← board

Track guessed as P from the FAILING STEP — line 1 of 2, ./compiler/pascal26 test/test_stackless_gen.pas /tmp/test_stackless_gen26, which names test/test_stackless_gen.pas. Not from the job's name or its src: those describe what the job is ABOUT, and this job's recipe spans 2 source file(s). The ranker reads frontmatter, so this line — not the body — decides who works it; correct it if the guess is wrong.

origin/master has advanced 2 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-core#src:test/test_stackless_gen.pas at cf9b14600039 in step 1/2, ./compiler/pascal26 test/test_stackless_gen.pas /tmp/test_stackless_gen26 (auto-filed by twatch)

Repro

tools/testmgr.py --tier native --job 'test-core#src:test/test_stackless_gen.pas' at cf9b14600039c2f62d7251b0e05330fb74827be9

Range

The named sha cf9b14600039 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 cf9b14600039, last good e7a805d13a09, 1 commit(s) in range — the watcher narrows this by idle bisect; check tstate/TSTATE.md for the current range.

Log tail

pascal26:144: error: compiler error: call to a runtime stub that was never emitted (code offset 0 is the ELF entry point). A frontend driver is missing its stub-emission call for the current flags/target.
(tail)
pascal26:144: error: compiler error: call to a runtime stub that was never emitted (code offset 0 is the ELF entry point). A frontend driver is missing its stub-emission call for the current flags/target.
  near: , ' ' ) ; writeln ; >>> end . unit

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

Log

Closed 2026-09-04 — fixed at HEAD, verified by running the job's own comparison

Re-verified at HEAD 0d090cd1d, compiler sha256 1968c7a7da57, as the ticket's own staleness banner asks. Both steps of the recipe pass:

Consistent with frankb-78's 203b8a8e8 (synthesised try/finally in the token pre-scan), which is the plausible fix; I did not bisect to confirm attribution, so treat the cause as likely rather than established.

Two traps worth recording, because both nearly closed this wrongly in opposite directions.

The ticket names step 1/2 as the failing step. Step 1 is the COMPILE, and it passes now — so "is the failing step fixed?" answers yes while saying nothing about step 2, which is where the actual assertion lives. An auto-filed regression names the step that reported the failure, not the step that tests the behaviour, and closing on the named step alone would have been a green about the wrong half of the recipe.

Going the other way, my own first read said the output was missing its last three lines — because I had printed it through head -6. The truncation was in my instrument, not the program. It answered correctly about the first six lines and I read it as the whole output. Running the job's real comparison is what settled it; a partial view of the right data is as misleading as the wrong data.

track: P was GUESSED from the failing step's path, as its own banner says, and was never corrected. Recorded here so the guess is not mistaken for a finding by anyone reading the done/ entry.