»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.
00:00 fhelmberger left 00:01 icetooth joined
emilper just in case it is important: tested gsl and cblas (gslcblas in fact) from C with the functions that I need and it works; from p6 on moar it crashes apparently at random, either here: 00:04
stat("/home/emilper/bin/rakudo/install/languages/perl6/lib/dynext/libperl6_ops_moar.so", 0x7fffdb607040) = -1 ENOENT (No such file or directory)
stat("/home/emilper/bin/rakudo/install/languages/perl6/runtime/dynext/libperl6_ops_moar.so", {st_mode=S_IFREG|0664, st_size=35474, ...}) = 0
--- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_ACCERR, si_addr=0x7f861f4a80f0} ---
+++ killed by SIGSEGV +++
or here:
mmap(NULL, 2164320, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 10, 0) = 0x7fe6cce68000
mprotect(0x7fe6cce77000, 2093056, PROT_NONE) = 0
mmap(0x7fe6cd076000, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 10, 0xe000) = 0x7fe6cd076000
close(10) = 0
mprotect(0x7fe6cd076000, 8192, PROT_READ) = 0 00:05
--- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x2} ---
+++ killed by SIGSEGV +++
... in case it matters :)
timotimo a gdb backtrace would probably be more helpful
maybe a build with ASAN?
emilper ... how do I do that ? literate in just c enough to read the code
timotimo ah 00:06
emilper ... I can find out :) and come back tomorrow
timotimo if you configure.pl your moarvm with --debug=3 --optimize=0 you can just start your moarvm with "gdb --args /.../.../moar blah.moarvm --foobar --barbaz"
then type "run" and when it crashes, "bt full" 00:07
emilper will do and return tomorrow, very late her e 00:08
timotimo good night! :)
and thanks for your time
emilper thank you
00:09 emilper left
psch only the proper return missing for tr///, but that really seems the hardest to me by now... 00:10
i mean, i guess i could slap a .distance or something onto Str and call that with a :op<callmethod> or somesuch, but i'm not sure that's the right way
the other way i can think of is building an ast that computes the distance, which i think looks even wronger as a solution 00:11
am i missing something obvious?
timotimo it'd probably be nicer to calculate the distance as you're doing the changes 00:13
psch tr/// calls trough to .trans
timotimo oh
psch so that's kinda hard
timotimo well, if you do it after .trans is finished, you can return an object that can do less work if it's evaluated as .Bool vs .Int 00:14
psch m: my $s = 'abc'; say $s.trans: 'a..c' => 'd..f' 00:15
camelia rakudo-moar a7b4cb: OUTPUT«def␤»
00:15 thou left
psch so i guess that still works if i evaluate it as .Int from .trans 00:15
it's just hard for me to tell when i can put changes into higher-level parts of rakudo and when they should stay in the grammar or actions 00:16
hm, does that actually sound sensible though? having Str.trans return Str or Int 00:17
timotimo: or am i misunderstanding your suggestion? 00:18
timotimo the return value is supposed to be an int, right? 00:19
that tells how much the string changed?
psch yes, for tr///, which effectively is always in smartmatch context
as in, i can't think of a way to use tr/// that's not either implicite against $_ or explicite with ~~
00:20 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg
timotimo so if we only look at the result in a boolean context, we can skip lengthy calculations of the actual value 00:20
00:23 ajr_ left
psch well i'm not yet doing the calculations for the cases where they have to be done 00:24
i assume you mean code like "if tr/// { }" could skip them, which makes sense
but "$c = $f ~~ tr///" obviously needs them and i don't know how to do them
well, except for the two ideas, which i don't feel competend enough to decide between 00:25
maybe i should just sleep on it... :) 00:28
timotimo sounds good 00:32
psch how'd i do that though? i mean, returning an object that only does the calculation if it stands in a context that wants an Int 00:34
eh, never mind that now. i actually should rather sleep 00:35
thanks for the input though timotimo++
00:35 psch left
timotimo well, the object could store the original string and the result string. in .Int it could return $!result //= do { #`( calculate ) }; 00:36
00:37 raiph left
BenGoldberg Isn't that what Proxy is for? 00:41
00:54 jnap1 joined, jnap left 00:57 thou joined
icetooth is perl6 going to eventually be used for web apps? 01:01
tadzik I'd expect so
it already is for some
01:04 FROGGS_ left
icetooth i'm trying to figure out what language to use for learning about web programming, but i can't find any solid info on how to use perl6 with apache, i'm assuming its not there yet? 01:04
tadzik oh, you're looking in the wrong directon 01:05
you shouldn't ever be concerned with apache when doing web development
that's a different layer of things, not your problem
I recommend reading up on PSGI, if you're a Perl-son 01:06
Perl perlson
ah, you get the idea
anyway, there's this new fancy trend these days, to keep your web app abstracted away from the way it's later being deployed
we have several PSGI implementations in Perl 6 too 01:07
github.com/tadzik/bailador is something I pulled off, which is quite useful
icetooth ahh interesting, i was only familiar with the mod_whatever scripting method 01:08
tadzik yeah, everyone's kinda moving away from that
PSGI is a wonderful idea
icetooth looks like its going to be another long night for me. thanks for the info! 01:10
01:13 Su-Shee left, klapperl_ joined
tadzik have fun :) I really have to call this night a night 01:13
01:15 Su-Shee joined 01:16 Khisanth left 01:17 klapperl left, FROGGS_ joined, Khisanth joined 01:18 jnap1 left
BenGoldberg icetooth, If you're writing web apps in perl, and you if *really* want to use Apache to deploy it, there exists a mod_psgi. 01:22
That's for perl5, though, not perl6 (yet). 01:23
icetooth i'm reading their readme right now actually, was thinking about using it for simplicity sake. since i seem to be stuck in the early 00's and all 01:24
doh
01:25 Akagi201_ joined
timotimo actually 01:27
i see a new trend where you run a little web server with your app and just let another web server handle stuff like caching and such
so you'll have something like mod_proxy hand your requests to your we bapp
web app*
because why have another protocol if we can just use http all the way through 01:28
01:31 uniejo joined
icetooth all i wanted was a nice little exercise in learning sql and a new scripting language, now its getting ridiculous :[ 01:34
dalek pan style="color: #395be5">perl6-examples: aeddf63 | (Yanick Champoux)++ | 99-problems/P06-scottp.pl:
typo in comment
 01:36
pan style="color: #395be5">perl6-examples: d051dac | timo++ | 99-problems/P06-scottp.pl:
Merge pull request #16 from yanick/patch-1

typo in comment
thank you!
01:40 thou left, thou joined 01:44 btyler left 02:16 raiph joined 02:18 jlaire left 02:19 jlaire joined 02:20 xiaomiao left 02:25 xiaomiao joined 02:27 telex left 02:28 telex joined 02:30 noganex joined 02:31 BenGoldberg left 02:33 noganex_ left
timotimo something i just thought of (while trying to fall asleep): i don't think we have a way to call function pointers from NativeCall yet 02:44
but binding OpenGL, for example, will require us to take function pointers and call them, because that's how you use OpenGL extensions
02:49 logie joined 03:02 kaare_ joined 03:04 grondilu joined 03:11 araujo left 03:17 molaf joined 03:21 xragnar_ joined, xragnar left, xragnar_ is now known as xragnar 03:22 rindolf joined 03:24 vendethiel left, kaleem joined 03:25 vendethiel joined
grondilu timotimo: isn't that what is done in 08-callbacks? 03:27
03:32 molaf left 03:35 btyler joined 03:37 [Sno] joined 03:38 uniejo left 03:40 logie left 03:42 kaleem left, kaleem joined 03:43 orafu joined 03:50 SamuraiJack joined
segomos_ timotimo: id promise.new doesnt start a new thread, what is the threadpoolschwduler.loads telling me? i thought that was a threadpool and it seems to increase with every new connection and promise created . . what am i misunderstanding (so i can go read more)? 03:52
if / threadpoolscheduler 03:53
04:02 kaare_ left 04:08 btyler left 04:16 anaeem1 joined 04:17 anaeem1 left, anaeem1 joined 04:23 [Sno] left 04:26 kaare_ joined 04:28 chenryn joined 04:31 rurban joined 04:36 Alula_ left
thou icetooth: if you're still around, there's an older implementation of a pretty full Plack (perl5 PSGI stack) in perl6 which, I think, hasn't gotten enough awareness. github.com/lopnor/p6-plackdo 04:38
yoleaux 12:43 MDT <raiph> thou: Your blog says "Let me know" but I see no obvious way to comment. What I'm sure anyone reading it wants to know is what you found out about control of your grammar's backtracking.
thou Yay, I got Perl6 added to mustache.github.io/ . 04:48
04:49 Alula_ joined, dayangkun joined, vendethiel left
sjohnson howdy. 04:50
04:50 rurban left 04:55 rurban joined, rhr joined 05:03 rurban left 05:05 kaleem left, rurban joined 05:06 kaleem joined 05:07 rurban left 05:10 rurban joined 05:12 rurban left 05:14 rurban joined 05:16 rurban left 05:17 raiph left 05:25 kaleem left 05:26 anaeem1 left, anaeem___ joined 05:28 rurban joined 05:32 gfldex joined 05:33 rurban left, grondilu_ joined 05:34 anaeem___ left, anaeem1 joined 05:36 grondilu left 05:38 vendethiel joined 05:48 rurban joined 05:50 [Sno] joined 05:53 rurban left, woolfy joined 06:05 lizmat left, kurahaupo joined 06:06 Alina-malina left, woolfy left 06:07 frettled joined 06:08 ivanshmakov left 06:20 denis_boyun_ joined
dalek kudo/nom: ffe6391 | coke++ | tools/autounfudge.pl:
enable autounfudge of moar, ala JVM
 06:29
kudo/nom: 813f6cf | coke++ | t/spectest.data:
run more spectest files

  (all tested on moar)
06:31 dayangkun left 06:34 dayangkun joined 06:44 FROGGS_ left 06:45 rindolf left 06:48 rindolf joined 06:49 rurban joined 06:54 rurban left
masak morning, #perl6 07:01
07:01 Alina-malina joined 07:02 dayangkun left 07:06 dayangkun joined 07:09 FROGGS joined, zakharyas joined 07:11 ivanshmakov joined 07:13 mr-foobar left 07:17 igorsutton joined 07:18 darutoko joined 07:19 properl6y joined 07:20 properl6y left, properl6y joined 07:21 dmol joined
PerlJam good morning masak 07:26
07:34 virtualsue joined 07:36 mr-foobar joined 07:44 SamuraiJack left 07:50 gtodd left, rurban joined 07:51 gtodd joined 07:53 rurban1 joined 07:54 rurban left 07:55 SevenWolf left 07:57 rurban1 left, brrt joined 08:01 ivanshmakov left, ivanshmakov joined, SamuraiJack joined 08:02 SamuraiJack left 08:07 SamuraiJack joined 08:09 SamuraiJack left, SamuraiJack joined 08:17 Woodi left, Woodi joined 08:23 icetooth left 08:27 Alina-malina left 08:28 pecastro joined 08:31 dakkar joined, hoverboard left 08:41 tommi joined 08:49 denis_boyun_ left
timotimo grondilu_: no, 08-callbacks is about turning a perl6-level sub into something that looks like a function pointer to C things 08:49
08:49 properl6y left
timotimo grondilu_: i want to cast a function pointer i get from a C-bound sub (so, essentially a 64bit integer) into a sub callable from perl6 08:49
segomos_: Promise is just the interface to get to know if/when a task has finished. in addition, the "then" method exists to kick off another task when a promise gets fulfilled and this also creates a new promise on the side 08:50
so in this case we always have a Callable that gets .cue'd to the Scheduler alongside a Promise to signal what happens in that Callable 08:51
i hope that helps a bit?
08:52 tommi left 08:54 rurban joined 08:58 rurban left, spider-mario joined 08:59 anaeem1 left 09:08 mr-foobar left, fhelmberger joined
grondilu_ timotimo: 09:13
timotimo: ok
09:13 grondilu_ is now known as grondilu 09:14 grondilu left, grondilu joined
timotimo jnthn: is there some easy way to teach zavolaj to do function pointers -> sub call "casting"? 09:30
09:32 stux left
jnthn Oh...if we get a function pointer returned? 09:32
I...think that is NYI and not quite trivial.
But possible.
timotimo OK 09:33
use case is OpenGL extensions
09:41 stux joined, kaleem joined 09:42 denis_boyun_ joined
grondilu
.oO( having an OpenGL port in Perl6 would be pretty cool )
 09:49
timotimo it'll be somewhat useless without good mutable native arrays, though 09:51
especially since the glBegin + glVertex/glColor/glTexCoord/... + glEnd pattern has been deprecated for a long time now. 09:52
i'm suspecting native/compact/sized arrays will harmonize very well with our jit 09:54
09:54 rurban joined
jnthn I'm sure we can harmonize them. :) 09:56
grondilu r: sub prefix:<^^>(Int $n) { 1 .. $n }; say ^^10 09:58
camelia rakudo-jvm 813f6c: OUTPUT«(timeout)» 09:59
..rakudo-{parrot,moar} 813f6c: OUTPUT«1..10␤»
09:59 rurban left 10:01 pochi joined 10:02 pochi_ left
dalek p/osr: cb1ed30 | jonathan++ | src/vm/moar/QAST/QASTOperationsMAST.nqp:
Emit osrpoint OSR hint ops in loop bodies.
 10:09
timotimo j: say "alive"
camelia rakudo-jvm 813f6c: OUTPUT«alive␤»
10:21 FROGGS left 10:23 FROGGS joined 10:24 Alina-malina joined 10:26 carlin joined 10:29 Alina-malina left, Alina-malina joined 10:34 chenryn left 10:38 chenryn joined 10:48 kaleem left 10:56 rurban joined 10:57 brrt left 11:00 rurban left
nwc10 www.blogger.com/comment.g?blogID=3...7640322884 -- As far as I know, a majority of the benchmarks we use have never been ported to Python 3. So it's far more complicated than just push a switch. 11:03
11:04 kaleem joined 11:06 logie joined 11:15 SamuraiJack left 11:17 uniejo joined 11:22 rindolf left 11:24 dmol left 11:29 uniejo left, uniejo joined 11:41 uniejo left 11:46 logie left 11:54 pdcawley left 11:57 rurban joined
tadzik hmm 12:00
feather seems to be ipv6-dead
I could only connect via v4 12:01
12:01 uniejo joined
tadzik (talk about first world problems) 12:01
12:01 rurban left 12:03 uniejo left, uniejo joined 12:04 dolmen joined
moritz tadzik: ssh -6 feather # works fine for me 12:10
tadzik hm, I had to go through -4
12:10 chenryn left
tadzik may be the office connection then 12:10
moritz tadzik: nopaste traceroute6 feather.perl6.nl output?
you can also try a 'ping6 google.com' 12:11
they are pretty reliable :-)
tadzik gist.github.com/tadzik/aa203978950b4422a1cc
ping to google works alright
moritz tadzik: traceroute looks pretty much the same for me starting from gw-atom86, but for me it works 12:12
tadzik: looks like it's actually a problem with feather 12:13
FROGGS moritz: does feather support polish ipv6?
tadzik do we have a special ipv6? :D
it always worked
FROGGS *g*
moritz ip6tables -nvL on feather doesn't show any suspicious rules 12:15
just
Chain INPUT (policy ACCEPT 2510K packets, 362M bytes) pkts bytes target prot opt in out source destination
3549K 482M sshguard all * * ::/0 ::/0
12:17 chenryn joined 12:20 lizmat joined, woolfy joined 12:23 Ven_ joined 12:27 logie joined 12:29 kaare_ left 12:33 chenryn left, Alina-malina left
zengargoyle is sshguard similar to fail2ban? did you muck your password enough to get blocked? :P 12:34
12:37 dmol joined 12:38 brrt joined
lizmat good *, #perl6! 12:43
12:44 guru joined, guru is now known as Guest5826, uniejo left
lizmat wonders whether there is a way to EVAL a string into a serialized format, like you get with --output 12:44
so that you can later just load the bytecode ? 12:45
12:45 Guest5826 is now known as ajr_
Ven_ lizmat: --output isn't just like shell's > ? 12:45
lizmat no, not afaik
12:47 logie left
Ven_ oh, I should try it sometimes then. 12:47
Having a specced AST format would help, I guess
12:47 logie joined
moritz a big difference is that if you use a redirect, and compilation fails, you truncate the output file (if it existed before) 12:48
timotimo lizmat: with nqp::getcomp you can get an instance of the Compiler object
Ven_ because EVAL() dumping bytecode would make it impossible to be ported anywhere
12:48 mr-foobar joined
timotimo then you can do the stages by yourself 12:48
12:48 jnap joined
Ven_ well, you're working on inner stuff :-). 12:48
lizmat I guess I could write the EVALcode to a file called Foo.pm, then shell out and do a ./perl6-m --target=mbc --output=Foo.moarvm Foo.pm 12:49
12:49 SamuraiJack joined
lizmat timotimo: ok, I'll look in that direction 12:49
the reason I'm asking, is that I'm thinking of having the internal index for a CURL::Installation to be serialized code 12:50
12:53 logie left 12:54 sqirrel joined
timotimo serialized code always has a security implication to it, but this is about installing code from places, so ... maybe not that important 12:55
12:59 rurban joined
lizmat timotimo: actually, this is a very good point :-( 13:00
I guess I have to make triple sure that no executable code is leaking from the META6.json file 13:01
into this index 13:02
13:03 rurban left 13:06 chenryn joined
TimToady top o' the * to you 13:12
moritz \o 13:13
lizmat TimToady o/
moritz o/
13:14 LLamaRider joined
lizmat will be heading down to hackathon room around 10am (OLT) 13:17
(Orlando Local Time)
13:19 chenryn left 13:21 cosimo left 13:22 cosimo joined 13:24 chenryn joined
FROGGS lizmat: is it important at this stage to turn the CURL::Installation database into something not human readable? 13:26
13:27 Alina-malina joined 13:29 rurban joined
lizmat yes, out of performance reasons 13:29
Ven_ which reminds me, are there low-hanging fruits ?
lizmat Ven_: yes
someone suggested yesterday that it might be a good idea to create a list of links to the Synopses
Ven_ a list of links ? 13:30
lizmat for a newbie to use to learn Perl 6
colomon jnthn, etc: module smoke testing is still mostly to completely broken under parrot and moar. It's been over a week since I had a successful complete smoke run.
13:30 cosimo left
Ven_ lizmat: oh, hmm. 13:30
lizmat :-)
13:30 cosimo joined
Ven_ well, that goes in the scope of my talk 13:30
lizmat you having gone through that process recently, seem like a good candidate :-) 13:31
FROGGS colomon: that is very weird, since I never have problem running the star module tests for these platforms for example
Ven_ yes, yes. I also point to people when I try and get them into it.
Should I make it a tour, rather, though ?
colomon FROGGS: how are you running the star module tests?
Ven_ "what do you want to learn if you're new here"
lizmat yup, sounds good :-)
colomon FROGGS: I've had failures building panda under parrot at least 75% of the time I've tried. 13:32
FROGGS I fork star, also fork MoarVM/nqp/rakudo into the star dir, and fake up a star release, unpack that tarball and compile and test
Ven_ Which makes me think I should write a non-p5-beginner tutorial, because we don't have those.
colomon FROGGS: define "test"?
FROGGS make modules-test 13:33
lizmat ++Ven_
Ven_ has finished his medical appointment week and now will have time to do stuff before sleeping o/
colomon FROGGS: hmmm. not sure how that works. I think it's possible that, of the two errors I'm seeing, neither would should up in your procedure. 13:34
Ven_ lizmat: where'd you post such a tutorial ?
That probably can be in a blog
FROGGS colomon: that might exlpain it, yeah
so it might be that the problem is somewhere in between of you testing infrastructure and rakudo
brb 13:35
TimToady lizmat: where is the hackathon? 13:39
lizmat one of the rooms on the other side of the corridor from where the YAPC::NA was 13:42
colomon FROGGS: hmmm, problem may have been because rakudobrew was giving me an error that didn't show up in my log? Tweaked smoke script and trying again.
lizmat on the second floor 13:43
13:43 chenryn left
colomon FROGGS: nope, that didn' 13:44
that didn't help at all.
shift_pmc() not implemented in class 'Continuation' 13:47
in block at bin/panda:21
in sub MAIN at bin/panda:19
13:47 chenryn joined
colomon arrrgh 13:50
none of these errors are consistent
13:50 rurban left
colomon I'm running "rakudobrew build-panda" and sometimes it works and sometimes it gives an error like the above. 13:51
13:52 btyler joined
colomon ==> Installing JSON::Tiny 13:52
Segmentation fault (core dumped)
that's new
tadzik :)
timotimo wow :S 13:53
colomon next run, successful install 13:54
next run, shift_pmc() not implemented in class 'Continuation' 13:56
moritz and why would it be implemented in that class :-)
14:00 kaare_ joined
timotimo arnsholt: will you have tuit supply in the future? there's lots and lots of unresolved issues in zavolaj that may benefit from a quick status update at the very least 14:02
for example, does this still apply? gist.github.com/arnsholt/4963237 14:04
14:06 chenryn left, araujo joined
colomon next run, successful install :\ 14:06
FROGGS colomon: that shift_pmc thingy is a new "feature" of parrot 6.5.0 I think 14:07
colomon next run, shift_pmc() not implemented in class 'Int' in block at bin/panda:21 14:08
14:13 treehug88 joined
colomon FROGGS: hmm… and my last successful smoke test was the day parrot 6.5.0 was released... 14:13
FROGGS that does not explain your moar problems though
though, we should ticket the parrot regression 14:14
colomon FROGGS: moar problem seems to be the same one from months ago -- installing File::Find from panda breaks the universe. 14:16
FROGGS: I would feel better if someone else could duplicate the parrot problem. 14:17
FROGGS colomon: do you export/set PERL6LIB or RAKUDOLIB by any chance?
colomon FROGGS: nope
14:20 sqirrel left 14:21 chenryn joined
Ven_ oh, introductory tutorial on github. That could be a good idea. 14:24
14:27 virtualsue left 14:29 [Sno] left
colomon tries to build parrot rakudo on his Mac see if the issue occrus there, and discovers he cannot because of some sort of libiconv.2.dylib issue. :\ 14:29
FROGGS ohh yes, the pain of working with a mac... 14:31
timotimo joys*
14:32 virtualsue joined 14:33 rindolf joined
colomon jvm build fails with java.lang.OutOfMemoryError: Java heap space 14:34
in run (gen/jvm/stage2/QRegex.nqp:395)
14:34 molaf joined
colomon starting to think this is a conspiracy to make everyone run moar…. 14:35
14:38 bluescreen10 joined 14:39 chenryn left
hoelzro ahoy #perl6 14:40
colomon o/
Ven_ o/ 14:41
14:45 kaleem left, ajr_ left, guru joined, guru is now known as Guest87331
colomon hopes the hackathon is going well 14:45
14:46 Guest87331 is now known as ajr_ 14:48 chenryn joined
TimToady well, lizmat++ seems to have wandered off 14:54
we didn't get there till 10:30
so pretty much I'm the P6 hackathon at the moment... 14:55
jnthn Well, then you'll just have to be crazy productive :P 14:57
masak you can do it!
PerlJam TimToady: so .... what are you hacking on? :) 14:58
TimToady I'm looking at ren1us++'s gist 15:01
which does indeed seem to indicate a bug
jnthn link? :)
masak url? :)
15:01 chenryn left
jnthn öl? :) 15:01
TimToady gist.githubusercontent.com/ChoHag/...stfile1.p6
if I add 'is rw' to the attribute, doesn't make any difference 15:02
ChoHag Oooh fix that one :)
TimToady if I change nextsame to callsame and try to get its return, it seems to be returning 0 args
ChoHag That's my blocker.
TimToady oops, that was ChoHag++ 15:03
sorry, mixing up my P6 fuzzers :)
jnthn uh, where is that nextsame meant to be going?
moritz what does that nextsame dispatches to?
jnthn oh, it's being used to wrap
TimToady the accessor
which wants to be a rw wrapper 15:04
as I said, the attr should be marked 'is rw is wtf', but still doesn't work with that
ChoHag Yeah I tried that. 15:05
15:06 denis_boyun_ left
jnthn Agree it needs the "is rw". Also agree it appears to be a bug. 15:07
Well, 2 bugs... 15:10
TimToady I'm surprised callsame seems busted
you'd think we'd have a test for its return value
ChoHag I *think* it's the wrap which strips the container somehow. 15:11
I've no idea what I'm looking at though.
TimToady thinks its the callsame/nextsame 15:12
because my (\c) = callsame complains about expected 1 but got 0 15:13
dalek kudo/nom: ef3bb15 | jnthn++ | src/core/Routine.pm:
wrap should not force decontainerization.
jnthn With that patch, it works with callsame. 15:14
TimToady jnthn++
jnthn I can see nextsame/nextwith are wrong in this regard; the fix isn't immediately obvious. Will have to poke a bit more.
spectest(s) for this welcome :) 15:15
15:16 raiph joined
TimToady well, my (\c) ::= callsame, I meant 15:16
jnthn bbiab 15:17
15:18 igorsutton left
gtodd lizmat: "Introduction to NQP" ? 15:22
lizmat: oops sorry I had scrolled back in my terminal and Ctrl-s'ed I think that conversation was from days ago 15:23
TimToady jnthn: the wrapper still seems to integer with BUILD and default values, though assignment works now 15:24
s/integer/interfere/ #grr
what's up with my brane?
Ven_ it's being hacked on 15:25
gtodd overzealous spell-checking ? (at biological level)
15:26 Ven_ left
gtodd your brain assumes you mean to type integer .. a word you have used more often than interfere ... your biologically innate sorting algorithms interfered with typing 15:27
lizmat was held up at breakfast but will be joining shortly
TimToady well, the program was using integers, anyway 15:28
15:28 woolfy left 15:29 lizmat left 15:30 zakharyas left
moritz integer interference 15:30
carlin integerference 15:31
15:31 sqirrel joined 15:32 lizmat joined 15:36 chenryn joined 15:42 labster joined 15:43 chenryn left 15:44 slavik left 15:45 hoverboard joined
lizmat is spectesting jnthn latest patch 15:46
15:46 SamuraiJack_ joined
ajr_ Google are apparently replacing their Dalvik VM with a new one called "ART". They claim it's entirely compatible. 15:47
www.theregister.co.uk/2014/06/26/go...nd_beyond/ 15:48
We'll see.
15:48 chenryn joined 15:49 slavik joined, virtualsue left 15:50 SamuraiJack left
dalek ast: e257fed | TimToady++ | S06-advanced/wrap.t:
add tests for wrapped rw accessors
 15:50
timotimo ajr_: you can already use ART if you activate the developer options on your phones 15:51
i've already run it for a few months on my nex4 and have equipped my newer devices with ART as well
ajr_ Good.
timotimo i didn't do measurements at all, but allegedly it gives a nice speed boost
ajr_ Obviously not an issue then. 15:52
timotimo it *seems* stable enough 15:53
15:59 kaleem joined 16:05 brrt left, fhelmberger left, chenryn left 16:06 kaleem left, virtualsue joined 16:08 chenryn joined 16:09 SamuraiJack_ left, SamuraiJack_ joined
gtodd carlin: integerference ++ ... that makes two people who have used this word :) 16:11
if we can create a coherent definition meme creation will be facilitated ... since people mix integers and reals/floats in code all the time (when rounding and numbers for time and duration - didn't Knuth once say 100 * 0.1 is rarely equal to 10 or something?) this phenomena could be called "integerference" perhaps? But IANAM 16:17
16:17 dayangkun left 16:18 rindolf left
dalek ast: 6f33411 | (Elizabeth Mattijsen)++ | S02-magicals/ (3 files):
Fudge DISTRO/PERL/VM for parrot

For some reason, only a todo seems to corrupt the internal state, and thus cause segfaults		 16:19
16:22 Rotwang joined, Rotwang left, Rotwang joined 16:26 chenryn left
ChoHag Given gist.github.com/ChoHag/e1a910d6ac454aadb7b3, which now works (thanks!), is there any way I can tell that SomeTrait has been applied to the $!x attribute? 16:28
moritz ChoHag: well, you can introspect the attributes, and check if it does the role with ~~ SomeTrait 16:30
16:30 molaf_ joined
moritz foo.^attributes.grep(*.name eq '$!x) ~~ SomeTrait 16:31
TimToady s/')'/')/ 16:32
16:33 molaf left
TimToady and your new gist doesn't work here... 16:33
16:36 FROGGS left, MilkmanDan left 16:37 MilkmanDan joined
TimToady oh, I never noticed that the wtf trait was trying to set rw instead of an 'is rw' 16:37
but anyway, doesn't work here with that either...
16:38 dakkar left
gtodd timotimo: with my Android 4.4 system I needed to tap 7 times on a secret button to turn on developer options ... which I only turned on because I thought it was necessary for perl ;-) (it is not) 16:38
16:39 dmol left
segomos_ gotta get that new amazon phone 16:39
TimToady works if I change the 'nextsame' to 'callsame' and return it implicitly
doesn't work with an explicit 'return callsame;' 16:40
16:40 chenryn joined
jnthn Is the thing the callsame's in marked "is rw"? 16:45
16:47 guru joined 16:48 guru is now known as Guest32177 16:49 ajr_ left, chenryn left, SamuraiJack_ left 16:50 SamuraiJack_ joined
dalek ast: d48718f | TimToady++ | S06-advanced/wrap.t:
test more ways of returning a wrapped rw accessor
 16:51
TimToady the wrapper is marked 'is rw'
I dunno what calls apply_handles... 16:52
16:52 virtualsue left
TimToady I presume that's deep magic of some sort 16:52
16:52 Guest32177 left 16:57 xinming left
jnthn method compose in Attribute, iirc 16:57
16:58 SamuraiJack_ left, xinming joined 17:00 lizmat left, lizmat joined
ChoHag Can I overwrite the equivalent of the set accessor of an attribute? 17:00
jnthn There is no set accessor 17:01
There's just accessors that "is rw" and ones that ain't.
vendethiel only a writable container :P
jnthn Right, it returns the container.
17:02 treehug88 left
lizmat lunch& 17:03
17:05 labster left
jnthn goes to acquire noms 17:06
17:07 rindolf joined 17:11 carlin left, klapperl joined 17:12 labster joined, klapperl_ left
segomos_ btyler: i was able to get wrk to run 10 connections on 10 threads against HTTP::Server::Async 17:19
17:19 virtualsue joined, troydm left
segomos_ with p6-j 17:20
17:22 labster left 17:23 dmol joined 17:25 wookie joined
timotimo gtodd: yes, i think it's usually the "build number" in the "about phone" section 17:25
i activated dev options because i want to do a bit of android development
17:26 carlin joined
timotimo wrk? 17:28
segomos_ ab replacement that is supposed to be better, btyler found it 17:29
btyler like ab, but better
segomos_ github.com/wg/wrk
it's on mac brew if you have osx 17:30
vendethiel where could I publish an equivalent to "p5 to p6" for non-p5ers ? Or a good tool for writing that
timotimo oh, neato.
vendethiel ST is pretty bad at that game ...
segomos_ probably on aptitude or linux package managers too if you have linux
masak is listening to Knutsen & Ludvigsen
moritz masak: Ronja has discovered them recently, so I've been listening to them alot recently 17:31
gtodd timotimo: it took about 20 mintues to reboot :)
timotimo "ST"? 17:32
vendethiel timotimo: sublime text
masak moritz: I queued them up when I saw you listening to them in the backlog. they've now rolled around.
segomos_ btyler: moarvm still craps out but it looks like some kind of unhandled, nonresumable, thread error. haven't really dug into it
masak moritz: tell Ronja I ratherr enjoy their music ;)
rather* 17:33
vendethiel: publish it somewhere where people can link to it :)
vendethiel: a blog post, a gist, a wiki page... 17:34
vendethiel i don't have a blog, though
segomos_ vim and git blog
vendethiel no vim 17:35
masak what segomos_ said.
no vim!? that's sad :/
here, have a nickel. buy yourself a vim.
vendethiel like, emacs I can tolerate; vim no way.
carlin heretic!
masak likes both 17:36
segomos_ haven't ever used emacs
masak Emacs gets the programmability really right. 17:37
vim nails composable actions.
gtodd www.perl6.org/compilers/ !! the curvy rounded drop shadow bits are missing from the red and yellow squares!!
segomos_ programmability?
meaning the ease at which you can write plugins for it?
gtodd masak: and Vim::X by yannick :) 17:38
btyler segomos_: I can still repeatably hang by having wrk connect with more connections than the number of threads in rakudo
segomos_ vendethiel: scriptogr.am is something i've been checking out, you can blog with markdown
btyler: did you pull down the latest? 17:39
masak segomos_: Emacs is in a very real sense an introspectable programming environment with a text editor implemented on top of it.
vendethiel doesn't customize emacs
masak o.O 17:40
btyler segomos_: yep, this is with latest
vendethiel well, I knew I'd only use it for 3 weeks, soo.
carlin emacs is a great operating system, if only it had a decent text editor
btyler and an up-to-date rakudo-j
masak carlin: I for one rather like its text editor.
segomos_ let me rebuild rakudo-j and retry
gtodd there might be either a vim programmable via perl6 extension ... or an editor written perl6 ... but that would still be different than Emacs ... 17:41
segomos_ ctrl+r in vim is usually good enough for me 17:43
ctrl+n rather 17:45
17:45 labster joined
masak I used to think that. 17:49
segomos_ maybe i'll give emacs a try
17:50 labster left 17:51 guru joined, guru is now known as Guest20834
gtodd sometimes if you squint scala code almost looks like is should actually run with perl6 ... docs.scala-lang.org/cheatsheets/ 17:51
17:52 Guest20834 is now known as ajr_
carlin needs sigils 17:55
segomos_ btyler: odd.. gist.github.com/tony-o/65e8d9b880e86b2ea8b6 17:56
carlin they even use $_ but without the $
segomos_ after burn in on the jvm that time ratchets down to 50ms area for me
btyler segomos_: is that with the RAKUDO_MAXTHREADS env?
segomos_ nah, i have that variable unset 17:57
btyler ok
likewise
segomos_ got an idea, i'lll check back in later 18:01
btyler segomos_: ok -- let me know if there's anything I can do to help get some insight
18:02 FROGGS joined
FROGGS o/ 18:03
18:06 treehug88 joined
vendethiel segomos_: thanks, I'll try that 18:08
I'm pretty much just looking for a blog to publish 18:09
I'll maybe translate it in french. at that point, I translated a terrifyingly big fork of phpBB, LiveScript's documentation, so I might as well... :)
18:23 dayangkun joined, smls joined 18:37 [Sno] joined, chenryn joined 18:41 darutoko left 18:45 chenryn left, [Sno] left, dayangkun left, dmol left, rindolf left, molaf_ left, LLamaRider left, gfldex left, jlaire left, zengargoyle left, vike left, sivoais left, nebuchadnezzar left, ggoebel111116 left, jercos left, treehug88 left, MilkmanDan left, slavik left, bluescreen10 left, cosimo left, mr-foobar left, gtodd left, frettled left, kurahaupo left, telex left, Khisanth left, thilp left, erkan left, robinsmidsrod left, mtj_ left, ChoHag left, nnunley left, raydiak left, mhasch left, ilbot3 left, djanatyn left, perigrin left, ccverg left, timotimo left, REPLeffect left, atrodo left, yogan left, tokuhirom left, muraiki left, go|dfish left, firnsy left, FOAD left, pmichaud left, hugme left, retupmoca left, Bucciarati left, TimToady left, lestrrat left 18:46 chenryn joined, [Sno] joined, dayangkun joined, treehug88 joined, dmol joined, rindolf joined, MilkmanDan joined, molaf_ joined, slavik joined, bluescreen10 joined, cosimo joined, LLamaRider joined, mr-foobar joined, gtodd joined, frettled joined, kurahaupo joined, gfldex joined, telex joined, jlaire joined, Khisanth joined, thilp joined, erkan joined, zengargoyle joined, timotimo joined, robinsmidsrod joined, ChoHag joined, REPLeffect joined, vike joined, sivoais joined, atrodo joined, mtj_ joined, nebuchadnezzar joined, ggoebel111116 joined, nnunley joined, raydiak joined, yogan joined, ccverg joined, tokuhirom joined, jercos joined, mhasch joined, ilbot3 joined, djanatyn joined, perigrin joined, muraiki joined, go|dfish joined, firnsy joined, lestrrat joined, TimToady joined, Bucciarati joined, retupmoca joined, hugme joined, pmichaud joined, FOAD joined, wilhelm.freenode.net sets mode: +v hugme, telex left 18:47 telex joined 18:48 zamolxes left, apejens left, rjbs left, [Coke] joined, apejens joined, segomos joined, rjbs joined, PerlJam joined, zamolxes joined, broquaint joined, Rounin joined, nickpatch joined, jtpalmer joined 18:51 LLamaRid1r joined, sivoais_ joined, jercos_ joined, zengargo1le joined 18:52 jlaire_ joined, gfldex_ joined 18:53 telex left 18:56 hoverboard is now known as totesboard, totesboard is now known as hoverboard, chenryn left, [Sno] left, dayangkun left, dmol left, rindolf left, molaf_ left, LLamaRider left, gfldex left, jlaire left, zengargoyle left, vike left, sivoais left, nebuchadnezzar left, ggoebel111116 left, jercos left 18:58 donaldh joined 18:59 vike joined, rindolf joined, dayangkun joined 19:00 telex joined 19:01 anaeem1_ joined 19:05 PerlJam left, sqirrel left, kid_ joined
dalek kudo-star-daily: b3af78f | coke++ | log/ (13 files):
today (automated commit)
 19:06
rl6-roast-data: 93003b6 | coke++ | / (4 files):
today (automated commit)
19:11 raiph left 19:14 rindolf left
vendethiel okay, scriptogr.am is exactly why I dislike client js. 19:14
segomos vendethiel: ? 19:17
vendethiel segomos: their dashboard stuff is soooo slow.
moritz ooh, looks like there's a new .pm group, 'hardware.pm' 19:18
a new candidate for a release name :-)
19:19 chenryn joined, [Sno] joined, dmol joined, molaf_ joined, ggoebel111116 joined
TimToady used to use emacs, but it wore out his pinky finger 19:20
segomos "58 useful emacs shortcut keys" lol..only 58 :-) 19:22
dalek ast: 9bb5c8c | (Elizabeth Mattijsen)++ | S06-advanced/wrap.t:
Adjust text count
 19:23
19:23 nebuchadnezzar joined
masak TimToady: what did you use the pinky finger for in Emacs? I'm curious. 19:23
vendethiel masak: ctrl ?
TimToady ^N ^P
this was kinda before they invented arrow keys :) 19:24
masak oh. I have my Caps Lock rewired as Ctrl.
vendethiel masak: I have my pinky for both caps lock and ctrl, though
TimToady well, my Ctrl was already in the right spot, but that didn't help :)
masak :)
vendethiel (well, it's probably less stressful for the pinky)
and I can't quite imagine not having arrow keys, haha
TimToady it's not so much the occasional Ctrl keys, but having to hold it down while pounding other keys 19:26
so, for instance, in vim I type Ctrl-] in preference to Esc
but that's brief
and doesn't induce RSI
segomos no problems with shift+; ? 19:27
TimToady I don't type ::::::::::: very often
segomos that is p common for me, maybe i'm overly enthusiastic about modifying the buffer with perldo though
vendethiel thought about typing ctrl-] everytime on his french keyboard 19:28
segomos to exit emacs i do ctrl+c and then ctrl+x ? is that really the best way?
19:28 itz joined
TimToady that's the only way I know 19:29
jdv79 is that the same as :wq?
geekosaur other way around
TimToady ZZ makes a little more sense :)
geekosaur ctrl-x, ctrl-c
segomos it's the same as just :q and if you've modified the buffer then it comes with a bunch of prompts
geekosaur it will prompt for you to save any unsaved buffers 19:30
TimToady always has autosave set, so even a ^Z to suspend will save
jdv79 suffering
TimToady yes, ZZ is basically :wq, but faster
ZZ also works in less 19:31
Juerd 21:26 < TimToady> so, for instance, in vim I type Ctrl-] in preference to Esc
Interesting. I often use esc to avoid typing alt :P
Got better with my new thinkpad though, where Esc is left of F1 instead of one row above it. 19:32
TimToady well, I'll use my Compose key to write «» in preference to ^K in vim :)
Juerd I never really use ^K because it only works in vim :)
19:32 [particle] joined
Juerd Compose works in irssi and everywhere else 19:32
19:33 [particle]1 left
segomos ahh in vim i can just use alt+\ and alt+shift+\ to type «» 19:35
actually outside of vim i can do that too
19:35 jlaire_ is now known as jlaire
anocelot I didn't know alt+\ worked! :) 19:36
segomos alt+; does the three dots too … 19:37
19:37 chenryn left 19:40 hoverboard left
TimToady I think most Usians think that using a Compose key is somehow unamerican... 19:41
segomos m: $*SCHEDULER = ThreadPoolScheduler.new(:max_threads(Inf)); my $p = Promise.new; start { sleep 500; $p.keep(True); }; await $p;
camelia rakudo-moar ef3bb1: OUTPUT«Failed to unbox object to integer␤ in method new at src/gen/m-CORE.setting:17977␤ in method initialize at src/gen/m-CORE.setting:18161␤ in method cue at src/gen/m-CORE.setting:18114␤ in method start at src/gen/m-CORE.setting:18341␤ in sub st…»
vendethiel can I say that Rakudo* is (contains) "a stable version of Perl 6"? 19:42
or is that wishful thinking :) ?
(soon-to-be)?
lizmat *is
it's stable in most parts
segomos r: $*SCHEDULER = ThreadPoolScheduler.new(:max_threads(Inf)); my $p = Promise.new; start { sleep 500; $p.keep(True); }; await $p;
vendethiel lizmat: my sentence is actually "r* is a distribution containing stable p6 and modules"
lizmat just the async has still some quirks, I would say
camelia rakudo-jvm ef3bb1: OUTPUT«Cannot unbox argument to '$permits' as a native int␤ in method new at gen/jvm/CORE.setting:17918␤ in method initialize at gen/jvm/CORE.setting:18102␤ in method cue at gen/jvm/CORE.setting:18055␤ in method start at gen/jvm/CORE.setting:18282␤ …» 19:43
..rakudo-parrot ef3bb1: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Undeclared names:␤ Promise used at line 1␤ ThreadPoolScheduler used at line 1␤Undeclared routines:␤ await used at line 1␤ start used at line 1. Did you mean 'spurt'…»
..rakudo-moar ef3bb1: OUTPUT«Failed to unbox object to integer␤ in method new at src/gen/m-CORE.setting:17977␤ in method initialize at src/gen/m-CORE.setting:18161␤ in method cue at src/gen/m-CORE.setting:18114␤ in method start at src/gen/m-CORE.setting:18341␤ in sub st…»
segomos is that a bug or can i not really set ThreadPoolScheduler max_threads to Inf ? 19:44
vendethiel oh, it's been 2 months since the last R* ?
is there a list of modules comprised in r* ?
TimToady was trying to figure out why we were talking about the R language here... 19:46
segomos rakudo *
vendethiel TimToady: I've left perl and am doing infographics now :)
segomos i don't think he's talking about the statistical software
TimToady well, I suppose it's potentially just a dialect of P6 19:47
vendethiel so you're going to abduct me anyhow, eh ?
jnthn segomos: YOu can't pass Inf there.
segomos: No, I'm not going to fix that. That's insane. 19:48
segomos i thought you or timotimo suggested i try doing that
jnthn Should probably shove a type constraint in to make it fail sooner though.
19:48 labster joined
jnthn I'm allowed to be insane some of the time... :P I thought I suggested hunting what was hogging threads... :) 19:48
vendethiel I don't understand why you can't use Inf as a native Int, really ;-). There's only one letter that's different :P
segomos jnthn: you did and i traced that down and fixed that part, i still hang waiting on promises to return 19:49
vendethiel: it fails on a comparison of int to inf
vendethiel segomos: I'm kidding :) 19:50
TimToady it's only, what, 25¢ different
19:50 kid_ left
segomos lol 19:50
vendethiel TimToady: considering here, "France Telecom" had "unlimited 300 mo 3g" phone subs, I need to correct you 19:51
TimToady m: say ord('F') +^ ord('T')
camelia rakudo-moar ef3bb1: OUTPUT«18␤»
TimToady yeah, 25¢
(2 bits)
m: say (ord('F') +^ ord('T')).base(2) 19:53
camelia rakudo-moar ef3bb1: OUTPUT«10010␤»
TimToady there we go
segomos that looks like -2 19:54
TimToady in spots 19:56
well, at least in two spots 19:57
m: say (-2).base(2); # curious
camelia rakudo-moar ef3bb1: OUTPUT«-10␤»
segomos btyler: i've added, in the ab.pl6 file, how to increase the threadpool available for the server - that should work for now
doesn't do 2s compliment? 19:58
19:58 dayangkun left
TimToady yes, but it's hard to print an infinite number of 1's 19:58
segomos gues 2s compliment of 10010 is actually -14
m: Inf.base(2).say 19:59
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!===␤Cannot coerce Inf or NaN to an Int␤»
segomos m: (Inf - 1).base(2).say
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!===␤Cannot coerce Inf or NaN to an Int␤»
TimToady our arbitrary Int is 2's complement, and it's a trade secret how we store the infinite sequence of 0's for positive or 1's for negative :) 20:00
segomos lol 20:01
i think i'm good with http::server::async for right now..it's not great but it's usable 20:02
ie, i'd like to figure out how to make it handle not exhausting the threadpool 20:04
20:04 donaldh left
segomos jnthn: can $!max_threads be exposed in ThreadPoolScheduler? 20:05
jnthn segomos: Yes, no problem with that being introspectable
20:05 brrt joined 20:06 fhelmberger joined
TimToady m: say so -1 +& $_ for 1,2,4,8....* 20:07
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!=== Error while compiling /tmp/nf7Z9KFd6m␤Confused␤at /tmp/nf7Z9KFd6m:1␤------> say so -1 +& $_ for 1,2,4,8....*⏏<EOL>␤ expecting any of:␤ dotty method or postfix␤»
20:07 breinbaa1 joined
TimToady m: say so -1 +& $_ for 1,2,4,8...* 20:07
camelia rakudo-moar ef3bb1: OUTPUT«(timeout)True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤True␤Tr…»
TimToady m: say +so -1 +& $_ for 1,2,4,8...*
camelia rakudo-moar ef3bb1: OUTPUT«(timeout)1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1␤1…»
segomos jnthn: not sure i understand that
TimToady see, it's 1's all the wy up :)
*way 20:08
20:10 fhelmberger left
FROGGS it is one's more all the way right 20:10
20:11 itz left
TimToady well, irc is vertically challenged 20:12
20:12 jercos_ is now known as jercos
jnthn segomos: I'll just patch it :) 20:12
TimToady lizmat++ for turning her laptop on end :D
segomos i've been trying some of the ^.attributes.get_value 20:13
FROGGS goes back to Archive::Tar 20:15
segomos ahh..should be ^.attributes[].get_value
lizmat FROGGS++
TimToady you mean .^ ? 20:16
segomos m: say $*SCHEDULER.^attributes
camelia rakudo-moar ef3bb1: OUTPUT«Mu $!queue Mu $!thread_start_semaphore int $!loads Mu $!loads_lock Mu $!initial_threads Mu $!max_threads int $!started_any Callable &!uncaught_handler␤»
segomos yea that is what i mean
m: say ThreadPoolScheduler.^attributes[5].get_value($*SCHEDULER) 20:18
camelia rakudo-moar ef3bb1: OUTPUT«16␤»
20:18 orafu left
dalek kudo/nom: ca77a3d | jnthn++ | src/core/ThreadPoolScheduler.pm:
Make initial/max threads introspectable.

As requested by segomos++.		 20:19
kudo/nom: f2060ea | jnthn++ | src/core/ThreadPoolScheduler.pm:
Ensure thread counts are Ints.
segomos thank you jnthn
20:19 orafu joined
vendethiel scriptogram's code rendering is :( 20:21
it doesn't handle GHM, it seems 20:22
20:22 dayangkun joined 20:23 brrt left, orafu left 20:24 orafu joined 20:26 molaf_ left
lizmat vendethiel: example ? 20:29
20:33 kaare_ left
timotimo GHM is github flavored markdown? 20:34
20:35 smls left
vendethiel timotimo: yeah, GFM* would be more correcT. 20:37
FROGGS fsdo correcT :P
vendethiel lizmat: that's supposed to be on different lines : i.imgur.com/TgyAk0O.png 20:38
20:38 orafu left, orafu joined
vendethiel m: say Inf.WHAT 20:39
camelia rakudo-moar ef3bb1: OUTPUT«(Num)␤»
lizmat m: say $Inf.WHAT 20:40
camelia rakudo-moar ef3bb1: OUTPUT«(Num)␤»
masak m: say $$Inf.WHAT 20:41
camelia rakudo-moar ef3bb1: OUTPUT«(Num)␤»
masak m: say $$$Inf.WHAT
camelia rakudo-moar ef3bb1: OUTPUT«(Num)␤»
vendethiel $$$ ? what is this, perl ? 20:42
masak m: for 0..Inf -> $N { EVAL 'say ' ~ '$' x $N ~ 'Inf.WHAT' }
camelia rakudo-moar ef3bb1: OUTPUT«(timeout)(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)␤(Num)…» 20:43
masak om num num
jnthn :P
vendethiel
.oO( It must've taken a pretty high number of tries to iterate on that joke )
TimToady m: my term:<∞> = Inf; say ∞
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!===␤Type 'term:<∞>' is not declared␤at /tmp/uGyYl3px1l:1␤------> my term:<∞>⏏ = Inf; say ∞␤Malformed my␤at /tmp/uGyYl3px1l:1␤------> my term:<⏏∞> = Inf; say ∞␤ …»
masak is a bit tired
labster m: $Inf = 3
yoleaux 13 May 2014 22:07Z <retupmoca> labster: take a look at github.com/labster/perl6-File-Find...tes/pull/2 when you get a chance :)
camelia ( no output )
TimToady m: my \term:<∞> = Inf; say ∞
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!=== Error while compiling /tmp/pTkPHQJGma␤Term definition requires an initializer␤at /tmp/pTkPHQJGma:1␤------> my \term⏏:<∞> = Inf; say ∞␤ expecting any of:␤ scoped declarator␤»
jnthn Hmm 20:44
TimToady n: my \term:<∞> = Inf; say ∞
camelia niecza v24-109-g48a8de3: OUTPUT«Inf␤»
TimToady niecza++
jnthn m: my \∞ = Inf; say ∞
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!=== Error while compiling /tmp/wd6G8_w9VB␤Malformed my␤at /tmp/wd6G8_w9VB:1␤------> my ⏏\∞ = Inf; say ∞␤ expecting any of:␤ scoped declarator␤»
labster oh hi yoleaux (merged that a while back)
FROGGS hi labster
labster hi FROGGS. I’m here at YAPC::NA, trying to hack Lingua::Number back into working shape. 20:45
TimToady m: sub term:<∞>() { Inf }; say ∞
camelia rakudo-moar ef3bb1: OUTPUT«Inf␤»
TimToady m: sub term:<∞>() { Inf }; say ∞()
camelia rakudo-moar ef3bb1: OUTPUT«Cannot find method 'postcircumfix:<( )>'␤ in block at /tmp/hg37I7T1cR:1␤␤»
vendethiel I'm not sure if the tutorial should be newbie-oriented or developer-oriented
FROGGS labster: is it fun? I can't tell :o) 20:46
vendethiel maybe I should make a learnXinYminutes version ?
lizmat XinY sounds fine to me
labster That’s your problem for not coming this year :P
vendethiel I'll finish writing that introductory one for now :)
FROGGS labster: I know :/
lizmat I once started on some Lightning Talks called "Perl 5 -> 6 Nuggets" 20:47
FROGGS but making a trip that far is kinda hard for me
(small kids, small woman, you see)
vendethiel is there a multiline-input stuff for the REPL ?
like ^J ?
FROGGS ␤ <-- ? 20:48
or does that only work here?
vendethiel FROGGS: no clue how to do that on windows :/
lizmat m: say $Nan.WHAT
moritz that's a camelia feature
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!=== Error while compiling /tmp/OshV0vaW9t␤Variable '$Nan' is not declared␤at /tmp/OshV0vaW9t:1␤------> say $Nan.WHAT⏏<EOL>␤ expecting any of:␤ method arguments␤»
TimToady I think that's only for irc
lizmat m: say $NaN.WHAT
camelia rakudo-moar ef3bb1: OUTPUT«(Num)␤»
TimToady the REPL is really kind of done wrongish
it should be called out of the parser when it knows it's at a suitable state to interpret 20:49
20:50 Rotwang left
TimToady it's rather like intuiting ; in those languages that do 20:50
labster m: $NaN = "definitely not a number”; say $NaN
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!=== Error while compiling /tmp/Ckvomnz5ab␤Unable to parse expression in double quotes; couldn't find final '"' ␤at /tmp/Ckvomnz5ab:1␤------> aN = "definitely not a number”; say $NaN⏏<EOL>␤ expect…»
20:50 orafu left
labster m: $NaN = "definitely not a number"; say $NaN 20:50
camelia rakudo-moar ef3bb1: OUTPUT«definitely not a number␤»
timotimo m: my \NaN = "definitely not a number"; say NaN; 20:51
camelia rakudo-moar ef3bb1: OUTPUT«NaN␤»
timotimo m: my term:<NaN> { "nope" }; say NaN
camelia rakudo-moar ef3bb1: OUTPUT«===SORRY!===␤Type 'term:<NaN>' is not declared␤at /tmp/vhjBvOJGer:1␤------> my term:<NaN>⏏ { "nope" }; say NaN␤Malformed my␤at /tmp/vhjBvOJGer:1␤------> my term:<Na⏏N> { "nope" }; say NaN…»
timotimo m: my sub term:<NaN> { "nope" }; say NaN
camelia rakudo-moar ef3bb1: OUTPUT«nope␤»
20:53 raiph joined
vendethiel okay, so, that's my first take on what I want it to look like : scriptogr.am/vendethiel/post/and-pe...came-forth 20:56
fell free to .tell me my mistakes, cause I sure made a lot
(I decided to keep the name I used for my presentation, because I like the sound. not sure why ...)
jnthn "you can use say, like so" - but code below uses print instead. 20:58
vendethiel oh, hahaha. thanks :) 20:59
I should've expected this
jnthn Also, in the operator examples, we tend to put whitespace around infixes in Perl 6 by default
May be wise to set newbies off in that direction too
vendethiel almost pasted all his blogpost in IRC 21:00
fixed, as well
21:01 orafu joined
segomos vendethiel: your second code block mentions 'say' and then uses 'print' 21:04
vendethiel segomos: jnthn++ made me fix this already :)
segomos ah missed that 21:05
FROGGS continue BLOCK in P5 is like a LAST phaser? 21:08
segomos phaser? 21:09
FROGGS m: for ^3 { say $_; LAST { say 42 } }
camelia rakudo-moar ef3bb1: OUTPUT«0␤1␤2␤42␤»
FROGGS m: for ^3 { say $_; FIRST { say 42 } }
camelia rakudo-moar ef3bb1: OUTPUT«42␤0␤1␤2␤»
FROGGS m: for ^3 { say $_; NEXT { say 42 } }
camelia rakudo-moar ef3bb1: OUTPUT«0␤42␤1␤42␤2␤42␤»
segomos interesting 21:10
FROGGS ahh no, continue is like NEXT
lizmat indeed, check perldoc -f continue
21:11 Alula_ left 21:12 Alula_ joined, uniejo joined 21:13 rurban joined
segomos btyler: with jnthn's change, you should be able to run HTTP-Server-Async with a sane number of max_threads on the scheduler (the default) without it hanging in jvm. no matter what wrk settings you use. you will get a lot of errors if you try to use it with high concurrency and low threads 21:14
btyler segomos: that mirrors what I found poking at it this afternoon
oh, I need to update :) 21:15
labster w00t, unbusted Lingua::Number
FROGGS labster++ 21:16
labster FROGGS: anything I can do for Archive::Tar?
FROGGS labster: well, if you want to port it when I am sleeping :o) 21:17
it is here: github.com/FROGGS/p6-Archive-Tar
labster ok will do lol
FROGGS the test I use is: gist.github.com/FROGGS/b0104f6f211880db6a07 21:18
about 50% of the code is translated, and it already does something 21:19
(for that simple test)
21:19 rurban left
dalek kudo/nom: 36265c2 | (Elizabeth Mattijsen)++ | src/core/ (2 files):
Remove hack for $NaN: constant: seems to work now
 21:20
21:23 uniejo left, uniejo joined, amkrankruleuen left
labster thanks for the commit bit FROGGS, hopefully I’ll be able to use it 21:28
FROGGS :o)
I am trying to list the content of that tar now...
21:29 virtualsue left
labster I’m currently reading S11 so I can fix the module compunit stuff that lizmat broke for me 21:30
21:32 PerlJam joined
FROGGS cool, except for a wrongish encoded filename I can get the of files... 21:36
and something is not a number but should one it seems...
gnight o/
lizmat gnight FROGGS 21:39
timotimo .tell vendethiel "Just enter submit to evaluate the code you typed." that "enter submit" sounds weird/wrong 21:45
yoleaux timotimo: I'll pass your message to vendethiel.
timotimo .tell vendethiel i'm not sure if this is the right thing to tell a newbie, but the reason why { say "hello world!" } only outputs "hello world!" and not 1 (the return value of say) is because that piece of code wrote something to stdout and then the return value will not be printed 21:46
yoleaux timotimo: I'll pass your message to vendethiel.
21:46 ren1us joined 21:47 emilper joined
segomos oo nativecall+libxml, cool! 21:49
sergot m: sub a(*%x) { say %x; }; a(b => 'b'); a('a' => 'a'); 21:50
camelia rakudo-moar f2060e: OUTPUT«("b" => "b").hash␤Too many positional parameters passed; got 1 but expected 0␤ in sub a at /tmp/22U3tLLD0h:1␤ in block at /tmp/22U3tLLD0h:1␤␤»
sergot is it ok? 21:51
jnthn Yes; if you quote the left hand side it's no longer a named arg.
But instead a (positional) Pair argument 21:52
Whether you find that ok is an entirely different matter I guess, but it's how it is... ;-)
sergot jnthn++ thanks 21:55
21:56 brrt joined
lizmat gnight jnthn! 21:56
brrt i guess good night too :-)
emilper hello 21:57
here are the details about the p6 segmentation fault when using libgsl with NativeCall that I mentioned yesterday: gist.github.com/emil-perhinschi/5a...05af7f7439
PerlJam sergot: that's one of the reasons I tend to prefer the colon pair form. You're slightly less likely to accidentally quote the LHS ;) 21:58
brrt so, in deserialisation 21:59
i'll see if i can spot something fishy 22:00
timotimo data=0x0 22:01
looks fishy to me
dalek kudo/nom: f445f8a | (Elizabeth Mattijsen)++ | src/core/ (15 files):
Get rid of the $Inf hack

Turns out +Inf breaks in the early settings somehow in the "mast" stage with: Code ref does not exist in serialization context
22:03 ajr_ left, LLamaRid1r left 22:04 breinbaa1 left
brrt no, data may be 0 22:04
if it's zero it is taken from the current compilation unit
is that silly... maybe yes 22:05
timotimo OK
sergot PerlJam++: right, thanks :) 22:06
brrt so, stables are deserialized before objects, but after closures 22:09
ok, direct problem is that cc is 0 22:12
cc is MVMContainerConfiguration
it is acquired by a string key, that string key is read from the serialized data
supposing that it isn't the whole stream that is corrupt, i'd like to know what that string is 22:13
and.. i'd also like to know what keys are in fact in the container registry 22:14
emilper is that something I can do ? don't know much gdb 22:15
brrt you could set a conditional breakpoint on line 2137 of src/6model/serialization.c when i is 9 22:18
then step until line 1937 (or set a breakpoint there)
22:20 Bucciarati left 22:21 Bucciarati joined
brrt emilper - it may also help if you make this an issue rather than a gist :-) 22:22
emilper ... in which project ? moarvm ? 22:24
brrt yes
it seems to me that nativecall is either creating corrupt serialized bytecode or that it relies on a container configuration that isn't yet setup 22:26
22:26 bluescreen10 left
brrt but that depends on the value of our string :-) 22:26
if it's something weird like null, then it's likely to be bytecode corruption earlier 22:27
if it is a human-readable string, it's likely a setup issue
or even a typo issue, those happen :-) 22:28
and i sure hope it is the latter
emilper if it is not obvious I never messed up with moarvm code before or written much C either 22:31
github.com/MoarVM/MoarVM/issues/108 22:32
22:32 uniejo left
brrt well, fortunately there are people who have :-) 22:32
22:32 rurban joined
brrt seriously you've been a help just finding this out :-) 22:33
and posting the stack trace, etc
22:34 uniejo joined
brrt not quite enough people are using perl6-on-moarvm-with-NativeCall for us to have found all bugs just yet 22:35
timotimo definitely 22:37
22:39 vendethiel- joined 22:40 vendethiel left, treehug88 left 22:41 lizmat left
emilper (gdb) break src/6model/serialization.c:2137 if i = 9 22:43
Breakpoint 1 at 0x7ffff79b56c5: file src/6model/serialization.c, line 2137.
(gdb) break src/6model/serialization.c:1937
Breakpoint 2 at 0x7ffff79b4b28: file src/6model/serialization.c, line 1937.
(gdb) run
Starting program: /home/emilper/safedelete/bin/moar --libpath=/home/emilper/safedelete/languages/nqp/lib/ --libpath=/home/emilper/safedelete/languages/perl6/runtime --libpath=/home/emilper/safedelete/languages/nqp/lib/ /home/emilper/testing_ground/perl6/native_call.moarvm
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Breakpoint 1, MVM_serialization_deserialize (tc=0x6035e0, sc=0x7ffff62f69a8, string_heap=0x7ffff62f6ae0, codes_static=0x7ffff62f6b18, repo_conflicts=0x7ffff62f6aa8, data=0x0) at src/6model/serialization.c:2137 22:44
2137 deserialize_stable(tc, reader, i, MVM_sc_get_stable(tc, sc, i));
(gdb) c
Continuing.
Unhandled exception: No STable at index 9
at <unknown>:1 (/home/emilper/safedelete/languages/nqp/lib/ModuleLoader.moarvm::36)
22:44 vendethiel- left
emilper from <unknown>:1 (/home/emilper/testing_ground/perl6/native_call.moarvm::7) 22:44
[Inferior 1 (process 10982) exited with code 01]
(gdb)
The program is not being run.
22:45 uniejo left
emilper perl6-on-moarvm-with-NativeCall is gorgeous, it makes it worth my time to learn C 22:48
22:49 labster left
brrt grumpy me would say that C already makes it worth your time to learn C :-) 22:50
hmm 22:51
that.... looks a bit more like corruption, then, to me
because first the STable was there an now it isn't
22:52 spider-mario left, lizmat joined
emilper you are right, just perl 5 helps more with deadlines :-/ 22:52
brrt that is probably true :-) 22:53
brrt knows deadline stress only too well]
23:03 gfldex_ is now known as gfldex 23:05 btyler left
emilper NativeCall looks to me like the "killer app" for p6 23:05
23:05 wookie left
brrt really? that's really cool to hear :-) 23:08
i wasn't involved in nativecall, mind you, but still
23:09 logie joined 23:11 matija left, sjohnson left, tadzik left, sjohnson joined 23:12 stux left, tadzik joined, stux joined 23:13 matija joined, logie__ joined, logie__ left 23:16 brrt left 23:18 logie__ joined, logie__ left 23:22 rurban left 23:24 brrt joined, brrt left 23:33 immortal joined 23:35 erkan left 23:38 xenoterracide joined 23:42 gfldex left 23:53 jnap left, logie left 23:58 epochbell joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!