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-nilpy#src:test/test_nilpy_print_arg_eval_order.npy@1 red at 9df2717684a3 (auto-filed by twatch)
- Type: regression (auto-filed by Track T watcher, host xeon). Untriaged.
- Found: 2026-08-04T05:03:29Z
- Test source: test/test_nilpy_print_arg_eval_order.npy
Repro
tools/testmgr.py --tier full --job 'test-nilpy#src:test/test_nilpy_print_arg_eval_order.npy@1' at 9df2717684a330d02747d73cffe34ffb575ca9ac
Range
bad 9df2717684a3, last good unknown, 0 commit(s) in range — the watcher narrows this
by idle bisect; check tstate/TSTATE.md for the current range.
Log tail
ok: /tmp/testmgr-scratch-288097/test_nilpy_printorder26 [code=1325814B data=32916B bss=9564B procs=1147]
Stub ticket: signal only. Track T agent (face 2) enriches or a dev track takes it from the repro line.
Fixed 2026-08-04 (claude-AN — I caused it)
Not a code regression: c8093ef11 overwrote an existing test file.
test/test_nilpy_print_arg_eval_order.npy already existed — it is the test for
[[bug-nilpy-print-emits-arguments-before-evaluating-later-ones]], wired at
Makefile line 1110 with its own expectation ("label = ERR", the star-unpack
rows, …). I wrote a new test for a different print bug under the same name and
added a second Makefile recipe for it, so the file passed my recipe and failed
the original's.
The compiler change in c8093ef11 is correct and unaffected; only the test file
was clobbered.
Restored the original from c8093ef11^ and renamed mine to
test_nilpy_print_container_arg_freshness.npy (+ .expected), which is what it
actually tests: that a container argument's TEXT is produced after every
argument is evaluated. Both recipes now run their own file, and both pass.
The check that would have caught it: the two names were close enough to read
as the same concern — "print arg eval order" vs "print evaluates args before
writing" — so ls test/ | grep <topic> before creating a test file is worth
more than it looks. A new test that needs a NEW Makefile recipe beside an
existing one for the same file is the tell.
- 2026-08-04 — resolved, commit 6d78481bb.