refactor(vm)!: remove publication and logical guard opcodes
Intrepid/Prometeu/Runtime/pipeline/head This commit looks good

Remove INDEX_GUARD, ARRAY_CHECK_LENGTH, STORAGE_PUBLISH and FIELDS_PUBLISH
from decoding, metadata, tooling, verification and execution. Keep ARRAY_LEN
and overlap-safe ARRAY_COPY with complete-group ownership and early charging
of calculable copy work, including refused attempts.

Update specifications, producer handoff, qualification evidence and lessons.
Share collection test support and remove redundant Copy clones for Clippy.

Validation: make ci-domains passed with 1,095 workspace tests, five separately
executed debugger socket tests and all 15 unchanged coverage domain gates.
VM release, resource diagnostics and layer/quality suites also passed.

BREAKING CHANGE: PBX using opcodes 0xA7, 0xA8, 0xAA or 0xAB is rejected,
including unreachable instructions. Consumers must adapt their emission and
collection algorithms or operational guarantees.
This commit is contained in:
2026-10-23 23:37:56 +01:00
parent 69ad1931de
commit dcf6561ea7
27 changed files with 1407 additions and 1013 deletions
+10 -15
View File
@@ -192,22 +192,17 @@ STRUCT_NEW (0x00A0), STRUCT_GET (0x00A1), STRUCT_SET (0x00A2), ARRAY_NEW (0x00A3
See the [random contract](../specs/runtime/02e-vm-owned-random-core.md) for Random BuiltinId 0x0016, ObjectKind tag 9, intrinsics 0x4000–0x4003, atomic state mutation, bounded 4+8*N costs and rejection-limit trap 0x00000015. Random is managed by the VM heap and existing GC, without host entropy or PBX changes.
## Logical guards and atomic storage publication
## Typed array range copy
INDEX_GUARD (0xA7), ARRAY_CHECK_LENGTH (0xA8), ARRAY_COPY (0xA9) and
STORAGE_PUBLISH (0xAA) provide generic logical validation, complete stored-group
transfer and prepared atomic publication. Their encoding, typed operands, exact
faults, overlap semantics, costs and preflight/commit guarantees are defined by
the [canonical instruction contract](../specs/runtime/02-vm-instruction-set.md#logical-guards-and-atomic-storage-publication). No new Registry category, syscall, collection algorithm or safepoint is introduced.
## Atomic publication of multiple Struct instances
FIELDS_PUBLISH (0xAB) atomically publishes complete typed fields to multiple
source-selected Struct destinations. Its schema is an ordinary Registry Struct;
no new category, request object, syscall or safepoint is introduced. Exact nullable
destination rules, stack effects, validation, costs and commit semantics are in the
[canonical instruction contract](../specs/runtime/02-vm-instruction-set.md#atomic-publication-of-multiple-struct-instances).
ARRAY_COPY (0xA9, Array TypeId u32) consumes source, sourceStart, target,
targetStart, count without a result. Complete groups and owners survive overlap
and failed preparation. Charge 4+2*N*W before ordinary validation for available
Int64 count in 0..u32::MAX; other counts/underflow charge base 4. Checked cost
overflow charges base and traps SizeOverflow. The
[canonical contract](../specs/runtime/02-vm-instruction-set.md#typed-array-range-copy)
defines fault precedence, scratch and commit. ARRAY_LEN (0xA4) and physical array
checks remain. Encodings 0xA7/0xA8/0xAA/0xAB reject at load even in dead code;
no replacement publication/guard or transaction is provided.
## Identity/hash and catalog schema 2