Persist cartridge identity and slot configuration, add MemcardBoot and the MemcardBuffer SDK, and consume the runtime's negotiated memcard host controls with protected teardown. Qualify save/reopen and version upgrades through distinct configured runtime processes. Include specs, examples, regression evidence and the consolidated discussion lesson.
Memcard integration evidence
Owner: Studio, with compiler/pbs and compiler/general propagation. Final result: all accepted D1..D7 criteria executed. Summary identifies exact provider source, binaries, schemas, producer artifacts, platform and commands.
final-regression-and-dist.log: full regression, installed knowledge, real DEV JavaFX tests, memcard process proof and existing headless process compatibility.process-dev.log,process-release.log,process-dist.log: actual PBS/Studio shipper/Gateway sessions using the configured executable, both optimizer variants and separate writer/reader processes across versions 1 and 2. Logs identify each PBX/source/manifest and unchanged image hash. Dev/release folders contain isolated development binary layouts; only dist is an extracted archive.native-dev.log,native-release.log: 46 observations each against production Firmware/SystemOS/FileMemcard, including the installed-JAR-produced fixture. These observations are deliberately separate from OS-process persistence proof.regression-and-ui.log,headless-focused.log: consumer protocol/MCP/authority and UI confirmation checks during implementation.history-2026-10-29/: historical initial carrier failure and successful buffer qualification before the provider's public process extension. Old pending-gate text and clean-provider claims in this directory refer only to that earlier date.
Final ordinary regression: 1,118 tests across 201 suites, zero failures/errors, 13 skipped. Installed knowledge: 2 passed. JavaFX DEV: 5 passed. Existing configured headless: 4 passed. Memcard process: 2 passed in each of dev, release and dist. Native production-host: 46 observations in each runtime profile.
Runtime source/configuration files and commits were not modified by this work. The selected provider was an existing delivery whose exact source hashes match runtime DSC-0068/LSN-0073 evidence. It was uncommitted during qualification and was independently committed during finalization as a4bec5bc4577850543f57c79f5124ca00169e1b3; the recorded source hashes still match and its final working tree is clean. Binary and archive hashes identify the extracted installation separately from its source baseline commit.
The discussion CLI manages completion/index/artifact removal. It has no evidence
relocation operation, so the old workflow evidence directory was moved here before
CLI housekeeping. discussion validate passed before and after housekeeping;
git diff --check passed. Local lesson/spec/memcard-handoff references resolve.
No other-platform, physical power-cut or complete reference-game deployment claim is made. DSC-0086 retains the latter qualification.