github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
00:01 patrickz left 02:38 fake_space_whale left 03:21 lucasb left 06:48 domidumont joined, domidumont left 06:49 domidumont joined 07:08 robertle joined 07:26 patrickb joined 07:34 zakharyas joined 08:08 lizmat joined 08:35 lizmat left 08:36 lizmat joined 08:52 fake_space_whale joined 09:34 scovit_ left 09:36 lizmat left 09:37 lizmat joined, lizmat left, lizmat_ joined 09:42 sena_kun joined, scovit joined 10:09 domidumont left 10:25 lizmat_ is now known as lizmat
kawaii Just bumping here to ask about the next MoarVM release :) 10:39
11:14 zakharyas left 11:30 brrt joined
timotimo hmm. learn more about signal handling so i can make sure heap snapshot writing won't get interrupted in the middle of a "blob"? 11:37
perhaps a signal handler that reacts to sigint by saying "send sigint again to immediately terminate" and otherwise would finish up the littlest bit of code before aborting 11:38
or mask sigint for the duration of a piece of work, i imagine the signal handler would "immediately" trigger after the signal is unmasked?
probably something i should worry about ... later :) 11:41
brrt :-) 11:47
timotimo so moarvm now writes the new heap snapshot format, at least i think it does :D 11:53
and i've got a slript that reads one zstd stream and gives back what was left in the input buffer so parsing can continue there
11:56 domidumont joined 12:51 brrt left 13:02 zakharyas joined 13:11 lizmat left, lizmat_ joined 13:19 lizmat_ is now known as lizmat 13:43 domidumont1 joined 13:47 domidumont left 13:54 patrickb left 13:55 patrickb joined
Guest2775 timotimo, jnthn: any theories as to what might be happening here gist.github.com/dogbert17/f902f4a8...c227eeaa17 14:18
jnthn Guest2775: Looks like a premature free, maybe 14:23
14:24 domidumont joined 14:26 domidumont1 left
Guest2775 jnthn: it's this piece of code gist.github.com/MasterDuke17/074d0...57745dd191 14:37
in the shell script (9999999.sh) I have added a 'sleep 25' line in order to hit the timeout logic in the script 14:38
perhaps it has something to do with the sig handlers mentioned in the line 'whenever signal(SIGTERM).merge: signal(SIGINT)' 14:39
14:51 fake_space_whale left 14:58 robertle left 15:25 Altai-man_ joined 15:26 sena_kun left 15:32 sena_kun joined 15:33 Altai-man_ left 15:36 Altai-man_ joined 15:37 Altai-man_ left 15:38 sena_kun left 15:44 robertle joined 16:04 sena_kun joined 16:16 patrickb left 16:51 domidumont left 17:05 lizmat left 17:24 domidumont joined 17:36 zakharyas left 19:07 domidumont left 19:23 patrickb joined 19:37 fake_space_whale joined 19:47 scovit left 19:50 lizmat joined 20:10 zakharyas joined 20:26 lizmat left 20:35 patrickb left 20:36 lizmat joined 20:53 zakharyas left 21:26 lizmat left 21:31 lizmat joined 21:36 scovit joined 21:37 lizmat left 23:07 lizmat joined, sena_kun left 23:13 lizmat left