It's from the "default" branch (unreleased 3.x) instead of release-2.x.y, and certainly not intended for release-2.x.y.
Quoting the commit, it seems to rely on the assumption that "DNS processes are already reaped in libpurple". This was true for the default branch around 2012 (I asked the author, apparently some queue stuff was introduced a few commits earlier) and is sort-of-true for the default branch now that it uses GResolver which is thread-based instead of fork-based.
It's not a safe assumption for release-2.x.y in its current state, it needs child process reaping.
Hi, pidgin dev here, got a report from an ubuntu artful user that it generated a bunch of zombie processes.
I noticed a patch named "hg-remove- SIGCHLD- handler. patch" that only exists in ubuntu and not in debian, and seems to be a backport of https:/ /bitbucket. org/pidgin/ main/commits/ 161320946afdd6b c331026fa774e1f 7bc680f12c
It's from the "default" branch (unreleased 3.x) instead of release-2.x.y, and certainly not intended for release-2.x.y.
Quoting the commit, it seems to rely on the assumption that "DNS processes are already reaped in libpurple". This was true for the default branch around 2012 (I asked the author, apparently some queue stuff was introduced a few commits earlier) and is sort-of-true for the default branch now that it uses GResolver which is thread-based instead of fork-based.
It's not a safe assumption for release-2.x.y in its current state, it needs child process reaping.
I recommend removing this patch.