#!watchflakes
post <- goos == "darwin" && (
log ~ `fatal error: found pointer to free object` ||
log ~ `fatal error: found bad pointer in Go heap` ||
log ~ `fatal error: workbuf is not empty` ||
log ~ `fatal error: workbuf is empty` ||
log ~ `schedule: holding locks` ||
log ~ `fatal error: bad summary data` ||
log ~ `fatal: morestack on gsignal` ||
log ~ `fatal error: fault` ||
log ~ `SIGBUS: bus error` ||
log ~ `runtime\.\(\*mheap\)\.allocNeedsZero` ||
(log ~ `panic: runtime error: invalid memory address or nil pointer dereference` && log ~ `^\s*runtime\.sigpanic.*\n\s+.*\n\s*runtime\.sweepone`) ||
(log ~ `panic: runtime error: invalid memory address or nil pointer dereference` && pkg == "cmd/compile/internal/ssa") ||
(log ~ `SIGSEGV: segmentation violation` && log ~ `^\s*runtime\.spanOfUnchecked\(`) ||
log ~ `undefined: s4ring` ||
log ~ `internal compiler error:`)
We have seen a recent rise in flakes that look like memory or disk corruption on darwin builders (amd64 and arm64).
At the moment, I'm assuming these are issues with the builders themselves, but it's also possible that there is a bug in Go itself.
We have seen a recent rise in flakes that look like memory or disk corruption on darwin builders (amd64 and arm64).
At the moment, I'm assuming these are issues with the builders themselves, but it's also possible that there is a bug in Go itself.