»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
RabidGravy well don't do that then 00:00
00:01 kurahaupo left
timotimo you can't really write every algorithm to be branchless :) 00:01
not without pain, at the very least :D
00:02 obfusk left
RabidGravy no but you have an asymetric architecure and clever compilers 00:02
00:03 obfusk joined
RabidGravy anyhow I'm whacked 00:03
toodles
timotimo toodles, mister gravy
00:04 Skarsnik left 00:08 RabidGravy left 00:20 kid511 joined 00:22 kid51 left 00:24 itcharlie joined 00:30 huggable left, huggable joined 00:43 pierre_ joined 00:46 tbrowder left 00:48 hotel joined, BenGoldberg joined, pRiVi left, cpage_ joined 00:49 SquireOfGothos left, avar left, kurahaupo joined, hotel|lethargy left 00:50 avar joined, avar left, avar joined, literal joined 00:52 pRiVi joined 01:01 hotel_california joined 01:02 hotel left 01:05 cdg left, hotel_california is now known as hotel, cdg joined
ZoffixWin stmuk, what needs to happen to have my blog posts that show up on pl6anet.org/ to be from perl6.party and not from blogs.perl.org? [Eventually] the .party posts will look better and they also offer in-browser modifiable and runnable code examples, so I rather have that as default. 01:07
hotel how do you use panda on windows again? 01:09
I have to make some batch file right?
01:10 cdg left
ZoffixWin hotel, does panda command not work|? 01:12
hotel no, it hangs
and apparently this has been the case since 2013 perl6maven.com/getting-started-with...on-windows
ZoffixWin :/ 01:13
hotel, what Perl are you on? Did you use R* installation? 2016.04?
hotel yeah
ZoffixWin And what exact command does it hang on? (What Windows are you on, BTW)?
hotel I think it's because windows tries to call the shell script and doesn't know what to do
running panda update, or any panda command
>windows X 01:14
ZoffixWin X?
Oh 10
ZoffixWin only has 7
hotel hm 01:15
timotimo there are 7 windows in my apartment, too
hotel weird, the fix on that link worked on my other computer 01:16
ZoffixWin fwiw, `panda` "works" on Win7, but fails to fetch HTTP::Tinyish for some reason :S i.imgur.com/Dsfzffp.png 01:18
Oh... do I need some sort of git thing on Windows? 01:19
hotel git-scm.org
ZoffixWin That doesn't answer my question :D 01:20
ZoffixWin doesn't use Windows for anything other than gaming
01:21 mr-foobar left
ZoffixWin I'm just saying we have no mention of needing git things on Download page. If it's a requirement, it should be mentioned 01:21
ugexe zef does not require git on windows, but will use it if available fwiw
hotel huh? puu.sh/oHMQy/a054f301d3.png
01:22 Ben_Goldberg joined
itcharlie zef bro 01:22
01:22 BenGoldberg left
itcharlie thats what you need 01:22
01:22 Ben_Goldberg is now known as BenGoldberg
hotel teach me the ways 01:22
ZoffixWin ugexe, how to get zef without panda or rakudobrew?
01:22 BenGoldberg left
ZoffixWin or git things :P 01:22
ugexe ZoffixWin: download the .zip from github
ZoffixWin Right. Makes sense. ugexe++ 01:23
hotel ZoffixWin, I guess they assume that if someone is using perl6 on windows they already have some sorta git thing? ¯\_(ツ)_/¯
ugexe zef can also install .zip archive links from github
ZoffixWin hotel, well... try zef: github.com/ugexe/zef
hotel yeah I'm there :)
01:24 Ben_Goldberg joined
hotel man, windows really hates files without extensions 01:26
gfldex i just got `Failed to open file /usr/local/src/rakudo/install/share/perl6/site/precomp/.lock: permission denied` after rebuilding Rakudo
here the stacktrace: gist.github.com/gfldex/ec7a4acd1a2...230d4a27a4 01:27
ugexe i suppose there is still a flaw though: if the source-url of a module is a git link (and all of them are that are not on cpan) it won't try to guess its archive url. this means you would have to `zef install github.com/xxx/yyy/master.zip` for each module (or manually s/// the urls in the package list) 01:28
hotel ugexe, I don't even anymore: puu.sh/oHNck/af950d0f43.png
01:28 Ben_Goldberg left 01:30 Ben_Goldberg joined
hotel wow puu.sh/oHNhB/6f0c2c66e5.png 01:30
also Im dumb
ugexe heh
stupid unzipper putting it in a stupid folder of the same name as the child folder
hotel yeah 01:31
01:32 hotel is now known as hotel|chagrin 01:34 sno left, sno joined 01:35 Actualeyes joined
hotel|chagrin interestingly, in order to get cmd to recognise that zef is on the path, I have to call it zef.pl6.bat 01:38
ugexe that is from rakudo itself i believe 01:42
hotel|chagrin putting zef.bat (which calls the other bat) in rakudo/bin works though...
ugexe here is a modified $HOME/.zef/store/p6c/packages.json that has all the git urls turned into their .zip gist.github.com/ugexe/c408d997fcf5...f0889a9bd3
hotel|chagrin but anyway, zef works so thanks :) 01:43
01:43 hotel|chagrin is now known as hotel
ZoffixWin \o/ 01:43
ugexe strange about the .bat though, i'm able to invoke it with just `zef` 01:44
01:44 ilbot3 left
hotel yeah dunno 01:44
now I have to make this stop puu.sh/oHNZh/1fac35871e.png
ugexe the js backend? wth 01:45
hotel atom perl6 plugin thing idk
ugexe oh
hotel m: say comb "aesthetics"; 01:46
camelia rakudo-moar 780f07: OUTPUT«5===SORRY!5=== Error while compiling /tmp/DpRaj37jSg␤Calling comb(Str) will never work with proto signature ($,, $?)␤at /tmp/DpRaj37jSg:1␤------> 3say 7⏏5comb "aesthetics";␤»
01:47 ilbot3 joined
ugexe m: say comb //, "aesthetics" 01:48
camelia rakudo-moar 780f07: OUTPUT«5===SORRY!5=== Error while compiling /tmp/qVZXxTgZTp␤Null regex not allowed␤at /tmp/qVZXxTgZTp:1␤------> 3say comb //7⏏5, "aesthetics"␤»
hotel it works if you use your own sub
ZoffixWin m: say comb /./, "aesthetics";
camelia rakudo-moar 780f07: OUTPUT«(a e s t h e t i c s)␤»
hotel m: sub foo($t) { return $t.comb; } say foo "hello";
camelia rakudo-moar 780f07: OUTPUT«5===SORRY!5=== Error while compiling /tmp/FEm7QYd0td␤Strange text after block (missing semicolon or comma?)␤at /tmp/FEm7QYd0td:1␤------> 3sub foo($t) { return $t.comb; }7⏏5 say foo "hello";␤ expecting any of:␤ infix␤ …»
hotel m: sub foo($t) { return $t.comb }; say foo "hello";
camelia rakudo-moar 780f07: OUTPUT«(h e l l o)␤»
ZoffixWin : 01:49
:\
hotel, that's nonsense.
hotel how so?
ugexe looks like method comb and sub comb are missing that relation 01:50
ZoffixWin hotel, comb sub always needs a matcher. Method form breaks up into characters without a matcher.
hotel, so instead of actually giving it a matcher, you're creating a whole 'nother sub just to call it with a default matcher that breaks up into characters
ugexe no matcher method comb has been touted for splitting a string into individual letters though
hotel ^ 01:51
ZoffixWin It uses substr, yes: github.com/rakudo/rakudo/blob/nom/...tr.pm#L121
ugexe creating a new sub for it though is indeed pointless
m: say "hello".comb
camelia rakudo-moar 780f07: OUTPUT«(h e l l o)␤»
hotel hey I'm learning don't bite my head off
ZoffixWin ¯\_(ツ)_/¯ 01:52
ZoffixWin is just talking
ugexe i dont think anyone knows all of perl6, dont sweat it
hotel haha
ZoffixWin FWIW: perl6.party/post/20160425-Perl6-Comb-It
surprisingly, I never bother to mention any usecase of the sub form of .comb :/ 01:53
hotel I mean, of course I wouldn't make a sub to call comb normally, I was just making sure stuff worked :P
01:54 colomon left
ZoffixWin hotel, IMO, you should rakudobug this. I see no reason why sub comb(Str) should not break up into chars: github.com/rakudo/rakudo/blob/c403...ol.pm#L338 01:54
01:54 Ben_Goldberg left
ZoffixWin For consistency with the method form, it should 01:54
hotel what do you mean?
like it shouldn't require a match? 01:55
grondilu tadzik: panda question: do I still have to run ./bootstrap.pl everytime I recompile rakudo
?
01:55 Ben_Goldberg joined
ZoffixWin hotel, right. comb "blah"; should behave the same as "blah".comb; And it's possible, but simply not implemented. Report it: github.com/rakudo/rakudo/#reporting-bugs 01:55
hotel alright, wilco
ZoffixWin hotel++ 01:56
ugexe or overengineer a solution with &comb.wrap or MONKEY-TYPING
ZoffixWin :)
01:57 Ben_Goldberg left
grondilu nevermind I had a syntax error in my META.info 01:58
01:58 Ben_Goldberg joined
ugexe you can probably get a PR in with `multi sub comb(Cool $input, $limit = *) { $input.comb($limit) }` at github.com/rakudo/rakudo/blob/c403...ol.pm#L342 (untested) 02:00
that proto ($, $, $?) might not like that though 02:01
02:01 BenGoldberg_ joined, Ben_Goldberg left, BenGoldberg_ is now known as Ben_Goldberg 02:02 Ben_Goldberg is now known as BenGoldberg, cpage_ left
hotel protos are default signatures? 02:03
ZoffixWin That creates ambiguity between comb Int, $input. No $limit should be there. As in method form: github.com/rakudo/rakudo/blob/c403...ol.pm#L246
02:04 BenGoldberg left
ZoffixWin still has no idea wtf protos are :P 02:04
hotel oh
02:05 BenGoldberg joined
hotel sets out to build an http server out of pieces of scrap metal and rusty tools 02:05
ZoffixWin heh... hotel++ # nice rt signature :PO
ugexe ($, $, $?) means 2 required positionals and an optional 3rd positional (like perl5)
ZoffixWin ugexe, what it means is clear, but why is it needed? What does it do? 02:06
ZoffixWin submits a PR to fix the comb() stuff along with the tests in roast
ugexe proto itself is for handling dispatching, so it would affect nextsame/callsame etc type stuff
hotel I didn't think it was clear
ZoffixWin hotel, it's good enough 02:07
ugexe it also gives you a single item to `is export` and whatnot
hotel no I mean what the protos mean
ZoffixWin Ah
ugexe or do stuff like `proto foo { {*}.say }; multi method foo { "42" }` 02:08
ZoffixWin Yeah, I still have no clue. Even with ugexe's explanation. Wouldn't dispatching handle the same with multies and what not?
ZoffixWin leaves this for "another day"
ugexe if you have multis from multiple roles/classes then i believe it helps act as the authority
hotel like a template or something?
ugexe thats how I use them 02:09
02:09 BenGoldberg left
hotel cool 02:10
02:10 BenGoldberg joined
hotel forgot how to import modules 02:10
ZoffixWin use blah;
ugexe `proto foo { {*}.say }; multi method foo(Int $a) { "42" }; multi method foo(Str $a) { "fourty two" }` -> xxx.foo(1) (says 42) xxx.foo("x") (says fourty two) 02:11
grondilu also the lexical scope must be defined on the proto
ugexe so you can also put some top level shared code in them
grondilu you can't do 'our multi'
hotel oh it's imperative not gerund
ZoffixWin++
ugexe s/.say/.IO/ and it will always return an IO::Path 02:12
02:14 cpage_ joined
hotel m: my constant H = 10; say H; 02:16
camelia rakudo-moar 780f07: OUTPUT«10␤»
02:16 noganex_ joined
hotel so constants don't need sigils? 02:16
ZoffixWin nope 02:17
m: my \Δ = 10; say Δ;
camelia rakudo-moar 780f07: OUTPUT«10␤»
hotel is \ shorthand for constant? 02:19
ZoffixWin I wouldn't say shorthand. I think it's something to do with having containerless variables or something
02:19 noganex left
ZoffixWin is too drunk to give good advice :P 02:19
hotel again? lol 02:20
ZoffixWin I'm quitting tomorrow! I promise!
:P
02:22 pierre_ left
hotel tomorrow's friday though 02:23
can't expect you to :P
ZoffixWin For me, today is Friday :P I booked tomorrow off. 02:24
hotel what are the colon things in method calls again?
oh gg
ZoffixWin Like foo.bar: 'blah'? They let you omit using parentheses
hotel my $listen = IO::Socket::INET.new(:listen, :localport(3333));
ZoffixWin Oh, those are just pairs
my $listen = IO::Socket::INET.new(listen => True, localport => 3333); 02:25
hotel interesting
ZoffixWin named args
m: sub foo (:$name, :$num) { say "\$name is $name and \$num is $num" }; foo :name<Foo> :num<42> 02:26
camelia rakudo-moar 780f07: OUTPUT«$name is Foo and $num is 42␤»
ZoffixWin m: sub foo (:$name, :$num) { say "\$name is $name and \$num is $num" }; foo :name<Foo> :42num
camelia rakudo-moar 780f07: OUTPUT«$name is Foo and $num is 42␤»
ZoffixWin (you don't need commas if all you got are named args) 02:27
ugexe m: sub foo(:a(:b(:$c))) { say $c; }; foo( a => 42 ) # also can do aliases
camelia rakudo-moar 780f07: OUTPUT«42␤»
ZoffixWin :o
That's awesome ugexe++ 02:28
02:28 kid511 left
ugexe handy for writing short versions of MAIN args 02:28
02:29 cpage_ left
ugexe perl6 -e 'sub MAIN(:o(:$out)!) { say $out }' # Usage: -e '...' -o|--out=<Any> 02:30
02:32 nowan_ left 02:33 nowan joined
ZoffixWin this is annoying... spectest in a VM eats so much CPU it interrupts my music in Windows host :/ Windows-- 02:41
ZoffixWin spins up 20-core 98GB-RAM Linode box just because he can o/ 02:42
ugexe building jvm rakudo on windows would freeze my entire machine for a few seconds ever 10 seconds or so
also spins the fan up faster than any game i've put it through :x 02:43
hotel some spectests don't work on windows iirc 02:45
02:48 pierre_ joined
hotel is there a standard for using single quotes vs double quotes? I know double quotes do some extra fancy stuff, so is there really a reason to use single quotes? 02:50
ZoffixWin uses single quotes when he doesn't expect to have "extra fancy stuff" to happen 02:51
hotel is there a benefit to doing that though? 02:52
ZoffixWin Lack of bugs due to accidental interpolation
hotel ah
ZoffixWin Lack of bugs due to accidental interpolation 02:53
m: say "Email me at [email@hidden.address]
camelia rakudo-moar 780f07: OUTPUT«Email me at [email@hidden.address]
ZoffixWin :/
Well, at least that was my reasoning back when using Perl 5 :P
ugexe m: say "C:\t\foo" 02:54
camelia rakudo-moar 780f07: OUTPUT«C: oo␤»
02:54 Ben_Goldberg joined, BenGoldberg left
ugexe email addresses always get me in perl5 too 02:54
02:54 Ben_Goldberg is now known as BenGoldberg
ZoffixWin Ah, right, in P6 you need to have [] after arrays when interpolating 02:54
ugexe we get this one instead tho 02:55
m: say "{"
camelia rakudo-moar 780f07: OUTPUT«5===SORRY!5=== Error while compiling /tmp/h8d2QyyEn0␤Unable to parse expression in double quotes; couldn't find final '"' ␤at /tmp/h8d2QyyEn0:1␤------> 3say "{"7⏏5<EOL>␤ expecting any of:␤ double quotes␤ term␤»
hotel what are the rules for doing the thing where $foo.bar "baz"; works?
ZoffixWin Yup. Any time you try to put CSS in a interpolating heredoc :P
m: class F { method bar ($x) {"Foo $x"}}; my $meow = F.new; sasy "$meow.bar("42")" 02:57
camelia rakudo-moar 780f07: OUTPUT«5===SORRY!5=== Error while compiling /tmp/m0WoB2HqNq␤Undeclared routine:␤ sasy used at line 1. Did you mean 'say'?␤␤»
ZoffixWin ..... "sasy"? 0.o
Ohhh
m: class F { method bar ($x) {"Foo $x"}}; my $meow = F.new; say "$meow.bar("42")"
camelia rakudo-moar 780f07: OUTPUT«Foo 42␤»
ZoffixWin use just () for method calls with no args 02:58
hotel rip in peace puu.sh/oHREW/24b14fb92e.png
apparently the HTTP::Headers module doesn't add in the HTTP/1.1 part
ugexe m: class Foo { method bar { 42 }; }; say Foo.new."bar"() # ?
camelia rakudo-moar 780f07: OUTPUT«42␤»
ugexe ah 02:59
03:00 cpage_ joined
ZoffixWin For a 20-core 98GB-RAM box, Perl 5 is sure taking its sweet ass time installing and testing itself :/ 03:00
hotel are you on a server? 0.0 03:01
03:02 milwaukee joined
hotel also downloadmoreram.com because it sounds like you don't have enough 03:02
ZoffixWin yeah. I'm on Linode.com
pfft... what do you mean!? I have 98GB of it! :D
Default perlbrew perl5 tests are just stupid 03:03
ZoffixWin blames mst for all of this
Why? Just 'cause I can! :P
hotel just download some more ram ^^
ugexe to be fair, HTTP/1.1 isnt a header line. then again neither is the status line but that is there (200 OK) 03:05
03:05 BenGoldberg left
hotel oh good point 03:06
ugexe well, unless it in `Status: 200` or whatever format
03:06 milwaukee left
hotel is dumb again 03:07
03:07 hotel is now known as hotel|chagrin, BenGoldberg joined
ugexe i imagine his intention though is to have the http api separate so you can use 1.0 *or* 1.1. but its doubtful to work with 2.0 03:07
03:07 milwaukee joined
hotel|chagrin pshah who uses 2.0 anyway 03:09
03:09 BenGoldberg left
hotel|chagrin also, for some reason, writing "HTTP/1.1 200 OK" to the client's connection makes it so that nothing else is written 03:10
ugexe need more streams to get those ads to you faster
03:10 BenGoldberg joined
ugexe maybe its only feeding a \n instead of \r\n? 03:11
if thats supposed to be the entire thing, it needs to be \r\n\r\n 03:12
03:12 cpage_ left
hotel|chagrin gist.github.com/HotelCalifornia/54...9c27530eca 03:13
ugexe ah, yep 03:14
headers are separated from body by a double \r\n (\r\n\r\n)
hotel|chagrin oh
cool 03:15
ugexe you are sending a single \n
03:15 yqt left
hotel|chagrin did not know that 03:15
ugexe many servers accept \n in place of \r\n i think though, so \n\n may work
github.com/ugexe/Perl6-Grammar--HT...-grammar.t
there are some valid example requests
hotel|chagrin I'm reading a tutorial and translating from python 03:16
the guy used a docstring for his response so I missed the second newline
ugexe you can also use that grammar via `say Grammar::HTTP.parse($http-request)` to see if its a valid request 03:17
and if its not, put a `use Grammar::Tracer;` above it and it'll show you exactly how it gets parsed and where the match failed 03:18
hotel|chagrin oh that's what that does
03:18 hotel|chagrin is now known as hotel 03:24 cpage_ joined 03:26 grondilu left 03:27 cdg joined 03:28 BenGoldberg left 03:29 cognominal left, BenGoldberg joined
ZoffixWin time TEST_JOBS=1000 make spectest 03:31
real 3m51.235s
hotel well, guess it's time for bed 03:32
03:32 hotel is now known as hotel|netteiru
hotel|netteiru night 03:32
ZoffixWin Failures: gist.github.com/zoffixznet/8b630dc...adfe65565e
night
time TEST_JOBS=100 make spectest 03:33
1m53.111s
Failures: gist.github.com/zoffixznet/5ee59ff...dca01e6f75
03:33 dwarring left, BenGoldberg left 03:35 BenGoldberg joined 03:39 huggable left
ZoffixWin :o 03:39
Fuck me.... Deleted the wrong Linode by mistake lol 03:40
alcohol--
Oh well... free upgrade! 03:41
03:41 cdg left 03:42 sno left, sno joined 03:44 sno left, sno joined 03:48 astj joined 03:49 ZoffixWin left 03:53 khw left 03:56 Actualeyes left, BenGoldberg left 04:01 AlexDaniel left 04:07 aries_liuxueyang left 04:08 aries_liuxueyang joined 04:09 cdg joined 04:26 pierre_ left 04:32 jack_rabbit joined 04:47 cdg left 04:53 buharin joined 04:57 Cabanossi left 05:00 Cabanossi joined, astj_ joined 05:02 cgfbee left 05:03 astj left, zacts left 05:05 macsnowball left, zacts joined 05:06 Actualeyes joined 05:13 astj joined 05:14 ufobat joined 05:17 astj_ left
ufobat morning :) 05:18
05:23 buharin left
sortiz \o ufobat 05:23
05:24 aries_liuxueyang left, aries_liuxueyang joined 05:26 rurban joined, pierre_ joined
MadcapJake trying to understand channels more and I am getting "Tried to read() on a socket from outside its originating thread" gist.github.com/MadcapJake/c29b76d...orkers-pl6 05:29
05:29 nwc10 joined, Actualeyes left
nwc10 they're going to need a bigger hanger. 05:30
05:31 pierre_ left 05:33 itcharlie left 05:36 buharin joined 05:45 buharin left 05:47 Actualeyes joined 06:01 rurban left 06:02 pierre_ joined
nwc10 twitter.com/elonmusk/status/728459808270000128 :-) 06:02
06:11 Ven joined 06:20 firstdayonthejob joined 06:25 CIAvash joined 06:32 wamba joined
Ven > Invalid GC status observed; aborting 06:39
06:39 firstdayonthejob left
Ven ^ anyone used GTK::Simple in the past 6 months observed something akin to that? 06:40
who*
06:51 sjoshi joined, jack_rabbit left 06:57 pierre_ left 07:00 sjoshi left
masak mornz, #perl6 07:00
07:00 pierre_ joined
masak Ven: I got some segfaults recently when running on top of HTTP::Server::Tiny. but nothing GC-related 07:01
07:01 domidumont joined 07:04 abraxxa joined 07:05 domidumont left 07:06 domidumont joined 07:10 [particle] left 07:11 jack_rabbit joined, [particle] joined 07:12 sjoshi joined 07:15 xxpor joined 07:17 domidumont left 07:18 maybekoo2 joined, Relsak joined 07:21 jack_rabbit left 07:23 pierre_ left 07:31 pierre_ joined 07:47 rindolf joined, g4 joined 07:55 domidumont joined 07:56 chansen__ joined 07:58 chansen__ is now known as chansen_ 07:59 sno left 08:00 darutoko joined 08:01 sno joined 08:02 RabidGravy joined 08:04 jnap__ joined 08:06 tharkun left, jnap__ is now known as jnap_
tadzik grondilu: I don't think it's necessary anymore; what happens if you don't? :) 08:08
08:11 Ven left 08:19 kurahaupo left 08:24 vendethiel left
stmuk ZoffixWin: submit a PR for github.com/stmuk/pl6anet.org/blob/...perlanetrc 08:33
08:34 tharkun joined
stmuk BTW your use of a comma rather than colon meant irssi didn't alert me there! 08:34
08:41 Emeric joined
nadim Morning all 08:42
moritz \o 08:47
RabidGravy ero 08:51
08:53 pierre_ left
psch o/ 08:53
m: role R { multi method foo(::?CLASS:U: ) { "foo" } }; class C does R { multi method foo(::?CLASS:D: ) { "bar" } }; say C.^can('foo').perl; say C.new.^can('foo').perl;
camelia rakudo-moar 780f07: OUTPUT«(method foo (::T $: | is raw) { #`(Method|75023040) ... },)␤(method foo (::T $: | is raw) { #`(Method|75023040) ... },)␤»
psch i kind of wish there were two elems in that method list 08:54
m: role R { multi method foo(::?CLASS:U: ) { "foo" } }; class C does R { multi method foo(::?CLASS:D: ) { "bar" } }; say C.foo; say C.new.foo 08:55
camelia rakudo-moar 780f07: OUTPUT«foo␤bar␤»
psch it works, but i need do know if i already added the :D candidate, 'cause i'm doing that during BUILD and only want to add when it's not around yet
s/do/to/
m: role R { multi method foo(::?CLASS:U: ) { "foo" } }; class C does R { submethod BUILD { self.^add_multi_method('foo', my method (::?CLASS:D: ) { "bar" }) } }; say C.foo; say C.new.foo # ah, this is what i'm actually doing 08:57
camelia rakudo-moar 780f07: OUTPUT«foo␤Cannot call foo(C: ); none of these signatures match:␤ (C:U $: *%_)␤ in block <unit> at /tmp/OD8LWVDckN line 1␤␤»
psch ...well, minus "the role is also applied during BUILD" :)
09:01 Ven joined 09:05 pierre_ joined
jnthn psch: Wouldn't that add a multi method to the class (static) on every instantiation (dynmaic)? 09:06
psch jnthn: well, yes. that's why i want to know how many candidates i already have, so i don't do that if it's been done with this specific Class + Role combination already
jnthn Also add_multi_method doesn't actually add it right away, it stores it up until compose time, to see if we end up with a proto 09:07
And doing such things at runtime will make the code need "no precompilation" 09:08
To work out reliabily, at least.
psch hm, so i probably just stick with different names for the methods
jnthn If you want to find out about what candidates a given proto controls, though, there's .candidates 09:09
09:09 nwc10 left
psch ...i though i tried that and didn't get anything 09:09
maybe did it on the List from find_method i guess
jnthn find_method doesn't return a List...that's can
psch ...i'll finish my coffee :) 09:10
jnthn ;-)
Not through my first today yet :)
09:11 john51 left
psch well, the scenario is that i have a class C, that during BUILD mixes in one of multiple roles. some of these roles supply methods that return C with a different role mixed in 09:11
09:12 john51 joined
psch well, actually BUILD adds the method, the role just knows which different role it has to be 09:12
and i had though "hey, maybe you can just reuse the method name for :U and :D so :U returns the role that have to be mixed in and :D returns the list of methods that return a C+{$thatrole}" 09:13
'cause that'd be nice from a naming perspective :)
09:13 wamba left
psch but yeah, i guess i'll just pull that into {$methodname}-roles and {$methodname}-methods 09:14
09:18 sno left 09:21 wamba joined
RabidGravy I'm having unhealthy thoughts about implementing something with POSIX message queues 09:23
09:28 pierre_ left, pierre_ joined
RabidGravy actually I'm not sure it would work due to mq_notify requiring a struct one of who's members is a callback 09:30
09:31 CIAvash left 09:32 pierre_ left 09:35 cognominal joined 09:39 pierre_ joined
sortiz RabidGravy, Umm, I think that the SIGEV_SIGNAL case maybe work. The SIGEV_THREAD case requires MoarVM special support, thought. 09:47
09:53 domidumont left 09:54 wamba left 10:01 Actualeyes left 10:12 kaare_ joined 10:14 pierre_ left 10:19 Actualeyes joined 10:29 maybekoo2 left
psch m: my $async = -> @a { say @a; sleep 1; say @a }; my @b = ^5; start { $async(@b) }; sleep 0.5; @b = 5..10; sleep 2 # how do i get ^5 both times..? 10:29
camelia rakudo-moar 780f07: OUTPUT«[0 1 2 3 4]␤[5 6 7 8 9 10]␤» 10:30
psch pointies don't default to 'is copy'?
moritz no
psch m: my $async = sub (@a) { say @a; sleep 1; say @a }; my @b = ^5; start { $async(@b) }; sleep 0.5; @b = 5..10; sleep 2
camelia rakudo-moar 780f07: OUTPUT«[0 1 2 3 4]␤[5 6 7 8 9 10]␤» 10:31
psch m: my $async = sub (@a is copy) { say @a; sleep 1; say @a }; my @b = ^5; start { $async(@b) }; sleep 0.5; @b = 5..10; sleep 2
camelia rakudo-moar 780f07: OUTPUT«[0 1 2 3 4]␤[0 1 2 3 4]␤»
psch ...and @-s don't either, apparently
10:32 boegel left 10:34 boegel joined 10:35 colomon joined 10:36 TEttinger left 10:46 wamba joined 10:49 sno joined 11:01 espadrine_ joined, espadrine_ is now known as espadrine
masak psch: a thing *is* copied in the above code. it's just not the thing you expect. 11:08
m: sub foo(@a) { say(@a); bar(); say(@a) }; my @x = 1, 2, 3; sub bar { @x = 4, 5 }; foo() 11:09
camelia rakudo-moar 780f07: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Qh_U7BspgF␤Calling foo() will never work with declared signature (@a)␤at /tmp/Qh_U7BspgF:1␤------> 3my @x = 1, 2, 3; sub bar { @x = 4, 5 }; 7⏏5foo()␤»
masak m: sub foo(@a) { say(@a); bar(); say(@a) }; my @x = 1, 2, 3; sub bar { @x = 4, 5 }; foo(@x)
camelia rakudo-moar 780f07: OUTPUT«[1 2 3]␤[4 5]␤»
psch moritz: well, yes. the array is copied, which copies "which container does @a[0] point to"
masak ^ simpler example, without threads
psch: I'm over here
psch masak: whoops, sorry :/
masak no, the array *reference* is copied
which isn't intuitive, but it's consistent
psch purposefully avoided the word "reference"
instance probably?
masak you don't get two arrays, you get two references to the same array
psch anyway, yeah, i get it :) 11:10
masak I wouldn't avoid the word "reference" here, since it's key to understanding what's going on :P
11:10 wamba left
psch masak: well, i don't like using it 'cause it's pragmatically overloaded from Perl 5 11:11
masak fairy nuff
psch as in, "list of arrayrefs" for ([1,2], [3,4])
masak that's the same meaning in Perl 6 in my mind, fwiw
11:12 nchambers is now known as sillytime
psch then i probably have my own, weird, pragmatic for "arrayref" which means something more Perl 5-specific 11:12
masak you should look into that :P 11:13
psch m: sub f($x, @a is copy) { say $x; say @a }; my @c = ^5; my &g = &f.assuming(*, @c); g(1); @c = 5..^10; g(1)
camelia rakudo-moar e3ded7: OUTPUT«1␤[0 1 2 3 4]␤1␤[5 6 7 8 9]␤»
psch that's closer to my actual problem, fwiw. taking the thready bits with me was not necessary... 11:14
not sure i can close over @c there
m: sub f($x, @a is copy) { say $x; say @a }; my @c = ^5; my &g = &f.assuming(*, my @ = @c); g(1); @c = 5..^10; g(1)
camelia rakudo-moar e3ded7: OUTPUT«1␤[0 1 2 3 4]␤1␤[0 1 2 3 4]␤»
psch actually yeah nvm
11:14 wamba joined
psch hm, curiously that doesn't actually work in my actual code :/ 11:16
masak I've come to the reluctant conclusion that the only actual solution to reference/aliasing/ownership problems (besides going 100% immutable) is what Rust is doing
moritz how well does it work in Rust? 11:17
from what I've heard, it seems to place a very heavy burden on the type system
masak it does force you to think about borrowing/ownership in about the same way as any strict type system forces you to think about return types and parameter types 11:18
11:18 boegel left
masak beyond that, I'm not sure. ENEEDMORERUSTEXPERIENCE 11:18
awwaiid How about selective-immutable, OCaml style?
masak awwaiid: where can I read more about that? 11:19
11:19 boegel joined
masak awwaiid: all I know is that in Perl 6, we don't do a good enough job of tracking immutability. 11:19
moritz: probably github.com/servo/servo is a great showcase of a "real-world" project using borrowing in actual situations 11:20
awwaiid I might have made it sound more like a thing than it is -- basically in OCaml things are immutable by default and you go out of your way to use refs. So it is common to use Okaski's purely functional data structure (www.amazon.com/Purely-Functional-St...521663504)
masak: OCaml itself is written in OCaml, might be interesting to do a read through their code 11:21
masak aha.
well, I'd file that under "going 100% immutable", in a way
and refs are basically a mechanism for you to opt out of the immutability
I would say such a solution is not an option for Perl 6, because it'd change the core nature of the language too much 11:22
psch gist.github.com/peschwa/f60774251b...7ffd7265f7 # the first call to .riff inside the start in L19 has the original @chords, the second has the new one
awwaiid hm. yeah -- even their built-in hash is immutable
11:23 kid51 joined
awwaiid psch: what does "@inner-chords[$++ % *]" mean? 11:23
masak awwaiid: "modulo length of array" 11:24
awwaiid I mean ... my guess is it is modding an incrementing thing
masak yes, an incrementing anon state variable
moritz $ is an anonymous state var
so it's just giving an incrementing sequence
psch awwaiid: yeah, it's incrementing a state variable and modulos it the array length to not go out of bounds
awwaiid ah. so $++ is incrementing the anon state. gotchya
thanks
masak m: my @things = <a b c>; for ^10 { @things[$++ % *] = $++ }; say @things 11:25
camelia rakudo-moar e3ded7: OUTPUT«[9 7 8]␤»
11:25 grondilu joined
moritz : my @things = <a b c>; say @things[$++ % *] for ^10 11:26
m: my @things = <a b c>; say @things[$++ % *] for ^10
camelia rakudo-moar e3ded7: OUTPUT«a␤b␤c␤a␤b␤c␤a␤b␤c␤a␤»
masak m: my @things = <a b c>; say @things[$++ % *] and @things[$++ % *] = "bacon" for ^10 11:27
camelia rakudo-moar e3ded7: OUTPUT«a␤b␤c␤bacon␤bacon␤bacon␤bacon␤bacon␤bacon␤bacon␤»
awwaiid yeah. Makes sense. Kinda dense, but now that I'm looking at it I guess the line above it is for debugging/watching, so this would maybe normally be a one-liner 11:31
11:33 Ven left 11:34 Ven joined 11:38 finanalyst joined 11:49 vendethiel joined 11:50 kurahaupo joined 11:53 kid51 left
psch oh, i see my error in the gist, actually 11:54
the .assuming is ran repeatedly
jdv79 assumptions are dangerous 11:58
11:59 sillytime is now known as nchambers 12:00 [Sno] joined 12:03 sno left 12:05 finanalyst left 12:06 rindolf left 12:07 wamba left, [Sno] left 12:10 sno joined 12:15 abraxxa left, ZoffixWin joined, ZoffixWin left, ZoffixWin joined 12:16 rindolf joined
sortiz wondering why people use @foo almost only for Array and %foo only for Hash. Fear of immutability? Perl5 habit? 12:17
ZoffixWin Oh man. Waking up, reading logs and seeing this: irclog.perlgeek.de/perl6/2016-05-06#i_12441561
ZoffixWin feels sorry for AlexDaniel spending time teaching huggable all the tricks
sortiz, what else do you use them for? 12:18
moritz m: say Match ~~ Associative
camelia rakudo-moar d9803a: OUTPUT«False␤»
moritz I guess you could use them for Set as well 12:19
and @ for Blob/Buf
sortiz ZoffixWin, my @foo is Array by default, but any Positional can be used, and the same for my %foo, where the default is Hash, but any Associative is valid. 12:20
moritz doc.perl6.org/type/Positional the type graph shows some arrows twice
like NFKC => Uni
oh, probably because they appear twice in type-graph.txt 12:21
ZoffixWin sortiz, oh. I guess Perl 5 habit then :P I didn't know they could do that
dalek osystem: 5ed988f | titsuki++ | META.list:
Add App::AizuOnlineJudge to ecosystem

See github.com/titsuki/p6-App-AizuOnlineJudge
osystem: 68a991f | titsuki++ | META.list:
Merge pull request #203 from titsuki/add-aoj

Add App::AizuOnlineJudge to ecosystem
c: 413a693 | moritz++ | type-graph.txt:
Remove double defintion of Uni and related types
12:21 domidumont joined, cdg joined
grondilu I'm trying to use panda on windows and it fails installing a package. Error seems to point at a lack of "prove" on my system. I thought it was shipped in since it must have used it to install whatever aleaday is install, musn't it? 12:22
already is installed 12:23
moritz grondilu: how did you install panda?
grondilu I got it from rakudo*
(I suppose)
moritz grondilu: from the MSI?
grondilu yes
moritz grondilu: then it might not even have run prove on installation
grondilu though I did try to install rakudo (not star) previously 12:24
so maybe there are remnants or something
moritz do you have p5's prove installed?
stmuk grondilu: the windows panda doesn't work without --notest see www.perl6.org/downloads/ errata
grondilu ok
confirmed. Works with --notests 12:26
sortiz m: my @l is List.new: |<a b c d e>; my @b is Buf; @b.push(|(1..10)); dd @l,@b; # For example ZoffixWin 12:27
camelia rakudo-moar d9803a: OUTPUT«()␤Buf.new(1,2,3,4,5,6,7,8,9,10)␤»
stmuk panda doesn't call prove correctly on Windows even when prove is installed .. zef is recommended
grondilu never used zef 12:28
stmuk also see blogs.perl.org/users/itcharlie/2016...e-zef.html 12:29
ZoffixWin sortiz, but what's the point of using @ in there?
moritz, do you have easy access to this channel's IRC logs that you could zip up everything from November to today and give it to me? If I grep them for commands to huggable and then replay those to the new instances, I should be able to restore its database. 12:31
sortiz Mark them as Positional or Associative. Ie capables of subscripting.
ufobat ZoffixWin, i'd try to look into this this afternoon. 12:32
moritz ZoffixWin: I can try come up with something 12:33
ufobat ZoffixWin, have you tried something like "say $app.renderer.perl"?
ZoffixWin ufobat, no rush. I accidentally destroyed my server last night so I'll be a bit busy :P
ufobat oh sh... :(
sortiz ZoffixWin, Otherwise, why use different sigils in the first place?
ZoffixWin m: my $l = List.new: |<a b c d e>; say $l[2..3] 12:34
camelia rakudo-moar d9803a: OUTPUT«(c d)␤»
ZoffixWin sortiz, to get the default hash/array stuff. To me, sigils feel a bit misdesigned in Perl 6. Maybe I don't get what their purpose is
sortiz, the $l above has subscribting capabilities, so I'm unsure what benefit there is in using a @l 12:35
sortiz m: my @l := <a b c>; .say for @l; my $l = <f g h>; .say for $l; 12:36
camelia rakudo-moar d9803a: OUTPUT«a␤b␤c␤(f g h)␤»
psch m: sub f(+@a) { say @a.perl }; my @a = <a b c d e>; my $l = List.new: |<a b c d e>; f @a; say f $l 12:37
camelia rakudo-moar d9803a: OUTPUT«["a", "b", "c", "d", "e"]␤[("a", "b", "c", "d", "e"),]␤True␤»
psch -spurious say
that's what's underlying sortiz++ example
moritz ZoffixWin: select ilbot_lines.line FROM ilbot_lines JOIN ilbot_day ON ilbot_day.channel = 38 WHERE ilbot_day.day >= '2015-11-01' AND ilbot_lines.line LIKE '%huggable' ORDER BY ilbot_lines.timestamp;
ZoffixWin: does that look workable to you? or do you need more info?
like, nickname, timestamp, whatever?
ZoffixWin moritz, looks good. 12:38
sortiz, why did you use := instead of normal =?
12:38 MilkmanDan joined, andreoss joined
andreoss m: "".ord.chr.say 12:39
camelia rakudo-moar d9803a: OUTPUT«␤»
moritz ZoffixWin: moritz.faui2k3.org/tmp/huggable.txt.gz I hope that helps 12:40
sortiz m: my @a = <a b c>; my @l := <a b c>; say @a.WHAT,@l.WHAT;
camelia rakudo-moar d9803a: OUTPUT«(Array)(List)␤»
lizmat m: "".ord.say
camelia rakudo-moar d9803a: OUTPUT«Nil␤»
andreoss for me it gives warning `Use of nil in a numeric context`
should it?
sortiz ZoffixWin, ^^^
lizmat m: Nli.chr
camelia rakudo-moar d9803a: OUTPUT«5===SORRY!5=== Error while compiling /tmp/JXsA3HLtxr␤Undeclared name:␤ Nli used at line 1. Did you mean 'Nil'?␤␤»
moritz .uni Q 12:41
lizmat m: Nil.chr
camelia rakudo-moar d9803a: OUTPUT«Use of Nil in numeric context in block <unit> at /tmp/rs7mKswII4 line 1␤»
12:41 Skarsnik joined
ZoffixWin moritz, that looks weird :P 12:41
moritz ZoffixWin: agreed
lizmat andreoss: I think the warning is correct
andreoss m: "".ord.chr.say
camelia rakudo-moar d9803a: OUTPUT«␤»
ZoffixWin moritz, should that be LIKE '%huggable%' ?
lizmat "".ord returns Nil, Nil.chr attempts to use Nil as a number
moritz ZoffixWin: ah, yes
andreoss lizmat: what happens with ^Q inside the quotes? 12:42
ZoffixWin sortiz, thanks
lizmat andreoss: ^Q ?
moritz ZoffixWin: will re-hug :-)
my mysql instance is just super slow :/ 12:43
andreoss m: "".chars.say
camelia rakudo-moar d9803a: OUTPUT«5␤»
ZoffixWin m: "".uninames.say 12:44
camelia rakudo-moar d9803a: OUTPUT«(DEVICE CONTROL ONE DEVICE CONTROL ONE DEVICE CONTROL ONE DEVICE CONTROL ONE DEVICE CONTROL ONE)␤»
moritz oh, forgot a JOIN condition
ZoffixWin: please re-download from the same URL 12:45
andreoss m: say elems qw[ ]
camelia rakudo-moar d9803a: OUTPUT«3␤»
ZoffixWin moritz++ got it. Thanks a lot! 12:46
andreoss m: say (qw[ ])».ord.chr.say
camelia rakudo-moar d9803a: OUTPUT«␤True␤»
andreoss m: say (qw[ ])».ord.chr.join.say
camelia rakudo-moar d9803a: OUTPUT«␤True␤»
andreoss m: say (qw[ ])».ord 12:47
camelia rakudo-moar d9803a: OUTPUT«(17 17 17)␤»
andreoss m: say (qw[ ])».ord».chr.join
camelia rakudo-moar d9803a: OUTPUT«␤»
andreoss some warnings are suppressed for camelia? 12:50
jnthn m: say (qw[Q Q Q]).ords».chr.join
camelia rakudo-moar d9803a: OUTPUT«Q Q Q␤»
jnthn Ah, qw 12:51
But on a string I suggest .ords rather than .comb>>.ord
moritz andreoss: no, camelia doesn't suppress anything 12:52
12:54 [Coke] left, rindolf left 12:55 [Coke] joined, wamba joined 12:58 Sgeo left 13:02 itaipu joined
andreoss seems that Linenoise ignores literal symbols 13:03
i.e and so on 13:04
sortiz Yep, Linenoise have serious problems with Unicode
13:05 rindolf joined
andreoss is control symbol from ASCII 13:05
moritz ASCII is a subset of Unicode :-) 13:06
andreoss of UTF-8 rather
moritz ASCII is both an encoding and a character repertoire 13:07
13:07 itaipu left
moritz in the latter role, it's a subset of Unicode 13:07
in the former role, of UTF-8
13:08 itaipu joined 13:16 cdg left 13:17 cdg joined 13:18 hotel joined
masak moritz: wow, I had never thought of that. 13:18
so ASCII numbers are *both* characters and codepoints, sort of?
moritz masak: aye 13:19
13:20 sufrostico joined
masak life was simpler back then 13:21
perlpilot but then we had to grow up
masak speak for yerself
:P
hotel Hey ZoffixWin, they're tearing apart the bug i reported :(
tadzik I'm still mildly amused when people treat me as a responsible grownup 13:22
perlpilot masak: that's almost the defining characteristic of growing up: things aren't as simple as they once were.
13:24 itaipu left
masak perlpilot: I'm inclined to agree, provided I'm also allowed to point out that I really enjoy staying in touch with my inner child 13:24
ZoffixWin hotel, what's the URL again? 13:25
hotel rt.perl.org/Ticket/Display.html?id=128077 13:26
I'm not smart enough to argue, lol
ZoffixWin hotel, I had tests and fixes for it last night, but I got too wasted before I committed anything
hotel Haha
Well, time for school. Bbl
13:27 hotel left, grondilu left 13:28 itaipu joined 13:31 kaare__ joined, kaare_ left 13:33 itaipu left, wamba left
perlpilot masak: sure! Inner children are awesome! Let them out as often as you can. :) 13:33
13:33 kaare_ joined
timotimo but if you let them out, are they still "inner"? 13:33
13:33 domidumont left 13:34 perlawhirl joined
masak Children/InsideOut.pm 13:36
13:37 kaare__ left 13:39 itaipu joined, g4 left 13:40 AlexDaniel joined 13:41 wamba joined 13:43 kaare_ left, kaare__ joined 13:47 pierre_ joined 13:48 zakharyas joined 13:51 prammer joined
dalek href="https://modules.perl6.org:">modules.perl6.org: 861d999 | (Richard Hainsworth)++ | templates/root/index.html.ep:
Update index.html.ep

Add href to Citations Index page.		 13:51
href="https://modules.perl6.org:">modules.perl6.org: 6714309 | (Zoffix Znet)++ | templates/root/index.html.ep:
Merge pull request #59 from finanalyst/master

Add link in Modules page template to Citations Index page.
ZoffixWin Any word on the merging of github.com/rakudo/rakudo/pull/743 ? 13:53
I stupidly made it on `nom` branch and now it's kinda annoying to when I try to clone the repo again. 13:54
I guess it's learning time to figure how to work around that \o/ 13:55
It depends on this merge to roast: github.com/perl6/roast/pull/113 13:56
13:58 perlpilot left, perlpilot_ left 14:00 perlpilot joined, smls joined
Ulti lizmat I notice you are replacing a lot of if/else with conditionals under the commit "streamline" is this actually a performance improvement? 14:06
masak conditionals -- as in ?? !!, yes?
Ulti yup
ternary operators 14:07
:P
timotimo yes, it is. it's a guaranteed "doesn't have curlies"
Ulti is that not something spesh can capture that there is a conditional with only assignment and optimize away the difference? 14:08
14:08 kaare__ is now known as kaare_
timotimo it's something the static optimizer should do, but it doesn't always do it yet 14:08
Ulti by curlies is that producing a closure every time even for control flow? 14:09
timotimo not necessarily
14:09 finanalyst joined
timotimo it can definitely cause lexical variables to stay lexical rather than be lowered to locals 14:10
masak I definitely like that we're reaping these short-term performance wins in the setting
timotimo if you're interested, you can look at the exact code we generate from those things with the spesh log
masak but sometimes I feel like we should keep a list of "code-degrading optimizations" we're making, so that in the future we can roll them back as the optimizer gets better at handling them 14:11
at least if we care about long-term setting niceness
timotimo anyway, ifs and curlies give us bigger code, which can push our simpler methods above the static inlining threshold
that won't trigger too often, but when it does, it's quite good 14:12
ugexe MadcapJake: data from a handle in one thread cant be used in another thread 14:14
it might appear to work sometimes, but thats because its really using the same thread (even if you put start { } around it) 14:15
Ulti thanks for the explanation
I'll remember that if I have some tight loops with that form of assignment 14:16
ugexe MadcapJake: github.com/MoarVM/MoarVM/issues/165
14:17 sjoshi left
timotimo alternatively, feel free to hack on the static optimizer :) 14:17
perlawhirl Ulti: you did the perl6 commit for GlotIO, yeah?
Ulti yup 14:18
perlawhirl can the container export RAKUDO_ERROR_COLOR=0, so that compiler errors disaply nicely
right now they look like this: m===mSORRY!m===m 14:19
Ulti not sure but I can take a look
a better plan is not default terminal errors :P
all the graphical editors with erorr highlighting get that issue too
perlawhirl Well, for most terminal users it's a bonus, and doesn't cause issues if they don't use colours
timotimo why does the thing provide a tty if it can't do ansi colors? :P
Ulti could check for a tty and not use colours also
perlawhirl bet yeah, on by default seems an odd choice
Ulti but guess thats not portable 14:20
timotimo it already does, doesn't it?
Ulti dont think so
14:20 mr-foobar joined
ZoffixWin +1 on fixing GlotIO output to not include trash characters (whatever method used :P) 14:21
perlawhirl i think default-off colours is the most sane way
14:21 astj left
literal how do I succintly flatten the return values of zip()? basically I wanted to do something like for @foo Z @bar -> $k, $v { ... } 14:21
timotimo huh, you are right
teatime lots of things check for tty to decide whether to do colors
timotimo but we do have "isatty" now, i think
teatime perlawhirl: so I'd have to alias perl6="perl6 --color=auto" ? 14:22
ah well, I already have like a dozen of those, so another won't hurt :)
geekosaur m: *IN.t.so.say
camelia rakudo-moar 7db7b4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/rAd6s9RaLA␤Two terms in a row␤at /tmp/rAd6s9RaLA:1␤------> 3*7⏏5IN.t.so.say␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ statement modifier␤ …»
ZoffixWin literal, isn't -> ($k, $v) { do the trick?
geekosaur whoops
m: $*IN.t.so.say
camelia rakudo-moar 7db7b4: OUTPUT«False␤»
literal ZoffixWin: ooh, indeed, that works. Thanks
perlawhirl Ulti: not it's an ENV var... so i think somehow the contanier literally needs to 'export RAKUDO_ERROR_COLOR=0' when it's created 14:24
my words aren't working, but i think that made sense
ahh, and i was responding to teatime, not Ulti 14:26
14:26 pierre_ left
teatime oh wut 14:27
perlawhirl not sure how you'd fix it in something like Atom... maybe if you export the env var it just werks? 14:28
ugexe m: say Rakudo::Internals.NUMERIC-ENV-KEY("RAKUDO_ERROR_COLOR")
camelia rakudo-moar 7db7b4: OUTPUT«Nil␤»
14:29 khw joined
perlawhirl if I may seque to another question, it seems NCurses does not like ANSI colors. is it known, or am i doing something wrong 14:30
teatime umm, I don't understand the question
perlawhirl i was just trying to use Term::ANSIColor'd strings with NCurses, and it spat out junk
teatime oh.
geekosaur yes 14:31
perlawhirl yes...? it has problems?
geekosaur ncurses has its own color management. it does not parse ANSI sequences, it *creates& them
perlawhirl oh... ok
geekosaur this is expected, it is not a bug, and trying to modify it so that it can un-ANSI just so it can re-ANSI would be frustrating and a source of bugs and unexpected behavior
if you want ANSI color with ncurses, use the ncurses color mechanisms 14:32
teatime what is the default flavor of cream soda? the red one.
perlawhirl alright.. i'll check the source / ncurses docs
geekosaur teatime, primary flavor is vanilla, red *may* have cherry added but not as primary flavor 14:33
teatime no, it's not cherry... I was thinking vanilla was probably the answer... vanilla isn't usually red, though, then? 14:34
geekosaur naturally? it's dark brown
but that's in fairly concentrated form; depending on the source, the more diluted form could well be reddish 14:35
teatime not actual vanilla. vanilla cream soda :)
geekosaur heh
cream soda may be clear, brownish, or red/reddish
Ulti given things /win 2 14:37
wah 14:38
geekosaur with "real" cream sodas this reflects amount of vanilla and possibly other flavorings, and the source of the vanilla. these days its all artificial x.x
*it's
Ulti perlawhirl: yeah I get what it needs, Docker most config is done through the environment its the preferred way to do things
perlawhirl cool beans 14:39
geekosaur oh, I should note that the curses color stuff looks overengineered. sadly, at the time when it was added, that was not actually overengineered; it reflected the actual limitations of terminals 14:41
14:41 grondilu joined
geekosaur things got a LOT simpler when terminals standardized on ANSI X3.64 14:41
teatime are you referring to 256-color mode and pallette setting 14:42
geekosaur but that only happened when real physical terminals mostly went away
teatime rather than the 24-bit color escape codes
geekosaur COLOR_PAIR and such
teatime 'cause the 24-bit escapes aren't widely supproted
geekosaur you are not talking about the (n)curses level 14:43
which is what perlawhirl was asking about
teatime no, I am
I am asking if you are referring to ncurses not using the 24-bit escapes
geekosaur I am referring to the curses API 14:44
teatime nevermind.
geekosaur whether it knows about and can use the escapes depends on whether the terminal description (terminfo) describes them in a way that (n)curses can understand
perlawhirl hrm, i might just use Term::Screen via Inline
geekosaur ...and since that too dates from the days of physical terminals, its ability to express such things is limited (it's possible but a lot of work) 14:45
teatime terminfo evolves, and is actually part of ncurses. 14:46
geekosaur yes, I am aware of that
Ulti perlawhirl: github.com/prasmussen/glot-containers/pull/7 14:48
that should do it I think
unfortunately I can't test it since the deployment of glot.io is fairly epic 14:49
14:49 itaipu left
Ulti since the Docker image needs to get rebuilt we might force a 2016.04 update too :) 14:49
perlawhirl 5 repo's combine to make it... kinda like Ultron
Ulti yup 14:50
perlawhirl 2016.04 bump is a good idea
Ulti had to go through three of them just to find the Dockerfile :'(
14:50 Relsak left
Ulti another plan is to actually implement the colour codes being converted to coloured <span> tags :3 14:51
14:51 molaf joined
perlawhirl pfft, they should made html understand ansi escape sequences in the first place :P 14:54
timotimo yes 14:55
smls Hm, the REPL fails to start up for me
It prints "To exit type 'exit' or '^D'" and then hangs
timotimo are you inside a "panda look"? 14:56
smls what's that?
timotimo it grabs a module for you and drops you in the module folder in a shell 14:57
smls No, I'm in my home folder
timotimo it sadly doesn't unlock the repo, so if you start a perl repl it'll usually hang waiting for the repo to become unlocked
awwaiid you should be able to see that if you do `strace perl6` -- you'll see it blocking on a .lock file 15:00
if that is the case, that is
So I see that "Bug #127190 for perl6: Perl 6 has 0 elements ( .hyper.grep )" is already filed -- is there any sort of work-around? 15:02
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=127190
awwaiid m: say ^10 .race.grep: *.is-prime 15:03
camelia rakudo-moar 7db7b4: OUTPUT«()␤»
smls m: .say for ^10 .race.map({ next if .is-prime; $_ }) 15:05
camelia rakudo-moar 7db7b4: OUTPUT«0␤1␤4␤6␤8␤9␤»
psch m: ^10 .race.grep( *.say );
camelia ( no output )
awwaiid I'm working on a talk submission and was going to use that as an example :) 15:06
psch whatever .grep does with whatever .race produces seems somewhat wrong... :)
hoelzro wishes ack supported P6 regexes
timotimo awwaiid: no way at the moment, sorry :(
awwaiid ok cool
timotimo except if somebody goes in and fixes it, of course
geekosaur psch, I think that;s a known bug in race?
psch geekosaur: yeah, the tickets been linked a bit ago 15:07
awwaiid there are a ridiculous number of other features that I can pick to highlight
psch geekosaur: i was just curious, which is why i put the *.say into the grep, to see what arrives there anyway
timotimo yeah, though of course parallelism features are extra juicy
geekosaur ob "you have a problem and you decide to use threads. nonw owy ou yoauv e hatwvo e ptrowbo plerombsle imns teinrtleearlveeda.ed." 15:08
ugexe heh 15:09
timotimo :)
awwaiid oh wait, if I put in a grpe it works 15:10
m: (^10).race.map(* + 0).grep(*.is-prime)
camelia ( no output )
awwaiid m: (^10).race.map(* + 0).grep(*.is-prime).list.say
camelia rakudo-moar 7db7b4: OUTPUT«(2 3 5 7)␤»
teatime heh, how much code do you think will break when, at some point in the future, » etc. begin actually multi-threading :) 15:11
ugexe something about map itemizing differently than grep such that hypers worked slightly different on each
timotimo teatime: for the longest time we've made the execution order of >> things randomized
it no longer is randomized, if i recall correctly
teatime I was gonna say, ».say works for me when I do it
timotimo but since hyper still gives you the result order you wanted, it won't be that bad 15:12
only side-effects will be scrambled
hoelzro m: 'f3cd5fa' ~~ /<-:Letter - digit>/ 15:15
camelia ( no output )
hoelzro m: 'f3cd5fa' ~~ /<-:Letter - Digit>/
camelia rakudo-moar 7db7b4: OUTPUT«Method 'Digit' not found for invocant of class 'Cursor'␤ in block <unit> at /tmp/RW10TqzI9S line 1␤␤»
hoelzro why is titlecase ok for Letter, but not for Digit?
timotimo wouldn't you need to :Digit?
because digit is also a method on Cursor?
hoelzro ahhh
timotimo whereas :Letter is a unicode property thing
hoelzro m: say '.' ~~ /<-:Letter - :Digit>/ 15:16
camelia rakudo-moar 7db7b4: OUTPUT«Nil␤»
teatime dunno if relevant but <digit> also already pre-exists
hoelzro that *should* match, right?
I'm trying to match any character that's neither a letter nor a digit
teatime no, . isn't in :Letter
oh nm
hoelzro unless the leading - is complementing (:Letter - :Digit)
...in which case it would still match, right? 15:17
smls timotimo: After nuking rakudobrew and rebuilding, my REPL works again. 15:20
timotimo interesting. good to know, i guess
awwaiid hm 15:22
ZoffixWin m: say '.' ~~ /<-[\w]+[_]>/ 15:23
camelia rakudo-moar 7db7b4: OUTPUT«｢.｣␤»
hoelzro awwaiid: did you see my comment on your PR? 15:24
awwaiid hoelzro: yes! But I haven't gotten a chance to fix it up yet. I missed .05 yet?
hoelzro: think it would be ok to print out "Entering fallback REPL" if it does the fallback? I'm want us to know/debug any times it falls back 15:25
15:25 tomboy65 joined 15:26 dolmen joined, yqt joined
hoelzro awwaiid: .05 is a way's away =) 15:27
I think that's a good idea
15:27 ufobat left
hoelzro something like "failed to load REPL.pm: $ERROR\nEntering fallback REPL" 15:27
15:27 tomboy64 left
timotimo hoelzro: well, very big changes should go in loooong before the release, to be fair 15:28
hoelzro timotimo: agreed
I'm thinking of even waiting until .05 is out to merge, to maximize our window 15:29
awwaiid ok cool. I'll get the fallback back in there over the weekend. 15:33
15:33 skids joined 15:36 Ven left 15:41 finanalyst left 15:46 Relsak joined
awwaiid Request for review/edits/suggestions on my Strange Loop talk submission -- gist.github.com/awwaiid/6d9c996557...943fde504c 15:49
15:50 domidumont joined
perlpilot awwaiid++ 15:51
awwaiid: experimention -> experimentation 15:52
15:54 ZoffixWin left
awwaiid oh right, spelling 15:54
perlpilot: updated 15:55
perlpilot awwaiid: How long is the talk? This is a session or a workshop? 15:56
hoelzro awwaiid: sounds good (both the weekend work and SL talk idea)
smls awwaiid: Aren't the concurrency primitives taken from (or heavily inspired by) C#/.NET, as opposed to complete innovations? 15:57
awwaiid perlpilot: I'm putting in for a 40 minute talk. I could do a workshop, but felt like this is more likely to get accepted.
perlpilot awwaiid: So, is it going to be more of a survey of neat Perl 6 things or are you just going to focus on a few and mention that there are other neat things available? 15:58
awwaiid smls: don't know! I think the innovation is more about how they are integrated in. start {} I think is c#, but .race.map looks more like clojure core.async (standard libs that have the same rough api but work on an async/concurrent datatype)
perlpilot: I'm leaving it open for myself, but I think the best use of time would be to do a deeper-dive into like 2 nonstandard features and a survey/bief-mention of some others. 15:59
15:59 dolmen left
perlpilot awwaiid: Sounds excellent! 15:59
awwaiid I think I'll make that clear on the thing 16:00
perlpilot awwaiid++
I also like how you say "the production-ready Rakudo implementation of Perl 6". The more we talk about the production-readiness, the better 16:02
awwaiid Yeah. That also let me talk about the back-ends 16:03
gist updated with a note about deep-dive into 2 advanced/nonstandard topics
smls [Coke], masak, other RT veterans: It looks like we have a large number of open tickets that were automatically submitted for roast test failures, about a year ago. Are they still needed? 16:06
(e.g. #124685)
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=124685
RabidGravy iirc it wasn't just failures but all the fudges 16:16
smls What was the idea behint RT'ing them all? 16:21
Isn't it more useful for implementors to run roast themselves and check the *current* failures/fudges? 16:22
RabidGravy yeah, you'd have thought so
psch smls: that's not always really feasible. r-j spectest can take well over two hours, depending on machine and load 16:23
RabidGravy if there is a big list of those tickets it might be worth grepping the roast for the numbers and just closing the ones that ain't there any more
because I think the plan was that the tickets should be in the source 16:24
16:24 samb1 joined
smls psch: But is anyone keeping the list up-to-date (i.e. closing no longer applicable ones)? 16:24
16:24 pnu_ left 16:26 pnu_ joined, milwauke` joined, prammer left
awwaiid "Your talk proposal was successfully submitted" . I'll let y'all know if it gets accepted and then everyone can come over :) 16:26
16:26 prammer joined
psch smls: well, i do usually comment on a ticket when i it, and usually our bug admins are attentive enough to react to that (when i don't forget to cc the mailing list...) 16:27
16:27 avalenn left
psch +fix 16:27
jnthn smls: Certainly I closed a bunch of them when I dealt with the tests they were about in some way
smls ok
16:28 avalenn joined, apejens joined, tharkun left 16:29 llfourn left 16:30 zakharyas left, llfourn joined, milwaukee left 16:31 Amnez777 left 16:32 kmwallio left
smls RabidGravy: It looks like it's 244 in total, that are still open 16:32
(search link here: gist.github.com/smls/eed26639e60f1...14856f41c) 16:33
16:33 kmwallio joined 16:35 Amnez777 joined, Emeric left, tharkun joined 16:39 sufrostico left 16:43 gtodd1 left 16:44 cpage_ left 16:45 sufrostico joined 16:46 newbie1 joined 16:47 ZoffixWin joined
ZoffixWin How can I just pop up a bunch of supplies and tell my program to sit and chill (and just repond to those supplies whenever they happen)? I tried by just sticking this loop {} but I now see my app is using 100% of CPU: github.com/zoffixznet/perl6-Bailad...ilador#L16 16:47
16:48 takadonet joined
timotimo you need "react" 16:49
16:50 yqt left
ZoffixWin I don't know how. The program exits right away 16:50
mst isn't this what forever/whenever are for?
timotimo mst: whenever goes into a "supply" or "react" block to listen to other supplies doing stuff 16:51
ZoffixWin hm. lemme try
timotimo yeah, you'd want to whenever your supplies inside the react block 16:52
jnthn ZoffixWin: Something like this: gist.github.com/jnthn/b29e1628f21e...58c79a62af
perlpilot isn't aware of a "forever" block, but that would be a nice self-documenting alternative to loop {} 16:53
ZoffixWin Thanks. jnthn++ timotimo++ mst++
16:56 itaipu joined
mst ah, right, so you use 'react' to say 'within this, we're setting up reactive stuff' then 'whenever' inside that uses the taps to do things, and the reason zoffix' stuff was exiting was he hadn't added any listeners? 16:59
timotimo that sounds like the thing, yeah
an empty react block would immediately shut down, i expect
m: react { }; say "done"
camelia rakudo-moar ac36d2: OUTPUT«done␤»
mst oh damnation, I just helped somebody implement one of those horrible auto restarter things 17:00
mst feels dirty now
timotimo hah
ZoffixWin hahahahaaa
mst, why is it horrible?
Manually restarting is horrible
mst Ctrl-C, up, enter doesn't bother me 17:01
ZoffixWin On EVERY change?
mst having my console log full of errors that are all complete lies so it obscures the real ones does bother me
tadzik haha, oh yes
"why doesn't it work? Oh,because this time it failed to recompile so nothing changed" 17:02
mst more "what's wrong now? oh, my app server screen is fully of errors, is that a real problem or just a failed recompile from three minutes ago when I saved the first file out of three I needed to change for this to compile again?"
I usually end up turning it off before I get as far as hitting tadzik's failure mode, but 'cold start the app please' is one of the standard ways I've found for letting junior devs discover they just wasted an hour 17:03
17:03 perlawhirl left 17:04 firstdayonthejob joined
mst of course, newbies *still* think it's a good idea and it makes them happy so I'd much prefer the feature existed so long as nobody mind me turning it off as hard as possible the first time I run into it :D 17:04
ZoffixWin No, I always know if it's a real problem because on each restart my restarter issues a message so all I need to follow is from end to the first restart message. 17:05
17:05 itaipu left
ZoffixWin And the error really should be chucked into the browser... 17:05
17:06 espadrine left, tharkun left, tharkun joined, itaipu joined
tadzik ahaha 17:07
smls m: say "abbbbb" ~~ /a b+/
camelia rakudo-moar ac36d2: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at /tmp/pdoXjL8i5l:1␤ ------> 3say "abbbbb" ~~ /a7⏏5 b…»
smls ^^ Isn't that warning a little overzealous? 17:08
ZoffixWin Probably here: github.com/ufobat/Bailador/blob/ma...pp.pm#L115 Just pop the exception into the the message
timotimo smls: it can really help in real life circumstances
i've been saved by it at least twice so far
ZoffixWin smls, IMO yes, but I'm of the minority opinion :) It doesn't *always* happen. There's only some sequence of tokens that trigger it. 17:09
smls I'd consider putting a space there, good practive/style
a little weird to have the compiler advise me to do the opposite... :P
ZoffixWin smls, right, but the way to write that is / 'a' 'b'+ /
smls That's *too* verbose for my taste... 17:10
ZoffixWin m: say "abbbbb" ~~ m:Perl5:x/a b+/ 17:11
camelia rakudo-moar ac36d2: OUTPUT«(｢abbbbb｣)␤»
timotimo m: say "abbbbb" ~~ m:/a 'b'+/
camelia rakudo-moar ac36d2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YhoQq6rq8d␤Colons may not be used to delimit quoting constructs␤at /tmp/YhoQq6rq8d:1␤------> 3say "abbbbb" ~~ m:7⏏5/a 'b'+/␤ expecting any of:␤ colon pair (restricted)␤»
timotimo m: say "abbbbb" ~~ m/a 'b'+/
camelia rakudo-moar ac36d2: OUTPUT«｢abbbbb｣␤»
ugexe use all the modifiers 17:12
timotimo m: say "abbbbb" ~~ m/'a' b+/
camelia rakudo-moar ac36d2: OUTPUT«｢abbbbb｣␤»
perlpilot wonders when that :Perl5 regex modifier will become problematic (similar to how PCRE on #perl is problematic)
ZoffixWin it doesn't support character classes. That's one problem I encountered 17:13
*named character classes [[:punct:]] and stuff like that
timotimo yeah, we didn't build it too far yet
perlpilot "best" might be to just use Inline::Perl5 for it now 17:14
timotimo yah
smls timotimo: Isn't the version without space *more* likely to trip up people? i.e. a newbie looking at /foo+/ might think it repeats the whole word. 17:16
17:16 itaipu left
timotimo huh. not sure 17:16
mst perlpilot: eh, you can always bounce people over to us :)
17:17 itaipu joined, cdg left
AlexDaniel m: say ‘abbbbb’ ~~ /‘a’ b+/ 17:17
camelia rakudo-moar ac36d2: OUTPUT«｢abbbbb｣␤»
geekosaur the real fun will be when people confuse m:p5 with pcre >.> 17:18
AlexDaniel say ‘abbbbb’ ~~ m+‘a’ b\++
m: say ‘abbbbb’ ~~ m+‘a’ b\++
camelia rakudo-moar ac36d2: OUTPUT«False␤»
ZoffixWin m: say ‘abbbbb’ ~~ m+‘a’ b++ 17:19
camelia rakudo-moar ac36d2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/VpQJqR_SQF␤Missing required term after infix␤at /tmp/VpQJqR_SQF:1␤------> 3say ‘abbbbb’ ~~ m+‘a’ b++7⏏5<EOL>␤ expecting any of:␤ prefix␤ term␤»
timotimo oh btw, it wouldn't be terribly hard to implement a m:PCRE :) 17:20
AlexDaniel m: say ‘abbbbb’ ~~ m＋‘a’ b+＋
camelia rakudo-moar ac36d2: OUTPUT«｢abbbbb｣␤»
mst timotimo: SET THE CONFUSATRON TO ELEVEN
ZoffixWin m: say ‘abbbbb’ ~~ /a ​b+/;
camelia rakudo-moar ac36d2: OUTPUT«5===SORRY!5===␤Unrecognized regex metacharacter ​ (must be quoted to match literally)␤at /tmp/7R3zM6shSZ:1␤------> 3say ‘abbbbb’ ~~ /a7⏏5 ​b+/;␤Unable to parse regex; couldn't find final '/'␤at /tmp/7R3zM6shSZ:1␤------> 3say …»
ZoffixWin awwww
timotimo :) :)
perlpilot mst: Except currently you'll just be educating them on the fact that m:Perl5// is a bit of a lie. 17:21
(Or ... I'll be doing that :)
17:21 ZoffixWin left 17:22 itaipu left
MadcapJake ugexe: thanks for the explanation and link! But I'm not really sure where I am accessing something outside of the thread. The only thing I could think of would be @collected (which I could actually put inside the last start block) 17:24
17:24 lichtkind_ left, andreoss left, cpage_ joined
ugexe MadcapJake: link? i didnt look at any code 17:24
17:24 newbie1 left
MadcapJake Otherwise maybe it's the channels I am passing as references to the workers, but without being able to do that, I am struggling to see the usefulness of channels 17:24
gist.github.com/MadcapJake/c29b76d...orkers-pl6
in debug, L16 (the $!chunks.poll call) is what sends the first "unable to read" error) 17:26
Though it doesn't always fail there, for some odd reason. 17:27
ugexe if $!chunks contains anything from another thread, then thats where its finally read
its not always failing there because sometimes it will happen in the same thread it originated in
AlexDaniel m: my $x = ‘’; for ^100000 { $x .= chop }; say ‘hello’ ~ $x ~ ‘world’ # awww, it works 17:28
camelia rakudo-moar ac36d2: OUTPUT«helloworld␤»
MadcapJake Well $!chunks is created inside the class but not inside the start block (inside of submethod BUILD)
17:29 itaipu joined, sufrostico left
ugexe right, but what about anything $!chunks contains? if any of its elements were created in a different $*THREAD.id then that error happens 17:29
MadcapJake yeah that's gotta be it then. so you can't communicate across channels? that seems really limiting
what's send/receive/poll for if you can't do that? 17:30
ugexe yea its a pita, and a bug
well you can communicate across channels, just not with data that originated from handles i believe
17:30 wamba left
MadcapJake yeah it's pulled from a text file 17:31
AlexDaniel m: ‘hello’.starts-with(‘’).say
camelia rakudo-moar ac36d2: OUTPUT«True␤»
AlexDaniel m: ‘hello’.ends-with(‘’).say 17:32
camelia rakudo-moar ac36d2: OUTPUT«True␤»
17:32 squirrel_ joined 17:33 squirrel_ left, sufrostico joined
ugexe you pretty much have to roll your own ipc... socket client/server is one way 17:34
jnthn It's the *handles* that can't be read on another thread 17:35
mst moarvm needs to expose something approximately like p3rl.org/IO::Pipely I think
or whatever the underlying libuv abstraction is
jnthn A Str or Buf can happily travel between threads.
AlexDaniel wonders why Str.indent exists
MadcapJake jnthn: then what's going on because that's all I'm passing is strings
mst AlexDaniel: anti-leftpad insurance
mst really wants access to libuv's unix domain socket / win32 named pipe abstraction 17:36
ugexe i dunno, i've never been able to get a buf from a socket on thread A to read on thread B
AlexDaniel mst: oh, well… OK, it makes sense. I guess
MadcapJake L38 calls lines on the handle and sends each line to a provider who rotates it between a series of workers 17:37
AlexDaniel that's actually pretty smart
mst xkcd.com/1676/ # ARBITRARY LENGTH NON BREAKING SNAKE 17:38
timotimo yes 17:39
mst ISAGN
17:39 hjst joined
MadcapJake timotimo: did you see the latest version of my shakespeare character counter? 17:40
AlexDaniel m: ‘🐍🐍🐍🐍hello’.indent(-4) # snakes not supported 17:41
camelia rakudo-moar ac36d2: OUTPUT«Asked to remove 4 spaces, but the shortest indent is 0 spaces in block <unit> at /tmp/IanqvZxiAO line 1␤»
timotimo MadcapJake: i have not. though i might write my own version later today, if you give me your source data :)
MadcapJake ocw.mit.edu/ans7870/6/6.006/s08/lec...speare.txt (5.3MB file) 17:42
timotimo thanks 17:43
MadcapJake I deleted the header and footer text too
timotimo: my latest version was an attempt to follow your advice of rotating between workers, so I'd be interested to know if I implemented it how you envisioned (though it doesn't work due to some weird thread problem that I don't understand) 17:44
[Coke] smls: yes, because any todo or fudged test in roast needs a ticket.
timotimo oh
[Coke] (or to be removed)
timotimo oh, is that workers.pl? 17:45
MadcapJake yeah
[Coke] (that said, the person who opened all those tickets did so on their own with only a little guidance. they've been greatly cleaned up since they were opened.) 17:46
timotimo i don't know why you have a central start block that divides data from $provider into all the workers' chunks
17:46 cpage joined
timotimo rather than having all workers .receive on the $provider 17:47
MadcapJake timotimo: I know why I did that: cus I don't know how to do this :P
timotimo :D
MadcapJake I thought that I needed to send to each worker separately, I didn't know I could have them all listen to one channel and it would divy up the work automatically 17:50
[Coke] (RT) if folks are looking to help work on the queue, I have a gist somewhere that would help. 17:51
17:51 zostay left, cpage_ left
[Coke] gist.github.com/coke/ac078396e8f216b83e9a ... needs some work, I'll polish that slightly. 17:52
smls [Coke]: I was just casually looking through old tickets... :P 17:55
17:55 zakharyas joined
smls For those that have been fixed, is leaving a comment and marking them "testneeded" still the best course of action? 17:55
[Coke] yes. 17:56
smls If it's just an LTA error, is it ok to mark it 'resolved' directly? (I did that today.)
[Coke] I mean, writing the test and closing out the ticket is best, but at least that, yes.
I think that still needs a test, no. 17:57
smls ok
[Coke] we have to verify that the warning is, in fact, NLTA.
thanks for going through tickets.
Here's another thing: if there's no [TAG] in the subject, update the subject with one if there's an obvious existing one it should be.
17:58 zostay joined
[Coke] gist.github.com/coke/9acb95ea87fab...3a2a6953ac 17:59
(there's all the current ones)
18:01 dj_goku left, dj_goku joined
AlexDaniel m: say ‘⋛’.uniname 18:11
camelia rakudo-moar ac36d2: OUTPUT«GREATER-THAN EQUAL TO OR LESS-THAN␤»
MadcapJake lol why is there a symbol for that 18:12
AlexDaniel MadcapJake: there's another one
m: say ‘⋚’.uniname
camelia rakudo-moar ac36d2: OUTPUT«LESS-THAN EQUAL TO OR GREATER-THAN␤»
timotimo that just means "those two things are orderable somehow"
perlpilot and depending on which one you use ... shows your bias 18:14
18:18 takadonet left, Joe_Perl joined
Joe_Perl Hello 18:18
timotimo hey there 18:19
Joe_Perl So i pretty much love perl6 but...
AlexDaniel m: .say for (0..0x1FFFF ==> grep { .uniname ~~ /MONKEY/ } ==> map {.chr.uniname}) 18:20
camelia rakudo-moar ac36d2: OUTPUT«MONKEY␤MONKEY FACE␤SEE-NO-EVIL MONKEY␤HEAR-NO-EVIL MONKEY␤SPEAK-NO-EVIL MONKEY␤»
AlexDaniel why so many monkeys?
timotimo Joe_Perl: uh-oh, here it comes :)
Joe_Perl Perl6 would be my first programming language i seriosly learn and i want to make sure that if i put my time into then for some reason decide i want to learn a new language I wont be totally messed up by Perl6. 18:21
Im just a hobiest but might transition into the IT field 18:22
MadcapJake the only thing Perl 6 will do is make you realize other languages aren't very powerful :)
18:22 lostinfog joined
timotimo all languages are equally powerful as long as they can do what brainfuck can :P 18:23
Joe_Perl I love Grammers they seem like a good compromise on Regex.
AlexDaniel compromise :o 18:24
MadcapJake timotimo: powerful in terms of the programmer not the computer
timotimo heh heh 18:25
MadcapJake Joe_Perl: Perl 6 grammars aren't a compromise on regex they are a tool that makes using regexes easier in some circumstances (where you need to parse things larger than one line) 18:26
AlexDaniel m: say ‘abbbbb’ ~~ m/‘a’ b+/ # you can still do your regexy stuff 18:27
camelia rakudo-moar ac36d2: OUTPUT«｢abbbbb｣␤»
MadcapJake grammars are to regexes what a general is to soldiers
mst wow, that's the worst metaphor I've seen in weeks 18:28
18:28 maybekoo2 joined
MadcapJake lol worst really? how is that so bad? let's see you come up with one :P 18:29
18:30 itaipu left
MadcapJake grammars organize regexes like a general organizes their soldiers, regexes capture text just like soldiers capture enemy combatants :) 18:30
perlpilot MadcapJake: What are the lieutenants and sergeants in this metaphor? they're the one who actually talk to soldiers. :-) 18:31
18:31 itaipu joined
MadcapJake perlpilot: those are your tokens and rules! the actual regex literals are the soldiers :P 18:31
ugexe but what about actions
MadcapJake orders
perlpilot ugexe: obviously they are the intelligence community
18:31 cdg joined
MadcapJake hahaha 18:31
perlpilot MadcapJake: Work at this a bit and it could become a really good lightning talk :) 18:32
18:32 zakharyas left 18:33 cdg left, cdg joined
perlpilot "Text as the enemy" works fairly well in any case 18:33
18:34 ufobat joined
MadcapJake yeah it does! I think it's a pretty good metaphor! 18:34
AlexDaniel well, if only it was part of military training
perlpilot speaking of training ... where's bootcamp in this metaphor?
MadcapJake Grammar::Debugger is bootcamp :) 18:35
perlpilot and what branch of the service are these soldiers in? Air Force? Marines? Navy? Army?
mst nah, bootcamp is for infantry
perlpilot Coast Guard?
18:35 itaipu left
mst bootcamp is an interactive grep tutorial 18:35
Joe_Perl At what scale (page traffic, database size etc.) would you need to worry about any speed deficits in perl6?
MadcapJake lol yes!
mst Grammar::Debugger is either officer training school, or more likely 'high command's post mortem team'
MadcapJake lol
perlpilot Joe_Perl: The question is a tad vague. 18:36
MadcapJake I'd say this is either Army or Coast Guard (two branches most likely to be capturing rather than just pummeling :)
perlpilot Then what are the Army Rangers? 18:37
MadcapJake unicode properties :) 18:38
perlpilot MadcapJake: you're certainly starting to live up to your nick ;)
MadcapJake hahaha it actually was a nickname I had at work a few years back :) 18:39
18:42 itaipu joined
Joe_Perl I suppose what i have in my head is the story that " Twitter had to switch to scala because Ruby was too slow at scale" Im not saying im going to have a company that big its just a theoretical question. 18:42
mst Joe_Perl: yeah, and twitter would never have got that far if they hadn't used ruby first for fast prototyping 18:43
ugexe that same story holds true for how many companies?
mst Joe_Perl: you're learning exactly the opposite lesson to the truth here
Joe_Perl Im willing to learn
mst I just explained the point. what wasn't clear?
Joe_Perl So perl6 would not hold up at " twitter scale"? 18:44
ugexe if you get big enough then chances are you can afford to handle those consequences. if you don't then you didn't waste a bunch of time/effort on what essentially is premature optimization
Joe_Perl ugexe: good point 18:45
mst Joe_Perl: if you think that question matters, you didn't listen to me.
Joe_Perl: right, ugexe is making the same point I already did
twitter switched *when it got big enough to matter*
until then, worrying about twitter scale is masturbation, not programming
RabidGravy :-O 18:46
ugexe are we web scale enough?
mst and as with other forms of masturbation, best done privately and not in the office
timotimo we're not scale enough
RabidGravy it's a clear sign you're a team player
;-)
18:46 itaipu left, Joe_Perl left 18:47 pmurias joined
AlexDaniel
.oO( web scale masturbation )
 18:47
ugexe perl6 is great for prototyping at the very least
pmurias AlexDaniel: a large part of the internet is devoted to that ;)
mst AlexDaniel: funnily enough, a large part of youporn's early rise was because they had more efficient CDN usage than the competitition because of the tools they got via catalyst 18:48
timotimo perl catalyst, eh?
perigrin yes. 18:49
timotimo that's a neat story
why doesn't perl brag about porn sites all the time? :P
tadzik it is!
AlexDaniel huggable: dunno
perigrin because it was later re-written in PHP
AlexDaniel :(
tadzik (because they switched to php)
perigrin but with a much more sound architecture
tadzik: strangely they didn't, the company they sold the IP to did.
tadzik ah 18:50
mst well, because it was bought out by another porn company with an existing PHP codebase and an existing PHP team
at which point the port was totally reasonable
mst remembers dropping into HN comments and telling people to stop complaining, I don't like PHP but they were still absolutely in the right business wise
18:51 grondilu left
perigrin I think they green-fielded the re-write ... but yeah 18:51
RabidGravy back in the last century if you had perl on your CV you were going to get a call from a recruiter about working on some "adult" web site at least once a week
ugexe the first time i was yelled at by mst was for not pasting my entire error logs, which i had edited out because i was developing a catalyst app for a porn company lol
RabidGravy :)
mst perigrin: yes, but *existing team* 18:52
perigrin mst: yeah, no the re-write was entirely justified. 18:53
mst if they'd listened to me the catalyst rewrite would've finished but *shrug*
they paid their bills while not listening
timotimo fantastic
RabidGravy I have a love/hate relationship with clients like that 18:54
tadzik heh, sex.pl was run by perl for a while too
but it was just a perl script generating a static html
mst ugexe: heh, I can see your reasoning but I hope in hindsight it's obvious why my brain was going "I doooon't caaaare I just want to fiinnd your buuuuggg" :D 18:55
18:56 Amnez777 left, edenc joined
timotimo tadzik: of course sex.pl would be run by perl :P 18:56
AlexDaniel or prolog…
timotimo or that
18:58 Amnez777 joined
RabidGravy or pl/1 19:00
19:00 domidumont left 19:01 prammer left, ZoffixWin joined
ZoffixWin AlexDaniel, no huggable for a while. I accidentally nuked my entire server last night :) irclog.perlgeek.de/perl6/2016-05-06#i_12441561 19:02
Luckily, moritz++ gave me all #perl6 logs mentioning addressing her, so I can reconstruct her database by re-issuing those commands.
I hope you didn't add a ton of factoids in /msg :P 19:03
ufobat ZoffixWin, i just tried your OO example. and it works for me :-(
AlexDaniel ZoffixWin: where are your backups?
19:03 edenc left, edenc joined
ZoffixWin heh, good question :P 19:04
Technically, I didn't lose any data tho :P
sortiz m: my @a := List.from-iterator((loop { }).iterator); Mu; # Reify delayed, good. 19:05
camelia rakudo-moar ac36d2: OUTPUT«WARNINGS for /tmp/1ZGRZkg8yw:␤Useless use of constant value Mu in sink context (line 1)␤»
sortiz my @a := Array.from-iterator((loop { }).iterator); Mu; # Why try to reify?
m: my @a := Array.from-iterator((loop { }).iterator); Mu; # Why try to reify?
camelia rakudo-moar ac36d2: OUTPUT«(timeout)WARNINGS for /tmp/19TRs_9jEO:␤Useless use of constant value Mu in sink context (line 1)␤» 19:06
rindolf ZoffixWin: hi ,sup? 19:07
ZoffixWin \o
19:09 azawawi joined
azawawi hi 19:09
yoleaux 5 May 2016 21:50Z <arnsholt> azawawi: It's maintained in the sense that I'll accept patches and such, and help out to the extent that I have time. But I'm super-duper busy ATM, sadly, so I'm not actively hacking on it ATM
sortiz Any good reason why Array try to reify its iterator at creation?
rindolf azawawi: hi, sup?
ZoffixWin ufobat, are you sure you tried the class version? I've just tried it again of a clean perl6/Bailador install and I get the template error, because it's using wrong template renderer 19:10
azawawi rindolf: done
ufobat ZoffixWin, i updated the issue
with my example
azawawi just finished reading zmq ffi perl5 code
ZoffixWin ufobat, oh, well, it does start, but try going to localhost:3000/ 19:11
ufobat it prints the renderer and its the ::Extended
ZoffixWin k, lemme try running your version 19:12
ufobat ZoffixWin, i added a say self.renderer.perl; in my / route and it printed the ::Extended as well 19:13
ZoffixWin I'll take a closer look once I'm done eating :) 19:15
ufobat enjoy the dinner :)
i've got a question regarding this issue. github.com/ufobat/Bailador/issues/79 . I assume that this erorr happends when you declare a class twice? 19:17
19:18 jjido joined 19:22 Relsak left
arnsholt azawawi: Speaking of ZMQ, maybe I should just give you a commit bit? 19:22
sortiz For anyone curious I'm asking what is the case of github.com/rakudo/rakudo/blob/1ee2...ray.pm#L53
azawawi arnsholt: cool :)
arnsholt You azawawi on GitHub as well? 19:23
azawawi yup
github.com/azawawi/
moritz ufobat: yes, it happens when two files have an our-scoped symbol Store
arnsholt azawawi: There you go. Have fun! =)
azawawi arnsholt: thanks :) 19:24
arnsholt: going to finish the book and then start hacking at it soon
ufobat moritz, could that happen if you install a module and have it in the lib directory at perl6 -Ilib .... for example?
moritz ufobat: well, a "use Store" should only load one version, either the installed *or* the local oen 19:25
azawawi arnsholt: been looking at what we need to support namely api.zeromq.org/4-1:_start and api.zeromq.org/3-2:_start
ufobat allright, thanks :) 19:27
19:28 grondilu joined
azawawi hmmm... I dont understand one thing. why does Debian skip versions packages.debian.org/sid/libzmq5 (libzmq5, libzmq3, no libzmq4) 19:30
moritz TIL that webkit uses WTF as an internal namespace: webkit.org/blog/6161/locking-in-webkit/
AlexDaniel azawawi: libzmq3 version is 4.0.5 19:31
azawawi: maybe it is 4? 19:32
moritz "WTF stands for Web Template Framework"
AlexDaniel libzmq5 version is 4.1.4… hmm…
azawawi AlexDaniel: no i understand that... but why did they skip using libzmq2 and libzmq4? is that normal?
AlexDaniel azawawi: well the only reason to have different versions of the same program is if both are needed at the same time. E.g. python2 and python3 19:34
19:34 itaipu joined
AlexDaniel azawawi: I don't know why that happened in this particular case though 19:34
but it could be anything
azawawi AlexDaniel: it seems the maintainer likes odd version numbers :) 19:35
rindolf moritz: heh, heh. 19:42
19:42 pmurias left 19:44 pmurias joined 19:51 azawawi left
ugexe ZoffixWin: i think thats a rakudo/CUR bug. i replied with more details 20:00
20:03 jjido left, uruwi left 20:04 grondilu left
ufobat good night 20:06
20:06 ufobat left
ugexe it mostly boils down to CUR wants to handle all things precomp, so if something needs to be re-precompiled its CURs job to do it. total reinstallation (the only way a package manager can re-precompile in this situation) involves things other than precompiling that don't need to happen 20:08
mst part of me continues to wonder if that was a good plan, but OTOH trying to maintin precomp out of step with rakudo would be its own set of nightmares 20:09
ugexe if a workaround is implemented (such as reinstalling all reverse depends) it puts less pressure to fix what is clearly a problem
mst right, if CUR is going to own precomp, there should be a re-precomp method, which will re-precomp revdeps, I guess 20:10
or do you mean that re-precomp should've triggered during installation of something and didn't?
ugexe i would think it should have already happened; that the precomps get invalidated and thus re-precompiled 20:11
mst right
but we should probably expose a way to force it
because tish and tyops hapne
ugexe yeah. an api to precomp would be nice, especially for doing bad things like parallel precompilation 20:12
to quote latarian milton: its fun to do bad things 20:14
timotimo "shiny; let's be bad guys" 20:15
20:16 |2701 joined
lizmat masak: re irclog.perlgeek.de/perl6/2016-05-06#i_12443272 : I'll gladly convert these optimizations back to "proper" Perl 6 code :-) 20:17
once we're sure the optimizer can handle them 20:18
like I've just done with (almost) all i = i + 1 -> ++i changes I did yesterday
20:19 maddingue left, maddingu1 joined
AlexDaniel lizmat: I was so happy to see that :) 20:20
20:21 Amnez777 left 20:22 Amnez777 joined 20:23 kaare_ left 20:25 infina left 20:27 infina joined, infina left, infina joined 20:30 darutoko left, itaipu left
[Coke] timotimo++ serenity reference. 20:30
masak lizmat: cool. that's what I was hoping :) 20:34
in a perfect world, all the "natural" ways to write things will also be fast, and then we should be using them in the setting ;) 20:35
lizmat yup, agree :-) 20:36
masak 'night, #perl6
lizmat night masak 20:37
sortiz night masak
20:39 itaipu joined 20:45 sufrostico left 20:48 bazzaar joined
bazzaar \o 20:49
.oO wonders if any good soul is enthused by the potential of using a perl6 grammar to parse a schema-only DB_dump file ? 20:52
mst interesting idea, you might want to look at the recdescent etc. parsers in the perl5 SQL::Translator module for bits to steal 20:55
timotimo yeah, especially stealing tests is worth a lot i bet
bazzaar thanks for the pointers, I'll take a look. 20:56
20:57 hotel|netteiru is now known as hotel
ZoffixWin Found my issue for why Bailador's class-y interface was failing me last night. I still had `use Bailador` in the file and it was setting up a default app and not my own or whatever. 20:59
hotel How do I give something optional arguments? or should I just use a multi?
bazzaar I've written and debugged such a grammar, ... only tested it against my own db schema though so probably not exhaustive
ZoffixWin hotel, named args are optional by default, positional args, just add '?' after them
hotel named argsare the ones with colons right?
actually I'm probably better off using a multi 21:00
ZoffixWin m: sub foo ($a, $b?, :$yes) { say quietly "$a $b $yes" }; foo 42; foo 42, 45; foo 42, 45, :72yes;
camelia rakudo-moar ac36d2: OUTPUT«42 ␤42 45 ␤42 45 72␤»
ZoffixWin hotel, oh, or add a default value: 21:01
hotel yeah I know all about those, I come from python
ZoffixWin m: sub foo ($a, $c = 2, $b?, :$yes) { say quietly "$a $c $b $yes" }; foo 42; foo 42, 45; foo 42, 45, 555; foo 42, 45, :72yes;
camelia rakudo-moar ac36d2: OUTPUT«42 2 ␤42 45 ␤42 45 555 ␤42 45 72␤»
21:02 TEttinger joined
ZoffixWin I think I'm gonna repeat my Perl 5 mistake and make my own Perl 6 web framework.... 21:04
Bailador just doesn't fit into my brain. 21:05
moritz port Mojolicious to p6?
21:05 jjido joined
mst have a look at my Web::Simple stuff for ideas as well simply because it's probably the most flexible perl5 dispatcher in existence 21:05
and is already kinda slang-like in its own way
bazzaar Wondering if anyone has a PG schema of their own, that they could run past my perl6 grammar, to flush out additional bugs
moritz bazzaar: what's a PG schema? 21:06
geekosaur postgresql
hotel so does this make sense? gist.github.com/HotelCalifornia/de...c4755aa421
ZoffixWin mst, will do
mst ZoffixWin: I suspect the controller lifecycle stuff from mojo will be worth grabbing, plus look at how catalyst uses role composition etc. to be more DRY for large scale apps, plus Web::Simple for my ideas of "a better dispatcher" (Mojo's dispatcher is a much easier API to an about-as-powerful-as-catalyst dispatcher, which is a different achievement) 21:07
moritz bazzaar: github.com/moritz/quelology/blob/m...schema.sql old and nothing fancy in there
also github.com/moritz/ilbot/blob/maste...postgresql 21:08
21:08 Skarsnik left
moritz this one contains a function at least :-) 21:08
bazzaar moritz: wow, that was quick :-) .... will definitely use those, I see some things not in my schema 21:10
hotel anyone? 0w0 21:12
ZoffixWin hotel, not really. On lines 2 and 3 you're declaring class variables, you know that? 21:15
bazzaar sql [in the db_dump] to build the db relations is relatively simple to parse, and I'm guessing would be quite alike between diff db_servers, and no database access necessary to be able to do all sorts of fun stuff with the db relations in perl6
hotel yeah?
ugexe hotel: it looks ok. you could also use `submethod BUILD` instead of new: submethod BUILD(Str :$!server-addr!, Int :$!server-port = 8080) { $!socket = ... } 21:16
hotel what's the difference between BUILD and new?
timotimo very big difference 21:17
for one, every class in the chain of derivation has their BUILD called
but only a single method new will be called
geekosaur new means you have to implement and orchestrate all the superclass building and such. BUILD is invoked byt he default new when it's time to initialize your class, with the superclasses already done
timotimo (the difference for submethods is that they don't get derived to subclasses) 21:18
hotel hm
timotimo BUILD is actually invoked via "self.bless", which your own method new will most likely call, too
21:18 ZoffixWin left
hotel But if I don't have a hierarchy... 21:19
ugexe the short answer is you need to use `new` when you want to pass positional arguments, not all named
hotel ah
timotimo you can't get "no" hierarchy. you have to derive at least from Mu, but if you don't give any target to derive from, you'll derive from Any
Xliff_ In P6 BUILD is the better equivalent of an OO constructor, rather than new(). 21:20
timotimo but Any doesn't do much in its construction, so it's n big deal
Xliff_ BUILD == assign values to attributes. new == perform actions required for the construction of a class instance.
At lest that's how I've begun to see it.
ugexe someone contrive a reason to sneak a BUILDALL in 21:21
Xliff_ ugexe, and BUILDALL does...?
ugexe call in the opposite order of BUILD, after its been initialized 21:22
21:22 hotel_california joined
hotel_california so then the body is the same in this case? 21:22
ugexe you dont need to `bless` in BUILD
21:22 hotel left, hotel_california is now known as hotel
hotel has a garbage connection 21:22
ah okay
Xliff_ Am I correct in assuming that bless is called by new, implicitly? 21:23
s/bless/self.bless/
ugexe BUILD is easier to subclass, so unless you really want positionals in your new you probably want to use BUILD
21:24 pmurias left
ugexe if you declare your own `new` method you have to call self.bless(|%_) yourself 21:24
mst hotel: short version: you should be using BUILD 21:25
hotel thanks all
mst hotel: if you think you can't use BUILD, you're probably wrong, ask here
hotel: then if we can't find a way to fix it without writing new(), then we can discuss doing it the hard way 21:26
hotel didn't know about it before 5 minutes ago
21:32 Sgeo joined 21:34 itaipu left 21:35 sortiz left, jjido left 21:38 jjido joined 21:42 yqt joined
hotel fried his brain, will try again later 21:45
ugexe i posted a comment to your gist with an untested example 21:46
21:46 bazzaar left
hotel nah, I got that part, just trying to translate from python to p6 while following a spec 21:46
but thanks 21:47
cxreg the more I learn Go the more I wish it was Perl 6 21:49
21:52 jjido left 21:56 kaare_ joined 21:58 mr-foobar left 21:59 mr-foobar joined 22:04 kaare_ left
AlexDaniel
.oO( the more you learn GO the more you wish it wasn't Go? )
 22:04
Go*
cxreg there are nice parts but those equivalent parts are also nice in p6 22:11
and the icky parts are better in p6 :p
AlexDaniel
.oO( use perl 6 then? )
cxreg don't you live in a nice bubble :p heh 22:12
RabidGravy I do :) apart from some tiny bits of C, Perl 5 and Javascript I've done nothing else but Perl 6 for a year 22:17
(I probably should get a proper job at some point ;) 22:18
22:20 rindolf left
cxreg I've been binging on "recent stuff that people are hiring people to do" 22:24
there are precisely zero jobs asking for perl 6 experience
maybe one day
hotel what is the point of keeping a Supplier around? can I just say my $supply = Supplier.new.Supply;? 22:26
smls m: say "a" eqv "a"
camelia rakudo-moar ac36d2: OUTPUT«1␤»
smls ^^ Why is this returning 1 instead of True? 22:27
hotel why not use eq?
smls It's a golfed-down example :P 22:28
hotel: Re. Supplier, you need it because you cannot call .emit on a Supply directly.
geekosaur isn't there a ticket for that? some kind of nqp shenanigans iirc
22:29 protium joined
hotel ooh, I get it now 22:29
smls++
cxreg m: say "a" eq "a" 22:32
camelia rakudo-moar ac36d2: OUTPUT«True␤»
hotel "The server MUST provide a sane Supply that emits Blob objects..." but only suppliers can emit? 22:33
cxreg m: say ("a" eqv "a") === True 22:34
camelia rakudo-moar ac36d2: OUTPUT«False␤»
cxreg m: say ("a" eq "a") === True
camelia rakudo-moar ac36d2: OUTPUT«True␤»
cxreg o_O
hotel lol
smls hotel: When you call the .emit method on the Supplier, it becomes available to all subscribers of the Supply. 22:35
cxreg i think that's a regression, i have an old build that returns True
hotel so basically I'm just supposed to have both
or I guess I have to anyway
just wondering why it says the supply should emit stuff 22:36
smls it is intended to mean the same thing
just doesn't use the language very precisely 22:37
hotel gotcha
smls You hand the Supply to the down/up-stream code, and keep the Supplier locally in your lexical scope where you want to emit stuff on the Supply.
ugexe cxreg: maybe related to this? github.com/rakudo/rakudo/commit/fe...c2e23c3a40 22:39
22:40 wamba joined
cxreg the issue isn't with ===, but with what eqv returns 22:40
ugexe er, sorry its the commit *after* that 22:41
github.com/rakudo/rakudo/commit/46...99ae02c2b9 22:42
cxreg yes 22:46
at least it looks like a likely culprit
ugexe i got True for 2014.04 fwiw 22:47
hotel =:= ?
cxreg i had True in 2015.12
hotel m: say 10 =:= 11
camelia rakudo-moar ac36d2: OUTPUT«False␤»
hotel m: say 10 =:= 10 22:48
camelia rakudo-moar ac36d2: OUTPUT«True␤»
cxreg hotel: compares aliased containers
ugexe er 2016.04 heh
cxreg oh heh
ugexe i.e. it had to have happened in the last few weeks
hotel what are these so-called "containers" people keep talking about? 22:49
geekosaur m: my $a; say $a.VAR.^name 22:50
camelia rakudo-moar ac36d2: OUTPUT«Scalar␤»
hotel m: my $a; say $a.VAR 22:52
camelia rakudo-moar ac36d2: OUTPUT«Any␤»
cxreg m: say "a" =:= "a" 22:53
camelia rakudo-moar ac36d2: OUTPUT«True␤»
hotel bork I can't think anymore 22:54
22:54 hotel is now known as hotel|out
geekosaur before you run away too far, consider 22:57
m: my $a; say $a.^name
camelia rakudo-moar ac36d2: OUTPUT«Any␤»
hotel|out so it helps with introspection?
geekosaur a variable like $a normally gives you its value. but it's actually bound to a container --- that is how it can be mutable 22:58
cxreg say my name say my name
smls hotel: Basic value types like strings (Str) and integers (Int) are immutable in Perl 6. Item containers (also called "scalars") are what makes it possible to use Perl 6 for imperative programming nonetheless.
E.g. when you write «my $i = 42;», it stores an Int object representing the number 42, in the item container represented by the variable $i.
And then when you call «$i++», it does not modify the Int object (which is immutable), but replaces it with another Int object representing the number 43.
Every normal $ variable, and every slot of a @ variable, represents such a container. 22:59
22:59 firstdayonthejob left
hotel|out so /that's/ why you can't access the characters in a Str 23:00
interesting nonetheless
AlexDaniel m: my $s = ‘abc’; $s.substr-rw(1, 1) = ‘z’; say $s 23:08
camelia rakudo-moar ac36d2: OUTPUT«azc␤»
AlexDaniel smls: how would you explain this? :)
23:08 kid51 joined 23:09 BenGoldberg joined
geekosaur don;t see how that proves smls wrong. substr's result doesn't know where it came from; substr-rw's is a proxy that knows and can make alterations to the container thereby 23:11
AlexDaniel geekosaur: no-no, it doesn't. I just thought that after a great explanation he can also give another great explanation for this :) 23:12
ugexe m: use MONKEY-TYPING; role StrList { method AT-POS($a) { return self.substr($a, 1) }; }; augment class Str { also does StrList; }; my $str = "abcd"; say $str[2] 23:13
camelia rakudo-moar ac36d2: OUTPUT«c␤»
23:19 smls left 23:21 zakharyas joined 23:22 sufrostico joined 23:23 jjido joined
hotel|out my phone's doing this weird thing where it drops or downloads corrupted audio packets when streaming but playback continues with long gaps of silence anyway 23:27
if only android was built on perl6, this kind of thing wouldn't happen 23:28
23:28 jjido left 23:31 ZoffixWin joined, ZoffixWin left, ZoffixWin joined
literal m: class Foo is Str {}; my $foo = Foo("dsfsdf"); say $foo 23:32
camelia rakudo-moar ac36d2: OUTPUT«Cannot find method 'Foo'␤ in block <unit> at /tmp/gTydPtmEfn line 1␤␤»
literal how should I be doing this?
ZoffixWin I need to generate a date in HTTP Last Modified header format. I don't see any convenient methods in Dateish/DateTime... are there any subs that could make it easier than this: 23:33
my $mod = DateTime.new("main.css".IO.modified); say $mod; printf "%s, %02d %s\n", <Mon Tue Wed Thu Fri Sat Sun>[.day-of-week-1], .day, <Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec>[.month-1] given $mod;
literal, Foo.new
literal ZoffixWin: is there no way to get a short-hand constructer like Str()? 23:34
ugexe method CALL-SAME($a) { self.new($a) } 23:35
er CALL-ME
ZoffixWin m: class _Foo is Str {}; sub Foo { _Foo.new: $^a }; my $foo = Foo("dsfsdf"); say $foo 23:36
camelia rakudo-moar ac36d2: OUTPUT«Default constructor for '_Foo' only takes named arguments␤ in block <unit> at /tmp/c_4MS2cjqk line 1␤␤»
ZoffixWin wonders what that throws that error
literal I see. new() requires a named argument, what is it for Str? docs don't mention it
ugexe m: class Foo { has $.str; method CALL-ME($str) { self.new(str => $str); }; }; my $foo = Foo("asdf"); say $foo
camelia rakudo-moar ac36d2: OUTPUT«Foo.new(str => "asdf")␤»
ZoffixWin neat 23:37
23:37 wamba left 23:38 grondilu joined
literal doesn't stringify like Str though 23:39
or .gistify
23:39 cdg left
ugexe $.str handles /*/, its all in what you make it 23:40
maybe you really want to do role Stringy instead of inherit from Str
23:40 cdg joined
literal I see didn't know about Stringy 23:42
hotel|out roles are like interfaces right? 23:45
23:45 hotel|out is now known as hotel
ZoffixWin <ZoffixWin> perlbot, eval: strftime '%a, %d %b %Y %T GMT', localtime 23:45
<perlbot> ZoffixWin: Fri, 06 May 2016 23:45:15 GMT
Is there similar simplicity in Perl 6?
There just gotta be. I refuse to believe the person who decided we needed .weekday-of-month didn't think to add an easy way to format datetimes :/ 23:47
hotel m: strftime "%a, %d %b %Y %T GMT", localtime 23:48
camelia rakudo-moar ac36d2: OUTPUT«5===SORRY!5=== Error while compiling /tmp/sG8tFMlJeP␤Undeclared routines:␤ localtime used at line 1␤ strftime used at line 1␤␤»
hotel shrugs
ZoffixWin Hmmmm
23:49 cdg left
hotel doc.perl6.org/type/Dateish#method_formatter ? 23:49
if I'm reading that right you can set it to your own function? 23:50
:S
ZoffixWin hotel, yes, but that function would have to be huge like this: github.com/rakudo/rakudo/blob/nom/...ime.pm#L10
Which is what I'm trying to avoid :P
It probably should belong in POSIX module: modules.perl6.org/dist/POSIX 23:51
The strftime thing that is
hotel wow
can't you use nativecall then? 23:54
or do you need it to be really cross-platform? :P
ZoffixWin Why, is it not cross-platform in C? 23:55
ugexe i think he just wants all the date-style directives for %Y etc for sprintf
hotel idk 23:56
guess it is then
so can't you do a native call to strftime?
ZoffixWin tries to hack something up and will submit a PR to POSIX
hotel lol
23:57 hotel is now known as hotel|out
ZoffixWin hotel, well, I don't know C... So that will involve "learning things" :) 23:57
hotel|out man7.org/linux/man-pages/man3/strftime.3.html
BenGoldberg Technically, you don't need to learn C. You just need to learn the types used by C. 23:58
ZoffixWin Sure :)
BenGoldberg Some of it is a bit hard, I mean, 'char' is a type whose purpose is to hold a chararacter. Mind blowing! 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!