.List;
public final class PbsDocumentationTextBlockNormalizer {
private static final String DELIMITER = "\"\"\"";
private PbsDocumentationTextBlockNormalizer() {
}
public static String normalizeLexeme(final String lexeme) {
if (lexeme == null || !lexeme.startsWith(DELIMITER) || !lexeme.endsWith(DELIMITER) || lexeme.length() < 6) {
return "";
}
return normalizeContent(lexeme.substring(DELIMITER.length(), lexeme.length() - DELIMITER.length()));
}
public static String normalizeContent(final String rawContent) {
final var lines = splitLines(rawContent == null ? "" : rawContent.replace("\r\n", "\n").replace('\r', '\n'));
removeIncidentalBoundaryLines(lines);
final int commonIndent = commonIndent(lines);
if (commonIndent > 0) {
for (int i = 0; i < lines.size(); i++) {
final var line = lines.get(i);
if (line.isBlank()) {
lines.set(i, "");
} else {
lines.set(i, line.substring(Math.min(commonIndent, leadingWhitespace(line))));
}
}
}
return String.join("\n", lines);
}
private static List<String> splitLines(final String text) {
final var lines = new ArrayList<String>();
final var parts = text.split("\n", -1);
for (final var part : parts) {
lines.add(part);
}
return lines;
}
private static void removeIncidentalBoundaryLines(final List<String> lines) {
if (!lines.isEmpty() && lines.getFirst().isBlank()) {
lines.removeFirst();
}
if (!lines.isEmpty() && lines.getLast().isBlank()) {
lines.removeLast();
}
}
private static int commonIndent(final List<String> lines) {
var common = Integer.MAX_VALUE;
for (final var line : lines) {
if (line.isBlank()) {
continue;
}
common = Math.min(common, leadingWhitespace(line));
}
return common == Integer.MAX_VALUE ? 0 : common;
}
private static int leadingWhitespace(final String line) {
var count = 0;
while (count < line.length() && Character.isWhitespace(line.charAt(count))) {
count++;
}
return count;
}
}
@@ -8,6 +8,9 @@ import p.studio.compiler.models.IRFunction;
import p.studio.compiler.models.IRGlobal;
import p.studio.compiler.models.IRGlobalOrigin;
import p.studio.compiler.models.IRGlobalVisibility;
import p.studio.compiler.models.IRLifecycleCallableReference;
import p.studio.compiler.models.IRLifecycleCallableRole;
import p.studio.compiler.models.IRLifecycleFileDeclaration;
import p.studio.compiler.models.IRReservedMetadata;
import p.studio.compiler.models.IRSyntheticCallableKind;
import p.studio.compiler.models.IRSyntheticFunction;
@@ -215,6 +218,12 @@ public final class PbsFrontendCompiler {
if (fileInitExecutable != null) {
executableFunctions.add(fileInitExecutable);
}
final var lifecycleDeclaration = lowerLifecycleDeclaration(
fileId,
effectiveModuleId,
lifecycleArtifacts,
fileInitExecutable,
executableFunctions);
return new IRBackendFile(
fileId,
functions,
@@ -224,7 +233,8 @@ public final class PbsFrontendCompiler {
reservedMetadata,
effectiveModulePool,
ReadOnlyList.wrap(callableSignatures),
executableLowering.intrinsicPool());
executableLowering.intrinsicPool(),
lifecycleDeclaration);
}
public String callableShapeSurfaceOf(final PbsAst.FunctionDecl functionDecl) {
@@ -306,33 +316,19 @@ public final class PbsFrontendCompiler {
final var frameDecl = lifecycleArtifacts.frameDecl();
final var firstGlobalDecl = lifecycleArtifacts.firstGlobalDecl();
final var fileInitAnchorSpan = initDecl != null ? initDecl.span() : firstGlobalDecl == null ? null : firstGlobalDecl.span();
final var fileInitAnchorName = initDecl != null ? initDecl.name() : firstGlobalDecl == null ? "" : firstGlobalDecl.name();
final var ordinaryFileInit = frameDecl == null ? initDecl : null;
final var fileInitAnchorSpan = ordinaryFileInit != null
? ordinaryFileInit.span()
: firstGlobalDecl == null ? null : firstGlobalDecl.span();
final var fileInitAnchorName = ordinaryFileInit != null
? ordinaryFileInit.name()
: firstGlobalDecl == null ? "" : firstGlobalDecl.name();
if (fileInitAnchorSpan != null) {
syntheticFunctions.add(new IRSyntheticFunction(
moduleId,
fileInitCallableName(fileId),
IRSyntheticCallableKind.FILE_INIT_FRAGMENT,
new IRSyntheticOrigin(fileId, fileInitAnchorName, fileInitAnchorSpan)));
syntheticFunctions.add(new IRSyntheticFunction(
moduleId,
moduleInitCallableName(moduleId),
IRSyntheticCallableKind.MODULE_INIT,
new IRSyntheticOrigin(fileId, fileInitAnchorName, fileInitAnchorSpan)));
}
if (initDecl != null && frameDecl != null) {
syntheticFunctions.add(new IRSyntheticFunction(
moduleId,
projectInitCallableName(moduleId),
IRSyntheticCallableKind.PROJECT_INIT,
new IRSyntheticOrigin(fileId, initDecl.name(), initDecl.span())));
}
if (frameDecl != null) {
syntheticFunctions.add(new IRSyntheticFunction(
moduleId,
frameWrapperCallableName(moduleId),
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(fileId, frameDecl.name(), frameDecl.span())));
}
return ReadOnlyList.wrap(syntheticFunctions);
}
@@ -343,8 +339,10 @@ public final class PbsFrontendCompiler {
final PbsLifecycleArtifacts lifecycleArtifacts,
final ArrayList<IRBackendExecutableFunction> executableFunctions,
final ArrayList<CallableSignatureRef> callableSignatures) {
if (lifecycleArtifacts.firstGlobalDecl() == null
&& lifecycleArtifacts.initDecl() == null) {
final var ordinaryFileInit = lifecycleArtifacts.frameDecl() == null
? lifecycleArtifacts.initDecl()
: null;
if (lifecycleArtifacts.firstGlobalDecl() == null && ordinaryFileInit == null) {
return null;
}
final var syntheticCallableName = fileInitCallableName(fileId);
@@ -352,8 +350,8 @@ public final class PbsFrontendCompiler {
callableSignatures.add(new CallableSignatureRef(moduleId, syntheticCallableName, 0, "() -> unit"));
final var instructions = new ArrayList<IRBackendExecutableFunction.Instruction>();
final var moduleInitSource = lifecycleArtifacts.initDecl();
if (moduleInitSource != null && moduleInitSource != lifecycleArtifacts.frameDecl()) {
final var moduleInitSource = ordinaryFileInit;
if (moduleInitSource != null) {
final var initExecutable = resolveLifecycleExecutable(executableFunctions, moduleId, moduleInitSource.name());
if (initExecutable != null) {
instructions.add(new IRBackendExecutableFunction.Instruction(
@@ -390,6 +388,58 @@ public final class PbsFrontendCompiler {
span);
}
private IRLifecycleFileDeclaration lowerLifecycleDeclaration(
final FileId fileId,
final ModuleId moduleId,
final PbsLifecycleArtifacts lifecycleArtifacts,
final IRBackendExecutableFunction fileInitExecutable,
final ArrayList<IRBackendExecutableFunction> executableFunctions) {
final var fileInitReference = fileInitExecutable == null
? null
: new IRLifecycleCallableReference(
moduleId,
fileId,
fileInitExecutable.callableId(),
IRLifecycleCallableRole.FILE_INIT_FRAGMENT,
fileInitExecutable.span());
final var projectInitSource = lifecycleArtifacts.frameDecl() == null
? null
: lifecycleArtifacts.initDecl();
final var projectInitExecutable = projectInitSource == null
? null
: resolveLifecycleExecutable(executableFunctions, moduleId, projectInitSource.name());
final var projectInitReference = projectInitExecutable == null
? null
: new IRLifecycleCallableReference(
moduleId,
fileId,
projectInitExecutable.callableId(),
IRLifecycleCallableRole.PROJECT_INIT,
projectInitSource.span());
final var frameSource = lifecycleArtifacts.frameDecl();
final var frameExecutable = frameSource == null
? null
: resolveLifecycleExecutable(executableFunctions, moduleId, frameSource.name());
final var frameReference = frameExecutable == null
? null
: new IRLifecycleCallableReference(
moduleId,
fileId,
frameExecutable.callableId(),
IRLifecycleCallableRole.FRAME_ROOT,
frameSource.span());
final var declaration = new IRLifecycleFileDeclaration(
moduleId,
fileId,
fileInitReference,
projectInitReference,
frameReference);
return declaration.isEmpty() ? null : declaration;
}
private IRBackendExecutableFunction resolveLifecycleExecutable(
final ArrayList<IRBackendExecutableFunction> executableFunctions,
final ModuleId moduleId,
@@ -443,27 +493,6 @@ public final class PbsFrontendCompiler {
return "__pbs.file_init$f" + fileId.getId();
}
public static String moduleInitCallableName(final ModuleId moduleId) {
return "__pbs.module_init$m" + normalizedModuleIndex(moduleId);
}
public static String projectInitCallableName(final ModuleId moduleId) {
return "__pbs.project_init$m" + normalizedModuleIndex(moduleId);
}
public static String frameWrapperCallableName(final ModuleId moduleId) {
return "__pbs.frame_wrapper$m" + normalizedModuleIndex(moduleId);
}
public static String bootGuardGlobalName(final ModuleId moduleId) {
return "__pbs.boot_guard$m" + normalizedModuleIndex(moduleId);
}
private static int normalizedModuleIndex(final ModuleId moduleId) {
final var normalized = moduleId == null ? ModuleId.none() : moduleId;
return normalized.isNone() ? -1 : normalized.getIndex();
}
private int safeToInt(final long value) {
if (value > Integer.MAX_VALUE) {
return Integer.MAX_VALUE;
@@ -50,7 +50,7 @@ public enum PbsSemanticKind {
public static PbsSemanticKind forToken(final PbsToken token) {
return switch (token.kind()) {
case COMMENT -> COMMENT;
case STRING_LITERAL -> STRING;
case STRING_LITERAL, DOC_TEXT_BLOCK -> STRING;
case INT_LITERAL, FLOAT_LITERAL, BOUNDED_LITERAL -> NUMBER;
case TRUE, FALSE, NONE, SOME, OK, ERR -> LITERAL;
case VOID, OPTIONAL, RESULT -> BUILTIN_TYPE;
@@ -78,6 +78,7 @@ public final class PbsAst {
}
public sealed interface AttributeValue permits AttributeStringValue,
AttributeDocTextBlockValue,
AttributeIntValue,
AttributeBoolValue,
AttributeErrorValue {
@@ -89,6 +90,11 @@ public final class PbsAst {
Span span) implements AttributeValue {
}
public record AttributeDocTextBlockValue(
String value,
Span span) implements AttributeValue {
}
public record AttributeIntValue(
long value,
Span span) implements AttributeValue {
@@ -132,6 +138,7 @@ public final class PbsAst {
ReturnKind returnKind,
TypeRef returnType,
TypeRef resultErrorType,
ReadOnlyList<Attribute> attributes,
LifecycleMarker lifecycleMarker,
Block body,
Span span) implements TopDecl {
@@ -143,18 +150,21 @@ public final class PbsAst {
ReadOnlyList<FunctionDecl> methods,
ReadOnlyList<CtorDecl> ctors,
boolean hasBody,
ReadOnlyList<Attribute> attributes,
Span span) implements TopDecl {
}
public record ContractDecl(
String name,
ReadOnlyList<FunctionSignature> signatures,
ReadOnlyList<Attribute> attributes,
Span span) implements TopDecl {
}
public record HostDecl(
String name,
ReadOnlyList<FunctionSignature> signatures,
ReadOnlyList<Attribute> attributes,
Span span) implements TopDecl {
}
@@ -169,18 +179,21 @@ public final class PbsAst {
public record ServiceDecl(
String name,
ReadOnlyList<FunctionDecl> methods,
ReadOnlyList<Attribute> attributes,
Span span) implements TopDecl {
}
public record ErrorDecl(
String name,
ReadOnlyList<String> cases,
ReadOnlyList<Attribute> attributes,
Span span) implements TopDecl {
}
public record EnumDecl(
String name,
ReadOnlyList<EnumCase> cases,
ReadOnlyList<Attribute> attributes,
Span span) implements TopDecl {
}
@@ -190,6 +203,7 @@ public final class PbsAst {
ReturnKind returnKind,
TypeRef returnType,
TypeRef resultErrorType,
ReadOnlyList<Attribute> attributes,
Span span) implements TopDecl {
}
@@ -3,5 +3,6 @@ package p.studio.compiler.pbs.lexer;
public enum LexErrors {
E_LEX_INVALID_CHAR,
E_LEX_UNTERMINATED_STRING,
E_LEX_UNTERMINATED_DOC_TEXT_BLOCK,
E_LEX_INVALID_STRING_ESCAPE,
}
@@ -17,6 +17,7 @@ public final class PbsLexer {
IDENTIFIER,
NUMBER,
STRING,
DOC_TEXT_BLOCK,
LINE_COMMENT
}
@@ -52,6 +53,7 @@ public final class PbsLexer {
case IDENTIFIER -> scanIdentifierState();
case NUMBER -> scanNumberState();
case STRING -> scanStringState();
case DOC_TEXT_BLOCK -> scanDocTextBlockState();
case LINE_COMMENT -> scanLineCommentState();
}
}
@@ -123,7 +125,15 @@ public final class PbsLexer {
}
addToken(match('=') ? PbsTokenKind.SLASH_EQUAL : PbsTokenKind.SLASH);
}
case '"' -> state = LexerState.STRING;
case '"' -> {
if (peek() == '"' && peekNext() == '"') {
advance();
advance();
state = LexerState.DOC_TEXT_BLOCK;
return;
}
state = LexerState.STRING;
}
default -> {
if (isDigit(c)) {
state = LexerState.NUMBER;
@@ -206,6 +216,23 @@ public final class PbsLexer {
state = LexerState.DEFAULT;
}
private void scanDocTextBlockState() {
while (!isAtEnd()) {
if (peek() == '"' && peekNext() == '"' && peekAfterNext() == '"') {
advance();
advance();
advance();
addToken(PbsTokenKind.DOC_TEXT_BLOCK);
state = LexerState.DEFAULT;
return;
}
advance();
}
report(LexErrors.E_LEX_UNTERMINATED_DOC_TEXT_BLOCK, "Unterminated documentation text block");
state = LexerState.DEFAULT;
}
private void scanLineCommentState() {
while (!isAtEnd() && peek() != '\n') {
advance();
@@ -250,6 +277,17 @@ public final class PbsLexer {
return source.codePointAt(nextIndex);
}
private int peekAfterNext() {
if (isAtEnd()) return '\0';
final int first = source.codePointAt(current);
final int secondIndex = current + Character.charCount(first);
if (secondIndex >= source.length()) return '\0';
final int second = source.codePointAt(secondIndex);
final int thirdIndex = secondIndex + Character.charCount(second);
if (thirdIndex >= source.length()) return '\0';
return source.codePointAt(thirdIndex);
}
private boolean isAtEnd() {
return current >= source.length();
}
@@ -17,6 +17,7 @@ public enum PbsTokenKind {
FLOAT_LITERAL,
BOUNDED_LITERAL,
STRING_LITERAL,
DOC_TEXT_BLOCK,
COMMENT,
// Import keywords.
@@ -84,6 +84,10 @@ final class PbsAttributeParser {
final var token = cursor.previous();
return new PbsAst.AttributeStringValue(token.lexeme(), context.span(token.start(), token.end()));
}
if (cursor.match(PbsTokenKind.DOC_TEXT_BLOCK)) {
final var token = cursor.previous();
return new PbsAst.AttributeDocTextBlockValue(token.lexeme(), context.span(token.start(), token.end()));
}
if (cursor.match(PbsTokenKind.INT_LITERAL)) {
final var token = cursor.previous();
final var parsedLong = parseLongOrNull(token.lexeme());
@@ -108,28 +108,28 @@ final class PbsDeclarationParser {
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> pendingAttributes) {
if (cursor.match(PbsTokenKind.STRUCT)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "struct declarations");
return parseStructDeclaration(declareToken);
rejectUnsupportedDeclarationAttributes(pendingAttributes, "struct declarations");
return parseStructDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.CONTRACT)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "contract declarations");
return parseContractDeclaration(declareToken);
rejectUnsupportedDeclarationAttributes(pendingAttributes, "contract declarations");
return parseContractDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.SERVICE)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "service declarations");
return parseServiceDeclaration(declareToken);
rejectUnsupportedDeclarationAttributes(pendingAttributes, "service declarations");
return parseServiceDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.ERROR)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "error declarations");
return parseErrorDeclaration(declareToken);
rejectUnsupportedDeclarationAttributes(pendingAttributes, "error declarations");
return parseErrorDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.ENUM)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "enum declarations");
return parseEnumDeclaration(declareToken);
rejectUnsupportedDeclarationAttributes(pendingAttributes, "enum declarations");
return parseEnumDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.CALLBACK)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "callback declarations");
return parseCallbackDeclaration(declareToken);
rejectUnsupportedDeclarationAttributes(pendingAttributes, "callback declarations");
return parseCallbackDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.GLOBAL)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "global declarations");
@@ -139,14 +139,14 @@ final class PbsDeclarationParser {
return parseConstDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.HOST)) {
rejectAttributesBeforeUnsupportedTopDecl(pendingAttributes, "host declarations");
rejectUnsupportedDeclarationAttributes(pendingAttributes, "host declarations");
if (isNotInterfaceMode()) {
final var end = declarationTerminatorDelegate.consume();
report(cursor.previous(), ParseErrors.E_PARSE_RESERVED_DECLARATION,
"'declare host' is reserved and not supported in ordinary source modules");
return new PbsAst.InvalidDecl("reserved declare host", span(declareToken.start(), end));
}
return parseHostDeclaration(declareToken);
return parseHostDeclaration(declareToken, pendingAttributes);
}
if (cursor.match(PbsTokenKind.BUILTIN)) {
consume(PbsTokenKind.TYPE, "Expected 'type' in 'declare builtin type' declaration");
@@ -211,14 +211,16 @@ final class PbsDeclarationParser {
span(implementsToken.start(), end));
}
private PbsAst.HostDecl parseHostDeclaration(final PbsToken declareToken) {
private PbsAst.HostDecl parseHostDeclaration(
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> attributes) {
final var name = consume(PbsTokenKind.IDENTIFIER, "Expected host declaration name");
consume(PbsTokenKind.LEFT_BRACE, "Expected '{' to start host declaration body");
final var signatures = new ArrayList<PbsAst.FunctionSignature>();
while (!cursor.check(PbsTokenKind.RIGHT_BRACE) && !cursor.isAtEnd()) {
var attributes = ReadOnlyList.<PbsAst.Attribute>empty();
var signatureAttributes = ReadOnlyList.<PbsAst.Attribute>empty();
if (cursor.check(PbsTokenKind.LEFT_BRACKET)) {
attributes = attributeParser.parseAttributeList();
signatureAttributes = attributeParser.parseAttributeList();
}
if (!cursor.match(PbsTokenKind.FN)) {
report(cursor.peek(), ParseErrors.E_PARSE_INVALID_DECL_SHAPE,
@@ -226,12 +228,13 @@ final class PbsDeclarationParser {
cursor.advance();
continue;
}
signatures.add(parseFunctionSignature(cursor.previous(), attributes));
signatures.add(parseFunctionSignature(cursor.previous(), signatureAttributes));
}
final var rightBrace = consume(PbsTokenKind.RIGHT_BRACE, "Expected '}' to end host declaration body");
return new PbsAst.HostDecl(
name.lexeme(),
ReadOnlyList.wrap(signatures),
attributes,
span(declareToken.start(), rightBrace.end()));
}
@@ -308,7 +311,9 @@ final class PbsDeclarationParser {
return fields;
}
private PbsAst.StructDecl parseStructDeclaration(final PbsToken declareToken) {
private PbsAst.StructDecl parseStructDeclaration(
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> attributes) {
final var name = consume(PbsTokenKind.IDENTIFIER, "Expected struct name");
consume(PbsTokenKind.LEFT_PAREN, "Expected '(' in struct declaration");
final var fields = parseStructFields();
@@ -340,6 +345,7 @@ final class PbsDeclarationParser {
ReadOnlyList.wrap(methods),
ReadOnlyList.wrap(ctors),
hasBody,
attributes,
span(declareToken.start(), end));
}
@@ -382,10 +388,17 @@ final class PbsDeclarationParser {
final var methods = new ArrayList<PbsAst.FunctionDecl>();
final var ctors = new ArrayList<PbsAst.CtorDecl>();
while (!cursor.check(PbsTokenKind.RIGHT_BRACE) && !cursor.isAtEnd()) {
var attributes = ReadOnlyList.<PbsAst.Attribute>empty();
if (cursor.check(PbsTokenKind.LEFT_BRACKET)) {
attributes = attributeParser.parseAttributeList();
}
if (cursor.match(PbsTokenKind.FN)) {
methods.add(parseFunctionLike(cursor.previous(), ReadOnlyList.empty()));
methods.add(parseFunctionLike(cursor.previous(), attributes));
continue;
}
if (!attributes.isEmpty()) {
reportAttributesNotAllowed(attributes, "Attributes are valid only on struct methods in struct bodies");
}
if (cursor.match(PbsTokenKind.CTOR)) {
ctors.add(parseCtorDeclarationInBody(cursor.previous()));
continue;
@@ -411,17 +424,19 @@ final class PbsDeclarationParser {
span(ctorToken.start(), body.span().getEnd()));
}
private PbsAst.ContractDecl parseContractDeclaration(final PbsToken declareToken) {
private PbsAst.ContractDecl parseContractDeclaration(
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> attributes) {
final var name = consume(PbsTokenKind.IDENTIFIER, "Expected contract name");
consume(PbsTokenKind.LEFT_BRACE, "Expected '{' to start contract body");
final var signatures = new ArrayList<PbsAst.FunctionSignature>();
while (!cursor.check(PbsTokenKind.RIGHT_BRACE) && !cursor.isAtEnd()) {
var attributes = ReadOnlyList.<PbsAst.Attribute>empty();
var signatureAttributes = ReadOnlyList.<PbsAst.Attribute>empty();
if (cursor.check(PbsTokenKind.LEFT_BRACKET)) {
attributes = attributeParser.parseAttributeList();
signatureAttributes = attributeParser.parseAttributeList();
if (isOrdinaryMode()) {
reportAttributesNotAllowed(attributes, "Attributes are not allowed in ordinary .pbs source modules");
attributes = ReadOnlyList.empty();
reportAttributesNotAllowed(signatureAttributes, "Attributes are not allowed in ordinary .pbs source modules");
signatureAttributes = ReadOnlyList.empty();
}
}
if (!cursor.match(PbsTokenKind.FN)) {
@@ -430,30 +445,41 @@ final class PbsDeclarationParser {
cursor.advance();
continue;
}
signatures.add(parseFunctionSignature(cursor.previous(), attributes));
signatures.add(parseFunctionSignature(cursor.previous(), signatureAttributes));
}
final var rightBrace = consume(PbsTokenKind.RIGHT_BRACE, "Expected '}' to end contract body");
return new PbsAst.ContractDecl(name.lexeme(), ReadOnlyList.wrap(signatures), span(declareToken.start(), rightBrace.end()));
return new PbsAst.ContractDecl(name.lexeme(), ReadOnlyList.wrap(signatures), attributes, span(declareToken.start(), rightBrace.end()));
}
private PbsAst.ServiceDecl parseServiceDeclaration(final PbsToken declareToken) {
private PbsAst.ServiceDecl parseServiceDeclaration(
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> attributes) {
final var name = consume(PbsTokenKind.IDENTIFIER, "Expected service name");
consume(PbsTokenKind.LEFT_BRACE, "Expected '{' to start service body");
final var methods = new ArrayList<PbsAst.FunctionDecl>();
while (!cursor.check(PbsTokenKind.RIGHT_BRACE) && !cursor.isAtEnd()) {
var methodAttributes = ReadOnlyList.<PbsAst.Attribute>empty();
if (cursor.check(PbsTokenKind.LEFT_BRACKET)) {
methodAttributes = attributeParser.parseAttributeList();
}
if (!cursor.match(PbsTokenKind.FN)) {
if (!methodAttributes.isEmpty()) {
reportAttributesNotAllowed(methodAttributes, "Attributes are valid only on service methods in service bodies");
}
report(cursor.peek(), ParseErrors.E_PARSE_INVALID_DECL_SHAPE,
"Service body accepts only method declarations");
cursor.advance();
continue;
}
methods.add(parseFunctionLike(cursor.previous(), ReadOnlyList.empty()));
methods.add(parseFunctionLike(cursor.previous(), methodAttributes));
}
final var rightBrace = consume(PbsTokenKind.RIGHT_BRACE, "Expected '}' to end service body");
return new PbsAst.ServiceDecl(name.lexeme(), ReadOnlyList.wrap(methods), span(declareToken.start(), rightBrace.end()));
return new PbsAst.ServiceDecl(name.lexeme(), ReadOnlyList.wrap(methods), attributes, span(declareToken.start(), rightBrace.end()));
}
private PbsAst.ErrorDecl parseErrorDeclaration(final PbsToken declareToken) {
private PbsAst.ErrorDecl parseErrorDeclaration(
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> attributes) {
final var name = consume(PbsTokenKind.IDENTIFIER, "Expected error name");
consume(PbsTokenKind.LEFT_BRACE, "Expected '{' to start error body");
final var cases = new ArrayList<String>();
@@ -463,10 +489,12 @@ final class PbsDeclarationParser {
consume(PbsTokenKind.SEMICOLON, "Expected ';' after error case label");
}
final var rightBrace = consume(PbsTokenKind.RIGHT_BRACE, "Expected '}' to end error body");
return new PbsAst.ErrorDecl(name.lexeme(), ReadOnlyList.wrap(cases), span(declareToken.start(), rightBrace.end()));
return new PbsAst.ErrorDecl(name.lexeme(), ReadOnlyList.wrap(cases), attributes, span(declareToken.start(), rightBrace.end()));
}
private PbsAst.EnumDecl parseEnumDeclaration(final PbsToken declareToken) {
private PbsAst.EnumDecl parseEnumDeclaration(
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> attributes) {
final var name = consume(PbsTokenKind.IDENTIFIER, "Expected enum name");
consume(PbsTokenKind.LEFT_PAREN, "Expected '(' after enum name");
@@ -522,7 +550,7 @@ final class PbsDeclarationParser {
final var rightParen = consume(PbsTokenKind.RIGHT_PAREN, "Expected ')' after enum cases");
final var semicolon = consume(PbsTokenKind.SEMICOLON, "Expected ';' after enum declaration");
return new PbsAst.EnumDecl(name.lexeme(), ReadOnlyList.wrap(cases), span(declareToken.start(), Math.max(rightParen.end(), semicolon.end())));
return new PbsAst.EnumDecl(name.lexeme(), ReadOnlyList.wrap(cases), attributes, span(declareToken.start(), Math.max(rightParen.end(), semicolon.end())));
}
private PbsAst.FunctionDecl parseFunctionLike(
@@ -543,6 +571,7 @@ final class PbsDeclarationParser {
returnSpec.kind(),
returnSpec.returnType(),
returnSpec.resultErrorType(),
pendingAttributes,
lifecycleMarker,
body,
span(fnToken.start(), body.span().getEnd()));
@@ -585,7 +614,9 @@ final class PbsDeclarationParser {
return token;
}
private PbsAst.CallbackDecl parseCallbackDeclaration(final PbsToken declareToken) {
private PbsAst.CallbackDecl parseCallbackDeclaration(
final PbsToken declareToken,
final ReadOnlyList<PbsAst.Attribute> attributes) {
final var name = consume(PbsTokenKind.IDENTIFIER, "Expected callback name");
consume(PbsTokenKind.LEFT_PAREN, "Expected '(' after callback name");
final var parameters = typeParser.parseParametersUntilRightParen();
@@ -600,6 +631,7 @@ final class PbsDeclarationParser {
returnSpec.kind(),
returnSpec.returnType(),
returnSpec.resultErrorType(),
attributes,
span(declareToken.start(), semicolon.end()));
}
@@ -651,6 +683,20 @@ final class PbsDeclarationParser {
reportAttributesNotAllowed(attributes, "Attributes are not allowed before " + targetSurface);
}
private void rejectUnsupportedDeclarationAttributes(
final ReadOnlyList<PbsAst.Attribute> attributes,
final String targetSurface) {
for (final var attribute : attributes) {
if ("Doc".equals(attribute.name())) {
continue;
}
p.studio.compiler.source.diagnostics.Diagnostics.error(context.diagnostics(),
ParseErrors.E_PARSE_ATTRIBUTES_NOT_ALLOWED.name(),
"Only [Doc] is allowed before " + targetSurface,
attribute.span());
}
}
private void reportAttributesNotAllowed(
final ReadOnlyList<PbsAst.Attribute> attributes,
final String message) {
@@ -676,14 +722,16 @@ final class PbsDeclarationParser {
return PbsAst.LifecycleMarker.NONE;
}
if (!isOrdinaryMode()) {
reportAttributesNotAllowed(attributes, "Attributes are not allowed before top-level functions");
return PbsAst.LifecycleMarker.NONE;
}
PbsAst.LifecycleMarker marker = PbsAst.LifecycleMarker.NONE;
for (final var attribute : attributes) {
final PbsAst.LifecycleMarker nextMarker;
if ("Doc".equals(attribute.name())) {
continue;
}
if (!isOrdinaryMode()) {
reportAttributesNotAllowed(attributes, "Only [Doc] is allowed before interface executable methods");
return PbsAst.LifecycleMarker.NONE;
}
if ("Init".equals(attribute.name())) {
nextMarker = PbsAst.LifecycleMarker.INIT;
} else if ("Frame".equals(attribute.name())) {
@@ -1,6 +1,7 @@
package p.studio.compiler.pbs.semantics;
import p.studio.compiler.models.SourceKind;
import p.studio.compiler.pbs.PbsDocumentationTextBlockNormalizer;
import p.studio.compiler.pbs.PbsFrontendCompiler;
import p.studio.compiler.pbs.ast.PbsAst;
import p.studio.compiler.source.Span;
@@ -19,6 +20,7 @@ public final class PbsDeclarationSemanticsValidator {
private static final String ATTR_INTRINSIC_CALL = "IntrinsicCall";
private static final String ATTR_INIT_ALLOWED = "InitAllowed";
private static final String ATTR_ASSET_LOWERING = "AssetLowering";
private static final String ATTR_DOC = "Doc";
private static final Set<String> RESERVED_ATTRIBUTES = Set.of(
ATTR_HOST,
@@ -27,7 +29,8 @@ public final class PbsDeclarationSemanticsValidator {
ATTR_BUILTIN_CONST,
ATTR_INTRINSIC_CALL,
ATTR_INIT_ALLOWED,
ATTR_ASSET_LOWERING);
ATTR_ASSET_LOWERING,
ATTR_DOC);
private final NameTable nameTable;
private final PbsConstSemanticsValidator constSemanticsValidator = new PbsConstSemanticsValidator();
@@ -60,6 +63,7 @@ public final class PbsDeclarationSemanticsValidator {
for (final var topDecl : ast.topDecls()) {
if (topDecl instanceof PbsAst.FunctionDecl functionDecl) {
validateDocAttributes(functionDecl.attributes(), "function declarations", diagnostics);
if (interfaceModule) {
reportInterfaceNonDeclarativeDecl(
functionDecl.span(),
@@ -82,6 +86,7 @@ public final class PbsDeclarationSemanticsValidator {
}
if (topDecl instanceof PbsAst.StructDecl structDecl) {
validateDocAttributes(structDecl.attributes(), "struct declarations", diagnostics);
binder.registerType(structDecl.name(), structDecl.span(), "struct");
if (interfaceModule && (structDecl.hasBody() || !structDecl.methods().isEmpty() || !structDecl.ctors().isEmpty())) {
reportInterfaceNonDeclarativeDecl(
@@ -94,6 +99,7 @@ public final class PbsDeclarationSemanticsValidator {
}
if (topDecl instanceof PbsAst.ServiceDecl serviceDecl) {
validateDocAttributes(serviceDecl.attributes(), "service declarations", diagnostics);
binder.registerType(serviceDecl.name(), serviceDecl.span(), "service");
binder.registerValue(serviceDecl.name(), serviceDecl.span(), "service singleton");
binder.registerVisibleTopLevelSurface(serviceDecl.name(), serviceDecl.span(), "service");
@@ -102,10 +108,12 @@ public final class PbsDeclarationSemanticsValidator {
}
if (topDecl instanceof PbsAst.ContractDecl contractDecl) {
validateDocAttributes(contractDecl.attributes(), "contract declarations", diagnostics);
binder.registerType(contractDecl.name(), contractDecl.span(), "contract");
if (interfaceModule) {
for (final var signature : contractDecl.signatures()) {
validateReservedAttributeTarget(signature.attributes(), "contract signature", diagnostics);
validateDocAttributes(signature.attributes(), "contract signatures", diagnostics);
validateReservedAttributeTargetExceptDoc(signature.attributes(), "contract signature", diagnostics);
}
}
validateContractDeclaration(contractDecl, binder, rules);
@@ -113,6 +121,7 @@ public final class PbsDeclarationSemanticsValidator {
}
if (topDecl instanceof PbsAst.HostDecl hostDecl) {
validateDocAttributes(hostDecl.attributes(), "host declarations", diagnostics);
binder.registerHostOwner(hostDecl.name(), hostDecl.span(), "host owner");
validateHostDeclaration(hostDecl, binder, rules, interfaceModule, diagnostics);
continue;
@@ -125,18 +134,21 @@ public final class PbsDeclarationSemanticsValidator {
}
if (topDecl instanceof PbsAst.ErrorDecl errorDecl) {
validateDocAttributes(errorDecl.attributes(), "error declarations", diagnostics);
binder.registerType(errorDecl.name(), errorDecl.span(), "error");
rules.validateErrorDeclaration(errorDecl);
continue;
}
if (topDecl instanceof PbsAst.EnumDecl enumDecl) {
validateDocAttributes(enumDecl.attributes(), "enum declarations", diagnostics);
binder.registerType(enumDecl.name(), enumDecl.span(), "enum");
rules.validateEnumDeclaration(enumDecl);
continue;
}
if (topDecl instanceof PbsAst.CallbackDecl callbackDecl) {
validateDocAttributes(callbackDecl.attributes(), "callback declarations", diagnostics);
binder.registerType(callbackDecl.name(), callbackDecl.span(), "callback");
rules.validateCallbackDeclaration(callbackDecl);
continue;
@@ -187,6 +199,7 @@ public final class PbsDeclarationSemanticsValidator {
false);
final var methodScope = PbsCallableScope.structMethods(ownerNameId);
for (final var method : structDecl.methods()) {
validateDocAttributes(method.attributes(), "struct methods", diagnostics);
validateCallable(
methodScope,
method.name(),
@@ -226,6 +239,7 @@ public final class PbsDeclarationSemanticsValidator {
final DiagnosticSink diagnostics) {
final var methodScope = PbsCallableScope.serviceMethods(nameTable.register(serviceDecl.name()));
for (final var method : serviceDecl.methods()) {
validateDocAttributes(method.attributes(), "service methods", diagnostics);
validateCallable(
methodScope,
method.name(),
@@ -494,12 +508,13 @@ public final class PbsDeclarationSemanticsValidator {
final PbsAst.ConstDecl constDecl,
final boolean interfaceModule,
final DiagnosticSink diagnostics) {
validateDocAttributes(constDecl.attributes(), "const declarations", diagnostics);
final var builtinConstAttributes = attributesNamed(constDecl.attributes(), ATTR_BUILTIN_CONST);
for (final var attribute : constDecl.attributes()) {
if (!isReservedAttribute(attribute.name())) {
continue;
}
if (ATTR_BUILTIN_CONST.equals(attribute.name())) {
if (ATTR_BUILTIN_CONST.equals(attribute.name()) || ATTR_DOC.equals(attribute.name())) {
continue;
}
reportInvalidReservedAttributeTarget(
@@ -540,6 +555,7 @@ public final class PbsDeclarationSemanticsValidator {
private void validateHostSignatureAttributes(
final PbsAst.FunctionSignature signature,
final DiagnosticSink diagnostics) {
validateDocAttributes(signature.attributes(), "host signatures", diagnostics);
final var hostAttributes = attributesNamed(signature.attributes(), ATTR_HOST);
final var initAllowedAttributes = attributesNamed(signature.attributes(), ATTR_INIT_ALLOWED);
final var assetLoweringAttributes = attributesNamed(signature.attributes(), ATTR_ASSET_LOWERING);
@@ -550,7 +566,8 @@ public final class PbsDeclarationSemanticsValidator {
if (ATTR_HOST.equals(attribute.name())
|| ATTR_CAPABILITY.equals(attribute.name())
|| ATTR_INIT_ALLOWED.equals(attribute.name())
|| ATTR_ASSET_LOWERING.equals(attribute.name())) {
|| ATTR_ASSET_LOWERING.equals(attribute.name())
|| ATTR_DOC.equals(attribute.name())) {
continue;
}
reportInvalidReservedAttributeTarget(
@@ -597,12 +614,13 @@ public final class PbsDeclarationSemanticsValidator {
private void validateBuiltinTypeAttribute(
final PbsAst.BuiltinTypeDecl builtinTypeDecl,
final DiagnosticSink diagnostics) {
validateDocAttributes(builtinTypeDecl.attributes(), "builtin type declarations", diagnostics);
final var builtinTypeAttributes = attributesNamed(builtinTypeDecl.attributes(), ATTR_BUILTIN_TYPE);
for (final var attribute : builtinTypeDecl.attributes()) {
if (!isReservedAttribute(attribute.name())) {
continue;
}
if (ATTR_BUILTIN_TYPE.equals(attribute.name())) {
if (ATTR_BUILTIN_TYPE.equals(attribute.name()) || ATTR_DOC.equals(attribute.name())) {
continue;
}
reportInvalidReservedAttributeTarget(
@@ -645,12 +663,13 @@ public final class PbsDeclarationSemanticsValidator {
private void validateIntrinsicCallAttributes(
final PbsAst.FunctionSignature signature,
final DiagnosticSink diagnostics) {
validateDocAttributes(signature.attributes(), "builtin method signatures", diagnostics);
final var intrinsicAttributes = attributesNamed(signature.attributes(), ATTR_INTRINSIC_CALL);
for (final var attribute : signature.attributes()) {
if (!isReservedAttribute(attribute.name())) {
continue;
}
if (ATTR_INTRINSIC_CALL.equals(attribute.name())) {
if (ATTR_INTRINSIC_CALL.equals(attribute.name()) || ATTR_DOC.equals(attribute.name())) {
continue;
}
reportInvalidReservedAttributeTarget(
@@ -770,6 +789,47 @@ public final class PbsDeclarationSemanticsValidator {
}
}
private void validateDocAttributes(
final ReadOnlyList<PbsAst.Attribute> attributes,
final String targetSurface,
final DiagnosticSink diagnostics) {
final var docAttributes = attributesNamed(attributes, ATTR_DOC);
if (docAttributes.isEmpty()) {
return;
}
if (docAttributes.size() > 1) {
for (int i = 1; i < docAttributes.size(); i++) {
reportDuplicateReservedAttribute(docAttributes.get(i), ATTR_DOC, diagnostics);
}
}
validateDocAttributeShape(docAttributes.getFirst(), targetSurface, diagnostics);
}
private void validateDocAttributeShape(
final PbsAst.Attribute attribute,
final String targetSurface,
final DiagnosticSink diagnostics) {
final var args = validateNamedArguments(attribute, Set.of("markdown"), Set.of(), diagnostics);
if (args == null) {
return;
}
final var value = args.get("markdown");
if (!(value instanceof PbsAst.AttributeDocTextBlockValue docValue)) {
p.studio.compiler.source.diagnostics.Diagnostics.error(diagnostics,
PbsSemanticsErrors.E_SEM_MALFORMED_RESERVED_ATTRIBUTE.name(),
"Doc.markdown on %s must be a documentation text block".formatted(targetSurface),
attribute.span());
return;
}
final var normalized = PbsDocumentationTextBlockNormalizer.normalizeLexeme(docValue.value());
if (normalized.isBlank()) {
p.studio.compiler.source.diagnostics.Diagnostics.error(diagnostics,
PbsSemanticsErrors.E_SEM_MALFORMED_RESERVED_ATTRIBUTE.name(),
"Doc.markdown on %s must not be empty".formatted(targetSurface),
attribute.span());
}
}
private Map<String, PbsAst.AttributeValue> validateNamedArguments(
final PbsAst.Attribute attribute,
final Set<String> requiredNames,
@@ -911,6 +971,21 @@ public final class PbsDeclarationSemanticsValidator {
}
}
private void validateReservedAttributeTargetExceptDoc(
final ReadOnlyList<PbsAst.Attribute> attributes,
final String targetSurface,
final DiagnosticSink diagnostics) {
for (final var attribute : attributes) {
if (!isReservedAttribute(attribute.name()) || ATTR_DOC.equals(attribute.name())) {
continue;
}
reportInvalidReservedAttributeTarget(
attribute,
"Attribute '%s' is not valid on %s".formatted(attribute.name(), targetSurface),
diagnostics);
}
}
private void reportInvalidReservedAttributeTarget(
final PbsAst.Attribute attribute,
final String message,
@@ -6,12 +6,22 @@ public record PbsEditorialCompletionCandidate(
String label,
PbsEditorialSymbolKind kind,
String detail,
String origin) {
String origin,
String documentation) {
public PbsEditorialCompletionCandidate(
final String label,
final PbsEditorialSymbolKind kind,
final String detail,
final String origin) {
this(label, kind, detail, origin, "");
}
public PbsEditorialCompletionCandidate {
label = requireText(label, "label");
kind = Objects.requireNonNull(kind, "kind");
detail = normalize(detail);
origin = normalize(origin);
documentation = normalize(documentation);
}
private static String requireText(
@@ -482,10 +482,11 @@ public final class PbsEditorialSupportService {
}
for (final var method : structInfo.methods().entrySet()) {
completions.put(method.getKey(), new PbsEditorialCompletionCandidate(
method.getKey(),
PbsEditorialSymbolKind.METHOD,
formatCallableDetails(method.getValue()),
receiverType.name()));
method.getKey(),
PbsEditorialSymbolKind.METHOD,
formatCallableDetails(method.getValue()),
receiverType.name(),
documentationFromCallables(method.getValue())));
}
}
} else if (receiverType.kind() == Kind.SERVICE || receiverType.kind() == Kind.CONTRACT) {
@@ -494,10 +495,11 @@ public final class PbsEditorialSupportService {
: Optional.ofNullable(model.contracts.get(receiverType.name())).map(ContractInfo::methods).orElse(Map.of());
for (final var method : methods.entrySet()) {
completions.put(method.getKey(), new PbsEditorialCompletionCandidate(
method.getKey(),
PbsEditorialSymbolKind.METHOD,
formatCallableDetails(method.getValue()),
receiverType.name()));
method.getKey(),
PbsEditorialSymbolKind.METHOD,
formatCallableDetails(method.getValue()),
receiverType.name(),
documentationFromCallables(method.getValue())));
}
}
return List.copyOf(completions.values());
@@ -587,7 +589,7 @@ public final class PbsEditorialSupportService {
formatCallableDetails(methods),
receiverType.name(),
signaturesFromCallables(methods),
""));
documentationFromCallables(methods)));
}
}
if (receiverType.kind() == Kind.SERVICE || receiverType.kind() == Kind.CONTRACT) {
@@ -602,7 +604,7 @@ public final class PbsEditorialSupportService {
formatCallableDetails(signatures),
receiverType.name(),
signaturesFromCallables(signatures),
""));
documentationFromCallables(signatures)));
}
}
return Optional.empty();
@@ -663,7 +665,7 @@ public final class PbsEditorialSupportService {
detail,
topDeclSymbol.origin(),
signatures,
"");
documentationFromTopDecl(topDeclSymbol.decl()));
}
private PbsEditorialResolvedSymbol constructorSymbolFromTopDecl(
@@ -678,7 +680,68 @@ public final class PbsEditorialSupportService {
detail,
topDeclSymbol.origin(),
signatures,
"");
documentationFromTopDecl(topDeclSymbol.decl()));
}
private String documentationFromCallables(final List<CallableSymbol> callables) {
for (final var callable : callables) {
if (!callable.documentation().isBlank()) {
return callable.documentation();
}
}
return "";
}
private String documentationFromTopDecl(final PbsAst.TopDecl topDecl) {
return documentationFromAttributes(attributesOf(topDecl));
}
private ReadOnlyList<PbsAst.Attribute> attributesOf(final PbsAst.TopDecl topDecl) {
if (topDecl instanceof PbsAst.FunctionDecl functionDecl) {
return functionDecl.attributes();
}
if (topDecl instanceof PbsAst.StructDecl structDecl) {
return structDecl.attributes();
}
if (topDecl instanceof PbsAst.BuiltinTypeDecl builtinTypeDecl) {
return builtinTypeDecl.attributes();
}
if (topDecl instanceof PbsAst.ServiceDecl serviceDecl) {
return serviceDecl.attributes();
}
if (topDecl instanceof PbsAst.HostDecl hostDecl) {
return hostDecl.attributes();
}
if (topDecl instanceof PbsAst.ContractDecl contractDecl) {
return contractDecl.attributes();
}
if (topDecl instanceof PbsAst.CallbackDecl callbackDecl) {
return callbackDecl.attributes();
}
if (topDecl instanceof PbsAst.EnumDecl enumDecl) {
return enumDecl.attributes();
}
if (topDecl instanceof PbsAst.ErrorDecl errorDecl) {
return errorDecl.attributes();
}
if (topDecl instanceof PbsAst.ConstDecl constDecl) {
return constDecl.attributes();
}
return ReadOnlyList.empty();
}
private String documentationFromAttributes(final ReadOnlyList<PbsAst.Attribute> attributes) {
for (final var attribute : attributes) {
if (!"Doc".equals(attribute.name()) || attribute.arguments().isEmpty()) {
continue;
}
final var argument = attribute.arguments().getFirst();
if (!"markdown".equals(argument.name()) || !(argument.value() instanceof PbsAst.AttributeDocTextBlockValue docValue)) {
continue;
}
return p.studio.compiler.pbs.PbsDocumentationTextBlockNormalizer.normalizeLexeme(docValue.value());
}
return "";
}
private PbsEditorialCompletionCandidate completionFromTopDecl(final TopDeclSymbol topDeclSymbol) {
@@ -690,7 +753,8 @@ public final class PbsEditorialSupportService {
topDeclSymbol.localName(),
symbolKind(topDeclSymbol.decl()),
detail,
topDeclSymbol.origin());
topDeclSymbol.origin(),
documentationFromTopDecl(topDeclSymbol.decl()));
}
private List<PbsEditorialSignature> signaturesForTopDecl(final PbsAst.TopDecl topDecl) {
@@ -1155,6 +1219,7 @@ public final class PbsEditorialSupportService {
functionDecl.returnKind(),
functionDecl.returnType(),
functionDecl.resultErrorType(),
functionDecl.attributes(),
functionDecl.lifecycleMarker(),
functionDecl.body(),
functionDecl.span());
@@ -1166,6 +1231,7 @@ public final class PbsEditorialSupportService {
structDecl.methods(),
structDecl.ctors(),
structDecl.hasBody(),
structDecl.attributes(),
structDecl.span());
}
if (topDecl instanceof PbsAst.BuiltinTypeDecl builtinTypeDecl) {
@@ -1177,13 +1243,13 @@ public final class PbsEditorialSupportService {
builtinTypeDecl.span());
}
if (topDecl instanceof PbsAst.ServiceDecl serviceDecl) {
return new PbsAst.ServiceDecl(newName, serviceDecl.methods(), serviceDecl.span());
return new PbsAst.ServiceDecl(newName, serviceDecl.methods(), serviceDecl.attributes(), serviceDecl.span());
}
if (topDecl instanceof PbsAst.HostDecl hostDecl) {
return new PbsAst.HostDecl(newName, hostDecl.signatures(), hostDecl.span());
return new PbsAst.HostDecl(newName, hostDecl.signatures(), hostDecl.attributes(), hostDecl.span());
}
if (topDecl instanceof PbsAst.ContractDecl contractDecl) {
return new PbsAst.ContractDecl(newName, contractDecl.signatures(), contractDecl.span());
return new PbsAst.ContractDecl(newName, contractDecl.signatures(), contractDecl.attributes(), contractDecl.span());
}
if (topDecl instanceof PbsAst.CallbackDecl callbackDecl) {
return new PbsAst.CallbackDecl(
@@ -1192,13 +1258,14 @@ public final class PbsEditorialSupportService {
callbackDecl.returnKind(),
callbackDecl.returnType(),
callbackDecl.resultErrorType(),
callbackDecl.attributes(),
callbackDecl.span());
}
if (topDecl instanceof PbsAst.EnumDecl enumDecl) {
return new PbsAst.EnumDecl(newName, enumDecl.cases(), enumDecl.span());
return new PbsAst.EnumDecl(newName, enumDecl.cases(), enumDecl.attributes(), enumDecl.span());
}
if (topDecl instanceof PbsAst.ErrorDecl errorDecl) {
return new PbsAst.ErrorDecl(newName, errorDecl.cases(), errorDecl.span());
return new PbsAst.ErrorDecl(newName, errorDecl.cases(), errorDecl.attributes(), errorDecl.span());
}
if (topDecl instanceof PbsAst.GlobalDecl globalDecl) {
return new PbsAst.GlobalDecl(newName, globalDecl.explicitType(), globalDecl.initializer(), globalDecl.span());
@@ -146,7 +146,8 @@ final class PbsFlowSemanticSupport {
List<CallableParameter> parameters,
List<TypeView> inputTypes,
TypeView outputType,
Span span) {
Span span,
String documentation) {
}
record CallableParameter(
@@ -282,7 +283,8 @@ final class PbsFlowSemanticSupport {
method.returnKind(),
method.returnType(),
method.resultErrorType(),
method.span()));
method.span(),
documentationFromAttributes(method.attributes())));
}
structs.put(structDecl.name(), new StructInfo(fields, methods));
return;
@@ -306,7 +308,8 @@ final class PbsFlowSemanticSupport {
signature.returnKind(),
signature.returnType(),
signature.resultErrorType(),
signature.span()));
signature.span(),
documentationFromAttributes(signature.attributes())));
}
structs.put(builtinTypeDecl.name(), new StructInfo(fields, methods));
return;
@@ -321,7 +324,8 @@ final class PbsFlowSemanticSupport {
method.returnKind(),
method.returnType(),
method.resultErrorType(),
method.span()));
method.span(),
documentationFromAttributes(method.attributes())));
}
services.put(serviceDecl.name(), new ServiceInfo(methods));
serviceSingletons.put(serviceDecl.name(), TypeView.service(serviceDecl.name()));
@@ -337,7 +341,8 @@ final class PbsFlowSemanticSupport {
signature.returnKind(),
signature.returnType(),
signature.resultErrorType(),
signature.span()));
signature.span(),
documentationFromAttributes(signature.attributes())));
}
// Host owners are value singletons with callable members, same access shape as services.
services.put(hostDecl.name(), new ServiceInfo(methods));
@@ -354,7 +359,8 @@ final class PbsFlowSemanticSupport {
signature.returnKind(),
signature.returnType(),
signature.resultErrorType(),
signature.span()));
signature.span(),
documentationFromAttributes(signature.attributes())));
}
contracts.put(contractDecl.name(), new ContractInfo(methods));
return;
@@ -367,12 +373,14 @@ final class PbsFlowSemanticSupport {
functionDecl.returnKind(),
functionDecl.returnType(),
functionDecl.resultErrorType(),
functionDecl.span()));
functionDecl.span(),
documentationFromAttributes(functionDecl.attributes())));
return;
}
if (topDecl instanceof PbsAst.CallbackDecl(
String name, ReadOnlyList<PbsAst.Parameter> parameters, PbsAst.ReturnKind returnKind,
PbsAst.TypeRef returnType, PbsAst.TypeRef resultErrorType, Span span
PbsAst.TypeRef returnType, PbsAst.TypeRef resultErrorType,
ReadOnlyList<PbsAst.Attribute> ignoredAttributes, Span span
)) {
final var symbol = callableFrom(
name,
@@ -380,7 +388,8 @@ final class PbsFlowSemanticSupport {
returnKind,
returnType,
resultErrorType,
span);
span,
"");
callbacks.put(name, new CallbackSignature(symbol.inputTypes(), symbol.outputType()));
return;
}
@@ -477,7 +486,8 @@ final class PbsFlowSemanticSupport {
final PbsAst.ReturnKind returnKind,
final PbsAst.TypeRef returnType,
final PbsAst.TypeRef resultErrorType,
final Span span) {
final Span span,
final String documentation) {
final var input = new ArrayList<TypeView>(parameters.size());
final var semanticParameters = new ArrayList<CallableParameter>(parameters.size());
for (int index = 0; index < parameters.size(); index += 1) {
@@ -490,7 +500,21 @@ final class PbsFlowSemanticSupport {
type,
parameter.span()));
}
return new CallableSymbol(name, semanticParameters, input, callableReturn(returnKind, returnType, resultErrorType), span);
return new CallableSymbol(name, semanticParameters, input, callableReturn(returnKind, returnType, resultErrorType), span, documentation);
}
private String documentationFromAttributes(final ReadOnlyList<PbsAst.Attribute> attributes) {
for (final var attribute : attributes) {
if (!"Doc".equals(attribute.name()) || attribute.arguments().isEmpty()) {
continue;
}
final var argument = attribute.arguments().getFirst();
if (!"markdown".equals(argument.name()) || !(argument.value() instanceof PbsAst.AttributeDocTextBlockValue docValue)) {
continue;
}
return p.studio.compiler.pbs.PbsDocumentationTextBlockNormalizer.normalizeLexeme(docValue.value());
}
return "";
}
private TypeView callableReturn(
@@ -5,24 +5,15 @@ import p.studio.compiler.messages.BuildingIssueSink;
import p.studio.compiler.messages.FESurfaceContext;
import p.studio.compiler.messages.FrontendPhaseContext;
import p.studio.compiler.models.IRBackend;
import p.studio.compiler.models.IRBackendExecutableFunction;
import p.studio.compiler.models.IRGlobal;
import p.studio.compiler.models.IRGlobalOrigin;
import p.studio.compiler.models.IRGlobalVisibility;
import p.studio.compiler.models.IRHiddenGlobalKind;
import p.studio.compiler.models.IRSyntheticCallableKind;
import p.studio.compiler.pbs.PbsFrontendCompiler;
import p.studio.compiler.pbs.PbsReservedMetadataExtractor;
import p.studio.compiler.pbs.semantics.PbsFlowSemanticsValidator;
import p.studio.compiler.pbs.stdlib.InterfaceModuleLoader;
import p.studio.compiler.pbs.stdlib.ResourceStdlibEnvironmentResolver;
import p.studio.compiler.pbs.stdlib.StdlibEnvironmentResolver;
import p.studio.compiler.source.Span;
import p.studio.compiler.source.diagnostics.DiagnosticSink;
import p.studio.compiler.source.identifiers.CallableId;
import p.studio.compiler.source.identifiers.FileId;
import p.studio.compiler.source.identifiers.ModuleId;
import p.studio.compiler.source.tables.CallableSignatureRef;
import p.studio.utilities.logs.LogAggregator;
import p.studio.utilities.structures.ReadOnlyList;
@@ -168,7 +159,7 @@ public class PBSFrontendPhaseService implements FrontendPhaseService {
private IRBackend mergeCompiledSources(
final ArrayList<CompiledSourceFile> compiledSourceFiles,
final Set<ModuleId> failedModuleIds,
final Map<ModuleId, Set<ModuleId>> moduleDependencyGraph) {
final Map<ModuleId, ReadOnlyList<ModuleId>> moduleDependencyGraph) {
final var irBackendAggregator = IRBackend.aggregator();
final var blockedModuleIds = blockedModulesByDependency(failedModuleIds, moduleDependencyGraph);
@@ -178,438 +169,70 @@ public class PBSFrontendPhaseService implements FrontendPhaseService {
}
irBackendAggregator.merge(compiledSource.irBackendFile());
}
return synthesizePublishedLifecycle(irBackendAggregator.emit());
irBackendAggregator.lifecycleModuleOrder(lifecycleModuleOrder(
compiledSourceFiles,
blockedModuleIds,
moduleDependencyGraph));
return irBackendAggregator.emit();
}
private IRBackend synthesizePublishedLifecycle(final IRBackend baseBackend) {
final var callableSignatures = new ArrayList<>(baseBackend.getCallableSignatures().asList());
final var executableFunctions = new ArrayList<>(baseBackend.getExecutableFunctions().asList());
final var globals = new ArrayList<>(baseBackend.getGlobals().asList());
final var fileInitFunctionsByModule = new LinkedHashMap<ModuleId, List<IRBackendExecutableFunction>>();
final var projectInitAnchorsByModule = new LinkedHashMap<ModuleId, String>();
final var wrapperNamesByModule = new LinkedHashMap<ModuleId, String>();
final var frameAnchorsByModule = new LinkedHashMap<ModuleId, String>();
for (final var executable : executableFunctions) {
if (!executable.callableName().startsWith("__pbs.file_init$")) {
continue;
}
fileInitFunctionsByModule
.computeIfAbsent(normalizeModuleId(executable.moduleId()), ignored -> new ArrayList<>())
.add(executable);
}
for (final var syntheticFunction : baseBackend.getSyntheticFunctions()) {
final var moduleId = normalizeModuleId(syntheticFunction.moduleId());
switch (syntheticFunction.kind()) {
case PROJECT_INIT -> projectInitAnchorsByModule.put(moduleId, syntheticFunction.origin().anchorCallableName());
case PUBLISHED_FRAME_WRAPPER -> {
wrapperNamesByModule.put(moduleId, syntheticFunction.callableName());
frameAnchorsByModule.put(moduleId, syntheticFunction.origin().anchorCallableName());
}
default -> {
}
private ReadOnlyList<ModuleId> lifecycleModuleOrder(
final ArrayList<CompiledSourceFile> compiledSourceFiles,
final Set<ModuleId> blockedModuleIds,
final Map<ModuleId, ReadOnlyList<ModuleId>> dependenciesByModule) {
final var availableModules = new LinkedHashSet<ModuleId>();
for (final var compiledSource : compiledSourceFiles) {
if (!blockedModuleIds.contains(compiledSource.moduleId())) {
availableModules.add(compiledSource.moduleId());
}
}
final ModuleId entryPointModuleId;
final String entryPointCallableName;
if (wrapperNamesByModule.size() == 1) {
entryPointModuleId = wrapperNamesByModule.keySet().iterator().next();
entryPointCallableName = wrapperNamesByModule.get(entryPointModuleId);
} else {
entryPointModuleId = baseBackend.getEntryPointModuleId();
entryPointCallableName = baseBackend.getEntryPointCallableName();
}
final var sortedModuleIds = new ArrayList<ModuleId>();
sortedModuleIds.addAll(fileInitFunctionsByModule.keySet());
for (final var moduleId : wrapperNamesByModule.keySet()) {
if (!sortedModuleIds.contains(moduleId)) {
sortedModuleIds.add(moduleId);
}
}
sortedModuleIds.sort(Comparator.comparing(moduleId -> moduleSortKey(baseBackend, moduleId)));
final var moduleInitCallableIds = new LinkedHashMap<ModuleId, CallableId>();
for (final var moduleId : sortedModuleIds) {
final var moduleInitName = PbsFrontendCompiler.moduleInitCallableName(moduleId);
final var fileInitFunctions = new ArrayList<>(fileInitFunctionsByModule.getOrDefault(moduleId, List.of()));
if (fileInitFunctions.isEmpty()) {
continue;
}
fileInitFunctions.sort(Comparator.comparingInt(fn -> fn.fileId().getId()));
final var callableId = new CallableId(callableSignatures.size());
callableSignatures.add(new CallableSignatureRef(moduleId, moduleInitName, 0, "() -> unit"));
moduleInitCallableIds.put(moduleId, callableId);
final var instructions = new ArrayList<IRBackendExecutableFunction.Instruction>();
var anchorSpan = fileInitFunctions.getFirst().span();
var anchorFileId = fileInitFunctions.getFirst().fileId();
for (final var fileInitFunction : fileInitFunctions) {
instructions.add(callInstruction(fileInitFunction, fileInitFunction.span()));
anchorSpan = fileInitFunction.span();
anchorFileId = fileInitFunction.fileId();
}
instructions.add(retInstruction(anchorSpan));
executableFunctions.add(new IRBackendExecutableFunction(
anchorFileId,
final var ordered = new ArrayList<ModuleId>(availableModules.size());
final var visiting = new HashSet<ModuleId>();
final var visited = new HashSet<ModuleId>();
for (final var moduleId : availableModules) {
appendDependenciesFirst(
moduleId,
moduleInitName,
callableId,
safeToInt(anchorSpan.getStart()),
safeToInt(anchorSpan.getEnd()),
0,
0,
0,
1,
ReadOnlyList.wrap(instructions),
anchorSpan));
availableModules,
dependenciesByModule,
visiting,
visited,
ordered);
}
CallableId projectInitCallableId = null;
if (!entryPointModuleId.isNone()) {
final var projectInitAnchor = projectInitAnchorsByModule.get(entryPointModuleId);
final var projectInitTarget = resolveExecutable(executableFunctions, entryPointModuleId, projectInitAnchor);
if (projectInitTarget != null) {
final var projectInitName = PbsFrontendCompiler.projectInitCallableName(entryPointModuleId);
projectInitCallableId = new CallableId(callableSignatures.size());
callableSignatures.add(new CallableSignatureRef(entryPointModuleId, projectInitName, 0, "() -> unit"));
executableFunctions.add(new IRBackendExecutableFunction(
projectInitTarget.fileId(),
entryPointModuleId,
projectInitName,
projectInitCallableId,
projectInitTarget.sourceStart(),
projectInitTarget.sourceEnd(),
0,
0,
0,
1,
ReadOnlyList.from(
callInstruction(projectInitTarget, projectInitTarget.span()),
retInstruction(projectInitTarget.span())),
projectInitTarget.span()));
}
}
final var frameExecutable = resolveExecutable(
executableFunctions,
entryPointModuleId,
frameAnchorsByModule.get(entryPointModuleId));
if (frameExecutable != null
&& entryPointCallableName != null
&& !entryPointCallableName.isBlank()
&& !entryPointModuleId.isNone()) {
final var existingBootGuard = globals.stream()
.filter(global -> moduleIdsMatch(entryPointModuleId, global.moduleId()))
.filter(global -> global.hiddenKind() == IRHiddenGlobalKind.BOOT_GUARD)
.findFirst()
.orElse(null);
final var bootGuardSlot = existingBootGuard != null
? existingBootGuard.slot()
: nextGlobalSlot(globals, entryPointModuleId);
final var wrapperCallableId = new CallableId(callableSignatures.size());
callableSignatures.add(new CallableSignatureRef(entryPointModuleId, entryPointCallableName, 0, "() -> unit"));
final var instructions = new ArrayList<IRBackendExecutableFunction.Instruction>();
instructions.add(getGlobalInstruction(bootGuardSlot, frameExecutable.span()));
instructions.add(pushBoolInstruction(true, frameExecutable.span()));
instructions.add(eqInstruction(frameExecutable.span()));
instructions.add(jumpIfTrueInstruction("bootstrap_done", frameExecutable.span()));
for (final var moduleId : sortedModuleIds) {
final var moduleInitCallableId = moduleInitCallableIds.get(moduleId);
if (moduleInitCallableId == null) {
continue;
}
instructions.add(new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.CALL_FUNC,
moduleId,
PbsFrontendCompiler.moduleInitCallableName(moduleId),
moduleInitCallableId,
null,
null,
0,
0,
frameExecutable.span()));
}
if (projectInitCallableId != null) {
instructions.add(new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.CALL_FUNC,
entryPointModuleId,
PbsFrontendCompiler.projectInitCallableName(entryPointModuleId),
projectInitCallableId,
null,
null,
0,
0,
frameExecutable.span()));
}
instructions.add(pushBoolInstruction(true, frameExecutable.span()));
instructions.add(setGlobalInstruction(bootGuardSlot, frameExecutable.span()));
instructions.add(labelInstruction("bootstrap_done", frameExecutable.span()));
instructions.add(callInstruction(frameExecutable, frameExecutable.span()));
instructions.add(retInstruction(frameExecutable.span()));
executableFunctions.add(new IRBackendExecutableFunction(
frameExecutable.fileId(),
entryPointModuleId,
entryPointCallableName,
wrapperCallableId,
frameExecutable.sourceStart(),
frameExecutable.sourceEnd(),
0,
0,
0,
2,
ReadOnlyList.wrap(instructions),
frameExecutable.span()));
if (existingBootGuard == null) {
globals.add(new IRGlobal(
frameExecutable.fileId(),
entryPointModuleId,
PbsFrontendCompiler.bootGuardGlobalName(entryPointModuleId),
"bool",
bootGuardSlot,
IRGlobalVisibility.MODULE,
true,
IRHiddenGlobalKind.BOOT_GUARD,
new IRGlobalOrigin(
frameExecutable.fileId(),
entryPointModuleId,
frameAnchorsByModule.getOrDefault(entryPointModuleId, frameExecutable.callableName()),
frameExecutable.span(),
"BOOT_GUARD"),
frameExecutable.span()));
}
}
final var normalizedGlobals = normalizeGlobalSlots(globals);
return IRBackend.builder()
.entryPointCallableName(entryPointCallableName)
.entryPointModuleId(entryPointModuleId)
.functions(baseBackend.getFunctions())
.syntheticFunctions(baseBackend.getSyntheticFunctions())
.globals(ReadOnlyList.wrap(normalizedGlobals))
.executableFunctions(ReadOnlyList.wrap(executableFunctions))
.modulePool(baseBackend.getModulePool())
.callableSignatures(ReadOnlyList.wrap(callableSignatures))
.intrinsicPool(baseBackend.getIntrinsicPool())
.reservedMetadata(baseBackend.getReservedMetadata())
.build();
return ReadOnlyList.wrap(ordered);
}
private int nextGlobalSlot(
final ArrayList<IRGlobal> globals,
final ModuleId moduleId) {
var nextSlot = 0;
for (final var global : globals) {
if (!moduleIdsMatch(moduleId, global.moduleId())) {
continue;
}
nextSlot = Math.max(nextSlot, global.slot() + 1);
}
return nextSlot;
}
private ArrayList<IRGlobal> normalizeGlobalSlots(final ArrayList<IRGlobal> globals) {
final var nextSlotByModule = new LinkedHashMap<ModuleId, Integer>();
final var normalized = new ArrayList<IRGlobal>(globals.size());
for (final var global : globals) {
final var moduleId = normalizeModuleId(global.moduleId());
final var slot = nextSlotByModule.getOrDefault(moduleId, 0);
nextSlotByModule.put(moduleId, slot + 1);
normalized.add(new IRGlobal(
global.fileId(),
moduleId,
global.name(),
global.declaredTypeSurface(),
slot,
global.visibility(),
global.isHidden(),
global.hiddenKind(),
global.origin(),
global.span()));
}
return normalized;
}
private IRBackendExecutableFunction resolveExecutable(
final ArrayList<IRBackendExecutableFunction> executableFunctions,
private void appendDependenciesFirst(
final ModuleId moduleId,
final String callableName) {
if (callableName == null || callableName.isBlank()) {
return null;
final Set<ModuleId> availableModules,
final Map<ModuleId, ReadOnlyList<ModuleId>> dependenciesByModule,
final Set<ModuleId> visiting,
final Set<ModuleId> visited,
final ArrayList<ModuleId> ordered) {
if (visited.contains(moduleId) || !availableModules.contains(moduleId)) {
return;
}
for (final var executable : executableFunctions) {
if (!callableName.equals(executable.callableName())) {
continue;
}
if (!moduleIdsMatch(moduleId, executable.moduleId())) {
continue;
}
return executable;
if (!visiting.add(moduleId)) {
return;
}
return null;
}
private IRBackendExecutableFunction.Instruction callInstruction(
final IRBackendExecutableFunction callee,
final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.CALL_FUNC,
callee.moduleId(),
callee.callableName(),
callee.callableId(),
null,
null,
callee.paramSlots(),
callee.returnSlots(),
span);
}
private IRBackendExecutableFunction.Instruction retInstruction(final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.RET,
"",
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction getGlobalInstruction(
final int slot,
final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.GET_GLOBAL,
"",
null,
null,
null,
slot,
null,
span);
}
private IRBackendExecutableFunction.Instruction setGlobalInstruction(
final int slot,
final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.SET_GLOBAL,
"",
null,
null,
null,
slot,
null,
span);
}
private IRBackendExecutableFunction.Instruction pushBoolInstruction(
final boolean value,
final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.PUSH_BOOL,
"",
null,
null,
null,
value ? "true" : "false",
null,
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction eqInstruction(final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.EQ,
"",
null,
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction labelInstruction(
final String label,
final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.LABEL,
"",
null,
null,
null,
label,
null,
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction jumpIfTrueInstruction(
final String targetLabel,
final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.JMP_IF_TRUE,
"",
null,
null,
null,
null,
targetLabel,
null,
null,
span);
}
private ModuleId normalizeModuleId(final ModuleId moduleId) {
return moduleId == null ? ModuleId.none() : moduleId;
}
private boolean moduleIdsMatch(
final ModuleId left,
final ModuleId right) {
final var normalizedLeft = normalizeModuleId(left);
final var normalizedRight = normalizeModuleId(right);
if (normalizedLeft.isNone() && normalizedRight.isNone()) {
return true;
for (final var dependency : dependenciesByModule.getOrDefault(moduleId, ReadOnlyList.empty())) {
appendDependenciesFirst(
dependency,
availableModules,
dependenciesByModule,
visiting,
visited,
ordered);
}
return !normalizedLeft.isNone()
&& !normalizedRight.isNone()
&& normalizedLeft.getIndex() == normalizedRight.getIndex();
}
private String moduleSortKey(
final IRBackend backend,
final ModuleId moduleId) {
if (moduleId != null && !moduleId.isNone()) {
final var index = moduleId.getIndex();
if (index >= 0 && index < backend.getModulePool().size()) {
final var moduleReference = backend.getModulePool().get(index);
final var joiner = new StringJoiner("/", moduleReference.project() + ":", "");
for (final var segment : moduleReference.pathSegments()) {
joiner.add(segment);
}
return joiner.toString();
}
visiting.remove(moduleId);
if (visited.add(moduleId)) {
ordered.add(moduleId);
}
return "";
}
private int safeToInt(final long value) {
if (value > Integer.MAX_VALUE) {
return Integer.MAX_VALUE;
}
if (value < Integer.MIN_VALUE) {
return Integer.MIN_VALUE;
}
return (int) value;
}
private Set<ModuleId> blockedModulesByDependency(
final Set<ModuleId> failedModuleIds,
final Map<ModuleId, Set<ModuleId>> dependenciesByModule) {
final Map<ModuleId, ReadOnlyList<ModuleId>> dependenciesByModule) {
final Map<ModuleId, Set<ModuleId>> dependentsByModule = new HashMap<>();
for (final var entry : dependenciesByModule.entrySet()) {
final var importer = entry.getKey();
@@ -7,13 +7,14 @@ import p.studio.compiler.source.tables.ModuleTable;
import p.studio.utilities.structures.ReadOnlyList;
import java.util.HashMap;
import java.util.LinkedHashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Set;
record PbsModuleAssembly(
ReadOnlyList<PbsParsedSourceFile> parsedSourceFiles,
Map<ModuleId, Set<ModuleId>> moduleDependencyGraph,
Map<ModuleId, ReadOnlyList<ModuleId>> moduleDependencyGraph,
ReadOnlyList<ModuleReference> canonicalModulePool,
Set<ModuleId> failedModuleIds,
ModuleTable moduleTable,
@@ -26,11 +27,11 @@ record PbsModuleAssembly(
if (moduleDependencyGraph == null || moduleDependencyGraph.isEmpty()) {
moduleDependencyGraph = Map.of();
} else {
final var copiedGraph = new HashMap<ModuleId, Set<ModuleId>>();
final var copiedGraph = new LinkedHashMap<ModuleId, ReadOnlyList<ModuleId>>();
for (final var entry : moduleDependencyGraph.entrySet()) {
copiedGraph.put(entry.getKey(), Set.copyOf(entry.getValue() == null ? Set.of() : entry.getValue()));
copiedGraph.put(entry.getKey(), entry.getValue() == null ? ReadOnlyList.empty() : entry.getValue());
}
moduleDependencyGraph = Map.copyOf(copiedGraph);
moduleDependencyGraph = java.util.Collections.unmodifiableMap(copiedGraph);
}
if (moduleTable == null) {
@@ -353,20 +353,24 @@ final class PbsModuleAssemblyService {
}
}
private Map<ModuleId, Set<ModuleId>> buildModuleDependencyGraph(
private Map<ModuleId, ReadOnlyList<ModuleId>> buildModuleDependencyGraph(
final ArrayList<PbsParsedSourceFile> parsedSourceFiles,
final ModuleTable moduleTable) {
final Map<ModuleId, Set<ModuleId>> dependenciesByModule = new HashMap<>();
final Map<ModuleId, LinkedHashSet<ModuleId>> dependenciesByModule = new LinkedHashMap<>();
for (final var parsedSource : parsedSourceFiles) {
final var moduleDependencies = dependenciesByModule.computeIfAbsent(
parsedSource.moduleId(),
ignored -> new HashSet<>());
ignored -> new LinkedHashSet<>());
for (final var importDecl : parsedSource.ast().imports()) {
final var moduleRef = importDecl.moduleRef();
moduleDependencies.add(moduleId(moduleTable, moduleRef.project(), moduleRef.pathSegments()));
}
}
return dependenciesByModule;
final Map<ModuleId, ReadOnlyList<ModuleId>> ordered = new LinkedHashMap<>();
for (final var entry : dependenciesByModule.entrySet()) {
ordered.put(entry.getKey(), ReadOnlyList.wrap(new ArrayList<>(entry.getValue())));
}
return java.util.Collections.unmodifiableMap(ordered);
}
private ModuleId moduleId(
@@ -1,61 +1,136 @@
[Doc(markdown = """
Represents a packed RGBA color.
""")]
[BuiltinType(name = "color", version = 1)]
declare builtin type Color(
pub raw: int
) {
[Doc(markdown = """
Builds a color from its packed integer representation.
- `raw`: packed color value.
""")]
[IntrinsicCall(name = "from_raw", version = 1)]
fn from_raw(raw: int) -> Color;
[Doc(markdown = """
Builds an opaque color from red, green, and blue channels.
- `r`: red channel.
- `g`: green channel.
- `b`: blue channel.
""")]
[IntrinsicCall(name = "rgb", version = 1)]
fn rgb(r: int, g: int, b: int) -> Color;
[Doc(markdown = """
Builds a color from red, green, blue, and alpha channels.
- `r`: red channel.
- `g`: green channel.
- `b`: blue channel.
- `a`: alpha channel.
""")]
[IntrinsicCall(name = "rgba", version = 1)]
fn rgba(r: int, g: int, b: int, a: int) -> Color;
[Doc(markdown = """
Parses a CSS-style RGBA color string.
- `value`: color text in HTML/CSS format.
""")]
[IntrinsicCall(name = "html_rgba", version = 1)]
fn html_rgba(value: str) -> Color;
[Doc(markdown = """
Builds an opaque grayscale color.
- `value`: grayscale channel value.
""")]
[IntrinsicCall(name = "gray_scale", version = 1)]
fn gray_scale(value: int) -> Color;
[Doc(markdown = """
Returns the packed RGB portion of this color.
""")]
[IntrinsicCall(name = "hex", version = 1)]
fn hex() -> int;
[Doc(markdown = """
Returns the alpha channel of this color.
""")]
[IntrinsicCall(name = "alpha", version = 1)]
fn alpha() -> int;
}
[Doc(markdown = """
Opaque black.
""")]
[BuiltinConst(target = "color", name = "black", version = 1)]
declare const BLACK: Color;
[Doc(markdown = """
Opaque white.
""")]
[BuiltinConst(target = "color", name = "white", version = 1)]
declare const WHITE: Color;
[Doc(markdown = """
Opaque red.
""")]
[BuiltinConst(target = "color", name = "red", version = 1)]
declare const RED: Color;
[Doc(markdown = """
Opaque green.
""")]
[BuiltinConst(target = "color", name = "green", version = 1)]
declare const GREEN: Color;
[Doc(markdown = """
Opaque blue.
""")]
[BuiltinConst(target = "color", name = "blue", version = 1)]
declare const BLUE: Color;
[Doc(markdown = """
Opaque yellow.
""")]
[BuiltinConst(target = "color", name = "yellow", version = 1)]
declare const YELLOW: Color;
[Doc(markdown = """
Opaque orange.
""")]
[BuiltinConst(target = "color", name = "orange", version = 1)]
declare const ORANGE: Color;
[Doc(markdown = """
Opaque indigo.
""")]
[BuiltinConst(target = "color", name = "indigo", version = 1)]
declare const INDIGO: Color;
[Doc(markdown = """
Opaque gray.
""")]
[BuiltinConst(target = "color", name = "gray", version = 1)]
declare const GRAY: Color;
[Doc(markdown = """
Opaque cyan.
""")]
[BuiltinConst(target = "color", name = "cyan", version = 1)]
declare const CYAN: Color;
[Doc(markdown = """
Opaque magenta.
""")]
[BuiltinConst(target = "color", name = "magenta", version = 1)]
declare const MAGENTA: Color;
[Doc(markdown = """
Fully transparent color.
""")]
[BuiltinConst(target = "color", name = "transparent", version = 1)]
declare const TRANSPARENT: Color;
@@ -1,35 +1,83 @@
[Doc(markdown = """
Low-level asset loading host bindings.
""")]
declare host LowAssets {
[Doc(markdown = """
Starts loading an addressable asset into a slot.
- `addressable`: asset reference.
- `slot`: destination slot.
""")]
[Host(module = "asset", name = "load", version = 1)]
[Capability(name = "asset")]
[AssetLowering(param = 0)]
fn load(addressable: Addressable, slot: int) -> (status: int, loading_handle: int);
[Doc(markdown = """
Reads the status of an asset loading operation.
- `loading_handle`: handle returned by `load`.
""")]
[Host(module = "asset", name = "status", version = 1)]
[Capability(name = "asset")]
fn status(loading_handle: int) -> int;
[Doc(markdown = """
Commits a completed asset loading operation.
- `loading_handle`: handle returned by `load`.
""")]
[Host(module = "asset", name = "commit", version = 1)]
[Capability(name = "asset")]
fn commit(loading_handle: int) -> int;
[Doc(markdown = """
Cancels an asset loading operation.
- `loading_handle`: handle returned by `load`.
""")]
[Host(module = "asset", name = "cancel", version = 1)]
[Capability(name = "asset")]
fn cancel(loading_handle: int) -> int;
}
[Doc(markdown = """
High-level asset loading commands.
""")]
declare service Assets {
[Doc(markdown = """
Starts loading an addressable asset into a slot.
- `addressable`: asset reference.
- `slot`: destination slot.
""")]
fn load(addressable: Addressable, slot: int) -> (status: int, loading_handle: int) {
return LowAssets.load(addressable, slot);
}
[Doc(markdown = """
Reads the status of an asset loading operation.
- `loading_handle`: handle returned by `load`.
""")]
fn status(loading_handle: int) -> int {
return LowAssets.status(loading_handle);
}
[Doc(markdown = """
Commits a completed asset loading operation.
- `loading_handle`: handle returned by `load`.
""")]
fn commit(loading_handle: int) -> int {
return LowAssets.commit(loading_handle);
}
[Doc(markdown = """
Cancels an asset loading operation.
- `loading_handle`: handle returned by `load`.
""")]
fn cancel(loading_handle: int) -> int {
return LowAssets.cancel(loading_handle);
}
@@ -1,16 +1,46 @@
[Doc(markdown = """
Low-level scene composer host bindings.
""")]
declare host LowComposer {
[Doc(markdown = """
Sets the scene camera position.
- `x`: camera x coordinate.
- `y`: camera y coordinate.
""")]
[Host(module = "composer", name = "set_camera", version = 1)]
[Capability(name = "composer")]
fn set_camera(x: int, y: int) -> void;
[Doc(markdown = """
Binds a scene bank for composer emission.
- `scene_bank_id`: scene bank identifier.
""")]
[Host(module = "composer", name = "bind_scene", version = 1)]
[Capability(name = "composer")]
fn bind_scene(scene_bank_id: int) -> int;
[Doc(markdown = """
Unbinds the active scene bank.
""")]
[Host(module = "composer", name = "unbind_scene", version = 1)]
[Capability(name = "composer")]
fn unbind_scene() -> int;
[Doc(markdown = """
Emits a sprite command into the active scene.
- `glyph_id`: glyph identifier.
- `palette_id`: palette identifier.
- `x`: screen x coordinate.
- `y`: screen y coordinate.
- `layer`: render layer.
- `bank_id`: asset bank identifier.
- `flip_x`: mirror horizontally.
- `flip_y`: mirror vertically.
- `priority`: ordering priority.
""")]
[Host(module = "composer", name = "emit_sprite", version = 1)]
[Capability(name = "composer")]
fn emit_sprite(
@@ -26,23 +56,53 @@ declare host LowComposer {
) -> int;
}
[Doc(markdown = """
High-level scene composition commands.
""")]
declare service Composer
{
[Doc(markdown = """
Sets the scene camera position.
- `x`: camera x coordinate.
- `y`: camera y coordinate.
""")]
fn set_camera(x: int, y: int) -> void
{
LowComposer.set_camera(x, y);
}
[Doc(markdown = """
Binds a scene bank for composer emission.
- `scene_bank_id`: scene bank identifier.
""")]
fn bind_scene(scene_bank_id: int) -> int
{
return LowComposer.bind_scene(scene_bank_id);
}
[Doc(markdown = """
Unbinds the active scene bank.
""")]
fn unbind_scene() -> int
{
return LowComposer.unbind_scene();
}
[Doc(markdown = """
Emits a sprite command into the active scene.
- `glyph_id`: glyph identifier.
- `palette_id`: palette identifier.
- `x`: screen x coordinate.
- `y`: screen y coordinate.
- `layer`: render layer.
- `bank_id`: asset bank identifier.
- `flip_x`: mirror horizontally.
- `flip_y`: mirror vertically.
- `priority`: ordering priority.
""")]
fn emit_sprite(
glyph_id: int,
palette_id: int,
@@ -1,67 +1,189 @@
import { Color } from @core:color;
[Doc(markdown = """
Low-level 2D graphics host bindings.
""")]
declare host LowGfx {
[Doc(markdown = """
Clears the render target.
- `color`: fill color.
""")]
[Host(module = "gfx2d", name = "clear", version = 1)]
[Capability(name = "gfx2d")]
fn clear(color: Color) -> void;
[Doc(markdown = """
Draws a filled rectangle.
- `x`: left coordinate.
- `y`: top coordinate.
- `w`: rectangle width.
- `h`: rectangle height.
- `color`: fill color.
""")]
[Host(module = "gfx2d", name = "fill_rect", version = 1)]
[Capability(name = "gfx2d")]
fn fill_rect(x: int, y: int, w: int, h: int, color: Color) -> void;
[Doc(markdown = """
Draws a line segment.
- `x1`: start x coordinate.
- `y1`: start y coordinate.
- `x2`: end x coordinate.
- `y2`: end y coordinate.
- `color`: line color.
""")]
[Host(module = "gfx2d", name = "draw_line", version = 1)]
[Capability(name = "gfx2d")]
fn draw_line(x1: int, y1: int, x2: int, y2: int, color: Color) -> void;
[Doc(markdown = """
Draws a circle outline.
- `x`: center x coordinate.
- `y`: center y coordinate.
- `r`: radius.
- `color`: outline color.
""")]
[Host(module = "gfx2d", name = "draw_circle", version = 1)]
[Capability(name = "gfx2d")]
fn draw_circle(x: int, y: int, r: int, color: Color) -> void;
[Doc(markdown = """
Draws a filled circle.
- `x`: center x coordinate.
- `y`: center y coordinate.
- `r`: radius.
- `border_color`: outline color.
- `fill_color`: fill color.
""")]
[Host(module = "gfx2d", name = "draw_disc", version = 1)]
[Capability(name = "gfx2d")]
fn draw_disc(x: int, y: int, r: int, border_color: Color, fill_color: Color) -> void;
[Doc(markdown = """
Draws a filled rectangle with a border.
- `x`: left coordinate.
- `y`: top coordinate.
- `w`: rectangle width.
- `h`: rectangle height.
- `border_color`: border color.
- `fill_color`: fill color.
""")]
[Host(module = "gfx2d", name = "draw_square", version = 1)]
[Capability(name = "gfx2d")]
fn draw_square(x: int, y: int, w: int, h: int, border_color: Color, fill_color: Color) -> void;
[Doc(markdown = """
Draws text at a screen position.
- `x`: left coordinate.
- `y`: baseline coordinate.
- `message`: text to draw.
- `color`: text color.
""")]
[Host(module = "gfx2d", name = "draw_text", version = 1)]
[Capability(name = "gfx2d")]
fn draw_text(x: int, y: int, message: str, color: Color) -> void;
}
[Doc(markdown = """
High-level 2D drawing commands.
""")]
declare service Gfx
{
[Doc(markdown = """
Clears the render target.
- `color`: fill color.
""")]
fn clear(color: Color) -> void
{
LowGfx.clear(color);
}
[Doc(markdown = """
Draws a filled rectangle.
- `x`: left coordinate.
- `y`: top coordinate.
- `w`: rectangle width.
- `h`: rectangle height.
- `color`: fill color.
""")]
fn fill_rect(x: int, y: int, w: int, h: int, color: Color) -> void
{
LowGfx.fill_rect(x, y, w, h, color);
}
[Doc(markdown = """
Draws a line segment.
- `x1`: start x coordinate.
- `y1`: start y coordinate.
- `x2`: end x coordinate.
- `y2`: end y coordinate.
- `color`: line color.
""")]
fn draw_line(x1: int, y1: int, x2: int, y2: int, color: Color) -> void
{
LowGfx.draw_line(x1, y1, x2, y2, color);
}
[Doc(markdown = """
Draws a circle outline.
- `x`: center x coordinate.
- `y`: center y coordinate.
- `r`: radius.
- `color`: outline color.
""")]
fn draw_circle(x: int, y: int, r: int, color: Color) -> void
{
LowGfx.draw_circle(x, y, r, color);
}
[Doc(markdown = """
Draws a filled circle.
- `x`: center x coordinate.
- `y`: center y coordinate.
- `r`: radius.
- `border_color`: outline color.
- `fill_color`: fill color.
""")]
fn draw_disc(x: int, y: int, r: int, border_color: Color, fill_color: Color) -> void
{
LowGfx.draw_disc(x, y, r, border_color, fill_color);
}
[Doc(markdown = """
Draws a filled rectangle with a border.
- `x`: left coordinate.
- `y`: top coordinate.
- `w`: rectangle width.
- `h`: rectangle height.
- `border_color`: border color.
- `fill_color`: fill color.
""")]
fn draw_square(x: int, y: int, w: int, h: int, border_color: Color, fill_color: Color) -> void
{
LowGfx.draw_square(x, y, w, h, border_color, fill_color);
}
[Doc(markdown = """
Draws text at a screen position.
- `x`: left coordinate.
- `y`: baseline coordinate.
- `message`: text to draw.
- `color`: text color.
""")]
fn draw_text(x: int, y: int, message: str, color: Color) -> void
{
LowGfx.draw_text(x, y, message, color);
@@ -1,85 +1,169 @@
[Doc(markdown = """
Root input device provider.
""")]
[BuiltinType(name = "input", version = 1)]
declare builtin type InputType() {
[Doc(markdown = """
Returns the gamepad input state.
""")]
[IntrinsicCall(name = "pad", version = 1)]
fn pad() -> InputPad;
[Doc(markdown = """
Returns the touch input state.
""")]
[IntrinsicCall(name = "touch", version = 1)]
fn touch() -> InputTouch;
}
[Doc(markdown = """
Represents the current gamepad state.
""")]
[BuiltinType(name = "input.pad", version = 1)]
declare builtin type InputPad() {
[Doc(markdown = """
Returns the up directional button.
""")]
[IntrinsicCall(name = "up", version = 1)]
fn up() -> InputButton;
[Doc(markdown = """
Returns the down directional button.
""")]
[IntrinsicCall(name = "down", version = 1)]
fn down() -> InputButton;
[Doc(markdown = """
Returns the left directional button.
""")]
[IntrinsicCall(name = "left", version = 1)]
fn left() -> InputButton;
[Doc(markdown = """
Returns the right directional button.
""")]
[IntrinsicCall(name = "right", version = 1)]
fn right() -> InputButton;
[Doc(markdown = """
Returns the A action button.
""")]
[IntrinsicCall(name = "a", version = 1)]
fn a() -> InputButton;
[Doc(markdown = """
Returns the B action button.
""")]
[IntrinsicCall(name = "b", version = 1)]
fn b() -> InputButton;
[Doc(markdown = """
Returns the X action button.
""")]
[IntrinsicCall(name = "x", version = 1)]
fn x() -> InputButton;
[Doc(markdown = """
Returns the Y action button.
""")]
[IntrinsicCall(name = "y", version = 1)]
fn y() -> InputButton;
[Doc(markdown = """
Returns the left shoulder button.
""")]
[IntrinsicCall(name = "l", version = 1)]
fn l() -> InputButton;
[Doc(markdown = """
Returns the right shoulder button.
""")]
[IntrinsicCall(name = "r", version = 1)]
fn r() -> InputButton;
[Doc(markdown = """
Returns the start button.
""")]
[IntrinsicCall(name = "start", version = 1)]
fn start() -> InputButton;
[Doc(markdown = """
Returns the select button.
""")]
[IntrinsicCall(name = "select", version = 1)]
fn select() -> InputButton;
}
[Doc(markdown = """
Represents the current touch input state.
""")]
[BuiltinType(name = "input.touch", version = 1)]
declare builtin type InputTouch() {
[Doc(markdown = """
Returns the touch contact as a button state.
""")]
[IntrinsicCall(name = "button", version = 1)]
fn button() -> InputButton;
[Doc(markdown = """
Returns the current touch x coordinate.
""")]
[IntrinsicCall(name = "x", version = 1)]
fn x() -> int;
[Doc(markdown = """
Returns the current touch y coordinate.
""")]
[IntrinsicCall(name = "y", version = 1)]
fn y() -> int;
}
[Doc(markdown = """
Represents one digital input button.
""")]
[BuiltinType(name = "input.button", version = 1)]
declare builtin type InputButton() {
[Doc(markdown = """
Returns true on the frame where the button becomes down.
""")]
[IntrinsicCall(name = "pressed", version = 1)]
fn pressed() -> bool;
[Doc(markdown = """
Returns true on the frame where the button becomes up.
""")]
[IntrinsicCall(name = "released", version = 1)]
fn released() -> bool;
[Doc(markdown = """
Returns true while the button is currently down.
""")]
[IntrinsicCall(name = "down", version = 1)]
fn down() -> bool;
[Doc(markdown = """
Returns how long the button has been held.
""")]
[IntrinsicCall(name = "hold", version = 1)]
fn hold() -> int;
}
[Doc(markdown = """
High-level access to player input devices.
""")]
declare service Input
{
[Doc(markdown = """
Returns the gamepad input state.
""")]
fn pad() -> InputPad
{
return InputType.pad();
}
[Doc(markdown = """
Returns the touch input state.
""")]
fn touch() -> InputTouch
{
return InputType.touch();
@@ -1,60 +1,134 @@
[Doc(markdown = """
Low-level logging host bindings.
""")]
declare host LowLog {
[Doc(markdown = """
Writes a log message at a numeric level.
- `level`: severity level.
- `message`: text to write.
""")]
[Host(module = "log", name = "write", version = 1)]
[Capability(name = "log")]
fn write(level: int, message: str) -> void;
[Doc(markdown = """
Writes a tagged log message at a numeric level.
- `level`: severity level.
- `tag`: application-defined tag.
- `message`: text to write.
""")]
[Host(module = "log", name = "write_tag", version = 1)]
[Capability(name = "log")]
fn write_tag(level: int, tag: int, message: str) -> void;
}
[Doc(markdown = """
High-level logging commands.
""")]
declare service Log
{
[Doc(markdown = """
Writes a trace log message.
- `msg`: text to write.
""")]
fn trace(msg: str) -> void
{
LowLog.write(0, msg);
}
[Doc(markdown = """
Writes a debug log message.
- `msg`: text to write.
""")]
fn debug(msg: str) -> void
{
LowLog.write(1, msg);
}
[Doc(markdown = """
Writes an informational log message.
- `msg`: text to write.
""")]
fn info(msg: str) -> void
{
LowLog.write(2, msg);
}
[Doc(markdown = """
Writes a warning log message.
- `msg`: text to write.
""")]
fn warn(msg: str) -> void
{
LowLog.write(3, msg);
}
[Doc(markdown = """
Writes a failure log message.
- `msg`: text to write.
""")]
fn failure(msg: str) -> void
{
LowLog.write(4, msg);
}
[Doc(markdown = """
Writes a tagged trace log message.
- `tag`: application-defined tag.
- `msg`: text to write.
""")]
fn trace(tag: int, msg: str) -> void
{
LowLog.write_tag(0, tag, msg);
}
[Doc(markdown = """
Writes a tagged debug log message.
- `tag`: application-defined tag.
- `msg`: text to write.
""")]
fn debug(tag: int, msg: str) -> void
{
LowLog.write_tag(1, tag, msg);
}
[Doc(markdown = """
Writes a tagged informational log message.
- `tag`: application-defined tag.
- `msg`: text to write.
""")]
fn info(tag: int, msg: str) -> void
{
LowLog.write_tag(2, tag, msg);
}
[Doc(markdown = """
Writes a tagged warning log message.
- `tag`: application-defined tag.
- `msg`: text to write.
""")]
fn warn(tag: int, msg: str) -> void
{
LowLog.write_tag(3, tag, msg);
}
[Doc(markdown = """
Writes a tagged failure log message.
- `tag`: application-defined tag.
- `msg`: text to write.
""")]
fn failure(tag: int, msg: str) -> void
{
LowLog.write_tag(4, tag, msg);
@@ -0,0 +1,21 @@
package p.studio.compiler;
import org.junit.jupiter.api.Test;
import p.studio.compiler.services.PBSFrontendPhaseService;
import static org.junit.jupiter.api.Assertions.assertEquals;
import static org.junit.jupiter.api.Assertions.assertInstanceOf;
import static org.junit.jupiter.api.Assertions.assertTrue;
class PBSFrontendProviderTest {
@Test
void shouldExposePbsSpecificationAndCompiler() {
final var provider = new PBSFrontendProvider();
assertEquals(PBSDefinitions.PBS.getLanguageId(), provider.specification().getLanguageId());
assertInstanceOf(PBSFrontendPhaseService.class, provider.compiler());
assertTrue(provider.languageService().isPresent());
assertInstanceOf(PBSFrontendLanguageService.class, provider.languageService().orElseThrow());
}
}
@@ -0,0 +1,54 @@
package p.studio.compiler.pbs;
import org.junit.jupiter.api.Test;
import static org.junit.jupiter.api.Assertions.assertEquals;
class PbsDocumentationTextBlockNormalizerTest {
@Test
void shouldNormalizeIncidentalBoundaryLinesAndIndentation() {
final var lexeme = "\"\"\"\n" +
" Draws a sprite.\n" +
"\n" +
" - `x`: screen x\n" +
" - `y`: screen y\n" +
" \"\"\"";
final var normalized = PbsDocumentationTextBlockNormalizer.normalizeLexeme(lexeme);
assertEquals("Draws a sprite.\n\n- `x`: screen x\n- `y`: screen y", normalized);
}
@Test
void shouldPreserveInternalLineBreaksAndMarkdownContent() {
final var content = "\n" +
" Paragraph one.\n" +
" \n" +
" `code_span()`\n" +
" Paragraph two.\n";
final var normalized = PbsDocumentationTextBlockNormalizer.normalizeContent(content);
assertEquals("Paragraph one.\n\n`code_span()`\nParagraph two.", normalized);
}
@Test
void shouldReturnEmptyStringForWhitespaceOnlyPayload() {
final var normalized = PbsDocumentationTextBlockNormalizer.normalizeLexeme("\"\"\"\n \n \t\n\"\"\"");
assertEquals("", normalized.trim());
}
@Test
void shouldNormalizeCarriageReturnLineEndings() {
final var normalized = PbsDocumentationTextBlockNormalizer.normalizeContent("\r\n\tLine one\r\n\tLine two\r\n");
assertEquals("Line one\nLine two", normalized);
}
@Test
void shouldReturnEmptyStringForMalformedLexeme() {
assertEquals("", PbsDocumentationTextBlockNormalizer.normalizeLexeme("\"not a text block\""));
}
}
@@ -4,8 +4,11 @@ import org.junit.jupiter.api.Test;
import p.studio.compiler.messages.Addressable;
import p.studio.compiler.messages.FESurfaceContext;
import p.studio.compiler.messages.HostAdmissionContext;
import p.studio.compiler.models.IRBackendExecutableFunction;
import p.studio.compiler.models.IRBackendFile;
import p.studio.compiler.models.IRHiddenGlobalKind;
import p.studio.compiler.models.IRGlobalVisibility;
import p.studio.compiler.models.IRLifecycleCallableRole;
import p.studio.compiler.models.IRSyntheticCallableKind;
import p.studio.compiler.models.SourceKind;
import p.studio.compiler.pbs.lexer.LexErrors;
@@ -53,7 +56,58 @@ class PbsFrontendCompilerTest {
}
@Test
void shouldExposeGlobalsAndSyntheticLifecycleArtifactsInBackendFile() {
void shouldNotLowerDocMetadataIntoRuntimeBackendArtifacts() {
final var undocumentedSource = """
declare global SCORE: int = 0;
fn add_one(value: int) -> int {
return value + 1;
}
[Init]
fn boot() -> void {
SCORE = add_one(41);
return;
}
""";
final var marker = "THIS_DOC_PAYLOAD_MUST_NOT_APPEAR_IN_RUNTIME_ARTIFACTS";
final var documentedSource = """
declare global SCORE: int = 0;
[Doc(markdown = \"""
Adds one to a value.
This deliberately large documentation payload is editor metadata only:
THIS_DOC_PAYLOAD_MUST_NOT_APPEAR_IN_RUNTIME_ARTIFACTS.
\""")]
fn add_one(value: int) -> int {
return value + 1;
}
[Doc(markdown = \"""
Initializes the score global.
\""")]
[Init]
fn boot() -> void {
SCORE = add_one(41);
return;
}
""";
final var compiler = new PbsFrontendCompiler();
final var undocumentedDiagnostics = DiagnosticSink.empty();
final var documentedDiagnostics = DiagnosticSink.empty();
final var undocumentedBackend = compiler.compileFile(new FileId(20), undocumentedSource, undocumentedDiagnostics);
final var documentedBackend = compiler.compileFile(new FileId(20), documentedSource, documentedDiagnostics);
assertTrue(undocumentedDiagnostics.isEmpty(), diagnosticsSummary(undocumentedDiagnostics));
assertTrue(documentedDiagnostics.isEmpty(), diagnosticsSummary(documentedDiagnostics));
assertEquals(runtimeFingerprint(undocumentedBackend), runtimeFingerprint(documentedBackend));
assertFalse(String.valueOf(documentedBackend).contains(marker));
}
@Test
void shouldExposeGlobalsAndTypedLifecycleDeclarationInBackendFile() {
final var source = """
declare global SCORE: int = 0;
@@ -76,12 +130,20 @@ class PbsFrontendCompilerTest {
assertFalse(fileBackend.globals().getFirst().isHidden());
assertEquals(IRHiddenGlobalKind.NONE, fileBackend.globals().getFirst().hiddenKind());
assertEquals("SCORE", fileBackend.globals().getFirst().origin().derivedFromUserSymbol());
assertEquals(4, fileBackend.syntheticFunctions().size());
assertEquals(1, fileBackend.syntheticFunctions().size());
assertTrue(fileBackend.syntheticFunctions().stream().anyMatch(fn -> fn.kind() == IRSyntheticCallableKind.FILE_INIT_FRAGMENT));
assertTrue(fileBackend.syntheticFunctions().stream().anyMatch(fn -> fn.kind() == IRSyntheticCallableKind.MODULE_INIT));
assertTrue(fileBackend.syntheticFunctions().stream().anyMatch(fn -> fn.kind() == IRSyntheticCallableKind.PROJECT_INIT));
assertTrue(fileBackend.syntheticFunctions().stream().anyMatch(fn -> fn.kind() == IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER));
assertTrue(fileBackend.syntheticFunctions().stream().allMatch(fn -> !fn.origin().anchorCallableName().isBlank()));
assertEquals(IRLifecycleCallableRole.FILE_INIT_FRAGMENT,
fileBackend.lifecycleDeclaration().fileInitFragment().role());
assertEquals(IRLifecycleCallableRole.PROJECT_INIT,
fileBackend.lifecycleDeclaration().projectInit().role());
assertEquals(IRLifecycleCallableRole.FRAME_ROOT,
fileBackend.lifecycleDeclaration().frameRoot().role());
final var fileInit = fileBackend.executableFunctions().stream()
.filter(fn -> fn.callableId().equals(fileBackend.lifecycleDeclaration().fileInitFragment().callableId()))
.findFirst()
.orElseThrow();
assertTrue(fileInit.instructions().stream().noneMatch(instruction -> "boot".equals(instruction.calleeCallableName())));
}
@Test
@@ -812,4 +874,62 @@ class PbsFrontendCompilerTest {
assertEquals(1L, pixel.fields().get(1).slotStart());
assertEquals(2L, pixel.fields().get(2).slotStart());
}
private static String runtimeFingerprint(final IRBackendFile backend) {
final var out = new StringBuilder();
out.append("functions:");
backend.functions().forEach(function -> out
.append(function.name()).append('/')
.append(function.parameterCount()).append('/')
.append(function.hasExplicitReturnType()).append(';'));
out.append("synthetic:");
backend.syntheticFunctions().forEach(function -> out
.append(function.moduleId()).append('/')
.append(function.callableName()).append('/')
.append(function.kind()).append(';'));
out.append("globals:");
backend.globals().forEach(global -> out
.append(global.moduleId()).append('/')
.append(global.name()).append('/')
.append(global.declaredTypeSurface()).append('/')
.append(global.slot()).append('/')
.append(global.visibility()).append('/')
.append(global.hiddenKind()).append(';'));
out.append("executables:");
backend.executableFunctions().forEach(function -> {
out.append(function.moduleId()).append('/')
.append(function.callableName()).append('/')
.append(function.callableId().getIndex()).append('/')
.append(function.paramSlots()).append('/')
.append(function.localSlots()).append('/')
.append(function.returnSlots()).append('/')
.append(function.maxStackSlots()).append('{');
function.instructions().forEach(instruction -> appendInstructionFingerprint(out, instruction));
out.append("};");
});
out.append("modules:").append(backend.modulePool()).append(';');
out.append("callables:").append(backend.callableSignatures()).append(';');
out.append("intrinsics:").append(backend.intrinsicPool()).append(';');
return out.toString();
}
private static void appendInstructionFingerprint(
final StringBuilder out,
final IRBackendExecutableFunction.Instruction instruction) {
out.append(instruction.kind()).append('(')
.append(instruction.calleeModuleId()).append(',')
.append(instruction.calleeCallableName()).append(',')
.append(instruction.calleeCallableId() == null ? "" : instruction.calleeCallableId().getIndex()).append(',')
.append(instruction.hostCall()).append(',')
.append(instruction.intrinsicCall()).append(',')
.append(instruction.label()).append(',')
.append(instruction.targetLabel()).append(',')
.append(instruction.expectedArgSlots()).append(',')
.append(instruction.expectedRetSlots())
.append(");");
}
private static String diagnosticsSummary(final DiagnosticSink diagnostics) {
return diagnostics.stream().map(d -> d.getCode() + ":" + d.getMessage()).toList().toString();
}
}
@@ -113,6 +113,55 @@ class PbsLexerTest {
diagnostics.stream().findFirst().orElseThrow().getCode());
}
@Test
void shouldLexDocumentationTextBlock() {
final var source = """
[Doc(markdown = \"""
Draws a sprite.
- `x`: screen x
- `y`: screen y
\""")]
""";
final var diagnostics = DiagnosticSink.empty();
final var tokens = PbsLexer.lex(source, new FileId(0), diagnostics);
final var docTextBlock = tokens.stream()
.filter(t -> t.kind() == PbsTokenKind.DOC_TEXT_BLOCK)
.findFirst()
.orElseThrow();
assertTrue(docTextBlock.lexeme().startsWith("\"\"\""));
assertTrue(docTextBlock.lexeme().endsWith("\"\"\""));
assertTrue(docTextBlock.lexeme().contains("Draws a sprite."));
assertTrue(docTextBlock.lexeme().contains("- `x`: screen x"));
assertTrue(diagnostics.isEmpty(), "Documentation text block should lex without diagnostics");
}
@Test
void shouldAllowSingleAndDoubleQuotesInsideDocumentationTextBlock() {
final var source = "[Doc(markdown = \"\"\"Uses \"quoted\" text and 'single' quotes.\"\"\")]";
final var diagnostics = DiagnosticSink.empty();
final var tokens = PbsLexer.lex(source, new FileId(0), diagnostics);
final var kinds = tokens.stream().map(PbsToken::kind).toList();
assertTrue(kinds.contains(PbsTokenKind.DOC_TEXT_BLOCK));
assertTrue(diagnostics.isEmpty(), "Quotes inside documentation text block should not need escaping");
}
@Test
void shouldReportUnterminatedDocumentationTextBlock() {
final var source = "[Doc(markdown = \"\"\"unterminated)]";
final var diagnostics = DiagnosticSink.empty();
PbsLexer.lex(source, new FileId(0), diagnostics);
assertTrue(diagnostics.hasErrors(), "Lexer should report unterminated documentation text blocks");
assertEquals(LexErrors.E_LEX_UNTERMINATED_DOC_TEXT_BLOCK.name(),
diagnostics.stream().findFirst().orElseThrow().getCode());
}
@Test
void shouldReportInvalidStringEscape() {
final var source = "\"bad\\q\"";
@@ -339,6 +339,82 @@ class PbsParserTest {
assertEquals(1, constDecl.attributes().size());
}
@Test
void shouldParseDocumentationTextBlockAttributeValuesInInterfaceModuleMode() {
final var source = """
[Doc(markdown = \"""
Represents a packed color.
\""")]
[BuiltinType(name = "color", version = 1)]
declare builtin type Color(
pub raw: int
) {
[Doc(markdown = \"""
Packs this color into an integer.
\""")]
[IntrinsicCall(name = "color.pack", version = 1)]
fn pack() -> int;
}
""";
final var diagnostics = DiagnosticSink.empty();
final var fileId = new FileId(0);
final PbsAst.File ast = PbsParser.parse(
PbsLexer.lex(source, fileId, diagnostics),
fileId,
diagnostics,
PbsParser.ParseMode.INTERFACE_MODULE);
assertTrue(diagnostics.isEmpty(), "Parser should accept documentation text block attribute values");
final var builtinType = assertInstanceOf(PbsAst.BuiltinTypeDecl.class, ast.topDecls().getFirst());
final var typeDoc = builtinType.attributes().getFirst().arguments().getFirst().value();
assertInstanceOf(PbsAst.AttributeDocTextBlockValue.class, typeDoc);
assertTrue(((PbsAst.AttributeDocTextBlockValue) typeDoc).value().contains("Represents a packed color."));
final var signatureDoc = builtinType.signatures().getFirst().attributes().getFirst().arguments().getFirst().value();
assertInstanceOf(PbsAst.AttributeDocTextBlockValue.class, signatureDoc);
assertTrue(((PbsAst.AttributeDocTextBlockValue) signatureDoc).value().startsWith("\"\"\""));
assertTrue(((PbsAst.AttributeDocTextBlockValue) signatureDoc).value().endsWith("\"\"\""));
}
@Test
void shouldPreserveDocumentationAttributesOnStructAndServiceMethods() {
final var source = """
declare struct Sprite() {
[Doc(markdown = \"""
Draws this sprite.
\""")]
fn draw() -> void {}
}
[Doc(markdown = \"""
Provides drawing commands.
\""")]
declare service Gfx {
[Doc(markdown = \"""
Clears the screen.
\""")]
fn clear() -> void {}
}
""";
final var diagnostics = DiagnosticSink.empty();
final var fileId = new FileId(0);
final PbsAst.File ast = PbsParser.parse(
PbsLexer.lex(source, fileId, diagnostics),
fileId,
diagnostics,
PbsParser.ParseMode.ORDINARY);
assertTrue(diagnostics.isEmpty(), "Parser should accept Doc attributes on struct and service methods");
final var structDecl = assertInstanceOf(PbsAst.StructDecl.class, ast.topDecls().getFirst());
assertEquals("Doc", structDecl.methods().getFirst().attributes().getFirst().name());
final var serviceDecl = assertInstanceOf(PbsAst.ServiceDecl.class, ast.topDecls().get(1));
assertEquals("Doc", serviceDecl.attributes().getFirst().name());
assertEquals("Doc", serviceDecl.methods().getFirst().attributes().getFirst().name());
}
@Test
void shouldRecoverFromMalformedReservedDeclarationsInInterfaceModuleMode() {
final var source = """
@@ -123,6 +123,49 @@ final class PbsEditorialSupportServiceTest {
assertEquals("Input", hostMethodHover.origin());
}
@Test
void shouldExposeDocMetadataOnTopLevelSymbolHover() {
final var source = """
[Doc(markdown = \"""
Runs the program.
- no parameters
\""")]
fn run() -> void { return; }
""";
final var ast = parseOrdinary(source);
final var hover = requireHover(source, ast, ReadOnlyList.empty(), "run");
assertEquals(PbsEditorialSymbolKind.FUNCTION, hover.kind());
assertEquals("Runs the program.\n\n- no parameters", hover.documentation());
}
@Test
void shouldExposeDocMetadataOnMemberHover() {
final var source = """
import { Gfx } from @sdk:gfx;
fn main() -> void {
Gfx.draw_pixel(1, 2, 3);
}
""";
final var ast = parseOrdinary(source);
final var supplementalTopDecls = supplementalTopDecls("""
declare host Gfx {
[Doc(markdown = \"""
Draws a single pixel.
\""")]
fn draw_pixel(x: int, y: int, color: int) -> void;
}
""");
final var hover = requireHover(source, ast, supplementalTopDecls, "draw_pixel");
assertEquals(PbsEditorialSymbolKind.METHOD, hover.kind());
assertEquals("Draws a single pixel.", hover.documentation());
}
@Test
void shouldResolveSignatureHelpForFunctionMethodAndConstructorCalls() {
final var source = """
@@ -14,15 +14,27 @@ class PbsInterfaceModuleSemanticsTest {
@Test
void shouldAcceptValidInterfaceModuleReservedMetadataShapes() {
final var source = """
[Doc(markdown = \"""
Packed color value.
\""")]
[BuiltinType(name = "color", version = 1)]
declare builtin type Color(
pub raw: int
) {
[Doc(markdown = \"""
Packs the color.
\""")]
[IntrinsicCall(name = "pack", version = 1)]
fn pack() -> int;
}
[Doc(markdown = \"""
Graphics host API.
\""")]
declare host Gfx {
[Doc(markdown = \"""
Draws one pixel.
\""")]
[Host(module = "gfx2d", name = "draw_pixel", version = 1)]
[Capability(name = "gfx2d")]
fn draw_pixel(x: int, y: int, c: Color) -> void;
@@ -43,6 +55,26 @@ class PbsInterfaceModuleSemanticsTest {
d.getCode().equals(PbsSemanticsErrors.E_SEM_MALFORMED_RESERVED_ATTRIBUTE.name())));
}
@Test
void shouldRejectDocOnBuiltinFields() {
final var source = """
[BuiltinType(name = "color", version = 1)]
declare builtin type Color(
[Doc(markdown = \"""
Raw color bits.
\""")]
pub raw: int
) {
}
""";
final var diagnostics = DiagnosticSink.empty();
new PbsFrontendCompiler().compileFile(new FileId(0), source, diagnostics, SourceKind.SDK_INTERFACE);
assertTrue(diagnostics.stream().anyMatch(d ->
d.getCode().equals(PbsSemanticsErrors.E_SEM_INVALID_RESERVED_ATTRIBUTE_TARGET.name())));
}
@Test
void shouldRejectInvalidReservedAttributeTargetsAndShapesInInterfaceModule() {
final var source = """
@@ -9,6 +9,106 @@ import static org.junit.jupiter.api.Assertions.*;
class PbsSemanticsDeclarationsTest {
@Test
void shouldAcceptDocOnNamedApiDeclarations() {
final var source = """
[Doc(markdown = \"""
Runs the program.
\""")]
fn run() -> void { return; }
[Doc(markdown = \"""
Holds player state.
\""")]
declare struct Player(x: int) {
[Doc(markdown = \"""
Updates player state.
\""")]
fn update() -> void { return; }
}
[Doc(markdown = \"""
Game services.
\""")]
declare service Game {
[Doc(markdown = \"""
Runs one game tick.
\""")]
fn tick() -> void { return; }
}
""";
final var diagnostics = DiagnosticSink.empty();
new PbsFrontendCompiler().compileFile(new FileId(0), source, diagnostics);
assertFalse(diagnostics.stream().anyMatch(d ->
d.getCode().equals(PbsSemanticsErrors.E_SEM_MALFORMED_RESERVED_ATTRIBUTE.name())));
assertFalse(diagnostics.stream().anyMatch(d ->
d.getCode().equals(PbsSemanticsErrors.E_SEM_INVALID_RESERVED_ATTRIBUTE_TARGET.name())));
}
@Test
void shouldRejectMalformedDuplicateAndEmptyDocAttributes() {
final var source = """
[Doc(markdown = \"""
First.
\""")]
[Doc(markdown = \"""
Second.
\""")]
fn duplicated() -> void { return; }
[Doc(text = \"""
Wrong argument.
\""")]
fn malformed() -> void { return; }
[Doc(markdown = \"""
\""")]
fn empty() -> void { return; }
""";
final var diagnostics = DiagnosticSink.empty();
new PbsFrontendCompiler().compileFile(new FileId(0), source, diagnostics);
assertTrue(diagnostics.stream().anyMatch(d ->
d.getCode().equals(PbsSemanticsErrors.E_SEM_DUPLICATE_RESERVED_ATTRIBUTE.name())));
final var malformedCount = diagnostics.stream()
.filter(d -> d.getCode().equals(PbsSemanticsErrors.E_SEM_MALFORMED_RESERVED_ATTRIBUTE.name()))
.count();
assertTrue(malformedCount >= 2);
}
@Test
void shouldRejectDocAliasesExtraArgumentsAndParameterTargets() {
final var source = """
[Document(markdown = \"""
Alias should not be accepted.
\""")]
fn alias() -> void { return; }
[Doc(markdown = \"""
Has an extra argument.
\""", format = \"""
markdown
\""")]
fn extra() -> void { return; }
fn parameter([Doc(markdown = \"""
Parameter docs are not allowed.
\""")] value: int) -> int { return value; }
""";
final var diagnostics = DiagnosticSink.empty();
new PbsFrontendCompiler().compileFile(new FileId(0), source, diagnostics);
assertTrue(diagnostics.stream().anyMatch(d ->
d.getCode().equals("E_PARSE_ATTRIBUTES_NOT_ALLOWED")));
assertTrue(diagnostics.stream().anyMatch(d ->
d.getCode().equals(PbsSemanticsErrors.E_SEM_MALFORMED_RESERVED_ATTRIBUTE.name())));
}
@Test
void shouldAllowFunctionOverloadsWithDifferentShapes() {
final var source = """
@@ -386,8 +386,10 @@ class PBSFrontendPhaseServiceTest {
assertTrue(diagnostics.isEmpty());
assertEquals(2, irBackend.getFunctions().size());
assertTrue(irBackend.getExecutableFunctions().stream().anyMatch(function -> "frame".equals(function.callableName())));
assertTrue(irBackend.getExecutableFunctions().stream().anyMatch(function ->
function.callableName().startsWith("__pbs.frame_wrapper$m")));
assertNotNull(irBackend.getLifecycleDeclaration());
assertEquals(IRLifecycleCallableRole.FRAME_ROOT, irBackend.getLifecycleDeclaration().frameRoot().role());
assertTrue(irBackend.getSyntheticFunctions().stream().noneMatch(function ->
function.kind() == IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER));
}
@Test
@@ -483,36 +485,19 @@ class PBSFrontendPhaseServiceTest {
BuildingIssueSink.empty());
assertTrue(diagnostics.isEmpty(), diagnostics.stream().map(d -> d.getCode() + ":" + d.getMessage()).toList().toString());
assertTrue(irBackend.getEntryPointCallableName().startsWith("__pbs.frame_wrapper$m"));
assertNotNull(irBackend.getLifecycleDeclaration());
assertEquals(IRLifecycleCallableRole.PROJECT_INIT, irBackend.getLifecycleDeclaration().projectInit().role());
assertEquals(IRLifecycleCallableRole.FRAME_ROOT, irBackend.getLifecycleDeclaration().frameRoot().role());
final var scoreGlobal = irBackend.getGlobals().stream()
.filter(global -> "SCORE".equals(global.name()))
.findFirst()
.orElseThrow();
assertEquals(IRGlobalVisibility.PUBLIC, scoreGlobal.visibility());
final var bootGuard = irBackend.getGlobals().stream()
.filter(global -> global.hiddenKind() == IRHiddenGlobalKind.BOOT_GUARD)
.findFirst()
.orElseThrow();
assertTrue(bootGuard.isHidden());
assertEquals(IRGlobalVisibility.MODULE, bootGuard.visibility());
assertTrue(bootGuard.name().startsWith("__pbs.boot_guard$m"));
assertEquals(1, bootGuard.slot());
final var wrapper = irBackend.getExecutableFunctions().stream()
.filter(function -> irBackend.getEntryPointCallableName().equals(function.callableName()))
.findFirst()
.orElseThrow();
assertEquals(11, wrapper.instructions().size());
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.GET_GLOBAL, wrapper.instructions().get(0).kind());
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.PUSH_BOOL, wrapper.instructions().get(1).kind());
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.EQ, wrapper.instructions().get(2).kind());
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.JMP_IF_TRUE, wrapper.instructions().get(3).kind());
assertTrue(wrapper.instructions().get(4).calleeCallableName().startsWith("__pbs.module_init$m"));
assertTrue(wrapper.instructions().get(5).calleeCallableName().startsWith("__pbs.project_init$m"));
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.PUSH_BOOL, wrapper.instructions().get(6).kind());
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.SET_GLOBAL, wrapper.instructions().get(7).kind());
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.LABEL, wrapper.instructions().get(8).kind());
assertEquals("frame", wrapper.instructions().get(9).calleeCallableName());
assertEquals(p.studio.compiler.models.IRBackendExecutableFunction.InstructionKind.RET, wrapper.instructions().get(10).kind());
assertTrue(irBackend.getGlobals().stream().noneMatch(global -> global.hiddenKind() == IRHiddenGlobalKind.BOOT_GUARD));
assertTrue(irBackend.getSyntheticFunctions().stream().noneMatch(function ->
function.kind() == IRSyntheticCallableKind.MODULE_INIT
|| function.kind() == IRSyntheticCallableKind.PROJECT_INIT
|| function.kind() == IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER));
}
@Test
@@ -572,16 +557,100 @@ class PBSFrontendPhaseServiceTest {
BuildingIssueSink.empty());
assertTrue(diagnostics.isEmpty(), diagnostics.stream().map(d -> d.getCode() + ":" + d.getMessage()).toList().toString());
assertTrue(irBackend.getEntryPointCallableName().startsWith("__pbs.frame_wrapper$m"));
assertNotNull(irBackend.getLifecycleDeclaration());
assertTrue(irBackend.getExecutableFunctions().stream().anyMatch(function ->
function.callableName().startsWith("__pbs.file_init$f")));
assertTrue(irBackend.getExecutableFunctions().stream().anyMatch(function ->
function.callableName().startsWith("__pbs.module_init$m")));
assertTrue(irBackend.getExecutableFunctions().stream().anyMatch(function ->
function.callableName().startsWith("__pbs.project_init$m")));
assertEquals(1, irBackend.getLifecycleDeclaration().modules().getFirst().fileInitFragments().size());
assertEquals(IRLifecycleCallableRole.PROJECT_INIT, irBackend.getLifecycleDeclaration().projectInit().role());
assertEquals(IRLifecycleCallableRole.FRAME_ROOT, irBackend.getLifecycleDeclaration().frameRoot().role());
assertTrue(irBackend.getGlobals().stream().anyMatch(global ->
"SCORE".equals(global.name()) && global.visibility() == IRGlobalVisibility.PUBLIC));
assertTrue(irBackend.getGlobals().stream().anyMatch(global -> global.hiddenKind() == IRHiddenGlobalKind.BOOT_GUARD));
assertTrue(irBackend.getGlobals().stream().noneMatch(global -> global.hiddenKind() == IRHiddenGlobalKind.BOOT_GUARD));
}
@Test
void shouldPublishDependencyOrderedLifecycleModulesInsteadOfNameOrFileIdOrder() throws IOException {
final var projectRoot = tempDir.resolve("project-lifecycle-dep-order");
final var sourceRoot = projectRoot.resolve("src");
final var zzzPath = sourceRoot.resolve("zzz");
final var aaaPath = sourceRoot.resolve("aaa");
Files.createDirectories(zzzPath);
Files.createDirectories(aaaPath);
final var zzzSource = zzzPath.resolve("source.pbs");
final var zzzBarrel = zzzPath.resolve("mod.barrel");
Files.writeString(zzzSource, """
declare global WARM: int = 1;
[Init]
fn warmup() -> void { return; }
""");
Files.writeString(zzzBarrel, """
pub fn warmup() -> void;
pub global WARM;
""");
final var aaaSource = aaaPath.resolve("source.pbs");
final var aaaBarrel = aaaPath.resolve("mod.barrel");
Files.writeString(aaaSource, """
import { warmup } from @app:zzz;
[Init]
fn boot() -> void {
warmup();
return;
}
[Frame]
fn frame() -> void { return; }
""");
Files.writeString(aaaBarrel, """
pub fn boot() -> void;
pub fn frame() -> void;
""");
final var projectTable = new ProjectTable();
final var fileTable = new FileTable(1);
final var projectId = projectTable.register(ProjectDescriptor.builder()
.rootPath(projectRoot)
.name("app")
.version("1.0.0")
.sourceRoots(ReadOnlyList.wrap(List.of(sourceRoot)))
.build());
registerFile(projectId, projectRoot, aaaSource, fileTable);
registerFile(projectId, projectRoot, aaaBarrel, fileTable);
registerFile(projectId, projectRoot, zzzSource, fileTable);
registerFile(projectId, projectRoot, zzzBarrel, fileTable);
final var ctx = new FrontendPhaseContext(
projectTable,
fileTable,
new BuildStack(ReadOnlyList.wrap(List.of(projectId))),
AppMode.Game);
final var diagnostics = DiagnosticSink.empty();
final var irBackend = new PBSFrontendPhaseService().compile(
ctx,
diagnostics,
LogAggregator.empty(),
BuildingIssueSink.empty());
assertEquals(0, diagnostics.errorCount(), diagnostics.stream().map(d -> d.getCode() + ":" + d.getMessage()).toList().toString());
final var declaration = irBackend.getLifecycleDeclaration();
assertNotNull(declaration);
assertEquals(2, declaration.modules().size());
final var firstModule = irBackend.getModulePool().get(declaration.modules().get(0).moduleId().getIndex());
final var secondModule = irBackend.getModulePool().get(declaration.modules().get(1).moduleId().getIndex());
assertEquals("zzz", firstModule.pathSegments().getFirst());
assertEquals("aaa", secondModule.pathSegments().getFirst());
assertFalse(declaration.modules().get(0).fileInitFragments().isEmpty());
assertEquals(IRLifecycleCallableRole.PROJECT_INIT, declaration.projectInit().role());
assertEquals(IRLifecycleCallableRole.FRAME_ROOT, declaration.frameRoot().role());
assertTrue(irBackend.getSyntheticFunctions().stream().noneMatch(function ->
function.kind() == IRSyntheticCallableKind.MODULE_INIT
|| function.kind() == IRSyntheticCallableKind.PROJECT_INIT
|| function.kind() == IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER));
}
@Test
@@ -73,7 +73,7 @@ public record BytecodeModule(
final var out = new ByteArrayOutputStream();
out.writeBytes(new byte[] { 'P', 'B', 'S', 0 });
out.writeBytes(new byte[] { 'P', 'B', 'X', 0 });
writeU16(out, version & 0xFFFF);
out.write(0);
out.write(0);
@@ -343,4 +343,3 @@ public record BytecodeModule(
}
}
}
@@ -40,7 +40,7 @@ public class LowerToIRVMService {
"IRBackend has no executable functions");
}
validatePbsLifecycleStructure(backend);
validateCommonLifecycleStructure(backend);
final var ordered = orderFunctions(backend);
final var funcIdByCallableId = new HashMap<CallableId, Integer>();
for (var i = 0; i < ordered.size(); i++) {
@@ -499,7 +499,9 @@ public class LowerToIRVMService {
return ReadOnlyList.wrap(ordered);
}
private void validatePbsLifecycleStructure(final IRBackend backend) {
private void validateCommonLifecycleStructure(final IRBackend backend) {
// Backend precondition on an already-assembled handoff. Declaration
// validation and materialization belong to LifecycleAssemblerService.
if (backend.getSyntheticFunctions().isEmpty()) {
return;
}
@@ -0,0 +1,532 @@
package p.studio.compiler.lifecycle;
import p.studio.compiler.messages.BuildingIssueSink;
import p.studio.compiler.models.*;
import p.studio.compiler.source.Span;
import p.studio.compiler.source.identifiers.CallableId;
import p.studio.compiler.source.identifiers.ModuleId;
import p.studio.compiler.source.tables.CallableSignatureRef;
import p.studio.utilities.structures.ReadOnlyList;
import java.util.ArrayList;
import java.util.HashSet;
import java.util.LinkedHashMap;
import java.util.Set;
public final class LifecycleAssemblerService {
public static final String PHASE = "lifecycle-assembly";
public static final String INVALID_DECLARATION = "COMMON_LIFECYCLE_INVALID_DECLARATION";
public IRBackend assemble(
final IRBackend backend,
final BuildingIssueSink issues) {
if (backend == null) {
report(issues, null, "IRBackend is required for lifecycle assembly");
return null;
}
final var declaration = backend.getLifecycleDeclaration();
if (declaration == null) {
return backend;
}
if (!validate(backend, declaration, issues)) {
return backend;
}
final var callableSignatures = new ArrayList<>(backend.getCallableSignatures().asList());
final var executableFunctions = new ArrayList<>(backend.getExecutableFunctions().asList());
final var syntheticFunctions = new ArrayList<>(backend.getSyntheticFunctions().asList());
final var globals = new ArrayList<>(backend.getGlobals().asList());
final var moduleInitCallableIds = new LinkedHashMap<ModuleId, CallableId>();
for (final var module : declaration.modules()) {
if (module.fileInitFragments().isEmpty()) {
continue;
}
final var moduleId = module.moduleId();
final var moduleInitName = moduleInitCallableName(moduleId);
final var moduleInitCallableId = new CallableId(callableSignatures.size());
callableSignatures.add(new CallableSignatureRef(moduleId, moduleInitName, 0, "() -> unit"));
moduleInitCallableIds.put(moduleId, moduleInitCallableId);
final var instructions = new ArrayList<IRBackendExecutableFunction.Instruction>();
IRBackendExecutableFunction anchor = null;
for (final var fragment : module.fileInitFragments()) {
final var executable = resolveExecutable(backend, fragment);
instructions.add(callInstruction(executable, fragment.origin()));
anchor = executable;
}
instructions.add(retInstruction(anchor.span()));
executableFunctions.add(new IRBackendExecutableFunction(
anchor.fileId(),
moduleId,
moduleInitName,
moduleInitCallableId,
anchor.sourceStart(),
anchor.sourceEnd(),
0,
0,
0,
1,
ReadOnlyList.wrap(instructions),
anchor.span()));
syntheticFunctions.add(new IRSyntheticFunction(
moduleId,
moduleInitName,
IRSyntheticCallableKind.MODULE_INIT,
new IRSyntheticOrigin(anchor.fileId(), anchor.callableName(), anchor.span())));
}
final var frameExecutable = resolveExecutable(backend, declaration.frameRoot());
final var entryPointModuleId = frameExecutable.moduleId();
CallableId projectInitCallableId = null;
if (declaration.projectInit() != null) {
final var projectInitTarget = resolveExecutable(backend, declaration.projectInit());
final var projectInitName = projectInitCallableName(entryPointModuleId);
projectInitCallableId = new CallableId(callableSignatures.size());
callableSignatures.add(new CallableSignatureRef(
entryPointModuleId,
projectInitName,
0,
"() -> unit"));
executableFunctions.add(new IRBackendExecutableFunction(
projectInitTarget.fileId(),
entryPointModuleId,
projectInitName,
projectInitCallableId,
projectInitTarget.sourceStart(),
projectInitTarget.sourceEnd(),
0,
0,
0,
1,
ReadOnlyList.from(
callInstruction(projectInitTarget, projectInitTarget.span()),
retInstruction(projectInitTarget.span())),
projectInitTarget.span()));
syntheticFunctions.add(new IRSyntheticFunction(
entryPointModuleId,
projectInitName,
IRSyntheticCallableKind.PROJECT_INIT,
new IRSyntheticOrigin(
projectInitTarget.fileId(),
projectInitTarget.callableName(),
projectInitTarget.span())));
}
final var entryPointCallableName = frameWrapperCallableName(entryPointModuleId);
final var bootGuardSlot = nextGlobalSlot(globals, entryPointModuleId);
final var wrapperCallableId = new CallableId(callableSignatures.size());
callableSignatures.add(new CallableSignatureRef(
entryPointModuleId,
entryPointCallableName,
0,
"() -> unit"));
final var wrapperInstructions = new ArrayList<IRBackendExecutableFunction.Instruction>();
wrapperInstructions.add(getGlobalInstruction(bootGuardSlot, frameExecutable.span()));
wrapperInstructions.add(pushBoolInstruction(true, frameExecutable.span()));
wrapperInstructions.add(eqInstruction(frameExecutable.span()));
wrapperInstructions.add(jumpIfTrueInstruction("bootstrap_done", frameExecutable.span()));
for (final var module : declaration.modules()) {
final var moduleInitCallableId = moduleInitCallableIds.get(module.moduleId());
if (moduleInitCallableId == null) {
continue;
}
wrapperInstructions.add(new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.CALL_FUNC,
module.moduleId(),
moduleInitCallableName(module.moduleId()),
moduleInitCallableId,
null,
null,
0,
0,
frameExecutable.span()));
}
if (projectInitCallableId != null) {
wrapperInstructions.add(new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.CALL_FUNC,
entryPointModuleId,
projectInitCallableName(entryPointModuleId),
projectInitCallableId,
null,
null,
0,
0,
frameExecutable.span()));
}
wrapperInstructions.add(pushBoolInstruction(true, frameExecutable.span()));
wrapperInstructions.add(setGlobalInstruction(bootGuardSlot, frameExecutable.span()));
wrapperInstructions.add(labelInstruction("bootstrap_done", frameExecutable.span()));
wrapperInstructions.add(callInstruction(frameExecutable, frameExecutable.span()));
wrapperInstructions.add(retInstruction(frameExecutable.span()));
executableFunctions.add(new IRBackendExecutableFunction(
frameExecutable.fileId(),
entryPointModuleId,
entryPointCallableName,
wrapperCallableId,
frameExecutable.sourceStart(),
frameExecutable.sourceEnd(),
0,
0,
0,
2,
ReadOnlyList.wrap(wrapperInstructions),
frameExecutable.span()));
syntheticFunctions.add(new IRSyntheticFunction(
entryPointModuleId,
entryPointCallableName,
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(
frameExecutable.fileId(),
frameExecutable.callableName(),
frameExecutable.span())));
globals.add(new IRGlobal(
frameExecutable.fileId(),
entryPointModuleId,
bootGuardGlobalName(entryPointModuleId),
"bool",
bootGuardSlot,
IRGlobalVisibility.MODULE,
true,
IRHiddenGlobalKind.BOOT_GUARD,
new IRGlobalOrigin(
frameExecutable.fileId(),
entryPointModuleId,
frameExecutable.callableName(),
frameExecutable.span(),
"BOOT_GUARD"),
frameExecutable.span()));
return IRBackend.builder()
.entryPointCallableName(entryPointCallableName)
.entryPointModuleId(entryPointModuleId)
.functions(backend.getFunctions())
.syntheticFunctions(ReadOnlyList.wrap(syntheticFunctions))
.globals(ReadOnlyList.wrap(normalizeGlobalSlots(globals)))
.executableFunctions(ReadOnlyList.wrap(executableFunctions))
.modulePool(backend.getModulePool())
.callableSignatures(ReadOnlyList.wrap(callableSignatures))
.intrinsicPool(backend.getIntrinsicPool())
.reservedMetadata(backend.getReservedMetadata())
.lifecycleDeclaration(declaration)
.build();
}
private boolean validate(
final IRBackend backend,
final IRLifecycleDeclaration declaration,
final BuildingIssueSink issues) {
var valid = true;
for (final var synthetic : backend.getSyntheticFunctions()) {
if (synthetic.kind() != IRSyntheticCallableKind.FILE_INIT_FRAGMENT) {
report(issues, synthetic.origin().span(),
"frontend handoff contains pre-materialized lifecycle artifact: " + synthetic.kind());
valid = false;
}
}
for (final var global : backend.getGlobals()) {
if (global.hiddenKind() == IRHiddenGlobalKind.BOOT_GUARD) {
report(issues, global.span(), "frontend handoff contains pre-materialized boot guard");
valid = false;
}
}
final Set<Integer> moduleIds = new HashSet<>();
final Set<Integer> lifecycleCallableIds = new HashSet<>();
for (final var module : declaration.modules()) {
final var moduleIndex = module.moduleId().isNone() ? -1 : module.moduleId().getIndex();
if (!moduleIds.add(moduleIndex)) {
report(issues, null, "duplicate lifecycle module: " + moduleIndex);
valid = false;
}
for (final var fragment : module.fileInitFragments()) {
valid &= validateReference(
backend,
fragment,
IRLifecycleCallableRole.FILE_INIT_FRAGMENT,
module.moduleId(),
lifecycleCallableIds,
issues);
}
}
if (declaration.projectInit() != null) {
valid &= validateReference(
backend,
declaration.projectInit(),
IRLifecycleCallableRole.PROJECT_INIT,
declaration.projectInit().moduleId(),
lifecycleCallableIds,
issues);
}
if (declaration.frameRoot() == null) {
report(issues, null, "executable lifecycle declaration is missing frame root");
valid = false;
} else {
valid &= validateReference(
backend,
declaration.frameRoot(),
IRLifecycleCallableRole.FRAME_ROOT,
declaration.frameRoot().moduleId(),
lifecycleCallableIds,
issues);
}
return valid;
}
private boolean validateReference(
final IRBackend backend,
final IRLifecycleCallableReference reference,
final IRLifecycleCallableRole expectedRole,
final ModuleId expectedModuleId,
final Set<Integer> lifecycleCallableIds,
final BuildingIssueSink issues) {
if (reference == null) {
report(issues, null, "missing lifecycle reference for role " + expectedRole);
return false;
}
var valid = true;
if (reference.role() != expectedRole) {
report(issues, reference.origin(), "incompatible lifecycle role: expected "
+ expectedRole + " but found " + reference.role());
valid = false;
}
if (!sameModuleId(reference.moduleId(), expectedModuleId)) {
report(issues, reference.origin(), "lifecycle reference module does not match declaration module");
valid = false;
}
if (reference.origin() == null || reference.origin().isNone()) {
report(issues, reference.origin(), "lifecycle reference origin is missing for role " + expectedRole);
valid = false;
}
final var callableIndex = reference.callableId().getIndex();
if (!lifecycleCallableIds.add(callableIndex)) {
report(issues, reference.origin(), "callable is assigned to multiple lifecycle roles: " + callableIndex);
valid = false;
}
if (callableIndex < 0 || callableIndex >= backend.getCallableSignatures().size()) {
report(issues, reference.origin(), "unresolved lifecycle callable id: " + callableIndex);
return false;
}
final var signature = backend.getCallableSignatures().get(callableIndex);
if (!sameModuleId(signature.moduleId(), reference.moduleId())) {
report(issues, reference.origin(), "lifecycle callable signature module mismatch: " + callableIndex);
valid = false;
}
if (resolveExecutable(backend, reference) == null) {
report(issues, reference.origin(), "lifecycle callable has no executable body: " + callableIndex);
valid = false;
}
return valid;
}
private IRBackendExecutableFunction resolveExecutable(
final IRBackend backend,
final IRLifecycleCallableReference reference) {
if (reference == null) {
return null;
}
for (final var executable : backend.getExecutableFunctions()) {
if (executable.callableId().getIndex() == reference.callableId().getIndex()
&& sameModuleId(executable.moduleId(), reference.moduleId())) {
return executable;
}
}
return null;
}
private void report(
final BuildingIssueSink issues,
final Span span,
final String message) {
final var effectiveSpan = span == null ? Span.none() : span;
issues.report(builder -> builder
.error(true)
.phase(PHASE)
.code(INVALID_DECLARATION)
.fileId(effectiveSpan.getFileId().isNone() ? null : effectiveSpan.getFileId().getId())
.start(safeToInt(effectiveSpan.getStart()))
.end(safeToInt(effectiveSpan.getEnd()))
.message(message));
}
private int nextGlobalSlot(
final ArrayList<IRGlobal> globals,
final ModuleId moduleId) {
var nextSlot = 0;
for (final var global : globals) {
if (sameModuleId(moduleId, global.moduleId())) {
nextSlot = Math.max(nextSlot, global.slot() + 1);
}
}
return nextSlot;
}
private ArrayList<IRGlobal> normalizeGlobalSlots(final ArrayList<IRGlobal> globals) {
final var nextSlotByModule = new LinkedHashMap<ModuleId, Integer>();
final var normalized = new ArrayList<IRGlobal>(globals.size());
for (final var global : globals) {
final var moduleId = normalizeModuleId(global.moduleId());
final var slot = nextSlotByModule.getOrDefault(moduleId, 0);
nextSlotByModule.put(moduleId, slot + 1);
normalized.add(new IRGlobal(
global.fileId(),
moduleId,
global.name(),
global.declaredTypeSurface(),
slot,
global.visibility(),
global.isHidden(),
global.hiddenKind(),
global.origin(),
global.span()));
}
return normalized;
}
private IRBackendExecutableFunction.Instruction callInstruction(
final IRBackendExecutableFunction callee,
final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.CALL_FUNC,
callee.moduleId(),
callee.callableName(),
callee.callableId(),
null,
null,
callee.paramSlots(),
callee.returnSlots(),
span);
}
private IRBackendExecutableFunction.Instruction retInstruction(final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.RET,
"",
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction getGlobalInstruction(final int slot, final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.GET_GLOBAL,
"",
null,
null,
null,
slot,
null,
span);
}
private IRBackendExecutableFunction.Instruction setGlobalInstruction(final int slot, final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.SET_GLOBAL,
"",
null,
null,
null,
slot,
null,
span);
}
private IRBackendExecutableFunction.Instruction pushBoolInstruction(final boolean value, final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.PUSH_BOOL,
"",
null,
null,
null,
value ? "true" : "false",
null,
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction eqInstruction(final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.EQ,
"",
null,
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction labelInstruction(final String label, final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.LABEL,
"",
null,
null,
null,
label,
null,
null,
null,
span);
}
private IRBackendExecutableFunction.Instruction jumpIfTrueInstruction(final String label, final Span span) {
return new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.JMP_IF_TRUE,
"",
null,
null,
null,
null,
label,
null,
null,
span);
}
private ModuleId normalizeModuleId(final ModuleId moduleId) {
return moduleId == null ? ModuleId.none() : moduleId;
}
private boolean sameModuleId(final ModuleId left, final ModuleId right) {
final var normalizedLeft = normalizeModuleId(left);
final var normalizedRight = normalizeModuleId(right);
if (normalizedLeft.isNone() || normalizedRight.isNone()) {
return normalizedLeft.isNone() && normalizedRight.isNone();
}
return normalizedLeft.getIndex() == normalizedRight.getIndex();
}
private String moduleInitCallableName(final ModuleId moduleId) {
return "__prometeu.lifecycle.module_init$m" + normalizedModuleIndex(moduleId);
}
private String projectInitCallableName(final ModuleId moduleId) {
return "__prometeu.lifecycle.project_init$m" + normalizedModuleIndex(moduleId);
}
private String frameWrapperCallableName(final ModuleId moduleId) {
return "__prometeu.lifecycle.frame_wrapper$m" + normalizedModuleIndex(moduleId);
}
private String bootGuardGlobalName(final ModuleId moduleId) {
return "__prometeu.lifecycle.boot_guard$m" + normalizedModuleIndex(moduleId);
}
private int normalizedModuleIndex(final ModuleId moduleId) {
final var normalized = normalizeModuleId(moduleId);
return normalized.isNone() ? -1 : normalized.getIndex();
}
private int safeToInt(final long value) {
if (value > Integer.MAX_VALUE) {
return Integer.MAX_VALUE;
}
if (value < Integer.MIN_VALUE) {
return Integer.MIN_VALUE;
}
return (int) value;
}
}
@@ -23,7 +23,8 @@ public class BuilderPipelineService {
final var analyses = List.<PipelineStage>of(
new ResolveDepsPipelineStage(),
new LoadSourcesPipelineStage(),
new FrontendPhasePipelineStage()
new FrontendPhasePipelineStage(),
new AssembleLifecyclePipelineStage()
);
final var compile = List.<PipelineStage>of(
@@ -0,0 +1,32 @@
package p.studio.compiler.workspaces.stages;
import lombok.extern.slf4j.Slf4j;
import p.studio.compiler.lifecycle.LifecycleAssemblerService;
import p.studio.compiler.messages.BuildingIssueSink;
import p.studio.compiler.models.BuilderPipelineContext;
import p.studio.compiler.workspaces.PipelineStage;
import p.studio.utilities.logs.LogAggregator;
@Slf4j
public final class AssembleLifecyclePipelineStage implements PipelineStage {
private final LifecycleAssemblerService assembler;
public AssembleLifecyclePipelineStage() {
this(new LifecycleAssemblerService());
}
AssembleLifecyclePipelineStage(final LifecycleAssemblerService assembler) {
this.assembler = assembler;
}
@Override
public BuildingIssueSink run(final BuilderPipelineContext ctx, final LogAggregator logs) {
final var issues = BuildingIssueSink.empty();
final var assembled = assembler.assemble(ctx.irBackend, issues);
if (!issues.hasErrors()) {
ctx.irBackend = assembled;
logs.using(log).debug("Common lifecycle assembly completed successfully!");
}
return issues;
}
}
@@ -17,13 +17,7 @@ public class FrontendPhasePipelineStage implements PipelineStage {
@Override
public BuildingIssueSink run(final BuilderPipelineContext ctx, final LogAggregator logs) {
final var frontedSpec = ctx.resolvedWorkspace.frontendSpec();
final var service = FrontendRegistryService.getFrontendPhaseService(frontedSpec.getLanguageId());
if (service.isEmpty()) {
return BuildingIssueSink.empty()
.report(builder -> builder
.error(true)
.message("[BUILD]: unable to find a service for frontend phase: " + frontedSpec.getLanguageId()));
}
final var provider = FrontendRegistryService.require(frontedSpec.getLanguageId());
final var projectTable = ctx.resolvedWorkspace.graph().projectTable();
final var fileTable = ctx.fileTable;
final var frontendPhaseContext = new FrontendPhaseContext(
@@ -36,7 +30,7 @@ public class FrontendPhasePipelineStage implements PipelineStage {
assetSurfaceContextLoader.load(ctx.resolvedWorkspace.mainProject().getRootPath()));
final var diagnostics = DiagnosticSink.empty();
final var issues = BuildingIssueSink.empty();
ctx.irBackend = service.get().compile(frontendPhaseContext, diagnostics, logs, issues);
ctx.irBackend = provider.compiler().compile(frontendPhaseContext, diagnostics, logs, issues);
logs.using(log).debug("Frontend phase completed successfully!");
adaptDiagnostics(diagnostics, issues);
return issues;
@@ -18,7 +18,7 @@ class BytecodeModuleTest {
assertEquals('P', bytes[0]);
assertEquals('B', bytes[1]);
assertEquals('S', bytes[2]);
assertEquals('X', bytes[2]);
assertEquals(0, bytes[3]);
final var sectionCount = readU32(bytes, 8);
@@ -77,4 +77,3 @@ class BytecodeModuleTest {
return ByteBuffer.wrap(bytes, offset, 4).order(ByteOrder.LITTLE_ENDIAN).getInt();
}
}
@@ -67,31 +67,31 @@ class LowerToIRVMServiceTest {
@Test
void lowerMustPublishSyntheticWrapperAtFunctionZeroAndKeepFinalRetInWrapperPath() {
final var backend = IRBackend.builder()
.entryPointCallableName("__pbs.frame_wrapper$m0")
.entryPointCallableName("__prometeu.lifecycle.frame_wrapper$m0")
.entryPointModuleId(new ModuleId(0))
.globals(ReadOnlyList.from(hiddenBootGuard(0)))
.syntheticFunctions(ReadOnlyList.from(new IRSyntheticFunction(
new ModuleId(0),
"__pbs.frame_wrapper$m0",
"__prometeu.lifecycle.frame_wrapper$m0",
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(new FileId(0), "frame", new Span(new FileId(0), 1, 10)))))
.executableFunctions(ReadOnlyList.from(
fn("__pbs.file_init$f1", "app", 30, ReadOnlyList.from(ret())),
fn("__pbs.module_init$m0", "app", 31, ReadOnlyList.from(
callFunc("app", "__pbs.file_init$f1", 30),
fn("__prometeu.lifecycle.file_init$f1", "app", 30, ReadOnlyList.from(ret())),
fn("__prometeu.lifecycle.module_init$m0", "app", 31, ReadOnlyList.from(
callFunc("app", "__prometeu.lifecycle.file_init$f1", 30),
ret())),
fn("__pbs.project_init$m0", "app", 32, ReadOnlyList.from(
fn("__prometeu.lifecycle.project_init$m0", "app", 32, ReadOnlyList.from(
callFunc("app", "boot", 10),
ret())),
fn("boot", "app", 10, ReadOnlyList.from(ret())),
fn("frame", "app", 20, ReadOnlyList.from(ret())),
fn("__pbs.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
fn("__prometeu.lifecycle.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
getGlobal(0),
pushBool(true),
eq(),
jmpIfTrue("bootstrap_done"),
callFunc("app", "__pbs.module_init$m0", 31),
callFunc("app", "__pbs.project_init$m0", 32),
callFunc("app", "__prometeu.lifecycle.module_init$m0", 31),
callFunc("app", "__prometeu.lifecycle.project_init$m0", 32),
pushBool(true),
setGlobal(0),
label("bootstrap_done"),
@@ -101,7 +101,7 @@ class LowerToIRVMServiceTest {
final var lowered = new LowerToIRVMService().lower(backend);
assertEquals("__pbs.frame_wrapper$m0", lowered.module().functions().get(0).name());
assertEquals("__prometeu.lifecycle.frame_wrapper$m0", lowered.module().functions().get(0).name());
assertEquals(IRVMOp.GET_GLOBAL, lowered.module().functions().get(0).instructions().get(0).op());
assertEquals(IRVMOp.PUSH_BOOL, lowered.module().functions().get(0).instructions().get(1).op());
assertEquals(IRVMOp.EQ, lowered.module().functions().get(0).instructions().get(2).op());
@@ -124,12 +124,12 @@ class LowerToIRVMServiceTest {
@Test
void lowerMustRejectWhenSyntheticWrapperEntrypointIsMissing() {
final var backend = IRBackend.builder()
.entryPointCallableName("__pbs.frame_wrapper$m0")
.entryPointCallableName("__prometeu.lifecycle.frame_wrapper$m0")
.entryPointModuleId(new ModuleId(0))
.globals(ReadOnlyList.from(hiddenBootGuard(0)))
.syntheticFunctions(ReadOnlyList.from(new IRSyntheticFunction(
new ModuleId(0),
"__pbs.frame_wrapper$m0",
"__prometeu.lifecycle.frame_wrapper$m0",
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(new FileId(0), "frame", new Span(new FileId(0), 1, 10)))))
.executableFunctions(ReadOnlyList.from(
@@ -148,11 +148,11 @@ class LowerToIRVMServiceTest {
.globals(ReadOnlyList.from(hiddenBootGuard(0)))
.syntheticFunctions(ReadOnlyList.from(new IRSyntheticFunction(
new ModuleId(0),
"__pbs.frame_wrapper$m0",
"__prometeu.lifecycle.frame_wrapper$m0",
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(new FileId(0), "frame", new Span(new FileId(0), 1, 10)))))
.executableFunctions(ReadOnlyList.from(
fn("__pbs.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
fn("__prometeu.lifecycle.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
callFunc("app", "frame", 20),
ret())),
fn("frame", "app", 20, ReadOnlyList.from(ret()))))
@@ -165,15 +165,15 @@ class LowerToIRVMServiceTest {
@Test
void lowerMustRejectWhenHiddenBootGuardIsMissing() {
final var backend = IRBackend.builder()
.entryPointCallableName("__pbs.frame_wrapper$m0")
.entryPointCallableName("__prometeu.lifecycle.frame_wrapper$m0")
.entryPointModuleId(new ModuleId(0))
.syntheticFunctions(ReadOnlyList.from(new IRSyntheticFunction(
new ModuleId(0),
"__pbs.frame_wrapper$m0",
"__prometeu.lifecycle.frame_wrapper$m0",
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(new FileId(0), "frame", new Span(new FileId(0), 1, 10)))))
.executableFunctions(ReadOnlyList.from(
fn("__pbs.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
fn("__prometeu.lifecycle.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
callFunc("app", "frame", 20),
ret())),
fn("frame", "app", 20, ReadOnlyList.from(ret()))))
@@ -186,16 +186,16 @@ class LowerToIRVMServiceTest {
@Test
void lowerMustRejectWhenSyntheticCallableOriginIsMissing() {
final var backend = IRBackend.builder()
.entryPointCallableName("__pbs.frame_wrapper$m0")
.entryPointCallableName("__prometeu.lifecycle.frame_wrapper$m0")
.entryPointModuleId(new ModuleId(0))
.globals(ReadOnlyList.from(userVisibleGlobal("__pbs.boot_guard$m0", 0)))
.globals(ReadOnlyList.from(userVisibleGlobal("__prometeu.lifecycle.boot_guard$m0", 0)))
.syntheticFunctions(ReadOnlyList.from(new IRSyntheticFunction(
new ModuleId(0),
"__pbs.frame_wrapper$m0",
"__prometeu.lifecycle.frame_wrapper$m0",
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(new FileId(0), "", Span.none()))))
.executableFunctions(ReadOnlyList.from(
fn("__pbs.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
fn("__prometeu.lifecycle.frame_wrapper$m0", "app", 40, ReadOnlyList.from(
callFunc("app", "frame", 20),
ret())),
fn("frame", "app", 20, ReadOnlyList.from(ret()))))
@@ -229,6 +229,35 @@ class LowerToIRVMServiceTest {
assertTrue(emissionOps.stream().allMatch(op -> op.span() != null));
}
@Test
void lowerMustAcceptManuallyConstructedCommonIRBackend() {
final var backend = IRBackend.builder()
.entryPointCallableName("tick")
.entryPointModuleId(new ModuleId(0))
.modulePool(ReadOnlyList.from(new ModuleReference("game", ReadOnlyList.from("runtime"))))
.executableFunctions(ReadOnlyList.from(
fnWithModule("tick", "game/runtime", 0, 10, ReadOnlyList.from(
callHost("timer", "poll", 1, 0, 0),
callIntrinsic("input.pad", 1, 0),
callFuncWithExpected("game/runtime", "helper", 20, 0, 0),
ret())),
fnWithModule("helper", "game/runtime", 0, 20, ReadOnlyList.from(
ret()))))
.intrinsicPool(ReadOnlyList.from(new IntrinsicReference("input.pad", 1)))
.build();
final var lowered = new LowerToIRVMService().lower(backend);
assertEquals(2, lowered.module().functions().size());
assertEquals("tick", lowered.module().functions().getFirst().name());
final var instructions = lowered.module().functions().getFirst().instructions();
assertEquals(IRVMOp.HOSTCALL, instructions.get(0).op());
assertEquals(IRVMOp.INTRINSIC, instructions.get(1).op());
assertEquals(IRVMOp.CALL, instructions.get(2).op());
assertEquals(1, instructions.get(2).immediate());
assertEquals(IRVMOp.RET, instructions.get(3).op());
}
@Test
void lowerMustRejectUnterminatedFunction() {
final var backend = IRBackend.builder()
@@ -421,7 +450,7 @@ class LowerToIRVMServiceTest {
return new IRGlobal(
fileId,
ownerModuleId,
"__pbs.boot_guard$m" + moduleId,
"__prometeu.lifecycle.boot_guard$m" + moduleId,
"bool",
0,
IRGlobalVisibility.MODULE,
@@ -0,0 +1,367 @@
package p.studio.compiler.lifecycle;
import org.junit.jupiter.api.Test;
import p.studio.compiler.messages.BuildingIssueSink;
import p.studio.compiler.models.IRBackend;
import p.studio.compiler.models.IRBackendExecutableFunction;
import p.studio.compiler.models.IRGlobal;
import p.studio.compiler.models.IRGlobalOrigin;
import p.studio.compiler.models.IRGlobalVisibility;
import p.studio.compiler.models.IRHiddenGlobalKind;
import p.studio.compiler.models.IRLifecycleCallableReference;
import p.studio.compiler.models.IRLifecycleCallableRole;
import p.studio.compiler.models.IRLifecycleDeclaration;
import p.studio.compiler.models.IRLifecycleModuleDeclaration;
import p.studio.compiler.models.IRSyntheticCallableKind;
import p.studio.compiler.models.IRSyntheticFunction;
import p.studio.compiler.models.IRSyntheticOrigin;
import p.studio.compiler.source.Span;
import p.studio.compiler.source.identifiers.CallableId;
import p.studio.compiler.source.identifiers.FileId;
import p.studio.compiler.source.identifiers.ModuleId;
import p.studio.compiler.source.tables.CallableSignatureRef;
import p.studio.compiler.source.tables.ModuleReference;
import p.studio.utilities.structures.ReadOnlyList;
import java.util.ArrayList;
import java.util.List;
import java.util.Arrays;
import static org.junit.jupiter.api.Assertions.*;
class LifecycleAssemblerServiceTest {
@Test
void assembleMustMaterializeNeutralLifecycleArtifactsInDeclaredOrder() {
final var backend = declarativeBackend();
final var issues = BuildingIssueSink.empty();
final var assembled = new LifecycleAssemblerService().assemble(backend, issues);
assertFalse(issues.hasErrors(), issues.asCollection().toString());
assertSame(backend.getLifecycleDeclaration(), assembled.getLifecycleDeclaration());
assertEquals("__prometeu.lifecycle.frame_wrapper$m1", assembled.getEntryPointCallableName());
assertEquals(1, assembled.getEntryPointModuleId().getIndex());
assertTrue(assembled.getSyntheticFunctions().stream().anyMatch(fn ->
fn.kind() == IRSyntheticCallableKind.MODULE_INIT
&& "__prometeu.lifecycle.module_init$m0".equals(fn.callableName())));
assertTrue(assembled.getSyntheticFunctions().stream().anyMatch(fn ->
fn.kind() == IRSyntheticCallableKind.PROJECT_INIT
&& "__prometeu.lifecycle.project_init$m1".equals(fn.callableName())));
assertTrue(assembled.getSyntheticFunctions().stream().anyMatch(fn ->
fn.kind() == IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER
&& "__prometeu.lifecycle.frame_wrapper$m1".equals(fn.callableName())));
assertTrue(assembled.getGlobals().stream().anyMatch(global ->
global.hiddenKind() == IRHiddenGlobalKind.BOOT_GUARD
&& global.isHidden()
&& "__prometeu.lifecycle.boot_guard$m1".equals(global.name())));
assertTrue(assembled.getSyntheticFunctions().stream().noneMatch(fn ->
fn.callableName().contains("__pbs.")));
assertTrue(assembled.getGlobals().stream().noneMatch(global ->
global.name().contains("__pbs.")));
final var wrapper = executable(assembled, "__prometeu.lifecycle.frame_wrapper$m1");
final var calledNames = wrapper.instructions().stream()
.filter(instruction -> instruction.kind() == IRBackendExecutableFunction.InstructionKind.CALL_FUNC)
.map(IRBackendExecutableFunction.Instruction::calleeCallableName)
.toList();
assertEquals(
List.of(
"__prometeu.lifecycle.module_init$m0",
"__prometeu.lifecycle.module_init$m1",
"__prometeu.lifecycle.project_init$m1",
"frame"),
calledNames);
assertEquals(
1,
calledNames.stream().filter("__prometeu.lifecycle.project_init$m1"::equals).count());
assertEquals(IRBackendExecutableFunction.InstructionKind.RET, wrapper.instructions().getLast().kind());
final var moduleInit = executable(assembled, "__prometeu.lifecycle.module_init$m0");
assertEquals(
List.of("file_init_dep"),
moduleInit.instructions().stream()
.filter(instruction -> instruction.kind() == IRBackendExecutableFunction.InstructionKind.CALL_FUNC)
.map(IRBackendExecutableFunction.Instruction::calleeCallableName)
.toList());
final var hostModuleInit = executable(assembled, "__prometeu.lifecycle.module_init$m1");
assertEquals(
List.of("file_init_host"),
hostModuleInit.instructions().stream()
.filter(instruction -> instruction.kind() == IRBackendExecutableFunction.InstructionKind.CALL_FUNC)
.map(IRBackendExecutableFunction.Instruction::calleeCallableName)
.toList());
}
@Test
void assembleMustInvokeProjectInitExactlyOnceAndExcludeItFromFileInit() {
final var assembled = new LifecycleAssemblerService().assemble(declarativeBackend(), BuildingIssueSink.empty());
final var wrapperCalls = executable(assembled, assembled.getEntryPointCallableName()).instructions().stream()
.filter(instruction -> instruction.kind() == IRBackendExecutableFunction.InstructionKind.CALL_FUNC)
.map(IRBackendExecutableFunction.Instruction::calleeCallableName)
.toList();
final var fileInitCalls = executable(assembled, "file_init_host").instructions().stream()
.filter(instruction -> instruction.kind() == IRBackendExecutableFunction.InstructionKind.CALL_FUNC)
.map(IRBackendExecutableFunction.Instruction::calleeCallableName)
.toList();
assertEquals(1, wrapperCalls.stream().filter("__prometeu.lifecycle.project_init$m1"::equals).count());
assertTrue(wrapperCalls.contains("frame"));
assertTrue(fileInitCalls.stream().noneMatch("boot"::equals));
final var projectInitWrapper = executable(assembled, "__prometeu.lifecycle.project_init$m1");
assertEquals(
List.of("boot"),
projectInitWrapper.instructions().stream()
.filter(instruction -> instruction.kind() == IRBackendExecutableFunction.InstructionKind.CALL_FUNC)
.map(IRBackendExecutableFunction.Instruction::calleeCallableName)
.toList());
}
@Test
void assembleMustNoOpWhenLifecycleDeclarationIsAbsent() {
final var backend = IRBackend.builder()
.entryPointCallableName("main")
.entryPointModuleId(new ModuleId(0))
.executableFunctions(ReadOnlyList.from(fn("main", 0, 0, 0, span(1))))
.callableSignatures(ReadOnlyList.from(new CallableSignatureRef(new ModuleId(0), "main", 0, "() -> unit")))
.build();
final var issues = BuildingIssueSink.empty();
final var assembled = new LifecycleAssemblerService().assemble(backend, issues);
assertFalse(issues.hasErrors());
assertSame(backend, assembled);
assertNull(assembled.getLifecycleDeclaration());
assertTrue(assembled.getSyntheticFunctions().isEmpty());
}
@Test
void assembleMustRejectMissingFrameRootWithoutRepair() {
final var backend = backendWithDeclaration(new IRLifecycleDeclaration(
ReadOnlyList.from(new IRLifecycleModuleDeclaration(new ModuleId(0), ReadOnlyList.empty())),
null,
null));
final var issues = BuildingIssueSink.empty();
final var assembled = new LifecycleAssemblerService().assemble(backend, issues);
assertTrue(issues.hasErrors());
assertSame(backend, assembled);
assertEquals(LifecycleAssemblerService.INVALID_DECLARATION, firstCode(issues));
assertTrue(firstMessage(issues).contains("frame root"));
}
@Test
void assembleMustRejectUnresolvedCallableIds() {
final var backend = backendWithDeclaration(new IRLifecycleDeclaration(
ReadOnlyList.empty(),
null,
reference(0, 1, 99, IRLifecycleCallableRole.FRAME_ROOT, span(3))));
final var issues = BuildingIssueSink.empty();
new LifecycleAssemblerService().assemble(backend, issues);
assertTrue(issues.hasErrors());
assertTrue(firstMessage(issues).contains("unresolved lifecycle callable id"));
}
@Test
void assembleMustRejectIncompatibleRoles() {
final var backend = backendWithDeclaration(new IRLifecycleDeclaration(
ReadOnlyList.empty(),
null,
reference(0, 1, 2, IRLifecycleCallableRole.PROJECT_INIT, span(3))));
final var issues = BuildingIssueSink.empty();
new LifecycleAssemblerService().assemble(backend, issues);
assertTrue(issues.hasErrors());
assertTrue(firstMessage(issues).contains("incompatible lifecycle role"));
}
@Test
void assembleMustRejectMissingOrigin() {
final var backend = backendWithDeclaration(new IRLifecycleDeclaration(
ReadOnlyList.empty(),
null,
reference(0, 1, 2, IRLifecycleCallableRole.FRAME_ROOT, Span.none())));
final var issues = BuildingIssueSink.empty();
new LifecycleAssemblerService().assemble(backend, issues);
assertTrue(issues.hasErrors());
assertTrue(firstMessage(issues).contains("origin is missing"));
}
@Test
void assembleMustRejectPreMaterializedDerivedArtifacts() {
final var backend = declarativeBackend(new IRSyntheticFunction(
new ModuleId(1),
"already_wrapped",
IRSyntheticCallableKind.PUBLISHED_FRAME_WRAPPER,
new IRSyntheticOrigin(new FileId(2), "frame", span(2))));
final var issues = BuildingIssueSink.empty();
final var assembled = new LifecycleAssemblerService().assemble(backend, issues);
assertTrue(issues.hasErrors());
assertSame(backend, assembled);
assertTrue(firstMessage(issues).contains("pre-materialized lifecycle artifact"));
}
@Test
void assembleMustRejectDuplicateModuleContributions() {
final var module = new IRLifecycleModuleDeclaration(
new ModuleId(0),
ReadOnlyList.from(reference(0, 1, 0, IRLifecycleCallableRole.FILE_INIT_FRAGMENT, span(1))));
final var backend = backendWithDeclaration(new IRLifecycleDeclaration(
ReadOnlyList.from(module, module),
reference(1, 2, 3, IRLifecycleCallableRole.PROJECT_INIT, span(2)),
reference(1, 2, 2, IRLifecycleCallableRole.FRAME_ROOT, span(2))));
final var issues = BuildingIssueSink.empty();
new LifecycleAssemblerService().assemble(backend, issues);
assertTrue(issues.hasErrors());
assertTrue(firstMessage(issues).contains("duplicate lifecycle module"));
}
@Test
void assembleMustRejectNullBackend() {
final var issues = BuildingIssueSink.empty();
final var assembled = new LifecycleAssemblerService().assemble(null, issues);
assertNull(assembled);
assertTrue(issues.hasErrors());
assertEquals(LifecycleAssemblerService.PHASE, issues.asCollection().iterator().next().getPhase());
}
private static IRBackend declarativeBackend(final IRSyntheticFunction... extraSynthetics) {
return declarativeBackend(null, extraSynthetics);
}
private static IRBackend backendWithDeclaration(final IRLifecycleDeclaration declaration) {
return declarativeBackend(declaration);
}
private static IRBackend declarativeBackend(
final IRLifecycleDeclaration declarationOverride,
final IRSyntheticFunction... extraSynthetics) {
final var depSpan = span(1);
final var hostSpan = span(2);
final var synthetics = new ArrayList<IRSyntheticFunction>();
synthetics.add(new IRSyntheticFunction(
new ModuleId(0),
"file_init_dep",
IRSyntheticCallableKind.FILE_INIT_FRAGMENT,
new IRSyntheticOrigin(new FileId(1), "warmup", depSpan)));
synthetics.add(new IRSyntheticFunction(
new ModuleId(1),
"file_init_host",
IRSyntheticCallableKind.FILE_INIT_FRAGMENT,
new IRSyntheticOrigin(new FileId(2), "boot", hostSpan)));
synthetics.addAll(Arrays.asList(extraSynthetics));
final var declaration = declarationOverride != null
? declarationOverride
: new IRLifecycleDeclaration(
ReadOnlyList.from(
new IRLifecycleModuleDeclaration(
new ModuleId(0),
ReadOnlyList.from(reference(0, 1, 0, IRLifecycleCallableRole.FILE_INIT_FRAGMENT, depSpan))),
new IRLifecycleModuleDeclaration(
new ModuleId(1),
ReadOnlyList.from(reference(1, 2, 1, IRLifecycleCallableRole.FILE_INIT_FRAGMENT, hostSpan)))),
reference(1, 2, 3, IRLifecycleCallableRole.PROJECT_INIT, hostSpan),
reference(1, 2, 2, IRLifecycleCallableRole.FRAME_ROOT, hostSpan));
return IRBackend.builder()
.entryPointCallableName("frame")
.entryPointModuleId(new ModuleId(1))
.modulePool(ReadOnlyList.from(
new ModuleReference("app", ReadOnlyList.from("zzz")),
new ModuleReference("app", ReadOnlyList.from("aaa"))))
.callableSignatures(ReadOnlyList.from(
new CallableSignatureRef(new ModuleId(0), "file_init_dep", 0, "() -> unit"),
new CallableSignatureRef(new ModuleId(1), "file_init_host", 0, "() -> unit"),
new CallableSignatureRef(new ModuleId(1), "frame", 0, "() -> unit"),
new CallableSignatureRef(new ModuleId(1), "boot", 0, "() -> unit")))
.executableFunctions(ReadOnlyList.from(
fn("file_init_dep", 1, 0, 0, depSpan),
fn("file_init_host", 2, 1, 1, hostSpan),
fn("frame", 2, 1, 2, hostSpan),
fn("boot", 2, 1, 3, hostSpan)))
.syntheticFunctions(ReadOnlyList.wrap(synthetics))
.globals(ReadOnlyList.from(new IRGlobal(
new FileId(1),
new ModuleId(0),
"SCORE",
"int",
0,
IRGlobalVisibility.MODULE,
false,
IRHiddenGlobalKind.NONE,
new IRGlobalOrigin(new FileId(1), new ModuleId(0), "SCORE", depSpan, "SCORE"),
depSpan)))
.lifecycleDeclaration(declaration)
.build();
}
private static IRLifecycleCallableReference reference(
final int moduleId,
final int fileId,
final int callableId,
final IRLifecycleCallableRole role,
final Span origin) {
return new IRLifecycleCallableReference(
new ModuleId(moduleId),
new FileId(fileId),
new CallableId(callableId),
role,
origin);
}
private static IRBackendExecutableFunction fn(
final String name,
final int fileId,
final int moduleId,
final int callableId,
final Span span) {
return new IRBackendExecutableFunction(
new FileId(fileId),
new ModuleId(moduleId),
name,
new CallableId(callableId),
(int) span.getStart(),
(int) span.getEnd(),
0,
0,
0,
1,
ReadOnlyList.from(new IRBackendExecutableFunction.Instruction(
IRBackendExecutableFunction.InstructionKind.RET,
"",
null,
null,
span)),
span);
}
private static IRBackendExecutableFunction executable(final IRBackend backend, final String name) {
return backend.getExecutableFunctions().stream()
.filter(function -> name.equals(function.callableName()))
.findFirst()
.orElseThrow(() -> new AssertionError("missing executable " + name));
}
private static Span span(final int fileId) {
return new Span(new FileId(fileId), 4, 16);
}
private static String firstCode(final BuildingIssueSink issues) {
return issues.asCollection().iterator().next().getCode();
}
private static String firstMessage(final BuildingIssueSink issues) {
return issues.asCollection().iterator().next().getMessage();
}
}
@@ -36,6 +36,22 @@ class BackendConformanceMatrixSpecTest {
"G19-5.2.6",
"G19-5.2.7",
"G19-5.2.8",
"G19-11.1",
"G19-11.2",
"G19-11.3",
"G19-11.4",
"G19-11.5",
"G19-11.6",
"PBS12-11",
"G20-4.1.1",
"G20-4.1.2",
"G20-4.1.3",
"G20-4.2.1",
"G20-4.2.2",
"G20-4.2.3",
"G20-4.3.1",
"G20-4.3.2",
"G20-4.3.3",
"G20-6.2",
"G20-6.3",
"G20-6.4",
@@ -75,7 +91,11 @@ class BackendConformanceMatrixSpecTest {
"PBS13-12.3.2",
"PBS13-12.4.1",
"PBS13-12.4.2",
"PBS13-12.5");
"PBS13-12.5",
"PBS13-12.7",
"PBS13-12.8",
"PBS13-12.9",
"G23-5.2");
@Test
void matrixMustExistAndContainRequiredRequirementRows() throws IOException {

Some files were not shown because too many files have changed in this diff Show More