»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:00 tinyblak left 00:03 b2gills1 is now known as b2gills 00:07 Gardner joined 00:11 Gardner left 00:12 Gardner joined 00:18 jepeway left 00:32 cognominal left 00:37 mr_ron left 00:41 mr_ron joined 00:57 elimik31 joined 01:06 aborazmeh joined, aborazmeh left, aborazmeh joined 01:16 elimik31 left, elimik31 joined
colomon Any ides what would generate a “Cannot invoke null object” error? error is occurring in a simple script which has worked for months… 01:21
*ideas
01:22 mr_ron left
colomon Argh! can’t install panda on 2015.04 because it doesn’t have unit. :\ 01:34
01:42 Gardner left, elimik31 left
colomon nor 2015.05 01:44
timotimo i would have mildly preferred putting in the semicolon form instead 01:51
at least for the time being
01:52 gfldex left
colomon semicolon form 01:55
?
wouldn’t a working patch be to use the { } form instead of unit? 01:57
PS apparently it’s not panda, it’s File::Find, Shell::Command, and JSON::Tiny
timotimo: any notion if the POD at the end of a module should be included in the { } 02:06
?
timotimo if it's not a #| or #= i don't think it matters 02:14
er, sorry, i meant the braces form of course
02:15 bjz left 02:17 kurahaupo1 joined 02:18 kurahaupo1 is now known as kurahaupo, yqt left 02:19 tinyblak joined
colomon Guess if I had permission to change… huh, guess I do have commit on JSON::Tiny 02:22
02:23 nys left 02:25 Sqirrel left 02:26 noganex joined 02:28 noganex_ left 02:30 cognominal joined, khisanth_ joined, Khisanth left
colomon Argh, is there something very wrong with rakudo causing these Cannot invoke null object messages? 02:31
colomon tries clean build
02:33 Sqirrel joined
timotimo hoelzro put the fix in, but we didn't bump nqp and rakudo yet 02:33
colomon well… if you can point me in the right direction, I’d be happy to do the legwork 02:34
(the fix is in moar, I take it?)
timotimo yeah
git describe --tags inside MoarVM's checkout will give you what to put into nqp's tools/build/MOAR_REVISION 02:35
the same dance inside nqp - after commiting the revision bump - gives you what to put into rakudo's tools/build/NQP_REVISION
02:37 AlexDaniel left 02:39 rmgk left 02:40 rmgk joined
dalek p: 436f66b | colomon++ | tools/build/MOAR_REVISION:
Use hoelzro++'s latest MoarVM patch.
 02:41
timotimo it might be a good idea to give a slightly more elaborate explanation 02:42
like "get the fix for null object invocations'
damn, mismatched quotes
colomon will do in the rakudo checkin 02:43
hoelzro timotimo: now you tell me =)
timotimo i tell you what exactly? the thing about the bump?
i wonder what else got bumped into this
oh
literally just the one patch
colomon yeah 02:44
almost done
hoelzro timotimo: about a better idea for a commit message
I linked to RT, but I should've included the error message 02:45
timotimo oh
the rt link is a good place to start
colomon bother, I’m having trouble with git.
hoelzro colomon: what kind of trouble?
timotimo hmm, maybe a userscript that links from github repositories connected to perl6 to rt.perl.org for those would be interesting?
a browser userscript, that is
colomon hoelzro: made my commit to rakudo, but when I push it just says Everythng up-to-date 02:46
hoelzro timotimo: that would be handy
timotimo colomon: that's very strange :)
colomon oh, I see
timotimo hoelzro: i don't know how to userscript :)
hoelzro timotimo: that makes two of us
;)
dalek kudo/nom: 7fc0782 | colomon++ | tools/build/NQP_REVISION:
Use hoelzro++'s latest MoarVM patch.

Get the fix for null object invocations.		 02:47
colomon wasn’t on nom
timotimo what branch were you on, then?
it would have had to push something, right?
colomon no branch
hoelzro detached HEAD, then? 02:48
colomon guess
timotimo ah
right 02:49
02:50 firefish5000 joined
skids After a stint of adding --configure-opts to "rakudobrew triplet" I was able to verify that that fix fixed Sum. 02:55
timotimo god damn it, all i do these days just breaks moar 02:56
hoelzro timotimo: could've been worse
timotimo how?
colomon hoelzro++ timotimo++ # update fixed my issue, now I can get back to $work 02:57
hoelzro also, we have to risk stability a bit to gain speed
timotimo i stole multiple hours of productivity from people, i bet
hoelzro timotimo: well, it doesn't block installation entirely, for one
timotimo if it had, i could have found it out immediately
hoelzro it could've been an awful GC, concurrency, or serialization bug
that random segfaults rakudo 02:58
and been *really* hard to track down
it took me less than 5 minutes to write a bisect script to find that commit
02:58 xinming left
hoelzro and I played bloodborne while it worked =) 02:58
timotimo something sloppy i recently did ended up sending jnthn on a wild goose-chase :\
02:59 xinming joined
colomon and my $work patch worked brilliantly. (p6 checks the test results.) 02:59
hoelzro timotimo: we've all made mistakes; it's better than no one contributing at all, right! 03:00
03:00 cognominal left
skids timotimo: stop beating yourself up, you contribute more than me or many others. 03:00
timotimo hm, may be 03:01
things are getting really real right now
with perl6 i mean
aint such a long time until christmas
jdv79 what does really real smell like? 03:02
hoelzro yeah, I know what you mean
September isn't far away
and that's *beta* time
jdv79 6.0.0 or 6a or whatever it'll be called will bomb inevitably, right?
timotimo tiny annoyances are piling up around me at the moment, so i'm a tiny bit on edge it seems 03:03
jdv79 i wouldn't be too worried about it
hoelzro I look at all the tickets in RT and get worried
timotimo at least there's a lot of churn in RT
hoelzro yeah
timotimo new tickets being added, but also tickets getting closed
hoelzro I feel like there are a lot of open tickets that are probably no longer valid
timotimo much better than having a permanently low number of tickets that don't get changed at all
jdv79 if 6b or 6.1.0 has serious troubles then i'll be concerned 03:04
timotimo i'm heading towards bed now
see you :)
hoelzro later timotimo 03:05
hoelzro goes to play Takenoko with wife and brother-in-law
03:17 steve_mc left 03:22 Celelibi left 03:23 Celelibi joined 03:31 steve_mc joined
skids phys.org/news/2015-06-year-old-algo...roven.html # summary: anything that solves edit distance in under N² would make NP-complete less hard than generally suspected. 03:38
03:43 steve_mc left 03:52 bjz joined 03:56 steve_mc joined 03:59 bjz left 04:02 Alina-malina left 04:09 telex left 04:10 telex joined 04:23 BenGoldberg left 04:30 bjz joined 04:36 davido_ left 04:38 davido_ joined
hoelzro timotimo: first cut of that userscript: gist.github.com/hoelzro/5c16c12baa9dc951511e 04:46
it *should* work
04:49 aborazmeh left 04:58 aborazmeh joined, aborazmeh left, aborazmeh joined 05:03 aborazmeh left 05:32 davido__ joined 05:34 davido_ left 05:39 bjz left 05:40 bjz joined 05:41 firefish5000 left 05:45 bjz left, bjz joined 05:46 Patterner joined 05:50 Psyche^ left 05:54 bjz left, gerry__ joined, bjz joined 05:56 xinming_ joined 06:00 xinming left, FROGGS[mobile] joined
FROGGS[mobile] my plan for today is to make this work: 06:01
my $cmd1 = run('cmd1', :out); run('cmd2', :in($cmd1.out)) 06:03
06:04 firefish5000 joined
FROGGS[mobile] in theory I only have to tweak run() in rakudo in the openpipe3 branch 06:05
skids That would be nice. 06:06
FROGGS[mobile] but now I wonder: is feed syntax usable here?
it also means I deprecate pipe() 06:07
skids I was under the impression that feed syntax is mosty intended to be a List thing.
FROGGS[mobile] ahh
and I should not forget to ask the libuv devs about how merging stdout/sdterr is supposed to work 06:11
if that is a low level thing at all 06:12
skids I believe in C it is done with freopen or something.
FROGGS[mobile] hmmm, I can grep for that in libuv's source 06:13
06:18 aborazmeh joined 06:19 aborazmeh left, aborazmeh joined 06:20 bjz left 06:21 bjz joined 06:26 FROGGS[mobile] left, FROGGS[mobile] joined 06:28 gerry__ left
raydiak S32-IO says pipe() no longer exists anyway and is replaced by IO::Pipe.pair 06:28
06:29 FROGGS joined
skids Ah, if no streams are being used the way to deal with it is dup/dup2 06:32
(in C)
06:33 gerry__ joined
FROGGS raydiak: yes, which also won't be true in future 06:35
what I imagine is that IO::Pipe will we a filehandle, returned by something that calls nqp::syncpipe() 06:36
and run() and shell() return a (IO::)Proc object, that has .in, .out and .err attributes of type IO::Pipe
skids Does run() start the cmd right away? Because then you have some time when cmd1 can output before cmd2 starts. 06:38
raydiak maybe it's buffered? 06:39
skids That would be overcomplicated.
versus starting cmd2 first and then later attaching cmd1 when you start it. 06:40
FROGGS now reads MoarVM/3rdparty/libuv/src/unix/process.c:267:static void uv__process_child_init(
skids: run() does not start when you capture any of <in out err> 06:41
but it does kick off when you dont capture anything
skids Ah, good.
FROGGS so my posted example might not work
06:42 aborazmeh left, domidumont joined
FROGGS so you either have to also print to the first command from P6, and close that handle, or read from cmd2 and close its handle 06:42
we'll see
raydiak also don't forget about the cases where people want to run pipelines of several commands 06:44
FROGGS raydiak: example?
(not a use case, but pseudo code) 06:45
raydiak um...well idk at midnight off the top of my head after a day like today, but like any shell script with more than one '|'
erm, not shell script but one-liner if you get my meaning 06:46
FROGGS raydiak: yes, that should work IMO
I will know today hopefully :o)
raydiak cool :) ++FROGGS
but off to bed for me...busy day and week, thankfully...just dropped in to say "Go team Camelia!" before sleep 06:47
FROGGS gnight raydiak :o) 06:49
06:49 aborazmeh joined, aborazmeh left, aborazmeh joined 06:50 domidumont left 07:02 kaare__ joined 07:08 lolisa joined 07:13 domidumont joined 07:14 RabidGravy joined 07:17 Sqirrel left 07:23 Akagi201 joined 07:32 aborazmeh left 07:33 skids left, lolisa left 07:44 Foxcool_ joined 07:45 Foxcool left, Foxcool_ left, Foxcool joined 07:47 espadrine joined
RabidGravy okay, so having fixed my bogus test data which led me off totally down the wrong path, theres definitely something funny going on with CArray[num64] that I can't see: 08:00
Adding -191.709945678711 at 57
Wrong kind of access to numeric CArray
in method ASSIGN-POS at lib/NativeCall.pm:333
can't replicate outside my code at the moment 08:03
FROGGS RabidGravy: does MVM_SPESH_DISABLE=1 help? 08:04
08:05 Alina-malina joined, darutoko joined
RabidGravy yes, yes it does 08:05
what does that do then? 08:06
FROGGS[mobile] disable dynamic optimizations and also the JIT 08:08
that explains why cant easily reproduce in a one liner... 08:09
the code must be "hot" in order to get speshed
RabidGravy ah, so that isn't really a solution going forward then
FROGGS[mobile] you can try wrapping it a while loop 08:10
no, but we need to golf it in order to fix it
in a while loop* 08:12
08:12 gerry__ left
RabidGravy while loop? my $f = True; while $f { # do the copy ; $f = False } doesn't help 08:15
the problem with replicating is that the data is an array of Nums which have already been copied from a CArray[num64] 08:16
FROGGS[mobile] then this needs to be part of the golf 08:23
RabidGravy er 08:26
'ang on, I don't think that'll work 08:27
08:30 fractionuser joined, fractionuser left
RabidGravy just about as short as I can get it: gist.github.com/jonathanstowe/9802...f7a6f992dc 08:36
it only every does it in the third loop and always does at the 58th element 08:37
I guess there may be opportunities to simplify the data 08:41
also should be noted that it only does it with CArray[num64] - not int16, int32 or num32 08:44
08:49 bjz left, diana_olhovik joined 08:54 gfldex joined, bjz joined 08:57 bjz left 08:59 cognominal joined
RabidGravy ah, no the data can be replaced with a "my @in = (^100).map({Num($_)});" 09:02
09:04 lolisa joined
RabidGravy FROGGS, I think that's as short as I can make it and still demonstrate that behaviour gist.github.com/jonathanstowe/9802...f7a6f992dc 09:04
I think I'm going to TODO the tests it is causing to fail, move on and RT that later unless someone has any further input 09:09
09:19 [Sno] left 09:20 rindolf joined
nine RabidGravy: it seems like your bug got introduced sometime within the last month (since commit 31f3db1) 09:23
RabidGravy: could only reproduce it after updating rakudo
RabidGravy Hmm
cheers for looking
FROGGS RabidGravy++ # that golfed script is good 09:25
RabidGravy the fact that it only happens on the third generation helps ;-) 09:26
09:28 lolisa left
nine bisecting... 09:28
RabidGravy I looked at the site of the exception and couldn't see anything obvious except to conclude that somewhere it was losing its inner-numiness 09:31
FROGGS nine++ 09:33
09:34 bjz joined 09:40 gerry__ joined 09:42 gerry__ left 09:52 [Sno] joined 09:53 bjz left
RabidGravy while I'm thinking of it, I think I'll add some tests for the "rw CArray getting populated by the library function" case in the NativeCall tests 09:53
as there doesn't appear to be any 09:54
09:56 domidumont left
nine d356462fed5706dd8fa6394dec26d1cced034cdb is the first bad commit, Author: Tobias Leich, "implement ignoremark, though needs tbd for interpolations" 09:58
Doesn't make much sense though :/
Oh but this commit bumps NQP version, so it may make sense after all 09:59
dalek c: 0de156d | paultcochrane++ | lib/Type/SetHash.pod:
Improve link name to Set/Bag Operators section
10:11 [Sno] left 10:12 [Sno] joined
FROGGS nine: I'd guess the problem is introduced in MoarVM 10:20
10:21 molaf joined
nine FROGGS: indeed. I should be able to bisect MoarVM without having to recompile Rakudo, shouldn't I? 10:27
FROGGS nine: correct
RabidGravy yeah the site of the exception is actually in the moar code 10:31
10:32 spider-mario joined
dalek c: 2ee4b15 | paultcochrane++ | lib/Language/ (2 files):
Fix link from classtut to Roles section in objects doc
 10:33
nine Except that nqp doesn't seem to like the stage0 changes 10:43
FROGGS nine: yes, there are changes where you have to rebuild nqp/rakudo, but these are rare 10:44
nine Unfortunately this is one of those changes. And when I rebuild Rakudo, it insists on checking out a specific MoarVM version making the bisect run useless. 10:55
colomon nine: that’s why I wrote the “triple” functionality into rakudobrew. Though you still have to do it by hand. 11:00
dalek c: f3be1d0 | paultcochrane++ | lib/Type/BagHash.pod:
Use a better link name to Set/Bag Operators section
 11:02
c: f25a270 | paultcochrane++ | type-graph.txt:
Add new exceptions to the type-graph
nine colomon: except if I remove the --gen-moar and --gen-nqp from config.status I guess 11:03
colomon nine: I meant, bisect by hand. the triple thing automatically builds the versions of rakudo, nqp, and moar you specify 11:04
I was working on just this sort of bug at the time.
FROGGS $ perl6 -Ilib -e 'use IO::Pipe; my $p1 = shell("dir", :out); my $p2 = shell("wc", :in($p1.out), :out); say $p2.out.lines' 11:06
110 219 3981
\o/
11:07 FROGGS_ joined
itz_ hmmm I'm a little sceptical about static analysis but "scan-build -- Clang static analyzer" looks interesting 11:08
11:11 FROGGS left
RabidGravy FROGGS, nice :-) 11:12
11:15 FROGGS_ is now known as FROGGS
FROGGS is happy 11:15
vendethiel
.oO( don't forget to water your FROGGS )
 11:19
11:24 rindolf left
dalek line-Perl5: 035eb53 | (Stefan Seifert)++ | / (2 files):
Support overloaded stringification of P5 objects

In Perl 5: package Foo; use overload '""' => sub { "foo!" }; In Perl 6: say $foo.Str(); # gives "foo!"		 11:37
vendethiel nine++
nine hoelzro++ for suggesting this in GH #12 11:38
11:44 spider-mario left, JimmyZ_ joined
JimmyZ_ nine: line 21 in test looks like useless? 11:45
11:46 gerry__ joined
nine JimmyZ_: indeed! 11:54
dalek line-Perl5: 9075f90 | (Stefan Seifert)++ | t/overload.t:
Remove left over debug code from test

Thanks to JimmyZ_++ for spotting this!
11:54 spider-mario joined 11:55 FROGGS left 11:57 JimmyZ_ left, nys joined 12:00 FROGGS joined
ab5tract FROGGS[mobile]: is that IO::Pipe behavior lazy? 12:03
timotimo hoelzro: you could check synbot6's code; it has an extra heuristic in it to make the "RT" in front not required and still not give terribly many false positives 12:07
12:07 [Sno] left
timotimo hoelzro: and you're replacing text inside the html, which may end up inside an attribute or something ... :\ 12:11
12:12 JimmyZ_ joined
RabidGravy Hmm, when you do @array1 ~~ @array2, what does it actually do? I have two arrays that look pretty similar but can't tell where the differ 12:13
psch RabidGravy: "what exactly" is in the ACCEPTS declaration — looks like a Whatever-aware deep comparison, on a glance 12:15
m: say [*] ~~ [1]
camelia rakudo-moar 7fc078: OUTPUT«5===SORRY!5=== Error while compiling /tmp/nTbJ2E_LUZ␤Expected a term, but found either infix ~~ or redundant prefix ~␤ (to suppress this message, please use a space like ~ ~)␤at /tmp/nTbJ2E_LUZ:1␤------> 3say [*] ~~7⏏5 [1]␤»
timotimo ~~ of arrays uses === to compare individual elements
psch m: my @a = *; my @b = [1]; say @a ~~ @b
camelia rakudo-moar 7fc078: OUTPUT«False␤»
psch m: my @a = *; my @b = [1]; say @b ~~ @a
camelia rakudo-moar 7fc078: OUTPUT«True␤»
psch +Whatever handling for the RHS
timotimo and it skips over whatever stars, like psch is demonstrating
psch so first-level identity, not deep comparison 12:16
timotimo++
m: my $a = 1; say $a === *; 12:23
camelia rakudo-moar 7fc078: OUTPUT«WhateverCode.new␤»
psch ...right
timotimo :P
12:31 AlexDaniel joined
FROGGS[mobile] ab5tract: fsdo lazy, yes 12:46
12:49 zakharyas joined 12:55 diana_olhovik left
FROGGS[mobile] ab5tract: you get filehandles you can read/write lines from/to 12:59
12:59 aborazmeh joined, aborazmeh left, aborazmeh joined
ab5tract The IO docs could use some love, but I'm not quite sure where lizmat++ is at with newio 12:59
13:01 bin_005 joined, aborazmeh left
ab5tract FROGGS[mobile]: I assumed so, but wanted to be sure. 13:01
psch: does that behavior surprise you? 13:02
13:02 JimmyZ_ left
psch ab5tract: which? Whatever currying ===? 13:02
no, i'm just a bit rusty :P 13:03
ab5tract yeah
:)
13:03 aborazmeh joined, aborazmeh left, aborazmeh joined
ab5tract I recently doc'd the behavior of Whatever and WhateverCode WRT to Whatevers stored in variables, so it's fresh in my mind 13:03
psch i'm also mostly cognitively loaded with trying to figure out how to invoke a Rakudo-level CompilationUnit... 13:04
13:04 coffee` left
psch 'cause that's the first step and having written-as-perl6, self-executing jars 13:04
s/and/in/
and i'd really like to be able to write perl6 and ship it to users who don't have rakudo-j installed 13:05
ab5tract That would be amazing
psch well, it'd still be huge jars
like, whole runtime of course
13:05 coffee` joined
psch although maybe there's a way to trim based on coverage for class files... 13:06
RabidGravy people with bigger brains than me: is there something about the number representation for Num that might cause it to be less accurate ±0.5 ? 13:07
around ^
psch not less accurate than IEEE754 in general, i'd hope 13:08
colomon RabidGravy: context? it’s basically just a standard floating point double. 13:09
RabidGravy I've got an array of 50 Nums and writing it into a wav file using libsndfile and then reading it back it's only the few frames around the ±0.5 around the third or fourth decimal place 13:10
that differ
13:10 aborazmeh left
moritz so this is not about arithmetics, but about (de)serialization 13:10
RabidGravy dunno, that's why I ask 13:11
it's six out of fifty which seems strange
moritz it's hard to tell without knowing how you do the comparison, input and output
iirc the default stringification loses precision 13:13
RabidGravy er, there is no stringification
13:13 domidumont joined
RabidGravy @array of nums -> CArray[num64] -> sf_writef_float() -> sf_readf_float -> CArray[num64] -> @array of nums 13:14
then comparing the starting array and the finishing array values numerically 13:16
13:20 domidumont left
moritz if you swap the "sf_writef_float() -> sf_readf_float" steps out for just copying in-memory, do you still get that behavior? 13:20
13:23 bin_005 left
vendethiel m: given "foobar" { say "it's $_'s turn" } 13:25
camelia rakudo-moar 7fc078: OUTPUT«5===SORRY!5=== Error while compiling /tmp/x5HFuntxNY␤Variable '$_'s' is not declared␤at /tmp/x5HFuntxNY:1␤------> 3given "foobar" { say "it's 7⏏5$_'s turn" }␤»
vendethiel yeaaah. about that :P
vendethiel uses $($_) instead
colomon would prefer {$_} in his code 13:28
13:31 zakharyas left
arnsholt $($_) has slightly different semantics, too, doesn't it? 13:32
RabidGravy: It's odd that you're getting rounding problems. How exactly are you computing your values and writing them to file? 13:33
13:35 BenGoldberg joined
RabidGravy arnsholt, I'm not computing them in the sense I took an existing WAV file read 25 frames into an array and printed the values to construct some test data for each of short, into, float and doubles and attempt to round-trip them through the appropriate libsndfile functions 13:37
arnsholt Right. So you read some stuff in with libsndfile, and then write them back out again with the same lib. So this is mostly NativeCall code, I assume? 13:38
RabidGravy er yes 13:39
arnsholt That's odd
RabidGravy I'm actually beginning to wonder whether the particular pattern I am seeing is an artefact of the dithering strategy libsndfile uses 13:40
colomon what is the range of values of the nums?
moritz and, uhm, don't wav files store integers? 13:41
RabidGravy yes, hence "the dithering strategy"
they don't actually have to store integers 13:42
the nums are in the range of ±1 13:43
let me try this with a floating point format, maybe I'm asking too much
moritz TIL
13:49 gerry__ left
colomon yeah, it definitely seems like it is very weird to have ±0.5 be the problem if the range is ±1 13:50
RabidGravy ignore me, there may be some issue somewhere but using a float format it gets it right
colomon Why don’t you forget libsndfile and just code up simple C to store a num64 array and let you retreive it later… 13:51
13:51 gerry__ joined
RabidGravy because libsndfile can open, read and write some 26 formats of sound files in 24 different encodings 13:53
hoelzro o/ #perl6
colomon RabidGravy: no, no, not to skip libsndfile as a goal, just to skip it and see if the nums transfer correctly. 13:54
13:54 domidumont joined
colomon ie rakudo problem or libsndfile problem/misuse. 13:54
ie *is it a rakudo...
psch o/ hoelzro 13:55
hoelzro timotimo: good points all around; I required the RT because GH will link #\d+ to its own issues
o/ psch
13:55 gerry__ left 13:58 davido__ left
timotimo ah, good point 13:58
but our RTs start somewhere in the thousands
13:59 davido__ joined 14:00 diana_olhovik joined
hoelzro true 14:00
14:01 Ven joined 14:02 gerry__ joined
hoelzro so maybe \d{3,}, then? 14:02
14:03 yqt joined
hoelzro heck, the lowest I've seen are 5 digits 14:03
14:04 gagalicious joined 14:05 diana_olhovik left
RabidGravy colomon, I've concluded that because when I write the same floats to a *float encoded WAV* and read them back as floats it comes back exactly the same that isn't rakudo's problem 14:06
I'm just either being thick or expecting to much of something or both 14:07
14:08 diana_olhovik joined
timotimo remember that float is 32bit and our num are 64bit 14:12
... or is that what you're already saying?
RabidGravy yep, the floats are going into and out of CArray[num32] 14:13
timotimo oh
do we actually properly support that?
RabidGravy it seems so
there is some trouble with num64/doubles but hey
14:14 molaf_ joined
RabidGravy on the whole I'm quite pleasantly surprised that it works so well 14:15
14:18 molaf left
RabidGravy it's going to be either more fun testing the sample rate conversion 14:22
14:25 gerry__ left 14:36 diana_olhovik left 14:42 zakharyas joined 14:46 rindolf joined 14:58 arumugam joined
arumugam p6: say 3; 14:59
camelia rakudo-moar 7fc078: OUTPUT«3␤»
arumugam hi
psch hi arumugam 15:00
arumugam sir
i would like to use perl6
i have written modules, programming in perl <6
psch rakudo.org/how-to-get-rakudo/ is a decent starting point, although not bleeding edge 15:01
especially windows is a few months behind, iirc
arumugam is there any possibility to create exe in perl6
psch i don't think so, no, but i'm not that involved in the backend that could support that
arumugam say if i would like to distribute my coding do i need the client to have perl 6 installed? 15:02
RabidGravy currently yes
arumugam i also couldnt find the activate state version for perl 6 15:04
RabidGravy er
psch there is no activestate version of Perl 6
arumugam er?
o.i.c
psch we have an MSI, you can find it on the link i wrote above
arumugam is there any new specific and feature rich ide for perl6 15:05
psch but as mentioned, it's a bit out of date, if i'm not misinformed
arumugam thank you psch
FROGGS[mobile] yes, latest msi is from march
IIRC
RabidGravy some Linux distributions have rakudo* of various uptodateness
15:06 espadrine left
arumugam dear all thank you for replying 15:06
AlexDaniel debian has 2014.07
arumugam: you can try perl6-mode in emacs 15:07
arumugam: that's not an IDE, but it might work for you
RabidGravy fedora has ~ 2015.04
AlexDaniel arumugam: especially if you already use emacs, if that's the case
arumugam emacs for windows
is that available only for linux? 15:08
how about padre does it support perl6
RabidGravy no
nor eclipse afaik
AlexDaniel RabidGravy: that's pretty new :) 15:09
RabidGravy yeah I was quite surprised when the update came through a few weeks ago 15:10
15:10 kurahaupo left
arumugam once again thank you geeks 15:11
15:12 arumugam left
AlexDaniel hm 15:12
15:12 gerry__ joined
AlexDaniel that was weird 15:13
psch hm 15:14
sometimes i'd like a mapping of qb_* functions to blocks
but then i try to imagine how to implement that and resign :s 15:15
s/functions/methods/
15:15 b2gills left
timotimo quickbasic functions? 15:15
psch no, generated methods in the jvm backend :P 15:16
QastBlock, i guess
*QAST
timotimo mhm :S 15:17
psch also 92 levels deep stacktrace with ~80% invokedynamic isn't helping :D
like, right now a "Cannot call method 'unshift' on a null object." points at an END block in CORE.setting 15:18
so i don't have @END_PHASERS, i guess? 15:19
15:19 b2gills joined
psch yeah, i'm probably still approaching this wrong :/ 15:20
15:25 gerry__ left
psch oh, no, approach might be right. we're creating @END_PHASERS in main.nqp and i'm skipping that 15:26
15:26 kurahaupo joined 15:30 diana_olhovik_ joined
FROGGS $ perl6 -Ilib -e 'use IO::Pipe; my $p1 = shell("dir", :out); my $p2 = shell("wc", :in($p1.out), :out($*ERR))' >/dev/null 15:35
110 219 3981
\o/ \o/
psch FROGGS++ 15:36
FROGGS $ perl6 -Ilib -e 'use IO::Pipe; shell("wc", :in("VERSION".IO.open), :out($*ERR))' 15:43
1 1 8
I'd say it is b