|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
00:08
MasterDuke joined,
p6bannerbot sets mode: +v MasterDuke
00:16
MasterDuke left,
MasterDuke joined,
herbert.freenode.net sets mode: +v MasterDuke,
p6bannerbot sets mode: +v MasterDuke
|
|||
| timotimo | i'm going through interp.c and i'm finding sooo many ops that are not being instrumented for allocation logging | 00:25 | |
| MasterDuke | timotimo++ | 00:33 | |
| could that somehow be marked in the oplist? | |||
| timotimo | i've considered that, too | 00:40 | |
|
01:00
avar left,
avar joined,
avar left,
avar joined,
p6bannerbot sets mode: +v avar
|
|||
| timotimo | tomorrow i should have an extra look at whether threads that don't participate in a given gc run are completely losing their data in the profiler at the moment | 01:32 | |
| like, since they don't show up in the gc's list of threads, are the number properly added to the threads that actually do the work? i don't think they are | 01:33 | ||
| and perhaps i should change it to record the data of all threads' nurseries and have the "participated" flag be just whether or not the thread was part of the parallel mark&sweep process | 01:34 | ||
|
02:35
avar left,
avar joined,
avar left,
avar joined,
p6bannerbot sets mode: +v avar
02:36
p6bannerbot sets mode: +v avar
05:10
lizmat left
05:27
lizmat joined,
p6bannerbot sets mode: +v lizmat
05:40
fake_space_whale left
07:17
domidumont joined,
p6bannerbot sets mode: +v domidumont
07:59
domidumont left
08:00
domidumont joined
08:01
p6bannerbot sets mode: +v domidumont
|
|||
| Geth | MoarVM: 2fdde4a210 | (Stefan Seifert)++ | src/spesh/optimize.c Fix malformed DU chain when smart coerce is optimized to unbox Needed to set the writer for the temp and destination registers. Thanks to nwc10++ for pointing this out! |
08:51 | |
| nine | nwc10: yep, was right about that. Fixed now :) | 08:52 | |
|
09:17
travis-ci joined,
p6bannerbot sets mode: +v travis-ci
|
|||
| travis-ci | MoarVM build errored. Stefan Seifert 'Fix malformed DU chain when smart coerce is optimized to unbox | 09:17 | |
| travis-ci.org/MoarVM/MoarVM/builds/450425375 github.com/MoarVM/MoarVM/compare/2...dde4a21014 | |||
|
09:17
travis-ci left
09:21
lizmat left
09:23
domidumont left
09:40
domidumont joined
09:41
p6bannerbot sets mode: +v domidumont
11:17
lizmat joined,
p6bannerbot sets mode: +v lizmat
11:27
lizmat_ joined,
p6bannerbot sets mode: +v lizmat_
11:30
lizmat left
11:38
lizmat_ is now known as lizmat
|
|||
| dogbert17 | nine: I'm testing your branch and getting suspicious valgrind output when running t/spec/S32-str/encode.t | 12:24 | |
| ==32612== Conditional jump or move depends on uninitialised value(s) | 12:25 | ||
| ==32612== at 0x5037808: validate_literal_operand (validation.c:260 | |||
|
12:30
brrt joined
12:31
p6bannerbot sets mode: +v brrt
|
|||
| brrt | \o | 12:31 | |
| nine++ - for finishing | |||
| having said that, istr that there was a reason we didn't have sp_findmeth in the expr JIT yet | 12:32 | ||
| I may be wrong about that | |||
| nine | It seems to work just fine ;) | 13:25 | |
| brrt: did you see my findings about expr jit of "set"? | 13:27 | ||
| brrt | no | ||
| did not | |||
| what's up with that? | |||
| yeah, maybe it wasn't sp_findmeth after all :-) | 13:28 | ||
| nine | Causes: MoarVM panic: Register types do not match between value and node in t/spec/S02-names-vars/list_array_perl.t when ran with MVM_SPESH_NODELAY=1 MVM_SPESH_BLOCKING=1 | ||
| brrt | oh, hmmm | 13:29 | |
| alright, I'll need to take a look I guess :-) | |||
| nine | I'd not be surprised if its about some signed/unsigned mismatch | ||
| brrt | :-) | 13:30 | |
| nine | My mbc backend seems to be one of the heaviest users of unsigneds we have | ||
| brrt | uhuh | ||
|
13:37
MasterDuke left
|
|||
| dogbert17 | nine: was the 'It seems to work just fine ;)' directed to brrt or do you not get the valgrind warnings? | 13:41 | |
| if so I must have made a mistake (or two :-) when building your branch | 13:42 | ||
| nine | dogbert17: was refering to brrt | 13:43 | |
| dogbert17 | phew :) | ||
|
13:55
patrickb joined
13:56
p6bannerbot sets mode: +v patrickb
|
|||
| brrt | nine: the sp_findmeth legacy jit template looks decidedly normal, so I'm sure I'm misrenembering things | 14:00 | |
| nine | good to hear :) | ||
| brrt | (I may be a bit paranoid around these things) | ||
|
15:15
fake_space_whale joined,
brrt left
15:16
p6bannerbot sets mode: +v fake_space_whale
15:52
geekosaur left
15:56
zakharyas joined
15:57
p6bannerbot sets mode: +v zakharyas
16:07
ItchyPlant joined,
p6bannerbot sets mode: +v ItchyPlant,
ItchyPlant left
16:16
lizmat left
16:27
lizmat joined,
p6bannerbot sets mode: +v lizmat
17:41
zakharyas left
20:21
domidumont left
20:26
fake_space_whale left
20:48
zakharyas joined
20:49
p6bannerbot sets mode: +v zakharyas,
zakharyas left
20:50
zakharyas joined,
p6bannerbot sets mode: +v zakharyas
21:01
ugexe left
21:13
lizmat left
21:14
Kaiepi left
21:15
Kaiepi joined,
p6bannerbot sets mode: +v Kaiepi
|
|||
| Geth | MoarVM: Kaiepi++ created pull request #991: Reimplement async socket introspection |
21:47 | |
| Kaiepi | \o/ | 21:50 | |
|
21:54
lizmat joined,
p6bannerbot sets mode: +v lizmat
22:17
zakharyas left
22:59
lizmat left
23:01
lizmat joined,
p6bannerbot sets mode: +v lizmat
23:16
MasterDuke joined,
p6bannerbot sets mode: +v MasterDuke,
MasterDuke left,
MasterDuke joined,
herbert.freenode.net sets mode: +v MasterDuke,
p6bannerbot sets mode: +v MasterDuke
23:42
patrickb left
|
|||