blog: the control that could not fail - #868
Merged
Merged
Conversation
Four gates had never been seen red. Writing their negative controls turned up one that certified the checking function but not the wiring from that function to the exit code -- so with main() neutered the gate printed OK on a broken catalog while its control reported every branch red. Also: my first attempt to measure this rewrote the returns inside the controls too, which made two sound controls look like vacuous passes. And the distinction the report turns on -- nine surviving mutants means nine controls that do not reach far enough, not nine broken gates. All six baseline-backed gates were checked by hand and are correct today. Refs gHashTag/t27#2465, #2467, #2468
PostMeta requires openQuestions and published; the new entry had neither,
and typecheck:ratchet caught it as src/data/blog/index.ts: 0 -> 1.
My own tsc check had missed it because I ran it from the repository root
rather than apps/website, where that tsconfig lives -- the same
wrong-working-directory error the post this entry describes is about,
made while adding the post.
Two unrelated things in the same commit, both stated rather than quiet:
- .claude/launch.json was created by the preview tooling and committed
by accident. It is not on main. Removed.
- typecheck-baseline.json 184/27 -> 179/26. The improvement is NOT mine
(SacredChemistryWidget.tsx, 5 -> 0, already improved on main, where
this gate is red for the same reason). Locking it in is the downward
direction and stops the gain sliding back unrecorded.
github-actions Bot
added a commit
that referenced
this pull request
Aug 22, 2026
blog: the control that could not fail (#868) * blog: the control that could not fail Four gates had never been seen red. Writing their negative controls turned up one that certified the checking function but not the wiring from that function to the exit code -- so with main() neutered the gate printed OK on a broken catalog while its control reported every branch red. Also: my first attempt to measure this rewrote the returns inside the controls too, which made two sound controls look like vacuous passes. And the distinction the report turns on -- nine surviving mutants means nine controls that do not reach far enough, not nine broken gates. All six baseline-backed gates were checked by hand and are correct today. Refs gHashTag/t27#2465, #2467, #2468 * fix(blog): the index entry was missing two required fields PostMeta requires openQuestions and published; the new entry had neither, and typecheck:ratchet caught it as src/data/blog/index.ts: 0 -> 1. My own tsc check had missed it because I ran it from the repository root rather than apps/website, where that tsconfig lives -- the same wrong-working-directory error the post this entry describes is about, made while adding the post. Two unrelated things in the same commit, both stated rather than quiet: - .claude/launch.json was created by the preview tooling and committed by accident. It is not on main. Removed. - typecheck-baseline.json 184/27 -> 179/26. The improvement is NOT mine (SacredChemistryWidget.tsx, 5 -> 0, already improved on main, where this gate is red for the same reason). Locking it in is the downward direction and stops the gain sliding back unrecorded.
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.
Eighth post of the gate-audit campaign, and the first where the defect was in the instrument I had just built.
Four gates had no negative control. Writing them produced thirty-one killed mutants — and one control that could not fail: it called the checking function directly, so with
main()'sreturn 1changed toreturn 0the gate printedOKon a catalog with a dangling reference and all seven cases still reported "RED, correct branch".The post argues the in-process design was not the mistake — it exists because a control run from the wrong directory once resolved its root to
/and nearly reversed a conclusion — and that the fix is a layer added, not replaced.It also carries the distinction the follow-up report turns on: nine gates with surviving mutants is nine controls that do not reach far enough, not nine broken gates. All six baseline-backed gates were checked by hand and go red correctly today.
en + ru.
build:ciclean,tscclean on the blog data.Refs gHashTag/t27#2465, gHashTag/t27#2467, gHashTag/t27#2468