»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
jnthn Only 2 days of vacation left /o\ 00:02
sisar was hoping for a good April fool's joke about perl6 04:46
sisar (well, the Wikipedia page was not a disappointment) 04:57
sorear defines alternating finite automata, lays the necessary groundwork for regexes with & in O(2^2^n) time : www.geocities.com/stockmeyer@sbcglo...et/alt.pdf 07:43
also contains a proof of the PSPACE-completeness of (a problem equivalent to) quantified boolean formulas, which is why I was looking for this 07:44
... mainly linking this because some months ago I conjectured that regexes with & were not in O(2^2^....^2^n) time for any finite power tower 07:45
nice to know I was wrong... still, I'm not about to implement 2-EXPTIME algorithms in niecza D:
samlh sorear: sorry for disappearing for ~a month now 07:54
phenny samlh: 28 Feb 19:16Z <sorear> tell samlh I am curious what version of Mono.Cairo.dll you have
samlh sorear: and i'm likely to disappear again
sorear Who are you? 07:55
samlh sorear: sorry, a windows user from a while back
sorear: just wanted to let you know that qx/notepad.exe/ and run("c:/windows/notepad.exe") now work on .net 07:58
sorear sweet 07:59
fsergot morning #perl6 o/ 08:45
moritz p6: say (1 - 0.5**128) == 1 09:14
p6eval rakudo 8d4f0c, niecza v16-17-g648c49b: OUTPUT«True␤»
..pugs: OUTPUT«␤»
moritz p6: say (1 - 0.5.FatRat**128) == 1
p6eval niecza v16-17-g648c49b: OUTPUT«False␤»
..pugs: OUTPUT«*** No such method in class Rat: "&FatRat"␤ at /tmp/pTPkOlmXVC line 1, column 6-25␤»
..rakudo 8d4f0c: OUTPUT«True␤»
moritz r: say (0.5.FatRat**128).WHAT 09:15
p6eval rakudo 8d4f0c: OUTPUT«FatRat()␤»
moritz r: say (1 - 0.5.FatRat**128).WHAT
p6eval rakudo 8d4f0c: OUTPUT«FatRat()␤»
moritz r: say (1 - 0.5.FatRat**128).perl
p6eval rakudo 8d4f0c: OUTPUT«FatRat.new(340282366920938463463374607431768211455, 340282366920938463463374607431768211456)␤»
tadzik r: say (1 - 0.5.FatRat**128).Num.perl
p6eval rakudo 8d4f0c: OUTPUT«1e0␤»
moritz oh, infix:<==> converts to Num
so, missing candidates. Easy to fix. 09:16
masak good antenoon, #perl6 09:47
masak 0.5.FatRat is another case of method casting. should we always write FatRat(0.5) instead? 09:55
dalek kudo/nom: 3218a72 | moritz++ | src/core/Rat.pm:
==-compare rats directly

before we would go through Num, which made (1-0.5.FatRat**128) == 1 produce True, falsely		 11:13
ast: 9d639e1 | moritz++ | S32-num/fatrat.t:
check that == for FatRat does not go through Num
 11:15
masak moritz++ 11:19
moritz note that the other comparison operators still suffer from the same problem 11:24
masak worth submitting a rakudobug over, you think? 11:25
I held back submitting the above one, because I had the feeling that you were already fixing it ;) 11:27
moritz sure, submit away 11:28
masak submits rakudobug
masak is looking at PAST and trying to fit the pieces together in his head 11:38
masak hm, mystified by the following: PAST::Val, and a bunch of other well-known PAST node types, don't seem to have a definition anywhere within the perl6/nqp repository. 11:50
are PIR and nqp flexible enough that they recognize a namespace even if it hasn't been declared anywhere? 11:51
and if so, what handles all the PAST::Val.new and PAST::Val.ACCEPTS calls everywhere?
moritz masak: I think they are inside the parrot repo 11:52
masak: compilers/pct/src/PAST/{Node,Compiler}.pir
masak ah. 'course. 11:56
moritz++
the ability to re-open any namespace from any PIR file constitutes a very "powerful" feature ;)
but it's no worse than Perl 5, I guess. 11:57
moritz aye
masak ah, yeah. this is the stuff. 11:58
masak looks like the contents of compilers/pct/src/PAST/{Node,Compiler}.pir, plus some things from the nqp repo's src/PAST/NQP.pir, make up what would need to be ported to QAST. 12:08
it's code describing the node layout, and a bunch of as_post methods to compile PAST nodes to POST.
heh, Java 7 allows the syntax "catch (IOException|SQLException ex) {" to avoid having to repeat the same catch clause with only the exception type differing. 12:15
Java 7 also introduces 0b binary literals.
moritz junctions! 12:22
tadzik :) 12:23
arnsholt Very limited junctions though =)
gfldex java junction: wiki.openttd.org/images/d/d0/3way_basic3.png 12:25
per6 junction: wiki.openttd.org/images/c/c0/QuadBr...nction.png
moritz www.perl-community.de/bat/poard/thread/17283 # German perl forum 12:26
tadzik eek 12:28
gfldex the aprils fool so far is from a german blog that never had a comment function 12:29
he added one with very advanced captcha
blog.fefe.de/?ts=b189f277
it's worth to hit shift-reload a few times
*the best 12:31
masak laughs at this 12:39
masak p6: enum Month <_ January February March April>; say Month("April") 13:41
p6eval niecza v16-17-g648c49b: OUTPUT«April␤»
..pugs: OUTPUT«*** No such subroutine: "&enum"␤ at /tmp/B2FCGdwNJJ line 1, column 1-44␤»
..rakudo 3218a7: OUTPUT«Method 'gist' not found for invocant of class 'Undef'␤ in sub say at src/gen/CORE.setting:6244␤ in block <anon> at /tmp/1NdbbJ5bjT:1␤␤»
masak hm. is this rakudobug known?
[Coke] I also am experience feather networking issues this past week or so. 13:46
leont FYI: bioperl.org/pipermail/bioperl-l/201...36407.html 13:48
tadzik seems legit 13:49
leont I'm not sure what to think about it though 13:50
tadzik he could have posted it on a better date :)
and, TPF is not on gsoc anyway :/
leont But Open Biology Foundation is ;-)
tadzik oh, nice :) 13:51
masak "implement Bioperl in Perl 6" sounds like a larger task than a GSoC task.
leont s/Biology/Bioinformatics/
masak: exactly
masak mind you, one might get off to a good start.
leont I mean, I'm sure perl6 is learnable in that amount of time, but I do assume that takes time too
masak I'd advise this person to get into contact with #perl6 ASAP. 13:52
leont I'm not sure how much useful stuff one would get done
Yeah, and #bioperl too􏿽x85
masak indeed, learning Perl 6 is not instantaneous. though it's quicker to learn with today's documentation than with last year's.
tadzik well, is it? 13:53
masak learning to write *idiomatic* Perl 6, which one would perhaps expect from a project like this, takes more than a summer :)
...but then again, "good start" and all that.
with TDD and refactoring, most projects can be lifted from the limitations of an initial design and faltering first steps of implementation.
leont Well, it's important to get foundations right for a project like this before people depend on it 13:54
BioPerl is scary Perl 5 from the mid-late 90s
Also, BioPerl as a whole is more than 2000 modules. You can't reimplement that with a dozen GSOC students. It would be fairly minimal in that regard. 14:00
masak right. 14:11
also, there's a bit of prior art, as far as I know.
pyrimidine hangs out here quite a bit. I seem to recall he wrote some bioperl6 code already. 14:12
probably thought about the design a bit, too.
arnsholt IIRC pyrimidine has a fair bit of bioperl6 code already 14:18
leont From a quick look, that code is mostly perl5 translated literally to perl6 14:21
Juerd [Coke]: Past week, really? We know about network issues the past 2 days.
tadzik yay, LWP::Simple now installs on nom 15:16
masak this should make lichtkind happy. 15:20
moritz phenny: tell lichtkind that LWP::Simple now works again on newest rakudo
phenny moritz: I'll pass that on when lichtkind is around.
tadzik (is panda-installable too)
masak makes ready to get on a train 15:24
[Coke] Juerd: perhaps I am projecting. 15:28
lichtkind cheers 16:13
phenny lichtkind: 15:20Z <moritz> tell lichtkind that LWP::Simple now works again on newest rakudo
lichtkind phenny: tell moritz thanks 16:14
phenny lichtkind: I'll pass that on when moritz is around.
moritz . 16:20
phenny moritz: 16:14Z <lichtkind> tell moritz thanks
moritz it's a real nuisance to write out all the comparison ops for rats 16:22
they all go like this: multi infix:<op>(Rational:D $a, Rational:D $b) { $a.numerator * $b.denominator op $b.numerator * $a.denominator } 16:23
moritz then another (Rational:D, Int:D) and a (Int:D, Rational:D) variant 16:23
moritz and I have to write those for <, <=, >, >=, <=> 16:23
I guess that could be done with macros 16:24
lichtkind moritz: its a module i have to install? kwp doesnt seems to be incore 16:31
lwp
but i recompile rakudo anyway :)
moritz lichtkind: yes, needs to be installed. panda can help you.
masak hello-on-a-train, #perl6! 16:34
tadzik hello masak
masak , with totally undiminished enthusiasm for IRC-ing in a moving vehicle
moritz: I was thinking "macros" also, when I read your complaint above. I think the interesting question, though, is: can it be done with *AST* macros? 16:35
and if not, which primitives are we lacking that would allow that? 16:36
moritz we'd need to be able to add holes into identifiers
but a textual macro would work in that simple case
masak sure it would. I'm just looking for ways to extend the niche of AST macros at the expense of textual macros. 16:38
lichtkind can you tell me about that failong subtest about loading?
i think i will also ask so question about postcircumfix and alike, while this is easy there are other cat i not quite understand 16:42
moritz ENOCONTEXT 16:43
which failing subtest about loading?
lichtkind i ran make test on raudo 16:47
and got t/02-embed/01-load.t ............... Failed 1/1 subtests 16:48
Araq what's perl6's way of providing genSym?
lichtkind what is gensym?
Araq a way to introduce a unique symbol for hygenic macros 16:49
lichtkind and by symbol you mean a operator? 16:53
Araq by symbol I mean symbol
identifier 16:54
lichtkind we have hygenic macros 16:55
Araq good
how do they work?
lichtkind strangelyconsistent.org/blog/macros...hey-really
a short overview
masak that blog post doesn't really address gensymming, though. 17:03
lichtkind yes but it explains that macros produce an ast 17:04
masak Araq: the short and slightly unsatisfactory answer is that careful separation of lexical scopes replaces gensymming in Perl 6.
lichtkind and this gets weaved into the program ast by compiler
masak at least by our current best understanding.
lichtkind: right, but there are lots of subtleties, starting with the fact that an AST must have a "context" much like a closure does. 17:05
Araq: at the end of masak.org/carl/gpw-2012-macros/talk.pdf there is an elucidation about how the lexical scopes conspire to make gensymming unnecessary. 17:06
unfortunately, there is no audio track with the PDF :) please ask more if things are not clear. 17:07
moritz olvvvvvv 17:19
-
masak o/ Ronja :) 17:20
moritz I wonder why the wikipedia page on hygienic macros doesn't mention careful scoping as a solution
\o
masak that struck me as odd too.
maybe it won't work :P
or maybe Wikipedia has, I dunno, some sort of omission. 17:21
moritz if it works, it seems indeed to be the cleanest solution 17:23
masak aye. 17:25
when I get to the point where it works in Rakudo, I might add the alternative to the list in the article ;)
Araq what is the alternative, masak? 17:39
masak well, let's make an example to make things concrete. 17:44
Araq careful scoping is no substitute for gensym
masak macro foo($x) { quasi { my $a = 1; say $a; say {{{$x}}} } }; my $a = 2; foo($a) 17:45
there are two variables called '$a' -- the one in the quasi block, and the one in the mainline.
generally, we'd need to gensym the one in the quasi block so that it won't collide with the one in the mainline. 17:46
but not in Perl 6.
Araq yeah but what if the macro wants to inject a new variable that is destructed at scope end (C++ like RAII)? 17:47
masak we generally handle destruction at scope end with a LEAVE phaser.
though it's still a little unclear to me which block the LEAVE phaser latches on to -- the quasi block, or the block in the mainline. 17:48
seems to me we would want it to latch on to the block in the mainline at least sometimes. this would imply that phasers in quasi blocks are "deferred" until macro application time, just like type declarations etc. 17:49
as to variable injection in general (with or without RAII), see the discuission on the COMPILING:: namespace in S06. 17:52
Araq link? 17:53
masak COMPILING:: grants you a kind of "un-hygiene", even for declarations if you want.
perlcabal.org/syn/S06.html#Macros
it's just that hygiene is switched on by default, because we think it's a nice default :) 17:54
lichtkind tadzik: while isntalling panda i got use of uninitialized value of type Command in string context 17:55
tadzik lichtkind: yeah, I know. I have no idea where that comes from :)
masak and gosh, I sure hope that careful scoping *is* a substitute for gensymming. :) but maybe we were talking past each other, because there certainly are ways in which it wouldn't be in other languages. in Perl 6, it is.
lichtkind tadzik: also the failing test in the fetcher? 17:56
tadzik lichtkind: if you look closely, there's no failing test
just a git error message that I don't know how to silence
Araq *shrug* I like gensym and dirty macros and scope injection 17:58
Teratogen dirty macros? 17:59
Araq un-hygienic macros 18:00
lichtkind tadzik: allright but i have to restart to isntall modules?
masak Teratogen: see en.wikipedia.org/wiki/Hygienic_macro for details. 18:01
Teratogen: the distinction doesn't really exist in C, though ;)
Araq brb
masak (I'm not aware of any gensymming in C. all C macros that I recall have been unhygienic.) 18:02
lichtkind tadzik: yes your right bit now i got: src/gen/Metamodel.pm 18:03
i mean Incorrect pre-compiled version of src/gen/Metamodel.pm loaded
tadzik (in a cross-platform way, that is) 18:04
lichtkind: oh, right
lichtkind: find ~/.perl6 -name '*.pir' -delete 18:05
there's something seriously wrong about module precompilation in rakudo
masak there is? please go on. 18:09
moritz it leads to weird errors and segfaults when loading some of the modules 18:10
masak yes, but what is it that's wrong? 18:12
and what RT ticket is this and how come I missed it? :) 18:13
lichtkind tadzik: thanks a lot it works now soemthing without error 18:14
lord and behold i might even write a perl 6 module :)
masak it's 'lo and behold' -- the 'lo' is an old shortening of 'look' ;) 18:16
lichtkind moritz: there is still failing tests it escape.t 18:17
moritz well, lichtkind has an idiosyncratic interpretation of every language I know :-)
lichtkind: which module?
lichtkind LWP::Simple 18:18
the one we talked about
moritz lichtkind: what revision are you on?
lichtkind moritz: you mena rakudo version? 18:19
moritz lichtkind: no, LWP::Simple
lichtkind moritz: i think panda fetches the newest out of git? 18:20
masak making it, um, anaphora, I think.
(because both "lo" and "behold" eĉpress the same idea)
express*
moritz lichtkind: I just cloned URI newly from git, and didn't get any test failurs 18:22
sorear good * #perl6 18:26
moritz good *, sorear 18:26
lichtkind moritz: how did you install ?
moritz lichtkind: I didn't install it. I cloned it from git, and ran 'make test'
previously I installed the prerequisites and removed all .pir files from ~/.perl6/lib 18:27
masak sorear! \o/ 18:28
sorear: how's C<handles> going?
lichtkind moritz: whats the secret of that pir removing? 18:29
i almost ran outa tuit for today
Teratogen when Araq said "dirty macros" I immediately thought "macro porn!"
lichtkind Teratogen: cs wise it is
moritz lichtkind: precompiling breaks stuff in weird ways :( 18:31
at least right now
sorear masak! \o/ 18:33
masak: I looked at Rakudo's implementation, that's about all I've done so far
sorear would kind of like to get niecza back over 100% this week... but that's probably not realistic.
Teratogen when Araq said "dirty macros" I immediately thought "macro porn!=jyukih 18:34
Araq "macro porn"? omg
Teratogen \,m. `17,6vg888bbbbbbbbbbbbbbbbbbbc6666666666666bbc6f456RFDRT6/?
masak sorear: ok.
Teratogen SORRY ABOUT THAT 18:35
err
sorry about that
ferret on keyboard
little varmint
masak Teratogen: I'm going to have to ask you to calm down. :/
sorear Teratogen++ # novel excuse
Teratogen it was my pet ferret Archimedes, masak!@ 18:36
he likes to get up on the keyboard and just kind of sit on it
benabik Teratogen: The many lines of apology is more irritating than the one line of random text.
Teratogen mashing keys as he does
sorear masak: Yesterday Teratogen was saying that when ey gets bored, ey makes noise on #perl6 to entertain emself
Teratogen sorear, hush
masak sorear: aye, I saw that.
tadzik masak: (precompilation) if we know what's wrong, it probably wouldn't be already :)
masak: but I know this since GPW. Loading precompiled SVG::Plot (for example) segfaults
masak Teratogen: we love to see you make progress as someone who contributes positively to the channel. right now you're not showing signs of that. 18:37
tadzik: I wasn't really aware until now.
Teratogen I'll try to do better masak
masak tadzik: it sounds familiar, but I haven't tried to precompile modules in a while, and so I haven't been affected. 18:38
masak hugs Teratogen
lichtkind moritz: ok i lay the problem down for now just want to finish grammar categories and special var 18:41
brrt in rakudo, is there a branch that does implement declared array dimensions? 18:54
moritz no 18:55
tadzik did you tackle the freenode riddle today? 19:03
masak heh. I thought I needed a dynamic variable in a module, but it turned out I just needed a lexical variable in that module. I wonder what made me confused in that particular way.
seems to me now that dynamic variables are mostly useful in situations that resemble traversal, or recursion. 19:04
masak de-trains 19:06
moritz masak: agreed about recursion and dynamic vars 19:09
or what STD.pm6 and rakudo's Grammar.pm use is to do some tiny bits of configuration on the calling location for stuff that goes deep into the call stack and that isn't worth passing around as objects all the time 19:11
lichtkind masak: whos idea was it to lowercase pod? 19:18
tadzik Damian's 19:19
I think
lichtkind and you know the reason?
damian is still around?
sorear damian is impossible to get rid of! 19:20
lichtkind here it is a good thing :)
Araq implicit parameters seem a better solution than dynamic vars
sorear Araq ! 19:21
tadzik I believe they serve different purposes
sorear Hi ! Haven't seen you fo a while
Araq well thanks, hi
lichtkind tadzik: only reason for me was it just looks better ebcasue special var have to stick out more and if pod do it too there is tto much uppercase and since is got even more part of the lang it should be lowercase like rest of keywords 19:22
tadzik damn, feather is completely unusable these days 19:35
sorear Juerd: is your company still being ddos'ed? 19:38
spider-mario woah, macros in rakudo? 19:41
perl6: macro hello($name) {quasi {say "Hello { {{{$name}}} }!"}}; hello "Perl" 19:42
p6eval niecza v16-17-g648c49b: OUTPUT«Use of uninitialized value in string context␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1226 (warn @ 3) ␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 234 (Mu.Str @ 10) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/boot/…
..pugs: OUTPUT«*** Undeclared variable: ("$name",MkPad (padToList [("$_",PELexical {pe_type = (mkType "Scalar"), pe_proto = <Scalar:0x7f46576b5b59>, pe_flags = MkEntryFlags {ef_isContext = True}, pe_store = <ref:0x7f46576b5091>}),("@_",PELexical {pe_type = (mkType "Array"), pe_proto = <Ar…
..rakudo 3218a7: OUTPUT«Hello AST<3843861497533412442>!␤»
spider-mario so, $name is an AST
that makes sense 19:43
sorear o/ spider-mario 19:43
Araq perl6: macro hello($name) {quasi {hello "Hello { {{{$name}}} }!"}}; hello "Perl" 19:44
p6eval niecza v16-17-g648c49b: OUTPUT«Use of uninitialized value in string context␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1226 (warn @ 3) ␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 234 (Mu.Str @ 10) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /home/p6eval/niecza/boot/…
..rakudo 3218a7: ( no output )
..pugs: OUTPUT«pugs: *** Cannot bind to non-existing variable: "$name"␤ at /tmp/xHkOsrr4f1 line 1, column 1␤»
spider-mario macros can’t call macros yet? 19:45
or is it just recursion that’s not handled?
dalek kudo/nom: 984dc7c | moritz++ | src/core/Rat.pm:
comperator ops for rats and rats and ints
spider-mario perl6: macro test1 {quasi {say "test1"}}; macro test2 {test1}; test2 19:46
p6eval niecza v16-17-g648c49b: OUTPUT«===SORRY!===␤␤Action method quasiquibble not yet implemented at /tmp/DgvqYWCRV2 line 1:␤------> macro test1 {quasi {say "test1"}⏏}; macro test2 {test1}; test2␤␤Action method quote:quasi not yet implemented at /tmp/DgvqY…
..rakudo 3218a7: OUTPUT«===SORRY!===␤Could not find sub &say␤»
..pugs: OUTPUT«test1␤*** ␤ Unexpected end of input␤ expecting bare or pointy block construct, ":" or identifier␤ Macro did not return an AST, a Str or a Code!␤ at /tmp/c9bzkAvbPm line 2, column 1␤»
spider-mario Could not find sub &say
???
tadzik spider-mario: I don't think quasi is there yet
spider-mario oh
quasi did work on a more complex example 19:47
perl6: macro test1 {quasi {say "test1"}}; macro test2 {quasi {test1}}; test2
moritz t/spec/S06-macros/macros-d1.t has the macro tests that rakudo currently passes
p6eval pugs, rakudo 3218a7: OUTPUT«test1␤»
..niecza v16-17-g648c49b: OUTPUT«===SORRY!===␤␤Action method quasiquibble not yet implemented at /tmp/3BgyLngsBN line 1:␤------> macro test1 {quasi {say "test1"}⏏}; macro test2 {quasi {test1}}; test2␤␤Action method quote:quasi not yet implemented at /t…
moritz anything that's not in there is likely to fail
spider-mario: if you start your queries with 'rakudo:' instead of 'perl6:
spider-mario yes, sorry
moritz ... instead of 'perl6:' you can avoid a lot of extra output 19:48
spider-mario I always forget then realize
moritz :-)
r: macro a() { quasy { say "foo" } }; a
p6eval rakudo 3218a7: OUTPUT«===SORRY!===␤Could not find sub &quasy␤»
moritz r: macro a() { quasi { say "foo" } }; a
p6eval rakudo 3218a7: OUTPUT«foo␤»
spider-mario but appearently, pugs is able to run the code too
spider-mario oh, I get it 19:49
moritz r: macro a() { quasi($x) { say "foo"; $x } }; a say 'foo'
p6eval rakudo 3218a7: OUTPUT«===SORRY!===␤Variable $x is not declared␤at /tmp/kBBUtSXrGj:1␤»
spider-mario say was executed at build time
which explains why it was not found
moritz but it should be available at compile time
r: BEGIN say 'foo'
p6eval rakudo 3218a7: OUTPUT«foo␤»
spider-mario oh, indeed
macro test {say "hello"}; test
oops
r: macro test {say "hello"}; test 19:50
p6eval rakudo 3218a7: OUTPUT«hello␤===SORRY!===␤Macro did not return AST at line 1, near ""␤»
spider-mario r: macro test {say "hello"; quasi {}}; test
p6eval rakudo 3218a7: OUTPUT«hello␤»
spider-mario hm…
r: macro test1 {say "hello"; quasi {}}; macro test2 {test1}; test2
p6eval rakudo 3218a7: OUTPUT«hello␤===SORRY!===␤Macro did not return AST at line 1, near ""␤»
spider-mario r: macro test1 {say "hello"; quasi {}}; macro test2 {test1()}; test2 19:51
p6eval rakudo 3218a7: OUTPUT«hello␤===SORRY!===␤Macro did not return AST at line 1, near ""␤»
moritz r: macro a() { quasi { 1 +1 } }; say a
p6eval rakudo 3218a7: OUTPUT«2␤»
spider-mario what am I missing?
pugs: macro test1 {say "hello"; quasi {}}; macro test2 {test1}; test2 19:52
p6eval pugs: OUTPUT«hello␤»
spider-mario pugs gets it right
and surprisingly quickly :D
timotimo huh, i can do "next" in a function that has no loop in it, if it's called from some place that has a loop?
i could just try instead of asking, actually 19:55
tadzik :) 19:55
timotimo that's pretty neat 19:56
i suppose the same mechanism makes "take" work in any function called transitively from inside a gather block
interestingly, next without a loop around it is not a problem to rakudo 19:57
p6: next
p6eval pugs: OUTPUT«*** Cannot use this control structure outside a 'loop' structure␤ at /tmp/YLNP69ye1w line 1, column 1 - line 2, column 1␤»
..niecza v16-17-g648c49b: OUTPUT«Unhandled exception: Illegal control operator: next␤ at /tmp/Bl9JeABqqZ line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3839 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3840 (module-CORE @ 65) ␤ at /home/p6eval/niecza…
..rakudo 3218a7: OUTPUT«next without loop construct␤ in block <anon> at src/gen/CORE.setting:303␤ in block <anon> at src/gen/CORE.setting:366␤ in block <anon> at /tmp/JaHkEfxacA:1␤␤»
timotimo p6: sub skip_some($a) { next if $a %% 2 }; skip_some(0); skip_some(1); skip_some(2); say "oh my"; 19:58
p6eval niecza v16-17-g648c49b: OUTPUT«Unhandled exception: Illegal control operator: next␤ at /tmp/2HqxGCY3RP line 1 (skip_some @ 4) ␤ at /tmp/2HqxGCY3RP line 1 (mainline @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3839 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3840…
..pugs: OUTPUT«*** ␤ Unexpected "%% 2"␤ expecting operator or "}"␤ at /tmp/crhaRZEgMA line 1, column 32␤»
..rakudo 3218a7: OUTPUT«next without loop construct␤ in block <anon> at src/gen/CORE.setting:303␤ in block <anon> at src/gen/CORE.setting:366␤ in sub skip_some at /tmp/B8gGQv2uBg:1␤ in block <anon> at /tmp/B8gGQv2uBg:1␤␤»
timotimo p6: sub skip_some($a) { next if $a %% 2 }; for ^9 { skip_some($_); say "oh my, $_"; } 19:59
p6eval rakudo 3218a7, niecza v16-17-g648c49b: OUTPUT«oh my, 1␤oh my, 3␤oh my, 5␤oh my, 7␤»
..pugs: OUTPUT«*** ␤ Unexpected "%% 2"␤ expecting operator or "}"␤ at /tmp/kQrh4NTe2E line 1, column 32␤»
timotimo mhm, mhm.
my rakudo doesn't complain if i just type "next", but i suppose there's a subtle difference between its repl and p6eval that causes this?
tadzik supposedly 20:00
REPL is probably just a loop, so next is likely to just work :)
timotimo oh, haha :D
that makes sense
yup, typing last makes the repl quit 20:01
that's amusing :)
tadzik :) 20:02
I like the feeling when something happens to work exactly the way you imagined it would
timotimo i've got that surprisingly often with perl6 so far
spider-mario I can confirm that 20:04
hopelessly writing some code, running it 20:05
and “what, it actually works?”
timotimo still, the amount of different "stuff" that one has to know to be good at perl6 (probably the same for perl5) is huge when compared to python, which is what i spent most of my time with in the last years
Araq hu? bugs in the REPL are intuitive? 20:06
timotimo python has much less syntax, most things that would have syntax in perl6 would have a function in python instead
spider-mario I guess being fluent at Perl 6 can take you further
timotimo i feel the same way, spider-mario, that's why i want to learn :) 20:07
Araq: it's intuitive that if the repl is just simply a repl, that the control exceptions you'd cause with "next" or "last" would cause the rep-loop to react 20:08
i guess fixing that is a simple matter of adding a CONTROL block in the rep-loop to complain about control exceptions coming from the eval statement without being handled 20:09
lichtkind is there a .add method? 20:11
Araq timotimo: I got that, just sayin'; it *is* a bug, albeit a minor 20:12
tadzik lichtkind: add what to what?
timotimo sure 20:13
dalek osystem: b0b2813 | (Pawel Pabian)++ | META.list:
Added GeoIP database support (currently Mac OS X only).
 20:14
tadzik oh, interesting 20:15
tadzik bbkr: does simple "foo.so" not work on OSX? 20:17
or foo.dynlib
tadzik wonders if there could be a macro for this 20:18
lichtkind tadzik: i mean alias to +
tadzik hmm
r: say 5.&infix:<+>(7)
p6eval rakudo 984dc7: OUTPUT«12␤»
tadzik \o/
bbkr tadzik: the point is that NativeCall cannot fallback to .dylib, it always wants .bundle, so I have to hardcode it until github.com/jnthn/zavolaj/issues/7 is resolved 20:19
tadzik I see 20:20
bbkr tadzik: after that i will remove also hardcoded path and it should work on any OS
tadzik I have a feeling there's a need for a mechanism which will make it easier to resolve things like this in modules
tadzik possibly something along "use Alien; sub foo is native(Alien::lib<foolib>) 20:24
and it will resolve things automagically. Could be doable
bbkr Native Call does it. is native("libGeoIP") works if I recompile library to .bundle on OS X and works also on Linux with .so files. 20:27
spider-mario perl6: my $name-with-hyphens = 42; say $name-with-hyphens 20:29
p6eval pugs: OUTPUT«*** No such subroutine: "&with"␤ at /tmp/DYunYTKIoZ line 1, column 1-27␤»
..rakudo 984dc7, niecza v16-17-g648c49b: OUTPUT«42␤»
tadzik oh, I see
bbkr so NativeCall is only "one issue away" from being truly multiplatform :) 20:30
tadzik I'm afraid the issue lives in nqp
or even in a dyncall library 20:31
bbkr I'm not experienced enough to debug this one on my own. If jnthn will reject issue on NativeCall repo I will report in NQP :) 20:33
tadzik I'm quite sure it's dyncall which lacks this 20:34
Araq what's NQP?
tadzik 3rdparty/dyncall/buildsys/bsdmake/osx.bsdmake
29:DLL_SUFFIX= .dylib
timotimo not quite perl, it's the language in which a lot of rakudo is written
benabik Araq: Not Quite Perl, the subset of Perl 6 that Rakudo is implemented in.
tadzik Araq: Not Quite Perl, a subset of Perl 6 in which the core Rakudo is implemented
oh wait, that's unrelated 20:35
(dynlib)
Araq a static subset of perl6?
benabik Araq: It's not static, the object system is just simpler.
Araq so it's quite unlike PyPy's implementation language? 20:36
tadzik bbkr: does dlopen("foo") Just Work for .bundle in OSX? 20:37
I think it's just dlopen() doing the hard work after all
lichtkind so damian said no reason 20:38
bbkr tadzik: DLL_SUFFIX lies :) NativeCall wants to fetch .bundle by default 20:40
tadzik bbkr: yeah, that DLL_SUFFIX was for building dyncall itself
bbkr: does `man dlopen' on OSX say anything about suffixes?
bbkr: github.com/perl6/nqp/blob/master/3...rwin.c#L54 20:41
it seems that it's just doing dlopen(), no magic
bbkr dlopen only mentions it can "can load dynamic libraries and bundles", nothing about suffixes 20:43
masak moritz: yeah, dynamic vars turn into a kind of couterweight to long parameter lists in recursion/traversal contexts. sort of the same trick as nesting subs and using outer vars, but in time rather than in space ;) 20:48
masak lichtkind: lower-casing pod was part of the whole TheDamian/tadzik/jnthn/pmichaud/masak hallway discussion at YAPC::EU 2011. 20:53
lichtkind: IIRC, Damian seemed to think that a variable like $=pod wasn't extraordinary enough to merit uppercase. 20:54
tadzik I was probably against this :P
If I wasn't I'd probably implement it right away to please Damian :) 20:55
masak well, you didn't prevent TheDamian from making the change then, or me from committing it later ;)
tadzik who am I to stand against you ;) Especially you Both
masak lichtkind: if you look at the commit I linked you to, you'll notice that a few things got lower-cased.
lichtkind yes i did 20:56
masak tadzik: I'll disregard the implication that I somehow posess Damian-like powers :P
lichtkind im mostly woth tadzik in this
but i like the way it looks
it just doesnt feel right syntactically
Araq you know, there is this thing called "case insensitivity"
masak well, both of you have write access to S26, I presume.
tadzik masak: why do you think I *still* haven't turned #= into #|? :P
masak Araq: :D
tadzik: that one *must* be laziness, though, because that's an improvement ;) 20:57
Araq it was a pretty good idea back in the days
lichtkind what is #= ?
tadzik masak: I still don't like this :)
masak lichtkind: buddy, you're in for a treat :)
Araq but then unfortunately unix did win
tadzik masak: but I didn't oppose very loud for I have no better idea
but again, I think having two kind of special comments is too much
masak Araq: I suddenly get the feeling you're not joking.
Araq: could be trolling, though. 20:58
Araq and now people think it's a feature that they need to care about details like that ...
tadzik lichtkind: it's the documenting comment
r: say ''
p6eval rakudo 984dc7: OUTPUT«␤»
tadzik I needed the ␤
Araq indeed, I'm not joking
spider-mario r: say
p6eval rakudo 984dc7: OUTPUT«␤»
Araq likes case insensitivity
lichtkind shit
tadzik r: #= I'm a hoarse! ␤ sub foo {}; say &foo.WHY
p6eval rakudo 984dc7: OUTPUT«Block::Declarator.new(WHEREFORE => sub+{<anon>} foo() { ... }, config => ().hash, content => Array.new("I'm a hoarse!"))␤»
tadzik r: #= I'm a hoarse! ␤ sub foo {}; say &foo.content
p6eval rakudo 984dc7: OUTPUT«Method 'content' not found for invocant of class 'Sub+{<anon>}'␤ in block <anon> at /tmp/bCkcInX1WW:2␤␤»
masak Araq: I would argue that case sensitivity is a feature, yes.
tadzik bleh 20:59
r: #= I'm a hoarse! ␤ sub foo {}; say &foo.WHY.content
p6eval rakudo 984dc7: OUTPUT«I'm a hoarse!␤»
masak spider-mario: that's a known rakudo regression.
std: say
p6eval std 5740094: OUTPUT«===SORRY!===␤Unsupported use of bare 'say'; in Perl 6 please use .say if you meant $_, or use an explicit invocant or argument at /tmp/u1msSXksw0 line 1:␤------> say⏏<EOL>␤Check failed␤FAILED 00:01 107m␤»
benabik ... p6 eval maps ␤ to newline?
spider-mario is it?
tadzik lichtkind: you can document ALL THE THINGS with this
spider-mario it’s actually the behavior I’d expect 21:00
masak spider-mario: yes, it is. see the STD output.
tadzik still. Go lives with just prefix-ish '//'. I think we _could_ live with just one kind of comment
and I'd prefer that one to be #=, like $= in $=pod
once I'll get a good idea, then actually implement it so you can't touch me and re-spec it agin 21:01
ahahaahaha!
Araq MASAK: no, not you, the uppercased one
;-)
tadzik like an evil overlord
masak tadzik: please please please prepare a spec patch, and we can talk over it.
tadzik masak: what I mean (or think), is that we probably don't need two kinds of doc-comments 21:02
but to prove this, I'd have to present a working solution
masak Araq: feels to me in a very short while you've conflated case-insensitivity of (a) file names, (b) variable names, (c) IRC nicks. this does not help your argument.
masak tadzik: all you need to present is a sufficiently clear S26 patch. 21:03
tadzik masak: I'd prefer to be sure that it'll make sense :)
Araq which argument? I have not even given an argument yet, just an example
tadzik so far it's mostly "I think this would be the easiest way" discussion
Araq an argument would be something like "psychology has proven that most if not all people remember the sound of a word, and not its spelling" 21:04
tadzik I'm thinking of having a "look, I got this working and it's the best" ace in my hand
masak tadzik: good. I think the spec shouldn't remain in a state where the guy implementing it feels uncomfortable. 21:05
sorear tadzik: I'm pretty sure th having two kinds of doc comments thing was my idea; I stole it from Haddock 21:06
masak and if it all moves in the direction of increased simplicity (which my changes did too, by the way!), all the better.
tadzik looks up Haddock
masak yeah, I remember the discussion starting around Haddock and its up/down doc comments.
benabik Two kinds of doc comments?
tadzik yeah, #= and #| 21:07
masak benabik: one sitting before a declaration, the other after it.
benabik Oh, how odd.
tadzik Haddock is a food fish.
Captain Haddock, a cartoon character in the The Adventures of Tintin series
duckduckgo is awesome
benabik tadzik: Haddock is a documentation system for Haskell
masak it's also a Haskell documentation standard.
tadzik yeah, I see
sorear Once upon a time, S26 had only one #= comment, and S26 had disambiguation rules that rivalled P5's implicit object syntax in confusingness 21:08
tadzik that's correct
lichtkind ah so its a comment that ends in $=pos?
masak yeah. that's why I said that my changes were a simplification.
lichtkind ah so its a comment that ends in $=pod?
tadzik "you can put #= before, after, or somewhere else and the compiler will be smart"
lichtkind: not really. It ends in the object being documented
see my example above
masak I'm assuming tadzik doesn't want to take us back to disambiguation land ;) 21:09
tadzik r: #= I'm a hoarse! ␤ sub foo {}; say &foo.WHY.content
p6eval rakudo 984dc7: OUTPUT«I'm a hoarse!␤»
tadzik nonono
remember, I'll be the one to get that working
masak that's the beauty of it :)
masak in other news, I don't like feather much right now. 21:16
tadzik yeah, I considered moving to my other shell 21:17
fsergot good night o/ :) 21:22
lichtkind good night 21:36
masak 'night
lichtkind i mean fsergot
i go when work is done :) 21:37
Araq fits your nick ;-)
masak :) 21:38
lichtkind im a machine :)
masak .oO( lichtkindbot ) 21:41
Araq wenn du um 4 noch hier bist, glaube ich dir 21:42
(if you're around still at 4am, I'll believe you)
tadzik considers marrying net::irc::bot to MuEvent 21:45
lichtkind araq: i slept 7o 8 hours the last 2 nights , now im resting and going early 21:49
what is muevent?
tadzik it's a feelalike of AnyEvent I wrote once
lichtkind beside we have now a online study group 21:50
masak phenny: fr en "meuvent"?
phenny masak: "move" (fr to en, translate.google.com)
tadzik see github.com/tadzik/MuEvent/blob/mas...example.pl
ohh, that makes it a pretty good name
masak r: macro test1 {say "hello"; quasi {}}; macro test2 {test1}; test2 21:52
p6eval rakudo 984dc7: OUTPUT«hello␤===SORRY!===␤Macro did not return AST at line 1, near ""␤»
masak huh.
spider-mario: thanks for finding this. I'll add it to the test suite, and find out what the underlying cause might be. 21:53
lichtkind looks like we can switch $~MAIN in #perl6 to german 21:55
benabik It won't be that less comprehensible for most people. :-) 21:56
masak gute Nacht, #perl6
Araq yeah and that's quite surprising 21:57
since you can't code in perl with a german keyboard layout 21:58
lichtkind gut anch mäsak :) 22:00
nacht
i really hardly get around the name since masak means trickster in czech 22:01
lichtkind but its a fishy slang word reaching from darling (mazan) into other meanings 22:02
*
sorear lichtkind: studying what? 22:08
lichtkind esoteric stuff :) 22:12
its a bit wannabe but hey you have to start where you are right :) 22:14
Juerd I've replaced a network cable that we suspect to be faulty 22:46
Hopefully this solves the connectivity issues
lichtkind is the cat statement_control for things like $a if ...; ? 22:55
sorear no, that's statement_mod_cond 23:00
statement_control is for prefix forms like 'for ... { ... }'
lichtkind thank you 23:01
and statement prefix seems to be same as modifier or not? 23:10
Juerd Well, that wasn't it, then :( 23:52
Plugged the cable into another switch port.
Araq good night guys 23:57
lichtkind good night 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!