github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
MasterDuke ugh, still getting `MoarVM panic: Internal error: zeroed target thread ID in work pass` when trying to --profile-stage=parse the rakudo build 00:55
Kaiepi finished my work on openbsd's libuv port upgrade! fingers crossed i don't have to do any more on it before i can get moarvm upgraded after the next release 01:32
02:02 Kaiepi left 02:39 squashable6 left 02:42 squashable6 joined 03:20 Kaiepi joined 05:41 squashable6 left 05:47 squashable6 joined 05:51 squashable6 left 05:55 squashable6 joined 06:18 zakharyas joined 06:37 zakharyas left 08:40 domidumont joined 08:53 domidumont left 09:36 brrt joined
brrt \o 09:36
dogbert17 o/ 09:50
09:50 brrt left 10:50 camelia left 11:23 sena_kun joined
lizmat wonders how difficult it would be to JIT nqp::strfromcodes ? 12:12
12:35 ggoebel joined
ugexe looks pretty straight forward 12:58
12:59 squashable6 left
ugexe you could just use e.g. typeparameterized as a template for the graph.c and core_templates.expr bits 13:00
13:05 brrt joined, squashable6 joined 13:57 Kaiepi left
brrt I'm going to rid myself of the MVM_JIT_STORAGE_GPR / MVM_JIT_STORAGE_FPR hack 13:59
because, the register id itself already encodes the regiser class
hmm, that doesn't entirely work. 14:04
ok, well, new plan
timotimo did anybody go ahead on strfromcodes yet? 14:05
jnthn No tme :)
brrt no, not me
there's actually a bunch of PRs that I haven't looked at 14:06
that are probably good
Geth MoarVM: fe472ce935 | (Kaz Wesley)++ | 3 files
enforce wb in get_attribute of C reprs

Fixes non-deterministic segfaults in anything that non-trivially uses CStruct/CPPStruct/CUnion reprs.
Also fixed a separate bug in CUnion: it wasn't caching child objects for attributes that are inlined.
Tested to fix the issue in MoarVM #800, Skarsnik/perl6-gumbo#5 and perl6 RT 131003. Probably also fixes MoarVM #751 and #1087
synopsebot RT#131003 [open]: rt.perl.org/Ticket/Display.html?id=131003 [SEVERE][SEGV] Heap corruption when using Gumbo
Geth MoarVM: 6e232cf2ca | (Jonathan Worthington)++ (committed using GitHub Web editor) | 3 files
Merge pull request #1119 from kazcw/c-repr-wb

enforce wb in get_attribute of C reprs
jnthn There's one less now
14:07 Kaiepi joined
brrt :-) 14:07
ugexe timotimo: im pushing a PR now :(
jnthn timotimo: Is the failing appveyor in github.com/MoarVM/MoarVM/pull/1110 a sign the PR ain't right? 14:08
Geth MoarVM/configurable-subsystems: 8ce8fade71 | (Timo Paulssen)++ | 3 files
confprog: implement add/sub/mul/div, negate.
timotimo ugexe: thank you
brrt jnthn: it depends, I think 14:09
we sometimes have intermittent failure
timotimo i think at that point windows wasn't building at all
===SORRY!===
910No backends specified on the command line.
911Please use --backends or --gen-moar
912
jnthn Hm, can we run it again? 14:10
timotimo i'll try
jnthn I can't find the button to do so
timotimo ok it's re-running
it's in samcv's namespace, so perhaps you just don't have permissions there 14:11
brrt seen samscv
timotimo it's supposed to be on the page that says "MoarVM", then the tab list with "Current build" and the jobs at the far bottom
brrt what's that command....
.seen samcv
yoleaux I saw samcv 25 May 2019 17:54Z in #perl6-dev: <samcv> the bad error message wasn't very helpful though
timotimo now who was it who was trying to get that PR tested on windows with me 14:12
was it you, ugexe?
anyway, it'll take about 4 more minutes for the other appveyor build to finish, then the mp_set one ought to start 14:13
ugexe yep
timotimo maybe it's necessary to rebase that pullrequest on latest master before restarting appveyor's attempt
OK the mp_set pr is building now 14:19
Geth MoarVM: ugexe++ created pull request #1120:
jit strfromcodes
14:23
14:38 brrt left
dogbert17 t/04-nativecall/06-struct.t fails for me 14:38
ugexe timotimo: the failing test passes on windows with your pr 14:43
Geth MoarVM: 4aa256ba56 | (Timo Paulssen)++ | src/math/bigintops.c
attempt to have fast mp_set and not broken windows
16:27
MoarVM: c374d7fd28 | (Jonathan Worthington)++ (committed using GitHub Web editor) | src/math/bigintops.c
Merge pull request #1110 from MoarVM/multiplatform_mp_set_int

attempt to have fast mp_set and not broken windows
16:28 domidumont joined
Geth MoarVM: 55b27c2aac | (Ben Davies)++ | 2 files
Fix miscellaneous compiler flag passing on OpenBSD

  -fno-omit-frame-pointer and -fno-optimize-sibling-calls were getting
overwritten by -fno-ret-protector the way this was originally written.
16:32
MoarVM: 3494e82dee | (Jonathan Worthington)++ (committed using GitHub Web editor) | 2 files
Merge pull request #1118 from Kaiepi/openbsd

Fix miscellaneous compiler flag passing on OpenBSD
MoarVM: 36469a622a | (Kaz Wesley)++ | 3 files
fix rakudo/rakudo#2965

You can't HAS a type before it's defined, whether it be stubbed or an attempt at recursive HASation. This changes the penalty from segfaulty death to merciful exception.
16:36
synopsebot RAKUDO#2965 [open]: github.com/rakudo/rakudo/issues/2965 I can't HAS stub?
MoarVM: 4e2ee39bef | (Jonathan Worthington)++ (committed using GitHub Web editor) | 3 files
Merge pull request #1116 from kazcw/cant-has-incomplete-type

fix rakudo/rakudo#2965
jnthn chuckles at the choice of symbol names in gist.github.com/MasterDuke17/ab710...13033ea3f8 :P 16:38
MasterDuke ha, i didn't even notice 16:41
Geth MoarVM: fbd48017a0 | (Daniel Green)++ | 3 files
Remove profiling overhead from profiles

By calculating the average cost of executing the profiling functions and then subtracting that times the number of entries from the exclusive cost of an HLL block.
MoarVM: dbbfe3d362 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 3 files
Merge pull request #1111 from MasterDuke17/remove_profiling_overhead_from_profile

Remove profiling overhead from profiles
MasterDuke jnthn++ on a merging spree 16:42
jnthn: gonna do an nqp/rakudo bump? 16:43
16:43 ggoebel left
Geth MoarVM: 6053919e0c | (Nick Logan)++ | 2 files
jit strfromcodes
16:46
MoarVM: fbb1ddd8e7 | (Jonathan Worthington)++ (committed using GitHub Web editor) | 2 files
Merge pull request #1120 from ugexe/jit-strfromcodes

jit strfromcodes
lizmat hmmm... smells like a bump's coming up :-) 16:47
jnthn MasterDuke: Could do a MoarVM one; if I get chance, I'll merge my NQP parse improvements later today, so could bump after that :)
MasterDuke oooh, looking forward to that
jnthn otoh, want to watch Portual vs Nizozemsko this evening also :) 16:49
Uhh...*Portugal, *Netherlands :)
lizmat: Did you mention that a load of time was spent in strfromcodes? 16:50
I suspect for the common case there's a speedup...
17:08 travis-ci joined
travis-ci MoarVM build passed. Jonathan Worthington 'Merge pull request #1120 from ugexe/jit-strfromcodes 17:08
travis-ci.org/MoarVM/MoarVM/builds/543432771 github.com/MoarVM/MoarVM/compare/d...b1ddd8e72c
17:08 travis-ci left 18:22 dogbert11 joined 18:23 dogbert17 left 18:49 domidumont left 19:08 AlexDaniel joined 19:36 sena_kun left 19:39 sena_kun joined 20:03 sena_kun left 20:17 brrt joined 21:23 squashable6 left 21:24 squashable6 joined 21:50 sena_kun joined 21:57 brrt left 22:02 Altai-man_ joined 22:03 sena_kun left 22:06 Voldenet left 22:10 Voldenet joined, Voldenet left, Voldenet joined 22:20 Altai-man_ left
Geth MoarVM: f2ee9baaf7 | (Ben Davies)++ | 9 files
Implement freemem and totalmem ops
23:05
MoarVM: 5f7dce2412 | (Ben Davies)++ | 2 files
Change MVM_platform_cpu_count count back to int, fix interp.c op order
MoarVM: ae5bddff3d | (Jonathan Worthington)++ (committed using GitHub Web editor) | 9 files
Merge pull request #1114 from Kaiepi/mem

Implement freemem and totalmem ops