github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nwc10 good *, #moarvm 05:08
05:41 linkable6 joined 06:17 lizmat_ joined 06:20 lizmat left 06:42 frost-lab joined 06:43 domidumont joined 07:28 MasterDuke joined 07:37 domidumont left 07:52 zakharyas joined, MasterDuke left 07:53 MasterDuke joined
MasterDuke blog.reverberate.org/2021/04/21/mu...eters.html interesting 07:57
08:01 sena_kun left
MasterDuke gcc.gnu.org/pipermail/gcc/2021-Apr...35891.html and some follow-up on computed goto vs tail call 08:03
08:03 sena_kun joined 08:09 sena_kun left 08:11 sena_kun joined 08:16 lizmat_ is now known as lizmat
MasterDuke now that the release has happened, i'm hoping github.com/MoarVM/MoarVM/pull/1474 and github.com/MoarVM/MoarVM/pull/1426 and github.com/MoarVM/MoarVM/pull/1470 can get some eyes on 08:16
jnthn: in the remove opt PR, github.com/MoarVM/MoarVM/pull/1426...667495190d is of particular question 08:18
08:24 kawaii left 08:30 kawaii joined 09:01 brrt joined
jnthn morning o/ 09:07
09:09 jnthn sets mode: -o jnthn
MasterDuke timotimo: i can't even get a dead simple example to work with tracy 09:10
nm, got it working 09:18
brrt moarning 09:28
nwc10 \o 09:29
09:36 brrt left
MasterDuke arg, but i haven't yet got it hacked into the moarvm build... 09:53
10:20 bisectable6 joined
MasterDuke how do i get the Makefile to just add 3rdparty/tracy/TracyClient.cpp to the compile? i.e., my simple test was just `gcc -DTRACY_ENABLE -g -march=native -I tracy/ -lstdc++ -lpthread -ldl -O2 -o at.out tracy/TracyClient.cpp test.c` and that worked (complied and generated data when run) 10:21
nwc10 MasterDuke: maybe something like git show 1b60490d2fa1e59db5e21ccfa1ec2b31142860cd build/Makefile.in 10:23
MasterDuke 'fatal: bad object 1b60490d2fa1e59db5e21ccfa1ec2b31142860cd' 10:24
nine MasterDuke: commit is from Apr 17, maybe git remote update?
MasterDuke ah, thanks 10:25
nwc10 "this week" :-)
10:28 sourceable6 joined
MasterDuke hm, that gets farther...sort of. if i `cp 3rdparty/tracy/TracyClient.cpp 3rdparty/tracy/TracyClient.c` then i get: 10:30
compiling 3rdparty/tracy/TracyClient.o
In file included from 3rdparty/tracy/common/TracySystem.cpp:41,
                 from 3rdparty/tracy/TracyClient.c:14:
3rdparty/tracy/common/TracySystem.hpp:8:1: error: unknown type name ‘namespace’
if i don't copy then i still get the `/home/dan/p6/MoarVM/src/main.c:281: undefined reference to `___tracy_emit_zone_begin'` 10:31
jnthn Um, sounds like that's C++ code, not C code. 10:40
MasterDuke yeah, it has an extern "C" 10:42
just got it built...let's see if it works...
looks like... 10:43
imgur.com/a/KM8DnCM 10:46
.tell timotimo imgur.com/a/KM8DnCM 10:47
m: say (^1_000_000).grep(*.is-prime).elems    # this is what i ran
camelia 78498
11:00 domidumont joined 11:10 MasterDuke left 11:12 MasterDuke joined
MasterDuke and the changes i made gist.github.com/MasterDuke17/7ddaf...ecc55a4a37 11:12
11:35 zakharyas left 12:27 brrt joined 12:39 frost-lab left 12:52 zakharyas joined 13:17 sena_kun left
lizmat and another Rakudo Weekly News hits the Net: rakudoweekly.blog/2021/04/26/2021-...leastable/ 13:25
13:28 notable6 joined
cog lizmat++ 13:30
Is 7.0 still a thing ? 13:31
13:32 zakharyas left, zakharyas joined, tellable6 joined
lizmat cog: it's a teaser :-) 13:34
13:39 MasterDuke left 13:42 MasterDuke joined 13:48 sena_kun joined 13:58 MasterDuke left, MasterDuke joined 14:33 brrt` joined 14:35 brrt left 14:38 brrt` is now known as brrt, bonsaikitten left, xiaomiao joined
MasterDuke hm, after adding the tracy tracking macros to MVM_(m|c)alloc and MVM_free, the collection program dies with `Instrumentation failure: Memory allocation event was reported for an address that is already tracked and not freed.`, but valgrind doesn't complain... 15:04
15:09 cog left
nine Maybe it missed a free? 15:11
MasterDuke hm, if i add --valgrind and then run under valgrind, valgrind itself dies after the program completes with `Memcheck: mc_main.c:5765 (vgMemCheck_is_valid_aligned_word): Assertion 'VG_IS_WORD_ALIGNED(a)' failed.` and then a valgrind internal stacktrace 15:14
nine: i'm not sure, still exploring this thing. got the same with with --full-cleanup
nine: unrelated i suspect, but github.com/MoarVM/MoarVM/issues/1105 might be up your alley 15:23
and still present 15:24
nine MasterDuke: looks to me like that's just missing cleanup of worker threads 16:02
part of my more_asan_fixes work in process
s/process/progress/
16:11 brrt left
MasterDuke cool 16:12
16:21 MasterDuke left 17:00 LizBot left, LizBot joined 17:11 domidumont left 17:40 zakharyas left 18:12 MasterDuke joined 18:27 brrt joined 18:58 patrickb joined 19:57 zakharyas joined 20:15 MasterDuke left 20:23 MasterDuke joined 20:53 zakharyas left 20:56 brrt left 21:09 patrickb left 22:05 dogbert11 joined 22:07 dogbert17 left 23:14 cog joined 23:28 dogbert11 left 23:29 dogbert11 joined 23:42 Kaiepi left