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.
00:07 vrurg joined
Geth rakudo: vrurg++ created pull request #2997:
Fix for line() not using @*ARGS within MAIN
00:54
01:10 lucasb left 01:58 squashable6 left 01:59 squashable6 joined, ChanServ sets mode: +v squashable6 04:34 AlexDaniel left 05:49 squashable6 left 05:51 squashable6 joined 06:30 patrickb joined
patrickb o/ 06:35
Is there a tag for pull requests to signify the PR is not meant to be merged (for whatever reasons) and can be ignored? 06:37
Typical case: "It's a nice idea, so I put it here, maybe one day I'll find time to work on this more."
The reason I'd like to have such a tag is, because then processing all PRs is an achievable goal. A PR can then either be rejected, worked on to be merged, or tagged with such a tag. No more clogging of the PR queue. 06:40
Name suggestions: "not-for-merge", "paused" 06:43
06:50 ZzZombo_ joined 06:52 ZzZombo left, ZzZombo_ is now known as ZzZombo
releasable6 Next release in ≈2 days and ≈11 hours. 3 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 07:00
rba kawaii: ecosystem-api.p6c.org is now redirecting to ecosystem-api.p6c.org 08:07
.tell kawaii ecosystem-api.p6c.org is now redirecting to ecosystem-api.p6c.org
yoleaux rba: I'll pass your message to kawaii.
rba and gtk-dlls.p6c.org is redirecting to gtk-dlls.p6c.org 08:08
kawaii rba: perfect, thank you! 08:13
yoleaux 08:07Z <rba> kawaii: ecosystem-api.p6c.org is now redirecting to ecosystem-api.p6c.org
kawaii AlexDaniel`: Blin running! 08:16
rba kawaii: what is Blin? 08:25
kawaii rba: a tool for testing all of the modules in the ecosystem (github.com/perl6/Blin) 08:26
to help identify any breaking changes we've made to rakudo
rba ahh. got it.
08:38 [TuxCM] joined
AlexDaniel` kawaii: sounds promising! 08:42
lizmat patrickb: in some places people use a [WIP] (Work In Progress]
prefix in the name of the PR
patrickb lizmat: That could also be a good name. How does one establish a new tag? Merely by using it? 09:01
Is the person creating a PR always able to add and remove tags? 09:04
09:04 squashable6 left
[TuxCM] Rakudo version 2019.03.1-628-gf05b1b080 - MoarVM version 2019.05-55-g8c85b4a46
csv-ip5xs1.163 - 1.163
csv-ip5xs-2058.064 - 61.412
csv-parser22.508 - 22.638
csv-test-xs-200.432 - 0.446
test6.116 - 7.034
test-t1.732 - 1.749
test-t --race0.784 - 0.847
test-t-2028.285 - 28.767
test-t-20 --race8.958 - 9.209
09:05
09:08 squashable6 joined, ChanServ sets mode: +v squashable6
lizmat patrickb: I wasn't thinking about tags, but that could be a solution also I guess 09:19
I've just added the WIP label
kawaii usercontent.irccloud-cdn.com/file/...-40-45.png 09:41
So uhhhh... This module is interesting
github.com/FROGGS/p6-Games-BubbleBreaker
Every time I run Blin (via a TTY, not SSH) this appears
when testing that module, and makes it impossible to continue
cc: AlexDaniel`
timotimo oh, is it trying to open a window or something? 09:43
kawaii yeah looks like it 09:49
but it never exits and prevents the rest of the tests from running
I'm rerunning Blin now via SSH instead of a TTY to see what happens 09:51
timotimo well, you should be able to kill it with a signal 09:54
09:54 [TuxCM] left 10:02 [TuxCM] joined
AlexDaniel` There is a list of modules that are ignored 10:09
On top of blin.p6 I believe
So just add it there, and open a ticket for the module saying that the tests are insane 10:10
10:13 [TuxCM] left
AlexDaniel` Also blin should be able to kill it after some time by itself, but your terminal will probably not return to the previous state 10:13
10:21 [TuxCM] joined 10:28 pmurias joined
pmurias nine: I got my CompUnit::Repository to export the thing returned from require 10:30
nine: finding the node_modules to look up things from is the next obstacle 10:31
ahh, I figured something out 10:32
10:34 ggoebel joined 10:41 pmurias left 10:53 pmurias joined 10:54 ggoebel left
nine pmurias: sounds good :) 10:54
10:55 [TuxCM] left 10:59 [TuxCM] joined 11:19 [TuxCM] left 11:37 [TuxCM] joined 11:39 pmurias left 11:46 [TuxCM] left 11:50 [TuxCM] joined 11:51 vrurg left 12:04 [TuxCM] left 12:05 vrurg joined 12:14 vrurg_ joined 12:15 vrurg left, vrurg_ is now known as vrurg 12:26 ilogger2 joined 12:31 Tux__ joined 12:36 ggoebel joined 12:39 vrurg joined, vrurg left
kawaii www.irccloud.com/pastebin/s2eoCcCV/ 12:45
AlexDaniel`: seen this before?
AlexDaniel` what the hell 12:47
kawaii: did it come from Blin itself?
or is it simply the output of some module that somehow got through?
kawaii looks like it, it killed Blin
AlexDaniel` This is not a good sign, I have never seen that
kawaii cc: jnthn ?
AlexDaniel` daaaaamn 12:48
kawaii: so try to run it again, I guess?
kawaii Yep just triggered another run
AlexDaniel` hopefully we'll see the same issue in one of the modules
but this is truly not a good sign, I have never seen Blin crash like this
12:49 vrurg joined
AlexDaniel` but! What is error code -11 ? 12:49
could it be like a process count limit or something?
timotimo: what's -11 ? 12:50
12:51 [TuxCM] joined 12:52 Tux__ left
kawaii AlexDaniel`: I'm assuming this is a more generic issue as opposed to an issue with a specific module since I went ahead and ran manual tests for all the modules in that snippet and they all passed with no regressions 12:56
AlexDaniel` kawaii: I'm trying to figure out what 11 is 12:57
kawaii: it can be something like out of memory
this is the right resource, right? docs.libuv.org/en/v1.x/errors.html
12:59 [TuxCM] left 13:00 [TuxCM] joined
vrurg Is it possible during compile time to install a stub which would be later turned into role/class/constant by require? Trying to figure out how to deal with R#2983 13:01
synopsebot R#2983 [open]: github.com/rakudo/rakudo/issues/2983 [regression] Cannot mix in non-composable type Any into object
ugexe 11 is EAGAIN
AlexDaniel` github.com/MoarVM/MoarVM/issues/1123
ugexe so lacked the resources to create another thread
kawaii ah fair enough 13:02
will give them VM some more to play with
AlexDaniel` kawaii: what's your cat /proc/sys/kernel/pid_max 13:03
ah no, that's the wrong thing…
maybe /proc/sys/kernel/threads-max ?
13:03 lucasb joined
kawaii AlexDaniel`: 31341 13:03
AlexDaniel` on some systems it is surprisingly low, for some reason 13:04
ugexe++ :) 13:07
ugexe: but I wonder, why wasn't it UV_ENOMEM? 13:08
or UV_EAI_MEMORY, I don't know the difference… 13:09
kawaii I hope my laptop doesn't set on fire from this, starting to get very hot from all this toasting :D
AlexDaniel` kawaii: I thought you had some beefy machine for it? :) 13:10
kawaii: how long does it take on your laptop?
kawaii Turns out the work-supplied server is way slower than my laptop
AlexDaniel` kawaii: also perhaps make sure that you have the latest rakudo for running blin itself, because IIRC json speedups are not in any release yet 13:11
so all the zef stuff is probably much faster now, which should help Blin a lot
kawaii I estimate that the work server would take about 8 hours for a full ecosystem test, seems to be about 3 on my laptop if it actually finishes without error :)
AlexDaniel` interesting 13:12
on a 24 core google compute instance it was just an hour, so your numbers make sense I think
kawaii yeah my laptop is 8 cores so that makes sense 13:16
13:16 pmurias joined 13:18 [TuxCM] left 13:21 [TuxCM] joined
ugexe AlexDaniel`: because pthread_create doesn't return those error codes 13:26
13:50 [TuxCM] left 13:51 [TuxCM] joined 13:52 vrurg left 14:01 [TuxCM] left 14:04 [TuxCM] joined, skids joined
Guest77572 interesting, one of [Tux] test files makes valgrind a bit angry 14:07
14:19 Tux__ joined 14:20 [TuxCM] left 14:27 Tux__ left 14:31 vrurg joined 14:35 [TuxCM] joined 14:37 vrurg_ joined, vrurg left, vrurg_ is now known as vrurg 14:41 vrurg_ joined 14:42 vrurg left, vrurg_ is now known as vrurg
Guest77572 there's something odd with the rakudo build methinks 14:44
if I want to run e.g. './perl6 t/spec/S04-phasers/in-loop.t' that works find, it automatically finds the Test dependency 14:45
but if I use perl6-valgrind-m or perl6-gdb-m it no lionger works unless I add '-Ilib' 14:46
*longer
14:48 AlexDaniel joined
AlexDaniel . 14:49
15:10 pmurias left 15:11 [TuxCM] left 15:30 vrurg left 15:31 vrurg joined 16:02 vrurg left 16:04 patrickb joined 16:48 [TuxCM] joined 16:59 vrurg joined 17:39 vrurg left 17:44 vrurg joined 17:47 vrurg_ joined 17:48 vrurg left, vrurg_ is now known as vrurg 18:16 vrurg left 18:22 vrurg joined
Geth ¦ problem-solving: AlexDaniel assigned to jnthn Issue require with non-sigiled arguments github.com/perl6/problem-solving/issues/43 18:43
patrickb releasable6: status 18:47
releasable6 patrickb, Next release in ≈2 days and ≈0 hours. 3 blockers. 221 out of 628 commits logged (⚠ 1 warnings) 18:48
patrickb, Details: gist.github.com/ef189b08fee13ccebc...884c3daf57
AlexDaniel kawaii: did it finish? :) 18:58
kawaii www.irccloud.com/pastebin/8FL3e2J8/ 19:00
AlexDaniel: ^
Yep, this is what I have
AlexDaniel kawaii: so it finished, right? 19:01
looks pretty good, you can just drag these files into a github gist :)
19:39 vrurg left 19:45 vrurg joined 19:56 pmurias joined 20:35 skids left 20:59 vrurg left, vrurg joined 21:03 lucasb left 21:24 vrurg left 21:35 vrurg joined 21:40 vrurg left
patrickb Phew. Finally ChangeLoged all my commits. 21:45
I wonder where all of the build-revamp commit logs are. They are neither reported by releaseable, nor listed in the ChangeLog Draft. 21:46
21:48 vrurg joined
patrickb releasable6: status 21:48
releasable6 patrickb, Next release in ≈1 day and ≈21 hours. 3 blockers. 257 out of 628 commits logged (⚠ 3 warnings)
patrickb, Details: gist.github.com/6e96abfa676f07482d...53652336d6
timotimo releasable doesn't seem to look into submodules
patrickb But all the commits to rakudo directly are also missing. e.g. github.com/rakudo/rakudo/commit/6e...2a6222c951 21:49
AlexDaniel interesting
21:49 vrurg left
AlexDaniel patrickb: doesn't seem to be missing though? 21:49
patrickb: releasable does complain that it is not logged 21:50
patrickb Searching for "6ee37" doesn't find anything when searching in the ChangeLog Draft and the releasable output. 21:51
AlexDaniel: Where do you see it?
AlexDaniel patrickb: screenshots.firefox.com/J7hA4Cuoq5...github.com
patrickb Ouch. I was obviously viewing a stale releasable gist... 21:53
AlexDaniel: Thanks for the screenshot!
22:11 pmurias left
MasterDuke timotimo: interesting, a random profile i just created shows <anon> gen/moar/stage2/NQPHLL.nqp:1825 as the top by inclusive. it's github.com/perl6/nqp/blob/master/s...#L195-L196 22:16
timotimo huh 22:18
that doesn't seem right
you don't have any confprog or so in there, right?
MasterDuke nope
timotimo did you just grab my moarvm changes? 22:19
MasterDuke yeah. but i'd be inclined to suspect my remove-profiler-overhead commits
oh, i think i have some profiles from before your commits, but after mine, that show the same thing 22:21
timotimo damn
does the value for "inclusive" there look bogus in some way?
MasterDuke no
oh 22:23
hm. it's greater than the total time reported in the overview tab 22:24
but less than `time` reports 22:25
timotimo `time` might include time from other threads
MasterDuke oops, i meant also greater than `time` reports 22:26
`time` said: real 0m2.174s user 0m2.336s sys 0m0.040s 22:27
the profile's total said: 1950.57ms
the top inclusive time was 3397.86ms (which had 1696.99ms exclusive) 22:28
the next inclusive was the usual `<unit-outer> -e:1`, with 1696.95ms inclusive 22:29
22:30 vrurg joined
MasterDuke well, time to sleep, hopefully will be able to investigate further tomorrow 22:31
did also get one instance where the top exclusive time in a profile was 18446744073709.54ms, but re-profiling the same code hasn't reproed it yet 22:32
AlexDaniel kawaii: so? :) 22:57
23:28 patrickb left