Skip to content

gated-caller, hermetic-check: name the population boundary — #544's instrument was outside the sentence, not inside it as a finding - #546

Merged
jobordu merged 1 commit into
mainfrom
dev1/subdir-boundary
Aug 21, 2026
Merged

gated-caller, hermetic-check: name the population boundary — #544's instrument was outside the sentence, not inside it as a finding#546
jobordu merged 1 commit into
mainfrom
dev1/subdir-boundary

Conversation

@jobordu

@jobordu jobordu commented Aug 21, 2026

Copy link
Copy Markdown
Contributor

Both tools glob non-recursively. tools/README.md keeps subdirectories out of scope deliberately, and that is not the defect — the defect is that N of M says nothing about it, so a reader takes M for the instrument count.

gated-caller    2 instruments exposing --self-test in architect-sweeps/
hermetic-check  3 suites in teamlead/

#544 is exactly this case

tools/architect-sweeps/prior-art.py is the instrument ARCHITECT reports as worked 6 of 6 when run, skipped once, and the skip shipped a duplicate.No caller audit in this repository could see it to say so — including the one that measured "1 of 36 sweeps has a gated caller" three hours earlier.

Which is worse than a finding. An instrument reported as UNCALLED is a row someone can act on. An instrument outside the population is invisible, and the count reads as complete.

NAMED, NOT INDEXED

Only directories and counts are printed — never file names — and nothing is executed. tools/teamlead/ is quarantined: not indexed, not silenced, nobody investigates it. ⇒ A count with a directory satisfies not silenced without indexing a single file, and the quarantine contributes 0 instruments to gated-caller's population in any case.

Controls

Asserted as a pair in both tools, because neither half fails alone: the line appears with a subdirectory passes if it were unconditional; absent when there is none passes if it never printed. Mutation-checked — unconditional line fails one assertion, a counter that never fires fails two. A subdirectory file not exposing --self-test must not inflate the count: the boundary reports the population, not the directory.

⚠ What this does not do

  • It does not widen the population. Subdirectories stay out of scope; they are now stated to be out of scope, with a number.
  • It does not establish whether gate-selftests.sh reaches subdirectories — I read its stated predicate, not a run.

Measured at ca9c2a1. — DEV1

instrument was not reported as uncalled, it was not reported

Both tools glob `tools/*.py` and `tools/test_*.py` non-recursively.
tools/README.md keeps subdirectories out of scope deliberately, and that
is not the defect. The defect is that "N of M" says nothing about it, so
a reader takes M for the instrument count and an instrument living in a
subdirectory is outside the sentence rather than inside it as a finding.

  gated-caller   2 instruments exposing --self-test in architect-sweeps/
  hermetic-check 3 suites in teamlead/

#544 is exactly this case. tools/architect-sweeps/prior-art.py is the
instrument whose skipped run shipped a duplicate, and no caller audit in
this repository could see it to say so — including the one that measured
"1 of 36 sweeps has a gated caller" three hours earlier.

NAMED, NOT INDEXED. Only directories and counts are printed, never file
names, and nothing is executed. tools/teamlead/ is quarantined — not
indexed, not silenced, nobody investigates it — and a count with a
directory satisfies "not silenced" without indexing a single file.

Both controls assert BOTH directions, because neither half fails alone:
"the line appears with a subdirectory" passes if it were unconditional,
"absent when there is none" passes if it never printed. Mutation-checked:
making the line unconditional fails one assertion, making the counter
never fire fails two.

A subdirectory file NOT exposing --self-test must not inflate the count —
the boundary reports the population, not the directory.

Measured at ca9c2a1.

DEV1
@coderabbitai

coderabbitai Bot commented Aug 21, 2026

Copy link
Copy Markdown

Warning

Review limit reached

@jobordu, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 51 minutes

Limit details: You’ve used the included review currently available.

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

How can I continue?

Wait for the limit to reset, then comment @coderabbitai review or push new commits to the PR.

An organization admin can change what happens after included review limits in Billing.

How do review limits work?

CodeRabbit enforces per-developer PR review limits within each organization.

For paid Pro and Pro+ reviews, CodeRabbit uses a developer's included PR review attempts over the past 7 days to set the current hourly allowance. At typical activity levels, the full plan allowance applies. Higher sustained activity can lower the allowance until earlier attempts leave the 7-day window.

Please refer docs for additional details.

Review details
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: ffbfa24e-8356-4805-9e67-8672fa8d30a3

📥 Commits

Reviewing files that changed from the base of the PR and between c4f2f96 and cd278e9.

📒 Files selected for processing (2)
  • tools/gated-caller.py
  • tools/hermetic-check.py

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

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.

1 participant