fix(ci): ledger the fifteen ports that do not generate - #5550
Merged
Merged
Conversation
check_specs_generate.py failed on master: 15 specs under specs/port/ do not generate with any backend, each merged while this step was already red. Making them parse is not a repair -- on #5473 seven were made to parse and their generated C and Zig still had dozens of errors per file -- so their ports are to be redone, tracked in #5549 with the error and the PR that added each one. Until then they join tools/specs_generate_baseline.txt by hand, with the compiler's own first line, which is how this ledger is meant to grow (--update-baseline refuses to grow it). 14 -> 29 entries. The gate passes on 769f325 (1212 specs, 1183 generate) and its self-check passes. Refs #5497 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
No spec newly fails to generate (known-broken set is ledgered)(Emit Bit-Exact Gate) fails on master: 15 specs underspecs/port/are refused by every backend. Each was merged while this same step was already red, so nothing stopped it.Why ledger and not repair
Making them parse is not a repair. On #5473 seven of them were made to parse, and the generated C and Zig still had dozens of errors per file. Each needs its port redone against the original source. #5549 lists the 15 with the compiler error and the PR that added each.
Change
The 15 lines are added to
tools/specs_generate_baseline.txtby hand, sorted, each with the compiler's first line asgenerates()reports it, under a comment that names #5549. That is how the tool means this ledger to grow:--update-baselinerefuses to grow it, and its self-check proves that refusal. 14 -> 29 entries.Evidence (769f325)
check_specs_generate.py:OK: 1212 specs, 1183 generate, 29 known-broken.--self-check: every verdict reaches the exit code.What CI will show next
With this step green, the job reaches the steps after it. Locally on 769f325,
verify_exhaustive,check_duplicate_agreement,verify_emit_bitexact(restored in #5498),verify_trainer_c,gft_backprop_microcode,fuzz_trainerandverify_igla_raceall pass.verify_multitargetfails for a real reason, #5506 (saddtraps in Zig). The job stays red until #5506 is decided, and from now on it is red for that reason.Refs #5497
🤖 Generated with Claude Code