00:11 |Tux| left 00:41 sena_kun left 00:55 |Tux| joined 00:56 sena_kun joined 01:38 |Tux| left 01:47 |Tux| joined 01:50 ZzZombo_ joined 01:53 ZzZombo left, ZzZombo_ is now known as ZzZombo
Geth_ rakudo: vrurg++ created pull request #3389:
Attribute package fix
02:05 Merfont left 02:06 Merfont joined 02:15 AlexDani` is now known as AlexDaniel, AlexDaniel left, AlexDaniel joined 02:19 |Tux| left
Xliff m: "aa".Int.say 02:20
camelia Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5aa' (indicated by ⏏)
in block <unit> at <tmp> line 1
Xliff m: 'z'.succ.say 02:24
camelia aa
02:24 |Tux| joined
Xliff m: 'z'.succ.say 02:24
camelia aa
Xliff m: 'Z'.succ.say
camelia AA
Xliff m: 'AA'.base(10).say 02:27
camelia No such method 'base' for invocant of type 'Str'. Did you mean any of these?

in block <unit> at <tmp> line 1
Xliff m: 52.base(26).say 02:28
camelia 20
Xliff m: 27.base(26).say
camelia 11
Xliff m: 25.base(26).say
camelia P
Xliff m: 25.base(26).^name.say 02:29
camelia Str
02:41 sena_kun left 02:57 sena_kun joined 03:00 |Tux| left 03:11 AlexDani` joined 03:12 AlexDaniel left 03:35 |Tux| joined
Geth_ rakudo: vrurg++ created pull request #3390:
Warn if attribute accessor cannot be installed
rakudo: vrurg self-assigned A method in a role masks a public attribute in a class doing that role github.com/rakudo/rakudo/issues/3382
0f4ef3cb5c | (Elizabeth Mattijsen)++ | src/core.c/Supply.pm6

  - make it a multi, to allow for more future functionality
  - one candidate for always chomping, one candidate with an option
  - reduce number of variables used, reorganize some conditions
  - let nqp::substr do more work
04:35 evalable6 left 04:36 travis-ci joined
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Make `succeed foo` 30% and `succeed foo,bar` 60% faster 04:36
travis-ci.org/rakudo/rakudo/builds/631090549 github.com/rakudo/rakudo/compare/b...f117502a93
04:36 travis-ci left 04:37 evalable6 joined 04:40 sena_kun left
zostay Hey, I'm looking for a new home for the raku vim plugin. I have been working on changes on my github account, but was considering moving it to the raku organization or creating a new vim-raku org for managing it. 04:43
Any suggestions?
04:57 Voldenet left, sena_kun joined 05:03 Voldenet joined, Voldenet left, Voldenet joined 05:09 squashable6 left 05:11 squashable6 joined
Xliff zostay: A vim-raku org might not be a bad idea. 05:34
05:53 squashable6 left 05:55 squashable6 joined 06:42 sena_kun left
lizmat re Travis: restarted the one failing job, and it turned green 06:51
06:57 sena_kun joined 07:19 ZzZombo left 07:26 squashable6 left 07:29 squashable6 joined 07:49 squashable6 left 07:50 squashable6 joined 08:42 sena_kun left
nine lizmat: but what failed and why? 08:48
08:58 sena_kun joined
lizmat nine: I didn't look, I only start to look when it doesn't go green after a restart 09:12
Travis just flaps too often
nine Which is why we should hunt down these issues 09:14
lizmat *sigh* you're right 09:15
nine And yes, I know exactly how hard this is :/ 09:16
lizmat but it's just bloody annoying having to go through oodles of output to figure out what went wrong
nine I'm currently looking at a backtrace of a segfault that only ever occurs on our production machine and that I've been waiting for several days for to appear again... 09:17
lizmat nine++ # more perseverance than me 09:21
nine Feels more like desparation :D 09:22
what the?! 09:52
Looking at my backtrace, I noticed one crucial difference between the production system and my local test: the segfault happens in the OpenSSL module's code, while I didn't use TLS in my tests. 09:53
So I changed that, fired up the test server in rr record, started my test script and.... segfault. In my test script!
lizmat huh?
nine Which I of course didn't run with rr record...
lizmat feels like some NativeCall weirdness ? 09:54
nine At least I started it in gdb, so now I know that both segfaults involve spesh's frame_walker, and both originate in OpenSSL code 09:56
And both are because the frame we're inspecting unexpectedly doesn't have a spesh_cand 09:57
lizmat nine: feels like an assert could be in place ?
nine Which is odd. It segfaults in the line MVMJitCode *jitcode = f->spesh_cand->jitcode; because f->spesh_cand is NULL. But the line right before that is: if (f->spesh_cand && f->spesh_cand->inlines) { 09:58
So, yeah, the check is already there!
lizmat then memory corruption by another thread ?
nine That kinda makes the most sense, yes. 09:59
Now I really wish I had run this in rr
lizmat was it the spesh thread that segfaulted, or another ? 10:00
nine no, a normal thread
lizmat so maybe spesh was doing something ?
nine No, but there are loads of other threads. Though all of them are waiting. But that doesn't mean too much 10:03
Let's see if I can repro when running in rr record... 10:04
Yes, indeed, I finally catch some segfaults. And....it's all different ones :( 10:16
lizmat so, definitely memory corruption by another thread :-(
nine That's the most likely explanation 10:22
I've definitely found a place where a different thread overwrote some data 10:27
I'm not sure I'm gonna resolve the bug(s), this decade... 10:31
lizmat with about 12 hours to go, I guess that would be a reasonable assumption :-) 10:32
nine Good thing there's a new one right around the corner :)
And I wanna spend most of those 12 hours with my wife playing games and eating delicious food :) 10:33
lizmat good plan! 10:34
10:41 sena_kun left
Geth_ rakudo: c63d8a87d1 | (Elizabeth Mattijsen)++ | src/core.c/Supply.pm6
Implement Supply.lines(:nl-in<foo>)

Allow a Supply to be split on the given line-ending. One could wonder if this shouldn't actually be implemented as a Supply.split.
nine Could also be a use-after-free issue 10:49
10:55 sena_kun joined 11:31 AlexDani` is now known as AlexDaniel, AlexDaniel left, AlexDaniel joined
Geth_ rakudo: 779b0a2ea3 | (Elizabeth Mattijsen)++ | src/core.c/Supply.pm6
Streamline Supply.words similar to Supply.lines

  - remove unused variables
  - use nqp::substr() logic to its fullest
  - don't keep nqp::chars value in variable
Didn't bother to look at any performance difference, don't expect much.
rakudo: 6bd0dd513f | (Elizabeth Mattijsen)++ | src/core.c/Supply.pm6
Remove superfluous invocant redefinitions

Not sure why they were there, but they are not needed.
nine Aha! By replacing MVM_realloc with a MVM_realloc_poisoned I wrote, I an clearly demonstrate a use-after-realloc issue! 12:29
MVM_realloc_poisoned overwrites the old memory area with 0xef which leads to an easily identifiable segfault when dereferenced 12:30
A neat trick I found when debugging all those GC issues :)
lizmat afk for a few hours& 12:31
12:41 sena_kun left 12:55 sena_kun joined 14:41 sena_kun left 14:57 sena_kun joined
AlexDaniel zostay: Raku org is a good place 15:07
zostay: unless you mean it to become a big project on its own 15:08
I don't understand where this idea of creating a separate org for every single projects comes from
but all it does is just make things more difficult
for me personally :) 15:30
nine lizmat: maybe I did find a fix this decade :) We'll see what jnthn++ will say to the proposal 15:54
[Coke] I wouldn't create a new org unless it was really needed. (I did it once back in the day, and wish I hadn't)
15:57 [TuxCM] joined
[TuxCM] Rakudo version 2019.11-319-g6bd0dd513 - MoarVM version 2019.11-103-g2ac1fa27f
csv-ip5xs0.793 - 0.844
csv-ip5xs-206.614 - 7.278
csv-parser42.372 - 43.501
csv-test-xs-200.445 - 0.453
test8.054 - 8.055
test-t1.941 - 1.997
test-t --race0.927 - 0.993
test-t-2033.189 - 33.824
test-t-20 --race11.577 - 11.669
(sorry, I posted the past days in #perl6-dev as I am using my laptop in the U.K. I did not look close enough) 15:58
16:01 Merfont left
Geth_ roast: 9fa8119812 | (Elizabeth Mattijsen)++ | S17-supply/elems.t
Remove silly test

  .elems on type objects generally gives "1", so why not here?
rakudo: cfb655bc71 | (Elizabeth Mattijsen)++ | src/core.c/Supply.pm6
Streamline Supply.elems

Into 2 separate multi methods, one for the timed version, and one for the version that doesn't use time. Also use nqp::op for time keeping to reduce timing overhead.
16:42 sena_kun left 16:47 [TuxCM] left 16:59 sena_kun joined
zostay Thanks for the tips. I'll just put it in raku for now and if it becomes a problem for some reason, I can always move it later. Happy New Year! 17:11
Xliff .seen timotimo 17:13
tellable6 Xliff, I saw timotimo 2019-12-24T14:11:41Z in #whateverable: <timotimo> AFKBBL
zostay Ah, well, I am part of the perl6 org on github, but it looks like only AlexDaniel has access to the raku org for the moment. 17:14
AlexDaniel what's your github handle
zostay zostay
AlexDaniel invite sent 17:15
there are 12 people right now
the problem is that I don't know how to send 332 invites…
zostay coolbeans
makes sense, thx 17:16
AlexDaniel I should just write a script to do that but I've been planning to do it for the last month :S
there's no API endpoint for adding users 17:22
so what am I supposed to do then huh
just curl it and hope there's no captcha, right? 17:23
Geth_ roast: 34ab3343b4 | (Elizabeth Mattijsen)++ | 2 files
Make tests less language name specific
rakudo: b1a68be413 | (Elizabeth Mattijsen)++ | src/core.c/Supply.pm6
Streamline Supply.head

  - separate into multi's
  - simplify the numbered one
rakudo: 604ca34a49 | (Elizabeth Mattijsen)++ | 11 files
Do a lot of "Perl 6" to "Raku" renaming

  - in comments
  - in error messages
  - in introspectable texts
AlexDaniel lizmat++ 18:14
18:26 toddr joined 18:41 sena_kun left 18:57 sena_kun joined
[Coke] AlexDaniel: create a new repo in the original that just says "request access to the new one" and click approve as they come in? 19:02
AlexDaniel [Coke]: I don't understand :) 19:03
you can't request access, as far as I know 19:04
so you'd end up sending more than a hundred invites by hand anyway
which is something I'd like to avoid…
[Coke] ah 19:39
Geth_ nqp: MasterDuke17++ created pull request #593:
Fix compiler warnings in main runner
20:00 travis-ci joined
travis-ci Rakudo build failed. Elizabeth Mattijsen 'Do a lot of "Perl 6" to "Raku" renaming 20:00
travis-ci.org/rakudo/rakudo/builds/631365959 github.com/rakudo/rakudo/compare/c...4ca34a4998
20:00 travis-ci left 20:12 squashable6 left 20:14 squashable6 joined
Geth_ rakudo: 186c4ce6c6 | (Elizabeth Mattijsen)++ | t/02-rakudo/14-revisions.t
Adapt 'make test' to name change as well

Note to self: always run "make test" as well, not just "make spectest"
20:22 squashable6 left, squashable6 joined
Geth_ rakudo: MasterDuke17++ created pull request #3393:
Fix compiler warnings in main runner
20:42 sena_kun left 20:58 sena_kun joined 21:28 |Tux| left 22:10 travis-ci joined
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Adapt 'make test' to name change as well 22:10
travis-ci.org/rakudo/rakudo/builds/631393762 github.com/rakudo/rakudo/compare/6...6c4ce6c64f
22:10 travis-ci left 22:15 |Tux| joined 22:42 sena_kun left 22:56 sena_kun joined 23:22 AlexDaniel left 23:24 AlexDaniel joined, AlexDaniel left, AlexDaniel joined
toddr I'm working on importing rt.perl.org tickets to github. 23:31
I'm hoping to map as many github IDs to people.
Does anyone know any of these folks and if so do they have a github ID? If so can you comment in the gist? gist.github.com/toddr/879988e0e4f6...3622ced9cb 23:33
23:35 |Tux| left
AlexDaniel toddr: github.com/rakudo/rakudo/blob/master/CREDITS 23:37
toddr: github.com/Raku/doc/blob/master/CREDITS
I guess you already checked these? :) 23:38
hmm doesn't look like it
23:38 squashable6 left
AlexDaniel for example Daniel Green → MasterDuke17 23:38
23:39 squashable6 joined
AlexDaniel Timo Paulssen → github.com/timo/ 23:39
23:41 |Tux| joined
AlexDaniel David Farrel is github.com/dnmfarrell 23:42
bartolin toddr: the last three entries in your gist are me: github.com/usev6
toddr I've been using the people listed in github/perl6 23:43
looking at credits thanks. Will refresh when I'm done 23:47
AlexDaniel there are also github.com/MoarVM/MoarVM/blob/master/CREDITS and github.com/perl6/nqp/blob/master/CREDITS but these probably don't have any new info 23:48
toddr perfect 23:49
AlexDaniel another source can be modules.raku.org/search/?q= (perhaps try to Ctrl+F the name of an author and if you're lucky it'll get you to the github repo of one of their modules) 23:51
toddr: ping me once refreshed, I'll try to find the rest
toddr some clearly don't have github based on the credits 23:52
23:52 Mutter joined
toddr ¯\_(ツ)_/¯ 23:52
will do
Mutter Yes, the ddgreen and Daniel green entries map to masterduke17 23:55
toddr [email@hidden.address] = Shitov? the credits are a little misleading 23:57
nm it's ashgti 23:59
the credits are wrong I think
AlexDaniel that's not Andrew I'm pretty sure
google says “John Harrison” but I don't know them