Skip to content

Delete the two restated sizes of the set of runs that report and refuse nothing - #399

Merged
iderex merged 1 commit into
mainfrom
the-pair-this-sentence-points-at-holds-four-398
Sep 17, 2026
Merged

iderex merged 1 commit into
mainfrom
the-pair-this-sentence-points-at-holds-four-398

Conversation

@iderex

@iderex iderex commented Sep 17, 2026

Copy link
Copy Markdown
Contributor

The issue this belongs to

Closes #398

What changed

Two sentences that gave the set of runs which report and refuse nothing a size
no longer give it one. CONTRIBUTING.md walks that set in a single passage and
declares the count there as a hand count over the entries beside it; that
passage is untouched. What is deleted is the second and third copy of its size:
the sentence about branch-health three hundred lines above it, and the header
comment of .github/workflows/cross-toolchain.yml.

Each site keeps the reason the number went rather than a corrected number, so a
reader who wants the size goes to the passage that holds the entries.

What failure it prevents

A reader deciding whether a green tick is a merge condition follows a pointer
that promises two entries and meets four. Both copies had already drifted, in
two different directions.

The branch-health sentence went stale. It was written when the passage below
held two:

git log origin/main -S'not a third member of the pair below' --format='%h %ad %s' --date=short -- CONTRIBUTING.md
724407c 2026-09-02 Report a workflow standing red on the default branch (#90)
git log origin/main -S'FOUR RUNS REPORT AND REFUSE NOTHING' --format='%h %ad %s' --date=short -- CONTRIBUTING.md
4396ab1 2026-09-16 Describe the two legs the gate section left out [#326]

The workflow comment was short on its first run rather than going stale. It
named two other runs and called its own leg the third, and mutation was
already in this gate nine days earlier and gates nothing:

for f in mutation.yml cross-toolchain.yml; do
  git log origin/main --diff-filter=A --format="%h %ad $f" --date=short -- .github/workflows/$f
done
00559cd 2026-08-27 mutation.yml
7587fa7 2026-09-05 cross-toolchain.yml
grep -n 'This leg gates nothing' .github/mutation/mutation.sh
592:  say "This leg gates nothing. A falling score obliges an issue and never a red gate, which is what taking the question off the merge path costs and what paying it looks like."

It was found by reading ## What the gate runs, and what each thing refuses
from the top against the section that corrected the count on 2026-09-16, which
moved the paragraph holding the entries and neither sentence pointing at it.

Evidence

The commit being pushed:

git rev-parse HEAD
f972439c69fd1d80c163ac9baec75cc7da48f97e
git diff --stat origin/main...HEAD
 .github/workflows/cross-toolchain.yml | 10 ++++++++--
 CONTRIBUTING.md                       | 16 +++++++++++++---
 2 files changed, 21 insertions(+), 5 deletions(-)

Neither file names a size for that set any more, and the passage that does still
does:

git grep -n -c 'the pair below' HEAD -- CONTRIBUTING.md ; echo "exit=$?"
exit=1
git grep -n 'are the two runs' HEAD -- .github/workflows/cross-toolchain.yml ; echo "exit=$?"
exit=1
git grep -n 'FOUR RUNS REPORT AND REFUSE NOTHING' HEAD -- CONTRIBUTING.md
HEAD:CONTRIBUTING.md:363:FOUR RUNS REPORT AND REFUSE NOTHING, which is deliberate rather than an

The path gate at this commit:

bash .github/doc-paths/doc-paths.sh check | tail -3
NOT MADE HERE: a tracked file that is not .md. Nothing else in this tree is prose today.

Every path these documents name resolves against the tracked set.
bash .github/doc-paths/doc-paths.sh check > /dev/null; echo "exit=$?"
exit=0

The workflow file still parses after the comment change:

python -c "import yaml; yaml.safe_load(open('.github/workflows/cross-toolchain.yml', encoding='utf-8')); print('yaml ok')"
yaml ok

What this does not cover

  • Whether branch-health belongs in that set at all. Its own paragraph says it
    never reddens for a red workflow and that its run does not go red for what it
    found, which is the property the set is named for, and the passage that walks
    the set does not include it. That would move the number to five and nothing
    here touches it. The edited sentence says so in its own text rather than
    asserting the membership either way.
  • The number four. I did not re-derive the membership of the set; I read the
    paragraph's declaration and the entries it walks. What this change is about is
    that two other sentences carried a size at all.
  • Any mechanism. Nothing in this repository reads a count written into prose, so
    a third stale copy tomorrow is refused by nothing and would be found by a
    reader again. No check is added here, and none is asked for.
  • The runs themselves. No workflow trigger, permission, job or step is changed;
    the only edit to .github/workflows/cross-toolchain.yml is inside its header
    comment.

Who has read it

Nobody other than me has read this change. The evidence above stands in place of
a second reading rather than beside one.

…se nothing

`CONTRIBUTING.md` walks that set in one passage and says of its own count there
that it is a hand count over the entries beside it. Two further sentences carried
a size for the same set and were not moved when it changed.

The sentence about `branch-health`, three hundred lines above, called the passage
below a pair and this leg a third member. It was written on 2026-09-02, when the
passage held two, and the passage has held four since 2026-09-16. The header of
`.github/workflows/cross-toolchain.yml` named two others and called its own leg
the third; `mutation` was in this gate nine days before that file arrived and
gates nothing, so that one was short on its first run rather than going stale
later.

Both numbers are deleted rather than corrected, with the reason written where the
sentence was. Nothing in this tree reads a count written into prose, so a
corrected copy is answered by nobody and the next leg to arrive makes a third
stale one. What the first sentence separated `branch-health` on is disclosed as
shared rather than kept: `cross-toolchain` refuses a survey that cannot be read,
which is the same shape as refusing a run that produced no report, and whether
`branch-health` belongs in that passage is left open rather than answered here.

Closes #398

Signed-off-by: Nils Lehnen <30603423+iderex@users.noreply.github.com>
@iderex
iderex merged commit c74cf54 into main Sep 17, 2026
29 checks passed
@iderex
iderex deleted the the-pair-this-sentence-points-at-holds-four-398 branch September 17, 2026 14:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Two sentences give the set of runs that report and refuse nothing a size, and the passage they point at has held four since 2026-09-16

1 participant