|
00:07
klapperl joined
|
|||
| Geth | MoarVM: jstuder-gh++ created pull request #843: Add 'slice' op |
01:06 | |
|
01:19
FROGGS joined
01:57
ilbot3 joined
01:59
FROGGS_ joined
05:51
AlexDaniel joined
06:29
lizmat joined
06:42
domidumont joined
06:48
domidumont joined
06:53
robertle joined
07:06
Ven`` joined
07:35
Ven` joined
07:57
Ven`` joined
08:30
zakharyas joined
08:35
zakharyas joined
10:34
zakharyas joined
10:59
Kaiepi joined
11:09
Ven`` joined
11:49
Kaiepi joined
11:51
Kaiepi joined
12:03
Ven`` joined
|
|||
| Geth | MoarVM/backtrace_uninline: a6c8806b3c | (Stefan Seifert)++ | src/core/exceptions.c Add inlined frames to backtraces Not yet working. The machinery is there but it fails to find the actual inlined candidate. |
12:18 | |
| samcv | night o/ | 12:44 | |
| see you all tomorrow | |||
|
12:48
zakharyas joined
14:36
FROGGS joined
14:37
tbrowder_ joined
14:59
Ven`` joined
15:04
lizmat joined
15:28
AlexDaniel joined
|
|||
| dogbert17 | hmm, the --profile option can still generate humungous profiles | 16:47 | |
| timotimo | yes, the fix for the inlining problem hasn't landed yet because it's also crashy | 16:48 | |
| dogbert17 | aha, any theories as to why it's crashy? | 16:51 | |
| timotimo | the code isn't getting the right inlines; nine's work on uninlining backtraces will be re-usable here, and i think he's gotten farther than i have | 16:52 | |
| dogbert17 | cool | 16:56 | |
|
17:18
domidumont joined
17:25
AlexDaniel joined
|
|||
| dogbert17 | heh, a coredump | 17:29 | |
| timotimo: what do you make of this? gist.github.com/dogbert17/95d345c7...d6526e0b3b | 17:45 | ||
| timotimo | good old memory corruption? | 17:46 | |
| dogbert17 | valgrind doesn't say anything if I turn off FSA_SIZE_DEBUG | 17:48 | |
| timotimo | oh, maybe it's a bug in fsa_size_debug? | 17:49 | |
| dogbert17 | perhaps we're not supposed to use that flag anymore? | 17:50 | |
| timotimo | if devs haven't kept it on recently, it could have bitrotted a little? | 17:51 | |
| things that violate the fsa size debug can lead to real crashes, so it should be kept in good repair | |||
| dogbert17 | haven't used it for a while but nwc10 found a bug in nqp recently | ||
| nwc10 | this is chance, but I happened to have this window open | 17:53 | |
| paste.scsys.co.uk/577152 | |||
| 3 NQP regression tests fail under ASAN when MoarVM is built with FSA_SIZE_DEBUG | |||
| timotimo | this could be fate! | ||
| nwc10 | t/p5regex/01-p5regex.t t/qregex/01-qregex.t t/serialization/01-basic.t | 17:54 | |
| no, I don't know when this first started | |||
| dogbert17 | uh oh, I hope AlexDaniel doesn't see this :) | 17:55 | |
| AlexDaniel | o_o | ||
| dogbert17 | I turned on the flag again and used ASAN, now it complains earlier: gist.github.com/dogbert17/458d1bcd...bea437fd1c | ||
|
17:59
lizmat_ joined
18:33
robertle_ joined
19:26
zakharyas joined
19:35
zakharyas joined
20:45
lizmat joined
|
|||
| dogbert17 | the problem seems to have surfaced after 2018.03 | 21:25 | |
|
23:06
Kaiepi joined
|
|||