test(extraction): harden agc_assembly extraction coverage (#857) - #931
Merged
Conversation
Contributor
- Migrated agc_assembly out of monolithic dicts. - Created ~136 tests for func_start, args, and _dependency_capture. - Verified valid, invalid, and pathological cases. - AGC regexes proved robust; no changes to language_standards.py required.
Closed
21 tasks
Reviewed the prior test-authoring pass (136 tests, claimed zero bugs found, "patterns already extremely robust") given every other language in this epic found real bugs. Found two significant, corpus-validated bugs the original review missed: - `func_start`'s opcode whitelist covered only 16 real instructions, missing ~25 opcodes this SAME file's own sibling rules (`branch`, `args`, `safety`, `state_mutation`) already recognize as legitimate -- CAF, TCF, XCH, LXCH, QXCH, AD, ADS, SU, MULT, DV, MASK, INCR, AUG, DIM, DAS, RELINT, EDRUPT, BZE, BMN, RESUME, RETURN, TCR, GOTO, RVQ. `CAF` (Clear and Add Fixed) alone has 94 occurrences in the real Apollo 11 corpus -- on par with CA/CS/TS as one of the most common AGC instructions -- and was entirely invisible as a subroutine anchor. - `args`'s register-coupling opcode list was missing AUG/DIM/INCR, real instructions this file's own `state_mutation` rule already recognizes, so `AUG A`/`DIM Q`/`INCR A` (a register explicitly coupled to a real opcode, exactly what this rule exists to capture) were invisible. Confirmed both fixes' real-world impact directly against the Apollo 11 (Luminary/Comanche) source corpus in language-crucible: func_start's total matches rose from 609 to 812 (+33%), with zero new false positives against data/constant pseudo-ops (OCT/OCTAL/DEC/2DEC/ADRES/EQUALS all correctly stay excluded -- those mark data declarations, not subroutine entries). Both fixes verified linear-time (no ReDoS). The original test suite's own `invalid` list literally asserted "MYLABEL\tTCF INTERNAL" must NOT match, with the comment "TCF is not in the opcode list" -- true of the code as written, false of real AGC assembly. Fixed the test file: moved TCF (and CAF/XCH/LXCH/AD/MASK/INCR/ RELINT) to valid cases, added dedicated regression tests for both bugs documenting the corpus-validated before/after counts, added AUG/DIM/INCR valid cases for args, and cleaned up two stray blank-line artifacts left in the old monolithic dict files by the migration. crucible_check.py showed a real diff, confirmed the only CONTENT changes (not global-aggregate/topological-coordinate ripple, an expected side effect of agc_assembly's structural mass shifting) were in agc_assembly itself -- real Apollo 11 subroutines (SPVAC, NEXTCORE, JOBWAKE3, SUPDXCHZ, VACFOUND) previously silently merged into a preceding function's body now correctly resolve as their own boundaries. Golden master fixtures re-blessed. Also rebased this branch onto latest main: it had accidentally bundled an already-separately-merged, unrelated kotlin fix (#899/#929) from before that work was cleanly split into its own PR. The rebase dropped the now-redundant duplicate commits automatically (git detected their patch content was already upstream). Epic #813 updated with 3 new recurring bug classes (58-60), including a meta-lesson about AI-authored test coverage validating against existing implementation behavior rather than real-world ground truth. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
squid-protocol
force-pushed
the
extraction-hardening-agc-assembly
branch
from
August 1, 2026 02:38
eb143ef to
a6bfb60
Compare
4 tasks
2 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR resolves #857 by hardening the extraction test coverage for
agc_assembly.Changes:
tests/extraction/languages/test_agc_assembly.pyfollowing the methodology inhow_to_harden_extraction.md.func_start,args,_dependency_capture), categorized across valid, invalid, and pathological cases.agc_assemblyfrom the old monolithic dicts (test_function_extraction_strict.pyandtest_dependency_extraction_strict.py).language_standards.pywere already extremely robust, and passed all tests flawlessly (no false positive captures or missing captures). Thus, no changes tolanguage_standards.pywere required.Resolves #857