Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_log/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by Zoffix on 27 July 2018.
00:14 Ven`` joined 00:15 lizmat left, p6bannerbot sets mode: +v Ven`` 00:20 lizmat joined, p6bannerbot sets mode: +v lizmat 01:05 Ven`` left 02:16 SqrtNegInf left
Geth rakudo: 5fcba5b366 | MasterDuke17++ (committed using GitHub Web editor) | src/Perl6/Optimizer.nqp
Fix typo in comment
 02:30
03:01 remi_ness left 03:05 martin__ joined 03:06 p6bannerbot sets mode: +v martin__ 03:09 ufobat left 03:15 remi_ness joined 03:16 p6bannerbot sets mode: +v remi_ness 04:08 SqrtNegInf joined 04:09 p6bannerbot sets mode: +v SqrtNegInf 04:49 MasterDuke left 06:22 martin__ is now known as ufobat 06:51 remi_ness left 07:45 brrt joined 07:46 p6bannerbot sets mode: +v brrt
lizmat Files=1264, Tests=87705, 378 wallclock secs (20.31 usr 5.98 sys + 2689.09 cusr 232.83 csys = 2948.21 CPU) 10:21
11:22 robertle_ joined 11:23 p6bannerbot sets mode: +v robertle_ 12:02 |Tux| joined, Tux__ left 12:03 p6bannerbot sets mode: +v |Tux| 12:05 ExtraCrispy left 12:07 ExtraCrispy joined, p6bannerbot sets mode: +v ExtraCrispy 12:35 Ven`` joined 12:36 p6bannerbot sets mode: +v Ven`` 12:53 MasterDuke joined, p6bannerbot sets mode: +v MasterDuke 12:54 MasterDuke left, MasterDuke joined, herbert.freenode.net sets mode: +v MasterDuke, p6bannerbot sets mode: +v MasterDuke 13:13 brrt left
lizmat notable6: weekly 13:37
notable6 lizmat, 12 notes: gist.github.com/a35ea99cb820cc65da...f6da606761
lizmat notable6: weekly reset 13:38
notable6 lizmat, Moved existing notes to “weekly_2019-01-14T13:38:26Z”
nine jnthn: wanna see something odd? This apparently used to work in our production code: method !watch-for(@folders, $watch_for) { for @folders -> $folder { whenever $folder.IO.watch -> $e { ... 13:46
jnthn nine: In 6.d the `whenever` must be lexically within the `supply` or `react` block 13:49
nine Apparently also in 6.c
Though I can't find a golf for that. 13:57
WAT? The golf is: use v6.c; unit class Foo; use File::Find::Rule:from<Perl5>; method !foo() { whenever "foo" { } } 14:00
The use File::Find::Rule:from<Perl5> makes the difference
timotimo does it accidentally export use v6.d somehow?
does it trigger the "we need $_ for perl5 backcompat modules" behaviour? 14:01
14:01 Ven`` left 14:09 brrt joined 14:10 p6bannerbot sets mode: +v brrt 14:18 Ven`` joined 14:19 p6bannerbot sets mode: +v Ven`` 14:32 titsuki left 14:35 titsuki joined 14:36 p6bannerbot sets mode: +v titsuki
lizmat reportable6: 2019-01-07 2019-01-14 14:40
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
lizmat, gist.github.com/7292e558989e1b9749...ca60e94686 14:41
AlexDaniel wow, a new ticket on RT… :) 14:43
lizmat yeah...daxim's right ? 14:45
should be closable by tomorrow... unless something weird happens
And another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/01/14/...it-spring/ 15:08
moritz: ^^
jnthn m: use Test; my $p = Promise.new; start react { my $c = Channel.new; whenever $c { }; $c.close; for $c.list { }; $p.keep; }; await Promise.anyof($p, Promise.in(5)); ok $p, "No hang when we close and then drain a channel we have a subscription to" 15:14
camelia ok 1 - No hang when we close and then drain a channel we have a subscription to
jnthn Hmm 15:16
m: use Test; react { my $c = Channel.new; whenever $c { }; $c.close; for $c.list { }; }; pass "Did not hang"
camelia (timeout)
jnthn m: use Test; my $p = start react { my $c = Channel.new; whenever $c { }; $c.close; for $c.list { } }; await Promise.anyof($p, Promise.in(5)); ok $p, "No hang when we close and then drain a channel we have a subscription to" 15:17
camelia ok 1 - No hang when we close and then drain a channel we have a subscription to
jnthn m: use Test; for ^5 { my $p = start react { my $c = Channel.new; whenever $c { }; $c.close; for $c.list { } }; await Promise.anyof($p, Promise.in(5)); ok $p, "No hang when we close and then drain a channel we have a subscription to ($_)" } 15:18
camelia ok 1 - No hang when we close and then drain a channel we have a subscription to (0)
ok 2 - No hang when we close and then drain a channel we have a subscription to (1)
ok 3 - No hang when we close and then drain a channel we have a subscription to…
jnthn OK, guess I have to test if with the `pass "we ddin't hang"` approach...
It's odd in that this was golfed out of a $dayjob app that wedged a thread, but that was off in the pool 15:20
I already have a local patch to address the bug, but was hoping for a slightly better test than "just pass if it doesn't hang"
Geth roast: 1aaa01506e | (Jonathan Worthington)++ | S17-channel/subscription-drain-in-react.t
Cover a Channel close/drain/subscription hang
 15:31
rakudo: 9b77459b8d | (Jonathan Worthington)++ | 2 files
Fix hang in Channel subscribe/close/drain

When we polled a Channel, we forgot to re-insert the CLOSE or FAIL message. This meant that under certain use patterns, such as when we had a subscription to a Channel, closed it, and then tried to drain it with a `for` loop, we could get a hang, since the draining loop never saw the CLOSE message.		 15:33
15:35 Ven`` left
lizmat jnthn: does that fix stackoverflow.com/questions/540885...gram-hangs ? 15:35
jnthn Doubt it 15:39
I don't recall that code using a channel
That fix was for a bug I golfed out of a complex $dayjob app 15:40
lizmat ah, too bad :-) 15:42
jnthn StackOverflow reputation is nice, but $dayjob puts food on the table... :-) Will see if I can reply to that one at some point though. 15:43
I'm curious what's going on
lizmat++ # weekly 15:50
16:05 brrt left
|Tux| Rakudo version 2018.12-213-g9b77459b8 - MoarVM version 2018.12-29-g1ff55bf1c
csv-ip5xs0.929 - 0.932
csv-ip5xs-207.029 - 7.348
csv-parser22.263 - 23.366
csv-test-xs-200.435 - 0.437
test7.632 - 7.645
test-t1.831 - 1.910
test-t --race0.824 - 0.860
test-t-2029.836 - 31.242
test-t-20 --race9.570 - 9.664
 16:08
16:22 brrt joined 16:23 p6bannerbot sets mode: +v brrt
Geth rakudo: fe38bdba62 | (Elizabeth Mattijsen)++ | 6 files
Allow parameterization of Bag/BagHash
 16:36
17:02 brrt left
AlexDaniel lizmat++ 17:07
Geth rakudo: 1c101a04f9 | (Stefan Seifert)++ | 2 files
Fix Inline::Perl5 failing to load when installed through the Staging repository

Marking the %quote_lang_cache as neverrepossess was not enough. It gets serialized anyway and the MVM_CF_NEVER_REPOSSESS on the object doesn't survive serialization/deserialization. This caused references to CompUnit::Repository::Staging's serialization context to end up in the installed Inline::Perl5::Parent when they shouldn't have. And That's why we ended up with ... (6 more lines)		 17:13
nine Nice to get that issue off my chest :) 17:14
lizmat nine++
17:16 patrickb joined 17:17 p6bannerbot sets mode: +v patrickb
nine Maybe this would be the right time to have a look at in-process precompilation again. I got stuck on that due to some repossession issues and I didn't have the debug foo back then to get around them. Maybe I have it now :) 17:19
timotimo i wonder if we can do in-process precompilation, but have it in one separate process anyway, so that memory usage from precomp won't stick around for long-running processes 17:20
nine Well once we can precompile multiple comp units in the same process without running into weird dependency issues, we have all the options for tricks like that :) 17:22
timotimo nice 17:23
patrickb nine: Thanks for the poke. I definitely will have an eye on performance. Currently I suspect the performance penalty of the code I inserted will be minimal (it's basically looking up two environment variables and minimal string manipulation) and I'm mostly concerned about the coupling of an "external" runner script and the internals of rakudo which I'd like not to have. 18:00
18:01 robertle_ left, AlexDaniel left
patrickb nine: There is an environment variable used in RepositoryRegistry.pm6 "RAKUDO_PREFIX" to specify the rakudo installation path. Would you object to renaming that to PERL6_HOME? 18:05
It is currently documented here: docs.perl6.org/programs/03-environ...ompilation 18:06
Geth rakudo: bcc8054a4d | (Elizabeth Mattijsen)++ | 5 files
Allow parameterization of Mix/MixHash
 18:18
18:19 gfldex left 18:45 AlexDaniel joined, p6bannerbot sets mode: +v AlexDaniel 18:46 ChanServ sets mode: +o AlexDaniel, AlexDaniel sets mode: -m , ChanServ sets mode: -o AlexDaniel 18:47 remi_ness joined 18:48 p6bannerbot sets mode: +v remi_ness 18:54 remi_ness left
releasable6 Next release in ≈4 days and ≈23 hours. 6 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
19:22 brrt joined, p6bannerbot sets mode: +v brrt 19:25 remi_ness joined 19:26 p6bannerbot sets mode: +v remi_ness
patrickb I'm off o/ 20:00
20:01 patrickb left 20:05 remi_ness left 20:30 brrt left 20:59 brrt joined 21:00 p6bannerbot sets mode: +v brrt 21:04 d4l3k_ joined, ChanServ sets mode: +v d4l3k_ 21:05 p6bannerbot sets mode: +v d4l3k_ 21:11 d4l3k_ is now known as dalek 21:22 gfldex joined 21:23 p6bannerbot sets mode: +v gfldex 21:42 brrt left
lizmat hmmm.. looks like we lost the zofbots ? 22:05
AlexDaniel ^^
lizmat just committed github.com/rakudo/rakudo/commit/54...32faf2402b
Make X::Trait::Unknown a bit more resilient 22:06
22:28 ExtraCrispy left
AlexDaniel timotimo: Geth? 23:23
lizmat: actually, the only zofbots were buggable, huggable and ZofBot 23:24
the last one is not very useful to everyone else
huggable can be easily brought up, I'll do it next weekend maybe
and tyil is working on buggable
lizmat: what we're seeing right now is hack being down…
I think? 23:25
actually, hack is up! 23:26
and only Geth is down…
23:27 Geth joined, ChanServ sets mode: +v Geth
AlexDaniel lizmat: I fixed it \o/ 23:27
AlexDaniel followed the instructions on github.com/perl6/geth/ 23:28
23:28 p6bannerbot sets mode: +v Geth
AlexDaniel ah, btw by zofbots I meant “those hosted by Zoffix” 23:28
of course there are more bots
23:32 noname__ left
AlexDaniel oooooooohh I just realized something 23:38
admin privileges on private repos are not a thing, apparently?
wrong channel :)

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!