← board

regression CASCADE: 15 jobs newly red at b45c759f9e65 (auto-filed by twatch)

Repro (start with a suspect, or any listed job)

tools/testmgr.py --tier full --job '<job>' at b45c759f9e65ec3fee4f71ad1e2ef26f65618bb1

Newly red jobs

Cascade stub: one signal for one event. Track T agent (face 2) or the owning dev track triages the root; individual tickets only for whatever remains red after the root is fixed.


Triage (2026-07-31, Track T)

Still red in all three full runs today (b45c759, 78847f9, 6104264) — the same 15-job set, no drift.

The c-conformance leg is explained and now split out to [[bug-c-main-missing-implicit-return-zero]]: 00211.c's main falls off the end without a return, and the C frontend emits no implicit return 0, so the exit status is stack garbage. It reported exit=188, 140, 76 across the three runs above — a genuine nondeterministic bug, not the flake it looks like. That contradicts this ticket's "ONE root cause" framing for that leg, hence the separate ticket; the split is deliberate.

Still unexplained: test-sqlite-threads-* (all four targets, x86-64 included) and test-lua-cross. The x86-64 sqlite-threads failure is the interesting one — it rules out a purely cross-target/qemu explanation.

Separately, a phantom episode worth recording: at 8b98dc33c6b9 the full tier reported NEW-RED for five unrelated core jobs (test_class.pas, cagg_init_local_b41.c, c_lua_opcode_decode_b132.c, test_cross_param_2darray.pas, test_managed_store_via_addr_b279.pas), all FIXED at the next native run. Only four commits landed between — three progress-docs edits and one perf(nilpy) change scoped to compiler/builtin/pylib.pas, which cannot touch any of them. The opt tier at that same sha was GREEN. That is a harness/build event, and it is the second today. Tracked as Track T tooling work, not a compiler bug.


CLOSED — the jobs are green (Track T, 2026-08-01)

Verified against live tstate at full-tier sha 832a0ad03776 (GREEN, 345.5s): every job this ticket names is passing, and xeon reports 0 failing jobs across the whole 1637-job matrix.

Closed as an auto-filed stub whose underlying failure is gone, not as work done here. Left in done/ rather than deleted so the signal history stays readable.

For the optdiff ones specifically: the compiler defect behind them was [[bug-c-wide-string-literal-narrow-in-value-context]], and the reason there are several tickets for one bug is [[bug-t-optdiff-shard-identity-is-positional]] — the failure migrated shard identity (5 -> 0 -> 2) as test files landed, re-filing itself each time. That ticket is still open and is the thing worth fixing; these stubs are its symptom.

Log