Skip to content

feat(tri): gates mutate --invert, the third axis (#2492) - #2499

Merged
gHashTag merged 3 commits into
masterfrom
feat/invert-operator
Aug 23, 2026
Merged

gHashTag merged 3 commits into
masterfrom
feat/invert-operator

Conversation

@gHashTag

Copy link
Copy Markdown
Collaborator

Two operators ask whether a gate can still reach its verdicts. This asks whether it reaches the right one.

A gate that fires its FAIL branch on a healthy tree and its OK branch on a broken one satisfies both return operators, and neither is looking at that.

Result

if C: → if not (C): on conditions whose body carries a verdict. Thirteen gates, one survivor: check_elab_ratchet's no-baseline branch — the same one both other operators leave, declared UNCOVERED with its reason.

Three axes now converge on one line. This is the first new instrument in this campaign to find nothing, and that is worth as much as the ones that found plenty: controls written over a week against two questions held against a third they were not written for.

The scoping is the design

Inverting a loop guard or a plumbing check makes the gate crash, and a control that reds on a traceback would score that as a kill — for the wrong reason, which is the mistake this command exists to catch.

Only conditions whose body holds a return 1..4, a raise SystemExit(N) or a FAIL print are sites. Measured over 19 mutations on four gates:

killed by message   19
killed by crash      0
survived             0

My verification probe was broken twice before it worked

It mutated inside the control functions — which the Rust implementation excludes and the throwaway Python probe did not — and its verdict label conflated not killed with killed by a crash, printing killed by crash for a mutant that had survived.

Both wrong in the same run, and the first output read as a finding about the tool. The probe written to verify a careful tool inherits none of its care, and is trusted anyway because it is short.

Six unit tests on the site selector, including the three exclusions: a body that only prints, a body that returns success, and anything inside a control function.

Refs #2492

Two operators ask whether a gate can still reach its verdicts. This asks
whether it reaches the RIGHT one: a gate that fires FAIL on a healthy
tree and OK on a broken one satisfies both of the others.

`if C:` becomes `if not (C):` on conditions whose body carries a verdict.
Thirteen gates, ONE survivor -- check_elab_ratchet's no-baseline branch,
the same one both other operators leave, declared UNCOVERED with its
reason. Three axes now converge on one line.

The scoping is the design. Inverting a loop guard or a plumbing check
makes the gate CRASH, and a control that reds on a traceback would score
that as a kill for the wrong reason. Only conditions whose body holds a
`return 1..4`, a `raise SystemExit(N)` or a FAIL print are sites.
Measured over 19 mutations on four gates: 19 killed by MESSAGE, zero
tracebacks.

My verification probe was broken twice before it worked -- it mutated
inside the control functions, which the Rust implementation excludes and
the probe did not, and its label conflated "not killed" with "killed by a
crash". The first output looked like a finding about the tool.

Refs #2492
…nd nothing (#2492)

--invert asks whether a gate reaches the RIGHT verdict, not whether it
can still reach one. Thirteen gates, one survivor: the branch both other
operators leave, declared uncovered.

That is the first new instrument here to find nothing, and it is worth as
much as the ones that found plenty -- controls written against two
questions held against a third.

Plus: scope an operator by what its kills would MEAN, and then measure
that the scoping worked. 19 mutations, 19 killed by message, zero
tracebacks.

And: the probe that checks a tool gets less care than the tool. Mine was
broken twice in one run and its first output read as a finding about the
tool. The throwaway scripts account for most of the seven broken
instruments in this campaign.

Refs #2492
@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-08-23 04:27:06 UTC

Summary

Status Count
Total Open PRs 6
PRs with Failing Checks 2
PRs with All Checks Green 4
READY 3
FAILING 2
PENDING 0

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=e905ba8eeb53 != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@github-actions

Copy link
Copy Markdown
Contributor

📓 NotebookLM Notebook linked to this PR

This notebook contains session context, decisions, and artifacts for this work.

The note was dated 2026-08-25 and check-now-freshness rejected it. The
real UTC date is 2026-08-23. I had been inferring the passage of days
from the number of iterations rather than reading a clock, and two
earlier notes carry the same invented drift.

Mechanism 8.2 inside out: the world did not move and I recorded that it
had. The gate caught it, which is the gate doing precisely its job, and
the first time in this campaign that an existing gate caught ME.

Refs #2492
@github-actions

Copy link
Copy Markdown
Contributor

📓 NotebookLM Notebook linked to this PR

This notebook contains session context, decisions, and artifacts for this work.

@github-actions

Copy link
Copy Markdown
Contributor

PR Dashboard

Generated at: 2026-08-23 04:37:29 UTC

Summary

Status Count
Total Open PRs 6
PRs with Failing Checks 2
PRs with All Checks Green 4
READY 3
FAILING 2
PENDING 0

Seal Status

  • ⚠️ STALE -- sha256(compiler.rs)=e905ba8eeb53 != manifest seal=87e5cbd3ad94.
    The committed NMSE numbers were certified against an older compiler.rs.
    Run scripts/reseal-check.sh locally for the two-step reseal command (advisory; not a merge gate).

@gHashTag
gHashTag merged commit b731f35 into master Aug 23, 2026
22 of 24 checks passed
@gHashTag
gHashTag deleted the feat/invert-operator branch August 23, 2026 04:38
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