AlexDaniel [Tux]: soo… what's up with that spike? 00:10
reportable6: 2018-03-26T00:00:00Z 2018-04-02T00:00:00Z 01:54
reportable6 AlexDaniel, OK, working on it! This may take up to 40 seconds
AlexDaniel, gist.github.com/334dff3a00592c29e1...7271cfe502
Geth_ rakudo: MasterDuke17++ created pull request #1682:
Micro-optimize List.roll($n)
02:55
rakudo: c71ff6642d | (Daniel Green)++ | src/core/List.pm6
Micro optimize List.roll($n)

Lexicals are faster than private attributes.
02:58
rakudo: f1b9c7daba | (Aleks-Daniel Jakimenko-Aleksejev)++ (committed using GitHub Web editor) | src/core/List.pm6
Merge pull request #1682 from MasterDuke17/micro_optimize_Lists_roll

Micro-optimize List.roll($n)
nqp: dd513ea040 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 9 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...8-ged4201e ed4201e debugserver: cut module names out of filenames 27a20d2 Merge pull request #826 from MoarVM/spesh-refactor-iffy 00a8e84 [Spesh] fix negated spesh of known type e7f985a [Spesh] Add optimize_unbox ... (5 more lines)
04:03
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...8-ged4201e
rakudo: affd5befbe | (Zoffix Znet)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 3 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....7-gdd513ea dd513ea [MoarVM Bump] Brings 9 commits 32fd547 Profiler Template: Allocations modal close by btn 6b49767 [JVM] Fix nqp::takedispatcher in non-void context
... (10 more lines)
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....7-gdd513ea
232c664238 | usev6++ | src/core/Exception.pm6
AlexDaniel so… I guess I'll add some discouraging messages to release_guide.pod for those who want to follow the process manually 07:29
I'm idling on dev channels of other langs, and I do backlog a bit when I have some time to waste 07:30
apparently someone found a vim swap file in the released rust src tar 07:31
xD
Geth_ rakudo: 916b41a219 | (Aleks-Daniel Jakimenko-Aleksejev)++ | docs/release_guide.pod
Discourage manual releasing

Because it's prone to human error and this was demonstrated countless times (and we've been using tools for more than a year now).
This is the first step in improving the document to be more useful for those who rely on tools. More commits will follow.
08:00
nqp: e095adcfe4 | (Jeremy Studer)++ | src/vm/moar/profiler/template.html
Profiler Template: Sort by time in call graph

Order routines in call graph tab by time instead of percent. Fixes subtle misordering of routines.
08:01
AlexDaniel ( github.com/perl6/whateverable/issues/300 ) 08:09
[Tux] sh: line 1: 18453 Segmentation fault (core dumped) perl6 -Ilib test.pl < empty.csv 2> /dev/null 08:44
Rakudo version 2018.03-148-g916b41a21 - MoarVM version 2018.03-68-ged4201e92
csv-ip5xs0.886 - 0.900
csv-ip5xs-208.525 - 8.555
csv-parser36.790 - 37.435
csv-test-xs-200.449 - 0.455
test8.952 - 9.400
test-t2.448 - 2.469
test-t --race1.066 - 1.196
test-t-2043.102 - 44.153
test-t-20 --race15.068 - 15.931
08:54
pmurias m: use nqp;say nqp::typeparameterized(Str:U); 09:19
camelia Cannot perform this operation on an uninstantiated definite type
in block <unit> at <tmp> line 1
AlexDaniel [Tux]: what is this segfault? Can you rakudobug it? 09:36
[Tux] I was unable to reproduce
AlexDaniel [Tux]: did you try running it in a loop? 09:59
[Tux] no 10:00
AlexDaniel do it please
or rakudobug the steps
*especially* if you've never seen this segfault before 10:01
[Tux] ran it in a 110x loop: no core dump 10:02
AlexDaniel and also given that we'll have rakudo star based on the next release
lizmat Files=1238, Tests=76257, 316 wallclock secs (15.63 usr 5.59 sys + 2181.46 cusr 220.39 csys = 2423.07 CPU)
AlexDaniel then maybe it needs some CPU load? 10:03
[Tux]: please rakudobug anyway
maybe dogbert17++ will use his magic to pull the segfault out of it
[Tux] Rakudo version 2018.03-148-g916b41a21 - MoarVM version 2018.03-68-ged4201e92
csv-ip5xs0.898 - 0.961
csv-ip5xs-208.569 - 8.622
csv-parser36.310 - 36.368
csv-test-xs-200.471 - 0.481
test8.994 - 9.135
test-t2.400 - 2.424
test-t --race1.042 - 1.055
test-t-2042.898 - 42.912
test-t-20 --race14.994 - 15.403
10:27
I started all over, removing all .precomp etc, but no core dumps
Geth rakudo: 7327a33a88 | (Elizabeth Mattijsen)++ | src/core/List.pm6
Further micro-optimize List.roll($n)

  - make decrementing part of the condition
  - save the final todo value in the iterator
12:13
dogbert17 ZOFFLOP: t/spec/S17-supply/syntax.t, tests #66 and #83 when run under some stress 13:10
AlexDaniel . 13:13
m: sub s { }; my $x = ‘abc’; $x ~~ s/.//
camelia 5===SORRY!5=== Error while compiling <tmp>
Malformed postfix call
at <tmp>:1
------> 3sub s { }; my $x = ‘abc’; $x ~~ s/.7⏏5//
AlexDaniel … should it be this way?
dogbert17 ZOFFLOP: t/spec/S32-list/pick.t 13:26
test #299: MoarVM panic: Must not GC when in the specializer/JIT 13:27
timotimo that's without my branch? 13:31
dogbert17 yup
gdb says all :) #7 0xb7be89fc in fix_wval (inlinee=<optimized out>, to_fix=0xb508cf90, to_fix=0xb508cf90, inliner=<optimized out>, tc=<optimized out>) at src/spesh/inline.c:227 13:33
your branch would probably fix this error 13:34
timotimo yeah, that's really the only place that can happen
it's supposed to :)
but i seem to recall it makes something worse somehow?
dogbert17 but we had this valgrind problem ...
I'm on 32-bit atm, I can give your branch a try and see what happens 13:35
the valgrind problem pops up immediately 13:38
timotimo did you unroot things some more and got a proper line number? 13:41
dogbert17 I only unrooted the stuff you showed me yesterday 13:43
timotimo i got the right line and an idea what's up 13:44
dogbert17 yay
huggable csv
huggable: csv 13:45
huggable dogbert17, github.com/Tux/CSV (see `bench csv` for how to run bench)
dogbert17 huggable: bench csv
huggable dogbert17, export PATH=`pwd`/install/bin:$PATH; cd CSV; for i in $(seq 1 10000); do echo 'hello,","," ",world,"!"'; done > /tmp/hello.csv; time perl6 -Ilib -MText::CSV test-t.pl </tmp/hello.csv
timotimo dogbert17: try this 13:47
AFKBBL
(commit pushed to moarvm)
dogbert17 :)
timotimo maybe we'll merge it then 13:48
dogbert17 ok, gimme a sec :) 13:50
the commit comment looked promising :)
timotimo: good news, the valgrind problem disappeared 13:53
timotimo do we use ZOFFLOP to refer to everybody? 14:14
dogbert17 I do :) 14:20
it can be clever if someone want's to grep the logs for failing test stuff 14:21
timotimo: your fix worked, perhaps you should merge ... 14:22
AlexDaniel releasable6: status 14:28
releasable6 AlexDaniel, Next release in ≈19 days and ≈4 hours. 4 blockers. 0 out of 149 commits logged (⚠ 9 warnings)
AlexDaniel, Details: gist.github.com/7151b08e51b61bce53...7cf1498f58
AlexDaniel ZofBot: hey look ↑, shitty bot is not shitty anymore! 14:29
ZofBot AlexDaniel, And he said, “In there,” and he pointed and there was a big room with a glass window on the other side of the train station door, and then he said, “Now, are you sure you know what you’re doing?” And said, “Yes
AlexDaniel ZofBot: yes what?
ZofBot AlexDaniel, huh? there is no conditional there though?
pmurias jnthn: do you have any hints why different modules get their own unique Str:U *unless* they use Hash[Str:U]? 16:16
jnthn pmurias: Umm...no, that sounds odd :S 16:17
pmurias how should I nopaste multiple files? 16:21
jnthn gist can handle multiple files
pmurias gist.github.com/pmurias/533d8fe7c6...13e1146b48 16:25
jnthn: that behavior is a bug? 16:44
jnthn Looking
pmurias: hmm...I thought coercion types were done with the parametric extensions 16:45
And so it should be interned at module load time
lizmat notable6: weekly 17:15
notable6 lizmat, 1 note: 2018-03-29T20:49:14Z <AlexDaniel>: gist.github.com/niner/a959cb893820...392d4c16cb
AlexDaniel wait what's that 17:16
lizmat notable6: weekly: reset
notable6 lizmat, No notes for “weekly:”
AlexDaniel xD
lizmat notable6: weekly reset
notable6 lizmat, Moved existing notes to “weekly_2018-04-02T17:16:26Z”
lizmat AlexDaniel: it's a WIP of nine
AlexDaniel ah
lizmat I guess not ready for the Weekly yet
otherwise he would have blogged about it I think :-) 17:17
pmurias jnthn: they are, but it doesn't work for some reason: github.com/rakudo/rakudo/blob/mast...OW.nqp#L32
lizmat reportable6: 2018-03-26T00:00:00Z 2018-04-02T00:00:00Z
reportable6 lizmat, OK, working on it! This may take up to 40 seconds
AlexDaniel I'm FASTER!!! gist.github.com/334dff3a00592c29e1...7271cfe502 17:18
lizmat: by the way, there's also a monthly: gist.github.com/45a2df11156178a841...d77296f174
reportable6 lizmat, gist.github.com/e9eb66568eb247ab38...4ced8c314d
lizmat and another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2018/04/02/...rl-6-idea/ 17:41
dogbert17 m: say sprintf "%.15g", log(2) 19:39
camelia 0.693147180559945
dogbert17 m: say sprintf "%.16g", log(2)
camelia 0.6931471805599454
dogbert17 m: say sprintf "%.17g", log(2)
camelia 0.6931471805599453
dogbert17 m: say sprintf "%.18g", log(2)
camelia 0.6931471805599454
dogbert17 hmm
lizmat hmmm... modules.perl6.org gives me a 500 ?? invoking the Zoffix 21:47
:-)
Zoffix got invoked 21:54
lizmat whee! :-)
timotimo prepare_cached failed: no such column: me.repo_url 21:55
/home/modules.perl6.org/modules.perl6.org/bin/../lib/ModulesPerl6/Model/Dists.pm line 55
but surely zoffix is already on the case \o/
Zoffix Yeah, I don't get why the [NEWDB] trigger didn't do the new db thing :/ 21:56
timotimo maybe some kind of race condition that made it skip over the trigger? 21:57
Zoffix lizmat: it's back up 22:02
lizmat Zoffix++
Zoffix Looks like it's cause there's no logic to handle [NEWDB] trigger on next run for when "exit when already running" exits. Need to redo it as a lock in the update script, rather than in the perl build script 22:09
Geth rakudo: 266f4a227b | (Elizabeth Mattijsen)++ | src/core/Awaitable.pm6
Micro optimize the Awaitable role

Using private methods in roles still incurs a substantial runtime overhead. By making the private methods public with UPPER_CASE names, we make the following code 10% faster with 5% fewer allocations:
   await do for ^100000 { start { } }
22:17
Zoffix goes out to do some shopping 22:49
FWIW, I paused cron on modules and currently rebuilding DB on the side, to work on the latest commit (right now the active commit on the lives site is a few commits behind master)
Gonna swap it to live site and re-enable cron when I come back 22:50