dogbert17 timotimo: when using MVM_SPESH_LIMIT and setting it to the first crashing value then the last specialization is indeed for Spesh of 'get-proxy' (cuid: 95, file: site#sources/FD28A8E22DFE16B70B757D9981C7B6C25543060C (HTTP::UserAgent):368) 00:20
00:29 greppable6 joined
timotimo i take it the spesh log is still ginormous? 00:35
it probably compresses *really* well
dogbert17 20 megs gzipped 01:21
02:57 ilbot3 joined
samcv that's pretty big 03:19
08:30 AlexDaniel joined 10:56 quotable6 joined 15:08 zakharyas joined 15:11 zakharyas joined 15:18 Kaiepi joined 16:10 zakharyas joined 18:06 zakharyas joined 18:25 zakharyas joined 18:33 domidumont joined 18:38 lizmat joined 18:54 zakharyas joined 19:53 buggable joined
MasterDuke i just tried doing a --profile-stage=optimize of the rakudo build, but i'm getting a segfault here github.com/MoarVM/MoarVM/blob/mast...rp.c#L5922 19:57
the backtrace is just #2 main, #1 MVM_vm_run_file, #0 MVM_interp_run 19:58
19:59 buggable joined 20:26 Kaiepi joined
timotimo i see a juicy problem with profile-stage 21:14
for you see, run_profiled goes through @END and call everything in there
er, no, not quite
21:30 Kaiepi joined
MasterDuke timotimo: thoughts on how to fix it? 22:19
timotimo it ended up not being an actual problem, then i got sidetracked 22:22
in other news, i'm real close to publishing my blog post 22:23
MasterDuke nice
timotimo this was fun, I hope I'll enjoy the next posts as much as I did this one 22:24
MasterDuke more power to you. through forced practice in school and work i've gotten not terrible at writing, but i still never enjoy it 22:37
timotimo i think the result is also not terrible 22:40
i just did a pass over the text splitting long sentences that were already multiple parts in the middle, so hopefully the text is more comfortable to read 22:43
22:58 buggable joined
timotimo wakelift.de/2018/03/05/delays-and-delights/ - my first TPF grant report post is up now 23:25
MasterDuke timotimo++ 23:29