Record that a green local rig run can mean half the rig never ran - #403
Open
nyblnet wants to merge 1 commit into
Open
Record that a green local rig run can mean half the rig never ran#403nyblnet wants to merge 1 commit into
nyblnet wants to merge 1 commit into
Conversation
Build size
Updated: |
test-sanitize.ts has a node half and a browser half. The browser half needs Chrome, says so when there is none, skips, and the run still ends green. sanitizeSvg needs a DOM, so it is covered ONLY by that half -- which is how a change to it passed five rigs locally and was caught by CI a minute later. The rule worth keeping is not "run the rigs" but "run the rig that covers the file you edited, and check WHICH HALF ran". A skip that prints like a pass belongs with the instruments this log already collects: the CSS transform on an inline element that reads back correctly and moves nothing, document.fonts.check() returning true for a missing font, rAF throttled to zero in a hidden tab. Filed first as "sanitizeSvg has no rig coverage", which was false. The correction is the useful artifact: "there is no coverage" would have sent someone to write a rig that already exists and works.
nyblnet
force-pushed
the
docs-half-a-rig
branch
from
August 30, 2026 19:13
058a7c4 to
bc3c361
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Docs only — one entry appended to
docs/DECISIONS.md.The finding.
scripts/test-sanitize.tshas a node half (pure functions,runs anywhere) and a browser half (drives real Chrome, because what it tests
needs a DOM). With no Chrome, the browser half says so, skips, and the run
still ends green.
sanitizeSvgneeds a DOM, so it is covered only by the browser half. That ishow a change to it passed five rigs locally and was caught by CI a minute later
with
✗ the svg <style> is kept(#402).The rule: not "run the rigs" but run the rig that covers the file you
edited, and check which half of it ran. Not how many passed — which ones.
It belongs with the instrument-lies family this log already collects: a CSS
transform on an inline element that reads back verbatim and moves zero pixels,
document.fonts.check()returning true for an uninstalled font, rAF throttledto zero in a hidden tab, an SVG rasteriser ignoring
font-family.The correction is the point. I first filed this as "
sanitizeSvghas no rigcoverage" — false. The coverage exists and works. "Your green local run may have
skipped it" is true and actionable; "there is no coverage" would have sent
someone to write a rig that already exists.
Raised by
bento-team-lead, who suggested promoting it rather than leaving itin the zone notes.
Note on scope:
docs/is scribe's zone, but the session ritual asks everyzone to promote durable findings to
DECISIONS.md, and this is append-only —no existing entry touched. Deliberately NOT touching root
CHANGELOG.md, whichis scribe's and has #400 open on it.
Checklist
DECISIONS.mdentries are a rebase, not a merge decision