[Coke] I love npm's module installer display 00:02
Zoffix I see. TODOing inside subtests basically is LTA. I couldn't even TODO inside a subtest in Perl 5. 00:09
I forget what I said a few days ago about fudging a subtest, but I may have been wrong. A subtest counts as 1 test, so fudging it should be fine ¯\_(ツ)_/¯ 00:10
Zoffix will attempt to make subtest fudgable soon
dalek ast: 346fa6c | (Zoffix Znet)++ | S19-command-line/repl.t:
[TODO FUDGE] REPL with -M with non-existant module exits fine

RT#128595		 00:11
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128595
dalek ast: eb77164 | (Zoffix Znet)++ | S19-command-line/repl.t:
Fix typo in test name
 00:13
Zoffix will attempt to make subtest fudgable soon 01:16
oops
dalek ast: a8afed7 | (Zoffix Znet)++ | S29-os/system.t:
[TODO FUDGE] run(:merge) on non-existent program does not crash

RT#128594		 01:22
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128594
Zoffix Perl 6 cake: pbs.twimg.com/media/CnMN6Q9WgAAxYSI.jpg:large 02:30
timotimo cakemilia 03:08
MasterDuke benchable: f583f22,110704d my $a = "a" x 2**16;for ^1000 {my $b = $a.chop($_)} 03:22
benchable f583f22=0.371098041534424 110704d=0.369231939315796
Zoffix benchable: 2015.06,2016.06 my $a = "a" x 2**16;for ^1000 {my $b = $a.chop($_)} 03:26
MasterDuke Zoffix: that was actually an error on my part, it wasn't supposed to show up here just yet 03:27
and we'll need AlexDaniel to run it
but that should be soon
AlexDaniel pretty soon, yeah
also note that I only went till 2015.10, so 2015.06 is not going to work (yet) :) 03:28
timotimo whoa, benchable 04:11
[Tux] This is Rakudo version 2016.06-239-g303e77f built on MoarVM version 2016.06-9-g8fc21d5 06:17
test 15.700
test-t 8.933
csv-parser 16.675
masak nice, less than 9 now 06:55
yoleaux2 10 Jul 2016 13:49Z <Zoffix> masak: should this be closed based on liz's comments or should there be more discussion? rt.perl.org/Ticket/Display.html?id...et-history
masak Zoffix: last I heard, TimToady++ was mulling over that one. dunno if something came out of that, though. 06:56
Zoffix: my use case for this is still alive and well: github.com/masak/nex/blob/62aaf422....pm#L7-L10 06:57
I don't know what the criterion is for keeping RT tickets open. my gut feeling tells me we can do better than the current status quo. but I also think lizmat++ makes a set of valid points. 07:00
timotimo what's "antipairs" again? 07:12
m: enum Foo <A B C D E>; say Foo::.antipairs.perl
camelia rakudo-moar 303e77: OUTPUT«(Foo::B => "B", Foo::A => "A", Foo::E => "E", Foo::D => "D", Foo::C => "C").Seq␤»
timotimo you want something different still, right?
travis-ci Rakudo build errored. Elizabeth Mattijsen 'Oops, forgot to add a type smiley' 09:40
travis-ci.org/rakudo/rakudo/builds/144281580 github.com/rakudo/rakudo/compare/7...1b934b2d61
Rakudo build errored. Elizabeth Mattijsen 'Array.BIND-POS 3.5x faster for non-reified elements 10:19
travis-ci.org/rakudo/rakudo/builds/144271600 github.com/rakudo/rakudo/compare/0...95de34e027
Rakudo build errored. Elizabeth Mattijsen 'Mention Array.AT-POS/ASSIGN-POS/BIND-POS improvements' 10:20
travis-ci.org/rakudo/rakudo/builds/144273212 github.com/rakudo/rakudo/compare/2...971f36cf78
lizmat good *, #perl6-dev 10:46
.tell Zoffix the tests you added to S29-os/system.t yesterday hang on OSX :-(
yoleaux2 lizmat: I'll pass your message to Zoffix.
lizmat which is what is causing the Travis fail messages 10:48
dalek ast: 4f52c6e | (Zoffix Znet)++ | S29-os/system.t:
Comment out hanging TODO fudge

The test hangs on OSX (per irclog.perlgeek.de/perl6-dev/2016-0...i_12832824 )
RT#128594		 10:52
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128594
Zoffix ACK 10:53
yoleaux2 10:46Z <lizmat> Zoffix: the tests you added to S29-os/system.t yesterday hang on OSX :-(
lizmat Zoffix++ # writing and commenting out for now :-) 10:54
m: use nqp; say nqp::sha1("abc") eq nqp::sha1("abc\0d") # disturbing
camelia rakudo-moar 303e77: OUTPUT«True␤»
jnthn lizmat: Wasn't a patch for that committed today? :) 10:59
lizmat yeah, but only in MoarVM afaik
jnthn Yeah
We only used it for hashing source code though, iirc
Which is relatively unlikely to contain a null :)
lizmat jnthn: it's also used in Blob.WHICH
jnthn Oh? 11:00
That sounds...fragile
And also unperformant :)
lizmat self.^name ~ '|' ~ nqp::sha1(self.join(","))
jnthn Or at least, relative to the quality of hashing we need on Blob.WHICH :)
wow, yeah, that's gonna...work but be slow :) 11:01
llfourn hopes nobody was using p6 to do pw management and allowing 0 in their passwords.
jnthn It's still relatively unlikely they'd have gone through a code path that hit the nqp::sha1 op :) 11:02
And the problem was restricted to that.
llfourn do we have SHA1 as a builtin? 11:03
(to Perl 6)
lizmat llfourn: only at NQP level, afaik
llfourn ok no worries then :) 11:04
jnthn Yeah, nqp::sha1 should not be used/exposed for general-purpose use 11:08
It only exists to hash source code files
brrt maybe call it nqp::hashcodefile then 11:15
just to obscure what it is, and to discourage people from relying on it
jnthn It may make it less tempting, yeah :) 11:16
And break anyone who did use it ;)
So we'll know about it :)
ooh, lunch time :) &
brrt lunch away 11:20
dalek kudo/nom: 6dbacb3 | lizmat++ | src/core/Array.pm:
Make Array.DELETE-POS about 7x faster

  - also do 6 allocations less per call
  - use setelems once instead of keeping popping
  - ready to handle Arrays with $!reified		 12:26
lizmat hmmm... s/with/without/ :-( 12:30
dalek kudo/nom: 66a5928 | lizmat++ | src/core/array_slice.pm:
Speedup %other check that I missed before
 12:33
lizmat afk&
[Coke] regarding "when do we close tickets" - RT #127789 is one of those tickets that needs to be decided by @larry; but I would say both masak and lizmat qualify there, and they disagree, so we need someone else to offer an opinion, I think. 12:34
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=127789
[Coke] BrokenRobot: github.com/coke/perl6-profiler-html now lets you round trip the old template (so I can start putting in a new template and running it with live data from a fresh profile run.) also added a gruntfile (not tied to the makefile yet) that lets me start manipulating the html/css/js 12:36
brrt lizmat: have you ever tested what the aggegrate effects of the 'Make $x about $n x faster' commits is? 13:02
also, why is nqp::if faster than just if?
BrokenRobot just if creates a block? 13:04
brrt hmmm
gfldex m: say do if ( 1 == 1) {};
camelia rakudo-moar 66a592: OUTPUT«Nil␤»
gfldex m: say do if ( 1 == 1) { 1 }; 13:05
camelia rakudo-moar 66a592: OUTPUT«1␤»
psch m: for ^10000 { $++ if 1 }; say now - BEGIN now; 13:20
camelia rakudo-moar 66a592: OUTPUT«0.0304687␤»
psch m: for ^10000 { if 1 { $++ } }; say now - BEGIN now;
camelia rakudo-moar 66a592: OUTPUT«0.0354006␤»
psch aalll noise vOv
|Tux| This is Rakudo version 2016.06-241-g66a5928 built on MoarVM version 2016.06-9-g8fc21d5 15:07
test 15.242
test-t 8.643
csv-parser 16.281
BrokenRobot New blog post "A Date With The Bug Queue or Let Me Help You Help Me Help You": perl6.party/post/A-Date-With-The-Bu...e-Help-You 15:09
perlpilot BrokenRobot++ 15:10
BrokenRobot: you could have just asked [Coke] about Lesson 1. I think he's gone through more bug reports than anybody. 15:11
BrokenRobot Challenge accepted! :) 15:41
Frameless bisect: " ".Int 16:25
bisectable Frameless: on both starting points the exit code is 0 and the output is identical as well
Frameless bisect: say " ".Int
bisectable Frameless: on both starting points the exit code is 0 and the output is identical as well
Frameless wtf
m: say " ".Int
camelia rakudo-moar 66a592: OUTPUT«0␤»
Frameless :o 16:26
On 2016.06-10-g7cff429 it crashes for the specific case of 1 space: gist.github.com/zoffixznet/41f4fd9...7b19265cf7 and there's a ticket reporting it: rt.perl.org/Ticket/Display.html?id=128543 16:27
bisect: " ".Int == 0 or die 16:28
bisectable Frameless: on both starting points the exit code is 0 and the output is identical as well
Frameless Broken bot
geekosaur less than ideal for this particular case, although if you know a failing version I think you can point to it?
timotimo that's right; check the irclog search for examples 16:29
geekosaur bisect: bad=7cff429 say " ".Int 16:30
Frameless I'm not following what you're suggesting... Right now it's searching from HEAD to some time in 2015 and it doesn't see the difference between June commits.
bisectable geekosaur: (2016-04-09) github.com/rakudo/rakudo/commit/576f975
Frameless :/ 16:31
Oh, ok, I see the problem. It was working then got broken then now it's working again
geekosaur yes. and running through every version to find one in which it fails would be rather expensive 16:32
timotimo yup, that's the difficulty with bisecting 16:33
now you can invert it and say that commit is "good" and we should find the commit that fixed it in the mean time ... or something like that
Frameless bisect: bad=7cff429,good=66a592 say " ".Int
bisectable Frameless: cannot find such “bad” revision
Frameless bisectable: help
bisectable Frameless: Like this: bisect: good=2015.12 bad=HEAD exit 1 if (^∞).grep({ last })[5] // 0 == 4 # RT 128181
Frameless bisect: bad=7cff429 good=HEAD say " ".Int 16:34
bisectable Frameless: “bisect run” failure
Frameless bisect: good=HEAD bad=7cff429 say " ".Int
bisectable Frameless: “bisect run” failure
Frameless It doesn't like good HEAD 16:35
Frameless snickers
bisect: good=7cff429 bad=HEAD my $x; try $x = " ".Int; say $x.defined and die 16:42
bisectable Frameless: exit code is 1 on both starting points, bisecting by using the output
Frameless: (2016-07-04) github.com/rakudo/rakudo/commit/e28305c
Frameless lool
So I did fix it after all :D
AlexDaniel poor bisectable was called broken :) 16:50
dalek ast: c10df83 | (Zoffix Znet)++ | S32-str/numeric.t:
.Int on string with nothing but spaces gives 0

RT#128543
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128543
AlexDaniel geekosaur: fwiw, it is much less expensive than most people think :) 16:54
in fact, it is so inexpensive that you'll probably see a bot for that in a few days… 16:56
Frameless github.com/rakudo/rakudo/compare/2......2016.06 17:01
m: say 1198 * 10 / 60
camelia rakudo-moar 66a592: OUTPUT«199.666667␤»
Frameless 3 hours for a piece of code that runs for 10 secs
Guess you could parallelize if you have a ton of cores. 17:02
AlexDaniel yeah, parallelizing is key 17:03
m: say 1500/(7)*0.5
camelia rakudo-moar 66a592: OUTPUT«107.142857␤»
AlexDaniel less than two minutes for any quick snippet 17:04
note that stuff typically runs in about 0.2-0.3 s on that server, so it will be even faster
Frameless *phew* Karen from TPF replied to my CLA email. My journey for sending legal paperwork is now complete :) "Thanks for sending it. I'm sorry about the problem with the address." 17:14
jnthn Frameless: Just heard from Karen about that. :) Commit invite issued. 17:33
Frameless "You are now a member of Rakudo Perl!" jnthn++ thanks :) 17:36
Frameless celebrates with the appropriate amount of fun
lizmat Frameless++ # welcome! 17:37
lizmat hopes to see many fine commits :-)
Frameless :)
timotimo yay 17:38
jnthn Indeed, welcome on board :)
Frameless \o/ 17:39
[Coke] tried to update github.com/rakudo/rakudo/wiki/contrib-Create-PR to reflect our current recommendations about PRs. Can anyone figure out why the code blocks aren't formatted like code? 17:43
also, feel free to edit it.
Also, rakudo currently has 42 PRs that need to be reviewed.
geekosaur that needs to be 4 spaces *in addition to* the indent for the itemized list 17:45
Frameless Using ``` seems to have done the trick as well
geekosaur so, I think, 6 or 7 spaces
[Coke] but the list is over at that point, no?
geekosaur that's a markdown corner case :/ 17:46
Frameless Oh number is now messed up
[Coke] Frameless++ geekosaur++
Frameless: we can just use 1) instead of 1. then
geekosaur has seen both behaviors, does not know which github uses
[Coke] I'm going to add "pending user" as a label for PRs so we can see which ones are waiting for the user to clean something up.
geekosaur and it may be doing something else because someone coded strictly to spec and this case falls into a hole 17:47
[Coke] jnthn: github.com/rakudo/rakudo/pull/232 is from 2014 and the last comment is that you're going to look at it locally. 17:49
jnthn [Coke]: In the end, fixed size arrays got done as part of the shaped arrays work (and yes, by me :P) 17:52
m: my @a[4]; @a[5] = 42
camelia rakudo-moar 66a592: OUTPUT«Index 5 for dimension 1 out of range (must be 0..3)␤ in block <unit> at <tmp> line 1␤␤»
jnthn [Coke]: So the PR can be closed 17:53
jnthn does so
[Coke] jnthn:^H++ 17:57
added "pending user" "has conflcits" and "needs review" labels. 17:59
timotimo "pending user" means we're waiting for the original reporter to react to a comment or so? 18:00
[Coke] yup 18:02
whereas needs review should mean a developer has to look at it. More labels are fine, just trying to get some light on these old PRs.
timotimo cool 18:03
[Coke] also tgging a few with failing travis as pending user, since I assume we're not merging anything that failed.
timotimo thanks, that seems like really helpful work 18:04
[Coke] might need to distinguish between "needs code review" and "needs spec review" 18:13
timotimo mhm, good idea
dalek kudo/nom: 04fbb41 | lizmat++ | src/core/Array.pm:
Remove one more use of ensure-allocated
 19:34
kudo/nom: ce4734f | lizmat++ | src/core/Str.pm:
Make Str.samemark about 50x faster

  - also fixes RT #128615
  - rewritten in nqp
  - made it match the documentation wrt to fewer pattern chars than string
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128615
timotimo 50x %) 19:38
Frameless :o 19:39
dalek kudo/nom: d59c1b8 | lizmat++ | docs/ChangeLog:
Mention Str.samemark in ChangeLog
 19:40
timotimo Frameless: did you have a different nickname recently? 19:41
TimToady yes, "MeowMix" :) 19:43
lizmat
.oO( sounds like "BrokenRobot" )
 19:44
Frameless timotimo: naaah :P 19:47
timotimo oh is it zoffix again? 19:48
lizmat no, it's Frameless :-) 19:51
geekosaur broken robot had a framing error? :p 19:52
dalek kudo/nom: 8bb4195 | lizmat++ | src/core/Hash.pm:
Handle Hash.Map properly
 19:55
lizmat
.oO( nothing that couldn't be Zoffixed )
gfldex why is it `Hash:U' and not `Hash:U:` ? 19:58
lizmat oops 19:59
gfldex: good catch 20:00
now why didn't the spectest catch that :-(
dalek kudo/nom: f5f545c | lizmat++ | src/core/Hash.pm:
Properly indicate invocant type, gfldex++
 20:08
jnthn Time to add another spectest... :) 20:27
dalek kudo/nom: e990823 | lizmat++ | src/core/Array.pm:
Make Array.List 5x as fast

  - also introduce the :view named, if you are sure the Array won't change
  - rewritten in nqp, does cow if no :view given
  - remove last use of ensure-allocated		 20:47
kudo/nom: 24cc914 | lizmat++ | docs/ChangeLog:
Mention Array.List improvement
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Make Array.DELETE-POS about 7x faster 20:59
travis-ci.org/rakudo/rakudo/builds/144442152 github.com/rakudo/rakudo/compare/3...bacb34525d
lizmat wow, that's a commit of 8+ hours ago 21:01
anyways, good night, #perl6-dev!
sortiz g'night lizmat! 21:03
dalek ast: 01728bd | (Zoffix Znet)++ | S32-hash/map.t:
.Map on Hash produces correct results
 21:36
Zoffix .tell [Coke] unsure if you have access to look into it, but it seems there's some email issues between RT and perl6-compiler list. For example lizmat replied on the mailing list to a bug report (to [email@hidden.address] but it did not show up on the ticket: RT#128615 Also, I've been commenting on a bunch of tickets with Bcc to perl6-compiler, but those don't get to mailing list 21:43
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128615
yoleaux2 Zoffix: I'll pass your message to [Coke].
dalek ast: 933588c | (Zoffix Znet)++ | S32-str/samemark.t:
.samemark with empty-string argument must throw

RT#128615		 23:05
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=128615
travis-ci Rakudo build passed. Elizabeth Mattijsen 'Speedup %other check that I missed before' 23:07
travis-ci.org/rakudo/rakudo/builds/144443432 github.com/rakudo/rakudo/compare/6...a5928ae101

Powered by Cro and the Raku® Programming Language.

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

Thank you!