|
github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm Set by AlexDaniel on 12 June 2018. |
|||
|
00:18
Kaiepi left
00:36
Kaiepi joined
00:54
Kaiepi left
00:57
Kaiepi joined
01:32
lucasb left
04:13
unicodable6 left,
shareable6 left,
releasable6 left,
sourceable6 left,
notable6 left,
greppable6 left,
coverable6 left,
nativecallable6 left,
quotable6 left,
bisectable6 left,
reportable6 left,
bloatable6 left,
benchable6 left,
committable6 left,
statisfiable6 left,
squashable6 left
04:14
sourceable6 joined,
statisfiable6 joined,
reportable6 joined,
bloatable6 joined
04:15
coverable6 joined,
nativecallable6 joined
04:16
unicodable6 joined,
quotable6 joined,
benchable6 joined,
bisectable6 joined,
squashable6 joined,
committable6 joined,
shareable6 joined,
greppable6 joined,
notable6 joined
04:17
releasable6 joined
04:32
squashable6 left
04:36
squashable6 joined
05:39
evalable6 left
05:41
evalable6 joined
06:48
domidumont joined
07:25
squashable6 left
07:28
squashable6 joined
07:53
domidumont left
08:08
domidumont joined
08:19
domidumont left
08:22
domidumont joined
08:59
zakharyas joined
09:00
sena_kun joined
09:02
AlexDani` joined
09:04
tobs` joined
09:05
harrow` joined,
moritz_ joined,
nine_ joined
09:06
hoelzro_ joined
09:10
sena_kun left,
AlexDaniel left,
nine left,
Geth left,
moritz left,
harrow left,
tobs left,
hoelzro left,
tadzik left,
tobs` is now known as tobs
09:17
tadzik joined
09:19
sena_kun joined
09:36
zakharyas left
09:49
hankache joined
09:59
zakharyas joined
10:13
hankache left
10:40
patrickb joined
10:54
sena_kun left
11:10
sena_kun joined
12:05
MasterDuke left
12:08
zakharyas left
12:55
sena_kun left
13:10
sena_kun joined
14:03
zakharyas joined
14:06
lucasb joined
14:19
robertle joined
14:40
brrt joined
14:55
sena_kun left
|
|||
| brrt | \o | 15:10 | |
| tellable6 | 2019-12-11T07:15:53Z #moarvm <nine> brrt: I'd be grateful for any such hints! My understanding of the JITs control flow is limited | ||
| brrt | uh, I've not been here for a while | ||
|
15:10
sena_kun joined
|
|||
| nwc10 | o/ | 15:19 | |
|
15:25
brrt left
|
|||
| nine_ | brrt: well, I did find it out in the end :) colabti.org/irclogger/irclogger_lo...-12-14#l32 | 15:36 | |
| tellable6 | nine_, I'll pass your message to brrt | ||
| nwc10 | nine_: The misunderstanding between me and rjbs is now correct, and rjbs kindly did all the dirty work for me and made an IO trial release for the fix we need: metacpan.org/changes/release/RJBS/IO-1.41_50 | 15:38 | |
| er, corrected | |||
| anway, mor eimportant: matrix.cpantesters.org/?dist=IO+1.41_50 | |||
| and I see some red that wasn't there earlier, so mmm.... | |||
| oh, test portability fail and unreleated. eg here's an older version: www.cpantesters.org/cpan/report/ace...be17da2fe0 | 15:40 | ||
| Guest38485 | nine_: did you see that |Tux| had some complaint today as well? | 15:57 | |
|
15:57
brrt joined
16:44
zakharyas left
16:54
sena_kun left,
brrt left
|
|||
| lizmat | Guest38485: I think nine acknowledged and reproduced that ? | 17:00 | |
|
17:09
domidumont left
17:10
sena_kun joined
|
|||
| nine_ | nwc10: sounds good! Thanks for the heads up | 17:14 | |
| Guest38485: yes, I think it's just a different symptom of the same issue | 17:15 | ||
| Huh, that's interesting: while I cannot provoke the issue by running the GC more often, it actually does the opposite: with a small nursery I cannot reproduce the issue at all | 17:19 | ||
|
17:21
nine_ is now known as nine
|
|||
| jnthn | I guess sometimes you can cause a thing to move into gen2 sooner, from which point on it doesn't move, and so "wrong nursery" checks won't fire. | 17:27 | |
| nine | That's a very good hint! | 17:28 | |
| Oddly the errors I get are all at Raku level about wrong types or an unexpected VMNull | 17:29 | ||
|
17:29
domidumont joined
|
|||
| nine | Most often: Native call expected return type with CPointer representation, but got a P6opaque (Scalar) | 17:29 | |
| jnthn | I assume it's not sensitive to spesh and friends? | 17:30 | |
| nine | Disabling spesh makes it more reproducible | ||
| Increasing MVM_NURSERY_SIZE _and_ MVM_NURSERY_THREAD_START drastically makes it go away, too. This kinda confirms the suspicion that it's a GC issue | 17:36 | ||
| (as now objects won't move because they stay in the nursery) | |||
| Luckily I do have a bit of experience with GC issues :) | |||
| But now I have to leave for some fitness training | 17:37 | ||
|
17:42
AlexDani` is now known as AlexDaniel,
AlexDaniel left,
AlexDaniel joined
18:00
brrt joined,
patrickb left
18:21
patrickb joined
18:29
brrt left
|
|||
| dogbert17 | nine: FWIW here's what I get: gist.github.com/dogbert17/b40c3033...63e0fa3a89 | 18:38 | |
|
18:40
Kaiepi left
18:45
Kaiepi joined
18:50
MasterDuke joined
18:53
sena_kun left
18:55
Kaiepi left
19:09
sena_kun joined
19:19
Kaiepi joined
19:21
Geth joined
19:30
domidumont left
20:13
zakharyas joined
20:54
sena_kun left
21:07
patrickb left
21:10
sena_kun joined
21:17
brrt joined
21:44
zakharyas left
22:01
Guest38485 left
22:12
sena_kun left
22:13
brrt left
22:27
MasterDuke left
22:53
Kaiepi left
|
|||