| japhb | Are there any segfault/crash bugs that made it to an actual nqp-m or r-m release (as opposed to only existing within a development cycle)? I can use tests for them to validate my stress testing code .... | 01:18 | |
| timotimo | i hope not! :P | 01:34 | |
| but you can get segfaults manually | |||
| like, for exapmle, you can box_I the value 0 into OpaquePointer and do stuff with it :) | 01:35 | ||
|
02:11
colomon joined
02:39
colomon joined
04:46
ggoebel111111111 joined
06:43
woolfy1 left
08:10
lizmat joined
09:04
lizmat joined
09:13
woolfy joined
09:14
lizmat joined
|
|||
| FROGGS | timotimo: see www.cpantesters.org/distro/S/SDL for the dynamic page | 09:16 | |
| tadzik: when you have time, can we muse about the connection of cpandatesters to github issues/RT? | 09:17 | ||
|
09:36
kjs_ joined
10:04
lizmat joined
10:36
lizmat_ joined
10:38
lizmat_ joined
10:50
FROGGS joined
11:00
kjs_ joined
11:44
kjs_ joined
|
|||
| timotimo | now that i actually look at this stuff, things get a tiny bit more complicated | 12:25 | |
| because the argument to nativecallinvoke is - at least in zavolaj - self, which is the role that boxes the nativecall object | 12:26 | ||
| so i have a little piece of indirection between spesh and the native call body that i should be inspecting | |||
|
14:08
woolfy1 joined
14:28
kjs_ joined
14:46
vendethiel joined
15:42
woolfy1 left
15:44
Ven joined
16:29
Ven joined
|
|||
| Ven | o/ | 16:29 | |
| timotimo | \ô/ | 16:31 | |
| tadzik | FROGGS: cpandatesters :D | 16:33 | |
| FROGGS: sure | |||
|
16:39
zakharyas joined
17:49
FROGGS_ joined
|
|||
| FROGGS_ | tadzik: :P | 18:00 | |
| tadzik: it would be awesome if we could link cpanda test reports to github issues.... like you can reference a commit | 18:01 | ||
| tadzik: though, I don't know how possible that is | |||
| tadzik: at least we should allow issue tracker links to bug reports, so one can see if a problem is already handled or not | 18:02 | ||
| for my P5 modules I often look again and again over test reports, because I cannot see that I already made a ticket for it | |||
| tadzik | heh, for that to be useful I need to write myself that tool to browse though github bugs in a sensemaking manner | 18:28 | |
| as in "show me open bugs in the repos created by me? | |||
| s/\?$/"/ | |||
| it's all I care about, and there's either no way at all or a very tricky way to achieve that | |||
| FROGGS_ | I would like to link to issues manually | 18:32 | |
| so I'd like to have some fields attached to a test report that I can edit | |||
|
19:59
lue joined
|
|||
| timotimo | i wonder if doing a full collection after having done a whole bunch of collections that have promoted almost nothing would make future collections faster because it potentially knocks out things in the gen2? | 20:15 | |
| t.h8.lv/p6profile/formula_profile_latest.html / t.h8.lv/p6profile/formula_profile_latest.pl | 20:21 | ||
| er | |||
| ww | |||
| dalek | arVM: 0809f66 | (Timo Paulssen)++ | src/jit/graph.c: can jit multicacheadd and multicachefind now. |
20:38 | |
| arVM: 824b284 | (Timo Paulssen)++ | src/jit/graph.c: tiny typo, |
20:39 | ||
| arVM: 15d3b75 | (Timo Paulssen)++ | src/jit/graph.c: third time's the charm! |
20:40 | ||
|
20:49
travis-ci joined
|
|||
| travis-ci | MoarVM build failed. Timo Paulssen 'tiny typo,' | 20:49 | |
| travis-ci.org/MoarVM/MoarVM/builds/39036545 github.com/MoarVM/MoarVM/compare/0...4b28497cea | |||
|
20:49
travis-ci left
20:50
lizmat joined
20:57
woolfy joined
20:59
vendethiel joined
|
|||
| timotimo | brrt: i implemented multicachefind and multicacheadd and now invokewithcapture is the blocker for a few interesting frames | 21:09 | |
|
21:30
lizmat_ joined
|
|||
| timotimo | brrt, this may be of interest: t.h8.lv/p6profile/jit_bails_per_frame_name.txt | 21:49 | |
|
22:45
Guest1419 joined
23:21
Guest1419 joined
|
|||