github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
timotimo cdn.discordapp.com/attachments/538...nknown.png - showing incoming refs of the STable of QAST::Var shows a whole bunch of QAST::Var objects appearing in a heap snapshot 02:33
Guest3346 .seen timotimo 15:06
tellable6 Guest3346, gist.github.com/e9190f0dccb3a64ca9...c52570c7ee
Kaiepi can someone review M#1190? 17:10
synopsebot M#1190 [open]: github.com/MoarVM/MoarVM/pull/1190 [IP6NS Grant] Close synchronous sockets properly
dogbert17 Ah, it's MasterDuke 18:38
been looking for you :) 18:39
MasterDuke heh. i saw that mention yesterday and wasn't sure why at first
then figured out it looked like the valgrind problem was called by my more-context-in-error-messages commits 18:40
dogbert17 it's about a valgrind issue with t/spec/S17-procasync/encoding.t
try it
MasterDuke yep, can repro 18:41
dogbert17 do you think it's been there all the time or are the changes to the file somehow involved? 18:42
btw, I'm looking for timotimo as well regarding another matter :) 18:43
MasterDuke i'm not 100% sure, but it looks like my code is involved, so i would assume it's recent 18:44
dogbert17 I believe so although I'm far from certain 18:45
MasterDuke looks like it might be the cast of `MVMGrapheme32 *buffer` into `char *` 18:48
dogbert17 if I go back to MoarVM b72a790b1b21c88c the problem vanishes
aha 18:49
timotimo o/ 19:31
dogbert17 hi timotimo 19:37
have a quick question for you? 19:38
is this code in use? github.com/MoarVM/MoarVM/blob/mast...ot.c#L1347 19:39
timotimo only if zstd is detected during Configure.pl time 19:40
dogbert17 cppcheck complains about lines 1364-5 claiming that 'entry' is not initialized. Could that be a problem? 19:41
timotimo ha 19:52
yeah, that could be kind of bad
on the other hand, i think entry isn't actually used?
does it highlight a place where it's used? 19:53
dogbert17 nope, I don't think the tool is that good 20:05
timotimo OK 20:10
because just looking over the code i couldn't find one
was potentially a left-over from copypasting the ver2 functions to make the ver3 ones
MasterDuke i probably won't have time to work on it again tonight, but the reason i used MVM_exception_throw_adhoc_free was because brrt didn't like repeating a free() call in every branch. is there a way to use a cast and/or MVM_exception_throw_adhoc_free that won't annoy valgrind? 20:57

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!