Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2016-02-06

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:05 Laurent__ joined #perl6
00:11 revhippie left #perl6
00:15 kivutar joined #perl6
00:23 travis-ci joined #perl6
00:23 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Streamline phaser handling
00:23 travis-ci https://travis-ci.org/rakudo/rakudo/builds/107363200 https://github.com/rakudo/rakudo/compare/91a5d11cb0a6...ec543e9d2cb5
00:23 travis-ci left #perl6
00:24 autark left #perl6
00:28 AlexDaniel ouch
00:29 timotimo Stage parse      : Label with no handler passed to newlexotic
00:29 timotimo huh
00:29 timotimo very strange
00:30 timotimo only on linux, only with moar master
00:37 vendethiel joined #perl6
00:40 MadcapJake thanks to skaji_++, evalbot on Slack can now handle multiline code blocks!
00:41 timotimo cool
00:42 AlexDaniel m: say 42;␤say 69 # by the way
00:42 camelia rakudo-moar ec543e: OUTPUT«42␤69␤»
00:48 MadcapJake timotimo, also skaji_ just pushed commit that allows camelia in any channel public/private! i'm building it right now
00:48 timotimo excellent
00:51 revhippie joined #perl6
00:51 Skarsnik MadcapJake, it use IRC for slack?
00:53 MadcapJake you can connect to slack via irc now
00:53 MadcapJake but no bridge yet
00:53 MadcapJake haven't had time to build it
00:55 addison_ joined #perl6
00:55 Skarsnik I mean camelia use the IRC bridge?
00:56 MadcapJake no, skaji_++ built this awesome docker container for camelia with some slack tweaks and i'm running that (currently on my machine)
01:00 Skarsnik well docker does not make camelia connect to slack xD
01:06 yeahnoob joined #perl6
01:06 AlexDaniel jvm: (;)
01:06 AlexDaniel j: (;)
01:06 camelia rakudo-jvm f99bdb: OUTPUT«===SORRY!===␤Method 'returns' not found for invocant of class 'BOOTInt'␤»
01:07 MadcapJake Skarsnik, right, custom code does :P the container is just an easy platform for deploying it
01:14 vendethiel joined #perl6
01:34 colomon joined #perl6
01:35 japhb *yawn*  Another week of $dayjob complete ... and now to enjoy the weekend.  :-)
01:36 AlexDaniel japhb: what's your plan?
01:36 japhb ... after, perhaps, a nap.
01:36 japhb AlexDaniel: I'm hoping to spend a little time hacking on my RPG GM assistance code.
01:37 japhb Been a while since I had a chance to work on that.
01:37 Hotkeys ooo
01:37 japhb You?
01:37 AlexDaniel japhb: oh well! I hoped that you have nothing to do so that you can fix hyper in rakudo :D
01:37 Timbus joined #perl6
01:38 * japhb chuckles
01:38 AlexDaniel japhb: otherwise I'd have to write ugly workarounds that run the same code multiple times until it actually works… my plan is not to write these workarounds :)
01:38 japhb :-)
01:38 japhb What's broken, btw?
01:39 japhb (As in, what's the broken behavior?)
01:40 AlexDaniel japhb: mainly this: https://rt.perl.org/Public/Bug/Display.html?id=127452
01:40 AlexDaniel m: dd @(^10 .hyper(:1batch).map: { sleep rand; $_ })
01:40 camelia rakudo-moar ec543e: OUTPUT«(0, 1, 2, 3, 4, 5, 6, 7, 8, 9)␤»
01:40 AlexDaniel m: dd @(^10 .hyper(:1batch).map: { sleep rand; $_ })
01:40 camelia rakudo-moar ec543e: OUTPUT«(0, 1, 2, 3, 4, 5, 6, 7, 8, 9)␤»
01:40 AlexDaniel m: dd @(^10 .hyper(:1batch).map: { sleep rand; $_ })
01:40 camelia rakudo-moar ec543e: OUTPUT«()␤»
01:41 japhb Woah, that's bad
01:42 AlexDaniel I also don't want to get rid of hyper in my code because someone smart will fix it tomorrow :)
01:42 AlexDaniel … or maybe not…
01:43 japhb Is it broken in 2016.01.1 or head or both?
01:44 japhb (I guess camelia confirms it's broken in head ...)
01:44 AlexDaniel japhb: before 2015.01 there was no such bug, but there was another one: https://rt.perl.org/Public/Bug/Display.html?id=127099
01:45 japhb I'm guessing by that being open, it's still a problem.  :-(
01:46 AlexDaniel japhb: yes, except that behavior changed
01:46 AlexDaniel m: my @result = ^1000 .hyper.map: * + 10; say @result ~~ @result.sort
01:46 camelia rakudo-moar ec543e: OUTPUT«True␤»
01:47 AlexDaniel ↑ this kinda works
01:47 AlexDaniel m: .say for (^10).hyper(:1batch).map: { sleep rand; $_ }
01:47 camelia rakudo-moar ec543e: OUTPUT«1␤»
01:47 AlexDaniel m: .say for (^10).hyper(:1batch).map: { sleep rand; $_ }
01:47 camelia rakudo-moar ec543e: OUTPUT«1␤»
01:47 AlexDaniel m: .say for (^10).hyper(:1batch).map: { sleep rand; $_ }
01:47 camelia rakudo-moar ec543e: OUTPUT«3␤»
01:47 AlexDaniel come on…
01:47 AlexDaniel m: .say for (^10).hyper(:1batch).map: { sleep rand; $_ }
01:47 camelia rakudo-moar ec543e: OUTPUT«0␤2␤3␤4␤1␤6␤5␤8␤7␤9␤»
01:47 AlexDaniel japhb: ↑ here
01:48 japhb Hmph
01:48 * AlexDaniel cries in the corner
01:49 japhb Yeah, I don't blame you.  Have you created spectests for these?
01:50 AlexDaniel japhb: oh, that's actually something I can do
01:54 AlexDaniel japhb: maybe that's something I'll do during this weekend :)
01:55 japhb Good idea!
01:57 ChristopherBotto joined #perl6
01:57 ChristopherBotto .seen TimToady
01:57 yoleaux I saw TimToady 5 Feb 2016 20:56Z in #perl6: <TimToady> someone could point out that p6 behaves properly with https://stackoverflow.com/questions/11695110/why-is-this-program-valid-i-was-trying-to-create-a-syntax-error
01:59 ChristopherBotto TimToady: I hope you are feeling better. I heard you were sick a while ago and the last several times I've checked the IRC logs you haven't been around
02:00 ChristopherBotto .tell TimToady: I hope you are feeling better. I heard you were sick a while ago and the last several times I've checked the IRC logs you haven't been around
02:00 yoleaux ChristopherBotto: What kind of a name is "TimToady:"?!
02:01 ChristopherBotto .tell TimToady I hope you are feeling better. I heard you were sick a while ago and the last several times I've checked the IRC logs you haven't been around
02:01 yoleaux ChristopherBotto: I'll pass your message to TimToady.
02:22 kid51_ joined #perl6
02:24 Timbus joined #perl6
02:53 kaare_ joined #perl6
03:02 Timbus joined #perl6
03:04 ugexe i think a recent commit broke nom on windows
03:04 ugexe https://ci.appveyor.com/project/ugexe/zef/build/1.0.66#L944
03:05 ugexe Label with no handler passed to newlexotic (during Stage parse)
03:07 ugexe previous working build is from 21 hours ago. so one of the commits from today
03:08 Hotkeys hmm
03:08 ugexe lizmat: ^
03:08 xpen joined #perl6
03:29 noganex joined #perl6
03:33 z8 joined #perl6
03:34 timotimo liz' most recent commit might have caused that, yeah
03:41 timotimo m: Syntax error!
03:41 camelia rakudo-moar ec543e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Y4zLJWGgwX␤Bogus postfix␤at /tmp/Y4zLJWGgwX:1␤------> 3Syntax error7⏏5!␤    expecting any of:␤        infix␤        infix stopper␤        postfix␤        statement end␤        statement …»
03:44 geekosaur perl 6 is so not into poetry mode >.>
03:44 timotimo m: say Perl, Cool with any Version
03:44 camelia rakudo-moar ec543e: OUTPUT«(Perl)(Cool)␤»
03:54 Hotkeys heh
03:55 Hotkeys timotimo: clever
03:56 timotimo i did not invent this
03:59 Hotkeys still clever
04:00 Hotkeys m: say Perl, Cool for any 1
04:00 camelia rakudo-moar ec543e: OUTPUT«(Perl)(Cool)␤»
04:02 Hotkeys m: Perl, Cool, any time with any 1
04:02 camelia rakudo-moar ec543e: ( no output )
04:02 Hotkeys m: say Perl, Cool, any time with any 1
04:02 camelia rakudo-moar ec543e: OUTPUT«(Perl)(Cool)any(1454731335)␤»
04:02 Herby_ joined #perl6
04:02 Herby_ Evening, everyone!
04:03 Herby_ \o
04:03 Herby_ m: say 1 .. 3
04:03 camelia rakudo-moar ec543e: OUTPUT«1..3␤»
04:03 Hotkeys hola
04:04 Herby_ How goes it?
04:05 Hotkeys m: say Perl, now with Hash
04:05 camelia rakudo-moar ec543e: ( no output )
04:05 Hotkeys kek
04:05 Hotkeys Herby_: playing with sentences in perl6
04:06 MadcapJake can you name subs with unicode?
04:06 Hotkeys I believe so
04:06 MadcapJake is there a special way to do it?
04:06 Hotkeys hmm
04:08 Hotkeys m: sub ❄ { say "$^a snowflakes }; ❄ 3;
04:08 camelia rakudo-moar ec543e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/iE_ZOzjKip␤Missing block␤at /tmp/iE_ZOzjKip:1␤------> 3sub7⏏5 ❄ { say "$^a snowflakes }; ❄ 3;␤    expecting any of:␤        new name to be defined␤»
04:08 MadcapJake yeah that's what i keep getting
04:08 Hotkeys m: sub term:<❄> { say "$^a snowflakes }; ❄ 3;
04:08 camelia rakudo-moar ec543e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/oaz1zPF1TY␤Unable to parse expression in double quotes; couldn't find final '"' ␤at /tmp/oaz1zPF1TY:1␤------> 3b term:<❄> { say "$^a snowflakes }; ❄ 3;7⏏5<EOL>␤    expecting any of:␤        …»
04:08 Hotkeys m: sub term:<❄> { say "$^a snowflakes" }; ❄ 3;
04:08 camelia rakudo-moar ec543e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ifosWmY562␤Two terms in a row␤at /tmp/ifosWmY562:1␤------> 3sub term:<❄> { say "$^a snowflakes" }; ❄7⏏5 3;␤    expecting any of:␤        infix␤        infix stopper␤        statement end…»
04:08 Hotkeys m: sub term:<❄> { say "$^a snowflakes" }; ❄(3);
04:08 camelia rakudo-moar ec543e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/vbKKzOesOL␤Calling term:<❄>() will never work with declared signature ($a)␤at /tmp/vbKKzOesOL:1␤------> 3sub term:<❄> { say "$^a snowflakes" }; 7⏏5❄(3);␤»
04:08 Hotkeys m: sub term:<❄> ($a) { say "$a snowflakes" }; ❄(3);
04:08 camelia rakudo-moar ec543e: OUTPUT«5===SORRY!5=== Error while compiling /tmp/KNSRFbKENJ␤Calling term:<❄>() will never work with declared signature ($a)␤at /tmp/KNSRFbKENJ:1␤------> 3 term:<❄> ($a) { say "$a snowflakes" }; 7⏏5❄(3);␤»
04:10 geekosaur m: sub prefix:<❄> ($a) { say "$a snowflakes" }; ❄(3);
04:10 camelia rakudo-moar ec543e: OUTPUT«3 snowflakes␤»
04:10 Hotkeys interesting
04:10 Hotkeys I guess that works
04:10 geekosaur I would expect term to be a binary operator
04:10 Hotkeys ah
04:10 MadcapJake i want one that could take multiple arguments though, not necessarily a prefix/infix
04:14 AlexDaniel m: my \term:<❄> = -> $a, $b, $c { say “$a, $b, $c snowflakes” }; ❄(4, 8, 15)
04:14 camelia rakudo-moar ec543e: OUTPUT«4, 8, 15 snowflakes␤»
04:14 AlexDaniel not exactly what you want, but…
04:15 MadcapJake interesting!
04:15 geekosaur hmmm, so actually infix is supposed to be the binary op
04:20 MadcapJake this isn't really gonna work how i want it to
04:20 AlexDaniel MadcapJake: honestly I don't think that using unicode sub names is a good idea
04:21 geekosaur hm
04:21 AlexDaniel MadcapJake: just measure the startup time…
04:21 MadcapJake hehe, fiiiine :P
04:21 geekosaur think I found a lie in S02
04:21 MadcapJake oh rly? didn't know it impacted that
04:21 AlexDaniel MadcapJake: it does, because basically you are changing the grammar
04:21 geekosaur tried to use statement_prefix:<❄> and it said there isn't such
04:21 MadcapJake oh ok, good point!
04:22 * MadcapJake thinks adulting is hard
04:22 geekosaur oh, do we still have that bug?
04:22 geekosaur (unicode was making grammar parse time explode)
04:22 geekosaur (I think there's still some hacks in the grammar to make various unicodey things not trigger it)
04:23 AlexDaniel no, the problem is not with unicode
04:23 AlexDaniel you can use term:<ascii!> and it will be as slow
04:24 geekosaur oh, guess I just can't use those that way then. (I know statement_control is there...)
04:24 AlexDaniel MadcapJake: +2s startup time here just for doing that… I don't think that it's worth it
04:25 AlexDaniel MadcapJake: perhaps this delay is not important if you're writing a module
04:26 AlexDaniel I think that precompilation will make it go away, but still…
04:28 MadcapJake well technically this is a module, but it's about measuring performance so probably not a good place to go all unicode crazy
04:28 Hotkeys lol
04:28 Hotkeys what unicode were you going to use?
04:29 MadcapJake ????
04:29 AlexDaniel what is that
04:29 Hotkeys I don't have a font that shows that
04:29 Hotkeys it's a microscope
04:29 MadcapJake lol
04:29 timotimo i don't see it
04:29 AlexDaniel .u ????
04:29 yoleaux U+1F52C MICROSCOPE [So] (????)
04:29 Hotkeys .u ????
04:29 yoleaux U+1F52C MICROSCOPE [So] (????)
04:30 AlexDaniel oh, it's a microscope
04:30 Hotkeys lol
04:30 timotimo ah, a microscope
04:30 MadcapJake it's so tiny
04:30 AlexDaniel geez, who would have thought…
04:30 MadcapJake you need a microscope just to see what it is ;)
04:30 Hotkeys I think it's good to not use very uncommon symbols :p
04:31 MadcapJake looks better in my code editor, I use size 15 font though
04:31 Hotkeys atom doesn't show it for me
04:31 Hotkeys might be my lack of a good font though
04:31 MadcapJake hm strange, it's your font prolly
04:31 MadcapJake I use Fira Code
04:32 MadcapJake (I designed language-perl6fe to work really well with all the ligatures that Fira Code has)
04:33 Hotkeys this font is nice
04:33 * MadcapJake back to boring old text based subs ;)
04:33 vendethiel joined #perl6
04:37 AlexDaniel m: sub ᖱᖲ { say ‘Hello world…’ }; ᖱᖲ
04:37 camelia rakudo-moar ec543e: OUTPUT«Hello world…␤»
04:38 MadcapJake how'd that work?
04:38 AlexDaniel MadcapJake: not necessarily
04:38 AlexDaniel MadcapJake: because these are letters
04:38 MadcapJake ahhh xD
04:38 AlexDaniel so any character with L* property should work
04:43 AlexDaniel m: sub ᖱ́ᖲ̀ { say ‘Hello world…’ }; ᖱ́ᖲ̀
04:43 camelia rakudo-moar ec543e: OUTPUT«Hello world…␤»
04:43 Hotkeys MadcapJake: thank you for this font
04:43 Hotkeys it is wonderful
04:44 MadcapJake Hotkeys, no prob! It's definitely my favorite programming font.
04:45 AlexDaniel MadcapJake: so you can even use a bunch of combining characters together with weird letters :)
04:45 MadcapJake AlexDaniel, haha wow!
04:58 Hotkeys MadcapJake: Can you add support for the hex formatting of fira?
04:58 Hotkeys it's very minor but it would look nice
04:58 AlexDaniel m: my @x = start { sleep 1; 42 }; @x.push: start { sleep 1; 69 }; say @x».result
04:58 camelia rakudo-moar ec543e: OUTPUT«[42 69]␤»
04:58 AlexDaniel m: my @x = start { sleep 1; 42 }; @x.push: start { sleep 1; 69 }; say @x».result
04:58 camelia rakudo-moar ec543e: OUTPUT«[42 69]␤»
04:58 AlexDaniel m: my @x = start { sleep 1; 42 }; @x.push: start { sleep 1; 69 }; say @x».result
04:58 camelia rakudo-moar ec543e: OUTPUT«[42 69]␤»
04:58 Hotkeys eg for 0xF
04:58 AlexDaniel come on…
04:59 Hotkeys ?
05:00 AlexDaniel it should break…
05:00 ZoffixWin m: say (start { sleep 1; 69 }).result
05:00 camelia rakudo-moar ec543e: OUTPUT«69␤»
05:00 Hotkeys why?
05:00 Hotkeys result waits
05:01 Hotkeys don't see why it would break
05:01 * ZoffixWin neither
05:01 AlexDaniel sure, but it does on my computer
05:01 Hotkeys oh
05:01 ZoffixWin Oh
05:01 AlexDaniel Cannot invoke this object (REPR: Null)
05:01 Hotkeys lol
05:01 ZoffixWin Works on my box
05:01 AlexDaniel with about 50% chance
05:01 Hotkeys works for me
05:01 AlexDaniel ZoffixWin: try it 10 times
05:02 Hotkeys AlexDaniel: http://i.imgur.com/YAS7pE1.png
05:02 AlexDaniel I can't reproduce it on camelia…
05:02 AlexDaniel Hotkeys: https://files.progarm.org/2016-02-06-070244_243x165_scrot.png
05:02 ZoffixWin AlexDaniel, worked 20 times: https://gist.github.com/zoffixznet/fce06ca21262d22924cb
05:03 vatzz61 joined #perl6
05:03 AlexDaniel geez, but there *is* a problem somewhere
05:03 Hotkeys Which version are you on AlexDaniel
05:03 ZoffixWin I'm on This is Rakudo version 2015.12-214-g775271d built on MoarVM version 2015.12-29-g8079ca5
05:03 AlexDaniel my rakudo is like 1 or 2 days old
05:03 AlexDaniel This is Rakudo version 2015.12-232-ga5fe347 built on MoarVM version 2015.12-29-g8079ca5
05:03 Hotkeys myself: This is Rakudo version 2015.12-232-ga5fe347 built on MoarVM version 2015.12-35-g940850e
05:04 AlexDaniel ha, look, almost
05:04 Hotkeys lol
05:04 * ZoffixWin builds a fresh one
05:04 Hotkeys I've heard the current version is breaking on windows
05:04 Hotkeys so I won
05:04 Hotkeys won't :p
05:04 AlexDaniel Hotkeys: not only…
05:05 Hotkeys breaking everywhere?
05:05 AlexDaniel yes, it's probably everywhere
05:05 Hotkeys ah
05:06 AlexDaniel ZoffixWin: I can't reproduce it on camelia… :/
05:06 ZoffixWin Cool, then the bug is fixed! :p
05:08 Hotkeys \o/
05:10 AlexDaniel ok guys can you try this?
05:10 AlexDaniel m: my @x; @x.push: start { sleep 1; 1 } for ^100; say @x».result;
05:10 camelia rakudo-moar ec543e: OUTPUT«Memory allocation failed; could not allocate 14208 bytes␤»
05:10 MadcapJake Hotkeys, on it!
05:11 AlexDaniel perhaps even try increasing 100 to something like 1000
05:12 Hotkeys worked
05:12 Hotkeys for 100
05:13 AlexDaniel I don't know what to do then
05:13 Hotkeys currently working on 1000
05:13 Hotkeys AlexDaniel: upgrade a tiny bit?
05:13 ZoffixWin Same. Works fine with freshly squeezed Perl 6 (single-core box) and on older one too (4-core box)
05:13 Hotkeys 1000 is still-- nvm
05:13 Hotkeys it finished
05:15 Hotkeys oh I see
05:15 Hotkeys I had manually bumped my moar revision
05:15 Hotkeys I was having issues or something I don't remember specifically what
05:17 ZoffixWin The a couple of iterations of 1000 version worked fine on both boxes too: http://i.imgur.com/0ecYPMb.png
05:17 Hotkeys yeah worked fine for me too
05:17 Hotkeys you probably just need to upgrade
05:18 Hotkeys I'm attempting the latest version
05:19 davido_ joined #perl6
05:19 AlexDaniel yeah, ok. I'll probably try again later once I upgrade
05:19 AlexDaniel thanks
05:21 Hotkeys ugexe: I was just able to build on windows fine
05:22 Hotkeys I used gmake instead of nmake if that's any help
05:23 Hotkeys http://i.imgur.com/IPnDRGR.png
05:24 Hotkeys oops
05:24 Hotkeys forgot to switch versions in rakudobrew
05:24 Hotkeys that version is the other build I had
05:24 Hotkeys I promise the current version built :p
05:25 Hotkeys http://i.imgur.com/0Ch1PKH.png
05:25 ugexe its not much help without any of the relevant build information. seeing as you are using rakudobrew i know you aren't using the same nqp version
05:26 Hotkeys fair enough
05:26 ugexe rather, same *moar* version
05:32 TeamBlast joined #perl6
05:32 ugexe `perl Configure.pl --gen-moar=HEAD --gen-nqp && gmake install` if this builds fine then its specific to the compiler toolchain. if not then maybe its moar/nqp related?
05:35 ugexe if only it didn't take so long to build on windows ugh...
05:39 sevvie joined #perl6
05:44 kurahaupo joined #perl6
05:49 Cabanossi joined #perl6
05:57 xpen joined #perl6
06:13 vendethiel joined #perl6
06:22 ugexe maybe it actually is moarvm? https://github.com/MoarVM/MoarVM/commit/b6e0e2a3876d30d356d9d2d7d1ec5edbd689eb53
06:25 ugexe it was with moar HEAD, so that would be it. the travis badge for rakudo is red as well from a failure with --gen-moar=master. lizmat your commits are have nothing to do with it^
06:29 st_iron joined #perl6
06:29 st_iron hi
06:30 AlexDaniel hello
06:30 st_iron say: "good morning"
06:30 st_iron oops
06:33 MadcapJake can roles use the `unit` modifier?
06:35 AlexDaniel m: say ‘good moarning’
06:35 camelia rakudo-moar ec543e: OUTPUT«good moarning␤»
06:35 MadcapJake har har
06:35 AlexDaniel MadcapJake: try it and see if it dwims :)
06:36 MadcapJake it doesn't that's why I ask xD
06:37 MadcapJake it says "unable to parse role definition
06:37 MadcapJake oops
06:37 TEttinger joined #perl6
06:37 MadcapJake It was an ID-10T error
06:38 MadcapJake specifically of the semi-colon variety
06:43 CIAvash joined #perl6
06:45 dupa joined #perl6
06:49 vendethiel joined #perl6
06:50 isBEKaml joined #perl6
06:51 isBEKaml OHHAI, does anyone see failures building rakudo HEAD?
06:51 isBEKaml https://gist.github.com/svatsan/0ca4b3aeee8eaef4fc5e
06:52 dupa hey are you all guys here very old?
06:52 hippie1 joined #perl6
06:52 isBEKaml dupa: why can't we all be both young and old? :-)
06:53 dupa isBEKaml, Perl is older than me
06:53 isBEKaml dupa: well, perl6 is probably younger than you
06:53 dupa sure
06:53 dupa ;d
06:53 dupa but I suppose almost all good perl6 programmers knows perl5
06:54 dupa thats why they choice perl6 for target cause young ppl like swift and go and ruby
06:54 dupa and php etc.
06:54 dupa old ppl like C and Perl
06:58 AlexDaniel dupa: I don't think that your assumptions are 100% correct :)
06:59 AlexDaniel e.g. I'm rather young
06:59 timotimo isBEKaml: lizmat's latest commit is probably responsible for that. could you check that out?
07:00 timotimo dupa: i'm a good perl6 programmer and i don't like perl5
07:00 timotimo also, perl is older than me
07:01 AlexDaniel anyway, who cares? :)
07:02 domidumont joined #perl6
07:02 timotimo right
07:04 timotimo i've grown up with the "hacker" mindset, more or less. so i'm ideally completely ignoring things like age, sex or sexual orientation, origin, ...
07:04 mr-foobar joined #perl6
07:04 isBEKaml timotimo: yes, I'm looking over commit stream right now
07:06 timotimo i should go to bed, i think
07:07 Hotkeys dupa: 18 here, perl6 is my current favorite lang
07:07 AlexDaniel \o/
07:08 Hotkeys |
07:08 Hotkeys / \
07:08 Hotkeys if you have users split from messages like hexchat and I think weechat does
07:08 Hotkeys that makes a stick figure
07:08 domidumont joined #perl6
07:08 Hotkeys if you don't that is unfortunate
07:09 AlexDaniel I'm decapitated, that's unfortunate
07:09 Hotkeys D:
07:10 Hotkeys AlexDaniel: http://i.imgur.com/5nt8mS1.png
07:10 Hotkeys :)
07:11 AlexDaniel m: sub o($){};     \ o /      1/
07:11 camelia rakudo-moar ec543e: ( no output )
07:13 ugexe timotimo: i think its the latest moarvm commits
07:21 geekosaur only makes a stick figure in hexchat if using a monospaced font
07:23 awwaiid joined #perl6
07:25 Hotkeys geekosaur: any sane person does
07:29 protium joined #perl6
07:33 isBEKaml ugexe: you may be right. Going back a few commits in rakudo still produce that error. It's somewhere in moarvm code, probably freeing up stuff a little too enthusiastically :-)
07:35 MadcapJake duda, I've never written perl 5
07:35 MadcapJake dupa ^
07:35 dupa MadcapJake, ;-)
07:35 MadcapJake I come from JavaScript land
07:36 PiL_ joined #perl6
07:37 MadcapJake If I wanted to create a seperate class's object, would I do this in `new` or `BUILD`?
07:37 MadcapJake My class has a Default that inherits and I want to default to creating a Default instance
07:38 timotimo well, the trick in construction really is that when you call self.bless in your new method, the self will be of potentially a more derived class than the class the method is defined in
07:39 timotimo but nothing stops you from return Default.new(...) or even return Default.bless(...)
07:39 timotimo which of the two is more proper for your use case really depends on the specifics
07:40 MadcapJake i'm just passing a few parameters, no fiddling at all
07:40 MadcapJake the function will just take some arguments and set them as parameters of the derived classes `new`
07:40 timotimo if Default has its own "method new" or you're checking self against being Default already, you can just return Default.new(...), otherwise return self.bless(...)
07:41 MadcapJake ok default doesn't have a new, it's just assumed to be there, so in that case it's just going to walk back up to my inherited `new` then right?
07:41 timotimo that's right
07:41 MadcapJake timotimo, thanks!
07:41 timotimo you're welcome!
07:42 MadcapJake can I do Default.bless?
07:42 timotimo yes
07:42 MadcapJake ok cool
07:42 timotimo then you might want to prevent a method new from being implemented in Default, perhaps?
07:43 timotimo because anything that gets put in there later on will not be run when you create a Default implicitly from BaseOfDefault.new
07:46 timotimo it's already been quite a while, but when i first understood the thing about self.bless, my mind was a little bit blown :)
07:47 dupa what WIP mean?
07:47 timotimo "work in progress" usually
07:48 timotimo http://design.perl6.org/S99.html - we have a little glossary where terms we use often that outsiders or newcomers may not know yet are defined and explained and such
07:48 timotimo S99:WIP
07:48 timotimo hm, isn't that how ... oh, synopsebot isn't back yet
07:48 timotimo now i remember i wanted to set it up! darn
07:49 TeamBlast joined #perl6
07:50 CIAvash joined #perl6
07:52 dalek specs: b10af0c | timotimo++ | S99-glossary.pod:
07:52 dalek specs: point out ".ast" is an alternative to ".made".
07:52 dalek specs: review: https://github.com/perl6/specs/commit/b10af0c107
08:02 CurtisOvidPoe joined #perl6
08:05 lokien_ joined #perl6
08:11 sortiz joined #perl6
08:12 darutoko joined #perl6
08:13 dupa joined #perl6
08:15 sortiz \o #perl6
08:16 MadcapJake joined #perl6
08:17 sortiz Is this the proper place for NQP questions?
08:18 vendethiel joined #perl6
08:18 isBEKaml Our resident NQP experts are probably not yet up/awake. Don't let that stop you from asking away
08:19 domidumont joined #perl6
08:22 sortiz isBEKaml, Thanks, nothing specific yet, I been reading the source, and seems that ops.markdown, among others, in nqp/docs is outdated
08:26 rindolf joined #perl6
08:31 isBEKaml sortiz: if you're diving into NQP, you can look at this for an intro: https://github.com/edumentab/rakudo-and-nqp-internals-course
08:31 firstdayonthejob joined #perl6
08:31 isBEKaml timotimo,ugexe: I've isolated the moar commit that broke the build. looking at a fix now.
08:36 sortiz isBEKaml, Yes, I started with that, really good as an intro.
08:38 CurtisOvidPoe joined #perl6
08:48 sortiz But I've found in rakudo's source some code that don't seems to have much sense, and I want to understand the why.
08:59 azawawi joined #perl6
08:59 azawawi hi
09:00 sortiz On the other side, lizmat's recent work show me that there are lots of oportunities for optimization. Well can be the case.
09:00 azawawi time p6doc -n Str   # => (real0m38.804s, user0m38.217s, sys0m0.511s)
09:01 Laurent_R joined #perl6
09:01 azawawi why is it soooo slow?
09:02 azawawi "This is Rakudo version 2016.01 built on MoarVM version 2016.01"
09:06 stmuk_ there was a patch to precomp pod but I don't think it was merged
09:06 azawawi let me check
09:06 dolmen joined #perl6
09:07 stmuk_ https://github.com/perl6/doc/pull/334
09:08 stmuk_ I'm getting p6doc -n Str  9.14s user 0.35s system 99% cpu 9.548 total
09:09 * azawawi tests it
09:09 stmuk_ and that's on a slow (but recent) Celeron
09:10 DoverMo joined #perl6
09:11 RabidGravy joined #perl6
09:12 azawawi hmmm...
09:12 azawawi i found the problem
09:12 _mg_ joined #perl6
09:12 azawawi p6doc-index build   # afterwards slows down p6doc considerably
09:13 azawawi removed index   # => 0m2.121s
09:13 stmuk_ ah its probably EVALFILE the index even if the flags don't specify its use
09:14 azawawi https://github.com/facebooknuclideapm/hyperclick  # im working on atom editor for help/links now
09:14 Sunbeam joined #perl6
09:15 stmuk_ I will look at that EVALFILE thing later if noone else gets there first
09:16 azawawi right now i have no way to rely on the index
09:16 azawawi and p6doc is not smart enough
09:17 stmuk_ the index is very slow ... I was hoping to replace it by moar serialisation but I haven't heard about that recently
09:18 * moritz leans towards an sqlite database these days
09:19 azawawi moritz: i agree
09:19 stmuk_ thats a good idea .. the NetBSD man page system uses its full text search and gets great search results
09:20 stmuk_ I suppose it would add DBIish as a dependency for p6doc (unless its just an option)
09:20 azawawi stmuk_++ # go go :)
09:20 bioexpress joined #perl6
09:20 stmuk_ RSN :)
09:22 azawawi the cool thing is that i could find errors way faster than before via atom lint :)
09:22 azawawi on Perl 6 projects that is
09:23 azawawi we need to define a p6 project lifecycle or build tasks
09:23 azawawi nobody wants to write "prove -v -e -Ilib"
09:24 azawawi simply "p6-builder test"
09:24 CurtisOvidPoe joined #perl6
09:25 DoverMo did anyone like my mascot idea for perl6?
09:25 azawawi or $some-cool-build-name {new|test|build|run|clean}
09:25 moritz DoverMo: I didn't see it
09:25 azawawi DoverMo: link it plz
09:26 sortiz Btw, p6doc-index is broken in windows, more '/' vs '\' assumptions.
09:26 DoverMo it was a river dolphin with han zi on it's body to represent the extinct chinese sub-species
09:27 RabidGravy why don't places use $*SPEC.dir-sep or whatever it is
09:27 moritz much of the p6doc code probably predates $*SPEC
09:30 sortiz Seems missing a doc about portability.
09:33 isBEKaml the commit that broke the build: https://github.com/MoarVM/MoarVM/commit/b6e0e2a3876d30d356d9d2d7d1ec5edbd689eb53
09:33 isBEKaml when do we normally free those handlers? For now, does it make sense to just revert it?
09:34 pmurias joined #perl6
09:34 dalek nqp: 83dc6d6 | (Pawel Murias)++ | src/vm/js/Utils.nqp:
09:34 dalek nqp: [js] Remove a piece of dead code.
09:34 dalek nqp: review: https://github.com/perl6/nqp/commit/83dc6d6766
09:35 xinming joined #perl6
09:39 Vitrifur joined #perl6
09:40 vendethiel joined #perl6
09:41 isBEKaml jnthn: ping
09:46 perlawhirl joined #perl6
09:51 kurahaupo joined #perl6
09:55 domidumont joined #perl6
10:03 ely-se joined #perl6
10:05 vendethiel joined #perl6
10:09 lizmat good *, #perl6!
10:11 pmurias hi
10:12 dalek nqp: f77aa4c | (Pawel Murias)++ | src/vm/js/QAST.nqp:
10:12 dalek nqp: [js] remove useless file
10:12 dalek nqp: review: https://github.com/perl6/nqp/commit/f77aa4c887
10:14 spider-mario joined #perl6
10:15 dupa joined #perl6
10:16 lizmat sortiz RabidGravy : in the newio branch, \ were always changed to / on input, so you didn't have to worry about which platform you worked on
10:16 lizmat (since Win can handle / just fine, or so I'm told)
10:17 RabidGravy At the OS level it can, the "shell" not so much
10:17 nine I seem to remember that it depends on which API you use
10:17 domidumont joined #perl6
10:18 RabidGravy though there used to be an API in DOS whereby you can change the directory separator for CMD.COM which I guess is long gone
10:18 RabidGravy (and the switch character too)
10:20 RabidGravy I've still got a DOS programmers reference upstairs somewhere
10:21 sortiz lizmat, Yup, but, for example we can't assume the results of .find has a mere '/' in windows, see https://github.com/perl6/doc/blob/master/bin/p6doc-index#L39
10:21 DoverMo left #perl6
10:21 lizmat RabidGravy: on Win, it would transform all / to \ before shelling out
10:22 lizmat sortiz: I see...
10:23 wamba joined #perl6
10:23 [Tux] test            22.699
10:23 [Tux] test-t          12.096
10:23 [Tux] csv-parser      51.792
10:25 dalek rakudo/nom: 6ade157 | lizmat++ | src/core/List.pm:
10:25 dalek rakudo/nom: Speed up List.sum by a few percent
10:25 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6ade15746e
10:25 * lizmat wonders what Travis will say about that
10:26 lizmat [Tux] : is that also a bit lower than before ?
10:26 lizmat test-t I mean?
10:27 RabidGravy it's been wibbling around the 12.1 area for a while now
10:28 azawawi do we have a performance graph for that?
10:28 iH2O joined #perl6
10:28 CurtisOvidPoe joined #perl6
10:29 bioexpress joined #perl6
10:29 Zero_Dogg joined #perl6
10:30 azawawi http://pasteboard.co/1lHRgQcE.png  # latest p6doc mock suggestions :)
10:31 stmuk_ originally MSDOS had a SWITCHAR option but it was abandoned but there are still traces
10:32 sortiz lizmat, In a more important question: I see your last commit to Capture, and thought that some nqp::isnull($!list) maybe can be eliminated.
10:32 RabidGravy I'd go an look in the book but can't be arsed, some INT 21 function
10:32 lizmat sortiz: I thought so too, initially, but there are other places in the core that create Captures without having a $!list
10:33 stmuk_ I think FreeDOS has it still http://help.fdos.org/en/hhstndrd/cnfigsys/switchar.htm
10:33 azawawi INT 21 # sweet memories :)
10:33 azawawi i still have a copy of PC Intern :)
10:34 sortiz But in BUILD, $!list seems initialized unconditionally
10:34 [Tux] lizmat: within nois level
10:34 [Tux] e
10:34 stmuk_ I have Peter Norton's "Shirt" book somewhere
10:34 lizmat :-(   but still  :-)
10:34 [Tux] 12.000 is oke-ish
10:35 iH2O left #perl6
10:38 vendethiel joined #perl6
10:39 sortiz lizmat, ie, in Capture.BUILD its assumed that @list is defined.
10:39 stmuk_ wonders how the docs are going to handle 6.d and 6.c
10:40 lizmat sortiz: but Hash.Capture directly creates a Capture object *without* $!list
10:42 _ramix_ joined #perl6
10:46 sortiz lizmat, I see, thanks.
10:48 * RabidGravy skips the changes feed tests in Sofa for speedy goodness
10:49 RabidGravy there, the tests take a minute less
10:56 RabidGravy Really need to get the EventSource client going, I *think* that it may be a matter of getting a "receive chunk" handler into H::UA and pushing the chunks to a supply
10:57 RabidGravy but I want to do all of the hooks stuff in H::UA in one go
10:57 RabidGravy unless someone else fancies doing it first
11:03 dalek doc: 2bc55d8 | (Salvador Ortiz)++ | bin/p6doc-index:
11:03 dalek doc: Use $*SPEC.dir-sep in p6doc-index
11:03 dalek doc: review: https://github.com/perl6/doc/commit/2bc55d8ba0
11:03 pangur joined #perl6
11:04 pangur crosspost: #emacs, #perl6 Anyone here use emacs to edit and compile perl6 code?
11:04 RabidGravy I'd be surprised if there wasn't
11:12 azawawi http://doc.perl6.org/ is the latest documentation... what about documentation for specific version (e.g. 2016.01)?
11:13 travis-ci joined #perl6
11:13 travis-ci Doc build failed. Salvador Ortiz 'Use $*SPEC.dir-sep in p6doc-index'
11:13 travis-ci https://travis-ci.org/perl6/doc/builds/107423805 https://github.com/perl6/doc/compare/7b80938dd02b...2bc55d8ba075
11:13 travis-ci left #perl6
11:14 RabidGravy azawawi, the documentation isn't currently versioned AFAIK
11:15 RabidGravy possibly should be tagged at release time
11:15 CurtisOvidPoe joined #perl6
11:24 vendethiel joined #perl6
11:39 kjs_ joined #perl6
11:56 ab5tract I don't think we've gotten to a point where we feel that it fully covers 6.c
11:56 ab5tract but it would make more sense to tag it to Perl 6 releases, not Rakudo ones
11:56 dalek doc: 1bc7a03 | (Salvador Ortiz)++ | bin/p6doc-index:
11:56 dalek doc: Revert "Use $*SPEC.dir-sep in p6doc-index"
11:56 dalek doc:
11:56 dalek doc: This reverts commit 2bc55d8ba075ce9671ed0ecfa535629956040c5c.
11:56 dalek doc: Must check first why travis-ci buid failed
11:56 dalek doc: review: https://github.com/perl6/doc/commit/1bc7a03a59
11:57 azawawi it would make also sense to make releases out of it so editors/IDEs can index that info
11:57 ab5tract azawawi: sure, but it's not Rakudo documentation. It's Perl 6 documentation.
11:57 azawawi so im finished with the UI for linking to Perl 6 documentation
11:57 azawawi and i cannot use it lol
11:58 azawawi p6doc -f say
11:58 azawawi p6doc -n say
11:58 ab5tract azawawi: ah, that's still broken?
11:58 ab5tract azawawi: it sounds like it might be time to return to the p6doc-server idea
11:59 azawawi or generate a json index that is a part of a release :)
11:59 azawawi bummer... wasted time on http://pasteboard.co/1lHRgQcE.png
12:00 ab5tract azawawi: AFAIR, the file that is generated by p6doc-index is just a hash declaration which gets EVAL'd
12:00 azawawi index.data is unusable
12:00 ab5tract you could try just calling to-json on that
12:00 wamba joined #perl6
12:00 azawawi tried lol
12:00 ab5tract ok
12:01 ab5tract oh man
12:01 ab5tract that is not time wasted
12:01 ab5tract let's get that working :)
12:02 ab5tract the plugin just needs a JSON version of index.data? is there a specification as to what the structure should look like?
12:03 ab5tract or does the plugin shell out to `p6doc -f $query` ?
12:03 azawawi i wanted to shell out
12:03 ab5tract so we just need to fix p6doc?
12:03 azawawi yup
12:03 azawawi :)
12:04 azawawi the worst thing for a developer is to see an IDE saying "Indexing..." :)
12:05 ab5tract well, the current speeds of p6doc when it was working would not impress anyone either
12:05 dupa joined #perl6
12:06 ab5tract because it slurps and evals that hash file every time
12:06 azawawi also one more thing, we need a human friendly index... which means
12:06 ab5tract obviously becoming slower and slower as the docs get better and better
12:06 azawawi i need to find the documentation of "say" by holding ctrl and then clicking on it :)
12:06 azawawi as simple as that
12:06 travis-ci joined #perl6
12:06 travis-ci Doc build failed. Salvador Ortiz 'Revert "Use $*SPEC.dir-sep in p6doc-index"
12:06 travis-ci https://travis-ci.org/perl6/doc/builds/107428713 https://github.com/perl6/doc/compare/2bc55d8ba075...1bc7a03a596c
12:06 travis-ci left #perl6
12:07 ab5tract azawawi: right, i see. so you need the index to know what you can shell out for
12:07 azawawi yeah that could be an optimization
12:07 vendethiel joined #perl6
12:08 kjs_ joined #perl6
12:08 azawawi use FooBar;  # when FooBar is clicked, we should its documentation
12:08 sortiz Oh my! Seems unrelated.. But why?!?
12:09 azawawi sortiz: seems like bad code
12:10 RabidGravy anyway, off out for a bit. toodles
12:10 ab5tract ciao RabidGravy!
12:10 azawawi RabidGravy: have fun
12:11 a3r0 joined #perl6
12:12 azawawi ab5tract: hmmm, i need to will use doc.perl6.org api right away :)
12:12 sortiz azawawi, I commit 2bc55d8 and fail, so I revert it in 1bc7a03, but failed again, see: https://travis-ci.org/perl6/doc/builds, I can't understand!
12:12 ab5tract azawawi: sorry, can you clarify a bit?
12:13 ab5tract sortiz: could be that it was already broken?
12:13 azawawi ab5tract: doc.perl6.org search bar
12:13 ab5tract azawawi: yeah, but are you saying that you will just use that instead for now?
12:13 ely-se joined #perl6
12:13 azawawi ab5tract: yup
12:14 ab5tract "i need to will use" <--- this was what confused me :)
12:14 ab5tract ok, so there's a workaround for now!
12:14 azawawi ab5tract: hopefully it will work
12:14 * azawawi takes a look at http://doc.perl6.org/js/search.js
12:14 sortiz ab5tract, Can be, but the build history in Travis don't show any evidence. I'm totally confused.
12:15 ely-se joined #perl6
12:17 jnthn isBEKaml: pong
12:19 SCHAAP137 joined #perl6
12:20 jnthn .tell isBEKaml I'm only about now and then during the weekend, so may be easier to .tell me what you want to ask :)
12:20 yoleaux jnthn: I'll pass your message to isBEKaml.
12:21 azawawi ab5tract: https://github.com/azawawi/atom-perl6-editor-tools/blob/master/lib/index.json
12:24 ab5tract azawawi: awesome!
12:28 kid51 joined #perl6
12:28 ab5tract .tell RabidGravy unfortunately the Tinky example workflow doesn't work on OS X either. I like the looks of this library though!
12:28 yoleaux ab5tract: I'll pass your message to RabidGravy.
12:30 kid51 Is there as yet a perl6 executable equivalent to 'perldoc'?
12:30 ab5tract kid51: p6doc, but it may be broken at this moment
12:31 kid51 If it did exist, I would have expected to find it along these other executables:
12:31 kid51 ls ~/rakudo-star/bin/
12:31 kid51 moar  nqp  nqp-m  perl6  perl6-debug-m  perl6-gdb-m  perl6-m  perl6-valgrind-m
12:31 isBEKaml .tell jnthn I was looking for some advice on fixing a broken commit: http://irclog.perlgeek.de/perl6/2016-02-06#i_11998442. No rush - enjoy your weekend :-)
12:31 yoleaux 12:20Z <jnthn> isBEKaml: I'm only about now and then during the weekend, so may be easier to .tell me what you want to ask :)
12:31 yoleaux isBEKaml: I'll pass your message to jnthn.
12:31 kid51 Hmm, I see it in: ~/rakudo-star/share/perl6/site/bin/p6doc
12:33 ab5tract kid51: i noticed that the binary was not where i expected it earlier today. i use rakudobrew and panda, and it seems that you see something similar with rakudo-star
12:34 geekosaur I have a p6doc in my (older) rakudobrew. seems something might be broken recently
12:34 ab5tract it needs some love. i proposed a reorganization of the CLI interface into something that fetches from an independent p6doc-server process
12:35 ab5tract but i haven't had the tuits
12:35 dupa joined #perl6
12:36 azawawi we need commands like the following https://golang.org/cmd/go/ in perl6... could be in panda/zef or new pkg
12:36 rindolf joined #perl6
12:36 ab5tract there is a pretty swazzy looking JSON-RPC library for perl 6 that we could use
12:37 kid51 $ p6doc -f push
12:37 kid51 In order to use unqualified sub and method names like "p6doc -f say"
12:37 ab5tract https://github.com/bbkr/jsonrpc
12:37 kid51 you will need to run "p6doc-index build" to build index.data.
12:37 kid51 Otherwise use "p6doc -f Type::Str.split" instead of "p6doc -f split" for now.
12:37 kid51 Not very user friendly
12:37 ab5tract kid51: I believe I've just pointed out that the tool needs some love :)
12:37 kid51 One wonders why p6doc-index build could not be done on installation
12:37 azawawi kid51: that and why wasnt it built when p6doc was installed :)
12:38 kid51 ab5tract: Yes, I'm trying to anticipate the complaints I'm going to get later today at our beginners study group
12:38 ab5tract kid51: right on. i would say that you might want to consider the CLI docs broken at the moment
12:38 azawawi ab5tract: then a good project is https://golang.org/cmd/go/ :)
12:39 azawawi ab5tract: i already built the UI for it... just make the command https://github.com/azawawi/atom-perl6-editor-tools#build-support
12:39 ab5tract in the context of a class
12:39 ab5tract the web interface has been given a lot more care and attention to detail
12:40 ab5tract s/class/study group/
12:44 rindolf joined #perl6
12:45 jnthn isBEKaml: Yeah, nwc10++ reported more into on the failure mode on #moarvm; reverted it, since it wasn't obvious how to do a more correct fix.
12:45 yoleaux 12:31Z <isBEKaml> jnthn: I was looking for some advice on fixing a broken commit: http://irclog.perlgeek.de/perl6/2016-02-06#i_11998442. No rush - enjoy your weekend :-)
12:45 azawawi ab5tract: any link on the topics you're going to cover in your beginners study group?
12:46 ab5tract azawawi: i think you want to ping kid51 :)
12:46 isBEKaml jnthn: yeah, I didn't know it was reverted, thanks
12:47 jnthn isBEKaml: I only just did it :)
12:47 musiKk joined #perl6
12:47 azawawi ab5tract: sorry got confused.. im writing perl6 in the vm and android on windows host :)
12:47 jnthn After looking into a correcter fix and deciding I'm not clever enough to figure it out today. Or at least, not at this part of today. :-)
12:48 azawawi kid51: any link on the topics you're going to cover in your beginners study group?
12:48 kid51 That will be discussed at Session 1.
12:48 kid51 There are several different tutorials on web.
12:49 kid51 And of course some who show up will want to Do Their Own Thing, while others will want to work more collectively
12:49 kid51 So, unknown yet.
12:49 kid51 Will report once we have something that is actually working
12:51 ab5tract kid51: that sounds like a good approach.
12:53 BenGoldberg joined #perl6
12:53 kid51 Here is the original call for this group:  http://thenceforward.net/perl/misc/nyc_perl6/20160116_00_call_for_group.txt
12:53 isBEKaml jnthn: okay - I was leaning towards reverting it, since I don't know how far up above that piece of code stands, where it needs to have handlers freed. C is a terrible flat-land :P
12:54 FROGGS joined #perl6
12:56 jnthn isBEKaml: Well, it's not that big a deal in so far as NQP_REVISION/MOAR_REVISION weren't bumped
12:57 jnthn Those give us a little breathing room :)
12:58 isBEKaml jnthn: ah, okay
12:58 dupa hey how can I use perl module without installing it via panda
13:01 azawawi kid51: thx for the info
13:04 franciskittu joined #perl6
13:05 vendethiel joined #perl6
13:06 sortiz Some change in rakudo between g8bd7ee6 and g6ade157 seems to broke the build for p6docs, I'm investigating..
13:06 lizmat which repo?
13:08 sortiz lizmat, This is Rakudo version 2015.12-242-g8bd7ee6 vs This is Rakudo version 2015.12-251-g6ade157
13:10 lizmat I wonder where that g comes from
13:10 lizmat doesn't look very hexadecimal to me
13:11 sortiz Yup, copied from travis-ci logs.
13:11 isBEKaml IIRC, it was just a prefix after which comes the revision string
13:11 lizmat initially I was wondering why I couldn't find those commit ID's in the git log
13:11 geekosaur git describe, and as far as I know it stands for git
13:12 sortiz In my rush don't noticed the 'g'
13:13 * geekosaur keeps grabbing those in other projects too, wish they had include a separator so select-word didn't always grab the g as part of the commit hash
13:14 geekosaur because everyone and their pet banth includes git describe in artifact ids :/
13:15 geekosaur but of course git doesn't *accept* it in a commit hash
13:16 sortiz For perl6/doc, the failure "This type does not support associative operations" is reported in https://github.com/perl6/doc/blob/master/lib/Perl6/Documentable/Registry.pm#L26
13:17 sortiz Thats is a common autovivification, that somehow now isn't working.
13:18 ab5tract
13:24 sortiz With a simple "%!cache{$by} = {};" before the for, the make succeeds, that file wasn't touched in weeks.
13:24 * jnthn wonders if the hash access opts maybe accidentally lost a :D on the invocant somewhere...
13:24 perlawhirl joined #perl6
13:24 jnthn Also if we do have a regression there, we clearly need more autoviv tests :)
13:25 magiczna_dupa I include bailador to project
13:25 magiczna_dupa but it seams
13:26 magiczna_dupa it has bugs
13:26 FatalNIX left #perl6
13:27 sortiz How I have to proceed?
13:29 ab5tract magiczna_dupa: I've never used it, but I have heard others say the same :(
13:30 lizmat jnthn : I'm afraid that's another case of attrinited given a false positive because something did a getattr before
13:30 Rotwang magiczna_dupa: great nick
13:31 lizmat jnthn: in other words, use of attrinited is making code fragile  :-(
13:32 lizmat reverting use of attrinited in hash<a>
13:32 Actualeyes joined #perl6
13:33 jnthn lizmat: Well, yeah, you have to be very consistent with it. It's designed precisely for construction time stuff.
13:34 lizmat well, I was consistent enough to not break any spectests  :-(
13:34 magiczna_dupa joined #perl6
13:36 perlawhirl joined #perl6
13:36 jnthn lizmat: Yeah, guess we need more of those ;)
13:36 jnthn Can revert it for now anyway
13:36 lizmat yeah, spectesting the revert atm
13:36 jnthn We may be able to find some more robust solution now we know there's something to be had optimization wise in that area.
13:43 dalek rakudo/nom: 7a51234 | lizmat++ | src/core/Hash.pm:
13:43 dalek rakudo/nom: Revert use of attrinited in Hash<a>
13:43 dalek rakudo/nom:
13:43 dalek rakudo/nom: This should fix problems in making docs, sortiz++ for reporting
13:43 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7a51234075
13:43 lizmat sortiz: could you check if this fixes the problem ?
13:44 sortiz lizmat, sure.
13:46 sortiz lizmat, give me a few minutes.
13:46 lizmat sure  :-)
13:51 sortiz lizmat, indeed fixed.
13:52 lizmat cool, thanks for checking!
13:53 FROGGS_ joined #perl6
13:53 dalek rakudo/nom: 04eac9b | lizmat++ | src/core/Str.pm:
13:53 dalek rakudo/nom: Fix magic inc/dec on beer mugs and camels
13:53 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/04eac9bc5d
13:54 raiph joined #perl6
13:57 * sortiz I need to understand well NQP ASAP
14:00 isBEKaml joined #perl6
14:00 ZoffixWin m: say so / <-[]+\s>* / for " ", "\t"
14:00 camelia rakudo-moar 7a5123: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YLd9cPWNdS␤Unable to parse expression in metachar:sym<assert>; couldn't find final '>' ␤at /tmp/YLd9cPWNdS:1␤------> 3say so / <-[]+7⏏5\s>* / for " ", "\t"␤    expecting any of:␤        prefix␤»
14:01 ZoffixWin I recall there was a way to specify "this character class, but minus this character"
14:01 Skarsnik joined #perl6
14:01 ZoffixWin Like \s but without space 32
14:02 dalek doc: a9fb33d | (Salvador Ortiz)++ | WANTED:
14:02 dalek doc: Add OS portability to WANTED
14:02 dalek doc:
14:02 dalek doc: And I need to kick start a travis-ci build :)
14:02 dalek doc: review: https://github.com/perl6/doc/commit/a9fb33d2c2
14:03 ZoffixWin m: say so / <-[ ]+[\s]> / for " ", "\t"
14:03 camelia rakudo-moar 7a5123: OUTPUT«===SORRY!===␤Iteration past end of iterator␤»
14:03 ZoffixWin :o
14:04 RabidGravy joined #perl6
14:05 ZoffixWin ah, I was close:
14:05 ZoffixWin m: say so / <[\s]-[ ]> / for " ", "\t"
14:05 camelia rakudo-moar 7a5123: OUTPUT«===SORRY!===␤Iteration past end of iterator␤»
14:05 ZoffixWin m: say so / <[\s] - [ ]> / for " ", "\t"
14:05 camelia rakudo-moar 7a5123: OUTPUT«===SORRY!===␤Iteration past end of iterator␤»
14:05 ZoffixWin m: so 'foo' ~~ / <[ a..z ] - [ f o ]> + /;
14:05 camelia rakudo-moar 7a5123: OUTPUT«WARNINGS for /tmp/wIJpl2RQFE:␤Useless use of "so " in expression "so 'foo' ~~" in sink context (line 1)␤»
14:05 ZoffixWin m: say so 'foo' ~~ / <[ a..z ] - [ f o ]> + /;
14:05 camelia rakudo-moar 7a5123: OUTPUT«False␤»
14:05 ZoffixWin wtf
14:06 ZoffixWin m: say so / <[\t] - [ ]> / for " ", "\t"
14:06 camelia rakudo-moar 7a5123: OUTPUT«===SORRY!===␤Iteration past end of iterator␤»
14:06 ZoffixWin m: say so / <[ \t ] - [x]> / for " ", "\t"
14:06 camelia rakudo-moar 7a5123: OUTPUT«False␤True␤»
14:06 ZoffixWin m: say so / <[ \t ] - [\ ]> / for " ", "\t"
14:06 camelia rakudo-moar 7a5123: OUTPUT«False␤True␤»
14:06 ZoffixWin m: say so / <[ \s ] - [\ ]> / for " ", "\t"
14:06 camelia rakudo-moar 7a5123: OUTPUT«False␤True␤»
14:06 vendethiel joined #perl6
14:06 ZoffixWin So it's an unescaped space inside a character class without anything in it.
14:07 isBEKaml m: say so /<[\s] - []>/ for " ", "\t"
14:07 camelia rakudo-moar 7a5123: OUTPUT«===SORRY!===␤Iteration past end of iterator␤»
14:07 isBEKaml same behaviour
14:08 ZoffixWin Sure, whitespace is insignificant
14:08 isBEKaml not what I meant - [ ] was treated as [] and didn't see a character searching it through
14:09 isBEKaml so it went past end of string. It should have some sort of check around those bounds
14:16 ZoffixWin Which string?
14:17 ZoffixWin I'd think it went in looking for characters in a character class, but there were none, hence the 'past end' message.
14:17 ZoffixWin I rather not add any checks for bounds if that means making the thing slower.
14:17 ab5tract o/ ZoffixWin
14:17 ZoffixWin ?
14:18 ab5tract I've enjoyed your blog posts a lot!
14:18 ab5tract Just waving hello, if that '?' was directed at me.
14:19 ZoffixWin After the vitriol you were spewing into my /msg a month or so ago, a wave of hello is suspicious.
14:20 lizmat .oO( a month is a lot of time on the internets )
14:21 RabidGravy two days is a long time on the internets
14:21 yoleaux 12:28Z <ab5tract> RabidGravy: unfortunately the Tinky example workflow doesn't work on OS X either. I like the looks of this library though!
14:21 lokien_ joined #perl6
14:22 ab5tract ZoffixWin: it was meant as a conciliatory gesture, in hopes of moving on..
14:22 ZoffixWin Very well... Thanks for liking my blog post.
14:22 RabidGravy ab5tract, it's file a bug against rakudo/moar/libuv then  :)
14:23 ab5tract RabidGravy: true
14:23 RabidGravy I'd go with checking that a minimal node.js example works to rule out the libuv part
14:24 cdg joined #perl6
14:26 ely-se joined #perl6
14:26 telex joined #perl6
14:27 travis-ci joined #perl6
14:27 travis-ci Doc build passed. Salvador Ortiz 'Add OS portability to WANTED
14:27 travis-ci https://travis-ci.org/perl6/doc/builds/107441749 https://github.com/perl6/doc/compare/1bc7a03a596c...a9fb33d2c2b5
14:27 travis-ci left #perl6
14:29 sortiz I can relax now :)
14:31 RabidGravy ab5tract, e.g. "var fs = require('fs'); fs.watch('.', function (event, filename){ console.log('event is: ' + event); });" if that works in node then rakudo/bug if not libuv bug
14:31 ab5tract RabidGravy: gotcha.
14:33 Deconachick joined #perl6
14:33 Deconachick rakudo
14:34 Deconachick .msg camelia p6
14:34 yoleaux Deconachick: Sorry, this command is admin-only.
14:35 sortiz till tomorrow #perl6
14:36 lizmat sortiz o/
14:38 MadcapJake joined #perl6
14:39 ab5tract cioa sortiz
14:40 vendethiel joined #perl6
14:44 Actualeyes joined #perl6
14:46 lizmat jnthn: wrt to additional tests, I'm not quite sure where to add them atm, if it's about things getting added for 6.c
14:47 lizmat or for 6.d.a for that matter
14:47 moritz I'm looking for volunteers to proof-read a German Perl 6 article for me (and give me feedback)
14:47 moritz hoelzro++ read a first draft already, but more eyes would be appreciated
14:48 moritz FROGGS_, timotimo, [ptc], nine: ^^
14:48 lizmat moritz: as long as you don't expect me to find any errors in German, I'm game :-)
14:49 dalek rakudo/nom: d25b51e | lizmat++ | src/core/Str.pm:
14:49 dalek rakudo/nom: Make magic inc/dec on strings 2x as fast
14:49 dalek rakudo/nom:
14:49 dalek rakudo/nom: By not needing an iterator to do initializations
14:49 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d25b51e6af
14:49 moritz lizmat: I expect that too, but nto from you :-)
14:50 rpburkholder joined #perl6
14:57 FROGGS[mobile] joined #perl6
14:58 FROGGS[mobile] moritz: can you email me details?
14:58 dalek rakudo/nom: dbc7115 | lizmat++ | src/core/Junction.pm:
14:58 dalek rakudo/nom: Remove superfluous return
14:58 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/dbc7115897
14:58 dalek rakudo/nom: 2fb2fbc | lizmat++ | src/core/ (2 files):
14:58 dalek rakudo/nom: Mark some more subs as not returning anything
14:58 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2fb2fbc7f5
14:59 moritz FROGGS[mobile]: sure
15:00 nine Why exactly can't the optimizer just remove the return if it's the last statement in a function?
15:02 moritz well, it'd have to verify that it resolves to &return from CORE
15:02 xinming joined #perl6
15:03 moritz and it also can't do it on "is rw" routines (because return decontainerizes
15:03 nine Which should be possible as we statically resolve subs at compile time, don't we?
15:03 moritz we do
15:06 moritz oh, and needs to check the absence of CONTROL blocks too
15:07 RabidGravy Hmm, if e.g. a PRE phaser is run, however it is added, if it returns false then  there should be a "Precondition failed" right?
15:08 moritz yes
15:08 moritz m: sub f() { PRE 0; say 'foo' }; f
15:08 camelia rakudo-moar 04eac9: OUTPUT«Precondition '0' failed␤  in sub f at /tmp/cSX4qG3I4D line 1␤  in block <unit> at /tmp/cSX4qG3I4D line 1␤␤»
15:09 cognome joined #perl6
15:10 TimToady joined #perl6
15:10 RabidGravy yeah, so if I do .add_phaser('PRE", { False }); in a trait that should work too (having confirmed it is run), because it doesn't fail
15:12 moritz note that routines that don't have phasers are codegen'd without support for phasers, for performance reasons
15:13 RabidGravy right, but the phaser does get run, it just doesn't boom
15:14 moritz RabidGravy: ah, the codegen adds the X::Phaser::PrePost
15:14 moritz RabidGravy: so you have to do that manually
15:15 RabidGravy m: sub trait_mod:<is>(Routine $r, :$phased!) { $r.add_phaser('PRE', sub (|c --> Bool) { say "PRE"; False; }); }; sub foo() is phased { say "HARR"; }; foo();
15:15 camelia rakudo-moar 2fb2fb: OUTPUT«PRE␤HARR␤»
15:15 RabidGravy ah
15:16 RabidGravy gotcha
15:20 AlexDaniel joined #perl6
15:23 virtualsue joined #perl6
15:37 kurahaupo joined #perl6
15:37 vendethiel joined #perl6
15:46 [Tux] will p6 get a num128?
15:46 dalek doc: 0bfe797 | (Steve Mynott)++ | bin/p6doc-index:
15:46 dalek doc: only display index file if it exists
15:46 dalek doc: review: https://github.com/perl6/doc/commit/0bfe797c5b
15:48 dalek rakudo/nom: 891d456 | lizmat++ | src/core/Mu.pm:
15:48 dalek rakudo/nom: Make basic object creation 5% faster
15:48 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/891d456b80
15:49 RabidGravy lizmat++ properly on a roll with this stuff
15:49 lizmat [Tux]: I wonder whether this has any affect on test-t
15:49 * [Tux] tries ...
15:58 [Tux] :( panda is littered with if expr { … return } else { … return } :(
15:59 [Tux] s/panda/Inline::Perl5/
15:59 [Tux] test            22.570
15:59 [Tux] test-t          12.073
15:59 [Tux] csv-parser      50.755
16:01 lizmat that's still within noise, but also lower, right ?
16:01 [Tux] yes
16:02 dalek doc: 2820630 | (Steve Mynott)++ | bin/p6doc:
16:02 dalek doc: EVALFILE too slow for usual case only use with -f flag
16:02 dalek doc: review: https://github.com/perl6/doc/commit/2820630f9b
16:02 [Tux] CSV::Parser was 52 earlier today
16:03 vendethiel joined #perl6
16:08 [Tux] and fixing Inline::Perl5 for -Duselongdouble is beyond my capabilities at the moment
16:10 jnthn [Tux]: num128 isn't on the roadmap for the near future
16:11 nine I guess num80 would be more precise
16:13 jnthn We get a lot of simplicity in MoarVM out of local/lexical storage being an array of 64-bit things (in a union type). I know that won't last forever (we'll likely also want to do complex value types at some point), but it'll be a large change.
16:14 jnthn lizmat: master branch of roast
16:14 jnthn lizmat: You seem to think 6.c is mutable
16:15 jnthn lizmat: Short of any errata that goes into the 6.c branch of roast, it's not.
16:15 jnthn Language releases are "snapshots" of the test suite at the release point, essentially.
16:16 [Tux] jnthn, was just asking as Inline::Perl5 is currently non-functional for perl5 -Duselongdouble (even if NV's are not used)
16:16 jnthn [Tux]: Is it common to build Perl 5 that way?
16:17 firstdayonthejob joined #perl6
16:17 jnthn (Just trying to get a picture of if this is a fairly normal configuration we should try to find a good solution for, or something less usual we can just do a bodge around for now...)
16:18 [Tux] FWIW *all* the production perl's *I* use since 5.18.1 are -Duselongdouble -Duse64bitall
16:19 jnthn OK
16:19 [Tux] I cannot tell for what others do
16:19 flussence Gentoo's is use64bitint=define, use64bitall=define, uselongdouble=undef
16:21 [Tux] OpenSUSE default is also -Uuselongdouble (uselongdouble=undef)
16:21 geekosaur mint (presumably also ubuntu) has it undef
16:21 geekosaur likewise os x
16:21 ugexe and debian
16:22 [Tux] my guess is the -Uuselongdouble is way more widespread than -Duselongdouble
16:22 jnthn OK
16:22 jnthn So, we should "fix it somewhow" for now, sounds like :)
16:23 [Tux] yes, but there are (for me) more important fixes with a higher priority
16:23 RabidGravy as a complete aside, can a LEAVE know *how* it is leaving a block?
16:24 [Tux] RT#127358 and utf8-c8
16:24 lizmat RabidGravy: not afaik
16:24 dalek doc: 96d063e | (Steve Mynott)++ | bin/p6doc-index:
16:24 dalek doc: make more windows friendly
16:24 dalek doc: review: https://github.com/perl6/doc/commit/96d063e7d5
16:24 lizmat [Tux]: I nudged jnthn yesterday already about that
16:24 lizmat [Tux]: it appears not quite an easy thing to fix generally
16:24 * [Tux] is aware
16:25 [Tux] if it was, it would have been fixed already  :)
16:25 Juerd lizmat: Thanks for the speed improvements. MQTT::Client's test suite went from 10 to 8,7 seconds since last week
16:25 lizmat Juerd: cool!
16:26 Juerd That means 8,6 of shavings left until it matches p5 ;)
16:26 jnthn [Tux]: It's pretty high on my list, but I'm still getting myself back into things.
16:26 Juerd But it's a tremendous start :)
16:27 jnthn [Tux]: As lizmat noted, I'm concerned the construction of it is fragile.
16:27 azawawi https://github.com/azawawi/atom-perl6-editor-tools#context-sensitive-help  # Context-sensitve help is here for Perl 6 :)
16:28 RabidGravy as the man said it's all about the "aggregation of marginal gains"
16:29 Juerd Who said that?
16:30 RabidGravy Dave Brailsford, the manager of the GB cycling team
16:30 Juerd That's probably why I never heard that one before :D
16:32 * geekosaur wonders if atom has matured past the memory leak that caused him to uninstall it
16:32 RabidGravy but the idea is sound, don't look for the one golden bullet, improve everything marginally
16:32 azawawi geekosaur: windows, linux or macosx?
16:32 geekosaur linux
16:33 geekosaur left it sitting for 2 days, it swapped everything else out
16:33 Juerd RabidGravy: I believe in profiling and that usually shows that there are golden bullets ;)
16:33 azawawi well i use all the time, the trick is to update lol
16:33 azawawi on windows, it auto-updates. On linux, it doesnt
16:33 Juerd RabidGravy: One of them is the regex engine. My test suite is very regex-heavy and that's probably why it's so slow.
16:34 Juerd In fact, it's all regexes
16:34 azawawi geekosaur: and... i noticed that chrome and atom can sometimes break each other
16:35 geekosaur so chrome's js engine isn't properly self-contained?
16:35 Juerd RabidGravy: But I can wait. Perl 5 is still great for my day to day production performance needs, and Perl 6 is great for just having fun :)
16:35 jnthn lizmat: Think I'll try and write up a FQA-style blog post on versioning stuff
16:35 jnthn lizmat: In an attempt to de-confuse things :)
16:35 azawawi geekosaur: it happened to me... i was apt-get upgrading chromium... and my atom broke
16:36 lizmat jnthn: excellent idea, with some practical examples, yes!
16:36 Juerd jnthn: I wonder, could a spectest be added to a released language version, if it is known to pass for all existing implementations?
16:36 geekosaur oh. that'd be debian's fault then, probably
16:37 geekosaur (debian has a very strict policy of sharing components that can be shared, so they'll have patched chromium and atom to use a common v8. v8 may not be designed for that...)
16:37 lizmat grrr.... got another 10% speed increase on basic object construction, but it fails 1 test  :-(
16:38 geekosaur (alhouhg that does still suggest v8 could do with more ability to have isolated instances)
16:39 Xor_ joined #perl6
16:39 Juerd lizmat: Just volkswagen the test ;)
16:40 lizmat yeah, the thought has crossed my mind  :-)
16:41 jnthn 1 test fail can be important :)
16:42 jnthn Juerd: I'd somewhat prefer not to
16:42 jnthn Juerd: It'll get in the way of seeing what errata a given language version had
16:43 Juerd Ah, yes, it would pollute the diff.
16:46 pmurias geekosaur: ninja patching of stuff by package maintainers sucks :/
16:46 lizmat sigh, the test doesn't fail when copied to a piece of code  :-(
16:48 lizmat hmmm.... seems I need to have Test loaded for the test to fail :-(
16:49 lizmat spooky
16:50 wamba joined #perl6
16:52 lizmat only when specifying a value for a native numeric attribute
16:53 lizmat $ 6 'class C { has num $.a }; C.new(a => 42e0)'   # golfed, is fine
16:53 lizmat $ 6 'use Test; class C { has num $.a }; C.new(a => 42e0)'
16:53 lizmat P6opaque: invalid kind in attribute bind
16:53 lizmat jnthn: ^^^ thoughts ?
16:56 pmqs_ I have an inner module that exports an enum. Want to export that enum from an outer module, but I'm struggling. This ticket, https://github.com/perl6/roast/issues/45 seems to imply that re-export is broken.
16:58 ugexe pmqs_: this might help: https://github.com/LLFourn/p6-CompUnit-Util
16:59 pmqs_ ugexe, thanks, I'll take a look
16:59 lizmat jnthn: fwiw, it feels like my change is tickling a native num bug
17:00 stmuk_ has anyone seen problems with write locks on precomp/.lock when running 'shell "perl6..."' inside another perl6 process?
17:01 khw joined #perl6
17:04 ugexe you can try to run the inner proc with -IFakeDir as the first argument and it might "fix" it (i use -Ilib/.precomp to hide it, as it will end up creating a lib/.precomp/.precomp)
17:05 pmqs_ ugexe, brilliant CompUnit::Util worked fine.
17:06 ugexe so like run($*EXECUTABLE, '-IFakeDir', '-Ilib', $file)
17:10 FROGGS joined #perl6
17:11 jnthn lizmat: That error implies it's nqp::bindattr-ing it instead of nqp::bindattr_n-ing it.
17:12 lizmat jnthn : https://gist.github.com/lizmat/fc376b2b5cb16619ab2b
17:12 stmuk_ ugexe++ # yes that works
17:12 lizmat but how would having use Test loaded or not make a difference ????
17:12 RabidGravy maybe it's a reverse volkswagen
17:12 timotimo perhaps %attrinit{$key_name} has different semantics from .AT-KEY($key_name) somehow?
17:13 lizmat only if %attrinit would be special
17:13 timotimo hrm
17:13 lizmat but it's not, it's just an ordinary hash created in Mu.bless
17:14 timotimo can you tell us which exact branch does the problem?
17:14 lizmat it fails at the top of the loop, which is even stranger
17:14 timotimo i mean, the one that does num isn't changed, is it?
17:14 timotimo oh, that's probably just line num reporting fail?
17:15 jnthn lizmat: Why all the removals of my $key_name = ...?
17:15 jnthn oh, it got inlined
17:15 lizmat yup
17:15 BenGoldberg joined #perl6
17:15 lizmat $ MVM_SPESH_DISABLE=1 6l 'use Test; class C { has num $.a }; C.new(a => 42e0)'
17:16 isBEKaml lizmat: that error you see with p6opaque is most likely a recent one, but you may already know this.
17:16 lizmat looks like it is spesh related  :-(
17:16 timotimo oh, you have a different patch that i haven't looked at
17:16 lizmat timotimo: https://gist.github.com/lizmat/fc376b2b5cb16619ab2b
17:16 isBEKaml lizmat: I don't see that error on my version here: This is Rakudo version 2015.12-251-g6ade157 built on MoarVM version 2016.01-17-g4671fae
17:17 lizmat like I said, I don't think there's a problem in the code per se, since it goes away with spesh disabled
17:17 timotimo that's open now
17:17 jnthn Yeah, looks like
17:18 timotimo uh oh, so it's likely something i messed up :S
17:18 jnthn Maybe you can unupmess it? :-)
17:18 timotimo downmess?
17:18 jnthn As long as you're careful not to just sidemess it :)
17:19 timotimo anyway, i'm glad about the 5% object creation speedup :)
17:20 rindolf joined #perl6
17:20 timotimo lizmat: when you reverted attrinited in Hash<a>, did that throw out a big speedup we got before? :(
17:20 lizmat no, not all of it, just a bit
17:21 * timotimo is trying to see the lennart pöttering keynote of fosdem and is having a really hard time
17:21 timotimo seems like they used the microphone on the camera
17:21 timotimo so you can hear the audience at about 50% volume relative to the speaker himself
17:21 timotimo and there's sufficient reverb to make listening really strenuous
17:21 timotimo i really hope perl devroom audio isn't as terrible
17:22 * lizmat is glad Tom videod all Perl devroom presentations with 2 camera's from close up front
17:22 timotimo that sounds good
17:22 timotimo good thinking, Tom!
17:22 lizmat (not using the FOSDEM equipment, I might add)
17:22 timotimo who is tom again?
17:22 timotimo browder?
17:22 lizmat Tom Bloor, see http://blogs.perl.org/users/shadowcat_mdk/2016/02/fosdem-2016.html
17:23 skids https://github.com/skids/perl6-Control-Bail now updated to not stack up closures accidentally.
17:24 skids Bickeshed still open, and I have no clue about some internals stuff so an eyeball by someone who does may point out something I did wrong.
17:24 dalek doc: f306dfc | (Steve Mynott)++ | bin/p6doc:
17:24 dalek doc: Use CompUnit::DependencySpecification to find pod file niner++
17:24 dalek doc: Specify -IFakeDir to prevent 'shell perl6..." blocking on a precomp write lock ugexe++
17:24 dalek doc: review: https://github.com/perl6/doc/commit/f306dfc301
17:26 skids Implementing slang actions method in perl rather than nqp had some ... interesting ... nuances.
17:31 timotimo lizmat: what's the difference between 6 and 6l on your machine?
17:31 timotimo oh, probably -Ilib?
17:31 lizmat alias 6='time perl6 -e'
17:31 lizmat alias 6l='perl6 --ll-exception -Mstrict -e'
17:31 timotimo nah, --ll-exception
17:31 lizmat hmmm...  the strict is apparently a leftover
17:32 timotimo okay, the first step, reproducing the error locally, is done
17:32 lizmat cool!
17:33 rpburkholder joined #perl6
17:34 lizmat timotimo: if I use a temp str for that if {}, the problem goes away...
17:34 lizmat closing in on it further
17:35 ely-se joined #perl6
17:36 lizmat if the if uses an nqp::existskey, it is also ok
17:36 ugexe m: my %a = :a(1), :b<2>, :c<3>; for %a -> $x { once { %a<b>:delete; %a<c>:delete; }; say $x.perl; };
17:36 camelia rakudo-moar 891d45: OUTPUT«:a(1)␤:sub(method perl (Mu:U $: *%_) { #`(Method|70481040) ... })␤:signature(:(Mu:U $: *%_))␤:num_types(1)␤:types((Mu,))␤:sub(method perl (Mu:U $: *%_) { #`(Method|70481040) ... })␤:constraints(())␤Cannot find method 'WHICH': no method cache …»
17:36 lizmat timotimo: for the bindattr_n case
17:37 timotimo hum, the block for bindattr_n seems to be invoked rather than inlined
17:38 lizmat yeah, it looks like adding just another line of dummy code in that block, fixes the problem
17:38 araujo joined #perl6
17:39 timotimo hm, let me do a check over our jit just to be safe
17:39 stmuk_ does p6doc work on windows now?
17:39 lizmat timotimo: just adding "my $something;" to the if block, fixes the issue
17:39 timotimo please verify that turning off the jit makes the problem go away?
17:40 lizmat MVM_JIT_DISABLE=1 right ?
17:40 araujo 春节快乐!
17:40 araujo :)
17:40 timotimo ah, yes. i think i see it.
17:41 Ben_Goldberg joined #perl6
17:41 magiczna_dupa joined #perl6
17:42 lizmat timotimo: indeed, with JIT disabled, it isn't a problem
17:42 timotimo false alarm :|
17:44 pmurias lizmat: are the Tom Bloor videos up?
17:44 lizmat not yet, afaik
17:46 timotimo i've found the line in the jit that fixes the problem by getting commented out
17:50 magiczna_dupa Im reading some module code and cannot find what mean the first line
17:51 lizmat magiczna_dupa: what is that line ?
17:51 magiczna_dupa unit role HTTP::Easy:ver<2.1.3>:auth<http://supernovus.github.com/>;
17:51 magiczna_dupa unit role HTTP::easy
17:51 timotimo i see now.
17:51 magiczna_dupa in import manual
17:52 magiczna_dupa is smh like scope
17:52 Skarsnik unit say that the fichier is what you declare after
17:52 magiczna_dupa in which you can export
17:52 geekosaur "unit role" - the whole file is a role
17:52 Skarsnik it avoid putting for example role R { } around everything
17:52 magiczna_dupa ohh
17:53 magiczna_dupa and why the role got that strange name
17:53 magiczna_dupa with version and auth
17:53 Skarsnik it's the full declaration if you want to add version and auth x)
17:53 magiczna_dupa but these variables ver and auth are unused?
17:54 Skarsnik for now I think yes?
17:54 geekosaur you can specify them when importing
17:54 geekosaur hm, for this maybe not
17:54 magiczna_dupa the same is with unit module? we write it instead wrapping all file into module
17:54 geekosaur yes
17:54 perlpilo1 magiczna_dupa: at some pont in the future you could use a role with the same name but be a different version or have a different auth in the same scope as that role
17:54 timotimo lizmat: just pushed to moar; can you try that out?
17:54 geekosaur because "role <whatever>;" (or "module", etc.) is a forward declaration
17:55 lizmat timotimo: in a momment
17:55 geekosaur auth will let you say *whose* HTTP::Easy you want, rather than forcing someone else's to take a different name
17:57 magiczna_dupa geekosaur, yeah but in use case we got use HTTP::Easy::PSGI;
17:57 dalek rakudo/nom: 7e25273 | lizmat++ | src/core/Mu.pm:
17:57 dalek rakudo/nom: Streamline BUILDALL for 8% faster object creation
17:57 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7e25273b43
17:57 dalek rakudo/nom: 5605d5f | lizmat++ | src/core/Mu.pm:
17:57 dalek rakudo/nom: Add dummy code to work around JIT issue
17:57 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5605d5f343
17:58 timotimo was the dummy code still necessary?
17:59 sjoshi joined #perl6
17:59 lizmat it was necessary without the Moar patch
18:00 lizmat I put it in a separate commit so that it's easily reverted
18:00 timotimo OK, so it will go away when MOAR_VERSION gets bumped
18:00 magiczna_dupa how to run tests in module?:D
18:00 lizmat timotimo: yup, that's the idea
18:00 timotimo the reason why the error occured was i thought "if we pass in the value and it's a float, we need the different calling convention!", but this really depends on the function being called, and this function declares the argument as MVMRegister
18:01 timotimo so when i checked for bindattr_n and used REG_VAL_F instead of REG_VAL, it ended up putting the value argument on a separate argument "stack", and the kind value slipped into the value value and it took whatever was in the register at the time for the kind instead
18:01 timotimo which ended up being not the right kind in this case
18:02 lizmat so it bombed: but in other cases, it would have just done the wrong thing, right ?
18:02 timotimo yes, it would have silently interpreted the kind as an int or pointer to a string
18:02 timotimo er, no
18:02 timotimo only if it was a bindattr_n or bindattrs_n
18:03 dalek perl6.org: eacd147 | (Steve Mynott)++ | source/downloads/index.html:
18:03 dalek perl6.org: Rakudo Star 2016.01 Errata
18:03 dalek perl6.org: review: https://github.com/perl6/perl6.org/commit/eacd1478c3
18:05 rpburkholder joined #perl6
18:06 timotimo any bindattr_n or bindattrs_n that ended up being jit compiled would either bomb because it didn't have the right value in the register that ends up being interpeted as "kind", or it'd have the "num64" kind in the "kind" parameter by sheer luck and set the same value every time
18:08 llfourn m: use parameters :D; # I want to have something like this. Is there any reason I shouldn't attempt to implement it?
18:08 camelia rakudo-moar 891d45: OUTPUT«5===SORRY!5=== Error while compiling /tmp/WbanWtsNzX␤use parameters not yet implemented. Sorry. ␤at /tmp/WbanWtsNzX:1␤------> 3use parameters :D7⏏5; # I want to have something like this. ␤»
18:08 timotimo llfourn: it can go in the ecosystem, in which case it's absolutely fine
18:09 llfourn timotimo: It would be quite hard to do from the ecosystem and it's NYI in rakudo so...?
18:09 timotimo hm, do we want to offer it in rakudo?
18:09 lizmat timotimo: confirm the problem is gone with most recent Moar
18:09 domidumont joined #perl6
18:09 maybekoo5 joined #perl6
18:10 timotimo lizmat: thanks. there definitely was a problem, just not sure it was 100% what caused this %)
18:10 * lizmat runs spectest to check for any other issues
18:10 lizmat hmmm... quite some Inline::Perl5 breakage now
18:11 timotimo huh!
18:11 llfourn timotimo: well at one point we did, but it never got finished. I'm not sure whether there was a good reason for that or just not enough time.
18:11 timotimo that's not good
18:12 lizmat $ perl6 t/spec/S01-perl-5-integration/basic.rakudo.moar
18:12 lizmat 1..15
18:12 lizmat moar(91724,0x7fff7a3f6300) malloc: *** error for object 0x10: pointer being freed was not allocated
18:12 timotimo hopefully it's not my patch that causes this?
18:12 lizmat no, I'm more expecting jnthn's recent moar patches
18:13 timotimo huh
18:13 lizmat that's why I put the temporary fix already in, as it will be some days before we can bump Moar
18:13 llfourn http://www.nntp.perl.org/group/perl.perl6.language/2015/10/msg36477.html # "FWIW you can now (as of a few days ago) control the default with
18:13 llfourn use invocant :D;"
18:13 timotimo right
18:14 geekosaur didn't that just get reverted?
18:14 lizmat llfourn: that was removed before 6.c
18:14 lizmat because it wasn't working correctly just yet
18:14 geekosaur [06 12:43] <dalek> MoarVM: Revert "Fix leak involving inlining handlers."
18:14 lizmat and we didn't want an incorrectly working version out there  :-)
18:14 geekosaur from #moarvm
18:15 llfourn lizmat: mmm so it's a good time to get it working now?
18:15 timotimo geekosaur: there were multiple patches, not all were reverted
18:15 llfourn (I'm going to attempt it if it is)
18:15 lizmat well, it would probably need to live in the .d setting
18:15 timotimo it's not really a setting thing, isn't it a grammar thing?
18:16 llfourn ah yeah didn't think about versioning
18:17 lizmat yeah, it's mostly a grammar / World thing, I think
18:18 timotimo now that we're promising 6.c, we implementors will be tortured a bunch more than we used to be :)
18:18 llfourn the NYIness comes from World
18:19 llfourn do we have a way to mark patches as .d or have a .d branch?
18:19 timotimo .c and .d live in the same codebase
18:19 yoleaux I don't know
18:19 llfourn (or spectest for that matter)
18:20 llfourn ah yeeh...
18:20 psy_ joined #perl6
18:21 timotimo otherwise we couldn't support "use 6.c" and "use 6.d"
18:21 llfourn are those version checking contructs available yet for the grammar?
18:21 timotimo i don't know :(
18:21 timotimo BBIAB
18:21 * lizmat is waiting for jnthn's blogpost about that  :-)
18:22 * llfourn adds himself to the waiting queue
18:22 llfourn ok i'll hold off on this then. Thanks.
18:24 dha joined #perl6
18:33 kid51 joined #perl6
18:36 magiczna_dupa joined #perl6
18:39 rpburkholder joined #perl6
19:01 vytas joined #perl6
19:02 dalek rakudo/nom: 463b6c8 | lizmat++ | src/core/Mu.pm:
19:02 dalek rakudo/nom: Give BUILD_LEAST_DERIVED the same treatment
19:02 dalek rakudo/nom:
19:02 dalek rakudo/nom: Not sure what this speeds up, but it should be again in the 5-10% range
19:02 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/463b6c8586
19:03 lizmat afk&
19:07 dha I've lost track, is there a comprehensive list of Perl 6 types? maybe broken down by category?
19:11 magiczna_dupa hey I am playing with perl6 modules
19:12 geekosaur dha, http://doc.perl6.org/type.html ?
19:12 magiczna_dupa and I don't get if I created module and use it
19:12 magiczna_dupa do I need to export sub or variable
19:12 magiczna_dupa to use it
19:13 DrForr That should be in the documentation.
19:13 magiczna_dupa http://doc.perl6.org/language/modules#Exporting_and_Selective_Importing
19:13 magiczna_dupa reading it
19:13 DrForr The documentation of the module, actually.
19:14 dha geekosaur - ah, right, of course. But not broken down by category.
19:14 magiczna_dupa DrForr, they write to place is export
19:14 magiczna_dupa then use Module; import default symbols
19:15 magiczna_dupa but as I check my example won't work
19:16 DrForr magiczna_dupa: Please nopaste or gist your code, I don't want to guess at what you're doing.
19:18 magiczna_dupa DrForr, http://paste.ofcode.org/ZpeJpZsc6TTV8uGyaSTpjK
19:20 DrForr What's the module filename? It should be SampleModule.pm or SampleModule.pm6
19:20 RabidGravy the stub sub with the export won't work, just but the is export on the sub, in this case you don't need the stub at all
19:21 RabidGravy m: module F { sub foo() is export { say "foo" }; import F; foo()
19:21 camelia rakudo-moar 463b6c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ou_84OANuM␤Missing block␤at /tmp/ou_84OANuM:1␤------> 3is export { say "foo" }; import F; foo()7⏏5<EOL>␤    expecting any of:␤        statement end␤        statement modifier␤        stateme…»
19:21 RabidGravy m: module F { sub foo() is export { say "foo" }}; import F; foo()
19:21 camelia rakudo-moar 463b6c: OUTPUT«foo␤»
19:21 RabidGravy the bare import there is because it can already see the module
19:22 RabidGravy if it was in another file you would use "use" instead
19:23 magiczna_dupa the filename is like module name
19:23 jnthn lizmat: Do you still get that build SEGV on Moar HEAD?
19:23 dha I should warn you all that we have about 16 people in a bar here working on learning Perl 6.
19:23 dha :-)
19:23 RabidGravy in your example also SampleModule::suck() won't work because the subroutine is lexically scoped by default so isn't accessible that way via the package
19:23 DrForr +=17
19:24 magiczna_dupa RabidGravy, what? what does it mean lexically scoped?
19:24 dha DrForr - also I've pointed people at your Grammar tutorial.
19:24 magiczna_dupa RabidGravy, it is private?
19:24 magiczna_dupa RabidGravy, scoped to current file?
19:25 molaf joined #perl6
19:25 RabidGravy it can't be seen outside the scope it is defined in, in the case the module { ... }
19:25 DrForr dha: Thanks. It's just episode 1.
19:25 jnthn m: role Perl6Learner { method goto-bar() { } }; role Gymnast { method goto-bar() { } }; class FlexibleProgrammer does Perl6Learner does Gymnast { }
19:25 camelia rakudo-moar 463b6c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/6epHDQWEap␤Method 'goto-bar' must be resolved by class FlexibleProgrammer because it exists in multiple roles (Gymnast, Perl6Learner)␤at /tmp/6epHDQWEap:1␤»
19:25 jnthn dha: Mini lesson on roles ;) ^^
19:25 dha I know. But we're here, so I figured I'd tell them about it. At least they know where to look for parts 2 and 3. :-)
19:26 DrForr I'll keep an eye on stats :)
19:26 dha jnthn - The error? :-)
19:26 jnthn magiczna_dupa: Lexical = scoped to the current code block (set of curly braces)
19:26 jnthn dha: Shows compile-time detection of conflicting methods in roles :)
19:27 dha *nod*
19:27 RabidGravy jnthn, I have a strange LTA regarding "must be resolved by class"
19:28 magiczna_dupa jnthn, have no idea how to change it
19:28 * DrForr cracks a bottle of framboise...
19:29 jnthn magiczna_dupa: Either by export/import, or just write our sub suck() ...
19:29 jnthn (The our means package scope)
19:29 jnthn RabidGravy: Whcih is?
19:29 jnthn *which
19:30 magiczna_dupa jnthn, yeah our is like global scope
19:30 magiczna_dupa and how to fix it by export?
19:31 magiczna_dupa I did is export and test export(:MANDATORY)
19:31 magiczna_dupa but it won't go
19:31 jnthn m: package Foo { sub bar() is export { say 'like this maybe' } }; import Foo; bar()
19:31 camelia rakudo-moar 463b6c: OUTPUT«like this maybe␤»
19:31 RabidGravy it's difficult to golf, but basically if I have a class that does two roles A and B with the same method foo, so in the class I do "method foo() { self.B::foo } " then in B I do "method foo() { self.A::foo }" it blows up with the "cannot concretise"
19:31 jnthn RabidGravy: Uh, does B do A?
19:32 RabidGravy no, but in the code it will only ever be applied to an A
19:32 autarch a fun error I cannot understand - https://gist.github.com/autarch/3c38c878972070310769
19:32 RabidGravy ah
19:33 jnthn I don't think that'll work out too well, 'cus roles are generic.
19:33 autarch This comes from trying to EVAL a string containing something serialized via .perl in another thread
19:33 jnthn On the type of the invocant
19:33 jnthn And they get made concrete at composition time
19:33 RabidGravy https://github.com/jonathanstowe/Sofa/blob/master/lib/Sofa/Document/Wrapper.pm
19:34 jnthn When we need to resolve a self.SomeRole:foo, then we look at all the roles we've made concrete, so we can call a real, instantiated method
19:34 RabidGravy let me see what happens
19:35 RabidGravy so it won't matter if I have a class that *must* do A get the role B which also has A ?
19:35 RabidGravy if you see what I mean
19:35 jnthn Think that should be fine.
19:36 dalek rakudo-star-daily: ce34419 | coke++ | log/ (9 files):
19:36 dalek rakudo-star-daily: today (automated commit)
19:36 dalek rakudo-star-daily: review: https://github.com/coke/rakudo-star-daily/commit/ce344196e5
19:36 dalek perl6-roast-data: 4cdbcc7 | coke++ | / (6 files):
19:36 dalek perl6-roast-data: today (automated commit)
19:36 dalek perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/4cdbcc7de8
19:36 dalek perl6-roast-data: 26f9ecd | coke++ | / (6 files):
19:36 dalek perl6-roast-data: today (automated commit)
19:36 dalek perl6-roast-data: review: https://github.com/coke/perl6-roast-data/commit/26f9ecd3b3
19:36 jnthn autarch: Looks like an internal compiler error, but hard to guess much more from that from the stack trace I'm afraid
19:36 autarch jnthn: well, I'm quite sure it comes from an EVAL call - but beyond that I'm not sure why it'd happen
19:36 RabidGravy m:  role A { }; role B does A { };   class C does B does A {}'
19:36 camelia rakudo-moar 463b6c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/6odFLYAzv8␤Strange text after block (missing semicolon or comma?)␤at /tmp/6odFLYAzv8:1␤------> 3B does A { };   class C does B does A {}7⏏5'␤    expecting any of:␤        infix␤        infix s…»
19:36 autarch I'll try to reduce it something reproducible with less code
19:37 RabidGravy m:  role A { }; role B does A { };   class C does B does A { }
19:37 camelia rakudo-moar 463b6c: ( no output )
19:37 jnthn autarch: The stack trace does match up with being in an EVAL.
19:37 RabidGravy well it compiles
19:37 jnthn autarch: I just can't intuit a lot more from it than that.
19:37 autarch yeah, I get that
19:39 wamba joined #perl6
19:40 ely-se joined #perl6
19:50 pdcawley joined #perl6
19:51 dha Have I just forgotten how to get a list of methods an object can do, or can you no longer do that since I last looked?
19:52 dha ah. found it.
19:52 * dha slaps self
19:52 moritz for the record: .^methods or .^methods(:local) or .^methods(:all), http://doc.perl6.org/routine/methods
19:53 DrForr Should be n the metamodel doc?
19:53 pmurias hoelzro: they still haven't merged your node-bigint pull request :(
19:55 magiczna_dupa hey I find next magic code for me :d
19:55 magiczna_dupa $!connection.recv( :bin )
19:56 moritz magiczna_dupa: which part of that don't you understand?
19:56 magiczna_dupa :bin
19:57 moritz it's a named argument to the recv method
19:57 dha Yeah, .^methods should probably be documented somewhere. Particularly as mop.pod indicates that most metamethods are all caps.
19:57 moritz and for recv, it means "return binary data"
19:58 moritz dha: meta method != method on a meta object
19:58 johnb_ joined #perl6
19:58 magiczna_dupa moritz, how?
19:58 magiczna_dupa method recv(IO::Socket:D: Cool $elems = Inf)
19:58 musiKk joined #perl6
19:58 dha moritz -  ok. Still, it should be documented somewhere. :-)
19:59 moritz dha: it is, see http://doc.perl6.org/routine/methods
19:59 bpmedley joined #perl6
19:59 magiczna_dupa I cant find whats IO::Socket:D but I suppose it is socket value
19:59 magiczna_dupa treated as Cool number or str
19:59 moritz magiczna_dupa: it's on IO::Socket::Async, I presume
19:59 Guest45613 Can someone recommend a good book that covers Perl 6's concurrency, parallelism, asynchrony?
20:00 moritz magiczna_dupa: ah no, the documentation of recv is just incomplete
20:00 moritz Guest45613: no books yet
20:00 jnthn IO::Socket::Async has no .recv method, fwiw
20:00 dha moritz -  ok, so without that link, and not knowing it to begin with, how would someone find it?
20:00 magiczna_dupa moritz, -.-
20:00 Guest45613 darn. ok
20:00 dha Maybe it should be a FAQ. Seems like something people would want to do, particularly if they're learning the language.
20:00 magiczna_dupa how this guy knows that recv got more arguments?
20:00 moritz Guest45613: https://perl6advent.wordpress.com/2015/12/14/day-14-a-nice-supplies-syntactic-relief-for-working-with-asynchronous-data/ is a nice intro though
20:01 magiczna_dupa dha, I was learning from perl6 intro
20:01 moritz magiczna_dupa: it's probably in the design document, or they read the implementation (which isn't too hard, considering that it's written in Perl 6 itself)
20:01 dha ok.
20:01 moritz dha: one could have thought "oh, I want a list of methods, so let's type "method" in the search box on doc.perl6.org and see what comes up"
20:01 Guest45613 thanks moritz, will have a read.  i looked at the talk in http://act.yapc.eu/ye2015/talk/6390 which got me really keen on learning more
20:02 moritz Guest45613: also http://doc.perl6.org/language/concurrency
20:02 Guest45613 ok
20:02 magiczna_dupa moritz, and as you said it is named argument
20:02 magiczna_dupa where then it is defined
20:02 magiczna_dupa it rather should me
20:02 magiczna_dupa recv(bin => True)
20:02 magiczna_dupa to flag binary mode
20:02 dha moritz -  yep, and that gives you a ridiculously long list of things. And if you just hit enter it takes you to the Method class.
20:03 ugexe https://github.com/rakudo/rakudo/blob/nom/src/core/IO/Socket.pm#L10
20:03 moritz magiczna_dupa: :bin is short for bin => True
20:03 magiczna_dupa moritz, oh ok
20:04 magiczna_dupa moritz, how can I fix documentation?
20:04 RabidGravy I'm just fixing it
20:04 Vitrifur joined #perl6
20:04 dalek doc: 6c23c2b | moritz++ | doc/Type/IO/Socket.pod:
20:04 dalek doc: IO::Socket.recv: document :bin
20:04 dalek doc: review: https://github.com/perl6/doc/commit/6c23c2b9b8
20:04 RabidGravy okay moritz was just fixing it :)
20:05 magiczna_dupa oh :D
20:05 moritz magiczna_dupa: but in general, it's all on github; each doc page has a link to the repo at the bottom
20:05 moritz magiczna_dupa: and you can send pull requests, or tell me your github ID, then I can invite you to the perl6 org
20:05 magiczna_dupa ohh so I need just PR
20:06 magiczna_dupa moritz, ok thanks :)
20:07 awwaiid joined #perl6
20:09 RabidGravy Dear couchdb, why is your "update function" interface so flexible, it's a right pain to test against
20:09 Hotkeys Lol
20:10 frobisher joined #perl6
20:10 jdv79 too much surface area?
20:10 RabidGravy it can do an update by PUT or POST, with a document ID or without a document ID it can take query parameters in the URI and body content of either form data or a json document
20:10 jdv79 so yes
20:11 Vitrifur joined #perl6
20:11 Skarsnik have fun RabidGravy
20:11 frobisher grr.
20:12 RabidGravy comings and goings
20:15 dalek doc: 7950162 | moritz++ | doc/ (2 files):
20:15 dalek doc: Document .^compose (at least for classes). Closes #360
20:15 dalek doc: review: https://github.com/perl6/doc/commit/7950162ef9
20:15 moritz github++ # "Closes #\d+" actually closing the issue
20:16 RabidGravy yeah I love that
20:16 yqt joined #perl6
20:17 dbrunton joined #perl6
20:17 RabidGravy I kept seeing the "Closed by a commit" in issues and it took me a while to work out what was going on
20:17 dalek doc: e4f40e6 | moritz++ | doc/Language/phasers.pod:
20:17 dalek doc: Language improvements in phasers.pod
20:17 dalek doc:
20:17 dalek doc: [Coke]++, Closes #381
20:17 dalek doc: review: https://github.com/perl6/doc/commit/e4f40e6e43
20:18 moritz m: say 'abc'.Uni
20:18 camelia rakudo-moar 463b6c: OUTPUT«Method 'Uni' not found for invocant of class 'Str'␤  in block <unit> at /tmp/2eU4HYU4wf line 1␤␤»
20:18 moritz m: say Uni('abc')
20:18 camelia rakudo-moar 463b6c: OUTPUT«Cannot find method 'Uni'␤  in block <unit> at /tmp/MPPwRgJrSx line 1␤␤»
20:19 moritz how do I creat an Uni?
20:19 Hotkeys Uni?
20:19 moritz Codepoint-level string (without Grapheme magic)
20:19 RabidGravy get some books and invite some students round
20:20 jnthn moritz: You coerce to one of the subclasses
20:20 jnthn moritz: Depending which NF* you want
20:20 moritz jnthn: ah
20:20 jnthn .NFC, NKFC, .NFD, etc.
20:20 moritz m: say 'abc'.NFC
20:20 camelia rakudo-moar 463b6c: OUTPUT«NFC:0x<0061 0062 0063>␤»
20:20 RabidGravy so 'CONTROL' and 'CATCH' phasers  won't work if added by .add_phaser right?
20:20 magiczna_dupa hey how can I test perl module?
20:21 magiczna_dupa as I checkout one and want to fix some things in it
20:21 jnthn RabidGravy: No, they involve some quite convulated code-gen
20:21 moritz magiczna_dupa: http://doc.perl6.org/language/testing
20:21 RabidGravy m: use Test;  ok 1;
20:21 camelia rakudo-moar 463b6c: OUTPUT«ok 1 - ␤»
20:21 RabidGravy there goes my plans for world domination
20:21 magiczna_dupa great ;-)
20:23 dalek doc: ba18428 | moritz++ | util/new-type.p6:
20:23 dalek doc: Fix new-type util
20:23 dalek doc: review: https://github.com/perl6/doc/commit/ba18428174
20:24 RabidGravy magiczna_dupa, if you are looking at making a module, as well as reading the great documentation you might want to peruse some of the fabulous modules that already exist to see how they work
20:24 virtualsue joined #perl6
20:28 PerlJam .oO( but ... but ... which ones are fabulous? )
20:29 Hotkeys All of them of course
20:29 RabidGravy all of them if people have been bothered to make them they are fabulous
20:30 magiczna_dupa RabidGravy, yeah sure thats what I am doing right now
20:30 RabidGravy I'm not convinced that Sofa is going to be fabulous however as I am rapidly losing the will to live writing the tests
20:32 magiczna_dupa RabidGravy, I think the good web framework should be created then maybe someone will start to use perl6 commercially
20:32 magiczna_dupa :P
20:32 RabidGravy why do you need a web framework to use it commercially?
20:32 magiczna_dupa cause web is popular ;D
20:32 moritz people are already using Perl 6 commercially
20:34 dha Oh?
20:34 magiczna_dupa ?
20:34 magiczna_dupa moritz, thats not true
20:34 magiczna_dupa ;D
20:34 DrForr And you know this ... how?
20:34 DrForr (not to moritz, sorry...)
20:35 magiczna_dupa DrForr, I was looking for job and no-one exist and more my cofriends in company don't like Perl
20:36 mst currently, where perl6 is being used commercially, it's because a developer wrote something in perl6 that turned out to be useful
20:36 moritz magiczna_dupa: masak and jnthn have repeatedly used Perl 6 for their work, for example. As did colomon
20:36 RabidGravy so that sample == "no one"
20:36 mst this is the standard way very early adoption happens
20:36 RabidGravy weird
20:36 colomon “does” colomon.  ;)
20:36 moritz mst: that's how it always starts
20:36 moritz colomon++
20:37 DrForr Sigh. Obviously a sample size of one proves a universal.
20:37 magiczna_dupa as I write swiss tool in Perl the guy which sit next to me said that he used perl last time in 2009 year and he won't to see moreover
20:39 moritz magiczna_dupa: this just proves that "the guy" doesn't use and like Perl, nothing more, nothing less
20:39 mst eh. most of my commercial code is perl5 and none of my customers are trying to replace it with something else
20:39 DrForr And who knows where he encountered Perl and what shape the code was in, and a *host* of other factors come into play, but since you've already made up your mind there's no point arguing.
20:40 mst magiczna_dupa: basically, I don't disagree that perl6 jobs aren't currently really a thing in terms of job adverts, but that doesn't mean people aren't using it at work at all.
20:41 mst DrForr: I don't think that's really fair. he's in an environment where lots of people are down on perl, and that's going to affect his perception. availability bias is a thing.
20:42 wamba joined #perl6
20:42 RabidGravy to be honest I've seen Perl 5 code that made me never want to see *software* again in the past.  But hey we won't speak ill of defunct companies.
20:42 mst I've seen perl5, C, javascript, make, sh, csh, python, and ruby code that all did that
20:42 mst and that's only off the top of my head
20:43 magiczna_dupa the thing which perl need it is something which already got a newcomer langs
20:43 magiczna_dupa like swift
20:43 RabidGravy yeah pick a language any language
20:43 magiczna_dupa good ide
20:44 DrForr mst: Yes, availability and selection bias, I do understand these principles. I'm just saying that the small sample size of people who have had contact with perl in no way generalizes to a sweeping statement that there is no commercial work for perl.
20:44 mst yes. but his actual thesis was "there's no jobs being advertised for it"
20:44 magiczna_dupa and onward syntax check
20:44 mst I believe pointing out that that can be true, and 'no commerical work' still be false, is more constructive
20:45 mst magiczna_dupa: in general, everybody thinks 'perl needs' whatever tool they most prefer using. then they don't write it, and wonder why the people who don't haven't written it for them yet
20:46 RabidGravy :)
20:46 magiczna_dupa mst, ye sure :-)
20:46 magiczna_dupa mst, ppl always says like that I know
20:47 ely-se joined #perl6
20:47 DrForr Yes, it is entirely true that there are few to no jobs advertised for it. consider the fact that I can count the number of weeks since it's been released on one hand. Surely that has *some* impact, if only because HR needs at least a *few* months' lead time to write ads.
20:47 RabidGravy and work out how many years of experience they can demand
20:48 mst DrForr: right. so I'm not sure why you think ranting at the user and saying they'll never change their mind is going to be more useful than pointing out they've missed the point :)
20:48 DrForr Mind you I imagine it'll take HR only a few months to demand workers with 5 years of experience.
20:48 RabidGravy it was a running joke back in the 2001-2 when they caught on to C# and were asking for five years of experience
20:48 mst corporate HR does tend to be hilarious
20:48 mst provided you like black comedy
20:49 DrForr I've just seen the pattern before, figured I'd get ahead of the curve :/
20:49 dalek doc: 76663f4 | moritz++ | / (6 files):
20:49 dalek doc: Basic documentation for Uni, NF{K,}{C,D}
20:49 dalek doc: review: https://github.com/perl6/doc/commit/76663f4aa9
20:49 AlexDaniel joined #perl6
20:49 jnthn .oO( I nearly already do have "10 years Perl 6 experience"... )
20:49 RabidGravy I dunno I most recently worked for a large financial institution, their HR was more like something Kafka may have come up with
20:49 * DrForr gets back to getting something done and kicking around a (bloody) web framework design.
20:50 RabidGravy well I guess I could wing it as I'm pretty sure I have patches in parrot more than ten years ago
20:51 dha Getting something done? What's that like?
20:51 magiczna_dupa DrForr, are you preparing web framework?
20:51 skids https://plus.google.com/+CybercitiBiz/posts/eyKgFkaaVFt
20:51 mst dha: almost, but not entirely, quite unlike today
20:51 DrForr I released the second version of Parrot, does that count? :)
20:51 dha Ah.
20:52 magiczna_dupa DrForr, I don't know what Parrot is ;d
20:52 DrForr magiczna_dupa: There's already Bailador, but I'm not going to restart that discussion.
20:53 magiczna_dupa DrForr, ye I see Bailador
20:53 wamba joined #perl6
20:54 dalek doc: bbcf876 | moritz++ | doc/Type/Str.pod:
20:54 dalek doc: Document Str.NFD and friends.
20:54 dalek doc:
20:54 dalek doc: Closes #365
20:54 dalek doc: review: https://github.com/perl6/doc/commit/bbcf876d71
20:54 tadzik hey, bailador was not exactly meant to be a production thing :)
20:54 mcmillhj joined #perl6
20:54 tadzik (also, nice nickname, as Rotwang said)
20:54 RabidGravy I think that horse has bolted matey
20:54 magiczna_dupa tadzik, o czesc
20:55 tadzik o/
20:55 * jnthn was trying not to say anything about the nick... :)
20:57 moritz http://doc.perl6.org/language/faq#Is_there_going_to_be_a_Perl_6_version_6.0.0%3F should we change that answer to "no, it's called v6.c"?
20:57 AlexDaniel “yes, it's called v6.c”?
20:58 magiczna_dupa hey is it okay to write one method over 100+ lines
20:59 magiczna_dupa :D
20:59 xenu usually no
20:59 magiczna_dupa I am looking for crazy module right now
20:59 magiczna_dupa ;s
20:59 stmuk_ the question probably should be "Is there a Perl 6 version 6.0.0?"
21:00 AlexDaniel magiczna_dupa: depends
21:01 stmuk_ and the answer "well yes and no. next question please?"
21:01 mst you *can* write a method that long, but please don't :)
21:02 AlexDaniel mst: Dunno, what if it has several lexical subs? :)
21:02 DrForr magiczna_dupa: Slang::Roman :)
21:02 dalek doc: e11e81a | moritz++ | doc/Language/faq.pod:
21:02 dalek doc: The FAQ answer for "Is there going to be a Perl 6 version 6.0.0?" is much simpler now
21:02 dalek doc: review: https://github.com/perl6/doc/commit/e11e81a3e4
21:02 ab5tract multi method dispatch with where constraints have definitely reduced the justifiability of 100+ methods
21:03 skids 100+ line methods --  100+ multimethods themselves though ... :-)
21:03 AlexDaniel moritz: uhh! Such a pity. It was a great answer
21:04 stmuk_ isn't 6.c intended to be supported by *all* post  2015.12 releases?
21:04 magiczna_dupa if I PR someones code do you think he will apply my work?
21:04 ab5tract skids: it can go the wrong way too, I'm sure. None of my p6 has gotten  that complex yet though..
21:04 magiczna_dupa if it will some more readable fixes
21:04 raiph joined #perl6
21:05 DrForr magiczna_dupa: ONe way to find out :)
21:06 PiL_ joined #perl6
21:06 AlexDaniel ab5tract: I've noticed that p6 code is rather dense. Some really complex stuff can fit into 5 lines. Line count doesn't matter much, I guess
21:08 AlexDaniel I mean bigger line count does not necessarily mean more complexity in Perl 6. But that might be more or less true in other languages
21:08 PerlJam stmuk_: why do you ask?
21:09 dalek perl6.org: 009f20d | (Steve Mynott)++ | source/downloads/index.html:
21:09 dalek perl6.org: suggest deleting p6doc.bat as windows fix for R* 2016.01
21:09 dalek perl6.org: review: https://github.com/perl6/perl6.org/commit/009f20d114
21:09 magiczna_dupa DrForr, also SDL2 bindings author not responds me
21:09 stmuk_ I haven't actually spun up a windows vm to check that but maybe it works :)
21:10 raiph .ask moritz Comments? When discussing versioning of p6doc ab5tract said "it's not Rakudo documentation. It's Perl 6 documentation." But p6doc already has lots of rakudo specific stuff in it, eg the FAQ, and I was about to add more with a Performance page that talks about Rakudo stuff.
21:10 yoleaux raiph: I'll pass your message to moritz.
21:12 RabidGravy I think I've got a 100+ line method somewhere or other but it's probably get a lexical sub and even a complete class definition in it
21:12 DrForr magiczna_dupa: You're taking a sample size of one person that won't respond to PRs and assuming that applies to the entire community. I deal with PRs on a daily basis, it's just a question of how much time that auhor has to devote to a given project.
21:12 AlexDaniel m: say [~] (0..0x1FFFF ==> grep { .uniname ~~ m/‘CUNEIFORM’ .* ‘SQUARED’ / })».chr # O_o
21:12 camelia rakudo-moar 463b6c: OUTPUT«????????????????????????????????????␤»
21:13 mst magiczna_dupa: different people are different.
21:13 mst magiczna_dupa: some authors are busy.
21:13 dalek doc: 49bceab | (Steve Mynott)++ | doc/Language/faq.pod:
21:13 dalek doc: fix typo and tweak 6.0.0 answer
21:13 dalek doc: review: https://github.com/perl6/doc/commit/49bceab6ba
21:13 RabidGravy magiczna_dupa, for reference the longest I waited for a PR to be accepted was eight months
21:15 magiczna_dupa use v6;
21:15 magiczna_dupa does it mean that author used perl interpreter
21:15 AlexDaniel .u ????
21:15 yoleaux U+1F18A CROSSED NEGATIVE SQUARED LATIN CAPITAL LETTER P [So] (????)
21:15 xenu if author is too busy or too lazy, you can always as him for commit bit
21:15 xenu usually they agree
21:15 xenu and fix the issues by yourself
21:15 RabidGravy I get very few PRs I can only assume that's because no-one uses any of my modules
21:15 AlexDaniel xenu: sometimes asking for a commit bit makes them accept your PR :)
21:16 AlexDaniel RabidGravy: or they “just work”
21:16 RabidGravy magiczna_dupa, all that means is that if the script/module/whatever is run like by Perl 5 it will crap out with a meaningful errror message
21:17 RabidGravy similar to:
21:17 RabidGravy m: use v7;
21:17 camelia rakudo-moar 463b6c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/CT5k2bJf6X␤No compiler available for Perl v7␤at /tmp/CT5k2bJf6X:1␤------> 3use v77⏏5;␤»
21:17 magiczna_dupa oh well it should be always used
21:18 RabidGravy rather than blundering on until it gets really confused by the syntax and craps out with some imcomprehensible error message
21:19 RabidGravy also, it can specify minor versions
21:19 RabidGravy m: use v6.c;
21:19 camelia rakudo-moar 463b6c: ( no output )
21:20 mst which can totally happen; we had somebody accidentally try to 'm:' on #perl yesterday and spent five minutes debugging his code as normal perl before we realised
21:20 RabidGravy m: use v6.d;
21:20 camelia rakudo-moar 463b6c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/ADdtSM0sSu␤No compiler available for Perl v6.d␤at /tmp/ADdtSM0sSu:1␤------> 3use v6.d7⏏5;␤»
21:20 AlexDaniel .u ⃞
21:20 yoleaux U+20DE COMBINING ENCLOSING SQUARE [Me] (◌⃞)
21:20 Skarsnik RabidGravy, gptrixie have quite 'big' routine x)
21:20 stmuk_ m: use v6.a
21:20 camelia rakudo-moar 463b6c: OUTPUT«5===SORRY!5=== Error while compiling /tmp/miYRqNtl_Z␤No compiler available for Perl v6.a␤at /tmp/miYRqNtl_Z:1␤------> 3use v6.a7⏏5<EOL>␤»
21:21 stmuk_ its probably not worth retrofitting :)
21:23 Psyche^ joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo