»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:00 Psyche^ joined
raiph m: say ٧၇௭႗ ~~ /7*/ 00:02
camelia rakudo-moar 2ad739: OUTPUT«｢7777｣␤»
raiph m: say ௭ 00:03
camelia rakudo-moar 2ad739: OUTPUT«7␤»
[Coke] lizmat: I run spectests with "prove -v -e t/fudgeandrun t/spec/..."
raiph .u ௭
yoleaux U+0BED TAMIL DIGIT SEVEN [Nd] (௭)
TimToady m: say ٧၇௭႗ 00:07
camelia rakudo-moar 2ad739: OUTPUT«7777␤»
TimToady wow, when did that happen? 00:08
m: say '٧၇௭႗' # just checking
camelia rakudo-moar 2ad739: OUTPUT«٧၇௭႗␤»
TimToady whew!
m: say ⅐ 00:09
camelia rakudo-moar 2ad739: OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at /tmp/N4rih2_AWJ:1␤------> 3say7⏏5 ⅐␤Bogus postfix␤at /tmp/N4rih2_AWJ:1␤------> 3say 7⏏5⅐␤ expecting any of:␤ infix␤ infix stopper␤ pos…»
TimToady aww
00:10 skids joined 00:11 Peter_R left 00:14 tardisx left
[Coke] m: say ¡Aye! 00:15
camelia rakudo-moar 2ad739: OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at /tmp/fpdjFnsGp7:1␤------> 3say7⏏5 ¡Aye!␤Bogus postfix␤at /tmp/fpdjFnsGp7:1␤------> 3say 7⏏5¡Aye!␤ expecting any of:␤ infix␤ infix stopper␤ …»
00:17 tardisx joined
zacts \o/ I just installed perl6 for the first time ever using rakudobrew 00:23
"a journey of a thousand miles begins with a single step..." I guess...
:-)
00:26 mr_ron left, hankache left, hankache joined
timotimo cool 00:27
let us know if you have any questions
00:27 mr_ron joined
zacts sure 00:28
00:29 hankache left
timotimo though i myself will be going to bed soon 00:29
00:52 softmoth_ left 00:53 aaronn left 01:04 perlawhirl left 01:08 raiph left, espadrine left 01:15 mr_ron left 01:16 raiph joined 01:17 mr_ron joined 01:23 sprocket left 01:31 softmoth_ joined 01:42 tardisx left
flussence timotimo / Zoffix : I'd advise not getting too eager with HSTS just yet; I don't think all the *.perl6.org subdomains are set up for HTTPS. 01:45
Zoffix I don't plan on messing with that any time soon :) 01:46
flussence just thought I'd mention it, I learned to be careful with it the hard way :)
Zoffix Sure :)
01:53 [particle] left 01:59 Herby_ joined
Herby_ Good evening, everyone! 01:59
02:06 tardisx joined
Zoffix \o 02:08
02:09 raiph left 02:18 xpen joined 02:19 sufrostico left 02:21 dayangkun joined 02:22 raiph joined 02:23 telex left 02:24 telex joined
llfourn can you say "defined but Nil is ok" with types. Like Maybe[Int] in p6 Type::Tiny 02:25
p5 Type::Tiny*
02:26 raiph left
llfourn m: my Int:D $a = Nil; say $a; 02:26
geraud isn't that Int:U ?
camelia rakudo-moar 2ad739: OUTPUT«(Int:D)␤»
llfourn geraud: but Int:U also matches Int 02:27
02:27 mr_ron left
llfourn m: class A { has Int:D $.a = Nil }; say A.new.a 02:29
camelia rakudo-moar 2ad739: OUTPUT«(Int:D)␤»
llfourn so Nil is ok in any :D type aye....
what if I want to say not an acutal Int that is never undef or Nil :\ 02:30
02:35 Gar joined, tardisx left, Gar left, geraud left 02:36 tardisx joined 02:37 yqt left 02:38 geraud joined 02:39 kid51 joined
Herby_ I'm new to programming, and have dipped my toes into Perl 5. I'm interested in learning Perl 6 02:42
Is there something similar to __DATA__ in perl 6?
llfourn Herby_: I don't think so. There is heredoc stuff though.
docs.perl6.org/language/quoting#Her...s%3A_%3Ato 02:43
02:43 finanalyst joined
llfourn oh and Welcome :) 02:43
Herby_ I'll take a look, and thanks!
finanalyst Hi. trying HTML::Template. Errors on tests when installing. A problem in Grammar.pm:42 02:45
Is this a known problem?
llfourn is trying it now 02:46
finanalyst: it actually worked for me :\
which test?
Zoffix finanalyst, how old is your perl? 02:48
finanalyst Zoffix, perl6 version 2015.09-381-g271e84c built on MoarVM version 2015.09-79-gee9fc2b 02:49
used panda update
llfourn finanalyst: windows?
finanalyst ubuntu 02:51
02:52 raoulvdberge left
llfourn finanalyst: can you do rakudobrew build moar? 02:52
Zoffix finanalyst, I'm on a fork of ubuntu and installs fine. My perl is newer, so maybe try upgrading perl? perl6 version 2015.09-442-ge303970 built on MoarVM version 2015.09-79-gee9fc2b
llfourn this is mine where it works: This is perl6 version 2015.10-220-g4988c70 built on MoarVM version 2015.10-61-g624d504
finanalyst ok will try thanks.
just come back to perl6 after a while. What is best way to upgrade?# 02:53
02:53 kaare_ joined
llfourn finanalyst: use github.com/tadzik/rakudobrew :) 02:53
Zoffix finanalyst, run rakudobrew build moar
finanalyst thanx. trying now 02:54
is there a way for panda to run the tests for a specific module? 02:56
02:57 kid51 left
llfourn finanalyst: you can run tests manually with panda look module_name 02:57
then perl6 -Ilib t/the_test_you want
Zoffix And you can panda look Some::Module to download it and drop a shell into it 02:58
02:58 Herby_ left, tardisx left
finanalyst llfourn and Zoffix, thank you very much. Upgrade solved problem 03:00
03:00 [particle] joined
llfourn \o/ 03:00
Zoffix \o/
finanalyst very impressed how much faster perl6 is now. A joy! 03:01
03:13 roguelazer joined 03:15 tardisx joined, sprocket joined 03:18 sprocket left 03:22 BenGoldberg joined 03:26 yuppie joined, yuppie left, yuppie joined, xpen left 03:29 herby_ joined, yuppie left 03:30 yuppie joined, yuppie left, yuppie joined, yuppie left
herby_ is modules.perl6.org the "CPAN" of perl 6? 03:30
Zoffix herby_, for now, yes
03:31 yuppie joined, yuppie left, yuppie joined 03:32 yuppie left, yuppie joined, yuppie left, yuppie joined
herby_ great, thanks 03:33
Zoffix gives yuppie a brand new CAT5 cable 03:34
03:36 mr_ron joined 03:48 herby_ left, xpen joined 03:53 kaare_ left
awwaiid wait... we have undef AND nil? 03:54
llfourn awwaiid: if you are refering to my confusion we have undef in that (Int) is undefined :)
m: say Int.defined 03:55
camelia rakudo-moar 2ad739: OUTPUT«False␤»
llfourn awwaiid: I read the design docs a bit and it seems that Nil is basically like a Failure 03:56
but without an error message
m: say \(Nil) ~~ :(Int:D)
camelia rakudo-moar 2ad739: OUTPUT«False␤»
llfourn so that won't work 03:57
m: class A { has Int:D $.a = Nil }; say A.new;
camelia rakudo-moar 2ad739: OUTPUT«A.new(a => Int:D)␤»
llfourn but that will work
BenGoldberg Nil is a value which magically disappears in most contexts.
awwaiid here and I was bragging that Int:D would only hold defined ints! 03:58
BenGoldberg It's not quite the same as () in perl5, but it's close.
llfourn awwaiid: it's right to describe it like that
As BenGoldberg++ said it depends on the context
you don't use it like undef in p5 03:59
03:59 tardisx left
BenGoldberg m: say [ 1, 2, Nil, 3 ] 03:59
camelia rakudo-moar 2ad739: OUTPUT«[1 2 (Any) 3]␤»
BenGoldberg m: say [ 1, 2, |Nil, 3 ]
camelia rakudo-moar 2ad739: OUTPUT«[1 2 (Any) 3]␤»
Zoffix Could powers-that-be give awwaiid merge perms to perl6 repos? They've been submitting a ton of PRs for the docs. Would be nice for them not to have to wait for someone to come around and merge. 04:00
BenGoldberg m: say [ 1, 2, Splat, 3 ]
camelia rakudo-moar 2ad739: OUTPUT«5===SORRY!5=== Error while compiling /tmp/uJHCnPiSLp␤Undeclared name:␤ Splat used at line 1␤␤»
Zoffix s/merge perms/commit perms/;
04:00 Actualeyes left
awwaiid Zoffix: I got merge perms yesterday 04:00
Zoffix Oh
awwaiid But I thought that was so I could merge other people's things :)
llfourn "The Nil value represents the absence of a value where there should be one, so it does not disappear in list context, but relies on something downstream to catch it or blow up." 04:01
awwaiid I'm used to the formality of two-human-merges at $work
Zoffix I don't see a perl6 org bande on your profile
awwaiid well. here, lmme merge this and we'll see...
04:01 Actualeyes joined
dalek c: 0a3175f | (Brock Wilcox)++ | doc/Language/syntax.pod:
More syntax.pod additions, especially about operators
 04:01
c: 7951121 | (Brock Wilcox)++ | doc/Language/syntax.pod:
Oops, fix where :D goes
c: e3d793c | (Brock Wilcox)++ | doc/Language/syntax.pod:
Merge pull request #191 from awwaiid/syntax-operator-basics

More syntax.pod additions, especially about operators
awwaiid seems like cheating :)
Zoffix m: my Int $x where * > 3 = 7; $x = 2
camelia rakudo-moar 2ad739: OUTPUT«Type check failed in assignment to $x; expected <anon> but got Int␤ in block <unit> at /tmp/b7TIsQGI0u:1␤␤»
awwaiid m: my Int $x where * > 3 = 7; $x = Nil; say $x 04:02
camelia rakudo-moar 2ad739: OUTPUT«(<anon>)␤»
llfourn qwerky isn't it :)
awwaiid Nil is looking a lot like the new null
llfourn awwaiid: but it's not like null for calling code 04:03
m: sub a (Int:D) { 'win'.say }; a(Nil);
camelia rakudo-moar 2ad739: OUTPUT«Type check failed in binding <anon>; expected Int but got Nil␤ in sub a at /tmp/e9p_wcEG4T:1␤ in block <unit> at /tmp/e9p_wcEG4T:1␤␤»
awwaiid so Zoffix and llfourn -- y'all just want me to go to town on random docs like it was a wiki?
m: my $x = Nil; sub a (Int:D) { 'win'.say }; a($x); 04:04
camelia rakudo-moar 2ad739: OUTPUT«Type check failed in binding <anon>; expected Int but got Any␤ in sub a at /tmp/mLaGBxALXg:1␤ in block <unit> at /tmp/mLaGBxALXg:1␤␤»
llfourn awwaiid: well like it was a very official wiki :D
Zoffix awwaiid, sure
awwaiid do you still want a PR to bounce through for easy undo? I'd still do a PR for anything I thought would be large or conversation-worthy 04:05
llfourn awwaiid: I'd say just go ahead. If you are re-structuring files you might want to PR 04:06
awwaiid k
is there a fast way to verify the pod markup? 04:07
llfourn awwaiid: not that I know of 04:08
awwaiid cool. well I'll keep tinkering :) 04:09
tomorrow. cya &
Zoffix awwaiid, I tend to do PRs only for things I want people to discuss (usually something large that has an impact on something)
llfourn I mean you could just Pod::To::HTML an individual page and see if you can get the CSS...cua
cya*
Zoffix awwaiid, the bot announces the commits, so I think it's pretty noticable what people are committing
bye
awwaiid [from the beyond: awesomeness, thanks!] 04:10
tadzik I just got an email with an invite to, effectivly, job speed dating 04:17
I have to admit, it is kind of cool, spam or not :D
04:17 tardisx joined
Zoffix heh 04:20
04:30 mr_ron left 04:52 BenGoldberg left 04:56 vendethiel joined 04:57 AlexDaniel left
Zoffix Haw! Browsing through pages of stock to find a bg pic for a p6 site and I come across this! :) depositphotos.com/6454910/stock-ill...ector.html 05:03
Hm. It actually sucks when placed on the page :P 05:05
05:19 vendethiel left
TimToady m: say Iterable === Any 05:25
camelia rakudo-moar 2ad739: OUTPUT«Method 'iterator' must be implemented by Iterable because it is required by a role␤ in any compose_method_table at gen/moar/m-Metamodel.nqp:2661␤ in any apply at gen/moar/m-Metamodel.nqp:2671␤ in any compose at gen/moar/m-Metamodel.nqp:2807␤ in…»
05:32 CIAvash joined
dalek href="https://modules.perl6.org/mojo-app:">modules.perl6.org/mojo-app: e4c1974 | (Zoffix Znet)++ | / (14 files):
Rudimentary working app
 05:33
05:47 Khisanth left, Actualeyes left
TimToady m: sub foo(Iterable $s) {...}; say &foo.signature 05:49
camelia rakudo-moar 2ad739: OUTPUT«Method 'iterator' must be implemented by Iterable because it is required by a role␤ in any compose_method_table at gen/moar/m-Metamodel.nqp:2661␤ in any apply at gen/moar/m-Metamodel.nqp:2671␤ in any compose at gen/moar/m-Metamodel.nqp:2807␤ in…»
05:49 colomon left
TimToady that tries to call Iterable.WHICH, and that forces role punning, which doesn't work here 05:50
m: say Iterable.foo
camelia rakudo-moar 2ad739: OUTPUT«Method 'iterator' must be implemented by Iterable because it is required by a role␤ in any compose_method_table at gen/moar/m-Metamodel.nqp:2661␤ in any apply at gen/moar/m-Metamodel.nqp:2671␤ in any compose at gen/moar/m-Metamodel.nqp:2807␤ in…»
TimToady doesn't matter which method you call on it
05:50 Khisanth joined 06:03 pyrimidi_ joined 06:05 khw left, Actualeyes joined, pyrimidine left 06:11 colomon joined 06:14 rindolf joined
zengargoyle Pod variable $=data not yet implemented. Sorry. :( 06:17
TimToady role punning bug filed as #126606 06:24
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126606
06:25 tardisx left 06:28 darutoko joined 06:36 finanalyst left 06:47 softmoth_ left 06:50 xfix joined, nightfrog left 06:52 nightfrog joined 06:58 [Sno] left 06:59 rindolf left 07:02 visiting joined, visiting left
dalek p: 94d5575 | usev6++ | src/HLL/sprintf.nqp:
Handle NaN and Inf in sprintf (directives e, f, g)
 07:07
p: 052389b | FROGGS++ | src/HLL/sprintf.nqp:
Merge pull request #253 from usev6/sprintf_nanorinf

Handle NaN and Inf in sprintf (directives e, f, g)
kudo/nom: 8517d87 | TimToady++ | src/ (4 files):
return typecheck now passes both Nil and Failure

Failure now considers itself to be derived from Nil, and the return typecheck bypass now is based on Nil. The upshot is that Nil is now regarded as the simplest form of failure, and people will be less tempted to use type objects to indicate failure.		 07:08
ast: b6cbc45 | TimToady++ | / (2 files):
baggy assignment fails slightly differently now
07:11 ggoebel2 left 07:13 skids left
dalek kudo/nom: f6d7f4d | TimToady++ | src/vm/jvm/runtime/org/perl6/rakudo/RakOps.java:
fix jvm too for Nil return
 07:13
07:15 ggoebel joined 07:17 CIAvash left, [Sno] joined 07:29 Actualeyes left
dalek ast: 6f6e494 | TimToady++ | S06-advanced/return.t:
ensure return typecheck passes Nil and Failure
 07:29
07:30 geraud left 07:37 eone joined 07:43 _mg_ joined
[Tux] lizmat, T::C broken in master. No time right now to dig 07:49
07:50 Actualeyes joined
[Tux] t/20_file is a short test (before the whole t/* starts barfing) 07:50
07:51 ely-se joined 07:53 _mg_ left 07:56 matt_ left 07:59 matt_ joined, matt_ is now known as Guest67892 08:02 _mg_ joined 08:03 abraxxa joined 08:13 espadrine joined 08:19 azawawi joined
azawawi hi #perl6 08:19
so now officially Perl 6 is broken again on my windows box 08:20
08:20 zakharyas joined
azawawi panda github.com/tadzik/panda/issues/245 08:20
zef github.com/ugexe/zef/issues/71 08:21
azawawi sighs
moritz azawawi: is there travis support for windows? 08:24
azawawi moritz: github.com/travis-ci/travis-ci/issues/216 08:25
moritz: a while back adamk got me accounts on microsoft vm servers (au) 08:26
moritz: to setup testing for padre on different platforms
moritz: as the ticket says it is licensing issues... though i think things should be different with windows 10 going forward 08:28
moritz: given that windows 10 is being handed as free to all windows 7 and 8 users
azawawi starts working on github.com/perl6/DBIish/issues/22
08:29 pippo joined
pippo o/ #perl6 08:29
azawawi pippo: hi :) 08:32
[Sno] moritz: sure I can document the options in Configure.pl 08:49
think I did it for rakudo and/or star ^^
lizmat good *, #perl6! 08:51
pippo lizmat: o/ 08:53
[Sno] moin lizmat :) 08:54
08:54 Actualeyes left 09:00 RabidGravy joined
lizmat [Sno] pippo o/ 09:02
Hotkeys is there a simple way to turn a list into an array 09:03
moritz Hotkeys: yes! 09:04
Hotkeys: assignment to an array variable, and calling .Array
Hotkeys woo
oh
I tried .array
that'll do it
moritz m: say (1, 2, 3).Array.perl
camelia rakudo-moar f6d7f4: OUTPUT«[1, 2, 3]␤»
Hotkeys capitalization is fun
moritz m: say [(1, 2, 3)].perl
camelia rakudo-moar f6d7f4: OUTPUT«[1, 2, 3]␤»
moritz also works
Hotkeys fancy 09:05
moritz (my @ = (1, 2, 3)).perl
m: (my @ = (1, 2, 3)).perl
camelia ( no output )
JimmyZ m: @(1,2,3).perl
camelia ( no output )
JimmyZ m: say @(1,2,3).perl
camelia rakudo-moar f6d7f4: OUTPUT«(1, 2, 3)␤»
moritz JimmyZ: precedence
erm, no
I think @(...) is actually .list
anyway, there's more than one way to do it :-) 09:06
JimmyZ m: say @([1,2,3]).perl
camelia rakudo-moar f6d7f4: OUTPUT«[1, 2, 3]␤»
JimmyZ m: say @((1,2,3)).perl
camelia rakudo-moar f6d7f4: OUTPUT«(1, 2, 3)␤»
Hotkeys I'll probably stick to .Array unless one of the others is more idiomatic 09:08
lizmat m: dd [1,2,3] # most ideomatic ? 09:09
camelia rakudo-moar f6d7f4: OUTPUT«Array $var = $[1, 2, 3]␤»
moritz it depends, really 09:10
if it's a long expression, I tend to favor method calls 09:11
for literals, I prefer [...]
09:11 dakkar joined 09:15 domidumont joined, xfix left
lizmat [Tux]: one commit to T::C, but needs more work 09:21
09:24 kjs_ joined 09:35 woolfy joined
woolfy Question from Perl6 on Facebook: "Is there a plan to include perl 6 in beaker notebook? I heard its grammars are great. " ( beakernotebook.com/ ) I never ever heard of BeakerNotebook. Anybody? 09:36
pink_mist err, I think that's a question better suited for the beakernotebook people 0_o 09:37
or so it seems to me at least
lizmat [Tux] T::C clean again, but with hacks :-( 09:40
[Tux] lizmat++
canary-coalmine
off to $work 09:41
dalek kudo/nom: c9cb7f4 | lizmat++ | src/core/Str.pm:
Fix running with PERL6LIB on JVM

Apparently, you cannot nqp::getattr into a native int on the JVM, even if the attribute in question *is* a native int.		 09:44
lizmat [Tux]: actually, the test showed that the canary has a hidden disease :-) 09:45
I just gave it a shot of penicillin
but it is not a long term solution :-)
moritz isn't there a separate nqp::getattr_i for native ints?
lizmat .tell [Coke] testing on JVM works again 09:46
yoleaux lizmat: I'll pass your message to [Coke].
09:47 eone left
lizmat moritz: I'll try that after I'm done with fitness 09:47
afk for a few hours&
09:50 abraxxa left, abraxxa1 joined 10:03 andreoss joined
azawawi moritz: what's the status of DBIish? Is it actively maintained or not? 10:07
moritz azawawi: it is 10:08
azawawi moritz: it seems to have been written by many authors over the years though since the coding style is not consistent 10:09
moritz azawawi: that's correct
azawawi moritz: with your permissions, im going to start refactoring it :) 10:10
andreoss are we going to have a real DBI any time soon?
10:14 dayangkun_ joined
andreoss i've read somewhere that DBI will be implemented in moarvm but cannot find it now 10:15
moritz azawawi: what do you want to refactor?
andreoss: the plan was to use the JDBC API as the interface between DBI and the DB backends
andreoss: but that's not a task I'm going to tackle, so I don't know when/if it'll happen
azawawi moritz: several classes are inside one file, moved to their own files (e.g. mysql)
moritz: mysql -> MySQL 10:16
moritz: kebab-case
moritz: add instructions for windows
moritz azawawi: as long as those classes are rather small, I'd rather keep them in separate files
azawawi: and I'm fine with adding kebab-cased APIs, but please don't break backwards compatbility 10:17
azawawi: instructions for windows are fine
s/separate/common/
10:17 dayangkun left
azawawi moritz: and improve documentation for newbies on different platforms 10:17
moritz: i will work on my own fork and then submit a PR.... to be safe :) 10:19
El_Che please don't go the python way, where it's stdlib is fairly inconsisting in the naming of methods and functions (capitals, not capitals, etc) 10:23
timotimo the stdlib is something that makes people sad
ely-se I don't want people to be sad. :( 10:24
arnsholt Well, with Perl 6 there's not really going to be all that much of a stdlib. Which is probably going to give us other problems instead =) 10:25
timotimo it's where code goes to die :)
oh hey arnsholt
azawawi moritz: which modules are using atm DBIish? 10:26
JimmyZ time to rename DBIish to DBI? :P
10:27 CIAvash joined
arnsholt o/ timotimo 10:27
moritz azawawi: dunno; some of my scripts (non-modules) use it 10:28
timotimo whenever i wanted to do something with databases from perl6, i used DBIish 10:30
[Sno] lizmat, moritz - confirming local test that perl6 runs on target :) 10:34
10:37 Actualeyes joined, andreoss left
timotimo m: say slurp() ~~ m:i:g/<[sno]>/ 10:44
camelia rakudo-moar c9cb7f: OUTPUT«(｢s｣ ｢n｣ ｢s｣ ｢o｣ ｢o｣ ｢n｣ ｢n｣ ｢n｣ ｢n｣ ｢s｣ ｢s｣ ｢n｣ ｢n｣ ｢n｣ ｢s｣ ｢o｣ ｢s｣ ｢o｣ ｢n｣ ｢n｣ ｢o｣ ｢s｣ ｢o｣ ｢N｣ ｢s｣ ｢s｣ ｢o｣ ｢n｣ ｢o｣ ｢n｣ ｢s｣ ｢n｣…»
timotimo m: say slurp() ~~ m:i:g:m/<[sno]>/
camelia rakudo-moar c9cb7f: OUTPUT«(｢s｣ ｢n｣ ｢s｣ ｢o｣ ｢o｣ ｢n｣ ｢n｣ ｢n｣ ｢n｣ ｢s｣ ｢s｣ ｢n｣ ｢n｣ ｢n｣ ｢s｣ ｢o｣ ｢s｣ ｢o｣ ｢n｣ ｢n｣ ｢o｣ ｢s｣ ｢o｣ ｢N｣ ｢s｣ ｢s｣ ｢o｣ ｢n｣ ｢o｣ ｢n｣ ｢s｣ ｢n｣…»
timotimo m: say (slurp() ~~ m:i:g:m/<[sno]>/)>>.Str.join
camelia rakudo-moar c9cb7f: OUTPUT«snsoonnnnssnnnsosonnosoNssononsnnnonnnsoosnsoosononnonsnsoNsoónsssssosonósnsnnnsnnnsooNoonsóosnonsonnnoossnnnnnsóSosoonnnsooSnSnoónnnonsnssonnnoooosnosssóoonnnnnnsoo␤»
timotimo m: say (slurp() ~~ m:i:g:m/<[tux]>/)>>.Str.join 10:45
camelia rakudo-moar c9cb7f: OUTPUT«ttúutuuuttúúúutttttuútuúttútttttutúuttttúttúTttutuututt␤»
timotimo m: say (slurp() ~~ m:i:g:m/<[coke]>/)>>.Str.join
camelia rakudo-moar c9cb7f: OUTPUT«CééeoCoecéEeocococoeeoccoceoeooeocoeoccoceococeoeóecceoeceoeóéeeeoooéoecóoeécoeceeoéééocoeeéeóeeooceeoeoooócoeéoeeoocoocecoceeóeooeoo␤»
pink_mist snusnu? 10:48
timotimo m: say (slurp() ~~ m:i:g:m/<[snu]>/)>>.Str.join
camelia rakudo-moar c9cb7f: OUTPUT«snsnúnnnussnnnsusnnsNuussúnúúnsnnnunnnssnssnnnnsnusúNsnsussssúsnsnsúnnnsnnnsNnssnunsnnúnssnunnnnsSsnnnsSnSnúnnnnsnssnnnúsnssusunnununnns␤»
timotimo m: say (slurp() ~~ m:i:g:m/<[snu]><[snu]>/)>>.Str.join(" ")
camelia rakudo-moar c9cb7f: OUTPUT«ún us nn Nu ún nn nn nn ún nn ns ún nu nn nn ún us nn␤»
timotimo m: say (slurp() ~~ m:i:g:ov:m/<[snu]><[snu]>/)>>.Str.join(" ")
camelia rakudo-moar c9cb7f: OUTPUT«ún us nn Nu ún nn nn nn ún nn ns ún nu nn nn ún us nn␤»
10:51 eliasr joined 10:55 azawawi left 10:56 Peter_R joined
TEttinger no no, I swear perl 6 is more legible this time! 10:58
and then we get say (slurp() ~~ m:i:g:m/<[snu]><[snu]>/)>>.Str.join(" ")
timotimo m: say (slurp() ~~ m/ :ignorecase :global :ignoremark <[snu]><[snu]> /).map(*.Str).join(" ") 10:59
camelia rakudo-moar c9cb7f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/U_YfeQz3dr␤Unrecognized regex modifier :global␤at /tmp/U_YfeQz3dr:1␤------> 3say (slurp() ~~ m/ :ignorecase :7⏏5global :ignoremark <[snu]><[snu]> /).map␤»
timotimo m: say (slurp() ~~ m:global:overlap/ :ignorecase :ignoremark <[snu]><[snu]> /).map(*.Str).join(" ")
camelia rakudo-moar c9cb7f: OUTPUT«ún us nn Nu ún nn nn nn ún nn ns ún nu nn nn ún us nn␤»
timotimo better?
TEttinger well, no, but it's still impressive 11:00
timotimo at some point all i can say is "well, it's not python or java" :)
TEttinger I primarily use clojure and people complain about how many non-alnum glyphs it uses 11:01
timotimo oh?
i haven't looked at clojure much, i know it uses ( ) [ ] ! and ? at least, and the -, too
TEttinger yeah, not perl people, but java/python people
timotimo though - could in some cases be called "alnum glyph"
or rather "honorary alnum glyph" 11:02
TEttinger yeah, it's a lisp, so you have () surrounding function/macro calls
timotimo oh, and ' for quote, of course
TEttinger but clojure also has [] for list-like vector literals, which mostly remove the need for quoted lists
timotimo i remember that part so far
m-expressions! :P
TEttinger {} for hash-maps, #{} for hash-sets, #() for anonymous fn shorthand, :keyword-or-something as a special interned string meant to be used as a key in an associative collection (map or set) 11:03
timotimo ah, that :foo thing is usually called "a symbol", right? 11:04
TEttinger not quite, symbols are things like +
or map
timotimo oh
TEttinger they're names of identifiers
usually they need the ' quote to be used
timotimo OK
TEttinger < and > notably just mean less than and greater than... except when used for arrow stuff 11:05
(-> 1 (* 2) (+ 3))
timotimo ah?
what kind of thing do you get from that?
TEttinger is the same as (+ (* 1 2) 3)
-> passes an arg through multiple function bodies
it's a relatively simple macro and perl if it doesn't already have it will as soon as macros are done 11:06
timotimo i don't quite understand why it'd pull the + out in front there
TEttinger prefix notation, that's a later call
timotimo oh, huh 11:07
a full-on shuffle
TEttinger pretty much
-> helps keep that stuff legible
timotimo oh, now i get it
11:07 _mg_ left
TEttinger there's a variant that puts the arg at the end instead of the beginning of the argument lists 11:07
timotimo you read it kind of like "with a 1, first multiply it by 2, then add 3 to it"
TEttinger yep!
timotimo now that makes a whole lot of sense
TEttinger it's one of the best uses for macros that makes things simpler, changing order to improve clariy 11:08
clarity
timotimo m: say 10.&infix:<*>(2).&infix:<+>(3)
camelia rakudo-moar c9cb7f: OUTPUT«23␤»
TEttinger macros often make things more complex
timotimo basically that :)
TEttinger yes
perl is amazing at the string stuff 11:09
timotimo even if this isn't a string of characters, but a string of operations :D
TEttinger yeah, mostly still amazed by how you had u with an accent without typing it up above
timotimo hah
that's ignoremark at work 11:10
any PCRE-using language ought to have that
TEttinger would it ever make the s with comma underneath?
or c with cedilla?
timotimo it could, it's just not in the input file
let me see how i type that with neo2
11:11 rindolf joined
timotimo m: say "çṣ" ~~ m:ignoremark:global/ <[sno]> /; 11:11
camelia rakudo-moar c9cb7f: OUTPUT«(｢ṣ｣)␤»
timotimo m: say "çṣ" ~~ m:ignoremark:global/ <[snoc]> /;
camelia rakudo-moar c9cb7f: OUTPUT«(｢ç｣ ｢ṣ｣)␤»
RabidGravy apparently, DBIish is depended on by "DB::Model::Easy", "Task::Star", "DB::ORM::Quicky", "Slang::SQL", "Lingua::Unihan"
TEttinger neat 11:13
11:14 ely-se left
timotimo for reference, this is the source string: 11:16
m: say slurp
camelia rakudo-moar c9cb7f: OUTPUT«Céad slán ag sléibhte maorga Chontae Dhún na nGall␤Agus dhá chéad slán ag an Eireagal ard ina stua os cionn caor is coll;␤Nuair a ghluais mise thart le Loch Dhún Lúich’ go ciúin sa ghleann ina luí␤I mo dhiaidh bhí gleanntáin ghlas’ G…»
TEttinger ohhh
11:18 dayangkun_ left
RabidGravy I think, JSON::Tiny wins in the most depended up module though 11:21
timotimo probably
also, JSON::Fast won't ever be able to catch up :P
RabidGravy I'll do a proper list later, going shopping in a bit
11:24 ely-se joined 11:28 _mg_ joined 11:31 CIAvash left 11:33 RabidGravy left, CIAvash joined
dalek kudo/nom: 6f8030b | lizmat++ | src/core/Str.pm:
Proper fix for JVM breakage, moritz++
 11:36
jnthn morning, #perl6 11:37
ooh, no messages :) 11:38
timotimo oh hey jnthn
lizmat jnthn o/ 11:39
|Tux| timotimo, any special reason you used my nick in that split, as in: do you have a question? 11:49
timotimo sorry, it's just the people who have [ ... ] around their nicknames 11:50
no disturbance intended
(just because [ ] used to be character classes) 11:51
woolfy jnthn ohey 11:52
11:53 TEttinger left
timotimo obey jnthn 11:53
11:53 CIAvash left 11:59 zacts left
jnthn timotimo: Indeed. Now, fix 10 RTs! :P 11:59
timotimo since you didn't tell me to fix 10 existing RTs, i'll just write up some useless RTs and fix them as NOTABUG :P 12:00
jnthn Which just shows that it doesn't matter what metric you pick for anything, people will find a way to game it. :-) 12:01
moritz ah, the funny stories about companies that tried to measure development productivity by number of commits 12:02
timotimo "you removed a thousand lines, so you owe us bigtime, buddy!"
12:03 _mg_ left, zacts joined, kjs_ left 12:04 xpen left
jnthn remembers the time he removed 300,000 lines of C# and 10,000 lines of SQL at one client... Using Perl 6 scripts to do hacky dependency analysis and instrumentation on the code. :) 12:04
hm, I wonder if rt.perl.org/Ticket/Display.html?id=122891 is magically fixed by \r\n being a grapheme :) 12:05
m: say "a\n" ~~ /^^\s*$$/
camelia rakudo-moar 6f8030: OUTPUT«Nil␤»
jnthn m: say "a\r\n" ~~ /^^\s*$$/
camelia rakudo-moar 6f8030: OUTPUT«Nil␤»
jnthn yay :)
For once it fixes something rather than breaks something :P 12:06
12:08 Skarsnik joined
dalek ast: db7683e | jnthn++ | S05-metachars/line-anchors.t:
Test for RT #122891.
 12:09
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=122891
12:10 xpen joined
jnthn One more xmas RT falls :) 12:13
jdv79 m: Array.new(|[456]).perl.say; Hash.new(|{:baz(456)}).perl.say 12:14
camelia rakudo-moar 6f8030: OUTPUT«[456]␤{}␤»
jdv79 that seems inconsistent to m
e
timotimo very good, now you'll only have to do another ~0.7 RTs today, jnthn ;)
jnthn I did 2 yesterday ;) 12:15
timotimo oh!
oops. we should turn off your internet connection for a day, in that case!
lizmat that wold leave .4 for today then :-)
*would
jnthn I'm going to kick rt.perl.org/Ticket/Display.html?id=123154 off the xmas RT list 'cus I fixed it on Moar, then it got re-opened and tagged JVM only 12:16
moritz +1 12:17
jnthn If anyone wants a hint where to fix it on JVM, I can provide one.
12:20 sufrostico joined
lizmat jnthn: perhaps add the hint to the ticket ? 12:21
if not for yourself down the line ?
12:24 _mg_ joined
jnthn lizmat: I'd have to look stuff up to give it :) 12:25
Though, could do I guess
lizmat it's sort of fresh in your mind now :-) 12:26
jnthn noted the Moar commit that fixed it on the ticket
jdv79 Hash.new((|$h)) # that sort of thing seems wonky
jnthn jdv79: If you add an extra pair of parens, you're making a Slip 12:27
jdv79 i have to in this case 12:28
becaues of what i mentioned above
maybe someday it'll make sense but to a newbie that does not
jnthn Right, |{...} passed named args, but the Hash constructor expects a list of pairs
jdv79 yet the array constructor can take [] or |[] just fine 12:29
just weird
jnthn |[] also results in positional arguments
jnthn will be gone for a few hours 12:30
jdv79 more to my point - why doesn't Hash.new(:a(123)) work?
it just feels wrong. probably me.
12:31 xpen left 12:39 ilogger2 joined, ChanServ sets mode: +v ilogger2
jdv79 but lunch i think. gettin tired fast. & 12:39
12:40 kaare_ joined 12:42 azawawi joined
azawawi hi 12:42
moritz: initial refactor github.com/azawawi/DBIish/tree/mas...lib/DBDish without pg 12:43
dalek kudo/nom: d2b4941 | lizmat++ | src/core/ (2 files):
PROP/PVALCODE internal subs to Rakudo::Internals
 12:47
12:48 caymanboy joined
Zoffix .tell moritz would it be possible to glance at update-modules.perl6.org that's run by cron? I just wanna check my modifications integrate well with it and will offer any updates to that file once the app is ready 12:51
yoleaux Zoffix: I'll pass your message to moritz.
lizmat $ 6 'dd Hash.new(:42a)'
Hash $var = {:a(42)}
jdv79: ^^^
$ 6 'dd Hash.new((:666b),:42a)' 12:52
Cannot mix positional and nameds in Hash.new
moritz Zoffix: perlpunks.de/paste/show/5641e88b.3530.2e0
yoleaux 12:51Z <Zoffix> moritz: would it be possible to glance at update-modules.perl6.org that's run by cron? I just wanna check my modifications integrate well with it and will offer any updates to that file once the app is ready
Zoffix Thanks.
grondilu m: say (foo => "bar") . hash<foo>; # not sure if &[.] is supposed to be reckognized here 12:55
camelia rakudo-moar d2b494: OUTPUT«5===SORRY!5=== Error while compiling /tmp/nxMPshQ9O8␤Missing required term after infix␤at /tmp/nxMPshQ9O8:1␤------> 3say (foo => "bar") . hash<foo>7⏏5; # not sure if &[.] is supposed to be ␤ expecting any of:␤ prefix␤ …»
12:55 Celelibi joined
grondilu m: say (foo => "bar") . hash . <foo>; 12:56
camelia rakudo-moar d2b494: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YDqHsTOASA␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/YDqHsTOASA:1␤------> 3say (foo => "bar") . hash . <7⏏5foo>;␤»
grondilu m: say (foo => "bar") . hash.<foo>;
camelia rakudo-moar d2b494: OUTPUT«5===SORRY!5=== Error while compiling /tmp/NeweuHrbky␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/NeweuHrbky:1␤------> 3say (foo => "bar") . hash.<7⏏5foo>;␤»
grondilu m: say (foo => "bar") . hash .{foo};
camelia rakudo-moar d2b494: OUTPUT«5===SORRY!5=== Error while compiling /tmp/w2mt5uyquk␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/w2mt5uyquk:1␤------> 3say (foo => "bar") . hash .{7⏏5foo};␤»
grondilu m: say (foo => "bar") .hash .{foo};
camelia rakudo-moar d2b494: OUTPUT«5===SORRY!5=== Error while compiling /tmp/MPhBzGvqhG␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/MPhBzGvqhG:1␤------> 3say (foo => "bar") .hash .{7⏏5foo};␤»
grondilu m: say (foo => "bar") .hash .{"foo"}; 12:57
camelia rakudo-moar d2b494: OUTPUT«5===SORRY!5=== Error while compiling /tmp/mJ0BhZwA7l␤Unsupported use of . to concatenate strings; in Perl 6 please use ~␤at /tmp/mJ0BhZwA7l:1␤------> 3say (foo => "bar") .hash .{7⏏5"foo"};␤»
12:58 s_kilk joined
s_kilk So, I was looking at the source-code for the UUID module, and noticed a cool looking '>>' operator github.com/retupmoca/P6-UUID/blob/...ID.pm6#L26 , but I can't find any reference to it on doc.perl6.org. Am I losing it, or is there no documentation for that op? I'm pretty sure I'm losing it. 12:59
moritz s_kilk: it's actually not >>, but >>. 13:01
s_kilk: and it means "call a method on each of these values, possibly in parallel"
s_kilk ah, now that's cool
moritz and yes, might not be documented properly 13:02
we are still missing lots of docs :(
s_kilk Ah, ok, that's understandable
is there a deeper source of documentation that might be more raw than doc.perl6.org, but closer to the spec?
probably the spec and test suite itself, really 13:03
moritz right 13:04
design.perl6.org
s_kilk thanks 13:05
And there it is: design.perl6.org/S17.html . thanks for the help. 13:07
13:12 s_kilk left
dalek kudo/nom: 7099c76 | lizmat++ | src/core/Map.pm:
Streamline Map.new, for better inlining
 13:16
[Coke] .botsnack 13:17
yoleaux 09:46Z <lizmat> [Coke]: testing on JVM works again
:D
synbot6 om nom nom
[Coke] (thanks, lizmat++)
lizmat in the end, it was a 4 char patch 13:18
13:20 telex joined
ely-se I wish yoleaux and synbot6 both replied to ".botsnack" with ".botsnack" 13:25
synbot6 om nom nom
moritz we've had bot loops in the past. It was no fun :( 13:29
ely-se nice :D
13:29 dbrock- joined
DrForr .botcrack 13:30
13:33 dbrock- left, dbrock- joined
timotimo lizmat: oh, .STORE will return self? 13:36
lizmat yes
timotimo good catch!
lizmat double checks 13:37
timotimo oh, also, will using the *% for Hash's new method prevent subclassing Hash? because usually you'll have self.bless or nqp::create(self) somewhere
er, as in: method new(*%contents) { return %contents }
13:38 RabidGravy joined
lizmat well, I tried that already, but *% does not have containers for values 13:38
timotimo oh!
azawawi moritz: so even $sth.fetchrow is not needed since we can say $sth.rows[0], right?
timotimo yeah, that's trouble
RabidGravy So the top ten most depended on modules are JSON::Tiny, XML, File::Temp, MIME::Base64, URI, File::Find, Digest, LWP::Simple, HTTP::Easy, LibraryMake 13:39
13:40 andreoss joined, kjs_ joined
moritz azawawi: depends on what you mean by "not needed" 13:40
azawawi moritz: deprecated that is
moritz azawawi: if you don't want the previous rows to be in memory, the .fetchrow interface might still be easier
azawawi moritz: one thing, why fetchrow, not fetch-row? :)
moritz azawawi: I don't think it's deprecated. See my earlier notice about staying backwards compatible
lizmat I haven't looked at DBIish, but would having the resultset being an Iterable (with an iterator) solve all of the issues ? 13:41
moritz azawawi: because it's part of the p5 DBI API emulation
13:41 dbrock- left
moritz azawawi: and if we don't encourage folks to use it, there's no need to rename the methods 13:41
azawawi moritz: i see
13:41 dbrock- joined
moritz don't modernize compatiblity APIs :-) 13:42
13:42 dbrock- left
azawawi moritz: so you want database code written in Perl to be easily moved to Perl 6, right? 13:42
moritz azawawi: that, and I don't want to break any current users
unless for *very* good reasons. "This is a shinier name" is not a very good reason. 13:43
azawawi moritz: in that case, why not introduce a compatibility role that does that?
but it is shiny :)
lizmat m: sub a(*%h) { dd %h; %h }; a(:42a)<a> = 666 # timotimo
camelia rakudo-moar 7099c7: OUTPUT«Hash $var = {:a(42)}␤Cannot modify an immutable Int␤ in block <unit> at /tmp/hB3i_iwaHl:1␤␤»
azawawi lizmat: i always say less code is better code until less code is not readable 13:44
timotimo right
moritz azawawi: IMHO we can have three APIs: a low-level API that a driver must provide, the compatiblity methods, and a shiny API (.rows or .all or so being a Seq)
timotimo so no problem :)
moritz azawawi: the latter two can be provided by roles
azawawi moritz: so basically i will make DBI::Easy :)
13:45 dbrock- joined
Skarsnik RabidGravy, interesting. maybe it could be show in a way in modules.perl6.org? I mean when I had to choose a JSON module I was like "wich on is the most commonly used/maintened/not a pet test 13:45
jdv79 an old cpan problem 13:46
moritz which would show CGI.pm with a huge number
even though one probably doesn't want that module for new code 13:47
lizmat is there a reason we don't have an Any.iterator? method iterator(Any:D:) { self.list.iterator } ? 13:48
jdv79 temporal decay mixed in may help
Skarsnik well for stuff like a JSON module it could be a tag: 'used in perl6 common tools' to mean it used in panda so it's probably well maintened 13:49
jdv79 no
azawawi Skarsnik: that's why we added github stars columns :) 13:50
jdv79 a lot of crappy code is used in common tools
ratings help. reviews would help.
moritz and the module with the most github stars has them because hackernews liked the readme 13:51
jdv79 stars are not good. exactly.
moritz (fwiw I like the readme too :-)
azawawi any ratings system can be eventually spoofed... look at google search results 13:52
jdv79 cpan, the ecosystem, has better solutions
RabidGravy well if anyone wants to play with the deps then it's gist.github.com/jonathanstowe/f198...3ed0b74b3e 13:54
azawawi jdv79: what about knowing whether this module has great documentation or not? 13:55
jdv79 namely metacpan's search
sounds like kwalitee 13:56
a lot of this has been done
Skarsnik is there a nice way to export/import the list of module installed? 13:57
Zoffix Skarsnik, panda list --installed 13:58
huh 13:59
Skarsnik, panda --installed list
13:59 azawawi left 14:02 zakharyas joined 14:04 kjs_ left 14:05 AlexDaniel joined
moritz awwaiid: it would help to ahve a rating system that doesn't spoof itself :-) 14:06
Skarsnik I don't want to be rude, but maybe remove 'pet'/deprecated module? I mean why there is so much JSON module? like JSON::Pretty, JSON::Fast, JSON::Faster? 14:09
Zoffix jdv79, was it you who made a blog post somewhere about MetaCPAN for P6? 14:10
jdv79 yeah. ive been distracted. hope to get back to it soon. 14:11
Zoffix I was wondering what, if any, progress has been made to make it. And is there a contributable repo?
jdv79 yeah
im at lunch on my phone. later we could talk more:)
Zoffix Sure
jdv79, I'm almost done with mojo-app powered version of modules.perl6.org; was just wondering how many bells and whistles I should bother adding to it :P 14:12
If there's a P6 MetaCPAN in works, I rather contribute to that. 14:13
mrf Skarsnik: The question is who decides which ones get removed? 14:14
RabidGravy I'd say it's too early to even think about that
Zoffix There will be rotting crap always; look at P5's CPAN. just let it happen :) 14:15
mrf I agree.
dalek osystem: 6a89179 | (Kamil Kułaga)++ | META.list:
Added Pod::Coverage

See: github.com/teodozjan/pod6-coverage		 14:19
Zoffix \o/ 14:20
14:21 yqt joined
andreoss m: say || \\(1) 14:25
camelia rakudo-moar 7099c7: OUTPUT«\(1)␤»
andreoss m: say || || \\(1)
camelia rakudo-moar 7099c7: OUTPUT«\(1)␤»
14:27 khw joined
andreoss m: say || \(\(1)) 14:28
camelia rakudo-moar 7099c7: OUTPUT«\(1)␤»
andreoss m: say ||( || \(\(1)))
camelia rakudo-moar 7099c7: OUTPUT«\(1)␤»
14:29 Skarsnik_ joined
andreoss how to resolve Capture of a Capture? 14:31
m: say ||( (|| \(\(1)))[0] )
camelia rakudo-moar 7099c7: OUTPUT«1␤»
14:33 mr_ron joined, Skarsnik_ is now known as Skarsnik
dalek href="https://modules.perl6.org/mojo-app:">modules.perl6.org/mojo-app: 48f1839 | (Zoffix Znet)++ | mojo-app/ (9 files):
App now passes all but one tests in t/02-app-home-page.t
 14:33
Zoffix Time to play Fallout 4 \o/
ugexe Zoffix: truuuu 14:35
andreoss m: \\\1
camelia rakudo-moar 7099c7: OUTPUT«Potential difficulties:␤ To refer to a positional match capture, just use $0 (numbering starts at 0).␤ Parenthesize as \(...) if you intended a capture of a single numeric value.␤ at /tmp/U2oaEMlUay:1␤ ------> 3\\\7⏏051␤»
andreoss strange warning
lizmat it's really for people coming from P5 14:38
14:40 virtualsue joined 14:42 dbrock- left, dbrock- joined 14:43 dbrock- left
zengargoyle i'm still a little afraid of POD6, don't know which parts of the spec are implemented or not. 14:44
14:45 dbrock- joined
zengargoyle like =data not being there yet but a #| comment magically putting a code snippet in my README.md 14:45
ugexe try { run("asdf"); }; say 1 # should this reach `say 1`?
moritz ugexe: ah, it's because run() returns something that, in sink context, blows up 14:46
a Proc object, to be exact 14:47
we've fixed that for Failure, but seemingly not for Proc
ugexe yea, but why do i have to catch it to prevent it from escaping the try
moritz because we haven't yet given it the same treatment as Failure
and we probably should
14:47 xpen joined
timotimo right, try turns on "use fatal" inside the block, and that doesn't seem to work properly with the proc thing 14:48
ugexe wouldnt calling method sunk (or the exception inside it) only once fix it?
flussence should, imo 14:49
timotimo ah, so it's just missing a little attribute "already-sunk"
moritz no
I don't think that try { run("asdf"); }; calls sink inside the try block at all 14:50
timotimo oh, of course, it's not sink that gets called
moritz because sink context doesn't propagate inwards (yet?)
timotimo i forget what our code-gen does, exactly
moritz it's only the outside that's sunk
timotimo p6fatalize 14:51
gee, thanks
14:51 rindolf joined
moritz "layer of indirection" 14:51
ugexe hmm. the other thing about it that bugs me is that i cant reuse the Proc that the exception gives me. if i try to use the Proc object outside of the catch later (to examine WHY it failed for instance) it will throw the exception again 14:52
timotimo ah, that's in the actions
14:52 sprocket joined 14:53 advwp joined
timotimo yeah, if istype($thing, failure) { $thing.sink } 14:53
so since run doesn't return a failure-istype-ing thing, we don't sink it inside the try 14:54
moritz maybe we need a common role for "blows up in sink context"
let's call it "Titanic"
ugexe adding a `so` before run makes it work as i would expect
PerlJam
.oO( SinkHole )
timotimo sure
flussence
.oO( Cat )
timotimo that forces the sinkness to happen earlier 14:55
not ideal, of course
cats only blow up in the sink if you turn on the water
and then, not all cats do that 14:57
flussence see? perfect metaphor for the average command line program! :)
lizmat timotimo: indeed, my sister has cats that sit in the sink waiting for you to put on the water tap 15:01
timotimo tee hee 15:03
the kitties i hang out with regularly don't like water too much
huf dont like it on them, or dont like it at all? most cats like to paw at water at least 15:06
timotimo i know they don't like it on them
letting drops fall onto one of them makes him twitch his back a bunch, but he doesn't jump up and run away or anything 15:07
ugexe ahh on windows that try { run "asdf" }; returns exit code -16 instead of -1 15:08
zengargoyle i had a kitten once that would stand outside the shower mewing until you let it in to play in the rain.
timotimo d'aaw
RabidGravy our cat was very surprised once when she was young, she jumped in the bath thinking it was empty to find that it wasn't 15:13
timotimo there's this image series of a very concerned cat that has discovered his/her owner (well, slave, really) fell into water (a bathtub) and calls for help on their behalf %) 15:15
15:17 mr-foobar joined 15:19 sprocket left 15:20 isBEKaml joined
DrForr Percy absolutely *loves* her showers. 15:21
timotimo oh lord, percy :D
isn't that also the name of a cat in some disney movie? 15:22
a dog, actually 15:23
DrForr In this case she's a bird, actually.
timotimo ah 15:24
i know birds like to bathe (by necessity, i imagine), but shower is new to me :)
though of course birds get showered every now and then by regular rain
DrForr Indeed, and with a daddy who can make it rain whenever he wants...
dalek rl6-roast-data: 7ba7df9 | coke++ | / (9 files):
today (automated commit)
 15:29
[Coke] RT-- 15:30
anything in the ecosystem that can deal with wget's saved cookie files? 15:35
timotimo not that i know, nah
except if it's a common format that we do have something for already 15:36
[Coke] do any of the http-aware modules do cookies? checking... 15:37
moritz I'm pretty sure LWP::UserAgent does
'cause I discussed cookie jar design with sergot++ back in the days
[Coke] LWP::UserAgent isn't listed on modules 15:38
dalek kudo/nom: ed10656 | peschwa++ | src/Perl6/Metamodel/BOOTSTRAP.nqp:
Correct caching logic for rwness-sensitive MMD.

  lizmat++ for pointing out this didn't work right yet.
[Coke] I assume it's provided by something else? 15:39
15:39 softmoth_ joined
psch almost feels like it could've been less effort to try and implement rwness-aware multicachefind... o.o 15:39
but that now actually passes S06-multi/by-trait.t, and roast was clean as well 15:40
15:40 skids joined
lizmat psch++ 15:40
dbrock- ui sai i ca'e salci 15:41
i ie tirna na la'a u'i gi'e te cusku
[Coke] moritz: do you mean HTTP::UserAgent? 15:42
dbrock- oops hahahaha sorry
moritz [Coke]: erm, yes 15:44
zengargoyle wonders idly if POD =data could be implemented by cloning =code and tweaking things.
Skarsnik HTTP::Ua has the http::cookies stuff x) 15:45
lizmat zengargoyle: that feels like that could work 15:46
[Coke] (nice having so much stuff already in the ecosystem, even if it's not everything)
RabidGravy yeah HTTP::UserAgent does the cookies, no persistent cookie jar as yet. There may also be a bug in the parsing for some content 15:47
Skarsnik also $.cookies is not writable ~~ 15:48
15:50 muraiki joined
RabidGravy Skarsnik, it is now 15:52
psch m: multi foo(Int $x is rw) { say "writable" }; multi foo(Int $x) { "readonly" }; say foo(42); foo my $ = 42;
camelia rakudo-moar ed1065: OUTPUT«readonly␤writable␤»
RabidGravy like right now 15:54
Skarsnik RabidGravy, I mean I did not test yet, but I am not I could write $ua.cookies.set-cookie("stuff");
[Coke] sees a weird error out of http::useragent: 15:56
err code: 0
error:00000000:lib(0):func(0):reason(0)
[Coke] wonders what tier that's even coming from.
RabidGravy nope, don't know
[Coke] I don't see docs on how to use a cookie jar with an http useragent. digging... 15:57
15:58 mr_ron left 15:59 isBEKaml left
RabidGravy something like $ua.cookies = HTTP::Cookies(file => $somefile, autosave => True); 15:59
15:59 isBEKaml joined
RabidGravy but dunno if that's well tested 15:59
16:01 Skarsnik left, virtualsue left, Skarsnik joined
Skarsnik weird, only freenode disconnect me w) 16:04
16:13 isBEKaml left 16:15 Begi1116 joined 16:20 retupmoca joined
dalek rl6-most-wanted: ae81cf8 | RabidGravy++ | most-wanted/modules.md:
Add a few more WIPS
 16:21
rl6-most-wanted: d23ea2c | RabidGravy++ | most-wanted/modules.md:
Merge branch 'master' of github.com:perl6/perl6-most-wanted
hoelzro o/ #perl6 16:23
RabidGravy rarr!
lizmat m: .say for Mu # is this a bug or not ? 16:24
camelia rakudo-moar ed1065: OUTPUT«Method 'map' not found for invocant of class 'Mu'␤ in block <unit> at /tmp/UQmw9mG9aU:1␤␤»
lizmat All tests successful.
Files=1074, Tests=50141, 236 wallclock secs (12.91 usr 3.81 sys + 1442.08 cusr 131.04 csys = 1589.84 CPU)
whee!
16:25 rarara_ joined 16:26 eliasr joined, sprocket joined
Skarsnik I think what should be added in most wanted feature : something to 'automatize' the creation of Binding modules x) 16:27
16:27 xpen left
dalek kudo/nom: 326a744 | lizmat++ | src/core/ (7 files):
You want an iterator? Just ask for one!

There were many places in the code doing a ternary checking for iterability, and if not, do .list.iterator instead of just .iterator . By putting an iterator method in Any, all of these ternaries can go, leading to less code to execute and better inlineability. Of course, since Nil is no longer Any, it needed its own iterator method.		 16:27
timotimo yeah, there's already a C parser, Skarsnik
get to work! ;) 16:28
Skarsnik I looked at it and was thinking of getting only the part that parse definition of type/struct/enum and function declaration to parse C header. I am pretty sure most of the time you don't even need to resolve macro or other #include. 16:31
stmuk_ m: ++4 16:32
camelia rakudo-moar 326a74: OUTPUT«Parameter '$a' expected a writable container, but got Int value␤ in block <unit> at /tmp/G_2TxCC4VA:1␤␤»
Skarsnik but I am not sure how to write/modify the grammar to ignore the other stuff 16:33
stmuk_ I'm seeing a similar error to that as a regression with nom since niceville for nativecall
Skarsnik why Parameter $a? 16:34
psch m: say &prefix:<++>.params[0].perl
camelia rakudo-moar 326a74: OUTPUT«Method 'params' not found for invocant of class 'Sub+{Precedence}'␤ in block <unit> at /tmp/Y0K3QoCDeZ:1␤␤»
moritz say &prefix:<++>.signature.params[0].perl
m: say &prefix:<++>.signature.params[0].perl
camelia rakudo-moar 326a74: OUTPUT«Mu $␤»
psch m: say &prefix:<++>.candidates[1].signature.params[0].perl
camelia rakudo-moar 326a74: OUTPUT«Mu:U $a is rw␤»
moritz that's the proto, I guess
psch++ is ahead of me :-) 16:35
psch stmuk_: the "is rw" dispatch stuff should be fixed since about an hour or two ago, did you pull and rebuild with latest?
(although i don't know what might've cause the regression in Niceville)
stmuk_ psch: I just built about then but will retry anyway 16:36
16:41 mr_ron joined
psch m: multi f(Str $a is rw) { }; multi f(Int $b is rw) { }; f "foo" # this is a bit unfortunate, but afaiu should be resolvable when we have multicachefind rw-aware 16:42
camelia rakudo-moar 326a74: OUTPUT«Parameter '$b' expected a writable container, but got Str value␤ in block <unit> at /tmp/nDYADmSkif:1␤␤»
psch note it's only the name itself that gets mismatched, not the dispatch
i guess it's also fixable right now, but that'd need quite bit more bookkeeping than i'd like to figure out for a hopefully-not-permanent solution 16:43
stmuk_ yes! it works now 16:44
I must have just missed the fix
16:45 zakharyas left
lizmat cycling& 16:45
16:48 ^elyse^ joined 16:51 ^elyse^ left
timotimo oof. when snarky commit messages are the only piece of happyness you can get out of a piece of coding 16:52
RabidGravy :-O 16:56
I don't suppose there is any existing code that turns 'Foo::Bar' into 'Foo/Bar.pm' ? 16:57
timotimo you have to look it up via the meta's provides section
RabidGravy nah, not that
literally having a module name in hand (that I may e.g. want to create) and turn it into a pathname 16:58
timotimo choose a shiny path with a password creation algo or something, then put it into the provides section of the meta file :P 16:59
RabidGravy :-\
17:00 yqt left
RabidGravy I'll just go with "($mod-name ~~ s:g/'::'/{ $*SPEC.dir-sep }/ ) ~ '.pm'" then 17:01
:)
timotimo what if the operating system you're on doesn't construct paths just by putting the dir-sep in between?
you should split by '::' and catpath the result, i think
17:01 kjs_ joined
Skarsnik why not pm6? 17:01
RabidGravy because I prefer .pm 17:03
timotimo you're so Rab!
Rabical! 17:04
RabidGravy Rabbitical
catfile though
there should be something which gives you '::' 17:09
timotimo m: say ":" x 2
camelia rakudo-moar 326a74: OUTPUT«::␤»
Skarsnik $*SPEC.modulesepforlazyrabid 17:10
zengargoyle Skarsnik: C::Parser doesn't do #define stuff, have to run it through gcc -E or such, then it might loose some of the #defines which should maybe turn into Enums. 17:11
zengargoyle thinking out loud 17:12
RabidGravy zengargoyle, I looked at it a while back with regard to doing that
ilmari
.oO( C::PreProcessor )?
timotimo you really want to do that to yourself? 17:13
or is that just something that uses gcc or clang for the job?
Skarsnik zengargoyle, well for extracting type/funct declarationg ignoring Preprocessor stuff could be enought x) 17:14
RabidGravy I got a bunch of scripts from extracting those parts from a .h file, but mostly hand made for each project 17:15
17:16 _mg_ joined
timotimo first i read "a bash script" 17:16
next thought was "NO WAY" 17:17
zengargoyle might be easier to take some other SWIG-like thing and add a Perl 6 backend?? 17:18
Skarsnik probably
zengargoyle not really sure
17:19 dbrock- left
Skarsnik I looked at SWIG but it confuse me, the documentation mentionned writing a small program and SWIG will bind the two language. Not generate the binding file x) 17:20
timotimo isn't the thing about swig that it'll generate C code to bind C++ libraries and such?
zengargoyle h2xs :P
timotimo haxors
17:21 ^elyse^ joined
Skarsnik It supposed to be able to write perl5 binding 17:21
timotimo yeah, probably via XS
because what else was there back then
17:22 rarara_ left
Skarsnik yes, but it generate the XS file, so it does seem a strech to assume we can make it generate perl6 nativecall code ~~ 17:23
zengargoyle think it'll be a multi-step process though. maybe parse information and dump somehow to give you the chance to rename things, patch up mistakes, etc. then use that to generate bindings. 17:26
17:26 _mg_ left
dalek p: b3c4902 | (Jens Rehsack)++ | / (3 files):
introduce support for sdkroot and sysroot to support cross-builds

To allow building for another location rather the final install prefix, regardless whether chroot based builds, cross-compile or canadian cross sdk is the goal, realize the difference between prefix, the sdk-root and the system root for final target.
Signed-off-by: Jens Rehsack [email@hidden.address]		 17:35
p: 9f87bb9 | (Jens Rehsack)++ | tools/build/install-jvm-runner.pl:
nqp-j: allow experienced user start with other libdir or jars

To allow dedicated use of another NQP_DIR or another NQP_JAR list for e.g. cross-compiling NQP.
Signed-off-by: Jens Rehsack [email@hidden.address]
p: d380395 | moritz++ | / (4 files):
Merge pull request #252 from rehsack/master

introduce support for staging directories during build
17:35 diana_olhovik joined 17:37 kjs_ left
PerlJam lizmat: re tensorflow, I've been thinking about it, but my problem is always one of time and timing. I wouldn't be unhappy if someone else did it though :) 17:37
timotimo hm? using tensorflow for what? 17:38
PerlJam timotimo: provide a perl6 interface for it via NativeCall 17:39
timotimo oh, right
llfourn incase anyone hasn't seen it: www.youtube.com/watch?v=kwxHXgiLsFE # that p6 'release talk' from a few weeks ago finally got uploaded 17:40
17:40 diana_olhovik left 17:41 hartenfels joined
timotimo i just watched it, but i had the feeling i had already seen it 17:42
17:43 hankache joined
dalek kudo/nom: bb49e08 | (Jens Rehsack)++ | / (3 files):
introduce support for sdkroot and sysroot to support cross-builds

To allow building for another location rather the final install prefix, regardless whether chroot based builds, cross-compile or canadian cross sdk is the goal, realize the difference between prefix, the sdk-root and the system root for final target.
Signed-off-by: Jens Rehsack [email@hidden.address]		 17:44
kudo/nom: 253e633 | (Jens Rehsack)++ | / (4 files):
builder: add capability to work with staging directories

To allow building for a staging root which might differ from host and tools to be used, be more careful about the PATH's to be used. This will allow cross-compiling rakudo using two builds of nqp: a native one in SDKROOT and a cross-compiled one in SYSROOT. Using host one remains untouched when run without --sysroot or --sdkroot, respectively.
Signed-off-by: Jens Rehsack [email@hidden.address]
kudo/nom: fa50048 | moritz++ | / (5 files):
Merge pull request #580 from rehsack/nom

add staging dir support for building
17:45 kjs_ joined
dalek ar: f0def6d | (Jens Rehsack)++ | / (3 files):
introduce support for sdkroot and sysroot to support cross-builds

To allow building for another location rather the final install prefix, regardless whether chroot based builds, cross-compile or canadian cross sdk is the goal, realize the difference between prefix, the sdk-root and the system root for final target.
Signed-off-by: Jens Rehsack [email@hidden.address]		 17:46
ar: dc1ba76 | moritz++ | / (3 files):
Merge pull request #53 from rehsack/master

add staging dir support for building
RabidGravy it's all going off 17:48
timotimo all bets are off?
jnthn So commit! 17:49
TimToady Not rollback!
timotimo gien ne va plus
gitien?
TimToady m: say Iterable === Any 17:50
camelia rakudo-moar 326a74: OUTPUT«Method 'iterator' must be implemented by Iterable because it is required by a role␤ in any compose_method_table at gen/moar/m-Metamodel.nqp:2661␤ in any apply at gen/moar/m-Metamodel.nqp:2671␤ in any compose at gen/moar/m-Metamodel.nqp:2807␤ in…»
TimToady jnthn: you saw this cute one? ^^^
jnthn TimToady: It's long known
timotimo jnthn: did you see that amazing thing TimToady put in the channel where a bunch of unicode things turn into 7 when say'd? 17:51
jnthn TimToady: Clearly, we can't pun a role with missing bits to delegate the method to it
TimToady: We currently have a list of methods that don't cause punning
timotimo: Yeah, ShimmerFairy++ implemented that a while back, no?
Argh
timotimo er, implemented?
jnthn Those configure changes have busted the build on Win32
timotimo it looked more like a bug
TimToady I didn't put it on channel, was just responding 17:52
jnthn I didn't take it as one
I suspect \d matches them
TimToady it's just one step more to make ⅐ return a Rat
jnthn And we know how to numify anything that has a Unicode numeric value
timotimo oh
i totally misread that!
dalek : b5919fc | PerlJam++ | misc/perl6advent-2015/ (2 files):
[advent] Get started on Advent 2015
 17:53
jnthn Well, it could be, but I'd be surprised given our liking of Unicode things ;)
timotimo i thought the string "foobar" where f, o, b, a and r are weird unicode numbers, turned into "777777" when printed out
PerlJam BTW, advent season will be here before you know it. Everyone start thinking about what they could write for perl6advent.wordpress.com/ and add their ideas to github.com/perl6/mu/tree/master/mi...dvent-2015
TimToady I don't consider it a bug, was just surprised it snuck in without me noticing :)
gee, I wonder what I should write for Christmas...
timotimo i only considered that a bug because i misread it, etc etc 17:54
TimToady maybe just "Here."
17:54 ribasushi joined
jnthn TimToady: github.com/rakudo/rakudo/blob/nom/....nqp#L2938 17:54
timotimo TimToady: i liked the way you ended the tack with something big
jnthn That's the exlucded list so far
TimToady: We can keep adding to it, but my "lists of things to remember" nerve is tingling :) 17:55
OK, guess I'd better try to unbust the Win32 build :/
Well, or configure
There's nothing I hate much more than build systems.
Well, there are things, but in software, build systems are pretty argh.
TimToady m: say Failure.^mro
camelia rakudo-moar fa5004: OUTPUT«((Failure) Nil (Cool) (Any) (Mu))␤»
17:56 kjs_ left
TimToady I think we'll just have to live with Nil not having the parens there 17:56
timotimo haha
i love the way Nil sticks out there
TimToady the only alternative is to make it (Nil) everywhere, which I don't want
timotimo remember when Nil disappeared in lists? would be hilarious if that happened right in there, too
17:57 andreoss left
TimToady m: sub foo (--> Int) { Nil }; say foo # now works more like Failure 17:57
camelia rakudo-moar fa5004: OUTPUT«Nil␤»
TimToady the attempt to move Nil outside of Any didn't pan out though... 17:58
well, gotta go to writing group, where they're making me work on PP6, or whatever it ends up being named 17:59
17:59 kjs_ joined
PerlJam TimToady++ though ... "making" you? :) 18:00
jnthn lol
Patching string concat with File::Spec->catfile caused the breakage
I love how making things better makes them worse :P
moritz jnthn: sorry, I only tested under linux 18:02
timotimo still hoping for the windows servers on travis-ci
they also have mac os servers, but i assume they're not going to give us as much coverage as a windows server would 18:03
server/vm/whatever
dalek : 39f612c | PerlJam++ | misc/perl6advent-2015/topic-brainstorming:
[advent] add a small topic brainstorm
p: 517c8f8 | jnthn++ | tools/lib/NQP/Configure.pm:
Avoid stupid catfile behavior, fixing Win config.
jnthn moritz: No worries. In you/[Sno]++'s defence, it *is* silly that it broke, and I'd not have caught this in code review either.
18:03 lostinfog joined 18:06 kjs_ left
jnthn Darn, importing the same fix to Rakudo's configure.pm didn't unbust things. 18:06
oh, there's another catfile 18:07
dalek kudo/nom: 409cdbb | jnthn++ | tools/lib/NQP/Configure.pm:
Unbust Configure on Windows.
18:08 andreoss joined
jnthn All fixed. 18:10
[Sno] jnthn: thanks, that's a weird behavior :/ 18:14
jnthn [Sno]: Ain't it just. 18:15
[Sno]: Anyway, np
18:16 domidumont joined 18:26 abaugher joined 18:33 zengargoylew joined
nine tadzik: ping 18:35
.tell tadzik github.com/tadzik/File-Find/pull/15 18:38
yoleaux nine: I'll pass your message to tadzik.
18:39 isBEKaml joined
grondilu how do I compare buffers? 18:39
m: say Buf.new(1) == Buf.new(1);
camelia rakudo-moar 409cdb: OUTPUT«True␤»
grondilu m: say Buf.new(1, 2) == Buf.new(1, 2);
camelia rakudo-moar 409cdb: OUTPUT«True␤»
grondilu m: say Buf.new(1) == Buf.new(2);
camelia rakudo-moar 409cdb: OUTPUT«True␤»
grondilu m: say Buf.new(1) ~~ Buf.new(2); 18:40
camelia rakudo-moar 409cdb: OUTPUT«False␤»
grondilu m: say Buf.new(1) ~~ Buf.new(1);
camelia rakudo-moar 409cdb: OUTPUT«False␤»
psch jnthn: i'm guessing the for loop in MVMMultiCache.c:289 is what'd need to learn about rwness..?
hartenfels grondilu try `eqv`
PerlJam nine: you know, if dir() had an option to recurse, File::Find could be obsolete :)
psch oh, unless MVM_multi_cache_add doesn't even keep track of that... 18:41
hartenfels m: say Buf.new(1) eqv Buf.new(2), Buf.new(1) eqv Buf.new(1)
camelia rakudo-moar 409cdb: OUTPUT«FalseTrue␤»
nine PerlJam: but File::Find also supports filtering 18:43
18:43 isBEKaml left
nine as does dir... 18:44
18:44 Begi1117 joined
PerlJam yep (dunno when that got added, but I noticed it the other day) 18:44
18:46 andreoss` joined 18:47 Begi1116 left, andreoss left, eliasr left 18:49 xfix joined, _mg_ joined 18:50 CIAvash joined 18:51 kjs_ joined 18:54 andreoss` left 18:57 andreoss joined 18:58 andreoss left
psch well, all those different lookup function in MVMMultiCache.c scare me :P 18:58
hankache how do i execute external commands?
18:59 andreoss joined
hankache the equivalent of exec() or system() in Perl 5 18:59
grondilu m: say Buf.new(1) eqv Buf.new(2);
camelia rakudo-moar 409cdb: OUTPUT«False␤»
grondilu m: say Buf.new(1) eqv Buf.new(1);
camelia rakudo-moar 409cdb: OUTPUT«True␤»
grondilu hartenfels: thanks
19:00 sufrostico joined
flussence «Stage parse : 81.476» -- was 89 a few days ago! 19:00
hartenfels grondilu: np
psch hankache: shell or run probably, see doc.perl6.org/routine/shell (and s/shell/run/ respectively) 19:01
hankache gracias psch 19:02
will look into it
jnthn psch: You'll need to update add and find, and I suggest using a bit flag for "is in a container" 19:03
psch: Much as we do for definedness 19:04
psch jnthn: how feasible is that for me without any understanding of spesh? assuming i need to take care of MVM_multi_cache_find_spesh as well 19:05
timotimo i think that's just a version of that that makes sure to not invoke anything and not fail or anything
jnthn psch: I don't think it'd need much spesh understanidng. 19:07
psch alright, i'll flail my subpar C around a bit then :P
19:08 ribasushi left
psch but i guess that means it also needs to pay attention to the new bit flag 19:08
jnthn Sure, you set it in add, and check it in find 19:09
I *think* there's a spare bit
19:10 andreoss left
jnthn arg_tup[i] = STABLE(arg)->type_cache_id | (IS_CONCRETE(arg) ? 1 : 0); 19:11
19:12 yqt joined, ribasushi joined
jnthn Those type cache IDs are 64 bits apart 19:12
19:12 eliasr joined
jnthn But some bits are used for native types too 19:13
So that may need tewakery to allow an extra bit
[Sno] jnthn: btw - would be great to see an example for windows bat files how to use default directories for nqp/perl6 bootstrappath 19:14
timotimo "64bits apart"? like, "aligned to a full 64bit"? :P
jnthn timotimo: uh...64 apart :) 19:15
[Sno]: I'm not really clueful about batch files, I'm afraid. 19:16
jnthn has to google every time he writes one :) 19:17
19:18 andreoss joined
[Sno] but you have a big advantage - a running system (even when I do it in a VM, I never know whether I broke it or it has to be done somehow else ^^) 19:18
jnthn True :) 19:23
19:25 _mg_ left 19:31 kjs_ left
hankache why does this work 19:35
m: my $command = run 'perl6', '-v', :out; print $command.out.slurp-rest;
camelia rakudo-moar 409cdb: OUTPUT«run is disallowed in restricted setting␤ in sub restricted at src/RESTRICTED.setting:1␤ in sub run at src/RESTRICTED.setting:14␤ in block <unit> at /tmp/C2liWPd1nC:1␤␤»
19:35 softmoth_ left
hankache oops 19:35
i can't use run here?
moritz no 19:36
hankache ah ok
moritz (you can scream, though :-)
hankache so why does this work: my $command = run 'perl6', '-v', :out; print $command.out.slurp-rest;
and not this my $command = run 'perl6 -v', :out; print $command.out.slurp-rest;
pink_mist because 'perl6 -v' is not a binary on your system, presumably? 19:37
moritz hankache: because you're trying to run a binary 'perl6 -v', and that doesn't exist
ugexe run takes an argument list, not a string
moritz hankache: you can use shell() though
which takes a single string that a shell will interpret and execute
hankache ok
19:38 _mg_ joined
hankache so if i use run i need to split the exec and the args 19:38
moritz only if you get them from a single string
... in which case you can also use shell instead of run
hankache i can't control the result with shell 19:39
that's why i used run
shell automatically outputs the result 19:40
jnthn You can capture the output stream instead. 19:41
shell('...', :out) or so
hankache moritz ugexe pink_mist thanks
jnthn will try it 19:42
jnthn indeed it works 19:43
psch bits 2 to 6 or so looked free, from a bit of gdb poking 19:44
and apparently i have it working spesh-less
hankache ok so any advantages of one overt the other?
over*
skids hankache: you're less likely to run into security isues if you sanitize the args and use run(). 19:45
hankache excellent. thanks skids 19:46
19:49 ^elyse^ left
AlexDaniel what skids said! 19:53
hankache: though there is not much sanitizing required, since args are passed as is
jnthn psch: Did you try some native arg multi-dispathces? 19:55
2-6 may be used in that
19:55 andreoss left 19:56 DrPete_ joined
hankache AlexDaniel but args will be coming from different places so instead of conacatenating theem in a string 19:56
it's easier to pass them as a list
psch m: multi f(int $x is rw) { "rwint" }; multi f(Int $x) { "Int" }; multi f(str $x) { "str" }; f 2; f my int $x = 5; # works locally as well
camelia ( no output )
AlexDaniel hankache: yes, and it is more correct :)
psch oh but eh
i'm not actually checking rwness nor output 19:57
jnthn :)
psch huh 19:58
AlexDaniel hankache: in perl5 though ｢system｣ will degrade to shell if you pass only one parameter, I'm happy that perl6 does not have such pitfall
hankache AlexDaniel run will not work wi 19:59
with one param
AlexDaniel hankache: by the way, if you have to pipe several commands you can still do that with run. See doc.perl6.org/type/Proc
ugexe perl6 -e 'run("ls")'
bin Build.pm hooks lib logotype META6.json README.pod t xt
psch well, what i'm actually seeing locally is that it doesn't dispatch to the rw cand at all, neither for int nor num 20:00
so uh, mabe i should actually try to figure out what all those bits mean :P 20:01
AlexDaniel ugexe: by the way, it should be noted that ｢ls｣ output is unparseable
flussence wishes he knew some nice, readable way to define/validate nativecall bitflag thingies in terms of p6 enums
AlexDaniel ugexe: which is why I dislike such examples… normally you will never see run(‘ls’) (or similar) in real code 20:02
RabidGravy flussence, the Audio::Sndfile does that a bit 20:04
flussence ooh, thanks. will have a look...
RabidGravy the file types are bit masked
20:05 mr_ron left 20:08 rindolf left
flussence well that code looks pretty nice and readable, even despite github's highlighting butchering most of the file :) 20:10
psch hm, not actually seeing a free bit here 20:11
one arg_tup came out to 100110111100001, and type_cache_id for another was 111101101000000 20:12
the only zeros in both are 5 to 1 (or 6 to 2, depending if the right most is 1 or 0...) and those are MVM_CALLSITE_ARG_*
(that's still not "i know what this means", but...) 20:13
jnthn psch: Yeah. I think that you can s/64/128/ for the type_cache_id increment, and then << 2 the native type flag (and >> 2 in find) 20:14
psch: And that frees bit 2 to be your container bit
psch jnthn: oh, so the highest bit is the one that can be shifted off? 20:15
jnthn The << 2 comment refers to the line like:
arg_tup[i] = (arg_type << 1) | 1;
arg_tup[i] = (arg_type << 2) | 1;
That
oops
It's << 1 now, I'm suggesting << 2
No, 2nd counting from the LSB 20:16
20:16 Begi1118 joined, kjs_ joined
psch jnthn: i'm still stuck on the type_cache_id increment... 6model.c has a + 64 there..? 20:17
20:17 Begi1117 left
jnthn Right, and a 64 in the atomic increment too 20:17
Make those both 128
That just means you get an extra certain-to-be-zero bit
timotimo .o( might as well turn that into a #define'd constant ) 20:18
jnthn Aye
We could also just atomic incr the type cache ID and then shift if by the constant
*shtift it
*shift
heh, good job I'm not writing code with this typing :P
psch o.o
flussence oh that's neat, «unit enum Foo» DWIM.
(well, in a kinda pointless way since all the important code comes before the semicolon, but still.) 20:20
timotimo sort of like a =begin end
or =finish or something
flussence looks like you can hang subs off it like a normal module. That seems like it'll come in useful for that bitfield thing I asked about... 20:22
20:23 zakharyas joined
timotimo interesting. will they just be inside its .WHO? ... like the enum values themselves? 20:24
that'd be a bit weird
flussence m: enum Fruit <apple pear> { our sub tomato {'tomato‽'} }; say Fruit::tomato 20:25
camelia rakudo-moar 409cdb: OUTPUT«5===SORRY!5=== Error while compiling /tmp/PBdSHRhL5S␤Unexpected block in infix position (missing statement control word before the expression?)␤at /tmp/PBdSHRhL5S:1␤------> 3enum Fruit <apple pear>7⏏5 { our sub tomato {'tomato‽'} }; say Fr…»
flussence doesn't work like that, though; hm. 20:26
ugexe m: my %x; %x<y>.grep(*.so)
camelia rakudo-moar 409cdb: OUTPUT«Cannot call iterator(Any: ); none of these signatures match:␤ (Any:D $: *%_)␤ in block <unit> at /tmp/t9Y6n14ABm:1␤␤»
flussence m: unit enum Fruit <apple pear>␤our sub tomato {'tomato‽'}␤say Fruit::tomato
camelia rakudo-moar 409cdb: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Ltq9xeP5w1␤Bogus statement␤at /tmp/Ltq9xeP5w1:2␤------> 3o7⏏5ur sub tomato {'tomato‽'}␤ expecting any of:␤ whitespace␤»
flussence m: unit enum Fruit <apple pear>; # missing semicolon ␤our sub tomato {'tomato‽'}␤say Fruit::tomato
camelia rakudo-moar 409cdb: OUTPUT«tomato‽␤»
20:27 softmoth_ joined
timotimo dude 20:27
that's weird! :)
flussence though that first error makes it sound like I'm doing something that's not *supposed* to work, so I'm not sure...
20:28 mr_ron joined
El_Che I attended a Python course at work where the instructor was very curious about Perl6. 20:29
dalek ast: 5df60df | usev6++ | S32-str/sprintf.t:
Expect 'Inf' when 'Inf' is passed to sprintf

that's more consistent than 'inf' or 'INF'
jnthn That's nothing, one of the Python teachers at my work has a Perl 6 commit bit :P
flussence It seems to work identically whether or not the “unit” keyword is there, so maybe that just gets ignored.
20:30 hankache left
ugexe my %x; %x<y>.grep(*.so) # is this correct in not working anymore? 20:30
El_Che jnthn: you win :)
flussence ah, yeah, that was it. $?PACKAGE never gets set. 20:31
lizmat m: my %x; %x<y>.grep(*.so)
camelia rakudo-moar 409cdb: OUTPUT«Cannot call iterator(Any: ); none of these signatures match:␤ (Any:D $: *%_)␤ in block <unit> at /tmp/4fapro0hIE:1␤␤»
ugexe m: sub foo (:$a) { say $a.grep(*.so) }; foo()
camelia rakudo-moar 409cdb: OUTPUT«Cannot call iterator(Any: ); none of these signatures match:␤ (Any:D $: *%_)␤ in sub foo at /tmp/JW77TShj2D:1␤ in block <unit> at /tmp/JW77TShj2D:1␤␤»
lizmat I guess I broke that... 20:32
20:32 ^elyse^ joined
lizmat ah...I know what's wrong 20:32
[Coke] m: for 1,2,3,4 -> $x { $x.say } ; # should we make a non-texas variant of the -> ?
camelia rakudo-moar 409cdb: OUTPUT«1␤2␤3␤4␤»
flussence the subs inside still get namespaced into the enum though, which *is* weird. feels like a bug I probably can't rely on (but would be nice if it stayed) 20:33
lizmat ugexe: testing a fix
give me a few mins
ugexe cool, glad you have an idea 20:34
lizmat m: Any.grep: *.so # golfed
camelia rakudo-moar 409cdb: OUTPUT«Cannot call iterator(Any: ); none of these signatures match:␤ (Any:D $: *%_)␤ in block <unit> at /tmp/pK8PsqkCga:1␤␤»
20:34 raoulvdberge joined
dalek kudo/nom: 2940709 | lizmat++ | src/core/Any.pm:
Type objects also want iterators, ugexe++ for spotting
 20:35
20:35 ^elyse^ left 20:36 rindolf joined
hartenfels Hey y'all, I made a module: github.com/hartenfels/Text-Markdown-Discount, can someone tell me how to add it to the module ecosystem? 20:37
RabidGravy make a PR adding your projects META.info to the github.com/perl6/ecosystem/blob/ma.../META.list 20:39
hartenfels RabidGravy: Alright, thank you. 20:40
20:45 telex left 20:46 telex joined
dalek osystem: 0c79e57 | (Carsten Hartenfels)++ | META.list:
Add Text::Markdown::Discount to ecosystem

See github.com/hartenfels/Text-Markdown-Discount		 20:48
osystem: 5ac2a71 | RabidGravy++ | META.list:
Merge pull request #90 from hartenfels/master

Add Text::Markdown::Discount to ecosystem
20:50 lostinfog left
[Coke] hartenfels++ 20:50
20:51 CIAvash left 20:53 rindolf left
PerlJam hartenfels++ indeed. 20:53
20:55 _mg_ left, kaare_ left
hoelzro hartenfels++ 20:55
20:56 domidumont left
PerlJam I wonder if at some point we'll get something like: use Clib :stdio; to get at fopen and friends. 20:56
zacts hello #perl6
PerlJam greetings zacts 20:57
zacts :-)
I figured out how to make an ASCII / UTF-8 Camelia >%<
I think it's cool, I might make a t-shirt of it for myself 20:58
with the >%< zoomed in
and like #!/usr/bin/env perl6 or something
with a 'say ">%<";' or something like that? but yeah 20:59
just an idea
timotimo zacts: the official utf-8 camelia is »ö« 21:00
zacts ah yeah? I didn't realize that... that's cooler
=)
hartenfels Oh that was quick. And thanks for the postincrements. 21:01
zacts I see it in the topic now though
21:03 zakharyas left
hartenfels PerlJam I'd definitely use those instead of rolling my own stdio bindings. Might be worth to start a Clib.pm6. 21:05
RabidGravy or stick it in the existing POSIX 21:06
21:06 ifim joined
timotimo i think it's been in camelia.txt before it was in the topic 21:06
hartenfels: do you know "markdeep"? is there any sense in wanting that as a perl6 module? i think its only(?) implementation is javascript :|
retupmoca github.com/cygx/p6-native-libc/ looks like it might have fopen stuff already 21:07
PerlJam sweet
21:07 abaugher left, jantore joined 21:08 abaugher joined, softmoth_ left
hartenfels timotimo I don't know it, but I guess it'd be glueable if you called into nodejs or some other JavaScript library. 21:09
retupmoca Oh neat, I'll just depend on that instead. 21:10
PerlJam though it doesn't look like Native::LibC groups exports ... it would be nice to tag things with :stdio :stdlib :string etc. IMHO rather than have to list off all the imports explicitly 21:11
hartenfels Also seems to be lacking fdopen, but both those should be an easy enough patch. 21:13
21:16 cygx joined
PerlJam is the construct "is native(Str)" guaranteed future proof? (i.e., it won't blow up because of an undefined value at some point) 21:16
cygx PerlJam: there already is Native::LibC in the ecosystem
hartenfels It's documented that way in the NativeCall documentation at least.
cygx PerlJam: I'm busy right now with Other Things, but if you have a feature request... 21:17
21:17 habamax joined
PerlJam cygx: I think hartenfels has one (maybe in the shape of a PR) 21:17
21:18 mr_ron left
PerlJam cygx++ btw. 21:18
hartenfels Soonish at least.
If I can think of a sane way to write a test for fdopen 21:19
PerlJam hartenfels: Care to write an advent post? perl6advent.wordpress.com/ (not right now, but around advent season) We could always use more articles. 21:22
cygx: you too re advent article. 21:23
hartenfels PerlJam: Sure, but about what? Most of what I learned was of previous advent calendar posts.
PerlJam hartenfels: your experience writing Text::Markdown::Discount. What it's like to use Perl 6 from your perspective. etc. 21:24
hartenfels: anything you think is neat about Perl 6
cygx note that fdopen is from POSIX, not the C standard lib
so I'd put it into a Native::POSIX module (but still populate the Native::LibC namespace so it gets mapped into libc::, which is how I organize things right now) 21:25
21:25 habamax left
cygx PerlJam: I can probably come up with something to write about 21:25
hartenfels PerlJam: Perl6 is neat all around, aside from its startup time. But I'll see if I can think of something interesting to write. 21:26
PerlJam cool
moritz idly wonders if CVE-2015-5307 means he should reboot master.p6c.org 21:27
hartenfels cygx: Oh, I wasn't aware of that. Then I should probably take that out of my module for the sake of portability anyway.
cygx hartenfels: it's available on Windows as _fdopen, so it is reasonably portable 21:28
it's just not part of the C language standard lib 21:29
21:29 espadrine joined, sufrostico left
cygx the standard lib doesn't do file descriptors 21:29
hartenfels That's what I checked too, so I just assumed it was standard. While the function on its own might work, I don't wanna depend on a Native::POSIX that might not install on systems the module would otherwise work on. 21:30
21:36 TEttinger joined 21:39 njmurphy joined, softmoth_ joined 21:40 espadrine_ joined, espadrine left
grondilu m: my @a; @a[3..4] = rand xx 2; say @a 21:44
camelia rakudo-moar 294070: OUTPUT«[]␤»
grondilu m: my @a; @a[3,4] = rand xx 2; say @a
camelia rakudo-moar 294070: OUTPUT«[(Any) (Any) (Any) 0.434058554835957 0.714992255868652]␤»
dalek ast: dca1fae | usev6++ | S04-declarations/my.t:
Add test for RT #87034
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...l?id=87034
grondilu m: say (my @)[3..4] = rand xx 2;
camelia rakudo-moar 294070: OUTPUT«()␤»
grondilu m: say (my @)[3..4] = 1, 2
camelia rakudo-moar 294070: OUTPUT«()␤»
grondilu m: say (my @)[3,4] = 1, 2 21:45
camelia rakudo-moar 294070: OUTPUT«(1 2)␤»
grondilu looks like a bug
m: say (my @)[my @ = 3..4] = 1, 2
camelia rakudo-moar 294070: OUTPUT«(1 2)␤»
grondilu shouldn't @array[$range] = @stuff work as expected? 21:46
awwaiid whaaa 21:47
psch grondilu: that's RT #124316
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=124316
awwaiid what is this (my @) thing? 21:48
psch currently works with explicit flattening, not sure what the desired behavior is
uh, well not in that anon case from the looks of it
21:48 n0xff joined
psch m: my @a; my @b; @a[0,1] = 42,43; say @a.perl; @b[^2] = 42,43; say @b.perl; @b[| ^2] = 12,13; say @b.perl # the case in the ticket 21:49
camelia rakudo-moar 294070: OUTPUT«[42, 43]␤[]␤[12, 13]␤»
grondilu awwaiid: anonymous array
awwaiid link to doc?
psch m: say (my @)[| (3..4)] = 1, 2 # just precedence...
camelia rakudo-moar 294070: OUTPUT«(1 2)␤»
psch although arguably that should be (Any), (Any), (Any), 1, 2..? 21:50
grondilu awwaiid: in S02: "Bare sigils may be used as placeholders for anonymous variables:"
synbot6 Link: design.perl6.org/S02.html#
awwaiid thanks!
grondilu btw anonymous variables were introduced relatively recently if I'm not mistaken, yet it's a feature I can't live without :) 21:51
21:51 mr_ron joined
Hotkeys in what situations would one use anonymous variables 21:52
awwaiid are there non one-liner use cases?
ya, what Hotkeys said
psch i liked the toggle sub someone brought up recently 21:53
or whatever it was called
arnsholt When you're implementing an interface method, and your implementation only needs some of the arguments
skids m: sub a { state $++; }; a().say; a().say
camelia rakudo-moar 294070: OUTPUT«0␤1␤»
arnsholt That's a use for anonymous variables
Or deconstructing a multiple-element list (a la my ($foo, $bar, $blech)), and you only need a couple of the values 21:54
awwaiid arnsholt: ah, so like _ in ruby and others, but kinda typed?
arnsholt $ is exactly like _
But with types, since Perl 6 has types, yeah
21:55 [Sno] joined
gfldex m: my ($,$foo) = 1,2; say $foo; 21:55
camelia rakudo-moar 294070: OUTPUT«2␤»
Hotkeys is there a prettier way to do inline for stuff with multiple variables (I know you can use $_ when you only need one
ie
er
eg
-> $k, $v { say "$k: $v" } for %stuff.kv
like python has the pretty 'foo(x) for k, v in stuff.kv' 21:56
gfldex m: sub oi {$ ^= 1}; say oi, oi; # i don't know what this does, but it seams not to work correctly
Hotkeys (that's some pseudo stuff)
21:56 spider-mario joined
camelia rakudo-moar 294070: OUTPUT«Memory allocation failed; could not allocate 78848 bytes␤» 21:56
Hotkeys I know you can also do 'for %stuff.kv -> $k, $v { say "$k: $v" }" 21:57
psch m: my $a ^= 1, 2; say $a.perl # gfldex
camelia rakudo-moar 294070: OUTPUT«WARNINGS:␤Useless use of constant integer 2 in sink context (line 1)␤one(one(), 1)␤»
Hotkeys which is nicer
psch gfldex: i think you're misunderstanding ^=
AlexDaniel Hotkeys: $^var1 $^var2 perhaps?
grondilu you can also use %stuff.pairs
Hotkeys oh
psch but yes, the overflow should probably get fixed... :l
Hotkeys can you do that @ AlexDaniel
retupmoca m: my %h = (a => 1, b => 2); for %h.kv { say "$^k: $^v" } 21:58
camelia rakudo-moar 294070: OUTPUT«a: 1␤b: 2␤»
Hotkeys grondilu: that was only an example
oh fancy
AlexDaniel Hotkeys: yeah, I love it
grondilu m: my %h = (a => 1, b => 2); for %h { say .key, ": ", .value } 21:59
camelia rakudo-moar 294070: OUTPUT«a: 1␤b: 2␤»
skids It's recursively building a Junction of and Junction of...
AlexDaniel Hotkeys: though I mostly use it in subs
Hotkeys grondilu: I didn't mean for hashes specifically :P
I meant just like
gfldex psch: that is very likely, but should that fail the way it does?
Hotkeys multiple variables in situations where one would use $_
grondilu Hotkeys: oh yeah, my bad
Hotkeys for one variable
psch gfldex: the issue is what skids++ says. i don't know if we can know that you don't want to recursively build something until it's too late
AlexDaniel Hotkeys: yeah, that's pretty much what $^var thingies do
grondilu yeah self declared variables are the way to go then 22:00
Hotkeys perl6 gets mad when I used placeholders in the repl
but only sometimes
skids m: sub oi {$ +^= 1}; say oi, oi; # maybe what was originally intended
camelia rakudo-moar 294070: OUTPUT«10␤»
Hotkeys like 22:01
m: my %h = (a => 1, b => 2); say "$^k: $^v" for %h.kv
camelia rakudo-moar 294070: OUTPUT«5===SORRY!5=== Error while compiling /tmp/l0SlDJaq4Y␤Cannot use placeholder parameter $^k in the mainline␤at /tmp/l0SlDJaq4Y:1␤------> 3my %h = (a => 1, b => 2); say "$^k7⏏5: $^v" for %h.kv␤»
psch Hotkeys: placeholders are effectively parameters that add themselves to the signature
Hotkeys am I stuck with proper for blocks
psch Hotkeys: the mainline doesn't have a signature
Hotkeys for that kind of thing
jdv79 wonders what "we can do better than cpan" means. 22:02
retupmoca m: my %h = (a => 1, b => 2); { say "$^k: $^v" } for %h.kv; # you can do it this way - it just needs to be a block
camelia rakudo-moar 294070: OUTPUT«a: 1␤b: 2␤»
jdv79 TimToady: ^
Hotkeys i guess
rip 22:03
lizmat jdv79: I think this is more about the cpan client (installation), then about PAUSE / CPAN / MetaCPAN 22:04
22:04 skids left
grondilu retupmoca: I'm a bit surprised this work. Shouldn't there be () after the block? 22:06
m: { say "$^a" } for <hello bye> 22:07
camelia rakudo-moar 294070: OUTPUT«hello␤bye␤»
grondilu The block is automatically executed.
retupmoca no, because adding () would try to call the block before the 'for' ever got to see it 22:08
grondilu oh yeah
psch m: { sub { say "$^a" }($_) } for <hello bye>
camelia rakudo-moar 294070: OUTPUT«hello␤bye␤»
retupmoca actually in this case it might try to call the block with 0 arguments for each iteration
psch m: sub { say "$^a" }($_) for <hello bye> # superstituous block even 22:09
camelia rakudo-moar 294070: OUTPUT«hello␤bye␤»
retupmoca right, what psch showed
psch m: { say "$^a" }($_) for <hello bye> # superstituous sub 22:10
camelia rakudo-moar 294070: OUTPUT«hello␤bye␤»
psch and if you add to that that $^a acts a signature builder it makes sense to work without the parens
if you add parens without $_ inside you get a "too few arguments"
m: { say "$^a" }() for <hello bye> # as here 22:11
camelia rakudo-moar 294070: OUTPUT«Too few positionals passed; expected 1 argument but got 0␤ in block at /tmp/EdWBb7q51z:1␤ in block <unit> at /tmp/EdWBb7q51z:1␤␤»
22:21 espadrine joined, espadrine_ left 22:22 xfix left
mrf .tell ugexe I am going to push these tests to a branch in my fork for you to look at. They are currently failing but I can't for the life of me work out why: github.com/ungrim97/Perl6-Grammar-...e/refactor 22:22
yoleaux mrf: I'll pass your message to ugexe.
22:22 zengargoylew left 22:27 llfourn joined
dalek c: 71f2532 | VZ++ | doc/Language/syntax.pod:
Fix examples showing underscores in numbers

It was somewhat confusing to claim that `1_000_000` and `10_00_00` evaluate to the same number, make sure that all numbers have the same number of zeroes, even if they have different numbers of underscores.		 22:28
c: 5a6575e | RabidGravy++ | doc/Language/syntax.pod:
Merge pull request #192 from vadz/patch-1

Fix examples showing underscores in numbers
Skarsnik _ in number? it's weird lol 22:29
RabidGravy he
llfourn I just watched the p6 announce vid. It was great :D 22:30
Larry menetioned a list of things that are definitely not going to be in 6.christmas made by jnthn. Anyone got a link? 22:31
psch llfourn: gist.github.com/jnthn/040f4502899d39b2cbb4 22:32
llfourn psch: cheers :)
22:33 tardisx joined 22:39 ifim left 22:40 Begi1119 joined 22:42 Begi1118 left 22:46 muraiki left 22:49 Begi1119 left, gfldex joined 22:54 yqt left, kjs__ joined 22:55 kjs_ left, kjs__ is now known as kjs_ 22:56 tardisx left, tardisx joined
grondilu m: multi infix:<*>(Real $k, $v where [&&] map { $_ ~~ Real }, @$v) { [ $k X* @$v ] }; say pi * [1, 2]; 22:58
camelia rakudo-moar 294070: OUTPUT«[3.14159265358979 6.28318530717959]␤»
grondilu m: multi infix:<*>(Real $k, $v where [&&] map { $_ ~~ Real }, @$v) { [ $k X* $v ] }; say pi * [1, 2];
^unexpected infinite recursion
camelia rakudo-moar 294070: OUTPUT«(timeout)» 22:59
23:05 kid51 joined
timotimo someone said "anonymous variables are a recent addition"; that's only the syntax for @[...] and %{...} i believe 23:07
grondilu I did say it. I think it's no more than a year or two. 23:13
(so "recent" was very relative) 23:14
ugexe mrf: my guess would be the use of | vs ||. `||` represents what the spec as written in the rfc wants 23:18
yoleaux 22:22Z <mrf> ugexe: I am going to push these tests to a branch in my fork for you to look at. They are currently failing but I can't for the life of me work out why: github.com/ungrim97/Perl6-Grammar-...e/refactor
dalek kudo/nom: 4422d81 | lizmat++ | src/core/ (10 files):
Don't use a !=:= b, but !(a =:= b)

Because it's at least 25x faster. Either this, or we may want to
  prime the grammar for !=:= (because they are used a lot in iterators,
and iterators imply hot paths, generally). Not directly noticeable speedup in spectest, but some iterator rich apps may notice improved performance.		 23:19
lizmat good night, #perl6!
Skarsnik should the vm/compilater do it if it's faster? x) 23:20
ugexe mrf: otherwise you can end up with obs tokens and whatnot that technically match longer, but are expected to match as a last resort if nothing else matches 23:22
23:22 spider-mario left, perlawhirl joined
perlawhirl is anybody home? 23:25
gfldex m: say 'hi perlawhirl!'; 23:26
camelia rakudo-moar 294070: OUTPUT«hi perlawhirl!␤»
zengargoyle sorta here
llfourn m: say "I'm always here";
camelia rakudo-moar 294070: OUTPUT«I'm always here␤» 23:27
perlawhirl hi... i'm trying to work out if i'm able to pass a new rule to a grammar
zengargoyle just precompiled panda which shaved ~1.5s off `panda list` for ~2x speedup
perlawhirl so i have the TOP rule... and tokens... can i override the TOP with a new rule made up of tokens that are defined in the grammar, and use that when i call .parse
llfourn perlawhirl: do you mean can you chose to start with a different rule other than TOP? 23:28
zengargoyle wonder why there's not a rakudobrew or panda --precompile flag
ugexe because precompile was taken out
23:28 rindolf joined
grondilu precompiling panda must be made manually? 23:29
perlawhirl kind of... except the only rule in the grammar is TOP. say i want to re-order the tokens, can i make a new rule and pass that to the grammar?
grondilu I mean doesn't it do it itself each time we ./rebootstrap or something?
perlawhirl eg. my rule thingy { <token2> <token1> }; MyGrammar.parse($string, rule => thingy); 23:30
something like that.
llfourn perlawhirl: you can just make a role { rule my-rule { ... } }
zengargoyle was there a reason behind taking out precompiling? (i think i remember whn it did)
llfourn perlawhirl: and then MyGrrammer.new does MyRole
zengargoyle perlawhirl: yes, there's an option to .parse to specify a rule to start with.
llfourn zengargoyle: though perlawhirl wants to use a specific rule object made at runtime 23:31
perlawhirl zengargoyle: I know you can have multiple rules defined in the grammar, and tell it which one to use... but what if i want to define another rule outside of the grammar
llfourn: i'll look into this role business
zengargoyle doc.perl6.org/type/Grammar#___top 23:32
llfourn perlawhirl: you can actually MyGrammar.^add_method('thingy',my rule thingy { ... }) # I think
but applying a role to an instance is better practice 23:33
zengargoyle can't you just access it as Grammar::rule
a Grammar is just a fancy class and a rule is just a fancy method
23:34 softmoth_ left, softmoth_ joined
llfourn zengargoyle: but you have to still call it on the class to get all the sub-regexes to work 23:34
s/class/grammar/
zengargoyle and/or subclass it or maybe add a role. not sure, haven't tried.
perlawhirl hmm, that sounds like it should work too... will play around a bit more 23:35
zengargoyle well, guess it does depend on what's in the rule, if it's only things like <ws> <alpha> etc. that are already defined outside of the grammar...
llfourn m: grammar G { }; role R { }; say G.new does R
camelia rakudo-moar 4422d8: OUTPUT«G+{R}.new␤»
llfourn perlawhirl: something like that :) 23:36
zengargoyle at least that's what i'd try. :P
llfourn zengargoyle: it kinda breaks encapsulation that way...plus wouldn't the rules have to be 'our'. 23:37
23:37 mr_ron left
zengargoyle me thinks rule called from outside of grammar might still work... wouldn't rule have closed over any references to other things in the grammar at compile time? 23:37
perlawhirl ok... so subparse is kind of what i want... this works: G.subparse($string, :rule<token1>)
<token1> is already defined in G
zengargoyle ah, yeah, probably our. 23:38
psch perlawhirl: what's your use-case, ooc?
perlawhirl howver, how can i do with multiple tokens, ie. G.subparse($string, :rule<token2><token1>)
zengargoyle but then methods don't need an our on them.
zengargoyle brain hurts.
llfourn m: grammar G { token fun { } }; G.^find_method('fun') is probably what you mean zengargoyle? 23:39
camelia rakudo-moar 4422d8: OUTPUT«5===SORRY!5===␤Null regex not allowed␤at /tmp/86zIr1B6SD:1␤------> 3grammar G { token fun { 7⏏5} }; G.^find_method('fun') is probably w␤Two terms in a row␤at /tmp/86zIr1B6SD:1␤------> 3{ token fun { } }; G.^find_method('fun')7⏏5 …»
llfourn m: grammar G { token fun { } }; say G.^find_method('fun') # is probably what you mean zengargoyle?
camelia rakudo-moar 4422d8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/WexCJB19C7␤Null regex not allowed␤at /tmp/WexCJB19C7:1␤------> 3grammar G { token fun { 7⏏5} }; say G.^find_method('fun') # is prob␤»
perlawhirl psch: just playing atm to learn... but a use-case might be a date parser
zengargoyle i'm sorta guessing you could also subclass the grammar to take other parameters and pass in your new rule.
llfourn perlawhirl: you can't do it wit multiple tokens afaik 23:41
psch perlawhirl: i'm not seeing it... :)
perlawhirl: the questions as asked are mostly solved with inheritance
perlawhirl: e.g. "can i replace TOP?" -"yes, inherit and replace TOP in the subgrammar"
llfourn perlawhirl: you have to apply a role with whatever you want and then parse from there 23:42
psch similarly, if you want to replace one specific rule you can also override it in the subgrammar
23:43 softmoth_ left
psch if you want to add a new rule, well that only makes sense if the grammar you want to parse it with knows where to try to match it, so you'd probably have to override TOP again 23:43
zengargoyle i'm thinking like a Grammar to parse like a reciept but then you start getting recipts with different date formats.
so same grammar but with a different <date>?
perlawhirl zengargoyle: yes, another good example use-case. ALL receipts have <item> <cost> <date> and <total>... but in a different order 23:44
llfourn zengargoyle: if you want to keep the old format as a possibility you can use a multi which could match either
zengargoyle also not sure of 'why not ust subclass'
23:44 ddrown joined
psch m: grammar A { token TOP { <a> }; token a { a } }; grammar B is A { token a { aa } }; say A.parse("a"); say A.parse("aa"); say B.parse("aa") # like this... 23:44
camelia rakudo-moar 4422d8: OUTPUT«｢a｣␤ a => ｢a｣␤(Any)␤｢aa｣␤ a => ｢aa｣␤»
psch note the (Any), which is A failing to parse "aa" 23:45
perlawhirl yes... I have seen (Any) plenty of times while playing with grammars :)
zengargoyle perlawhirl: btw, have you tried playing with Grammar::Tracer? 23:46
23:46 kjs_ left
perlawhirl zengargoyle: nope... will check it out 23:47
sprocket on the subject of grammars, i’ve just started writing my first one and had popped in here to ask if anyone had a good, recent example of one i might refer to :)
zengargoyle it's a part of Grammar::Debugger i think.
JSON::Tiny
ADT had an interesting Grammar/Actions 23:48
sprocket thanks! :) 23:49
zengargoyle probably any module that hase Parse in it's name :)
23:50 skids joined 23:55 mr_ron joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!