»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 moritz on 3 May 2013.
00:24 tgt left 00:25 pmurias left
masak 'night, #perl6 00:28
00:30 tgt joined
sorear oh hey, "return" isn't implemented. 00:32
00:35 benabik left 00:39 benabik joined 00:47 benabik left
sorear now it is! :D 00:50
mst :D 00:52
00:53 benabik joined 00:54 woosley joined
sorear o/ mst 00:55
01:02 FROGGS_ left 01:09 anuby joined 01:15 FROGGS_ joined 01:16 hypolin left, hypolin joined
colomon sees no sign of a recent push.... ;) 01:32
01:49 lustlife` joined 01:58 FROGGS__ joined 02:02 FROGGS_ left, bonsaikitten left 02:03 bonsaikitten joined
labster c'mon dalek, show us some commits! 02:03
02:04 Khisanth left
sorear colomon: i has no commit bit 02:06
colomon sorear: right. :( 02:07
labster All you have to do is send a letter 50 miles north, sorear.
I can merge pull requests, to rakudo anyway. 02:08
colomon as can I.
labster r: say "あ".succ # expected output: い 02:14
camelia rakudo b2072f: OUTPUT«あ␤»
labster r: say "π".succ 02:16
camelia rakudo b2072f: OUTPUT«ρ␤»
02:18 Khisanth joined
colomon has somehow wandered from getting ready to add more Set operator tests to reading about bicameralism. 02:20
labster read that as 'bicamelism' 02:21
how do you feel about mkdir returning a path object rather than True? 02:23
colomon one camel is enough, thank you
would it just return the path you just asked it to make? 02:24
labster yeah
colomon there's a certainl logic to that
"t/spec".path.mkdir.chdir
return Failure if it doesn't work? 02:25
labster exactly
colomon +1
labster currently a thrown failure until we get the "use fatal" in try blocks thing. I'm not in a mood to break panda. 02:26
02:27 benabik left 02:29 ldthien0 joined
colomon it's still somewhat broken anyway, I think... 02:30
02:32 Chillance joined
labster yeah, but that's more of a language design brokenness, I think 02:37
02:38 aindilis left, aindilis joined
labster Personally, I like the idea of soft failure, because it's somewhere between checking for return codes, and stopping the program in the middle. 02:39
colomon labster: no, I mean panda is somewhat broken
at least, I'm routinely getting weird results from it
labster It does that... yeah. 02:40
02:40 atroxaper joined 02:47 nitestryker joined 02:51 BenGoldberg left 02:56 ldthien0 left, ldthien0 joined, nitestryker left
colomon Don't suppose TimToady is out there somewhere? 03:05
03:10 BenGoldberg joined
labster I believe he's in Europe now, so not at this time of day 03:12
colomon ah 03:15
I am soldiering on, making weird but apparently working changes to Set
well, the first one worked, anyway 03:20
03:33 tgt left 03:34 preflex left, preflex_ joined, ChanServ sets mode: +v preflex_, preflex_ is now known as preflex
colomon has [∪] and [∩] working in Niecza, but is going to bed without checking them in. :) 03:53
o/
sorear o/ colomon
04:07 census joined 04:13 konundra left, lustlife` left 04:19 snarkyboojum joined 04:27 Psyche^ joined 04:30 Patterner left, Psyche^ is now known as Patterner 04:41 nitestryker joined 04:49 birdwindupbird joined 04:51 BenGoldberg left 04:54 zby_home_ joined 04:55 FROGGS__ left 04:56 odoacre joined 05:01 snarkyboojum left
sorear is starting on a long-term project to replace p6eval with something more robust 05:01
ssutch :) 05:02
lue more robust how? 05:11
05:21 census left 05:27 amoe joined
moritz sorear: something dogfooding? 05:29
sorear no. 05:30
05:30 amoe_ left 05:31 dmol joined
sorear I may just get a bigger VPS and move all of p6eval there 05:31
getting rid of the 1/3 split would probably help
05:38 FROGGS joined
bonsaikitten if y'all need some hosting help, I might have some spare capacity 05:43
Teratogen I have an amazon ec2 vps! 05:46
it's free!
(for a year)
05:52 dmol left 05:57 zby_home_ left 06:01 bluescreen10 left, bluescreen10 joined
sorear bonsaikitten: baseline memory usage is going to be lowesh, but it needs to tolerate spikes up to 2-3G without thrashing; disk requirements are probably dominated by the jdk1.7 06:02
bonsaikitten sorear: sounds tolerable 06:03
if you can live with a hard cap at, say, 3GB
06:06 zby_home joined
sorear is trying to get CATCH working in rakudo now 06:10
FROGGS ++sorear 06:12
06:15 zby_home left
sorear savors the irony that rakudo-jvm currently has working multiprompt delimited continuations but mostly not exceptions 06:33
FROGGS :P 06:34
06:38 kaleem joined
sorear nqp: try die() 06:40
camelia nqp: ( no output )
sorear die("foo") works correctly. try { nqp::die("foo") } works correctly. try { die("foo") } doesn't catch the exception 06:46
FROGGS O.o 06:47
sorear oh, I see, (I think)... p6 exception objects are getting thrown as control exceptions with an invalid type, so they can't be caught at all 06:55
07:01 ldthien0 left
dalek p: f16d02d | sorear++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/VMExceptionInstance.java:
Default new exceptions to EX_CAT_CATCH

I don't know if this is right, but it's less wrong than defaulting them to an invalid (0) category that is impossible to catch.		 07:02
timotimo_ early morning coding: first attempt at making @foo.>>.uc work
sorear would like to get to a state where you don't need --ll-exception all the time 07:05
FROGGS that would be cool, yes
timotimo_: what is your motivation? is it specced that way? 07:06
r: my @foo = 'A' .. 'F'; say @foo>>.uc
camelia rakudo b2072f: OUTPUT«A B C D E F␤»
timotimo_ it is specced that that's allowed 07:08
std: <foo bar baz quux>.>>.uc
camelia std d4cc5ab: OUTPUT«ok 00:00 41m␤»
timotimo_ std: (1, 2, 3, 4, 5).>>++
camelia std d4cc5ab: OUTPUT«ok 00:00 44m␤»
timotimo_ r: <foo bar baz quux>.>>.uc
camelia rakudo b2072f: OUTPUT«===SORRY!===␤Confused␤at /tmp/KSnFk4vbPj:1␤------> <foo bar baz quux>.⏏>>.uc␤ expecting any of:␤ dotty method or postfix␤»
timotimo_ r: <foo bar baz quux>>>.uc.perl.say
camelia rakudo b2072f: OUTPUT«("FOO", "BAR", "BAZ", "QUUX")␤»
timotimo_ i think >.>>. looks a bit nicer than >>>.
sorear n: <aa bb cc>.>>.uc.perl.say
camelia niecza v24-76-g3e65d84: OUTPUT«["AA", "BB", "CC"].list␤» 07:09
timotimo_ also, std allows unspace there
std: <foo bar baz quux>\ .>>\ .uc.perl.say
camelia std d4cc5ab: OUTPUT«ok 00:00 41m␤»
07:09 domidumont joined
timotimo_ well, it works locally now. 07:11
just need to run spectests now
FROGGS cool
timotimo_ was pretty darn easy, too :)
FROGGS timotimo_: btw, I fixed the &&= / .= bug
(in v5)
timotimo_ is the commit up already? i'd like to see what you had to change 07:12
FROGGS no, not yet
07:12 domidumont left
FROGGS the infix:<.=> must return a variable and needs to be rw 07:12
timotimo_ ah, an is rw was missing
i was *so* close :)
07:12 domidumont joined
FROGGS yeah, and .= (concat_op in actions) made a bind, I changed it to p6store 07:13
sorear looks like it's currently failing because printing a backtrace ultimately involves iterating over a range, and RangeIter is not happy 07:14
timotimo_ sorear, that's pretty cool!
FROGGS ahh, hehe
sorear ...because cmp doesn't work...
timotimo_ sorear: is that the only reason? :) 07:16
sorear timotimo_: no, there are other reasons hidig behind it I'm sure
timotimo_ expects 0 failed tests and a few todo passes 07:17
sorear i'm not even trying to run tests at this point 07:18
timotimo_ (my changes, not yours ;) ) 07:19
FROGGS meh, and I thought sorear is crosscompiling to timotimo_ *g*
moritz maybe we should deactive pretty backtraces in rakudo-jvm for now 07:22
*deactivate
timotimo_ oh, i probably don't get the todo passes because they were hard-skipped for rakudo 07:24
(my @r = (1, 4, 9))».++; - Method 'postcircumfix:<( )>' not found for invocant of class 'Int' - er?! 07:28
r: say (1, 2, 3)>>.++;
camelia rakudo b2072f: OUTPUT«No such method 'postcircumfix:<( )>' for invocant of type 'Int'␤ in sub METAOP_HYPER_CALL at src/gen/CORE.setting:14143␤ in block at /tmp/AdFckmwzgW:1␤␤»
timotimo_ there were tests for this, but they didn't fail because the whole block was skipped due to the .>> form missing
sorear peels back another layer 07:29
cmp is failing because Order.(1) doesn't work
timotimo_ i think the qast that gets generated from that is wrong 07:31
07:31 salv0 left
timotimo_ gist.github.com/anonymous/a60e3321e57046edaa39 - seems to not mention the ++ operator at all 07:32
FROGGS timotimo_: maybe the dump is incomplete 07:34
timotimo_ that would be a possibility for sure :(
FROGGS because it calls a &METAOP_HYPER_CALL
timotimo_: you could dump the @a>>++ form to compare 07:35
sorear r: say Order ~~ Enumeration
camelia rakudo b2072f: OUTPUT«False␤»
timotimo_ that calls QAST::Op(call &METAOP_HYPER_POSTFIX) with the list and then QAST::Var(lexical &postfix:<++>)
sorear r: say Order.^roles 07:36
camelia rakudo b2072f: OUTPUT«No such method 'roles' for invocant of type 'Perl6::Metamodel::EnumHOW'␤ in block at /tmp/aY78vovQzl:1␤␤»
timotimo_ it would still be a net win if i unskip the block, re-skip the >>.++ forms and put that up?
07:37 fhelmberger joined
moritz yes 07:38
sorear > Order.^methods(:local) 07:39
postcircumfix:<( )> perl roll pair kv pick Int enums Str Numeric gist <anon> <anon>
> Order.(2)
Method 'postcircumfix:<( )>' not found
umm
07:39 SmokeMac_ left
timotimo_ i don't think i understand skipping 07:40
don't i put a #?rakudo skip 'explanation' newline { newline code newline };?
moritz should work, and skip the block 07:42
timotimo_ does it only work on top-level blocks?
gist.github.com/anonymous/17fbd409820a3da23724 - looks like this currently
07:42 pr_ joined, SmokeMachine joined
timotimo_ the skip seems to generate its own braces around only the first line in the block or something? 07:43
do i need the #?DOES 3 thing?
nope, don't 07:44
07:44 sqirrel joined
moritz it might only work with top-level blocks 07:44
I remember some fudge strangeness with that
sorear I also need to do something about the rakudo-jvm startup time
timotimo_ right. i'll cut out the non-working parts temporarily
moritz sorear: port it to moarvm? :-) 07:45
sorear real 0m7.348s
user 0m13.399s
"on the plus side, it seems to be automatically using both my cores"
07:45 domidumont left 07:46 flaviusb joined
timotimo_ :D 07:46
07:46 domidumont joined
pr_ Hello, when I read lines from a DOS Ansi-Latin-1 file with Rakudo Perl 2013.05, I get a "malformed utf-8 error" on following block: for $fh.lines -> $line { say $line; }. Any help would be much appreciated. pr 07:47
moritz pr_: pass :encoding<Latin-1> to the open() call 07:48
pr_ That did it. Thanks. Any documentation on this anywhere? 07:49
nwc10 sorear: This is the JVM? I observed this. Hypothesis is that it's one core for real work, and one core for the GC
sorear: also, jnthn might have some thoughts on JVM startup time. 07:50
Parrot has to be better for something :-)
timotimo_ sadly, dotty postfix ops seems to be much more complicated
nwc10 (Although right now, lets not forget, Parrot is still most complete platform for Rakudo)
07:50 flaviusb left
moritz pr_: perlcabal.org/syn/S32/IO.html#Functions 07:51
07:51 hoelzro_ is now known as hoelzro
pr_ "method encoding": found it. merci. 07:53
moritz and the Str :$enc = "Unicode", 07:54
argument to open()
I guess we allow :encoding as a synonym for :enc
07:54 woolfy joined
timotimo_ i don't understand the reasoning behind dotty:sym<.*> adding a dispatch:<.> to the dottyop that is created from, for instance ++ in $foo.++ 07:57
(especially when dispatch:<.> doesn't seem to exist) 07:58
er, no, postfix:<.> is the one that doesn't exist
hm. 07:59
08:03 SamuraiJack left 08:06 flaviusb joined 08:08 SamuraiJack joined
sorear Wrong number of arguments passed; expected 1..-1, but got 0 08:09
08:10 flaviusb left 08:11 flaviusb joined
diakopter sorear: *giggle* 08:13
08:15 flaviusb left
sorear apparently nqp-jvm is systematically failing to pass the invokee as an argument to nonstandard invoke routines 08:18
so Order.(1) is getting mangled into 1() 08:19
08:19 ldthien0 joined, ldthien0 left 08:20 pr_ left
diakopter sorear: maybe it's interpreting -1 as Inf 08:21
(I just realized)
dalek ast: 5a7f157 | (Elizabeth Mattijsen)++ | S32-list/categorize.t:
First batch of additional categorize tests
 08:22
lizmat good morning #perl6 from Nancy, France
about to decommute :)
08:22 rindolf joined
FROGGS hi lizmat 08:23
08:27 flaviusb joined 08:31 flaviusb left
sorear o/ 08:33
08:35 woolfy left
lizmat on our way back home, with some sightseeing on the way… see you all in ~6 hours& 08:37
08:37 lizmat left 08:41 dakkar joined
dalek p: 932e096 | sorear++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/ (3 files):
Add invokee as first argument to invocation handlers for Parrot consistency and basic sanity
 08:41
timotimo_ sorear: suddenly everything up the chain works, too? 08:45
08:45 raiph left 08:46 flaviusb joined 08:51 dmol1 joined 08:53 dmol1 left
masak good antenoon, #perl6 08:53
diakopter hi masak
TimToady is glad to see that sorear++ and diakopter++ now know each other well enough that the occasional snippiness at each other's abbreviated style no longer jeopardizes the friendship :) 08:54
timotimo_ \o/
TimToady (backlogging, as you might surmise :) 08:55
sorear timotimo_: no, I get an inch futher 08:57
method postcircumfix:<( )>($ ($x)) {
$x ~~ ::?CLASS ?? $x !! self.^enum_from_value($x)
}
the good news is we can now properly call postcircumfix:<( )>
ACCEPTS is blowing up. 08:58
TimToady finds the pun on unacceptable to be unacceptable...
timotimo_ heads towards uni 08:59
on a bike
in 30 degC :(
TimToady don't get run over
diakopter brrrrr
sorear no wifi on the bike, eh
diakopter oh wait
timotimo_ there's bike lanes all the way to the uni, except for a few meters all in all :)
sorear i hear they have it on trains and planes now
TimToady diakopter: you were thinking Kelvin :P 09:00
timotimo_ oh, you probably meant actual humans *running* me over, like on foot
yeah, that could happen
i'll be careful, thanks!
09:00 pr_ joined
sorear timotimo_: try not to catch fire either 09:00
TimToady (when you halt)
sorear good antenoon masak 09:01
diakopter comes upon a way to determine if a non-gmail.com address is actually a gmail address 09:02
xmpp leaks
pr_ Hi, in Perl6, is it possible to conditionally create variables? eg, let's say I am reading a bunch of csv files and generating sql files to insert the data into a database. Each csv file has a different structure: whereas file 1 has 10 columns, file 2 has 12,etc. Is there a way to shorten/lengthen the number of declared variables based on the file name? eg, my ($v1,$v2) = split("\t"... in the case of file 1, my ($v1,$v2,$v3) in case the sc 09:05
TimToady generally that means you want an array or hash instead 09:06
and no, you can't change the number of lexically declared variables without something like an eval 09:07
pr_ Could a macro do this?
diakopter not at runtime
... if you read in the file at macro time........ zugh
sorear ok, I'm getting my current problem because subsignatures are busted. sub bob($ ($x)) { say nqp::isnull($x) }; bob(\5)' # 1 09:08
exactly how many perl 6 features does "1..5" use?
diakopter about avogadro's number 09:09
TimToady at least one, which is itself
since a 1..5 returns 2..5 when iterated
FROGGS is it right that this behaviour is now against the spec? 09:13
r: say 1 for Nil
camelia rakudo b2072f: ( no output )
sorear ah, I'm running into a NYI in the binder
09:14 ssutch left
TimToady FROGGS: yes, that should say 1 once now, not assume that Nil means () 09:14
FROGGS great, thanks
TimToady: do we still need all of these? Nil, Any, Mu, Whatever? I have problems remembering what to take when :/ 09:15
TimToady they all mean different things
musée & 09:17
09:18 kaleem left
FROGGS fair enough 09:19
09:19 kaleem joined 09:29 domidumont left
pr_ Hi, I am having an issue with long heredocs in p6 2013.5. Is this the right place to mention it? 09:30
timotimo_ didn't catch fire on the way :) 09:31
even got here on time
moritz pr_: yes, feel free to ask
sorear so why is it that the rakudo C code is free of line-end whitespace but the Java code is absolutely dripping with the stuff?
pr_ How do I post 10 lines of code here? 09:32
sorear pastebin-type site
nwc10 is there a test that fails if you add whitespace to the C code? And no such test for the Java code?
moritz pr_: use a paste bin, gist.github.com if you don't prefer something else
sorear constantly has to resist doing a find | xargs perl -ipe
moritz nwc10: there used to be a test, when rakudo was part of parrot 09:33
sorear nwc10: would jnthn kill me if I added such a test?
pr_ What's a "paste bin"?
nwc10 sorear: I don't know. To make it pass you'd have to remove all the current whitespace
moritz pr_: a website where you can enter your 10 lines of code, and give us the URL where we can view it
nwc10 It wouldn't be terrible on git blame
but I'd ask him first
moritz sorear: experience shows that he'll just commit his changes with tests failing :-) 09:34
nwc10 (you can blame without whitespace, but not on the gitweb interface)
we need to put a lock on his beer fridge which only releases when all tests pass
sorear pre-commit hook?
nwc10: heh
moritz is it actualy worth the effort? 09:35
sorear moritz: the { and } commands in vim don't work if you have whitespace on your blank lines, and fixing vim seems like more effort :p 09:36
nwc10 I think that having a test is worth it. Otherwise ugliness gradually slips in
sorear also I have my vimrc set up to highlight EOL whitespace
nwc10 and then the git blame history gets messed up with "tidy up whitespace" type commits
which 10 years from now will be hateful 09:37
09:37 daxim joined
pr_ I have just pasted the code here: gist.github.com/anonymous/5795786 09:38
masak pr_: looks like a function call... 09:39
r: my $foo = "foo"; say "$foo(1, 2, 3)"
camelia rakudo b2072f: OUTPUT«No such method 'postcircumfix:<( )>' for invocant of type 'Str'␤ in block at /tmp/zeZyjquCVc:1␤␤»
masak pr_: is this the error you get?
r: my $foo = "foo"; say "$foo\(1, 2, 3)"
camelia rakudo b2072f: OUTPUT«foo(1, 2, 3)␤»
masak pr_: try it like that.
moritz or try qq :s :heredoc 'EOT'; 09:40
then only scalars should be interpolated, no function call syntax
pr_ moritz: :s does not make any difference 09:41
sorear .ask jnthn Would it be insane for me to rename org.perl6.rakudo.Ops to RakOps or something so we don't have to fully qualify all the time?
yoleaux sorear: I'll pass your message to jnthn.
diakopter stumbles across this in the clogs irclog.perlgeek.de/perl6/2009-11-27#i_1775684
pr_ companyname is not a function call. It's just a database column name in a long heredoc. 09:42
masak diakopter: heh :)
pr_: if you write "$var(...)" in a string, even a heredoc, Perl 6 will read that as a function call. 09:43
pr_: you haven't said that's the problem, but I'm currently assuming it is.
moritz r: my $x = 'foo'; say "$x()"
camelia rakudo b2072f: OUTPUT«No such method 'postcircumfix:<( )>' for invocant of type 'Str'␤ in block at /tmp/zHsI5B0hVc:1␤␤»
moritz r: my $x = 'foo'; say q:s"$x()"
camelia rakudo b2072f: OUTPUT«No such method 'postcircumfix:<( )>' for invocant of type 'Str'␤ in block at /tmp/csCxmndlcT:1␤␤»
diakopter masak: when do comedies of errors become tragedies?
masak diakopter: depends on one's disposition, I should think. 09:44
09:44 chayin_ joined, chayin left
pr_ masak: putting a backslash before the left parens next to $csv_file did the trick. thx. 09:48
09:50 pr_ left
sorear sleep& 09:55
09:59 anuby left
timotimo_ interesting, i got to a point in debugging where $_.++ will generate a dottyop with an NQPMu as its .ast 10:13
is that due to backtracking fixing it later or is the grammar in that case faulty?
FROGGS I'd say the grammar doesnt fit the actions (or vice-versa) 10:14
10:16 rindolf left
timotimo_ i may be barking up the wrong tree entirely and there should be something more in the grammar than just dottyop:sym<.> and :sym<.*> 10:19
FROGGS can you paste a diff? 10:22
timotimo_ the only additions i made are some nqp::say's spread over dotty and dottyop 10:25
i'll need to relocate in a few minutes
FROGGS hmmm, okay
10:38 kresike joined
kresike hello all you happy perl6 people 10:38
FROGGS hi kresike 10:41
kresike FROGGS, o/ 10:44
10:46 domidumont joined
cognominal nr: my $i = 1; 'a' ~~ / 'a' ** $i / 10:48
camelia niecza v24-76-g3e65d84: OUTPUT«Potential difficulties:␤ Unsupported use of atom ** $i as separator; nowadays please use atom+ % $i at /tmp/CZeVkCIWmS line 1:␤------> my $i = 1; 'a' ~~ / 'a' ** $i⏏ /␤␤»
..rakudo b2072f: OUTPUT«===SORRY!===␤Quantifier quantifies nothing␤at /tmp/gbCq9vjr6B:1␤------> my $i = 1; 'a' ~~ / 'a' ** ⏏$i /␤ expecting any of:␤ postfix␤ infix stopper␤ infix or meta-infix␤ prefix or term␤ pref…
FROGGS cognominal: NYI, I already tried to do that without luck 10:49
cognominal FROGGS, thx
cognominal searches for a workaround 10:50
masak cognominal: match with * or +, then check the number of matches in a <?{ ... }> 10:51
10:51 odoacre left
FROGGS nr: regex atom { 'a' }; my $i = 1; 'a' ~~ / [<?{ $i--; } <atom>]* / 10:51
camelia niecza v24-76-g3e65d84: OUTPUT«===SORRY!===␤␤Methods must be used in some kind of package at /tmp/sMvKsVIfg4 line 1:␤------> regex atom ⏏{ 'a' }; my $i = 1; 'a' ~~ / [<?{ $i--; ␤␤Unable to parse metachar:sym<< >> at /tmp/sMvKsVIfg4 line 1:␤------> […
..rakudo b2072f: OUTPUT«Useless declaration of a has-scoped method in mainline␤===SORRY!===␤regex assertion not terminated by angle bracket␤at /tmp/gE5kjN8bTs:1␤------> 'a' }; my $i = 1; 'a' ~~ / [<?{ $i--; }⏏ <atom>]* /␤ expecting any of:␤ …
cognominal masak: indeed
10:52 odoacre joined
FROGGS nr: my regex atom { 'a' }; my $i = 1; say 'aaa' ~~ / [<?{ $i-- }> <atom>]* / 10:52
cognominal …using a non greedy qunatifier to avoid to much backtracking.
camelia rakudo b2072f, niecza v24-76-g3e65d84: OUTPUT«｢a｣␤ atom => ｢a｣␤␤»
FROGGS nr: my regex atom { 'a' }; my $i = 7; say 'aaa' ~~ / [<?{ $i-- }> <atom>]* / # it is almost good :o)
camelia rakudo b2072f, niecza v24-76-g3e65d84: OUTPUT«｢aaa｣␤ atom => ｢a｣␤ atom => ｢a｣␤ atom => ｢a｣␤␤»
10:53 ecocode joined
FROGGS well, you could make it fail when $i is nonzero 10:53
nr: my regex atom { 'a' }; my $i = 7; say 'aaa' ~~ / [<?{ $i-- }> <atom>]* <!{ $i }>/ # test
camelia rakudo b2072f, niecza v24-76-g3e65d84: OUTPUT«｢aa｣␤ atom => ｢a｣␤ atom => ｢a｣␤␤» 10:54
FROGGS hmmm
masak FROGGS: that doesn't seem very compatible with backtracking.
FROGGS true
masak: I run into something similar: I did something about heredocs, which added the stopper twice to a list because of a <?term> 10:56
masak oh! day after tomorrow is #masakism again! \o/ 11:06
an evening of great group hacking and discussion.
it seems that the core group of attendees will be smaller this time. maybe there will be more time for discussion and looking at individual solutions. 11:07
11:10 JimmyZ joined 11:16 atroxaper left 11:29 flaviusb left
cognominal by the way, should we say "forwardtracking" for non greedy quantifiers? :) 11:31
moritz "forthtracking" 11:32
cognominal I like FORTH, the language, but everytime the word is uttered, this zardozian image pops up in my mind :koti.mbnet.fi/manki1/mainoksia/Atari_forth.jpg 11:35
masak no, "backtracking" is fine, because the direction "back" pertains to regex-space, not to string-space. 11:36
cognominal yes, until non greedy qantifiers, backtracking applied to the two spaces. 11:38
moritz just spent two hours debugging a problem which turned out to be stale caches, proving the proverb about the two hardest problems in CS
cognominal moritz, what is the second hardest? 11:39
moritz cognominal: the proverb goes like this: The two hardest problems in CS are naming, cache invalidation and off-by-one errors
masak moritz: jnthn has started saying that the two hardest problems in CS are cache invalidation and cache invalidation :) 11:41
timotimo_ i declare making $_.++ work too hard for now.
masak moritz: (which is sufficiently autopunny for my taste) 11:42
timotimo_ github.com/rakudo/rakudo/pull/164 - pr'd my work so far :) 11:44
11:45 woosley left 11:47 census joined 11:48 rindolf joined
JimmyZ $_.++ works in alpha, for some reason, it lost in nom :P 12:03
12:03 Ulti left
timotimo_ :) 12:04
12:07 pmurias joined
pmurias one annoying thing about --target=ast is that for large files the source code for a given AST node takes a lot of screen space 12:08
any ideas how to fix that?
timotimo_ send it through a perl one-liner? :))
open it in a pager that won't do word wrapping?
moritz pmurias: AST simply isn't a terse, human-readable representation 12:09
timotimo_ moritz: the trouble is that each AST node will print the source code it was made from after itself 12:10
moritz but limited to max. 50 chars or so, no?
timotimo_ not sure 12:11
nope, it's full
moritz that's something that should be fixed, then
timotimo_ i can totally do that 12:12
pmurias my other idea is to spit out html+js so that things could be expanded on a click or something like that
timotimo_ if i put in the trimming in QAST::Node.dump, that won't be possible any more 12:13
pmurias timotimo_: emitting an interactive page directly from QAST::Node seems a better idea then parsing the textual output from --target=ast 12:15
so I wouldn't worry about that
timotimo_ you are suggesting an extra --target perhaps? 12:16
so that the content of the .dump method doesn't matter to you?
pmurias for an richer output I suggest an extra --target 12:17
moritz or an environment variable 12:18
12:19 skids left 12:21 Ulti joined 12:24 pmurias left, zakalwe left, zakalwe joined 12:25 pmurias joined
timotimo_ gist.github.com/timo/3cf2ef1bdd0b0dedd0d7 - my suggestion for autopun of the day 12:26
tadzik hehe 12:27
timotimo_ well, ./perl6-jvm -> use Test gives an error, so maybe that's the exception that flew? 12:28
(when trying to run the tests in t/01-sanity/ 12:29
and failing to get any running at all)
12:29 potatogim joined
FROGGS hehe 12:31
12:31 potatogim left
FROGGS timotimo_++ 12:31
12:33 timotimo_ is now known as timotimo 12:37 bluescreen10 left
timotimo still amazed by the rote speed of jakudo parsing 12:40
Stage parse : 91.597 <- parrot | jvm -> Stage parse : 37.542
moritz last I looked, 'use v6;' blew up on rakudo-jvm
timotimo ah, indeed
does "use v6;" do anything at all? i thought it was just a little nothing in the grammar or something 12:41
moritz it seems to do just a wee bit more 12:43
masak .oO( they whispered sweet nothings to each other -- mostly just no-op pragrams )
moritz are "pragrams" a mixture of "pragmas" and "programs"? 12:44
12:46 dayangkun joined 12:47 ecocode left
FROGGS use v6 is a noop in rakudo currently 12:49
moritz: about the blew-up, jnthn said something about that issue a week ago or so 12:50
[Coke] it's the main blocker at the moment to running the test suite. 12:52
(internal one, not spec.) 12:53
masak moritz: no, just a typo :)
12:53 snearch joined 12:57 daxim_ joined, daxim left
[Coke] another day of this channel being the only perl6 news on google alerts. 12:58
timotimo :( 12:59
masak r: my @stuff = 1, 2, 3, 4, 5; my @old = 1, 2, 3; say @stuff (-) @old 13:02
camelia rakudo b2072f: OUTPUT«set(4, 5)␤»
moritz r: my @stuff = 1..4; say @stuff (-) 0..3; 13:03
camelia rakudo b2072f: OUTPUT«set("1", "2", "3", "4")..3␤»
timotimo let in the new, throw out the old.
moritz r: my @stuff = 1..4; say @stuff (-) (0..3);
camelia rakudo b2072f: OUTPUT«set(4)␤»
timotimo haha :)
13:09 ajr joined, ajr is now known as Guest3673 13:10 Guest3673 is now known as ajr_ 13:11 atroxaper joined 13:17 tgt joined 13:18 tgt left 13:21 bonsaikitten left, PacoAir joined 13:22 bonsaikitten joined 13:24 konundra joined, tgt joined
timotimo fwiw, the error experienced by use v6 comes from my $v := $*W.find_symbol(["Version"]).new(~$<vfoo>); 13:25
and also, my $v = Version.new("v6"); will give a NPE
> my @parts = "v6".comb(/:r '*' || \d+ || <.alpha>+/); 13:26
java.lang.NullPointerException
it would appear .comb is broken and that causes version objects to not be creatable 13:27
does it sound terrible to add a #?jvm block that special-cases "v6"?
dalek ast: e924fe3 | (Solomon Foster)++ | S03-operators/set.t:
Tests for reduce on set union and intersection

Add a first few simple tests for reduce meta-op on union and intersection.		 13:28
colomon not at all surprised comb is broken... I imagine it relies on gather / take, and that's not there yet.
timotimo could be 13:29
i'm testing my ugly hack that checks if the string is "v6" ;)
colomon a temporary hack to make "use v6" work seems like a Very Good Thing.
13:31 salv0 joined
timotimo didn't get me much closer to the finish line it seems 13:34
weird, the find symbol + new thing still fails? 13:35
(probably because i misunderstood what text gets passed to the constructor) 13:38
13:39 FROGGS left 13:40 benabik joined 13:43 kaleem left 13:46 sqirrel left 13:49 kaare_ joined
colomon rn: say (set <a b c>) (|) (set <1 2 3>) (|) (set <z y x>) 13:49
camelia rakudo b2072f: OUTPUT«Too many positional parameters passed; got 3 but expected 2␤ in sub infix:<(|)> at src/gen/CORE.setting:13260␤ in block at /tmp/k4xPGuWPGG:1␤␤»
..niecza v24-76-g3e65d84: OUTPUT«Unhandled exception: Excess arguments to infix:<(|)>, used 2 of 3 positionals␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (infix:<(|)> @ 1) ␤ at /tmp/Ld1DNnVEs6 line 1 (mainline @ 6) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4530 (ANON @ 3) ␤ …
13:50 panchiniak_ joined
dalek ecza: 2b14288 | (Solomon Foster)++ | lib/CORE.setting:
Change set union and intersection to be proper list operators.
 13:50
colomon clearly still need more set op tests!
13:51 skids joined 13:52 FROGGS joined
colomon thinks he has fixed that one in Niecza... 13:53
masak colomon++ # somebody set up us the tests 14:00
s/set up/set op/
:)
colomon After that test above, I now suspect the new reduce on union and intersection tests I pushed will fail on Rakudo 14:02
When I did them, I assumed they would work.
14:02 kaleem joined 14:07 snearch left
dalek rl6-roast-data: 67df23f | coke++ | / (4 files):
today (automated commit)
 14:11
rl6-roast-data: 1339dce | coke++ | / (3 files):
today (automated commit)
[Coke] niecza has been dirty for 334 days. pugs has been clean for 5 days. rakudo has been clean for 2 days. 14:12
dalek kudo-star-daily: 9a83b7d | coke++ | log/ (5 files):
today (automated commit)
kudo-star-daily: 7455ddb | coke++ | log/ (5 files):
today (automated commit)
14:13 panchiniak_ left
timotimo java.lang.NoSuchMethodError: org.perl6.rakudo.Ops.p6routinereturn(Lorg/perl6/nqp/sixmodel/SixModelObject;Lorg/perl6/nqp/runtime/ThreadContext;)Lorg/perl6/nqp/sixmodel/SixModelObject; 14:17
my fault or known error?
timotimo runs the sanity test suite now 14:20
14:21 benabik left
FROGGS timotimo: FYI: the bind=>p6store wasnt needed for the v5-issue 14:23
14:26 benabik joined
timotimo i'll issue a pullrequest for the "use v6" fix now after a bit of testing 14:28
[Coke] timotimo++ 14:29
14:31 BenGoldberg joined, spider-mario joined
timotimo failing a couple, a few of the test files run through completely, though 14:32
[Coke] that's about what I saw when I just commented out 'use v6;' earlier.
timotimo mhm
[Coke] timotimo++
dalek : 4a5c04a | (Tobias Leich)++ | lib/Perl5/Actions.nqp:
dispatch operators to overload-able subs
: 00becb9 | (Tobias Leich)++ | lib/Perl5/Terms.pm:
added method P5Numeric
: b141e6f | (Tobias Leich)++ | lib/Perl5/overload.pm:
added more operators, and fallbacks based on <=>
: 15fec8a | (Tobias Leich)++ | STATUS.md:
update after Str.P5Numeric and overload.pm update
14:41 benabik left 14:47 benabik joined
dalek kudo-js: ea49f2f | (Pawel Murias)++ | src/QAST/Compiler/JavaScript.nqp:
Handle QAST::Var's with 'static' as decl.
 14:47
p: 1729750 | (Pawel Murias)++ | src/QAST/Var.nqp:
Make --target=ast print out the value of decl on QAST::Var.
 14:48
FROGGS it is awesome what is happening these days^^
[Coke] can all nqp opcodes be run both directly with nqp:: and via QAST::Op ? 14:50
dalek p: c2eba21 | (Timo Paulssen)++ | src/QAST/Node.nqp:
if the code part of a node is too long, truncate it.
timotimo here's that fix i forgot to actually push
[Coke] nqp::say() is invokable 3 ways, it seems: nqp::say, say, and via QAST. wondering if all opcodes are similarly functional. 14:51
while, on the other hand, looks like it has the builtin and the :op, is probably not invokable with nqp::while. 14:54
(well, it is invokable, but not in a hepful way. can someone write an example where it works?)
14:55 aindilis left
pmurias [Coke]: via QAST? 14:57
[Coke]: the difference between say and nqp::say is that say can handle multiple arguments 14:58
nqp: say("Hello","[Coke]");
camelia nqp: OUTPUT«Hello[Coke]␤»
pmurias what would be a good site for a GSoC project blog?
FROGGS blogs.perl.org? 14:59
masak +1 15:01
15:02 kaleem left
[Coke] nqp: nqp::say("thing", "2"); 15:03
camelia nqp: OUTPUT«Error while compiling block : Error while compiling op say (source text: "nqp::say(\"thing\", \"2\")"): Operation 'say' requires 1 operands, but got 2␤current instr.: '' pc 48612 (src/stage2/QAST.pir:17766) (src/stage2/QAST.nqp:2981)␤»
[Coke] pmurias: no, I'm wondering if you can write an example that shows a functioning while loop with nqp::while
nqp: nqp::while(); 15:04
camelia nqp: OUTPUT«Error while compiling block : Error while compiling op while (source text: "nqp::while()"): Method 'fresh_' not found for invocant of class 'RegAlloc'␤current instr.: '' pc 48612 (src/stage2/QAST.pir:17766) (src/stage2/QAST.nqp:2981)␤»
[Coke] (or if I should just abandon that, and assume that everything in docs/nqp-opcodes.txt is for QAST usage.
15:07 ajr joined, ajr_ left 15:08 ajr is now known as Guest33115 15:10 tgt left
pmurias [Coke]: why do you want to write a while in such a way? 15:12
15:12 birdwindupbird left 15:13 SmokeMac_ joined
pmurias [Coke]: my $i:=3;nqp::while($i,say($i),$i--) 15:15
15:15 SmokeMachine left
pmurias nqp: my $i:=3;nqp::while($i,say($i),$i--) 15:15
camelia nqp: OUTPUT«3␤2␤1␤»
kresike bye folks 15:16
15:16 kresike left
pmurias blogs.perl.org/users/pawel_murias/2...stone.html 15:16
15:17 domidumont left 15:19 BenGoldberg left
masak pmurias++ 15:23
pmurias just relised that I need to fix an updated test case to full pass the inchstone ;/ 15:26
15:27 BenGoldberg joined 15:30 bluescreen10 joined 15:32 Ulti left 15:36 kbenson left
[Coke] pmurias: I don't, particularly. Trying to come up with some prose for the leadin of new nqp opcode docs. 15:40
wondering why your invocation of nqp::while there has 3 args when it's "documented" as taking only 2
15:45 Guest33115 is now known as ajr_ 15:47 JimmyZ left 15:50 Ulti joined 15:54 bbkr_ left
pmurias [Coke]: the third argument is the 3rd argument in for (;;...) {} 15:56
[Coke]: I myself avoid looking at the current nqp opcode docs
[Coke] pmurias: if we're trying to get more people to use nqp, I'd rather fix the docs then tell people to avoid them. 15:58
*than*
15:59 fhelmberger left
[Coke] looks like i will have to avoid the existing ones when generating new ones, though. 15:59
pmurias having better docs would be great 16:00
[Coke]: I had to rewrite a lot of things in nqp-js as I had to figure a lot of things by trial and error rather then by having good docs avalible 16:09
dalek kudo-js: ce87e59 | (Pawel Murias)++ | nqp:
Update the npq submodule.
 16:10
kudo-js: 5e1cd98 | (Pawel Murias)++ | runtime.js:
Enable stringification of exceptions. Fix failing test 44.
16:10 atroxaper left 16:13 konundra left
TimToady pmurias++ # "inchstone" :) 16:14
diakopter [Coke]++ awesome error of the month Method 'fresh_' not found for invocant of class 'RegAlloc' 16:20
16:21 pjcj left 16:23 daxim_ left
sorear timotimo-- duplicating my work 16:28
[Coke] which work? making 'use v6' work?
16:29 pjcj joined
sorear timotimo: if you pull sorear/rakudo, you'll find that p6routinereturn exists, as does gather/take 16:29
[Coke]: yes
[Coke] ah. I also had no idea you were tackling that issue. 16:32
16:32 tgt joined 16:33 SamuraiJack left 16:36 ajr_ left
timotimo yays! :) 16:39
sorear: does use v6; work in that repo? 16:40
also: sorry :(
16:42 REPLeffect left, tgt left
sorear not quiiiite 16:43
comb works, though. :) 16:45
timotimo yays :)
what part doesn't work?
sorear eq. 16:46
16:47 konundra joined
timotimo ah, interesting 16:48
in the selftests, == with "0" and "0" doesn't work either
colomon sorear++ 16:51
dalek : ddc0fc6 | (Tobias Leich)++ | lib/Perl5/Actions.nqp:
added overload-able op /
: 68c800e | (Tobias Leich)++ | lib/Perl5/ (2 files):
call .P5Numeric for numeric operations
: 5c6edcf | (Tobias Leich)++ | STATUS.md:
call .P5Numeric for numeric operations
16:57 lizmat joined 16:58 tgt joined, abnorman joined 16:59 vk joined
lizmat good *, #perl6! 16:59
17:00 vaelxon left
masak lizmat! \o/ 17:00
lizmat back in Echt, really!
17:02 sivoais is now known as dr-faustus 17:04 dr-faustus is now known as sivoais, ajr joined, ajr is now known as Guest50214 17:09 tgt left
masak O ECHT? 17:10
lizmat jah!
perigrin Crums DM, Oh Echt, Oh Crikey -- Penfold. 17:11
lizmat perigrin!
17:18 dakkar left 17:20 fgomez left, woolfy joined
perigrin lizmat! 17:23
17:24 tgt joined
lizmat glad to see you're not just lurking on #perl6 :-) 17:24
perigrin no no no I make the occasional random and mildly snarky comment too!
lizmat community building++
perigrin :D 17:25
17:26 konundra left, census left
masak it's true. I checked the backlog. 17:28
colomon This is the perigrin who had dinner with us that first Saturday night at YAPC::NA? o/ 17:30
masak I believe so.
dalek ast: f66aba6 | (Elizabeth Mattijsen)++ | S32-list/classify.t:
Added named sub test, just like with categorize
timotimo sorear: well, $x == $x where $x is the string "0" won't work because Method 'Bridge' not found 17:33
lizmat colomon, masak: perigrin is also one of the YAPC::NA 2014 organizers :-) 17:34
we found out tonight is the last night ST Into Darkness is showing in our local service cinema
17:34 twigel left
masak perigrin: if YAPC::NA 2014 is anything like YAPC::NA 2013, then well done! :P 17:34
colomon lizmat: I knew that the perigrin we had dinner with was. :) Just wasn't 100% sure that was how he spelled his nick, if you know what I mean.
17:34 tgt left
lizmat so, we're off for some 3D spectacle(s), double baileys and 1 set of chicken nuggets and 1 set of cheese sticks 17:35
benabik Has a location for ::NA '14 been announced?
colomon timotimo: that's probably something to do with roles, I'm guessing?
benabik: Orlando
lizmat off to see some James Tiberius Kirk, reimagined 17:36
17:36 tgt joined
colomon lizmat: yay! I'm looking forward to seeing that someday... 17:37
timotimo colomon: but i'm able to "0".Numeric with no problem and it should dispatch through that, no? 17:42
colomon .... dunno. 17:43
17:45 tomyan joined, BenGoldberg left
timotimo well, it works with "0" == "0" but not with $x == $x 17:46
i don't know how to figure out if the role isn't added properly
$x.^methods is unfortunately empty
17:48 dmol joined 17:50 tgt left
pmichaud good morning, #perl6 17:50
yoleaux 14 Jun 2013 16:13Z <lizmat> pmichaud: Maybe it is an idea to have a 2nd optional parameter to .classify|.categorize to indicate the hash to classify/categorize into?
moritz timotimo: with does?
timotimo my $x = "0";? 17:51
17:52 panchiniak joined, shinobicl joined 17:53 konundra joined, fgomez joined 17:57 panchiniak left
colomon timotimo: I'll try to take a look at it 18:00
dalek rl6-roast-data: c6291af | coke++ | / (4 files):
today (automated commit)
 18:01
18:01 BenGoldberg joined 18:04 btyler joined 18:08 tgt joined 18:09 p5eval left 18:11 p5eval joined 18:14 tgt left 18:15 tomyan left
colomon timotimo: I'm not able to do "0".Numeric. :( 18:21
\
> "0".Numeric
java.lang.NoSuchMethodError: org.perl6.rakudo.Ops.p6routinereturn(Lorg/perl6/nqp/sixmodel/SixModelObject;Lorg/perl6/nqp/runtime/ThreadContext;)Lorg/perl6/nqp/sixmodel/SixModelObject;
timotimo you need sorear's code for that to work
sorear/rakudo 18:22
colomon ah.
timotimo i don't have access to my desktop right now (busted v6)
so i can't really work on that any more :|
18:24 minibota joined
colomon timotimo: Ah, I see it now. 18:26
> "0".Numeric
0
timotimo yes, but curiously, my $x = "0"; $x.Numeric may not work? 18:27
18:27 zby_home_ joined, tomyan joined 18:29 tomyan left
colomon timotimo: actually that works fine for me. 18:30
> my $x = "0"; $x.Numeric
0
timotimo oh?
try $x == $x
colomon $x == 0 works 18:31
but
> $x == $x
Method 'Bridge' not found
that's actually a very odd error 18:32
pmichaud seems like an indication that a multimethod is missing somewhere.
colomon JVM is just using the normal setting, right? I'm not looking at the wrong code here? 18:33
perigrin masak: I hope it will be. I hope it will be better actually :) 18:34
timotimo it does, yes 18:35
colomon > class A { method Numeric { say "Called Numeric;" 14; }; }; 18:36
VMArray representation does not implement at_pos_boxed
.....
looks like an endless loop of that error message
typo on my part, but weird error! 18:37
Fixed version works, and calls numeric correctly, but... 18:38
> A.new == 0
Called Numeric
False
> A.new == 14
Called Numeric
True
> A.new == $x
VMArray representation does not implement at_pos_boxed
....
so this is quite bizarre
ah, don't think I'd defined $x that time.
pmichaud: what's weird with this is works for many, many cases. It's only the specific case of $x == $x which triggers the Method 'Bridge' not found message. 18:40
*it works
18:43 tomyan joined, domidumont joined
colomon is guessing the rakudo debugger doesn't work in JVM yet? 18:44
> "0" == "0" 18:47
True
> my $x = "0"; $x == $x
Method 'Bridge' not found
oh, the debugger doesn't step through core anyway, does it? Hmm. 18:51
18:53 tgt joined 18:54 tgt left, cognominal left 18:55 cognominal joined, tgt joined 18:56 nyuszika7h left 18:57 groky left, bonsaikitten left 18:58 bonsaikitten joined 18:59 avar left 19:00 avar joined, avar left, avar joined, Gothmog_ left 19:01 Gothmog_ joined, nyuszika7h_ joined 19:04 broquaint left, broquaint joined 19:05 nyuszika7h_ is now known as nyuszika7h, groky joined 19:08 tomyan left 19:11 jercos left
sorear good * #perl6 19:11
pmichaud: !!!
masak sorear! \o/ 19:12
19:12 diakopter left 19:13 tomyan joined
sorear pmichaud: hi 19:14
19:15 jercos joined 19:16 diakopter joined 19:17 tgt left
sorear diakopter: got your mail 19:17
19:17 tomyan left
sorear FYI 19:17
colomon o/ 19:18
diakopter sorear: ok thx :) 19:19
sorear pmichaud: i would like to talk to you (or another rakudo org admin) about CLA procedures 19:20
to confirm that www.perlfoundation.org/contributor_..._agreement is correct, and get an elaboration on the email procedure you mentioned
19:21 minibota left, ixxvil joined
pmichaud sorear: it's correct. 19:21
ixxvil so did you guys get a logo for that compiler?
or whatever it was called then
it had a funny name to it that began with an s
sorear there's rakudoperl.org/wp/wp-content/upload...o-1001.png 19:22
pmichaud sorear: better is probably www.perlfoundation.org/attachment/legal/cla3.pdf
sorear oh, that's a logo for a compiler that starts with an s
oh, that's a logo for a compiler that starts with an r
TimToady do we have one that starts with s?
sorear sprixel I guess
diakopter -- 19:23
ixxvil na not rakudo
TimToady ixxvil: why do you want to know?
ixxvil well if they still need one 19:24
diakopter tries to make your nick to a roman numeral
TimToady there are very few conserved needs here :)
sorear ixxvil: i'm not sure who they is... I can't think of a proper p6 compiler whose name starts with an s
TimToady either the needs get filled, or they go away...
diakopter sorear: if you haven't deleted it, would you mind forwarding the mail to me? 19:25
sorear perlito, mildew, pugs, niecza, rakudo
ixxvil niecza
:D
pmichaud perl6.org/compilers/ # list of compilers
benabik 'n' neq 's'
sorear done
ixxvil i remember TimToady talking about kerning and things then
but i left half way through, life got busy
pmichaud maybe it's really "sniecza" :-P
ixxvil so did niecza get one? 19:26
pmichaud or it could be ('n' | 's') ~ "iecza"
sorear maybe ixxvil is thinking of "sorear" :D
19:26 pmurias left
pmichaud I'm afk for a bit 19:26
ixxvil yeha i think the head dev for that was some guy who had a nick that began with an s
TimToady has no opinion on compiler logos, only language logos :)
sorear pmichaud: nothing on email procedures 19:27
benabik pmichaud: 'sniecza' makes me want to reply gesundheit
ixxvil you need help with a language logo?
pmichaud oh, I don't know that there are official "email" procedures.
sorear ixxvil: gee, I wonder who that could be
ixxvil i thought perl6 was all good to go
TimToady ixxvil: no, not at the moment :)
sorear :)
pmichaud checks.
ixxvil alright!
sorear pmichaud: ah, I was wondering because I thought you mentioned them but I didn't see anything about it on the site 19:28
ixxvil TimToady: you know i gotta thank you though
pmichaud we've accepted email copies as provisional until the originals arrive in the mail.
but we still end up needing originals.
TimToady ixxvil: no you don't, but I don't mind if you do :)
sorear ok
ixxvil i dont know if you helped then or not but the important of kerning and things like that play a ajor role in the work i do now
pmichaud as I said a few days ago, I'll gladly send you a CLA and pre-addressed-stamped envelope :)
ixxvil i didnt see it then but 2 yrs since yeah
importance
sorear how long does the mail generally take?
pmichaud well, 3-5 days. 19:29
sorear pmichaud: it'll go faster if I use my own printer and stamps
but thanks for the offer
colomon sorear++
pmichaud but we'll grant the commitbit as soon as you let us know "it's mailed" :)
ixxvil i remember vaguely you mentionig about my kerning being off and i kinda ignored your comment then
:D
TimToady ixxvil: well, I learned about kerning from my wife--it's one of those things you start to notice, and then can't ever quit
ixxvil right
thats all i do now
sorear pmichaud: ok, I have everything I need from you right now
ixxvil lettering and tons of type
sorear keming, eh
ixxvil keming is part of it
os part of the game! 19:30
TimToady the slides at YAPC::NA talked about Tcl at one point, and that looked pretty awful between the T and the c
moritz we also trow around lots of types in Perl 6 :-)
pmichaud sorear: excellent. just let me know when it's mailed and we'll grant the commitbit :)
ixxvil what kinda type
i dont do type per se as in i dont design typefaces
i could if its paid though otherwise no way hose
oh well anyways good times 19:31
19:31 panchiniak joined
moritz ixxvil: types, as programmers use them 19:32
r: say Int.^mro
camelia rakudo b2072f: OUTPUT«(Int) (Cool) (Any) (Mu)␤»
ixxvil you mean mono?
monospace is just ugly heh but you guys need it more than i do 19:33
moritz no, as in "data type"
sorear always thought it was José
ixxvil moritz: sorry, out of my jurisdiction 19:34
TimToady that's the gardener's juristiction, hose A and hose B
*diction
ixxvil what does your wife do again? 19:35
grapic design?
19:35 REPLeffect joined
TimToady ixxvil: she's a calligrapher 19:35
ixxvil REALLY?
TimToady among other things
ixxvil can i see some work
TimToady well, she hasn't done much since college...
ixxvil thats all i do now
[Coke] TimToady: so your main problem with tcl is the kerning? ;)
ixxvil but i dont use traditional methods besides on paper
there was an article recently about the calligraphers payscale at the whitehouse 19:36
19:36 tomyan joined
ixxvil it's something like 90k 19:36
but pretty darn good work
TimToady [Coke]: it may have been something else beginning with T, so it's really the T's fault, not Tcl's 19:37
19:38 census joined
ixxvil where is this tcl you're talking about though? on their webpage? 19:38
19:38 panchiniak left
ixxvil Tcl Developer Xchange.... dat X 19:38
inconsistency in the weight.. that prolly needs to go but i dont think they care 19:40
19:40 tgt joined 19:42 tomyan left
dalek kudo-star-daily: 7eb872f | coke++ | log/ (5 files):
today (automated commit)
 19:46
[Coke] anyone want to update the modules for star for the next release?
19:47 panchiniak joined
ixxvil alright folks nice talking to you 19:47
later!
19:47 ixxvil left
masak oh, right. it's rakudo release week. 19:48
heh, I wonder what sucker signed up for this monthly release... :P 19:49
19:49 domidumont left
labster isn't there some sort of online training in the near future, too? 19:52
19:53 tgt left
masak labster: in two days, yes :> 19:56
looking forward to it! 19:57
lue hello world o/ 20:00
sorear o/
colomon oh, so it's the day before release day. no worries, then.
20:00 Guest50214 left
sorear needs to hurry up and break rakudo 20:00
colomon sorear: a grand ambition. 20:01
sorear once I have my commitbit I can join the "broke the build on release day" club 20:03
20:04 ssutch joined
labster All I ever managed to break was Parrot's GC. 20:05
TimToady All I ever managed to do was break the language... 20:06
ssutch HA
labster we are the king's horses and king's men in charge of putting camelia back together again. 20:08
timotimo wait a second
TimToady okay, I waited a second
labster I'm not sure who the king actually is. Maybe King Kong.
timotimo aaw, masak, i won't be able to make it to the workshop :( 20:09
masak timotimo: :(
TimToady The question is, who's to be master, that's all.
masak timotimo: please update github.com/perl6/mu/wiki/perl6-wor...-june-2013
20:09 kaare_ left
timotimo already done 20:10
masak TimToady: or, if you're in the Rakudo repository, who's to be nom, that's all.
timotimo++
sorear andreas koenig?
masak Simba?
TimToady en.wikipedia.org/wiki/The_King 20:13
labster added myself to the new masakism list 20:14
masak labster++ \o/
we're a smaller group this time around, but that means you'll each be given more attention :)
labster I know, I might even be able to ask detailed questions this time 20:16
masak looking forward to that :) 20:17
20:17 tgt joined
labster I think I have insufficient hubris as a programmer. Does anyone care if we return path objects from mkdir(), rename(), symlink(), etc. instead of True? (Still Failure on failure, obviously.) 20:19
TimToady use case? 20:23
labster "t/spec".path.mkdir.chdir
TimToady I guess that one is apparent to the reader, but symlink.chdir wouldn't be so clear 20:24
labster perhaps not on those.
I think rename returning the new file would be clear enough. 20:25
20:30 zby_home_ left
labster Symlink takes an argument, but I haven't really determined the exact interface there. Should it always take its argument as the destination path, or require something like .symlink(from=>"foo") or .symlink(to=>"bar") 20:30
ssutch .symlink(from=>"foo") or .symlink(to=>"bar") both seem to make sense to me 20:32
sorear .rmdir.chdir 20:33
TimToady \o/
symlink is not done so often that it needs multiple ways to do it
ssutch .rm.chdir would just fail, no such path or whatever 20:34
TimToady I'd stick with the POSIXy interface
labster just the function then, rather than a class interface.
lue
.oO(does &symlink fail on Windows/other systems without the magic of symlinks?)
 20:35
TimToady I'd add it to the class only if we can guarantee some greater portability that way
otherwise it falls into the category of a foolish consistency 20:36
masak +1 on POSIXy interface. 20:37
20:37 ssutch left
benabik NTFS has symlinks, IIRC 20:37
masak I'd much rather see cute method chaining experiments in user-provided modules.
sorear now that masak has said it, I agree 20:39
20:39 panchiniak left
TimToady chained mutators are not very P6ish 20:39
20:40 tgt left
TimToady P6 likes to keep its mutators at the statement level most of the time 20:41
20:41 ssutch joined
TimToady expressions are more for FP stuff 20:41
TimToady thinks side effects should look like assembly language :) 20:42
labster Well, that sounds good. I already have a module where I put all of my crazy ideas. 20:43
TimToady assignment being the notable exception...
20:43 tomyan joined
TimToady of course, someone coming from other backgrounds such as Ruby may have different expectations... 20:44
colomon TimToady: which is exactly why .match should not set $/
TimToady wishes .subst had a way of injecting $/ into a closure... 20:48
masak too
TimToady that's really the only reason I'd hold out for setting $/ on the method forms
sorear > 1 cmp 10 20:50
Increase
\o/
colomon \o/
20:50 vk left, xilo1 joined
colomon is okay with requiring subst users to make it -> $/ { ... } 20:51
sorear still no [1..10] 20:52
20:52 xilo1 left
dalek p: 38a8b97 | sorear++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
Add return value coercion support for jvm
 20:52
p: 6ba8f35 | sorear++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
Add experimental nqp::backtrace for structured access to jvm backtraces
dylanwh NTFS symlinks have oddly specific limitations that make me sad 20:53
sorear o/ dylanwh 20:54
20:54 grondilu joined 20:57 Juerd left, tgt joined
lizmat wonders whether TimToady has any units on his mind 20:57
20:58 betterworld left
sorear the java debugger likes to put commas in line numbers. [4] E940929BF74D90FB0CDF81E1E7A217FDFBD39661.qb_3637 (src/gen/CORE.setting:3,963) # this really freaks me out, and I can't quite put a finger on why 20:58
[Coke] why it does or why it freaks you out? 20:59
sorear why it bothers me so
20:59 konundra left
dylanwh fwiw, NTFS supports both hardlinks and symlinks. The limitations are that symlinks have a short path, and only administrator-types can make them. 20:59
lizmat suspects a European background in sorear
in Europe, the comma is a decimal point 21:00
dylanwh and you can only have 2**10 links to the same inode.
grondilu commas are the standard 10^3 separator in English, aren't they?
lizmat well, before we had computers and the American world view about decimal *points* forced upon us :-) 21:01
21:01 betterworld joined
sorear
.oO( line 3_963 )
 21:01
21:01 Juerd joined
grondilu maybe you want to install a "perl locale" or something :-) 21:02
sorear multi postfix:<-->(Num:D \a is rw) { # New failure mode: This is being called on an argument of type Int. What? Naturally, nqp::unbox_n is not happy. 21:03
21:04 btyler left
lizmat pmichaud: wrt extra parameter to classify|categorize: that's now moot with these methods being callable on hashes 21:05
21:05 btyler joined 21:07 snearch joined
pmichaud lizmat: (moot) excellent, I was hoping that would be the answer. :-) 21:14
I like the methods-callable-on-hashes approach.
that feels much more natural than the other ones.
21:15 pmurias joined
pmichaud (afk again) 21:15
dalek ast: b9eaf06 | (Elizabeth Mattijsen)++ | S32-list/categorize.t:
Some more categorize test refinements
 21:23
21:28 skids left 21:31 spider-mario left
dalek ecs: 70bad12 | (Elizabeth Mattijsen)++ | S32-setting-library/Containers.pod:
Fix typo
 21:35
21:41 dmol left
dalek ast: 95ba4f6 | (Elizabeth Mattijsen)++ | S32-list/classify.t:
Added some more tests when using typed hashes
 21:42
21:45 preflex left
TimToady nr: say '▁' ... '█' 21:45
camelia rakudo b2072f: OUTPUT«(timeout)» 21:46
..niecza v24-77-g2b14288: OUTPUT«▁ ▂ ▃ ▄ ▅ ▆ ▇ █␤»
TimToady rakudobug
lizmat submits rakudobug
21:46 tomyan left, Chillance left 21:47 preflex joined, ChanServ sets mode: +v preflex
dalek ast: 1b82ada | (Elizabeth Mattijsen)++ | S32-list/classify.t:
Some additional tests for classify, like categorize
 21:47
masak lizmat++ 21:48
21:49 PacoAir left 21:52 tomyan joined 21:56 tomyan left
lizmat is going to test the state of her own bed& 21:57
22:01 pochi left, pochi joined
sorear ranges work in the repl, because the repl is installing an exception handler that appears to resume everything. 22:05
22:09 bluescreen100 joined 22:13 bluescreen10 left 22:18 benabik left, benabik joined 22:20 bluescreen100 left
sorear oh, it's not a MMD problem, rakudo-jvm just doesn't like nqp::sub_n(foo, 1) very much 22:20
rakudo: nqp::unbox_n(1) 22:22
camelia rakudo b2072f: OUTPUT«This type cannot unbox to a native number␤ in block at /tmp/Njc1WlNvN9:1␤␤»
sorear rakudo: nqp::sub_n(1,2)
camelia rakudo b2072f: ( no output )
timotimo r: '▁'.succ.say 22:23
camelia rakudo b2072f: OUTPUT«▁␤»
timotimo r: '▁'.succ.succ.say
camelia rakudo b2072f: OUTPUT«▁␤»
timotimo r: '▁'.succ.succ.succ.say
camelia rakudo b2072f: OUTPUT«▁␤»
22:24 rindolf left
timotimo ah, that's how it works 22:24
sorear should that work? should I replace 1 with 1e0 in the setting, modify nqp::sub_n to coerce, or something else? 22:25
could also perhaps use Want on num literals
TimToady rosettacode.org/wiki/Sparkline_in_unicode is the page in question, btw
timotimo: it doesn't work by calling .succ on that
timotimo the range reify method uses .succ 22:26
succ looks into RANGEPOS to find out if it's something it should be counting up. if it isn't, it no-ops.
should i make it to work just like niecza if it's not part of a carry-able range thingie? 22:27
TimToady see S03:2024
it's not the range operation, it's the sequence operator, to begin with 22:28
timotimo oh!
derp :)
TimToady and the sequence operator treats single codepoints specially
(or it's supposed to...)
sorear > [1..10] 22:37
1 2 3 4 5 6 7 8 9 10
>
22:37 shinobicl left
TimToady -> bed & 22:42
sorear rakudo: nqp::print(nqp::getstderr(), "moo") 22:43
camelia rakudo b2072f: OUTPUT«===SORRY!===␤Error while compiling block (source text: "nqp::print(nqp::getstderr(), \"moo\")"): Error while compiling op call: Error while compiling block : Error while compiling op print: Operation 'print' requires 1 operands, but got 2␤»
sorear rakudo: nqp::printfh(nqp::getstderr(), "moo")
camelia rakudo b2072f: OUTPUT«moo»
22:48 census left 22:57 skids joined
dalek p: aefda07 | sorear++ | src/vm/jvm/runtime/org/perl6/nqp/runtime/Ops.java:
Allow nqp::join to work on native str arrays on JVM
 22:59
sorear backtrace printer is now working. ish./ 23:01
23:01 abnorman left
sorear working well enough to not eat the error messages 23:01
although it turns everything into a ===SORRY!=== and eats the backtraces
23:01 abnorman joined
masak sounds wonderful :P 23:02
benabik ===SORRY!=== I made you an error, but I ate it.
sorear sorear/rakudo/nom if anyone is interested. will have CLA in the mail in time for tomorrow's pickup 23:04
timotimo suffers from a very typical computer science and programming ailment: 23:07
too many layers of indigestion :(
masak lasagna for dinner, eh? :)
timotimo that would have been lovely; except perhaps for the fact that it was between 30 and 35 degC out today 23:10
i think i shouldn't have eaten mostly ice creat today
is anybody working on the "one codepoint sequence" thing right now or should i perhaps give it an actual shot? 23:17
23:20 BenGoldberg left 23:24 pmurias left 23:25 Ayiko left
timotimo oh, there's actually succ-with-carry semantics for multi-character strings 23:25
23:26 btyler left
masak yes, but less than there used to be, IIRC. 23:30
timotimo oh, i think i have misread actually
23:31 konundra joined
timotimo is my new & revised reading that it'll only ever happen if both end points are single-codepoint-strings correct? 23:31
masak rn: say "w".."ba" 23:35
camelia rakudo b2072f, niecza v24-77-g2b14288: OUTPUT«"w".."ba"␤»
masak rn: .say for "w".."ba"
camelia rakudo b2072f, niecza v24-77-g2b14288: ( no output )
masak maybe :) 23:36
rn: .say for "w".."ac"
camelia rakudo b2072f, niecza v24-77-g2b14288: ( no output )
23:36 jlaire left
timotimo don't you want ... instead 23:37
?
rn: say "w"..."ba"
camelia rakudo b2072f: OUTPUT«Decrement out of range␤ in method Str at src/gen/CORE.setting:10160␤ in method Str at src/gen/CORE.setting:876␤ in method Stringy at src/gen/CORE.setting:885␤ in sub infix:<eq> at src/gen/CORE.setting:1288␤ in sub infix:<eq> at src/gen/CORE.setting:1286␤ in m…
..niecza v24-77-g2b14288: OUTPUT«Unhandled exception: Magical string decrement underflowed␤ at <unknown> line 0 (KERNEL Str.pred @ 1) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3340 (ANON @ 4) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3374 (ANON @ 3) ␤ at /home/p6eval/niecza/…
timotimo hehe.
23:37 jlaire joined 23:39 snearch left
pmichaud .u ▁ 23:41
yoleaux U+2581 LOWER ONE EIGHTH BLOCK [So] (▁)
pmichaud .u █
yoleaux U+2588 FULL BLOCK [So] (█)
lue r: say NaN.Int; say Inf.Int; 23:42
camelia rakudo b2072f: OUTPUT«0␤Cannot coerce Inf to an Int␤ in method gist at src/gen/CORE.setting:10161␤ in method gist at src/gen/CORE.setting:893␤ in sub say at src/gen/CORE.setting:11047␤ in block at /tmp/l04qPlvnZA:1␤␤»
pmichaud say '▁' cmp '█'
r: say '▁' cmp '█'
camelia rakudo b2072f: OUTPUT«Increase␤»
23:42 nitestryker left
pmichaud r: say '▁' .. '█' 23:42
camelia rakudo b2072f: OUTPUT«"▁".."█"␤»
pmichaud r: say ('▁' .. '█').list
camelia rakudo b2072f: OUTPUT«(timeout)» 23:43
pmichaud r: say ('▁' ... *).[^5]
camelia rakudo b2072f: OUTPUT«▁ ▁ ▁ ▁ ▁␤»
pmichaud r: say ('▁'.succ)
camelia rakudo b2072f: OUTPUT«▁␤»
pmichaud r: say ('▁'.succ).ord 23:44
camelia rakudo b2072f: OUTPUT«9601␤»
pmichaud r: say '▁'.ord
camelia rakudo b2072f: OUTPUT«9601␤»
pmichaud say "\x65".succ
r: say "\x65".succ
camelia rakudo b2072f: OUTPUT«f␤»
pmichaud r: say "\x2581".succ 23:45
camelia rakudo b2072f: OUTPUT«▁␤»
pmichaud weird.
rjbs r: say (1 cmp 2).^WHAT
camelia rakudo b2072f: OUTPUT«===SORRY!===␤Cannot use .^ on a non-identifier method call␤at /tmp/z_j6wJyruS:1␤------> say (1 cmp 2).^WHAT⏏<EOL>␤ expecting any of:␤ method arguments␤»
rjbs r: say (1 cmp 2).WHAT
camelia rakudo b2072f: OUTPUT«(Order)␤»
rjbs Huh. I had not yet seen that type.
pmichaud Order is an enumeration.... Increase, Same, Decrease
rjbs Cool.
<pmichaud> Yes, Order is Cool.
pmichaud r: say Order ~~ Cool 23:46
camelia rakudo b2072f: OUTPUT«True␤»
rjbs :)
Yow, my rakudo will no longer run. 23:47
rjbs builds a new one!
23:47 BenGoldberg joined
lue I'm suddenly <stunned perplexed>.all by the fact that !Num.ACCEPTS(Int) . That feels wrong (coming from the point of view of math). 23:50
lue guesses someone will show him why that's The Right Way™ for programming
sorear lue: Num reflects an implementation, not a value set 23:51
when I say my Num $x I'm saying "$x points to a number stored as floating-point" 23:52
pmichaud r: say so Numeric.ACCEPTS(Int)
camelia rakudo b2072f: OUTPUT«True␤»
sorear it can't point to a number stored as integer, or decimal string
more practically, the multidispatch optimizer needs a representation guarantee in order to avoid coercion checks 23:53
timotimo i won't finish the sequence thing today.
lue OK. I think I conflated Num and Numeric then.

Powered by Cro and the Raku® Programming Language.

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

Thank you!