You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I wrote a fast writer draining the OS-pipe within microseconds. Problem persists
— FastWriter receives many chunks, the longest 16,571 bytes, the large one in less than 118μs, smaller in less than 20μs
— FastWriter receives additional small chunks after the big chunk
Apple baselessly asserts macOS the best operating system in the world, where writing to a terminal window could erratically take 2 seconds
ianlancetaylor
changed the title
exec.Cmd.Wait exec.Cmd.StderrPipe data-loss race condition go1.22.3 240525 darwin-arm64 14.5
os/exec: Cmd.Wait Cmd.StderrPipe data-loss race condition go1.22.3 240525 darwin-arm64 14.5
May 25, 2024
There is therefore an additional synchronizing wait required:
— in between successful [exec.Cmd.Start] and [exec.Cmd.Wait] where
— any stream obtained via [exec.Cmd.StderrPipe] [exec.Cmd.StdoutPipe] has to be read until io.EOF
the [exec.Cmd.Wait] subsequent to that does not await anything, it does cleanup
Go version
1.22.3
Output of
go env
in your module/workspace:What did you do?
Send SIGABRT to Go-launched sub-process while capturing standard error
Reproduction:
What did you see happen?
33% of cases, a cut-off stack trace:
sys_darwin.go:23 +0x58 fp=%
never fails while debugging
What did you expect to see?
on success, last line a complete line-terminated register line:
fault 0x19d6259ec
The text was updated successfully, but these errors were encountered: