IJHack / QtPass / 41477231087

93%

push

github

Executor: bound the wait after kill() in a blocking run (#2110)

* Executor: bound the wait after kill() in a blocking run

stopBlocking() waited forever after kill(). SIGKILL cannot be ignored,
but a child in uninterruptible sleep (a hung NFS or FUSE mount) exits
only when the kernel lets it, and until then held the calling thread
past the run's deadline. Wait one more grace period instead, log a
warning if the child is still there, and return -1 as before: a
blocking run now ends within its timeout plus twice the grace.

QProcess's destructor still waits for such a child, for up to 30 s;
the header says so.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VJygpMUf461yZgfAqwMJ6d
Signed-off-by: Anne Jan Brouwer <brouwer@annejan.com>

* Executor: the timeout does not cover QProcess::start() itself

start() can block, for example opening a FIFO a redirected channel
names, before the deadline is ever checked. No caller redirects one;
say so instead of promising a bound end to end.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VJygpMUf461yZgfAqwMJ6d
Signed-off-by: Anne Jan Brouwer <brouwer@annejan.com>

* Executor: "a FIFO named by a redirected channel"

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VJygpMUf461yZgfAqwMJ6d
Signed-off-by: Anne Jan Brouwer <brouwer@annejan.com>

---------

Signed-off-by: Anne Jan Brouwer <brouwer@annejan.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

2 of 4 new or added lines in 1 file covered. (50.0%)

10086 of 10880 relevant lines covered (92.7%)

398.0 hits per line