Track A from the job NAME test-arm32, not from its source. This job names a MECHANISM rather than a subject — the source it was fed (tools/compiler_srchash.sh) is what the mechanism was run ON, not what is being tested, so a lane guessed from it would be wrong by construction. The ranker reads frontmatter, so this line decides who works it; re-lane it if this job has changed what it covers.

origin/master has advanced 3 commit(s) since this sha. Re-verify at current HEAD before acting — the callback is tagged to the sha that was tested, which may no longer be the state of the tree.

regression: test-arm32#src:tools/compiler_srchash.sh at b29428afe251 in step 1/1, livesrc=$(tools/compiler_srchash.sh); \ stampsrc=$(sed -n 's/^srchash //p' compiler/.pascal26.fixedpoint); \ if [ -z "$… (auto-filed by twatch)

• Type: regression (auto-filed by Track T watcher, host seven, twatch 5666dc9dba23). Untriaged.
• Found: 2026-09-08T17:22:50Z
• Test source: tools/compiler_srchash.sh compiler/.pascal26.fixedpoint
• Failing step: line 1 of 1 of the job's recipe; it names tools/compiler_srchash.sh compiler/.pascal26.fixedpoint.

  livesrc=$(tools/compiler_srchash.sh); \ stampsrc=$(sed -n 's/^srchash //p' compiler/.pascal26.fixedpoint); \ if [ -z "$livesrc" ]; then \ echo "tools/compiler_srchash.sh produced NO source hash, so this check cannot run."; \ echo " An empty result means 'could not measure', never 'measured and they
  
  

Repro

tools/testmgr.py --tier full --job 'test-arm32#src:tools/compiler_srchash.sh' at b29428afe251160ab130ead03f9132d3345e778a

Range

The named sha b29428afe251 CANNOT be the cause — it touches no buildable file (docs / tickets / tstate only). It is the sha that was TESTED, i.e. the upper bound of an untested range; the cause is somewhere below it.

bad b29428afe251, last good 95fc8aff2016, 2 commit(s) in range — the watcher narrows this by idle bisect; check tstate/TSTATE.md for the current range.

Log tail

compiler/.pascal26.fixedpoint was written for DIFFERENT SOURCES than the tree has.
  stamp sources: deadbeef
  tree sources:  2c26ecae2d8e5fa3a75cd41bdf9df7033d50b5f3594ff92edadba9fb4de89eec
  Same 215 files on both sides, so a hashed file's CONTENTS changed.
  Nothing untracked or modified locally, so no stray file explains it:
  this stamp was written for a DIFFERENT TREE. Usual cause is a pull
  -- or a sync, which pulls -- with no rebuild after it.
A stamp NEWER than sources it does not describe is how this step
printed 'verified' three times in one day without building anything.
Recover with:  rm -f compiler/.pascal26.fixedpoint && make compiler/pascal26

Stub ticket: signal only. Track T agent (face 2) enriches or a dev track takes it from the repro line.

Log

• 2026-09-08 — auto-closed by the seven watcher: test-arm32#src:tools/compiler_srchash.sh passes at 5506794317a7 (tier full); it was red at b29428afe251. Reopening is by a fresh NEW-RED stub, since a second red is a second finding with its own range.