»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
flussence vim points out a missing \n in the statusbar and it's configurable whether it adds one, don't know about vanilla vi 00:16
segomos what branch of panda do i need for moarvm? 00:28
[Coke] kurahaupo: looks like emacs agrees with ecilpse 00:38
segomos [Coke]: what happens if you do ':set list' 00:40
geekosaur as far as emacs is concerned, newline is just another character. appending after one is perfectly legit 00:41
segomos www.linuxquestions.org/questions/li...le-718461/ 00:45
TimToady masak: smartmatching was introduced in A04 02:33
TimToady so early 2002 02:37
segomos is there a panda branch that works for moarvm or am i installing some of these modules manually for space invaders? 03:25
dwarring segomos: haven't tried space invaders myself yet ... 03:52
but latest panda does seem to be moar friendly 03:53
cd ~/git/panda # or whatever
perl6-m bootstrap.pl # works for me
then to install modules... 03:54
dwarring perl6-m ./panda install ACME::Blah 03:54
segomos everytime i run panda i get 'moar vm not supported' 03:57
i tried git reset --hard HEAD; git pull; #rebuild commands here
and still get the error
i've also tried blowing away the folder and re-cloning the git repo to no avail
dwarring hmm 03:59
are you able to bootstrap? 04:05
using either bootstrap.pl or rebootstrap.pl
segomos: ^^ 04:06
segomos it errors out saying moarvm isn't supported 04:09
dwarring it does like an old panda version somehow 04:16
segomos do i need to build it with jvm perl6 04:17
?
{tony@boondocks:~/Downloads/panda}% git reset --hard HEAD 04:18
HEAD is now at 05af361 Merge pull request #81 from kanatohodets/master
that is what i have from a git reset
dwarring same as mine 04:19
segomos {tony@boondocks:~/Downloads/panda}% perl6 bootstrap.pl
==> Bootstrapping Panda
==> Fetching File::Find
==> Building File::Find
moar is an unsuppored backend VM.
dwarring I've just been using: perl6-m ./rebootstrap.pl 04:20
...or perl6-j ./rebootstrap.pl # java 04:25
PerlJam not that I think it'll solve this particular problem, but you might want to use rakudobrew: github.com/tadzik/rakudobrew 04:28
segomos i'll give it a shot, reboostratp didn't work either 04:29
segomos PerlJam: does that need to be a sudo 'rakudobrew build' i fail make install 04:37
PerlJam What OS are you on? It should try to install in $HOME/.rakudobrew if you're not root
segomos osx 04:38
i'm not seeing a log file 04:39
PerlJam huh, dunno. So far I've been using it on various linux with success
segomos i'll bug timotimo tomorrow 04:40
make: write error 04:42
Failed running make install at /Users/tony/.rakudobrew/bin/rakudobrew line 23.
FROGGS_ segomos: search for "write error" in the irc logs, there seems to be a bug in (g)make on osx 05:45
lizmat fwiw, I have been seeing the "make install" error on OSX for a *long* time 06:17
I just run make install again, and it's fine
empirically, it only seems to happen if something in src/Perl6 needs to be compiled 06:18
those are my datapoints
breakfast& 06:20
masak morning, #perl6 06:32
lizmat++ # output bug
TimToady++ # A04 06:33
masak shopping, flight & 06:48
lizmat safe journeys, masak! 07:09
(and Kathy)
moritz still hasn't met Kathy :( 07:10
good morning everybody
kurahaupo Good ugt moritz 07:25
FROGGS morning #perl6 07:28
Ven 'morning, *. 07:31
sergot morning o/ 07:47
dalek ast: 7c27d66 | (Elizabeth Mattijsen)++ | S02-types/ (6 files):
Add tests for (Set|Bag|Mix)(|Hash).fmt
 07:48
lizmat morning everybody!
tadzik good morning 07:56
lizmat tadzik o/ 08:06
jnthn morning o/ 08:18
FROGGS jnthn: btw, I updated MoarVM/loop_labels about what you mentioned yesterday 08:21
dalek kudo/nom: a5103c3 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Remove internal references to $*OS

So that we don't generate deprecated warnings when just running core functions		 08:23
lizmat FROGGS jnthn o/
jnthn FROGGS: nice
I'm currently giving my slides another look over for any misteaks, then will get back to labels :) 08:24
lizmat
.oO( misteak = illegally procured tropical wood )
 08:25
FROGGS *g*
xiaomiao mi-steak? 08:28
moritz mis-teak 08:29
dalek ast: 5284455 | moritz++ | S32-exceptions/misc.t:
unfudge tests for RT 121534
 08:34
moritz m: try eval 'my class A {}; (-> &c, $m { A.new()(); CATCH { default { $m } } } )(A, "")'; say $!.perl 08:35
camelia rakudo-moar bc585f: OUTPUT«X::TypeCheck::Binding.new(symbol => "\&c", operation => "binding", got => A, expected => Callable)␤Saw 1 call to deprecated code during execution.␤================================================================================␤Sub eval (from GLOBAL)…»
moritz lizmat: there's something wonky with subtests on the JVM 08:38
paste forthcoming 08:39
lizmat: perlpunks.de/paste/show/537c6684.362a.1b5 08:40
lizmat moritz: fwiw, it more indicates a problem with POST than with the test itsels 08:53
*itself 08:54
and because it gets executed twice, the subtest fails because the number of tests is different from the (sub)plan 08:55
moritz but isn't it throws_like that executes the test twice? 08:57
j: my sub a { POST 0 }; a() 08:58
camelia rakudo-jvm bc585f: OUTPUT«(timeout)»
lizmat is firing up a perl6-j
moritz huh, the perl6-j repl prints Postcondition '0' failed twice 08:59
lizmat so the test failure is correct :-)
moritz lizmat: I still don't understand how throws_like $anything, ...; can make the subtests execute twice 09:00
sergot p: my $str; LEAVE { $str ~= '1' }; LEAVE { $str ~= '2'; die 'foo' } 09:02
camelia rakudo-parrot bc585f: OUTPUT«foo␤ in block at /tmp/Iz3KgdKDPp:1␤␤»
sergot p: my $str; try { LEAVE { $str ~= '1' }; LEAVE { $str ~= '2'; die 'foo' } }; say $str; 09:03
camelia rakudo-parrot bc585f: OUTPUT«21␤»
sergot m: my $str; try { LEAVE { $str ~= '1' }; LEAVE { $str ~= '2'; die 'foo' } }; say $str;
camelia rakudo-moar bc585f: OUTPUT«2␤»
dalek kudo/nom: d96ad15 | (Elizabeth Mattijsen)++ | src/core/Exception.pm:
X::Assignment::RO thrown correctly on JVM/Parrot
 09:10
lizmat moritz: found other problem and fixed ^^^ 09:11
FROGGS lizmat++ 09:16
sergot lizmat++ 09:17
lizmat moritz: looks like the code is only EVALled once, but throwing twice 09:26
moritz: suggest to rakudobug it and fudge for jvm 09:29
jnthn FROGGS: Thanks for fixes to MoarVM labels branch 09:31
FROGGS jnthn: well, of course
jnthn Got two functions I want to more carefully review for GC issues, then will be content :) 09:32
FROGGS k :o) 09:33
lizmat moritz: some more datapoints: it's two different exception objects being thrown
of the same type and content, but with different WHICH
jnthn FROGGS: yeah, it's fine 09:34
So, that's MoarVM bit reviewed :) 09:35
lizmat $ perl6 -e 'Backtrace.new.DUMP'
java.lang.StackOverflowError
FROGGS jnthn: I guess I should also rename the block_label to label_local in MoarVM/lib/MAST/Nodes.nqp:L599
jnthn FROGGS: Preferably 09:36
FROGGS done 09:37
jnthn In NQP, for the Parrot support: 09:40
512;513;514
those magic numbers are used in a couple of places. May want to just shove them in lexicals at the top of the file 09:41
jnthn ohh 09:42
a later commit did it :)
FROGGS jnthn: why don't you just compare against master? 09:43
FROGGS there is a lot of movement in the commits 09:43
jnthn FROGGS: I prefer to review commit by commit
FROGGS k
jnthn Yeah, I'm noticing...
FROGGS (that is why I wanna squish them)
jnthn A rebasing/cleaning up mighta been nice pre-review ;)
FROGGS next time, promised :o) 09:44
lol, "next time"
lizmat moritz: more datapoints: looks like the "exit_handler" is run twice, the number of POST blocks is 1 09:46
so the POST block did not get push twice to the phasers
*pushed
lizmat marking as rakudobug now 09:50
jnthn FROGGS: Anyway, don't see anything in NQP patches that worry me 09:54
FROGGS \ö/ 09:55
lizmat it's in #121935
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=121935
dalek ast: b1be572 | (Elizabeth Mattijsen)++ | S32-exceptions/misc.t:
Fudge POST throws_like test for #121935
 09:57
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=121935
sergot m: END { say 1 } 09:58
camelia rakudo-moar a5103c: OUTPUT«1␤»
sergot m: LEAVE { say 1 }
camelia rakudo-moar a5103c: OUTPUT«1␤»
sergot m: my $str; try { END { $str ~= '1' }; END { $str ~= '2'; die 'foo' } }; say $str;
camelia rakudo-moar a5103c: OUTPUT«(signal )(Any)␤»
sergot p: my $str; try { END { $str ~= '1' }; END { $str ~= '2'; die 'foo' } }; say $str; 09:59
camelia rakudo-parrot a5103c: OUTPUT«(Any)␤===SORRY!===␤foo␤foo␤current instr.: 'print_exception' pc 140304 (src/gen/p-CORE.setting.pir:59595) (gen/parrot/CORE.setting:11382)␤called from Sub 'throw' pc 413908 (src/gen/p-CORE.setting.pir:173269) (gen/parrot/CORE.setting:11238)␤ca…» 10:00
jnthn FROGGS: In Rakudo, are we really sticking the match object into the Label object?
FROGGS jnthn: do we O.o?
I'll have a look
jnthn + has Mu $!match;
And it's being used like on in gist 10:01
sergot m: { LEAVE { say 11 }; END { say 12 } } ; { LEAVE { say 2 } }
camelia rakudo-moar a5103c: OUTPUT«11␤2␤12␤»
FROGGS jnthn: look at rakudo/src/Perl6/Grammar.nqp:1108: token label {
and:
/home/froggs/dev/rakudo/src/Perl6/Actions.nqp:1142: if $*LABEL {
/home/froggs/dev/rakudo/src/Perl6/Actions.nqp:1143: $past.push(QAST::WVal.new( :value($*W.find_symbol([$*LABEL])), :named('label') ));
jnthn: but yeah, it keeps a match object, to get line numbers etc 10:02
to print that: 10:03
rakudo$ perl6-m -e 'FOO: while 0 {}; say FOO'
Label<FOO>(at -e:1, '⏏FOO: while 0 {}; say FOO')
jnthn FROGGS: Right; that's going to cause the whole Match tree beneath it to be serialized. :(
FROGGS eww
jnthn And match objects point to QAST
So all of the QAST tree too
jnthn Well, beneath that point anyway 10:04
FROGGS okay... so I should pull the information that I need out of it in token label
jnthn So, think that needs to change
Just pass in the surrounding strings to the constructor?
FROGGS yeah
jnthn Also, I think we should have src/core/Label.pm
FROGGS jnthn: yes
jnthn Rather than it going in PseudoStash :)
FROGGS I think I said that I'll do that when that what I did is not too off :o) 10:05
bah, what a crappy sentence :/
jnthn lol
FROGGS .oO( lol: redo lol ) 10:06
jnthn FROGGS: In MapIter, where the $nextiter is created, you forget to pass on the label 10:09
my $nextiter := nqp::create(self).BUILD($!listiter, $!block, $!flattens);
jnthn Meaning that once we've nommed the first chunk of the list, the map loses its label... 10:10
Probably can be tested using a for on a gather/take
FROGGS ahh!
okay, I'll add tests and the fix
jnthn \o/ 10:11
That's all I have. :)
FROGGS O.o
that means \o/?
yay :D
jnthn Well, means I'm happy enough to merge it once the things I mentioned are fixed. 10:12
FROGGS sure sure
jnthn Up to you as relesae manager if you want to do it the day before the release ;)
FROGGS will do that today
I will run star tests on that branch on all three backends beside spectesting it again, and will merge then if nothing pops up 10:13
jnthn Excellent. 10:13
FROGGS++
FROGGS is happy 10:14
lizmat m: END die
camelia rakudo-moar a5103c: OUTPUT«(signal )»
lizmat r: END die 10:15
camelia rakudo-jvm a5103c: OUTPUT«(timeout)»
..rakudo-moar a5103c: OUTPUT«(signal )»
..rakudo-parrot a5103c: OUTPUT«===SORRY!===␤␤No exception handler and no message␤␤current instr.: 'print_exception' pc 140304 (src/gen/p-CORE.setting.pir:59595) (gen/parrot/CORE.setting:11382)␤called from Sub 'throw' pc 413908 (src/gen/p-CORE.setting.pir:173269) (gen/parrot/…»
jnthn ooh, a SEGV
lizmat yup
apparently you're not allowed to die in the end
jnthn :P
lizmat it causes a rift in the space / time continuum 10:16
sergot++ for finding that one
as a comparison: $ perl -e 'END {die}'
Died at -e line 1.
END failed--call queue aborted.
jnthn m: END die; exit 1 10:17
camelia rakudo-moar a5103c: OUTPUT«␤ in block at /tmp/XDwdzgcuE0:1␤␤Unhandled exception: ␤ at <unknown>:1 (/home/p6eval/rakudo-inst-1/languages/perl6/runtime/CORE.setting.moarvm:print_exception:4294967295)␤ from src/gen/m-CORE.setting:11459 (/home/p6eval/rakudo-inst-1/langua…»
sergot :) 10:18
btw. I'm working on #121530 - could you give me some clues?
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=121530
lizmat S04:1406 10:20
synopsebot Link: perlcabal.org/syn/S04.html#line_1406
sergot $phasers are added correctly - I checked this. So it's something wrong with handling exceptions ( FROGGS++ )
He also suggested to delay that "die" after all LEAVE phasers. Where should we do this?
lizmat so parrot gets it right (the comment is a bit unclear) 10:21
jnthn yeah.
sergot someone? :) 10:26
jnthn It segfaults while checking a guard 10:26
'cus a real NULL sneaks in somewhere
That shouldn't happen, but seems we get it now and then. So for the sake of getting the Moar release more robust, I'll throw in checks for it 10:28
lizmat $ MVM_SPESH_DISABLE=1 perl6 -e 'END die'
===SORRY!===
Unhandled exception:
indeed, disabling SPESH fixes the pb
well, the segfault :-) 10:29
jnthn yeah
Well, it's easy to make spesh guards robust enough to survive it. 10:30
I'll do that for now because I think this won't be the only case we see this issue.
And I'd rather we don't have a segfaulty release
yeah, tht does is 10:32
*that, it
lizmat jnthn: the following sometimes works, sometimes segfaults and sometimes gives: 10:35
await do for ^1000 { start { 1 for ^10000; .print } }
When invoking max, Provided outer frame 0x7fac74606660 (MVMStaticFrame max) does not match expected static frame type 0x7fac74600650 (MVMStaticFrame ) 10:36
jnthn lizmat: Yeah, that error looks familiar; I've got other examples that produce it.
lizmat lunch& 10:37
jnthn Pushed a Moar patch for the sergot segfault 10:38
moritz "the sergot segfault" 10:47
moritz likes it
named bugs
sergot ;D
moritz much like the "double bacon bug"
tadzik oh yes :D 10:48
and a snowman comet 10:49
mr-foobar hey all, any idea about how peg's are different from perl6 rules ? 11:02
FROGGS mr-foobar: since I never played with PEGs, I can't tell 11:18
lizmat jnthn is on! 11:49
lizmat Learning by Porting Perl 5 modules to Perl 6 11:50
(first edition)
FROGGS hmmm??
that is a new talk :o) 11:51
(and a very awesome one most likely, in several ways)
tadzik "first edition" made me think it's a book :D
lizmat (first showing) then :-) 11:52
it's about CSS::Tiny
remove "use strict"
s/package/class/
FROGGS .oO( done ) 11:53
lizmat almost :-)
s/sub/method/ 11:54
and self is there :-) 11:55
get rid of unpacking parameters in favour of using a signature 11:57
custom error structure can go 12:01
lizmat try to slurp :-) 12:03
Ven `method ! for_valid_years ( YearlyStatistics : & thing_to_do ) {̀` why do you need to specify the receiver here ? 12:08
lizmat not sure you *need* it, but maybe it is needed to select the right candidate ? 12:09
timotimo o/ 12:16
lizmat timotimo o/ 12:17
timotimo Ven: the proper word for it is "invocant" 12:17
Ven timotimo: good. still, why so, in the role itself ?
timotimo where do i find that code again? 12:18
Ven hum; one of jnthn's slides
timotimo mhm
i seem to recall that example
well, dunno 12:20
maybe it's just there for clarity
FROGGS it is just an alias for 'self', no? 12:21
timotimo not in this case
lizmat sometimes you need to name the invocant
e.g. if you want to use it as the left hand side of a pair :-) 12:22
timotimo mhm
or if you have a method inside of a method and you want to disambiguate which "self" you are refering to
lizmat r: <[a..z]>.say # shouldn't this be syntactic sugar for ("a".."z").say ?
Ven www.jnthn.net/papers/2014-fosdem-p...-today.pdf
camelia rakudo-jvm d96ad1: OUTPUT«(timeout)»
..rakudo-{parrot,moar} d96ad1: OUTPUT«[a..z]␤»
Ven lizmat: seems ambigous 12:23
www.jnthn.net/papers/2014-fosdem-p...-today.pdf slide 52
lizmat r: /<[a..z]>/.WHAT.say # it means that range in a char class 12:25
camelia rakudo-{parrot,jvm,moar} d96ad1: OUTPUT«(Regex)␤»
lizmat r: say "a" ~~ /<[a..z]>/, say "A" ~~ /<[a..z]>/ 12:26
camelia rakudo-{parrot,jvm,moar} d96ad1: OUTPUT«Nil␤｢a｣␤True␤»
lizmat r: say "a" ~~ /<[a..z]>/; say "A" ~~ /<[a..z]>/ 12:27
camelia rakudo-{parrot,jvm,moar} d96ad1: OUTPUT«｢a｣␤␤Nil␤»
FROGGS lizmat: <...> in regexes is not just about <[...]> cclasses..., you can also do <[a..z]+[A..F]-:Lo> 12:29
timotimo <[a..z]> could become syntactic sugar, i don't see any conflicts off-hand ...
lizmat I know :-)
timotimo the < > syntax is supposed to be somewhat magical anyway
FROGGS so when we would make <[...]> equal in regexes and outside, then others might ask about all the other forms of assertions 12:30
lizmat I was just wondering: for symmetry it would make sense
wonder what TimToady's thoughts would be
mj41 Hi. My slides github.com/mj41/Perl6-Analytics-results ... jnthn, lizmat you are my heroes :-). Thanks for your work. 12:31
timotimo ooooh way cool
FROGGS mj41++ 12:32
lizmat mj41: you're welcome and thanks!
mj41++ indeed
Ven mj41++
jnthn mr-foobar: perlcabal.org/syn/S05.html#Overview explains Perl 6 grammars/PEGs 12:34
Sheesh, time went fast in that talk... 12:35
lizmat yes, you took 45 mins
:-)
jnthn Yeah...didn't feel like it at all
lizmat and only did 2/3 of the slides ?
jnthn lizmat: I figured I may not get into the grammar transform.
lizmat indeed .... and maybe that would be better in a separate presentation 12:36
FROGGS slide 15 is very interesting
jnthn lizmat: It probably needs a whole hour to do that too. But it's better to be overprepared than underprepared. :) 12:36
Ven (I get that slides aren't available atm :P ?) 12:36
lizmat indeed... could source material for a book, actually :-) 12:37
jnthn lizmat: ooh, yes 12:37
FROGGS mj41: that feature matrix is impressive! 12:40
mr-foobar jnthn: perl6 has these concepts of infix, prefix, postfix, circumfix, post circumfix ... are these native to peg or are they specific to rules ? 12:41
mj41 FROGGS: there is already link at perl6.org/compilers/features
including Niecza
FROGGS mj41: I know that one, but yours are nicer :o) 12:42
jnthn mr-foobar: Those are protoregexes - which are just another way to write an alternation that interacts well with inheritance.
mj41 FROGGS: I mean "Summary chart" link above "Legend" raw.githubusercontent.com/mj41/Per...odData.png
jnthn They work using the normal longest token and other tie-breaking rules
mr-foobar jnthn: just wondering how token postfix:sym<[ ]> translates to peg/rules 12:44
timotimo that means that the rule postfix has, among others, a rule named "postfix:sym<[ ]>" in which "<sym>" refers to "[ ]" 12:45
FROGGS mj41: ahh, this thingy is not very visible, missed it
timotimo they are meaningful to the EXPR parser, i believe
not really a "native" thing, as the EXPR parser itself is part of perl6's grammar (more exact: inside HLL) 12:46
jnthn mr-foobar: If you think of postfix as an alternation, then postfix:sym<[ ]> is contributing an alternative.
timotimo but it's not expressed in perl6 rules, it's a method that operates on Cursor objects directly
jnthn mr-foobar: There's also an operator precedence parser, but it's just a consumer of these
jnthn mr-foobar: That is, it does a methodcall like .postfix, which then tries all of the alternatives 12:47
timotimo jnthn: does it seem like a worthwhile optimization whatsoever to know that in a submethod, the type of self eqv $?CLASS?
jnthn Better to leave spesh to "know" that... 12:48
timotimo fair enough.
jnthn There's nothing to stop somebody actually grabbing it out of the submethod table and doing evil.
Since spesh is dynamic opt it can survive that. 12:49
timotimo so we just log the value of typeof self and see it's the very same type object 5 times in a row and the trace log thingie specializes for that?
jnthn 4 times, but yes 12:50
timotimo that seems fair :)
jnthn I build this presentation in a rather different way to usual; can find it at github.com/jnthn/css-tiny-presentation 12:52
FROGGS lol, someone put a label in asyncops.pm O.o
timotimo i'll be out renovating again today, maybe i'll have a bit of leftover brain space to crunch some questions 12:53
jnthn: got suggestions what i could be thinking about?
FROGGS git log --reverse -p --grep="^[^[][^m][^e][^t][^a]" <--- hehe
jnthn :P
Ven jnthn++ # no gh-pages yet :P ? 12:54
jnthn Ven: oh, that might make more sense than a PDF 12:55
jnthn wonders how to set that up..
Ven I think github has tutorial with a button
timotimo jnthn: you didn't actually put an index.html into the presentation dir :) 12:56
jnthn oh...
oh yeah
'cus it's a generated file
d'oh
Ven yea
lizmat FROGGS: was it me?
Ven hence the demand for a generated file in gh-pages :)
mj41 jnthn: My parser of github.com/mj41/Git-Repository-Log...aw.pm#L175 is a bit bigger :-). But not Perl 6 yet.
jnthn mj41: Yeah, mine does the bits I needed ;) 12:57
timotimo wow, that tool runs fast
FROGGS lizmat: no idea, just was a bit unexpected for me :o)
lizmat ah, yes
I put that there, so that I can optimiza that when we have labels
this was from the time were labels were already parsed, not too long ago 12:58
jnthn timotimo: Actually added index.html now
timotimo: Yes, couple of seconds for me :)
I'd be happy to see the tool taken and used by others :) 12:59
Ven jnthn: should I dl rakudo & clone the repo to build the index.html, or is it available somewher , 13:11
e?* 13:12
timotimo one of the slides is too tall for my screen :\
well, barely
jnthn Ven: Oh, I pushed the index.html already
just get lates 13:13
But any recent Rakudo should do it
timotimo lots of space above the title, but the code box touches the bottom
jnthn If you want to re-generate
timotimo: yeah...was first time I tried out using something != PowerPoint.
timotimo: Was useful given I wanted to generate the slides from the repo history
timotimo: But I'm not especially sold. :) 13:14
timotimo aye
nice slides :) 13:16
moritz URL?
Ven yay :) thanks jnthn
moritz behind on backloggin
Ven would still love a git checkout -b gh-pages && git push -u origin gh-pages :P 13:17
FROGGS moritz: github.com/jnthn/css-tiny-presentation 13:18
jnthn Ven: Done that :)
Sould be at jnthn.github.io/css-tiny-presentati...esentation but can take 10 mins or so 13:19
Ven yep 13:20
thankies :) 13:21
masak hello from back-in-Sweden :) 13:27
mj41 hello masak And thanks for visiting us. 13:28
FROGGS hi masak :o) 13:29
lizmat jnthn: after signal(SIGINT).tap.close, Control-C should not be caught anymore, should it ? 13:31
or are we missing the "closing" logic on the signal tap still ? 13:32
masak mj41: my pleasure.
jnthn lizmat: Indeed 13:33
I think may be missing in MoarVM *and* the thing to call nqp::cancel in Rakudo.
lizmat ok, that explains
was going to do a lightning talk about using signal to have a loop iteration finish 13:34
jnthn hmm 13:35
lizmat I will avoid the issue
jnthn debating whether to show his slide making from git diff tool in a lightning talk
Ven: Be happy: jnthn.github.io/css-tiny-presentati...sentation/ 13:36
Ven jnthn: hehe, I was actually reading them from a zip, but it's better with css and images :)
timotimo jnthn: it would be cool if you did
lizmat gist.github.com/lizmat/48b2c6e3a246eb9f0e25 13:37
mj41 time perl6 -e 'say $_ for 1..2_000_000;' > /dev/null 13:38
real0m49.057s
time perl -MMoose -e 'print "$_\n" foreach 1..2_000_000' > /dev/null
real0m0.622s
lizmat clearly we have optimisation to do with say 13:39
mj41: what does "time perl -MMoose -e1" say ?
tadzik jnthn: s/distnace/distance/ on slide 19 :) 13:40
mj41 time perl -MMoose -e"say"
real0m0.228s
pmichaud good morning, #perl6
yoleaux 12 Apr 2014 03:02Z <raiph> pmichaud: I've tweaked a rakudo.org link per average++'s comment at irclog.perlgeek.de/perl6/2014-04-12#i_8575582
mj41 time perl6 -e 'say'
real0m0.354s
lizmat pmichaud! 13:41
how are the robots?
moritz taking over the world!
pmichaud we're finally done with the season. :)
FROGGS hi pmichaud
Ven hai :) 13:42
pmichaud at the world championship, the team ended up with 2nd place in Programming, and 6th place in the robot game.
we're.... quite pleased.
jnthn o/ pmichaud :)
lizmat pmichaud++ and team ++
moritz pmichaud: congratulations to your team
jnthn lizmat: Yes, the say is certainly to blame 13:43
pmichaud I'm trying to build rakudo for parrot and getting an icu error. known? Do I need to go read docs? ;-)
masak pmichaud! \o/ 13:44
jnthn pmichaud: First that I've heard of it.
lizmat is icu installed?
jnthn pmichaud: Where are you getting it?
lizmat is it on OS X?
FROGGS pmichaud: add this symlink:
ll /usr/include/unicode
lrwxrwxrwx 1 root root 37 Nov 24 10:06 /usr/include/unicode -> /usr/include/x86_64-linux-gnu/unicode/
pmichaud kubuntu 14.04... afaik libicu is installed
jnthn FROGGS++ # knowing stuff
masak why the heck is that needed?
and what can we do about it?
pmichaud FROGGS: okay... that sounds like solving the wrong problem.
masak +1 13:45
jnthn masak: Use MoarVM :P
FROGGS masak: parrot uses icu-config to locate it, and icu-config tells that there would be a /usr/include/unicode which is not true anymore
instead, parrot should use pkg-config to query libicu details
pmichaud FROGGS: so, it's a debian package bug?
jnthn Ugh, so icu-config gives wrong answers?
sergot m: Bla::Bla.new
camelia rakudo-moar d96ad1: OUTPUT«===SORRY!===␤Could not find symbol '&Bla'␤»
FROGGS no
pmichaud: only pkg-config can give correct results, because the include paths depend on arch, which icu-config can't handle 13:46
timotimo jnthn: i seem to recall seeing that we do a wait for the output file descriptor to be "available" again after printing stuff
sergot I got this:
$ perl6 -e 'Bla::Bla.new'
No such method 'throw' for invocant of type 'Any'
pmichaud FROGGS: so, it's a parrot bug
FROGGS so, parrot should first look for pkg-config, and fall back to icu-config
timotimo i *think* we do that twice for the say, because we print the "\n" separately 13:47
FROGGS pmichaud: correct, and there is already an issue
sergot Ok, I know why,nvm
timotimo lizmat: could you benchmark print "$_\n" instead of say "$_"?
lizmat will do
pmichaud github.com/parrot/parrot/issues/1014 # okay, got it
pmichaud wonders if it's worth tracking that bug down or if he should just ignore Parrot. 13:48
FROGGS timotimo / lizmat: say is 34s here, print is 24s
masak jnthn: :P
lizmat timotimo: print "$_\n" about twice as fast
lizmat goes off to tweak &say
timotimo lizmat: but be careful; if the string is very long, the string concatenation operation may be super expensive, too 13:49
that's worth benchmarking, too.
FROGGS jnthn: that way of sliding is very cool! 13:49
masak pmichaud: I think it's worth it to the extent people (a) still care about Parrot, (b) still depend on Parrot. but those two are unknowns to me.
pmichaud masak: those are unknown to me also :) 13:50
lots of things are unknown to me at the moment... I've been away too long :)
masak pmichaud: it's really nice to have you back.
jnthn lizmat: gist.github.com/jnthn/41dbb7ab7150f6cfd41c
timotimo sadly due to the fact that $*OUT is dynamic, we can't turn a say into a print at compile time when the last bit of the string is constant and just put a \n into the string constant
jnthn lizmat: That gets about 1/3 of the time off
timotimo: This is all rather silly in so far as at the VM level we have a sayfh op 13:51
timotimo: Which I guess we are not using...
timotimo oh, right.
FROGGS hehe
pmichaud: it could take a while to read all the changelogs :o) 13:52
especially because they get longer and longer these days
timotimo jnthn: good idea to fetch the $*OUT only once 13:53
jnthn Grr, the C profiler I use doesn't support redirecting the output to NULL
uh, NUL
FROGGS just don't look at the output :P 13:54
jnthn So, some of it is costly IO, but even with it actually spewing the output to the console for real, 8.76% of the time is spent in MVM_string_replace o.O 13:56
FROGGS huh 13:57
pmichaud FROGGS: oh, I've been keeping up with a good bit of the commit messages.
Stage parse : 52.588 13:58
wow
that's... good.
FROGGS if you would enable CGOTO in the Makefile.in, it would drop a few more seconds
pmichaud moarvm++
cognominal pmichaud, what is a fate edge in a NFA? 13:59
FROGGS I've seen 43s once
pmichaud cognominal: I have no idea.
FROGGS (and without CGOTO I am usually between 48s and 50s)
pmichaud cognominal: I've been absent for almost a year, so don't expect me to have detailed answers right now :)
anyone going to yapc::na, ooc? 14:00
FROGGS me not
jnthn pmichaud: My box at home gets that in around 35s ;)
jnthn refrains from pointing out that pmichaud++ implemented fate edges in NFA :P
cognominal pmichaud: well, you may have it in mind. In github.com/perl6/nqp/blob/master/s...ex/NFA.nqp fate edges seem special cased.
cognominal pmichaud++ anyway 14:01
pmichaud refrains from pointing out that the past year has been... well, the past year. :)
jnthn cognominal: Fate edges mean "we have reached a match point"
cognominal ok
jnthn pmichaud: That and I've probably hacked on NFA sufficiently that memory may not help ;)
cognominal: That is, that particular path through the NFA was a success 14:02
cognominal: May happen many times for a given candidate if there's stuff like quantifiers involved
pmichaud iirc, "fate" just means this is a point where the current NFA could be spliced onto another subsequent one 14:03
colomon pmichaud! \o/ 14:06
[Coke] anyone want to be co-admin on planeteria? 14:07
PerlJam gutentag #perl6
pmichaud! LTNS! 14:08
pmichaud PerlJam: indeed
PerlJam [Coke]: What does "co-admin" mean for the prospective victim^Wvolunteer? 14:08
moritz [Coke]: me
PerlJam: it probably means that he/she can add and remove feeds 14:09
nwc10 good *, Pm 14:13
lizmat looking at a lightning talk about log4perl... now eh... isn't that a Supply ? :-) 14:15
lizmat with stuff added, granted that 14:16
jnthn lizmat: sounds kinda right :) 14:17
[Coke] (coadmin) you add or update feeds when some asks you. 14:21
masak "LTNS" sounds like a network protocol... :) 14:23
[Coke] moritz: now in the club.
PerlJam masak: it is ... kinda
FROGGS sounds oraclish 14:24
PerlJam [Coke]: Sounds low maintenance enough that I too volunteer. :) 14:27
jnthn lizmat: Here's a slightly more radical patch gist.github.com/jnthn/cd062dfa6bccf8f95e2d 14:42
jnthn (Which makes further savings.) 14:44
lizmat will look later, after breaking up the workshop :-) 14:48
jnthn nooooo....don't break up the workshop! :P
lizmat :-) 14:49
masak lizmat broke the workshop? 14:50
maybe you can get a refund...
lizmat haven't done it yet :-) 14:51
cognominal r: grammar A { token TOP { <x> }; proto token x {*}; token x:sym<a> { a }; token x:sym<b> { b }; }; $/ := A.parse("a"); say(?$/); my %h := $/<x>.hash(); say(+%h); for %h.keys { say($_) } 14:54
camelia rakudo-{parrot,jvm,moar} d96ad1: OUTPUT«True␤0␤»
masak m: my %h = foo => 1, bar => 2; say (-<< %h).perl
camelia rakudo-moar d96ad1: OUTPUT«("foo" => -1, "bar" => -2).hash␤»
lizmat Richard Jellinek at the CZPW just announce that PetaMem will convert the Word2Num and Num2Word modules to Perl 6 in the coming months
FROGGS \o/ 14:55
Convert all the things! /o/
lizmat end of CZPW& 14:56
jnthn PetaMem++ \o/ 14:56
masak Richard++ 14:57
PerlJam looks at metacpan.org/source/PETAMEM/Lingua...ord2Num.pm 15:01
jnthn eow &
PerlJam ah, it uses Parse::RecDescent, so they should be used to not-very-fast performance ;-) 15:02
anyway, if they all have a similar structure, then the conversion shouldn't take too long.
dalek kudo/loop_labels_test: 2ede82a | (Tobias Leich)++ | src/core/MapIter.pm:
pass along labels to $nextiter
 15:03
kudo/loop_labels_test: 6ab9ccd | (Tobias Leich)++ | src/ (2 files):
refactor Label to not keep an NQPMatch
PerlJam (heck, I bet if someone had organized a hackathon around it, they probably could have gotten it done immediately after czpw) 15:04
masak probably :)
anyway, when it happens, I think #perl6 will be happy to review the result ;)
PerlJam yep!
I'm half tempted to do the english one myself just 'cuz 15:05
masak PerlJam: I'm sure they wouldn't mind ;) 15:14
cognominal r: grammar A { token TOP { <x> }; proto token x {*}; token x:a { a }; token x:b { b }; }; $/ := A.parse("a"); say(?$/); for %($/) { say($_.key) } 15:16
camelia rakudo-{parrot,jvm,moar} d96ad1: OUTPUT«True␤x␤»
cognominal When matching a proto regex, is there a standard way to know from the match value which regex matched? Here: x:a 15:18
Ulti those PetaMem modules already appear to be doing Perl6esk things cpansearch.perl.org/src/PETAMEM/Lin...ord2Num.pm 15:19
oh thats for exports 15:20
masak dinner & 15:25
cognominal bbl& 15:53
mj41 jnthn: www.ceskypivnifestival.cz/en/czech/prague-2014/ ... sorry for noticing this so late :-( 16:35
FROGGS Exception in thread "Timer-571" java.lang.OutOfMemoryError: GC overhead limit exceeded # :o( 16:49
tadzik what 17:03
:D
"I'm the emulator of infinite memory, I'll throw an OOM for you" 17:04
or is it "GC took too long, so I'll just kill your program so you don't have a prolonged GC pause?" :P 17:05
flussence
.oO( "Halting problem preempted" )
 17:11
FROGGS jnthn: there is a regression in r-m*, LWP::Simple fails t/get-w3-latin1-utf8.t 17:50
FROGGS again some jvm gc weirdness (with a backtrace) gist.github.com/FROGGS/9453c10577b1b729acf2 18:05
FROGGS jnthn: spesh bug that causes the r-m* regression: gist.github.com/FROGGS/9228162347e3b4c03a18 18:24
dalek p: 1783628 | (Tobias Leich)++ | / (11 files):
implement labeled loops and throwing of labels as payload

Labels are lexcially installed as instances of type NQPLabel, and are glued to the exception handlers of 'while', 'until' and 'for' loops. When an exception with a label as payload is thrown, the handler that holds the identical label will be in charge.		 18:52
FROGGS m: for 1 { say &?BLOCK } 18:58
camelia rakudo-moar d96ad1: OUTPUT«===SORRY!=== Error while compiling /tmp/HgIBwtRZNa␤Undeclared routine:␤ &?BLOCK used at line 1␤␤»
dalek kudo/nom: 196b4ff | (Tobias Leich)++ | / (11 files):
implement labeled loops and throwing of labels as payload

Labels are lexcially installed as instances of type Label, and are glued to the exception handlers of e.g. 'while', 'until' and 'for' loops. When an exception with a label as payload is thrown, the handler that holds the identical label will be in charge. Introspecting the labels of a block is NYI, since we do not have &?BLOCK to yet, see S06:3047.		 19:03
synopsebot Link: perlcabal.org/syn/S06.html#line_3047
ast: 3ef752a | (Tobias Leich)++ | S04-statements/ (2 files):
add more tests for labels
 19:08
ast: 31ccb4c | (Tobias Leich)++ | S04-statements/ (2 files):
unfudge now passing label tests
FROGGS k, running jvm spectests now without the evalserver... 19:39
flussence std: / <-cntrl -[ \s \" \% \' \< \> \\ \` \{ \| \} ]> /; # I've ended up writing this in my code... any ideas how I can pretty it up? 19:42
camelia std ec3d853: OUTPUT«===SORRY!===␤Unsupported use of curly brackets; in Perl 6 please use square brackets at /tmp/0BJ8477_Zw line 1:␤------> / <-cntrl -[ \s \" \% \' \< \> \\ \` \{⏏ \| \} ]> /; # I've ended up writing thi␤Parse failed␤FA…»
flussence m: / <-cntrl -[ \s \" \% \' \< \> \\ \` \{ \| \} ]> /; 19:43
camelia ( no output )
flussence (stdbug?)
FROGGS m: say 'abc{' ~~ / <-cntrl -[ \s " % ' < > \\ ` { | } ]>+ /; 19:45
camelia rakudo-moar d96ad1: OUTPUT«｢abc｣␤␤»
FROGGS std: say 'abc{' ~~ / <-cntrl -[ \s " % ' < > \\ ` { | } ]>+ /; 19:46
camelia std ec3d853: OUTPUT«ok 00:01 126m␤»
FROGGS flussence: you don't need to escape them
flussence I feel dumb now :)
FROGGS hehe, you should not :o)
FROGGS but the error message of std feels a bit off anyway 19:47
it should not complain about that in character classes me thinks
flussence std: / <[ \s \" \' \\ '%<>`{|}' ]> / 19:48
camelia std ec3d853: OUTPUT«ok 00:01 123m␤»
flussence that works, keeps vim's syntax highlighting happy too :) 19:49
FROGGS but it is possible that it will warn/throw in future
because that could be an unintentional use of quotes 19:50
flussence good point... 19:51
vendethiel I wonder how I could get something with multi-line blocks from masak++'s ws-parsing stuff.. 19:58
masak mine? 19:59
oh, the indent stuff.
well, the multi-line block would just end up in the same Suite, no? 20:00
dalek ast: a835db1 | masak++ | / (4 files):
chase spec change about .later/.earlier and TimeUnit

This makes roast up-to-date with spec change 337f433398a070884e3258fc724c3b5f775fe085.
p: 8d37d8c | (Tobias Leich)++ | t/nqp/19-file-ops.t:
adjust lseek test because of a change in lseek itself

  '3' is not an invalid 'whence' anymore on some platforms, see:
  man7.org/linux/man-pages/man2/lseek.2.html
  "SEEK_DATA and SEEK_HOLE are nonstandard extensions also present in
Solaris, FreeBSD, and DragonFly BSD; they are proposed for inclusion in the next POSIX revision (Issue 8)."
kudo/nom: f244c39 | masak++ | src/core/Temporal.pm:
[Temporal.pm] implement later/earlier

Also kill off TimeUnit and change the API for .truncated-to, all according to spec change 337f433398a070884e3258fc724c3b5f775fe085.		 20:01
masak I wanted to get this one in before tomorrow's release. glad I did.
FROGGS masak++
masak this gets the basic thing in. a few comments, though:
FROGGS and so did I sneak something in too :P
masak (a) I didn't deprecate .delta, I just replaced it. someone wants to do better -- feel free. 20:02
(b) .truncated-to handles the 'minute' form but not the 'minutes' form. this is a bug, and it currently has no spectest coverage. I will fix if no-one beats me to it. 20:03
hm. that's it, I think. 20:04
oh, here, let me update the ChangeLog, too.
FROGGS k, I'll simulate a rakudo-star on these changes later today or tomorrow 20:05
flussence grr... I keep getting a "use of uninitialized value of type Any in string context" with no line info :(
oh, wait a minute... 20:06
flussence turns Grammar::Tracer on and finds it instantly
vendethiel masak: yeah, it'd, but I need to make a difference between a rule and selector
(basically for a css preprocessor)
dalek kudo/nom: f5e98b3 | masak++ | docs/ChangeLog:
[ChangeLog] mention removal of .delta and TimeUnit
 20:07
vendethiel I'll need to override line, that much is obvious 20:08
masak I would like to implement .squish(:expires), too, but it unlikely I'll make it before the release, because I'm teaching tomorrow. 20:09
dwarring r: my $value = True but False; say $value
camelia rakudo-jvm 196b4f: OUTPUT«(timeout)»
..rakudo-{parrot,moar} 196b4f: OUTPUT«False␤»
timotimo line numbers with warningsbwould he fantastic before the mvm release 20:10
masak rakudo-jvm is all "WARNING. PARADOX. MUST OVERHEAT."
dwarring heh 20:11
FROGGS not everything can happen before this release :o)
dwarring I'll put in an rakudobug 20:12
FROGGS so, let us just take it as "after the release is before the release" and do things whenever it fits
masak dwarring: just be aware, camelia has been known to timeout on rakudo-jvm. 20:18
dwarring: so it might just be an artifact. 20:19
FROGGS n: my $value = True but False; say $value 20:21
camelia niecza v24-109-g48a8de3: OUTPUT«True␤»
FROGGS m: my $value = True but False; say $value
camelia rakudo-moar 196b4f: OUTPUT«False␤»
FROGGS I guess it is about that
dwarring this expects 'True' - perl6advent.wordpress.com/2010/12/1...lse-truth/ 20:23
masak I kind of expect True, too. 20:24
dwarring r: my $value = True but False; say ?$value 20:25
camelia rakudo-{parrot,jvm,moar} 196b4f: OUTPUT«False␤»
masak because the *value* is still True, kind of. it's just the boolean aspect of it that has been overridden.
FROGGS n: my $value = True but False; say ?$value
camelia niecza v24-109-g48a8de3: OUTPUT«False␤»
FROGGS niecza++
masak: there are no star regressions due to your patch btw 20:26
masak I suspected that. 20:27
FROGGS gnight #perl6 20:32
masak 'n8, FROGGS 20:32
dalek ast: 2cee534 | (David Warring [email@hidden.address] | integration/advent2010-day (2 files):
adding 2010 advent days 19 and 21
 20:59
masak 'night, #perl6
vendethiel when you read all the advents post in one time (which I pretty much did, 2 times) you notice how great the difference it 21:00
is*, between 2009 and 2013.
colomon vendethiel: it's easy for those of us who were here all that time to get blase because the changes happened little by little. 21:03
vendethiel yeah, I guess :] more stuff to talk about 21:04
dagurval I'm getting "Spesh: failed to fix up handlers" with moarvm on Time::Duration tests. Should I report that somewhere? 21:07
colomon hmmm, wonder what the current timings of perl6advent.wordpress.com/2009/12/1...g-snowmen/ are like... 21:12
PerlJam colomon: I await your blog post that tells us ; 21:16
PerlJam er ;) 21:16
PerlJam The Gtk one could be ported to use Gtk::Simple probably and then it'd work on moar and parrot (maybe?) and not just jvm. 21:21
colomon …. you mean and not just niecza. ;) 21:37
Is Gtk::Simple actually up and running?
raiph colomon: jnthn.net/papers/2014-nlpw-reactive.pdf#page=40 # jnthn's very cute recent announce of GTK::Simple 21:40
yoleaux 13 May 2014 06:49Z <FROGGS> raiph: Hi, are you still working on the script that parses mirrored.by, or should I use one of these as a base? gist.github.com/raiph/63508a5f4ade000acc7f / gist.github.com/raydiak/11382643
colomon raiph: yes, I saw the announcement, and I was all ready to rip the Gtk::Simple code out of that and make it a module when I utterly failed to successfully install the relevant version of Gtk on my Mac. 21:41
raiph colomon: I imagine the 2014 advent calendar will have several speed focused entries; an update then of 2009's "Making Snowmen" sounds perfect as one of them :) 21:48
tadzik it'll contain at least one game :) 21:49
huf jnthn: typo on slide 19, distnace
timotimo there once was a tetris game, ISTR
tadzik there was 21:50
timotimo ooooh, Test is more colorful now 22:07
nice
panda didn't get updated for the latest TAP things we have? 22:08
tadzik we have TAP things? :D 22:09
timotimo gist.github.com/timo/f1cfec5a1c91cd793fba - who knows what's up here?
tap parse errors
tadzik uh, wtf 22:10
timotimo i'm glad you're up :) 22:11
tadzik reproducible?
timotimo let me try. 22:12
i don't think the tests used to be loud tests
the weird thing is 22:13
ok 1 - what should pass, passes
this line is red
and yes, i can reproduce it
tadzik is dies_ok running in a different thread or what? :O 22:15
timotimo no clue. 22:17
tadzik is that moar head? 22:18
timotimo yes 22:19
and nqp head and rakudo head
tadzik oh, hm
feel like bisecting something? :D
timotimo :|
donaldh_ jnthn: is there a way to cancel an async listen? Seems not, unless I am missing something.
timotimo it should return a cancellation handler somewhere 22:20
or is that only for other stuff?
donaldh_ It returns a supply of new connections. 22:21
timotimo oooh 22:22
my panda was dirty
tadzik: i think it was all my fault
could we get an environment variable to make panda run tests loudly?
timotimo oh 22:23
i know why that would fail
it tests the tester
i manipulated the tester to test loudly
it spews its output during testing
and prove gets massively confused 22:24
tadzik ah, there we go :) dirty panda 22:26
timotimo aye, dirty panda 22:27
panda could warn about that if it fails to rebootstrap from a dirty commit :)
perltricks hey all, am writing a Test file. Does the Test module has use_ok, can_ok etc? error message suggests not, but some online docs say yes 22:27
tadzik no, it does not 22:29
which online docs say that?
perltricks Thanks Tadzik. See here: feather.perl6.nl/~markstos/Test.pm.html 22:29
tadzik ...how did you find that/ 22:30
tadzik this is from 2006 :D 22:30
perltricks Googling for Perl 6 test articles ... ah yeah ancient history then ;) 22:30
tadzik google at its best again ;)
I'm sure I can find some more up-to-date docs for you 22:31
perltricks Is there an idiomatic way to do use_ok? would it be: "use module || fail; " 22:31
tadzik if use fails, an exception willl be throws 22:32
so you can juts do 'use Module; pass "module loaded"
'
ok, it seems that there are no fresh Test docs
perltricks what about testing for methods? Is there a "can" function? 22:32
tadzik every object has a can() method 22:34
so you can do this:
tadzik r: use Test; ok Int->can('gist'); ok Int->can('fibo') 22:35
camelia rakudo-{parrot,jvm,moar} f5e98b: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unsupported use of -> as postfix; in Perl 6 please use either . to call a method, or whitespace to delimit a pointy block␤at /tmp/tmpfile:1␤------> use Test; ok Int->…»
tadzik eeh?
tadzik oh, of course 22:35
r: use Test; ok Int.can('gist'); ok Int.can('fibo')
camelia rakudo-{parrot,jvm,moar} f5e98b: OUTPUT«ok 1 - ␤not ok 2 - ␤» 22:35
tadzik nnah
tadzik oh, ok :) 22:35
perltricks wonderful. Thank you 22:36
tadzik you're welcome 22:36
timotimo er 22:38
timotimo "use" works at compile time 22:38
you need to do it in an eval if you want to put it into tests
EVAL, even.
perltricks timotimo, thanks runtime would be better. but "EVAL use module;" returns undeclared name error. 23:02
timotimo aye, we have typed exceptions (output the value of $!.WHAT) 23:04
perltricks is EVAL documented anywhere? 23:06
timotimo in the synopses surely
www.google.com/search?ie=UTF-8&...e+Synopses 23:07
perltricks found it, thanks 23:08
ok got it now: BEGIN { EVAL('use module') } 23:11
timotimo er ... no. 23:25
what are you trying to achieve?
NativeCall gives a whole bunch of deprecation messages; are we going to change that up for the release or something? 23:26
timotimo i went ahead and pushed jnthn's Simple.pm6 into perl6/gtk-simple/lib/GTK/Simple.pm6 23:55
and added a META.info
timotimo my class Queue is repr('ConcBlockingQueue') { }; my $!queue := nqp::create(Queue) - does this have a more canon spelling yet? 23:58
apparently that's exactly how the setting does it 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!