lizmat Files=1301, Tests=109809, 211 wallclock secs (28.67 usr 8.10 sys + 2943.29 cusr 271.84 csys = 3251.90 CPU) 11:02
Geth nqp: 4c96df5436 | (Elizabeth Mattijsen)++ | src/QRegex/Cursor.nqp
Remove HLL dependency

The previous hack was apparently needed because Rakudo's Match.Int called it, rather than having its own HLL equivalent.
11:30
rakudo: 37ddd5984d | (Elizabeth Mattijsen)++ | 2 files
Make sure Match has its own Int coercer

Lack of this made it call the NQP's QRegex/Cursor one, which was originally not capable of handling Raku's Ints. Until about a year ago, when a dependency on HLL was hacked into NQP to fix this.
This commit bumps NQP to remove that hack, and introduces a HLL Match.Int that will make sure it won't descend into the depths of NQP anymore.
11:40
roast: 34ecff098f | (Elizabeth Mattijsen)++ | 11 files
Batch #24 of roast RT -> GH ticket migration
11:56
lizmat only 400 more files to go
sena_kun does anyone here run stresstest? more or less often 12:19
lizmat runs a *lot* of spectests, but no stresstest 12:22
sena_kun I think we have an issue then
because I see hangs on stresstest, but spectest is just ok
lizmat starts one
sena_kun I am running one from b787095358bcbac4c6326b99e9673e54ef5afd58 and t/spec/S17-promise/nonblocking-await.t is dead 12:23
t/spec/S17-promise/nonblocking-await.t ............................ ^C 12:28
Test run interrupted!
All tests successful
ok, this is... odd.
lizmat Files=1357, Tests=187216, 279 wallclock secs (39.23 usr 11.33 sys + 3854.07 cusr 366.25 csys = 4270.88 CPU) 12:34
no hanging for me
sena_kun tries again 12:36
lizmat needs to be afk, probably for the rest of the day
sena_kun lizmat: good luck 12:39
lizmat is back for a bit 15:06
sena_kun: is there something I can do for you in the next 30 mins or so ?
nine sena_kun: I seem to remember nonblocking-await.t being troublesome from time to time 15:20
sena_kun github.com/rakudo/rakudo/issues/3441 looks like a monster to me and is reproduced consistently for me even using master branches for everything. at the very least, it would be very cool if anyone could run releasable script to try out if it is something local or not. 15:28
lizmat sena_kun: confirmed the hang 15:31
sena_kun passing stresstest at some point just stops (usually it is the file specified, I got it a couple of times in a row, so not a rare flapper), stops eating CPU
lizmat: really? with sake all?
lizmat raku t/spec/integration/advent2011-day07.t after checking out the 6.c-errata branch 15:32
sena_kun hmm, I used zscript and it passed, I think
bisect comes to the rescue... 15:33
lizmat found the reason for the hang
git diff 15:34
sena_kun awaits interesting news 15:36
Geth roast/6.c-errata: 9a9ba99c1c | (Elizabeth Mattijsen)++ | packages/Advent/GrammarProfiler.pm
Add reference to method "raku" to prevent infinilooping

Not sure why this fixes it, but the module is playing some tricks with find_method, and so does the Mu.raku/perl logic, so maybe that is it.
sena_kun lizmat: committing to roast means to cherry-picking, I guess... 15:37
lizmat yeah, I'd cherry pick that one 15:38
lizmat has to go afk again, for the rest of the day
sena_kun lizmat: o/
s/to cherry/no cherry/ 15:40
sena_kun ping rba 18:28
sena_kun .tell rba hi! Can I use the blin-intended server for running spectest on various rakudo checkouts? That'll greatly help with releasing process, cause right now I am kind of short on computing resources. I started to setup a setup using rakudobrew && z script combo, but need some things (for example, make) to make things working. My questions are: 1)can I use the blin server this way? 2)can you help with setting up dependencies (at leas 19:12
t making sure rakudobrew build moar passes)?
tellable6 sena_kun, I'll pass your message to rba