Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
00:03 reportable6 left 00:04 reportable6 joined 05:36 greppable6 left, statisfiable6 left, coverable6 left, squashable6 left, notable6 left, linkable6 left, reportable6 left, shareable6 left, nativecallable6 left, committable6 left, tellable6 left, bloatable6 left, evalable6 left, quotable6 left, bisectable6 left, sourceable6 left, benchable6 left, releasable6 left, unicodable6 left 05:37 bloatable6 joined, sourceable6 joined, releasable6 joined, squashable6 joined, unicodable6 joined, bisectable6 joined 05:38 evalable6 joined, coverable6 joined, reportable6 joined, greppable6 joined, tellable6 joined, notable6 joined, committable6 joined 05:39 statisfiable6 joined, shareable6 joined, benchable6 joined, nativecallable6 joined 05:40 quotable6 joined, linkable6 joined 06:02 reportable6 left 06:04 reportable6 joined 06:10 Kaiepi left 06:12 Kaiepi joined 08:09 AlexDaniel left 08:33 linkable6 left 08:34 linkable6 joined 08:35 linkable6 left 08:38 linkable6 joined 09:03 patrickb joined
MasterDuke nine: do you not get an error locally with github.com/MoarVM/MoarVM/pull/1511 ? 09:07
i've gotten some random segfaults when building nqp and rakudo with it 09:18
and running rakudo tests, but i haven't been able to catch any in gdb yet 09:28
nine MasterDuke: huh...no, that PR is actually supposed to fix such things :D 09:52
No coredumps either? 09:53
MasterDuke hm, i can check
i always forget i have to use coredumpctl 09:54
gist.github.com/MasterDuke17/e028a...657eca4b8e 09:55
updated with a little more info 09:57
and some more 09:58
nine But, but, this is before my newly added code? 10:08
Btw I didn't really test it after changing it to MVM_free_null. 10:09
Can you change it back and test again?
10:27 discord-raku-bot left, discord-raku-bot joined
MasterDuke k 10:29
oh, but CI was showing all fails on your first version too 10:35
nine: same segfault 10:47
backtrace looks the same 10:48
10:52 Ven_de_Thiel joined
MasterDuke and no surprise since CI showed it too, but same results with clang 10:59
11:02 linkable6 left 11:05 linkable6 joined
nine Oh, now I see it, too 11:09
lizmat And another Rakudo Weekly News hits the Net: rakudoweekly.blog/2021/06/29/2021-26-r3/ 11:11
11:16 linkable6 left 11:17 linkable6 joined
nine MasterDuke: cought it in rr 11:44
Well, well, in MVMStaticFrame *sf = (MVMStaticFrame *)MVM_repr_at_pos_o(tc, check_frames, i); we actually get a VMNull instead of a static frame. 11:48
11:57 dogbert11 joined, dogbert17 left 12:02 reportable6 left 12:05 reportable6 joined 12:33 frost left
nine Ah, it's simply a good old GC issue 12:45
MasterDuke: pushed a fixed version :) 12:50
MasterDuke nice, and CI seems happy with it so far 13:03
nine slowly exhales 13:04
MasterDuke nine: that missing MVMROOT was a pre-existing problem wasn't it? i.e., the first version of your PR just made it more likely to trigger somehow 13:56
nine MasterDuke: no, because before my change there was no point at which we could have entered the GC. There is now with MVM_gc_mark_thread_blocked(tc); 13:58
MasterDuke ah
15:16 patrickb left 16:44 Ven_de_Thiel left 18:02 reportable6 left 18:03 reportable6 joined 21:05 m6502 joined 21:12 m6502 left 22:03 AlexDaniel joined 22:12 AlexDaniel left 22:28 AlexDaniel joined 22:33 m6502 joined 22:36 m6502 left, jgaz joined 22:58 jgaz left