Repository navigation
fix(gates): a deleted tracked spec is broken, not skipped; extract _prereq (#2492) - #2523
Merged
Merged
Conversation
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Contributor
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
#2520 gave a mechanical test: every
skipclaims that the thing missing is not the subject of this check. Fourteen calls in the tree; read each as that sentence and two do not survive.Two skips that were failures
Both are a spec file tracked in git.
verify_emit_bitexactskipped whengft_smul.t27/gft_sadd.t27were absent;verify_igla_racewhenternary_mac.t27was.Measured, not argued — renaming the file aside:
verify_emit_bitexact.pyverify_igla_race.pyA deleted spec is not a bare machine. It is the repository missing the thing the gate exists to verify, and the sentence "the missing spec is not the subject of this check" is simply false. Both now call
broken(): fatal with or without--require.--requireshould not be what saves you from a deleted source file.The extraction, at the one safe moment
Four hand-copied
skip()s had just been made to behave identically (#2518, #2519, #2520) — which is the only time a deduplication is a pure deduplication.tools/_prereq.pynow holds both words, with the rule that separates them written where the code is rather than in a comment inside one of four copies, which is exactly how the four came to disagree.The extraction broke a control on its first run, and the control said so
verify_multitarget's planted tree carried the script and not the module it now imports, so the child died at import with empty stdout — and the case refused to read that as a skip. A plant must carry everything the thing under test needs; adding a shared dependency changes what "everything" means, and the only cheap way to learn that is a control that fails loudly.Two assertions also moved: messages now name the script from
argvrather than a constant, soSKIP verify_multitargetbecameSKIP verify_multitarget.py. A control asserting exact text is a control that notices a message changing — which is the point, even when the change is mine.Verified
All five self-checks green.
verify_igla_race,verify_multitarget,verify_trainer_cstill exit 0 on a healthy tree. This PR touches four files in the twenty-minuteemit-bitexactjob, so it waits for that job rather than for the checks beside it.Refs #2492