perl6-projects.org/ | nopaste: sial.org/pbot/perl6 | evalbot: 'perl6: say 3;' | irclog: irc.pugscode.org/
Set by mncharity on 25 March 2009.
cspencer i'm trying to move the shift method over from Range.pir -> Range.pm, but i'm getting a "shift_pmc() not implemented in class 'Range'" error when i do 00:04
is there any way of associating the P6 shift method with the shift_pmc vtable method? 00:05
00:09 cspencer left 00:10 cspencer joined
brunoV pdc, this is a personal suggestion. Before coding *anything* in perl, ask first if it's already been done. Most of the time, it has, and you only have to glue CPAN modules together. 00:27
00:30 plash left
pdc that sounds like it would ruin the fun! 00:30
brunoV pdc, the fun is in looking at your non-trivial program working after 15 minutes of CPAN shopping 00:31
literal "and if you must reinvent the wheel, try to invent a better one"
pdc I take your point though, and reinventing the wheel is something I should eventually stop doing. But in C at least, I quite like coding stuff which I could just import if I wanted
good point brunoV. developing apps quicker would be great and is one reason I plan to learn a scripting lang 00:32
but I use reinvention as learning experiences. Doing a lot of work yourself is (imo) a great way to really get to grips with a language. That's important in C and I should think it is even more so in Perl 00:34
although I keep reading how you don't need to learn all of Perl to start writing useful code, just a small subset with what you need 00:36
00:36 xinming left
brunoV pdc, also, I don't know if they've suggested this to you yet, but www.perlmonks.org is a great place to join, ask question and look to from time to time 00:40
pdc I've read about and briefly visited it, thanks 00:41
00:44 fridim joined 01:00 davidad left 01:03 dKingston left 01:04 dKingston joined 01:14 amoc joined 01:22 plash joined 01:31 xinming joined 01:36 xinming_ joined 01:41 brunoV left, brunoV joined 01:44 xinming__ joined 01:47 xinming left, xinming__ is now known as xinming 02:00 justatheory joined 02:01 ZuLuuuuuu joined 02:02 xinming_ left
pasteling "ZuLu" at 193.140.225.206 pasted "Gives "too few arguments passed"" (5 lines, 97B) at sial.org/pbot/35865 02:03
ZuLuuuuuu Hi, why this code gives "too few arguments passed"? sial.org/pbot/35865
02:24 dKingston left 02:28 ZuLuuuuuu left 02:29 davidad joined 02:43 bacek_ joined 02:46 davidad left 02:47 meppl left 02:49 meppl joined, davidad joined 02:58 c9s_ joined 02:59 c9s_ is now known as c9s__, c9s__ is now known as c9s_
pugs_svn r26098 | diakopter++ | hgis 03:11
03:18 nihiliad left 03:30 meppuru joined, brunoV left
szabgab on perlcabal.org/ shouldn't the links to feather point to perlcabal (shall I remove the hostname from those links?) 03:33
03:34 Tene joined 03:35 sephee left 03:44 mjk joined 03:45 mjk left, mjk joined, mjk left 03:46 mjk joined, meppl left 03:49 justatheory left 03:51 orafu joined 04:03 |mjk| joined, SamB left 04:04 mjk left, |mjk| left 04:06 mib_531n77 joined, OuLouFu left, mib_531n77 left 04:09 mberends joined 04:20 cspencer left, frioux|home left
mberends @tell masak 1) I like Configure.p6 because it resembles rakudo/Configure.pl and proto/lib/Configure.pm. 2) Copying Configure.pm into $project/lib is good 3) PERL6LIB->RAKUDO_DIR iffy, might lead to Rakudo abuse. 4) agree Configure->make autorun off by default, but prefer a prompt asking preference instead of editing Configure.pm (git pull clobbers). 5) PERL6LIB good idea, but 'export' is often preferred to 'env', must find out why. 04:20
lambdabot Consider it noted.
04:21 mberends left 04:23 c9s__ joined 04:24 frioux|home joined 04:35 brunoV joined 04:40 c9s left, c9s__ is now known as c9s 04:42 c9s left 05:03 amoc left
s1n frioux|home: ping 05:09
05:15 amoc joined
pasteling "mibbit" at 192.117.127.193 pasted "just trying" (1 line, 12B) at sial.org/pbot/35869 05:24
05:27 Southen left 05:41 justatheory joined 06:06 FurnaceBoy left 06:10 DemoFreak joined 06:17 nihiliad joined 06:21 justatheory left 06:30 c9s_ is now known as c9s 06:51 masak joined
masak howdy, #perl6 06:51
lambdabot masak: You have 1 new message. '/msg lambdabot @messages' to read it.
masak @massage
lambdabot mberends said 2h 30m 49s ago: 1) I like Configure.p6 because it resembles rakudo/Configure.pl and proto/lib/Configure.pm. 2) Copying Configure.pm into $project/lib is good 3) PERL6LIB->RAKUDO_DIR
iffy, might lead to Rakudo abuse. 4) agree Configure->make autorun off by default, but prefer a prompt asking preference instead of editing Configure.pm (git pull clobbers). 5) PERL6LIB good idea,
but 'export' is often preferred to 'env', must find out why.
06:53 mberends joined
mberends masak: good euro-morning! 06:54
masak morn'!
szabgab mmm
masak I'm composing a quick answer, but now I can tell you directly instead of via lambdabot :)
mberends: 1) I tend not to like '.p6' endings. executables are executables. 3) Rakudo abuse, in what way? 4) go right ahead; I don't fully understand, but it sounds good. 5) not sure we can use 'export' in this instance. I've tried. 06:55
by the way, I'm now dogfooding proto. 06:57
should lead to an avalanche of improvements.
06:57 Southen joined
mberends masak: 1) let's meet in the middle with capitalized Configure 3) a) rakudo/ should not be a library dir in the long term, Test.pm is a homeless exception b) pointing PERL6LIB there establishes the wrong precendent that rakudo/ works as a library dir c) people able to but not understanding a) put more modules into rakudo/ 4) will do soon 5) will research 07:02
masak mberends: 1) deal. 3) agreed. I didn't think of it that way. let's do it your way, and complain loudly for a long-term solution. 07:05
07:06 payload left 07:07 fridim left
mberends masak: i've been offline about a week, but read most of the logs. keep up the dogfooding and proto improvements. Pod::Parser v2 passes 1 test ;) 07:12
masak mberends: thanks. yesterday was some short-and-sweet proto work, today will be Web.pm. good luck with Pod::Parser v2. 07:13
mberends: oh, and my sekkrit plan right now is to finish the Lobster, and then hook up November with HTTP::Server.
07:19 ejs joined 07:20 iblechbot joined 07:22 ejs left 07:23 ejs joined 07:24 payload joined
mberends masak: now the sekkrit's out ;) that's HTTP::Daemon currently, but nvm. I'll mibbit in and out during $class this week :) 07:26
masak ah, I thought HTTP::Daemon and HTTP::Server were two different entities. 07:29
anyway, I'll attempt to tame the daemon today.
07:42 brunoV left 07:48 baest joined 07:50 agentzh joined 07:54 mib_7bnr0h joined 07:55 mib_7bnr0h left 08:02 meppuru left 08:07 bacek_ left 08:09 davidad left 08:16 payload left
jnthn morning, all(#perl6) 08:19
masak jnthn: marnin'.
08:20 nihiliad left, pmurias joined 08:22 zamolxes joined
moritz_ OMG JUNCTIONS, wont't that autothread in some totally unpredictable way? :-) 08:22
masak moritz_: no, it won't compile due to the mismatched parenthesis :P 08:24
so we're safe, for now. 08:25
moritz_ :-)
masak mberends: I have a good mind to write a 'Perl 6 project Well-Thought-Out Practices' somewhere. but it's a bit of work to put everything into words, and I'm not sure it belongs in proto. it's more of a blog post. 08:27
08:28 cxreg joined
masak hm. I'm considering adding an 'all' pseudopackage to proto, which would expand to mean all Perl 6 projects. but do I want it to expand to all installed projects when updating, or do I want a separate 'installed' pseudopackage for that? 08:33
moritz_ the former
masak oki, good.
that one was my favourite. 08:34
08:37 kane_ joined, cognominal joined 08:40 Southen left
masak make some coffee, grab the newspaper, run './proto update all'... gist.github.com/90672 08:42
only works locally on my box right now.
08:59 cognominal left 09:03 cognominal joined
masak there, pushed. 09:08
now you can do './proto install all' as well. :) 09:09
szabgab OT, have any of you created screencast on linux ? I cannot manage to do it
09:16 ruoso left 09:18 PerL_MonK joined 09:19 payload joined
Matt-W is getting twitchy, having not done anything on Form for a while 09:37
I have plans, just... a lack of time
masak I know the feeling.
jnthn There exist people who don't have that feeling? 09:40
pugs_svn r26099 | pmurias++ | [re-smop] fixed bug in interpreter
literal mberends: I believe export is preferred to env because export is the bash way while env is the csh (or some such) way 09:41
masak literal: no, I think that env is the sh-agnostic way. 09:43
literal could be
masak literal: but export does not work in this case. 09:44
because each &run call makes its own subshell.
09:44 pmurias left
masak hm. maybe if the export and the actual command were put in the same &run call... 09:45
that might actually work.
but I'd like a resource saying that export trumps env before changing anything. :)
moritz_ if it starts a new subshell anyway, can't you just do 'VAR=VALUE command'? 09:46
masak moritz_: that's what we currently do. 09:47
except I've been told that doing that _without_ putting an 'env' before is bash-specific, so I stick an 'env' before.
moritz_ ok 09:48
masak shell scripting is gnarly.
moritz_ which is why we love perl :-)
eiro hello guies
masak eiro: oh hai! 09:49
Matt-W Indeed there's a reason why we have Makefile.pl instead of Makefile.sh
hi eiro
masak Matt-W: but the current well-thought-out practice is to have a Configure, written in Perl 6. :P 09:50
Matt-W yeah but the principle's the same :)
masak not many projects do that right now, though.
aye, same principle.
Matt-W I see it as more of a mid-term goal
09:51 Southen joined
masak rakudo: print chr($_) for <20013 25991 65281> 09:53
p6eval rakudo 0310a3: OUTPUT«中文！»
masak Rakudo++
jnthn .oO ( middle woman? ) 09:54
masak the Perl 5 executable gives the same result, but complains about "Wide character" thrice.
Matt-W incomplete unicode fonts --
moritz_ Unicode++ UTF-8++ (I have no idea what these characters mean ;-)
masak: start with the -CS option 09:55
literal masak: -CO :)
or that
Perl 5 likes to assume Latin 1
moritz_ perl -CS -wE 'print chr($_) for <20013 25991 65281>'
(-E is a perl 5.10-ism)
masak jnthn: no, "Chinese (language)". 中 does indeed mean middle, and you understandably confused 文 with 女
literal moritz_: you're missing a qw there 09:56
moritz_ literal: no
literal: <...> works in Perl 5 as well, for obscure reasons
literal oh right, as lons as you don't add some * and stuff
jnthn masak: Well, both look a bit like an ironing board....
literal long
moritz_ literal: actually it's a glob, but if you don't use meta characters like *, ?, {} etc it works
moritz_ too slow ;-)
masak jnthn: the reason China names itself "the middle kingdoms" goes back to early times when the country was a small kingdom among many.
rgs perl -CS -wE 'say chr for <2{0013,5991} 65281>' # golf ! 09:57
masak jnthn: I'm past the point where I see such visual similarities between different sinographs.
moritz_ rgs++
masak jnthn: the first has four strokes, the second has only three. 09:58
moritz_ if you want to golf, you can omit the -w and a space, obviously ;-)
literal rgs: "20013 25591" is actually shorter than "2{0013,5991}" :)
masak rgs++
literal: still, the idea is cool! :)
rgs well
moritz_ but it feels much more golfy ;-)
literal true 09:59
masak jnthn: will there be a Rakudo day this week? 10:01
jnthn masak: Samozrejme.
masak :)
masak thought so
jnthn Probably Thu or Fri.
10:01 iblechbot left
Matt-W likes Rakudo Day 10:01
Matt-W thinks everyone should have a Rakudo Day
Or maybe a Perl 6 Day 10:02
masak every day is Perl 6 day!
jnthn Planning on typed arrays and hashes - or at least a good bit of progress towards them - for Tue or Wed. 10:04
masak nice.
Rakudo Whatever-we'll-call it will be the best release ever! 10:05
s:2nd/' '/-/;
moritz_ masak: so far every rakudo release was the best release ever, at that time ;-) 10:06
masak of course.
nothing says propaganda needs to be logical.
:P
moritz_ jnthn: cool. I just submitted an article which mentions them as not yet implemented; if you do it this week, I might still be able to remove that notice ;-) 10:07
masak perlgeek.de/blog-en/perl-6/tidings-2009-04.html # for those who want to read it 10:08
Matt-W typed arrays and hashes
I thought that already worked...
moritz_ rakudo: my Int @a = <foo bar>; say @a[0] 10:09
p6eval rakudo 0310a3: OUTPUT«get_iter() not implemented in class 'Integer'␤current instr.: '_block14' pc 59 (EVAL_17:41)␤»
Matt-W oh dear
that is bad
jnthn masak: BTW, on Chinese, I didn't study any of it at all for over a year and a half. :-( And I was quite depressed during the month I did spend trying to learn some, so little stuck. 10:10
Would like to have another shot at it sometime though. :-)
masak jnthn: Chinese is a big undertaking. needs virtually undivided attention. at least from me. 10:11
it's not that it's very hard, it's just... a complete, internally consistent set of Other Stuff. a bit like Haskell.
Matt-W I suspect it's slightly harder than Haskell 10:12
Haskell doesn't have hundreds of different characters to learn to read :)
jnthn masak: Yeah, I don't see me being able to give it the time it'd need for serious progress for a while.
masak jnthn: I barely have the time to keep it fresh. 10:13
...while also working half-time and studying other things.
10:14 PerL_MonK left
masak Matt-W: I think you're right; to the extent they are comparable, Haskell is easier. then again, I'd say _all_ programming languages are easier to learn than _all_ spoken languages. 10:14
Matt-W masak: I would agree
masak perhaps with some rare exceptions, like Dis and Toki Pona. :)
Matt-W languages for human communication are generally vastly more complicated 10:15
10:15 Southen left
masak aye. even Esperanto is jarringly complex. 10:15
moritz_ so we'll invent a vastly more complex programming language, just to disprove the "natural languages are more cmplex" theory ;-) 10:16
jnthn I fear the complexity may be a required element of any natural language that's going to be successful in the mainstream.
(Constructed language, that is...)
Matt-W I think you need complexity in order to express the range of meanings which we use in everyday speech
jnthn Otherwise it may well lack the expressivity to make it good to use. 10:17
Matt-W Let alone in literature
masak jnthn: necessary, but not sufficient. exhibit A: all conlangs out there. :P
Matt-W There is a reason why so many synonyms survive in languages like English - because they don't mean *exactly* the same thing
masak Matt-W: and because people have one brain each.
sbp Toki Pona is no exception
moritz_ and because different people want to see different synonyms die ;-)
sbp the words are easy to learn 10:18
but most things are not expressed with single words
and in some of the larger units that you construct with them you can almost think of them as random phonemes, quite frankly
masak sbp: you have a point. Toki Pona does not express things more easily than other languages.
sbp right. like you need five or six words to say "duck"
Matt-W The concepts you want to express are just as complicated
sbp and even then it's not entirely distinguishable from swan, say
(it's something like: bird which sits above water) 10:19
Matt-W Could you cast an analogy there between languages which use the Latin alphabet to construct words, and languages which use a much wider range of characters to represent entire syllables or words?
10:20 orafu left
sbp yeah. there's a good page about this, hang on 10:20
10:20 OuLouFu joined
sbp Matt-W: www.zompist.com/yingzi/yingzi.htm 10:20
Matt-W sbp: that's really interesting, thanks 10:25
10:28 payload1 joined, payload left 10:34 SamB joined, Southen_ joined 10:35 sitaram joined 10:44 payload1 left, payload joined
masak rakudo: role Foo {}; role Bar {}; sub baz(Foo Bar $a) {}; say "should not be alive here before 6.0.0" 10:52
p6eval rakudo 0310a3: OUTPUT«should not be alive here before 6.0.0␤»
masak submits rakudobug
10:53 mib_yk3mtq joined
jnthn masak++ # thanks, that's on my to deal with list 10:59
11:01 ludan joined, mib_yk3mtq left
masak with pleasure. 11:01
moritz_++ for blogging about it. I didn't have a complete overview of the situation before.
rakudo: $*IN.readline # should this one wait for user input? 11:14
p6eval rakudo 0310a3: ( no output )
masak because it doesn't.
11:15 ludan left
masak rakudo: my IO $a; $a.readline # mwhahaha. 11:15
p6eval rakudo 0310a3: OUTPUT«Null PMC access in get_bool()␤current instr.: 'parrot;IOIterator;' pc 6971 (src/classes/IO.pir:162)␤»
masak submits rakudobug
rakudo: IO.readline # golf FTW 11:17
p6eval rakudo 0310a3: OUTPUT«Null PMC access in get_bool()␤current instr.: 'parrot;IOIterator;' pc 6971 (src/classes/IO.pir:162)␤»
11:31 payload1 joined 11:32 payload left
dalek kudo: a04f610 | (Moritz Lenz)++ | build/Makefile.in:
insert newline in Makefile.in, as suggested on p6c
 11:46
11:47 jeremiah_ joined
jeremiah_ rakudo: say "Hello, my name is foo" 11:48
p6eval rakudo 0310a3: OUTPUT«Hello, my name is foo␤»
jeremiah_ ^^ Gabor told me to say that.
masak rakudo: my $name = "jeremiah_"; say "hello, my name is $name" 11:50
p6eval rakudo 0310a3: OUTPUT«hello, my name is jeremiah_␤»
jeremiah_ But my name really is foo. 11:52
masak oh, sorry.
rakudo: my $name = "foo"; say "hello, my name is $name"
p6eval rakudo 0310a3: OUTPUT«hello, my name is foo␤»
jeremiah_ I am going to heckel you during your talk in Oslo masak 11:53
:)
masak is that so?
jeremiah_ Although I think it is spelled heckle.
masak as long as you behave, I don't mind heckling. 11:54
jeremiah_ :)
jnthn Oh awesome does that apply to me too? ;-0
masak jnthn: er.
jeremiah_ Yes, I will heckle you too.\
So sure, it applies.
jnthn jeremiah_: No no I meant could I heckel masak too.. ;-)
erm, however you spell it
masak :)
jnthn <- not a native English speaker, honest ;-)
masak allright, you can all heckle me. I'm sure I deserve it for some reason. :P 11:55
jeremiah_ What is your mother tongue? i386?
jnthn BBC Micro BASIC. ;-)
moritz_ so what does "heckle" mean?
jeremiah_ My mother tongue is cat.
heckle means to harass 11:56
jnthn Did that make learning lolspeak easier? ;-)
jeremiah_ Usually performers on the stage get heckled by people in the audience.
11:56 ihrd joined
ihrd hi there 11:56
jeremiah_ jnthn: Well, yes. But my real accomplishment is translating Ulysses into cat.
That was hard.
masak ihrd: OH HAI 11:57
moritz_ that's easy. `cat Ulysses.txt'
jeremiah_ heh
ihrd masak, HAI 11:58
masak ihrd: как дела? 11:59
jeremiah_ rakudo: use v6; 12:00
p6eval rakudo 0310a3: ( no output )
jeremiah_ rakudo: use v6; say '1..2'; say 'ok 1'; 12:01
p6eval rakudo 0310a3: OUTPUT«1..2␤ok 1␤»
12:01 c9s left
moritz_ rakudo: say (:foo).value.perl 12:08
p6eval rakudo 0310a3: OUTPUT«1␤» 12:09
moritz_ rakudo: say (:!foo).value.perl
p6eval rakudo 0310a3: OUTPUT«0␤»
ihrd masak, fine, I am working on me first patch for rakudo and this is make me happy :)
masak :)
masak is patching Rakudo too
moritz_ ihrd++
Matt-W likes to see happy people 12:14
masak happiness is like the GNU GPL: infectious.
moritz_ isn't it the other way round? ;-) 12:16
jnthn Yeah but so is flu. :-)
12:18 mib_dfzp3n joined 12:21 stevenhirsch joined
stevenhirsch rakudo: say "hello world" 12:22
p6eval rakudo 0310a3: OUTPUT«hello world␤»
12:24 amoc left
masak hello, stevenhirsch. 12:25
stevenhirsch hi masak
masak tips hat
stevenhirsch how are you today?
12:26 iblechbot joined
masak stevenhirsch: sun is shining, I'm hacking Rakudo and coding Perl 6. I'd say pretty good. :) 12:27
stevenhirsch: and you?
stevenhirsch cool! raining here in ny
Matt-W masak: ooh what are you hacking on?
masak Matt-W: sekkrit. :P 12:28
um, and don't backlog.
12:29 daemon left 12:32 Kisu left, Kisu joined
masak for those who don't want to wait until NPW to heckle me, here's a patch that implements IO.ins: gist.github.com/90740 12:34
it passes all spectests, and produces the right results.
but I feel it's not very good PIR, so I post it here, waiting for some kind soul to review.
Matt-W masak: why does it have =item readline and then implement a method called 'ins' 12:35
masak Matt-W: fixed. :)
I saw that one too after posting.
the gist has been corrected, and the patch as well. 12:36
literal .ins ? 12:38
oh, a method, not a file extension
masak aye. 12:39
one that I feel I need now and then.
walk & 12:40
12:41 Kisu left
Matt-W masak: I can't comment on the PIR as my own PIR is dreadful 12:41
12:41 Kisu joined
moritz_ ah, is that Perl 6's $. ? 12:41
the patch isn't very nice, but I don't know how to improve it ;-) 12:43
except writing it in Perl 6, of course
and have a private method inc_ins or so, and call that from pir
literal what does "ins" stand for? I read the Pod and I still don't get the name
moritz_ Industrial-strength Numeric counting-Solution 12:44
short INS, or ins
;-)
literal if only the past tense of read were more distinguishable from its present tense in English 12:45
12:45 payload1 left
literal .recorded works, though 12:46
but that sounds more like something that's been written
moritz_ .read-past-tense ;-) 12:47
moritz_ isn't up for serious discussions today
literal .has-read
moritz_ .I-HAS-READ 12:48
literal that would be the Perl6::LOLCAT grammar version
or...lolcat Setting 12:49
damn, github has cool graphs 12:52
jeremiah_ literal: URL? 12:55
I likes me pretty graphs. :)
literal github.com/rakudo/rakudo/graphs/impact
this for example
and other graphs there under the "Graphs" tab
jeremiah_ Whoah. That is cool.
moritz_ somehow I feel that overrepresents my impact 12:58
masak back from walk 12:59
ok, so we agree that my patch is bad. :)
and I was wrong about the spectests -- it actually makes a few of them fail. 13:00
masak investigates
13:00 pmurias joined
masak whoa... my patch introduces a really strange bug. :/ 13:07
I do 'perl6 t/spec/S16-filehandles/io.rakudo', and after the last line '# FUDGED!', Rakudo re-prints the plan, '1..58'. 13:08
if I add an 'exit 0' at the end of the test file, it stops doing that.
masak compiles to PIR and diffs
moritz_ rakudo: use Test; plan 1; isa_ok True, Bool::True 13:09
p6eval rakudo 0310a3: ( no output )
jnthn masak: Show me the patch? 13:10
masak gist.github.com/90740
the PIR diff says I have an extra 'say' call at the end :/
13:14 donaldh joined, ejs1 joined
jnthn + $I0 = ins 13:15
+ inc $I0
+ ins = $I0
inc ins # should do it
masak it didn't.
jnthn Or even better perhaps
'prefix:++'(ins)
masak because I didn't find a way to initialize it as 0.
and I got an error 'Undef mumble mumble inc'. 13:16
jnthn 'prefix:++'(ins) # should auto-viv
masak ok, trying that.
maybe I should follow moritz_++'s advice too, and put ins-inc'ing in a method... 13:17
moritz_ rakudo: say :!foo.value
p6eval rakudo 0310a3: OUTPUT«0␤»
13:23 araujo left, ejs left 13:28 Tene_ joined 13:38 iblechbot_ joined 13:40 Tene left
masak feh, running the spectests takes a lot of time nowadays... 13:41
moritz_ yes
I should really look into making it parallel again
masak again? what made it not parallel? 13:42
moritz_ the switch away from parrot's harness
masak oh.
PacoLinux for me, making the spectest on saturday took 29 mins, yesterday was nearby 70 mins 13:44
13:45 riffraff joined
masak next time I'll time it. 13:45
13:47 c9s joined
jnthn Ouch. That's...quite an increase. 13:48
masak two thirds of the files in t/spec are listed in t/spectest.data.
moritz_ that's both good and bad 13:49
13:49 c9s_ joined
moritz_ bad because it means we need many more tests 13:49
masak 348 out of 544.
13:49 c9s_ left
masak idly wonders if he could get a non-GSoC grant for writing S09 tests this summer 13:49
13:50 c9s_ joined, iblechbot left
literal you should apply for a TPF grant 13:50
moritz_ I guess the hague grants are limited to actualy implementation work
but I'm not sure about the policy 13:51
masak moritz_: I think so too.
13:51 c9s is now known as c9s__
masak would be fun to apply for a Hague grant, if I found some niche of Rakudo where I felt I could really make a difference. 13:51
13:51 ejs1 left
masak literal: maybe a TPF grant is the right answer for that. just going to finish my current one first. ;) 13:52
13:52 c9s_ is now known as c9s
moritz_ "Grants will be made to projects which obviously and concretely advance the completion of a Perl 6 implementation. Other Perl 6 projects, while potentially very worthwhile, are not the focus of Hague grants." 13:53
from www.perlfoundation.org/ian_hague_pe...ent_grants
that still leaves a bit space open for discussion 13:54
if some implementors assert that thorough tests are vital for this area of implementation
masak maybe I could write tests and then help implement them. 13:55
jnthn: changing to 'prefix:++'(ins) worked fine, all spectests pass now. do I commit? 13:56
jnthn masak: Can you show me final patch you're about to commit? 13:57
masak certainly. hold on. 13:58
jnthn S09 - I am pondering filing a proposal.
To work on some implementation.
Once I'm done with my current grant.
Having test coverage work going on at the same time would to me at least seem a nice thing.
masak I think I would like writing those tests.
gist.github.com/90763
13:59 payload joined
jnthn I think your tests would make a harsh mistress. :-) 14:02
masak: Patch looks OK to me. 14:03
You could btw have written method ins in PIR
erm
in Perl6
method ins { $!ins }
masak I'll do that, and then commit.
moritz_ would a 'has $.ins' also work? 14:05
jnthn not here 14:06
moritz_ because it's used from PIR before?
jnthn well, it may if the stage 1 promised to never use them...
moritz_ which adds unwanted interdepencies, right 14:07
jnthn If we totally pull IO out of the stage 1, OTOH (which would be fine) then yes, that would work then.
Yeah, I'm trying to keep the S1 clean of dependencies on the setting.
Not entirely sucessfully though. :-(
moritz_ make spectest takes 23 minutes here 14:08
masak would .ins be 'is export' or not?
moritz_ masak: don't think so 14:09
jnthn think not
masak good, then I don't think so either. 14:10
moritz_ democractic coding ;-)
PerlJam When we we know how many slots TPF has for gsoc? 14:11
masak the tyrrany of the #perl6 majority. :)
PerlJam (or, if "we" already know, how many slots did TPF get? :)
literal close to April 20th probably
14:14 brunoV joined, alester joined 14:15 Front_slash joined 14:19 dKingston joined
jnthn OK, time for me to go to pm group 14:21
ihrd jnthn: bye! 14:23
14:23 b2gills joined
PerlJam jnthn: have fun! 14:23
jnthn Will do! :-)
14:25 meppl joined 14:28 ihrd left 14:29 pjcj left 14:30 pjcj joined
pmichaud good morning #perl6 14:35
PerlJam good morning Pm
masak hi pm 14:37
dalek kudo: e83b2b8 | (Carl Masak)++ | src/ (2 files):
implemented IO.ins
 14:43
pmichaud (hague grants and tests) -- I think a case could be made that S09 tests are on the critical path. 14:44
14:45 brunoV left
literal masak: btw, what does "ins" stand for? 14:47
masak literal: the number of records (normally lines) that have been read in. 14:48
literal yes, but "ins" isn't a word :P
masak literal: neither is sprintf.
literal but it means something 14:49
moritz_ but it is short for "string print formatted" or so
literal is "ins" supposed to be a plural of "in" or something?
moritz_ that might be it
masak aye. 14:50
literal I would have gone for .nr (like Perl 5's $NR)
masak the alternatives were much worse, and then TimToady suggested 'ins'.
I immediately took a liking to it.
PerlJam literal: you mean like AWK's $NR *shudder*
literal I guess
PerlJam masak: so ... does that mean there's an IO.outs in our futures? 14:52
masak PerlJam: only time will tell. but S32/IO.pod seems to suggest that. 14:53
14:54 elmex joined 14:55 brunoV joined 15:00 exodist joined
mberends ahem: .ins and .outs: irclog.perlgeek.de/perl6/2009-03-14#i_984777 15:01
masak mberends++
but know that that was after throwing about alternatives for quite a while. 15:02
mberends yep
literal cool
15:04 cognominal left 15:06 mofino joined 15:09 justatheory joined 15:19 cognominal joined 15:20 donaldh left, donaldh joined
dalek kudo: 8f4dc52 | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 349 files, 8436 passing, 0 failing
 15:22
masak accidentally relies on the (nonexistent) laziness of =$*IN 15:23
PacoLinux make spectest back to 27m38.179s (maybe my machine was cpu busy yesterday :) ) 15:28
masak probably. 15:30
[particle] pacolinux: what was it before?
masak it took 28 minutes here.
PacoLinux saturday 29 mins, sunday 77 mins 15:31
masak today's short'n'sweet script: one that extracts all TODO comments out of a Perl script. here done for proto's installer: gist.github.com/90800
15:32 nihiliad joined
PerlJam your length tolerance on one-liners is clearly more than mine. 15:33
masak :P
it can probably be golfed a bit; a one-liner to me is simply a script that can be written on the command line. 15:34
PerlJam Me too, but there's a threshhold at which it becomes an actual file on disk and yours would have reached that threshhold for me. 15:35
literal my $line = $*IN.readline <-- can't this be written as my $line = =$*IN; ?
PerlJam (of course, it would have probably then started gaining features and such)
masak literal: yes.
PerlJam: the command line has an impressively short feedback loop. at least for me. 15:36
it has many disadvantages, too. but I like the short feedback loop.
15:44 Tene joined 15:46 dukeleto|work joined 15:47 mberends left
dukeleto|work mornin' 15:47
moritz_ hi there
masak oh hai 15:48
mofino Can you guys feel that? IT'S PERL SIX! 15:53
masak mofino: indeed. 15:54
15:54 pedrolance joined, pedrolance left
masak mofino: so, have you written your first module in Rakudo yet? :) 15:54
mofino no sir, I was just thinking about that 15:55
it'
dukeleto|work mofino: my sigil sense is tingling
mofino it's time to invest some effort in learning perl6
masak high time, I'd say.
mofino high time?! but it's only noon!
sitaram what's the best way for a perl5 old hand who's kinda lost touch to do that? (learn perl6. fast)
15:56 Tene_ left
masak mofino: use.perl.org/~masak/journal/38184 15:56
mofino: and that was in December. :)
dukeleto|work masak: is there a document describing how to create and package Rakudo modules?
masak dukeleto|work: don't know exactly what you mean, but we're working on similar things in proto.
dukeleto|work: it's quite similar to the Perl 5 way, I'd say. 15:57
sitaram: hanging around here is not a bad idea.
sitaram: otherwise, I can recomment moritz_'s blog, the specs, and Actually Writing Something in Rakudo.
pmichaud Parrot has its own ideas about how language modules should be packaged. I haven't quite decided if I agree with Parrot's notions.
sitaram masak: step 1: I 15:58
masak: step 1: I'm on it (hanging around here...!)
dukeleto|work pmichaud: which parrot notions are you not agreeing with?
brunoV hi all, I want to start playing around with rakudo. I checked out three weeks ago or so, how often do you think I should update?
sitaram step 2/3 -- I'll google and come back if I can't easily/obviously find them
step 4: will do -- thanks masak 15:59
pmichaud the idea that language-specific libraries belong under /usr/lib/parrot, I think.
masak sitaram: np. good luck!
and do ask, about anything.
15:59 c9s__ is now known as c9s_
masak brunoV: depends on what you want to do. 16:00
brunoV: nowadays, it's enough to take the monthly releases, unless you're VERY cutting-edge.
dukeleto|work brunoV: it changes every hour, updating once a day is usually enough to get the latest features
PerlJam brunoV: I update once a day or so just to see what's new. Many times I'm pleasantly surprised. Occasionally I'm bummed that there's nothing new :) 16:01
dukeleto|work brunoV: but if you are looking for stability, stick with monthly releases
16:01 cognominal left
brunoV masak, dukeleto, PerlJam, thanks! so you pull from github, right? 16:01
PerlJam brunoV: aye
masak brunoV: I pull from github, but the projects I develop are tested against the monthlies. 16:02
(so that other people won't have to pull from github if they don't want to)
16:03 Psyche^ joined, xinming left
brunoV thanks all, I'll update then 16:03
dukeleto|work brunoV: yes, I have a fork on github where I add features and then I do a "pull request" to the rakudo github user when I want them to check it out
16:04 xinming joined, exodist left
brunoV dukeleto, ooo but I am so far from messing with the source code! I'm barely starting to play with the language 16:04
16:04 xinming left, exodist joined
pmichaud note that pull requests tend to get lost or ignored right now. 16:04
at least, I'm generally not looking at them.
16:05 Whiteknight joined
dukeleto|work pmichaud: what is the preferred method now? patches and tickets on RT? 16:05
pmichaud Yes. 16:06
See the wiki pages that describe the preferred method.
dukeleto|work pmichaud: good to know
pmichaud: I thought that the github-squashed-feature-branch was also one of the submission methods, but I guess that was mortiz's suggestion 16:07
16:08 cognominal joined 16:09 Front_slash left
literal sitaram: this is definitely useful -> perlgeek.de/blog-en/perl-5-to-6/ 16:10
moritz_ dukeleto|work: after the move to git we were trying different methods; this was one of the attempts
PerlJam slowly kills himself with bratwurst
sitaram literal: I just saw that, and the other 2 categories in it, (perl-6 and perl-tips) -- thanks!
masak sitaram: found the spec yet? 16:11
masak is on his way to go swimming
sitaram not yet; still on one of the other pages + multitasking :(
literal perlcabal.org/syn
masak literal++ 16:12
sitaram thanks... (found moritz's blog easy enough of course...)
masak sitaram: read moritz_'s stuff first. the specs are for in-depth study.
sitaram (and it's the first hit on google for "perl 6 spec") 16:13
masak: looks like it; thanks
actually I did go through the apocalypses a long time ago, but forgot most due to non use, and even my perl 5 is rusty... 16:14
16:14 OuLouFu is now known as orafu
sitaram (looking forward to getting my hands dirty again) 16:14
masak sitaram: the apocalypses are superseded many times over. :)
literal the apocalypses are interesting reading
masak literal: they are indeed.
but they also show their age.
literal more down-to-earth than the specs, though outdated :P
16:15 Patterner left, Psyche^ is now known as Patterner
masak I think a thesis could be written merely on the diffs between the apocalypses and the spec. 16:15
and the rationales behind them.
sitaram hardly remembers any of the apocalypses so their being superseded doesn't matter :) 16:16
(except the regex one; since I used Parse::RecDescent a fair bit once) 16:17
16:17 smtms left 16:18 meppl left
masak sitaram: that'd be S05 nowadays. it's changed quite a bit. 16:18
16:18 masak left
sitaram regexes are what originally hooked me into perl back in the day, so I'll be able to grok them well enough; it's the rest of it that'll be a bit of a challenge :) 16:19
16:20 zamolxes left 16:23 dKingston left
moritz_ IMHO regexes/grammars are one of the killer features of Perl 6, and the one that's best implemented in Rakudo right now 16:23
PerlJam Is it specced somewhere that you can do array-unpacking in pointy blocks? my @aoa = [ [1,2],[3,4] ]; for @aoa -> [$a,$b] { ... } 16:24
pmichaud std: my @aoa; for @aoa -> [$a, $b] { ... } 16:25
p6eval std 26099: OUTPUT«ok 00:02 37m␤»
PerlJam Well, that STD likes it is good enough I guess :)
moritz_ PerlJam: yes, it is. Either in S04 or S06 16:26
16:31 spooneybarger joined 16:33 spooneybarger left 16:41 xinming joined
pugs_svn r26100 | lwall++ | Be explicit about design uncertainty wrt semantics of junctional collapse. 16:41
16:41 hercynium joined 16:45 orafu is now known as OuLouFu, OuLouFu is now known as orafu 16:51 kane__ joined, gbacon left 16:55 kane_ left 16:59 gbacon joined 17:10 donaldh left 17:11 sitaram left 17:12 riffraff left 17:13 c9s_ left 17:16 gdonald joined 17:31 gdonald_ joined, zamolxes joined 17:33 c9s_ joined 17:35 orafu left 17:40 eternaleye left, gdonald left 17:42 zamolxes left 17:46 gdonald joined 17:55 gdonald_ left 17:59 ejs joined, barney joined 18:01 gdonald_ joined 18:08 hercynium left 18:10 gdonald left 18:14 icwiener joined 18:16 gdonald joined 18:22 icwiener left 18:24 stevenhirsch left, hercynium joined 18:25 gdonald_ left 18:26 gdonald_ joined 18:28 PZt joined 18:29 ejs left 18:36 PhatEddy joined
PhatEddy perl: print "hello\n" 18:37
in perl one can print/stringify a qr// regex, is there anything similar for p6? 18:39
18:41 gdonald left
TimToady nothing is specced 18:41
[particle] ohm
PhatEddy For split() I would like to do some introspection to see if a regex does capturing. Trying to worry about cases like /(A)|B/. 18:43
Anyone have a thought on how? 18:44
pmichaud my guess would be to see if either list or hash context produces a non-empty list/hash 18:48
(of the resulting match)
for the regex itself... not sure how that would work.
I don't think we have that level of introspection yet. 18:49
18:50 gdonald_ left
PhatEddy The problem being that if the 'b' matches then nothing is returned but "perldoc -f split" indicates we should add an undef to the return list. 18:51
Seems like maybe a todo ... 18:52
pmichaud rakudo: say split(/b/, 'b').perl; 18:53
p6eval rakudo 0310a3: OUTPUT«["", ""]␤»
pmichaud doesn't look like "nothing is returned"
or if you mean from a match 18:54
TimToady rakudo: say split(/(b)/, 'b')
p6eval rakudo 0310a3: OUTPUT«␤»
TimToady rakudo: say split(/(b)/, 'b').perl
pmichaud rakudo doesn't know what to do with captures yet, no.
p6eval rakudo 0310a3: OUTPUT«["", ""]␤»
pmichaud (in split)
TimToady should possibly be an :option
pmichaud I'm not exactly sure what should happen there. 18:55
should the split parts be interleaved with the captured match objects?
PhatEddy I have a patch for the captures (it's an rt) but '1A2B3'.split(/(A)|B/ doesn't work right yet ...
pmichaud PhatEddy: I'm thinking the spec needs some work.
Perhaps I'm wrong... but I'd want to see a few test cases.
18:56 araujo joined
pmichaud for example, with '1A2B3'.split(/(A)|B/) ..... what should be returned? 18:56
TimToady spec says alternating strings and Match objects, basically 18:57
pmichaud because of the capture, or because of a flag, or ...?
TimToady that's why I suggested a flag :) 18:58
pmichaud Right.
we also have things like
'1A2B3'.split( / $<a>=(A) | B /)
and
TimToady and then it doesn't auto-trigger just because you split on /(x)$1/ or some such
PhatEddy in perl5 the example returns ['1', 'A', '2', undef, '3'] 18:59
TimToady right
pmichaud '1AX2B3'.split( / (A(X)?) | B /)
TimToady that's why it returns a Match in P6 19:00
if you want to get fancier, then a comb is probably better
but if Match behaves like Capture, then maybe it's responsive to list/slice context in terms of flattening 19:01
but I still think that argues for an explicit flag if you want the delimiter match
pmichaud anyway, Rakudo wants a clearer spec. 19:02
(for split)
TimToady then we don't have to introspect the Regex, which seems evil(ish)
not quite as dwimmy as p5, admittedly
pmichaud well, if it was a flag, then ['1', Match of A, '2', Match of B, '3'] would still do what was generally wanted in string contexts. 19:03
i.e., for '1A2B3'.split(/(A)|B/, :flag) { .say } # '1', 'A', '2', 'B', '3' 19:04
TimToady could possibly distinguish flags for the internal context too
:str :list :slice or some such
pmichaud I agree that match can be responsive to list/slice context
TimToady but interspersed Matches might be good enough
pmichaud I'm happy to provisionally adopt interspersed Matches for experimentation purposes 19:05
we just need a flag :-)
fwiw, in PHP it's PREG_SPLIT_CAPTURE_DELIM
PhatEddy pmichaud: should the 'B' really be captued in that last example? 19:06
TimToady :all maybe
pmichaud and would we want a similar option for comb ?
TimToady no
pmichaud okay, good. :-)
TimToady you can always write (.*?) to get that in comb
or some such
comb is really just syntactic sugar for m:g// 19:07
biab &
19:08 dukeleto|work left
PhatEddy Well the rt is 64062 and I was hoping to update it with the newest patches today or tomorrow. 19:13
For the moment I may note that the spec is in progress and add a delim flag I guess?
pmichaud yes. 19:14
19:25 meppl joined
TimToady the reason I suggested :all is that :delim (or :delims) seems to suggest returning *only* the delims 19:27
19:27 NoirSoldats left, plash left
TimToady and the possible abbreviations fo :delims are ambiguous in various ways 19:28
19:32 NoirSoldats joined
PhatEddy for the moment then I expect the delim flag will be ':all' 19:32
pmichaud agreed, :all 19:33
19:34 dukeleto|work joined
PhatEddy I didn't quite understand the other proposed flags but think I can muddle along until they show up in a spec someplace ... 19:37
19:48 dKingston joined
pugs_svn r26101 | lwall++ | [S32/Str] add :all flag to split 19:52
19:57 cdarroch joined 19:58 barney left
PhatEddy Sorry but looking now at the spec and rt 64062 the premise of the ticket, in terms of return values seems not quite right (rt.perl.org/rt3/Public/Bug/Display....?id=64062) 20:11
Where the initial ticket indicates dashes there should be match objects - no? 20:12
[particle] looks like i'm going to have time to work on S19 again very soon 20:13
which means i'll be writing tests. yay! 20:14
moritz_ so did the order of arguments to the sub form of split() just change? 20:19
pmichaud PhatEddy: go by the spec. The ticket has people thinking in terms of p5-split (which is different from p6-split)
moritz_: yes, it looks to me as though someone added 'is export' at some point in the spec's history 20:21
moritz_: I think that's probably an error... checking history.
no, that's not it. Apparently it's been 'is export' for quite some time. 20:23
20:28 sri_kraih_ joined
pmichaud apparently goes back to 2006: irclog.perlgeek.de/perl6/2006-09-24#i_-134460 20:31
I'm not sure where we picked up the p5-meme that split(/rx/, $val) should work. 20:32
maybe the spectests were incorrect and we followed those. 20:33
PhatEddy isn't split(/rx/, $val) the second split prototype in the spec? 20:34
PhatEddy dawns that the spec might be revised 20:35
pmichaud Note that the first (invocant) argument of both split functions in the current S32 are Str 20:36
split(/rx/, $val) would be looking for a function with a Regex as the first arg.
20:42 sri_kraih left 20:55 iblechbot joined 20:57 meppl left
moritz_ dukeleto: ping 21:00
21:07 Whiteknight left 21:08 iblechbot_ left 21:13 araujo left, araujo joined 21:28 PerL_MonK joined 21:32 payload left 21:34 [particle] left 21:42 nihiliad left 21:45 pmurias_ joined 21:46 pmurias_ left 21:49 jan_ left 21:50 pmurias_ joined 21:51 pmurias_ left 21:55 jan joined, pmurias_ joined, jan is now known as Guest59237 21:56 FurnaceBoy joined, Lyle joined
Lyle hi 21:56
PerlJam hello 21:57
Lyle Been playing about with Rakudo for the first time
21:57 pmurias_ left
PerlJam cool 21:57
Lyle Got it working through CGI with apache, but no luck with IIS 21:58
Settings are all right by IIS isn't picking up the output for some reason
21:58 meppl joined, dKingston left
Lyle I've put a request in IIS.net and emailed Jan (ActiveState) about it, but he's away for a week 21:58
21:59 dKingston joined, pmurias left, archpollux joined 22:00 archpollux left, pmurias joined 22:01 c9s_ left
Lyle Trying to get November to work on Vista as well 22:03
PerlJam awesome.
Lyle Not much luck yet, but I have managed to come up with 4 patches to help it build already :) 22:04
moritz_ Lyle: ah, that was you... does Rakudo build on vista without modifications now?
Lyle I haven't downloaded the current tree yet 22:05
I'll give it a go in a min :) I'm still downloading the zip, I'll have to teach myself git at some point
moritz_ no hurry; just let me know the result
22:23 payload joined 22:24 iblechbot left 22:32 PhatEddy left 22:35 Aisling joined 22:46 nihiliad joined 22:51 cognominal left
diakopter std: 3**7;0b1 22:58
p6eval std 26101: OUTPUT«ok 00:02 35m␤»
23:03 exodist left 23:05 pmurias left 23:09 ruoso joined
ruoso Hello! 23:12
diakopter ruoso: hi
ruoso so... what's new... 23:13
23:18 mikehh joined
diakopter in 'gone dark' mode 23:19
ruoso trying to get 3g modem to work on linux... 23:24
is the last change in junctions made to simplify "1 < 1|5 < 5"? 23:25
23:29 Whiteknight joined 23:30 amoc joined
Lyle moritz: yes the latest does build on vista without modification. Thanks :) 23:51
23:53 DemoFreak left

Powered by Cro and the Raku® Programming Language.

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

Thank you!