»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, or /msg camelia p6: ... | irclog: irc.perl6.org or colabti.org/irclogger/irclogger_logs/perl6 | UTF-8 is our friend!
Set by moritz on 22 December 2015.
00:00 chenryn left 00:07 cdg left 00:08 cdg joined
TEttinger .u SELFIE 00:10
yoleaux No characters found
00:10 RabidGravy left 00:12 cdg left 00:17 skingsbu left 00:21 Ven left
geekosaur not till the next unicode update :p 00:21
00:22 BenGoldberg joined, Zoffix left 00:23 BenGoldberg left, Zoffix joined 00:24 BenGoldberg joined 00:29 BenGoldberg left 00:30 BenGoldberg joined 00:32 BenGoldberg left 00:33 BenGoldberg joined 00:38 BenGoldberg left 00:40 BenGoldberg joined 00:42 BenGoldberg left 00:43 BenGoldberg joined 00:44 BenGoldberg left 00:45 BenGoldberg joined
skink Is there an official way to make arbitrary syscalls yet? 00:53
00:56 BenGoldberg left 00:57 yqt left, BenGoldberg joined 00:59 kurahaupo left
BenGoldberg skink, You can call arbitrary library functions using NativeCall. 01:03
skink BenGoldberg, It's not in a library file :) 01:04
01:04 BenGoldberg left 01:05 Zoffix__ joined 01:06 BenGoldberg joined 01:07 Zoffix__ is now known as sexy-coder-girl
skink BenGoldberg, It's not in a library file :) 01:09
01:09 Zoffix left 01:10 sexy-coder-girl is now known as Zoffix, Zoffix left, Zoffix joined
BenGoldberg How would you perform this syscall from a C program? 01:10
skink man7.org/linux/man-pages/man2/getrandom.2.html 01:11
01:11 kurahaupo joined
geekosaur syscalls can't really be invoked portably 01:12
01:12 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg 01:16 kurahaupo_ joined, kurahaupo` joined 01:17 espadrine left, BenGoldberg left 01:18 kurahaupo left 01:19 BenGoldberg joined
skink geekosaur, I already separated OS-specific functionality into separate modules 01:19
01:20 kurahaupo_ left
geekosaur man7.org/linux/man-pages/man2/syscall.2.html 01:25
01:28 huggable left
skink Disgusting :) 01:28
01:28 huggable joined, kurahaupo` left 01:32 _4d47 left 01:33 Zoffix left 01:34 TimToady left 01:35 Zoffix joined, BenGoldberg left, jdv79 left, ggoebel115 left, Celelibi left, jdv79 joined 01:36 remmie left, ggoebel115 joined, BenGoldberg joined 01:37 Praise left, pierrot_ joined, BenGoldberg left 01:39 BenGoldberg joined, Praise joined, Praise left, Praise joined, BenGoldberg left 01:40 TimToady joined 01:41 BenGoldberg joined, Celelibi joined, pierrot left, skink left, tbrowder joined
tbrowder back again, ref pod and docs 01:42
01:42 BenGoldberg left 01:43 BenGoldberg joined, araujo_ left 01:46 ilbot3 left, remmie joined
tbrowder pmurias: current p6 pod is not handled completely, e.g., tables. after loking at some other markup langs again, I think, given the present state of things, p6 pod (and tools) is okay but needs more attention. I have just submitted an e-mail to the perl6 users presenting my simple take on my impression of p6 pod and some possible projects. I would appreciate any commennts on that thread. 01:47
01:47 BenGoldberg left, ilbot3 joined
tbrowder bye all 01:48
01:48 tbrowder left, kurahaupo joined, BenGoldberg joined 01:49 BenGoldberg left 01:50 BenGoldberg joined 01:51 molaf left
Zoffix Any response to this FB thread? www.facebook.com/groups/1595443877...%22R%22%7D 01:55
"Looks like C++ is not supported yet. Is that correct?" "Yep, thats what I was referring to. It did not work for me. May be it is work in progress?"
01:55 BenGoldberg left 01:57 BenGoldberg joined, kanishka left 01:59 ssotka left 02:04 molaf joined 02:05 jack_rabbit joined, kurahaupo left, noganex joined 02:08 noganex_ left 02:11 kid51 left
tailgate are there examples of how to make a class/static method in a perl6 class? 02:12
02:12 dimon__ joined
dimon__ hey, what's the difference between "unit module Aaa;" and "unit package Aaa;" ? 02:13
geekosaur tailgate, our method foo(...) { ... }
02:19 BenGoldberg left 02:20 dimon__ left, BenGoldberg joined 02:21 dimon__ joined
dimon__ hey, what's the difference between "unit module Aaa;" and "unit package Aaa;" ? 02:21
02:21 BenGoldberg left 02:22 BenGoldberg joined 02:23 pierre_ joined
Zoffix Goood question. 02:25
The speculation on the subject seems a bit outdated: "package Foo; ... This form is illegal in a Perl 6 file. If you wish to have a file-scoped package, either use the brace form or declare it with the module keyword instead." 02:26
(and we now have `unit` keywords)
02:29 pierre_ left
dimon__ Zoffix: I see. 02:29
And if I have file1.pm6 and file2.pm6 in my library MyComp::MyLib, what should be the 1st on those file?
unit module MyComp::MyLib;
or
unit module MyComp::File1Class;
assuming that
I define those classes as 02:30
classMyComp::MyLib::File1Class
within those 2 modules
02:30 BenGoldberg left
dimon__ correction: class MyComp::MyLib::File1Class { .... } 02:30
or
class MyComp::MyLib::File2Class { .... }
sorry for the typos 02:31
02:31 BenGoldberg joined 02:32 jack_rabbit left 02:35 BenGoldberg left 02:36 BenGoldberg joined 02:47 BenGoldberg left 02:48 BenGoldberg joined 02:53 Ben_Goldberg joined, BenGoldberg left, Ben_Goldberg is now known as BenGoldberg 02:55 dimon__ left 02:56 BenGoldberg left 02:58 BenGoldberg joined 02:59 BenGoldberg left 03:00 BenGoldberg joined 03:01 |2701 left 03:08 cdg joined 03:12 cdg left 03:15 BenGoldberg left 03:16 BenGoldberg joined 03:17 BenGoldberg left 03:18 BenGoldberg joined 03:19 BenGoldberg left 03:21 BenGoldberg joined 03:22 ssotka joined 03:27 BenGoldberg left, dimon_ joined
dimon_ when I'm using a file-class declaration "unit class MyModule::Class1", how can I write "does RoleX" for Class1? 03:28
03:28 BenGoldberg joined
timotimo m: unit class FooBar; also does Positional; 03:29
camelia ( no output )
dimon_ timotimo: and if I need to export the module where "Positional" is defined? say, it's defined in MyModule 03:31
then this won't compile also does Positional; 03:32
timotimo export? you mean import?
you can put "use MyModule" in front
dimon_ neither this also does MyModule2::Positional;
I mean, MyModule2, not Module for Positional
timotimo i'm not sure how you mean that
dimon_ I mean, this won't compile "; also does MyModule2::Positional;" 03:33
timotimo so why not put "use MyModule2" in between?
03:33 khw left
timotimo oh, you mean you want Positional to be available as a member of the unit class' package? 03:33
dimon_ tnx
timotimo you'd have to do something like "my \Positional is export = Positional" 03:34
or something like that
dimon_ nope, I meant how to "use" MyModule2::Positional
before "also does MyModule2::Positional"
timotimo the thing after "use" is the file you want rakudo to look for in the filesystem
so if you want to get MyModule2::Positional, you'd usually "use MyModule2" 03:35
03:36 BenGoldberg left 03:37 BenGoldberg joined
dimon_ tnx 03:38
03:38 dimon_ left, BenGoldberg left 03:40 BenGoldberg joined, dimon_ joined, pierre_ joined
dimon_ by the way, if my library is available in "panda", and if I update its source code at github, will it be updated in panda automatically? 03:41
03:41 BenGoldberg left
dimon_ does panda use code from github of a package? 03:41
timotimo panda always pulls the code directly from github, yeah
dimon_ ok
03:41 dimon_ left
timotimo we will have something different in the future that's more similar to cpan and PAUSE 03:41
03:42 BenGoldberg joined 03:43 BenGoldberg left 03:45 BenGoldberg joined 03:48 BenGoldberg left 03:50 BenGoldberg joined 03:51 BenGoldberg left 03:52 BenGoldberg joined 03:53 pierre_ left 03:54 BenGoldberg left 03:55 BenGoldberg joined 03:56 BenGoldberg left 03:57 BenGoldberg joined 03:58 BenGoldberg left 04:05 ssotka left 04:06 BenGoldberg joined 04:07 BenGoldberg left 04:08 BenGoldberg joined 04:10 BenGoldberg left, mtj_ left 04:11 BenGoldberg joined 04:14 BenGoldberg left, mtj_ joined 04:15 BenGoldberg joined 04:17 BenGoldberg left 04:18 BenGoldberg joined 04:27 BenGoldberg left 04:32 BenGoldberg joined 04:39 dmc00 left 04:48 BenGoldberg left 04:50 Cabanossi left 04:51 Cabanossi joined 04:53 espadrine joined 04:54 BenGoldberg joined, pierre_ joined 04:58 BenGoldberg left 05:00 pierre_ left
dalek line-Perl5: 0dbc189 | (Zoffix Znet)++ | README.md:
Newer version of P5 available now
 05:02
line-Perl5: 9d70afd | lizmat++ | README.md:
Merge pull request #62 from zoffixznet/patch-1

Newer version of P5 available now
05:03 wamba joined, BenGoldberg joined 05:04 mr-foobar left 05:05 BenGoldberg left, dmc00 joined 05:07 dmc00 left 05:10 BenGoldberg joined 05:14 dmc00 joined, CIAvash joined 05:16 BenGoldberg left 05:21 dmc00 left 05:22 BenGoldberg joined 05:23 BenGoldberg left 05:28 huggable left, huggable joined 05:29 BenGoldberg joined 05:31 BenGoldberg left 05:36 BenGoldberg joined 05:39 BenGoldberg left 05:44 BenGoldberg joined 05:47 kaare_ joined 05:49 BenGoldberg left 06:01 domidumont joined 06:06 domidumont left 06:07 domidumont joined 06:21 spider-mario joined 06:23 grondilu joined 06:26 lizmat joined 06:27 espadrine left 06:32 kaare_ left 06:42 telex left 06:44 telex joined, wamba left 06:53 lizmat left 06:59 jack_rabbit joined 07:07 CIAvash left 07:16 molaf left 07:18 CIAvash joined 07:57 firstdayonthejob joined 08:10 RabidGravy joined 08:15 darutoko joined 08:16 kaare_ joined 08:20 Ven joined 08:30 grassass left 08:36 rindolf joined 08:38 grassass joined 08:41 pierre_ joined, sortiz left 08:44 Emeric joined 09:03 pierre__ joined, pierre_ left, Ven left 09:06 cpage_ joined
masak good almost-noon, #perl6 09:21
09:22 espadrine joined, buharin joined
RabidGravy our cat decided it was time for us to get up an hour and a half ago 09:22
buharin hiho
09:27 lue joined 09:28 huggable left, huggable joined, espadrine left, ShimmerFairy left, lue is now known as ShimmerFairy
grondilu just adding a 'BEGIN say "hi";' in one of my modules make it fail. Go figure. 09:33
can't reproduce it with a one liner though. 09:34
m: module A { our @a is export = rand xx * }; import A; say "ok";
camelia rakudo-moar ad8265: OUTPUT«ok␤»
grondilu m: module A { our @a is export = rand xx * }; BEGIN say "hi"; import A; say "ok";
camelia rakudo-moar ad8265: OUTPUT«hi␤ok␤»
grondilu something like that
it's for my clifford module (github.com/grondilu/clifford). I was trying to do some precomputation with a BEGIN block but even the simplest BEGIN generates: ==SORRY==Variable '@e' is not declared 09:36
also after this error removing the BEGIN say "hi"; is not enough. I have to manually remove the .precomp directory. 09:39
very LTA
RabidGravy grondilu, iirc doing IO in a module that may be precompiled causes badness
grondilu aren't all module susceptible to be precompiled? 09:40
*modules
RabidGravy that is if the IO will happen at compile time
yes
grondilu oh
hang on
grondilu tries with 'BEGIN sleep 1;' 09:41
nope, BEGIN sleep 1 causes the same error and requires manual remove of .precomp just the same. 09:42
grondilu wonders if BEGIN {} would just be ignored. 09:43
no, it's not ignored and it causes the error as well. 09:44
09:45 CIAvash left 09:47 buharin left, cpage_ left 09:49 CIAvash joined 09:50 zakharyas joined
grondilu :q 09:50
but seriously I have hard times figuring why adding 'BEGIN {}' in a module could make any difference whatsoever. 09:54
09:54 TEttinger left, skingsbu joined
dalek rl6-most-wanted: b49b9f9 | titsuki++ | most-wanted/modules.md:
Fix a Text-Tabs-Wrap link
 09:56
rl6-most-wanted: 53ac1e0 | titsuki++ | most-wanted/modules.md:
Merge pull request #31 from titsuki/fix-not-working-link

Fix a Text-Tabs-Wrap link
masak grondilu: I'd be quite interested to see a golf of what you're observing.
RabidGravy yeah I deeply suspect there is something else going on
llfourn writing to stdout during compile time used to be a problem but I thought it was more or less fixed now 09:57
09:57 kaare_ left
llfourn I always use BEGIN note "recompiling"; for that reason 09:58
grondilu masak: so would I. I've tried to make one but I failed. 10:01
I could try harder I guess.
10:01 pierre__ left 10:02 pierre_ joined
grondilu last time I tried I gave up when I realized the code I was executing was not the code I had typed because of the .precomp issue. 10:02
m: BEGIN say "compiled on " + now 10:08
camelia rakudo-moar ad8265: OUTPUT«5===SORRY!5=== Error while compiling /tmp/GRX4YX4MEy␤An exception occurred while evaluating a BEGIN␤at /tmp/GRX4YX4MEy:1␤Exception details:␤ Cannot convert string to number: base-10 number must begin with valid digits or '.' in '3⏏5compile…»
10:08 pmurias joined
grondilu silly me 10:08
m: BEGIN say "compiled on " ~ now
camelia rakudo-moar ad8265: OUTPUT«compiled on Instant:1463306963.554297␤»
grondilu ^I guess I could use somethink like that to check that my code is up-to-date or something. 10:09
dogbert17 good day #perl6 10:10
pmurias dogbert17: hi
dogbert17 pmurias: hi 10:11
noob question of the day, what does this syntax mean: my class X::Seq::Consumed { ... } # from src/core/Seq.pm
grondilu it's the standard way of making an exception IIRC 10:12
10:12 cpage_ joined
dogbert17 shouldn't ther be an 'is Exception' in there somewhere? 10:13
grondilu I think it was not considered necessary 10:14
there is no Exception class in Perl 6 10:15
m: say Exception
camelia rakudo-moar ad8265: OUTPUT«(Exception)␤»
grondilu oh there is
my bad
I don't know then. 10:16
RabidGravy dogbert17, that is a "forward declaration" 10:18
bartolin dogbert17: I think that is a stub definition for that class, which get fully defined in src/core/Exception (and there you get 'is Exception' as you expected) 10:19
RabidGravy it's actually defined in src/core/Exception.pm
dogbert17 RabidGravy++ bartolin++ many thanks
grondilu++ as well 10:20
RabidGravy it can do it like that (in separate files) because of the way that the CORE setting is compiled
dogbert17 aha, the reason for asking is that I want to try to fix github.com/perl6/doc/issues/515 10:21
hadn't seen the syntax before and thought, for some reason it meant 'not yet implemented' 10:22
RabidGravy yeah I was going to respond to that, I think the time is very soon that the exceptions should have their own section in the docs, on the grounds that if they are *all* documented it will swamp all the other types
dogbert17, it means "not implemented here" :) 10:23
or "yet"
dogbert17 RabidGravy: :) 10:24
and I'm guilty as charged :), have added several 'missing' exceptions to the docs in the last few days 10:25
RabidGravy that is to say that if by the time the whole thing is finished compiling there isn't a complete definition a "he following packages were stubbed but not defined" will be thrown
arnsholt In user code, you'd end up using that syntax if you have two types that depend on each other
10:26 araujo joined
arnsholt Like "class B {...}; class A { has B $.b; }; class B { has A $.a }" 10:26
bartolin dogbert17: you can read more about stub declarations in the speculations: design.perl6.org/S06.html#Stub_declarations
arnsholt (And no, that can't be split into two separate files)
dogbert17 thx all, very interesting discussion 10:27
10:27 aries_liuxueyang left 10:29 aries_liuxueyang joined
RabidGravy the thing with "document all the exceptions" is that there are greater than 250 defined in src/core/Exception.pm which is roughly more than the rest of the types 10:30
so a) the types page becomes rather large and b) they are mostly not all that interesting except in the context of *where* they are thrown 10:31
so having them in a separate section would prevent a) and allow the linking from the documentation for the things that throw them 10:33
dogbert17 RabidGravy: maybe I should hold off then until we have a solution ? 10:34
RabidGravy I'd carry on to be honest 10:35
dogbert17 RabidGravy: ok will do :) 10:36
RabidGravy I'd say that people agreeing will take forever or until someone just does it and it's better that things are documented, it can be re-arranged later
dogbert17 RabidGravy: how about =SUBTITLE 'Error due to trying to reuse a consumed sequence' is that intelligible or just bad english? 10:39
10:44 buharin joined 10:45 Actualeyes joined 10:47 kaare_ joined
gfldex the optimisations of the last 2 weeks have shaved off 13% execution time of my pod renderer 10:55
you totally improved my productivity!
11:01 Roamer` left 11:02 fildon joined
RabidGravy yay! 11:10
dalek c: 9728090 | (Jan-Olof Hendig)++ | / (2 files):
Added docs for X::Seq::Consumed
 11:11
RabidGravy yay!
dogbert17 :)
yesterday an odd thing happened when writing some docs, after having run htmlify.p6 my changes weren't there, i.e. in the generated html files. Had to nuke the 'precompiled' directory in order to make it work! 11:15
teatime waves g'morning.
dogbert17 o/ teatime 11:16
11:17 kid51 joined 11:18 kaare_ left 11:19 [TuxCM] joined 11:23 zakharyas left 11:27 pierre_ left 11:32 mr-foobar joined 11:36 nightfrog left 11:40 grassass left
llfourn hmm I have to keep deleting .precomp because I get "Missing or wrong version of dependency" often. Seems to have only started today. 11:46
I will see if I can bisect it 11:47
11:50 kurahaupo joined
llfourn seems to happen on 2016.03 as well o.o 11:56
RabidGravy :-\ 11:57
llfourn not using git so hard to know what I did to trigger it.
12:00 snarkyboojum joined 12:01 snarkyboojum left 12:04 mr-foobar left
Zoffix I was getting that when I installed a module and stuff depending on it and then got a local copy of the module to hack on. So basically the installed, precompiled modules were precompiled against the "old" installed module, not my local copy. 12:06
llfourn hmmm that interesting because a few of the compunits in what I'm working on do have the same name as some that are installed 12:07
Zoffix And I reported it: rt.perl.org/Ticket/Display.html?id=128088
12:08 snarkyboojum joined 12:12 pierre_ joined
tailgate if I have ((a -1) (e -5) (d -4) (c -3) (b -2)) how do I make it a hash of the form a => -1 e => -5 etc? 12:12
12:13 mr-foobar joined
llfourn tailgate: I'm not sure what you have there but try .hash 12:13
if it's a list of lists which is what it looks like you may have to flatten them first
12:13 _4d47 joined
llfourn m: ( <a,b>,<c d> ).perl.say 12:14
camelia rakudo-moar ad8265: OUTPUT«("a,b", ("c", "d"))␤»
llfourn ...
m: ( <a b>,<c d> ).perl.say
camelia rakudo-moar ad8265: OUTPUT«(("a", "b"), ("c", "d"))␤»
llfourn m: ( <a b>,<c d> ).hash.say
camelia rakudo-moar ad8265: OUTPUT«{a b => (c d)}␤»
llfourn m: ( <a b>,<c d> ).map(|*).hash.say 12:15
camelia rakudo-moar ad8265: OUTPUT«{a => b, c => d}␤»
llfourn ^ seems to do the trick
m: hash( (<a b>,<c d>) ).say 12:16
camelia rakudo-moar ad8265: OUTPUT«{a => b, c => d}␤»
llfourn ^ better
tailgate ah, I was using .Hash 12:17
llfourn m: %( (<a b>,<c d>) ).say # also works I think
camelia rakudo-moar ad8265: OUTPUT«{a b => (c d)}␤»
llfourn m: %( |(<a b>,<c d>) ).say # or this perhaps?
camelia rakudo-moar ad8265: OUTPUT«{a b => (c d)}␤»
llfourn ..nevermind 12:18
.Hash should work too I think
m: ( <a b>,<c d> ).map(|*).Hash.say
camelia rakudo-moar ad8265: OUTPUT«{a => b, c => d}␤»
llfourn at least there it does
12:19 snarkyboojum left 12:20 CIAvash left, [Tux] joined
tailgate what do | and |* mean? 12:21
llfourn | means turn the thing after it into a Slip ( an auto-flattening list )
* in this case means $_, which means each item in the map
so .map(|*) is short for .map({ |$_ }) 12:22
the problem is without it the .Hash gets two Lists as arguments where as what you want it four elements which will then pair up as key-values. 12:23
s/want it/want is/
12:33 pierre_ left 12:35 [TuxCM] left 12:46 mr-foobar left
tailgate thanks llfourn 12:47
llfourn np :)
12:47 mr-foobar joined 12:52 pierre_ joined 12:56 kanishka joined 13:00 Ven joined
Zoffix (╯°□°）╯︵ ┻━┻ 13:06
zef fails to install Inline::Perl5 with a non-descript error... tried with panda and it fails while trying to install prereqs because they're already installed -_- 13:07
timotimo hey zoffzoff
Zoffix hey
13:11 kaare_ joined
Zoffix Oh, pffftt... I forgot to perlbrew switch to proper perl, that's why Inline::Perl5 is failing. panda++ (actually showing the error) 13:13
13:14 CIAvash joined
teatime I am curious about how exactly Inline::Perl5 works / is implemented. 13:15
Zoffix I think it uses NativeCall to call into Perl5's C lib that handles XS.
RabidGravy yeah, it's something like that 13:18
teatime cool. 13:21
timotimo it might be easier to understand how Inline::Python works. i bet the python API is simpler than XS is :P
RabidGravy well the P5 "embedding" API is quite well documented perldoc.perl.org/perlembed.html 13:22
13:24 domidumont left 13:28 huggable left, huggable joined 13:29 Zoffix left 13:30 Zoffix joined
Zoffix What's a good monospace font that has good unicode support and doesn't look like shit? 13:30
RabidGravy just in case anyone is on a "search and document" mission for exceptions I've created github.com/perl6/doc/issues/517 to cover off the rest of them as there are loads
teatime Zoffix: PragmataPro, but it's not free, and actually is kindof expensive IMO. 13:31
Zoffix :(
I like free
13:31 kaare_ left
teatime Zoffix: Are you more interested in various scripts, or in Latin (and maybe Greek, Cryllic) + *symbols* ? 13:31
timotimo and emoji!
Zoffix Well, I'm interested in not seeing boxes with numbers in them: ¯\_(ツ)_/¯(╯°□°）╯︵ ┻━┻ 13:32
nine I think the basic parts of Inline::Perl5 and Inline::Python are equally easy to read
teatime Zoffix: Is this (primarily) for use in a terminal emulator (like xterm or rxvt-unicode)? 13:33
Zoffix This is what I see with the above: i.imgur.com/0G0HF0l.png
timotimo Zoffix: so no support for those boxed numbers emoji?
Zoffix I have 'Monospace' font that seems to have better support than Deja Vu, but it looks atrocious. 13:34
timotimo as a kid i loved "fixedsys" 13:35
but unicode wasn't so much a thing back then :P
teatime if you were on Linux, "monospace" would be a, like, "virtual" font, and inside the (terribly over-complicated) fontconfig config files (which you can customize via iirc ~/.fonts.conf), there would be defined a primary font and then a series of fallback fonts used to implement it. but I have nfc how that works on windows. 13:37
(it wouldn't surprise me if, for once, the Windows way to accomplish the same thing was actually more sane than the Linux way)
mst oh gods fonts 13:38
timotimo fonts are a Mighty Sucky Thing, aren't they, mst? :) 13:40
geekosaur I would suspect windows does much the same but through the registry and using uuids
13:42 Su-Shee joined
Su-Shee good afternoon. who's the person to talk to about the buildscript of rakudo-star? 13:42
Zoffix \o 13:43
Su-Shee Zoffix: is that a "it's me"? :)
Zoffix That's a hello :)
Su-Shee Zoffix: :) 13:44
Zoffix I dunno if there's one person. Just ask the question, I guess.
There's also #perl6-release, I believe.
Su-Shee Zoffix: too long, I just wanted to prepare said person(s) for a well not quite bug report, but for some buildscript cleanup thing mr shee is on the way of submitting to github...
CIAvash Zoffix: In my editor, it's falling back to DroidSansFallback to show those characters 13:45
Su-Shee Zoffix: or, rather not too long "something with DESTDIR support is broken" - but what exactly is too long :))
13:45 jameslenz joined
Zoffix :) 13:45
mst Su-Shee: heh, DESTDIR fixes are what got my first patch in 13:46
Su-Shee he tried to fix stuff, but after one fix the next problem came up so he stopped ;)
mst also didn't somebody create a perl6-dev channel for rakudo hacking?
timotimo we have #p6dev
mst oh for fuck's sake that's why I couldn't find it
can people please try and get the naming right? 13:47
mst sobs
timotimo right, it deviates from the ... :D
oh my
geekosaur iirc the release script was on a "this quick hack worked for me, if it breaks for you then you get to keep the pieces" basis
Zoffix +1 I too couldn't find it the first time for the same reason
RabidGravy there there, have a ponie
Su-Shee mst: I'm not quite up2date, I essentially just politely wanted to warn $person or if it is somebody I know, directly point person A to person B. ;)
13:47 Zoffix left
mst Su-Shee: oh, yeah, totally 13:48
Su-Shee geekosaur: well it DOES build things, but it has quirks making it hard to make proper packages with the use of DESTDIR. that's all.
13:48 Zoffix joined
Su-Shee I only know because I just made mr shee shorten the 5km long report. ;) 13:49
mst lol
timotimo Zoffix: perl6.party runs on rakudo? it seems surprisingly snappy compared to last we spoke
mst I don't think it does anymore
there was a memory leak, so he rewrote everything
Su-Shee mst: yeah well he was in the middle of explaining what pkgsrc is/does and putting in footnotes and that's never a good sign ;)
timotimo dang. 13:50
mst Su-Shee: probably not strictly necessary, we do actually know what DESTDIR is, it just doesn't always get exercised as often as it should
Zoffix timotimo, yeah huge memory leaks (irclog.perlgeek.de/perl6/2016-05-08#i_12449184), I saved the P6 version in a separate branch and rewrote the whole thing with Mojolicious. 13:51
Su-Shee mst: I'll just let him submit his stuff, what do I care if netbsd has a proper rakudo package build process :))
Zoffix Well, that and the fact that you can't have async stuff with Bailador. The in-browser code samples could take up to 20 seconds to run 13:52
timotimo Zoffix: were you still evaling the templates every time a page was requested?
13:52 Ven left
Zoffix CIAvash, the one download I found tells me "Doesn't look like a valid font file" when I try to install it on Windows :( 13:53
mst Su-Shee: do tell me once the PR's in, I'm something of a netbsd/pkgsrc fan and I'd quite like to keep an eye on it
Zoffix timotimo, yeah. Though first I had is cached trait on sub that gave me rendered templates. But it's not just that app that leaks. huggable leaks. Its restarter script leaks 13:54
timotimo mhm
Zoffix At least leaked at that time. I've not re-tried after the fixes you pushed to MoarVM
mst I'd've rewritten the restarter script in p5
timotimo the fixes i made were almost exclusively to fix --full-cleanup to give better results, fwiw 13:55
13:55 [Tux] left
Zoffix Ah 13:55
timotimo not for regular running
13:55 _mg_ joined
timotimo so you could use valgrind more helpfully 13:55
13:55 kurahaupo left
timotimo but i do agree we're not doing extremely well with memory usage at the moment 13:55
i guess i'll check out huggable and look if i find something obvious in the heap explorer 13:57
CIAvash Zoffix: I'm using this package for Arch Linux, www.archlinux.org/packages/communi...ttf-droid/
Zoffix It seems the leakage happens when she's addressed and generates a response
timotimo oh lord 13:58
cloned huggable, "panda installdeps ."
SSSPPPPAAAAAAAMMMMMmmmmmmm
CIAvash Zoffix: Not sure it helps, www.archlinux.org/packages/communi.../download/
mst SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM
timotimo ==> huggable depends on IRC::Client, IRC::Client, huggable, IRC::Client::Plugin::Factoid, IRC::Client
who sees the problem here ..
13:59 JustinHitla joined
Su-Shee mst: oh, it seems the problem is already reported at least twice. 13:59
tadzik I do :)
timotimo huggable depends on ...Plugin::Hug, which is provided by huggable
Zoffix huh
tadzik it's fixed in a branch, fwiw
wait what
13:59 kurahaupo joined
Su-Shee mst: github.com/rakudo/star/issues/65 and the link to the "Open" thing "Rakudo Star: Installation of modules doesn't respect DESTDIR" 13:59
timotimo IRC::Client::Plugin::Hug is in its depends section
Zoffix Oh, I confused provides with depends :)
timotimo hah
RabidGravy doh
mst Zoffix: you always were a little backwards 14:00
Zoffix Fixed...
timotimo xiffoz
fixfoz
fixzof
JustinHitla I have some text, and I want to get all words from it that contains only specific lettes, "a b c d e f g h i j k l m n" so words that fit are "leaf, leen" but not "leave, noun" 14:01
Zoffix JustinHitla, .words.grep: /<[tehletters]>/
timotimo so you'll write a regex like / << <[a b c d e f g h i j k l m n]> >> /
or that
kurahaupo "leave" would appear to contain both a and e.
JustinHitla so I don't even need perl ? I can use grep tool ?
Zoffix heh 14:02
JustinHitla, I guess... but .grep is a method in Perl 6
JustinHitla kurahaupo: yes, a and e are allowed
kurahaupo JustinHitla: do you mean "starts with"?
JustinHitla no, contains any of "allowed" letters
Zoffix m: say 'The stuff with words'.words.grep: /<[wu]>/
camelia rakudo-moar ad8265: OUTPUT«(stuff with words)␤»
JustinHitla I mean "consist" of
kurahaupo JustinHitla: so explain your "not" example 14:03
timotimo i forgot to put in a + after the <[ ... ]> in my regex
JustinHitla kurahaupo: other way around, if a word contains at least one letter not in the allowed list it doesn't fit
Zoffix for "consists of" just negate the match to exclude words that have letters you don't own
timotimo kurahaupo: leave contains v, which is not in the list of allowed letters
teatime pokes Zoffix
kurahaupo oh, comprises only those letters. just search for anything else and exclude the word
14:04 pierre_ left
kurahaupo like ! <[^a-n>] 14:04
>
timotimo no, we spell that <-[a-n]>
Zoffix: what's with the "use lib" in bin/huggable.p6? 14:05
kurahaupo grep -v '[o-z]' on the command line 14:06
timotimo oh
Zoffix timotimo, I was hacking on IRC::Client
timotimo won't "use lib 'lib'" in a bin/foo.p6 actually point to bin/lib? 14:07
JustinHitla kurahaupo: but first I need to split the text word by word then
Zoffix timotimo, I start it with perl6 bin/foo.p6
14:07 pierre_ joined
Zoffix timotimo, actually IRC::Client has a bug that might prevent you joining: github.com/zoffixznet/perl6-IRC-Cl...nt.pm6#L31 14:08
Sometiems messages get split onto multiple lines and that MOTD end message never gets reporter
*ed
This new keyboard is starting to get on my nerves :(
I don't know why there are such giant gaps between the keys 14:09
14:09 Su-Shee left
Zoffix (and keys themselves tiny) 14:09
timotimo hmm
JustinHitla Zoffix: you can always look at keyboard when typing
Zoffix watches timotimo descend into piles of half-arsed Zoffix-ware :) 14:10
timotimo the heap dump profile segfaults because it's stumbling over a somehow-invalid object
Zoffix JustinHitla, no thanks :)
RabidGravy over-rated 14:11
timotimo wonder how long it'll take my huggable to join my channel when run under valgrind
14:11 _mg_ left
timotimo aha! 14:12
14:12 _4d47 left
timotimo i now see it sloooowly process each individual message :D 14:12
Zoffix :) 14:15
timotimo it's about 1/2 into the initial welcome stuff
Zoffix How are you gonna kill it? 14:16
timotimo ctrl-c should be enough
i actually forgot --full-cleanup
that's not smart
Zoffix Won't doing that make valgrind show everything as leaked?
¯\_(ツ)_/¯
Zoffix loves the new single-key emoticons :D 14:17
timotimo i was hoping it'd show what or how the object got b0rked. but maybe i'll actually perl6-valgrind-m --profile=heap to get the crash reproduced guaranteed
kurahaupo JustinHitla: If you already have your words in a Perl (version 6) data structure, then this is the channel to ask. If they're in a file somewhere, ask #bash. If you don't know which version of Perl you're using, you're probably in the wrong place
timotimo the thing is, it's probably not the kind of leak valgrind would show that happens in the huggable case
it's probably objects being kept alive by something we've not thought about clearing out properly 14:18
for example, when spesh analyzes a piece of code, it'll log a bunch of values. if that piece of code doesn't run often enough to be fully specialized, that record will stick around until the program exits, keeping alive all those intermediate values that are otherwise uninteresting
i didn't have to kill huggable 14:20
14:20 |2701 joined
timotimo it took too long to respond to a PING and the connection was closed 14:20
Zoffix heh
CIAvash, thanks. Got the files, but they still don't install on Windows. 14:21
14:21 mr-foobar left
Zoffix teatime, heh, PragmataPro only costs 200 euros. That's not expensive for a font :) 14:27
At $work, I'm using Galaxy Polaris, costs $300 and we used to use Helios, which costs like $500 for all styles :) 14:28
timotimo so crashy :\
14:30 hotel_california joined
Zoffix Crashy :S 14:30
14:31 hotel left
timotimo we shouldn't be encountering b0rked objects in the heap profiler. something must have a pointer that's invalid and wouldn't be used normally but the heap analyzer is stumbling over it for some reason .. or something like that? 14:31
like, if the heap collector crashes, the regular GC should be crashing at the same point in a very similar way
14:32 Zoffix left, Zoffix joined 14:33 CIAvash left 14:43 cdg joined 14:44 crucialr_ joined, crucialr_ left, khw joined 14:45 Zoffix left, cdg left, Zoffix joined, crucialr_ joined
timotimo ooooh 14:47
could it be the heap profiler doesn't know about "reframe" yet?
14:47 kurahaupo left
timotimo that could totally be a thing 14:47
14:47 JustinHitla left
timotimo hm, but it mostly re-uses what the gc uses, too ... 14:48
in any case, it crashes reliably :P
Zoffix :) 14:49
14:51 pierre_ left
RabidGravy Oooh, I've found the original code to interface with www.velleman.eu/products/view/?coun...;id=351346 - so it might be able to knock something up without needing to bind all of libusb 14:52
Zoffix Heh: "You may write custom Windows (98SE, 2000, Me, XP) applications in Delphi" 14:54
14:57 darutoko left
RabidGravy I made some code years ago in P5 that used Device::USB but it never quite worked right and I was never sufficiently motivated to work out why 14:58
15:13 wamba joined 15:27 kanishka left 15:29 sortiz joined 15:30 kurahaupo joined
sortiz \o #perl6 15:31
RabidGravy harr! 15:37
15:40 itcharlie joined
RabidGravy github.com/jonathanstowe/Device-Ve...055-Native - but I think it would be easier just compile a little helper 15:40
15:43 pmurias left 15:48 dmc00 joined 15:49 dmc00 left
MadcapJake I wonder how difficult it would be to implement this for Perl 6: blog.skylight.io/introducing-helix/ 15:50
15:51 sena_kun joined, sena_kun left
mst was planning to do that for perl5 at some point 15:52
glad somebody made me some prior art
teatime Zoffix: fair enough; I'm not a designer (but I am quite interested in design and typography, just not well educated about them), nor do I ever buy fonts. But since most people only really want it for use in their own personal development environment (and not for use in print layouts, or as a webfont, or for embedding in an application they distribute), $200 seems kinda steep. However, it does have a very 15:53
accomodating (for a non-free font) license which allows all of the things I mentioned.
MadcapJake heh, yeah I have no experience with Rust but the author makes a great point that if you're going to provide a module that has some native extensions, shouldn't it be something that you can guarantee will not segfault? 15:54
teatime Zoffix: fwiw, I do beleive it's worth the price, like, intrinsically. just maybe not worth it for all of the people who would like to use it, if it were free.
RabidGravy MadcapJake, where would the fun in that be 15:57
MadcapJake haha 15:58
teatime Zoffix: also, I know you said it was OBE, but I sent you an add'l privmsg a while back. 16:01
16:04 stmuk_ is now known as stmuk, buharin left 16:08 cognominal joined 16:09 buharin joined 16:22 Zoffix left, Zoffix joined 16:44 itaipu joined
mspo how can you guarantee it won't segfault if the c code accesses memory it should not? 16:46
crashing is good
mst mspo: eh?
mspo: this is a thin C layer that binds to rust, where the compiler prevents 'accesses memory it should not' 16:47
mspo mst: okay
mst: what does it do when libfoo attempts to read outside?
mst what 16:48
mspo nevermind :)
mst helix is ruby -> (small C shim) -> rust
there is no libfoo involved
mspo k 16:49
16:50 skingsbu left
mst the whole point here is to have what would otherwise have been implemented in C, instead implemented in rust, so the compiler can slap you if you mess up your memory management 16:52
16:52 itcharlie left 16:53 crucialr_ left, itaipu left
|2701 I like ruby, but ruby and rust are so sjw now 16:54
mst fortunately, if you stay focused on the code, the politics of the people are basically irrelevant no matter where on the spectrum they fall 16:56
MadcapJake mspo: Rust is designed to be very safe and yes as mst states, this is about implementing something in *your* project in Rust for performance (and safety versus implementing it in C) 16:57
|2701 also, rakudo is already on top of llvm, would a helix-esque product actually make p6 any faster?
mst tends to regard ruby as a shinier, prettier perl5 except with terribly limited OO
|2701: um. you use the approach to optimise CPU bound stuff
MadcapJake rakudo is not on top of llvm
|2701 oh 16:58
I thought it was
mst perl6 is built for expressiveness, obviously a language like rust is going to perform better for the tight spots
|2701 well shit we should fix that
mst so even if it was using llvm, your question makes no sense
MadcapJake |2701: moarvm.org 16:59
mst rakudo uses moarvm, which is written in C, I'm not sure why you think 'on llvm' is a thing?
llfourn I've never heard of 'on llvm' either but this post sorta implies it's a thing wrt rakudo: brrt-to-the-future.blogspot.com.au/...-news.html 17:04
RabidGravy oh llvm can be an abstract JIT layer for compilers? cool 17:07
MadcapJake |2701: yes a helix-esque product used when some Perl 6 code bottlenecks would definitely improve your application's performance. The nice things about Rust vs NativeCall would be: [a] Rust is very safe, [b] Rust's compiler allows "zero-cost abstractions", and [c] Rust is modern and around the same performance of C. 17:08
|2701 rust is soaked in sjw nonsense though. it's possible to leverage llvm without resorting to Rust 17:09
17:09 itaipu joined
llfourn I think that the point is to use Rust not LLVM? 17:09
MadcapJake |2701: seriously that's the most ridiculous argument to not use a technology
|2701 ok 17:10
RabidGravy I'm not quite sure what the sjw thing has to do with any software
mst |2701: there are a couple of rust contribs who are visibly oriented in that direction ... and lots who aren't
put the tinfoil hat down and back away slowly
RabidGravy (I had to goofle mind)
mst it doesn't 17:11
there are apparently people out there who will refuse to use a piece of software because it doesn't have a code of conduct that aligns with their politics 17:12
|2701 absolutely
MadcapJake I'm not well-versed in LLVM but I don't think you can just "use" it, you have to create your project entirely around it and its tools.
mst |2701 is merely demonstrating that reversed stupidity is not intelligence by doing the exact same thing from the other direction
|2701 to call someone stupid because they disagree with you is itself stupid 17:13
but thanks
17:13 CIAvash joined
|2701 I never called you stupid nor insinuated anything of the sort 17:13
mst I didn't call you stupid
I called 'refusing to use a piece of software because it has an insufficiently $politics code of conduct' stupidity 17:14
I don't regard boycotting a piece of software for being "too SJW" as making sense any more than boycotting it for being "insufficiently SJW" 17:16
it's software, for flying fornication's sake
|2701 all the more reason politics should be kept out of it, imo
mst which is why your position is silly, yes :)
MadcapJake On top of that, to take a few people's outside-of-work perspective on life/liberty and equate it with an *entire* language project is also silly
17:16 notbenh left
|2701 mst, we can talk or you can call me names, but not both 17:17
i dont *particularly* care which
mst I'm not calling you names. if you want to be a professional offence taker by twisting words, go join feminist tumblr
|2701 its childish
you are though
llfourn he called "your position" silly....
mst precisely
MadcapJake How about we just leave this discussion completely off the table, it has *nothing* to do with Perl 6 and basically *nothing* to do with programming in general.
|2701 except that people are catching lawsuits and losing their jobs for not going along with this nonsense 17:18
mst ok, we're done here. this is not remotely on-topic.
|2701 so you *don't* want to sit around and call me names?
just want to be clear here 17:19
mst we're done here. this is not remotely on-topic.
MadcapJake |2701: no clarity needed, the discussion is over.
17:19 notbenh joined, damnlie left
RabidGravy but while we are being totally off-topic does someone want to bring me about ten grey doepfer patch cables (50cm) as I'm all out 17:20
MadcapJake In Wisconsin, this is about the time someone would say "How bout them Packers, eh?"
RabidGravy "weather's turned out nice again"
MadcapJake lol
psch .oO( did you see that ludicrous display last night? )
17:20 damnlie joined
arnsholt I got to show an American friend epic sax guy for the first time last time 17:21
Lulz were had!
MadcapJake arnsholt: what's epic sax guy?
MadcapJake googles
arnsholt www.youtube.com/watch?v=apav8vgfxMQ 17:22
mst MadcapJake: I hear they go well with binders
RabidGravy I hate to to think, puts me in mind of one of those Kenny G on ketamine with a delay pedal that always seem to pitch up at trance events 17:23
(see also crazy bongo guy at house clubs) 17:24
17:24 firstdayonthejob left
mst I believe the trick is to be on similar amounts of ketamine yourself 17:24
17:25 firstdayonthejob joined
RabidGravy :-O 17:25
MadcapJake lol arnsholt that is hilarious
grondilu damn eurovision is so silly 17:26
MadcapJake speaking of ketamine, apparently it's now going to be an anti-depressant :P
17:28 huggable left
grondilu how do I enumerate the elements of an enum? 17:30
RabidGravy .enums
grondilu m: my enum abc <a b c>; say abc.enums
camelia rakudo-moar ad8265: OUTPUT«Map.new((:a(0),:b(1),:c(2)))␤»
grondilu ok
llfourn m: my enum abc <a b c>; .say for abc::.keys # this too 17:31
camelia rakudo-moar ad8265: OUTPUT«a␤c␤b␤»
grondilu interesting syntax
m: my enum abc <a b c>; say abc::.WHAT 17:32
camelia rakudo-moar ad8265: OUTPUT«(Stash)␤»
grondilu never seen a (Stash) before
llfourn :: is just like .WHO
m: my enum abc <a b c>; .say for abc.WHO.keys # this too
camelia rakudo-moar ad8265: OUTPUT«a␤c␤b␤»
grondilu actually never quite tried to understand what WHO does
llfourn m: my enum abc <a b c>; say abc::c # this too
camelia rakudo-moar ad8265: OUTPUT«c␤»
llfourn gives you all child symbols of a package 17:33
grondilu k
llfourn m: .say for CompUnit::.keys
camelia rakudo-moar ad8265: OUTPUT«Handle␤PrecompilationDependency␤Loader␤DependencySpecification␤RepositoryRegistry␤PrecompilationUnit␤Repository␤PrecompilationStore␤PrecompilationId␤PrecompilationRepository␤»
grondilu not related at all, but I find it slightly confusing that the correct syntax for a multi-arg pointy block is -> $a, $b {...} while the syntax for a for loop with two parameters is for @a X @b -> ($x, $y) {...} 17:36
llfourn you don't need the ( ) do you? 17:37
grondilu I think you do
RabidGravy it unpacks differently
llfourn m: for ^10 -> $a,$b { }
camelia ( no output )
llfourn ah
m: for flat 5 X ^ 5 -> $a,$b { say $a,$b } # then do this? 17:38
camelia rakudo-moar ad8265: OUTPUT«50␤51␤52␤53␤54␤»
RabidGravy they are different things
so yeah, all good, not confusing at all 17:39
psch m: my @a = ^2 X ^2; for @a -> $a, $b { say "$a $b"; }; for @a -> ($a, $b) { say "$a $b" } 17:40
llfourn yeah it's kinda cool that you can unpack like that :)
camelia rakudo-moar ad8265: OUTPUT«0 0 0 1␤1 0 1 1␤0 0␤0 1␤1 0␤1 1␤»
psch m: my %h = :1a, :2b; for %h -> (:$key, :$value) { say "$key => $value" } # my fav example for unpacking 17:41
camelia rakudo-moar ad8265: OUTPUT«a => 1␤b => 2␤»
grondilu oh ok the syntax without the parens is ok but with a different semantics then? I'm fine with that.
it's easy to make mistake though 17:42
llfourn psch: oh you can unpack pairs :)
it's impossible not to learn new things about p6 every day. It's kinda scary >.<.
17:43 _mg_ joined
ugexe everytime i learn something new in perl6 though it forces me to forget some other perl6 thing 17:43
llfourn lol
it depends how strangely consistent the new thing is I think
RabidGravy until yesterday I hadn't realised that there were zippy meta-ops 17:45
i.e.
m: say <1 2 3 4> Z+ <5 6 7 8> 17:46
camelia rakudo-moar ad8265: OUTPUT«(6 8 10 12)␤»
RabidGravy which is cool
grondilu not related either : say I have cached function like sub f($n) { state $ //= do-stuff($n) }; does it make sense to evaluate it in void context in a BEGIN block to store precomputed values in the compiled file? Like BEGIN { f($_) for ^10 } ? 17:47
llfourn Z+ was one of the first bits of magic I ever saw
grondilu I mean the optimizer is not going to ignore the void context, is it?
llfourn m: say eager [1], -> @prev { [0, |@prev Z+ |@prev, 0] } ... 10; # pascals triangle
camelia rakudo-moar ad8265: OUTPUT«([1] [1 1] [1 2 1] [1 3 3 1] [1 4 6 4 1] [1 5 10 10 5 1] [1 6 15 20 15 6 1] [1 7 21 35 35 21 7 1] [1 8 28 56 70 56 28 8 1] [1 9 36 84 126 126 84 36 9 1])␤»
grondilu guesses not
llfourn grondilu: yeah sure why not? But that's what constant $x = ... is for 17:48
psch grondilu: note that this specific example will only assign $ for f(0) (unless that returns Any) 17:49
grondilu: i'm assuming that's because it's a minimized example :)
llfourn oh right. nvm me I didn't get you.
grondilu well yeah I badly edited that example. More (state @)[$n] = ... 17:50
well doing precomp was much easier than I thought in the end.
17:52 cognominal left
llfourn grondilu: I was going to suggest using a lazy list but then I remembered RT #127858 :( 17:53
synopsebot6 Link: rt.perl.org/rt3//Public/Bug/Displa...?id=127858
llfourn o/ #night 17:55
grondilu m: my enum abc <a b c>; say $_ ~~ abc for abc::.enums 17:56
camelia rakudo-moar ad8265: OUTPUT«Method 'enums' not found for invocant of class 'Stash'␤ in block <unit> at /tmp/zVMqMbrbxs line 1␤␤»
grondilu m: my enum abc <a b c>; say $_ ~~ abc for abc.enums
camelia rakudo-moar ad8265: OUTPUT«False␤False␤False␤»
grondilu ^how do I get the list of elements of type abc?
psch m: my enum abc <a b c>; say ::($_.key) ~~ abc for abc.enums 17:59
camelia rakudo-moar ad8265: OUTPUT«True␤True␤True␤»
psch i don't know if that does what's asked - i don't really understand the question
i also think it looks a bit horrible :S
grondilu m: my enum abc <a b c>; say ::(.key) for abc.enums 18:00
camelia rakudo-moar ad8265: OUTPUT«a␤c␤b␤»
grondilu that looks fine to me. Thanks.
grondilu is a bit surprised by the different order though. 18:01
psch grondilu: it's a Map, which is an immutable Hash, which are unordered
grondilu yeah
psch ...well, conceptually it's an immutable Hash, not in the isa sense 18:02
grondilu still it's a bit unexpected for an "enumeration"
mst m: my enum abc <a b c>; say ::(.key) for abc.enums.sort 18:03
camelia rakudo-moar ad8265: OUTPUT«a␤b␤c␤»
sortiz m: my enum abc <a b c>; .perl.say for abc::.values; # grondilu, may be what you are looking for? 18:15
camelia rakudo-moar ad8265: OUTPUT«abc::a␤abc::c␤abc::b␤»
18:16 domidumont joined
sortiz m: my enum abc <a b c>; say $_ ~~ abc for abc::.values; 18:17
camelia rakudo-moar ad8265: OUTPUT«True␤True␤True␤»
18:18 pmurias joined 18:19 BenGoldberg joined
grondilu m: my enum abc <a b c>; .say for abc.values 18:20
camelia ( no output )
grondilu m: my enum abc <a b c>; .say for abc::.values
camelia rakudo-moar ad8265: OUTPUT«a␤c␤b␤»
grondilu yep that's good too 18:21
sortiz m: my enum abc <a b c>; .perl.say for abc.^enum_value_list; # Or this?
camelia rakudo-moar ad8265: OUTPUT«abc::a␤abc::b␤abc::c␤»
18:23 a3r0 joined
sortiz grondilu, that last is ordered ^^ 18:23
18:27 dwarring joined
pmurias |2701: a helix like project wouldn't make Perl 6 itself any faster. It would allow using fast Rust code 18:28
|2701 I was making the point that LLVM is desired but Rust is not, due to sjw influence. it was already discussed, mst threw a temper tantrum and stomped off. 18:29
but yes, you are right 18:30
sortiz m: my enum abc <a b c>; say "$_ is { +$_ }" for abc.^enum_value_list; # Shows both numeric value and symbol name. 18:31
camelia rakudo-moar ad8265: OUTPUT«a is 0␤b is 1␤c is 2␤»
BenGoldberg m: my enum abc <a b c>; a.WHAT.say; 18:32
camelia rakudo-moar ad8265: OUTPUT«(abc)␤»
pmurias |2701: he was right in that bringing up sjw stuff has no place in the Perl 6 community 18:37
|2701 I disagree 18:38
but w/e
I actually can't use perl6 anyways since it's on github
grondilu (wth) 18:39
^scrap that, sorry
mst |2701: we already said the SJW stuff was done with. please do not attempt to restart the conversation again. 18:42
|2701 pmurias asked about it, I responded with a summary
mst you could have summarised the on-topic parts and skipped the politics 18:43
in future, please do so
18:43 BenGoldberg left
|2701 the conversation was ABOUT the politics, can you elaborate how I would've done that? 18:44
what would your phrasing have been>
mst "my objections to rust are non-technical and the channel has already decided that means they're off topic, so please let's not go there"
not difficult
pmurias feels guilty for bringing up that subject again :/ 18:45
|2701 no need to feel guilty, mst is the only one making a fuss over a conversation that was finished hours ago
18:45 BenGoldberg joined
|2701 you specifically brought it up again to whine about it and ask me not talk about it after I wasn't anyways 18:45
18:46 ChanServ sets mode: +q *!*@gateway/web/irccloud.com/x-jkxdsvxreqinfjvh
mst the only thing more annoying than irrational crusading feminists is irrational crusading anti-feminists 18:46
18:48 _nadim left
mspo github is political? 18:52
18:52 firstdayonthejob left
RabidGravy who knows? I'm completely confused 18:52
mst I've no idea what his problem with github was - should you care, you're welcome to /msg him to find out
mspo maybe that HR slideshow thing 18:53
could also be extreme GNU
mst I don't care
mspo whatevs ;)
mst I'm a great believer in there being more than one way to do it when it comes to personal politics too, and I'm not interested in anybody who wants to get absolutism all over the technology for whatever reason
grondilu amen 18:56
18:57 firstdayonthejob joined 19:02 MasterDuke joined, Actualeyes left
MasterDuke are there any heuristics for what makes faster nqp code? 19:04
timotimo that's difficult to say without knowing more about what you're looking at 19:05
pmurias MasterDuke: native types: 'my int $foo'
MasterDuke e.g., in Perl 6 (at least right now), getting rid of curlies by making things postfix if possible
grondilu MasterDuke: are you talking about pure nqp or nqp-doped Perl 6? 19:06
MasterDuke pure nqp
timotimo in pure nqp, our optimizer is better about removing curlies, for example 19:07
MasterDuke yeah, i've tried postifixing things and converting to ternaries where possible, no difference
timotimo are you looking at the output of --target=optimize to guide your experiments? 19:08
MasterDuke i'm timing 'perl6 --profile -e' 19:09
19:09 damnlie left
timotimo in that case you're not writing pure nqp code 19:09
you're still writing perl6 code, just using nqp ops directly
MasterDuke i'm trying to make the writing of the profile data to file faster 19:10
timotimo the optimization strategy for that is quite different
oh, that
OK
a c-level profile may also be interesting here. like with "perf record -g"
to see if it's doing something silly
sometimes we have profiles that end up spending 50% or more of the total time in just a single C function implemented in moarvm
19:10 damnlie joined
psch .oO( perl6 --profile -e'exec "$*PERL $program"' ) 19:10
oh wait 19:11
needs another --profile in the Str :P
timotimo that was the case when we were very sub-optimal about files with extremely long lines, for example
MasterDuke turtles all the way down
timotimo the profiler output was also quite a bit slower when it was concatenating the gigantic json string in place, whereas now it prints pieces to the output file and keeps bits in an array 19:12
that made a big difference, if i recall correctly
BTW, if you're okay with writing C instead of NQP, you could try picking up my (very old by now) moarvm branch "finite_callgraph_depth" and making that work 19:13
MasterDuke i tried concatenating a string and printing it occasionally, couple percent slower
timotimo i could never get it to output sensible values :(
pmurias MadcapJake: it seems rust allows making functions callable from C code (so NativeCall should work on them) 19:14
MasterDuke that's unlikely to work out well for either the code or me, i'm not a huge fan of c (nor very good at writing it)
timotimo OK
i've gotta go AFK for a bit
MasterDuke but thanks, it's at least something to investigate 19:15
timotimo it's about summing up data in the callgraph when nodes reach a certain depth
in theory, that calculation could be done in the nqp code that pre-processes the call graph 19:16
19:16 mr-foobar joined
timotimo it takes longer to recompile nqp, though, to get that code into action .. and on top of that you also have to recompile rakudo after nqp 19:17
when you change moarvm, you don't need to recompile nqp or rakudo
19:18 jack_rabbit left 19:24 pecastro left 19:26 pecastro joined 19:33 ssotka joined 19:34 domidumont left 19:38 pecastro left 19:39 pecastro joined
MadcapJake pmurias: that's interesting! I'll have to give it a try. 19:40
19:41 ssotka left
pmurias MadcapJake: the interesting/hard part would be automatic things so it's not necessary to manually write the wrapper (like with c libs) 19:42
19:45 dalek left, dalek joined, ChanServ sets mode: +v dalek, dalek left
timotimo o_O 19:45
19:46 dalek joined, ChanServ sets mode: +v dalek, dalek left
timotimo *seriously* 19:46
19:47 dalek joined, ChanServ sets mode: +v dalek, dalek left 19:48 dalek joined, ChanServ sets mode: +v dalek
timotimo removed a few auto-join channels 19:48
we're potentially using a very old version of this botnix thing, but ... damn, that's *stupid* 19:49
mst insufficient throttling 19:50
and, yes, yes it is
timotimo mst: what's a sensible value for its throttlebps value? currently it was set to 8768, i reduced it to 4096, but the comment says the default value is 256 19:52
mst I'd maybe try setting it to 3600 or so, for 'safely below 4096' and see where you get 19:53
it varies between networks, and of course getting reconnected to a closer server can suddenly make a previous throttle insufficient
19:54 ilbot3 left 19:55 ilbot3 joined 19:56 hotel joined 19:57 aries_liuxueyang left, hotel_california left
timotimo hm. dalek probably doesn't pick up configuration changes just because it gets kicked for excess flood 19:58
19:58 aries_liuxueyang joined
timotimo i'd have to reconnect, re-root, and reboot dalek ... and also add back these other channels to see if it still gets kicked 19:58
20:04 BenGoldberg left
arnsholt timotimo++ # looking into fixing dalek 20:07
timotimo arnsholt: i'm not going to touch the bot's code with a very long stick, though :)
arnsholt Even config changes count! =) 20:08
20:09 BenGoldberg joined
moritz I know that when I moved dalek from feather to hack, I tried a clean setup first 20:10
failed
20:10 buharin left, _mg_ left
moritz and then simply rsync'ed the directory over 20:10
a nasty mix of git and svn repos with lots of local modifications and untracked files
grepped for some paths, adjusted them to the new directory, and it worked
20:15 BenGoldberg left, xfix joined 20:17 itaipu left 20:18 wamba left
timotimo yeah, urgh ;( 20:20
20:23 pecastro left 20:30 cpage_ left 20:35 rindolf left 20:38 cognominal joined 20:39 rindolf joined 20:44 _nadim joined 20:49 Emeric left 20:56 yqt joined 21:20 hotel_california joined, hotel left 21:30 CIAvash left 21:33 TEttinger joined 21:41 gobol joined
RabidGravy better actually find the K8055 board in the morning to actually test that this code works 21:42
21:44 _dolmen_ joined 21:48 itcharlie joined 21:50 ssotka joined 21:54 _dolmen_ left 22:07 hotel joined, hotel_california left 22:16 cognominal left, pmurias left 22:17 skink joined 22:21 RabidGravy left 22:46 hotel_california joined, hotel left
dalek c: 6e28573 | (Jan-Olof Hendig)++ | doc/Language/classtut.pod:
More broken links fixed
 22:47
grondilu m: my $x; say $x.WHERE; say -> $y { $y }($x).WHERE; 22:48
camelia rakudo-moar ad8265: OUTPUT«43465288␤43465288␤»
grondilu m: my $x; ; my $p = -> $y { -> { $y } }($x); my $x = pi; say $p($x)(); 22:49
camelia rakudo-moar ad8265: OUTPUT«Potential difficulties:␤ Redeclaration of symbol $x␤ at /tmp/NRiDHZJGvc:1␤ ------> 3; my $p = -> $y { -> { $y } }($x); my $x7⏏5 = pi; say $p($x)();␤Too many positionals passed; expected 0 arguments but got 1␤ in block <unit> at /t…»
grondilu m: my $x; ; my $p = -> $y { -> { $y } }($x); $x = pi; say $p($x)();
camelia rakudo-moar ad8265: OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in block <unit> at /tmp/Pw89P_KaJI line 1␤␤»
grondilu m: my $x; my $p = -> $y { -> { $y } }($x); $x = pi; say $p(); 22:50
camelia rakudo-moar ad8265: OUTPUT«(Any)␤»
grondilu was expecting pi here somehow
22:51 |2701 left
skink Odd. I'm passing a C function a CStruct and its contents seem to get written all out of order, without a consistent pattern 22:51
Zoffix grondilu, you're closing over it before it has any value 22:52
well, not over *it* by over the $y with value of $x from when $x isn't yet a pie
m: my $x; $x = pi; my $p = -> $y { -> { $y } }($x); say $p();
camelia rakudo-moar ad8265: OUTPUT«3.14159265358979␤»
Zoffix m: my $x; $x = π; my $p = -> $y { -> { $y } }($x); $x = τ; say $p(); 22:53
camelia rakudo-moar ad8265: OUTPUT«3.14159265358979␤»
grondilu yeah but precisely I want to build a closure over an undefined value.
Zoffix Mission accomplished then :) You do get an undefined value back :D
grondilu m: my $x; $x = π; my $p = -> $y { -> { $y } }($x); $x = τ; say $p(); $x = sqrt(2); say $p(); 22:54
camelia rakudo-moar ad8265: OUTPUT«3.14159265358979␤3.14159265358979␤»
Zoffix m: my $x; my $p = -> { $x }; $x = π; say $p();
camelia rakudo-moar ad8265: OUTPUT«3.14159265358979␤»
Zoffix Did you mean this instead?
grondilu in my last example I was expecting pi and then sqrt'(2)
well I was not expecting it with that code but that illustrats my goal. 22:55
Zoffix That expectation is incorrect. You're passing undefined $x as $y parameter to a block that closes over *that $y*, which is undefined, and that block will always return that value. It's no longer linked to $x
My last example is how to close over $x and still keep getting new values when it changes. Your version is a construct made precisely to break that link. 22:56
grondilu hang on
that looks a lot like what I was trying to do in my first example. 22:57
22:57 Sgeo_ joined, caymanboy joined
grondilu m: my $x; my $p = -> $y { -> { $y } }($x); $x = pi; say $p(); 22:57
camelia rakudo-moar ad8265: OUTPUT«(Any)␤»
grondilu because I want the closure to be returned by a function. 22:58
22:58 firstdayonthejob left
Zoffix Interesting. Apparently my fancy keyboard doesn't support 𝑒 :/ 22:58
.u 𝑒
yoleaux U+1D452 MATHEMATICAL ITALIC SMALL E [Ll] (𝑒)
Zoffix Nothing happens when I assign it to a key :/ (and it's not rendering in my font) 22:59
grondilu that is I was expecting -> { $x } to behave exactly like -> $y { -> { $y } }($x)
22:59 caymanboy left
Zoffix grondilu, the first closes over $x, the second closes over $y. The first will return different results when $x changes. The second won't, because $y doesn't change., 23:00
geekosaur is that your keyboard or some other part of the system not handling it?
Zoffix m: my $x; $x = pi; my $p = -> $y is rw { -> { $y } }($x); $x = τ; say $p();
camelia rakudo-moar ad8265: OUTPUT«6.28318530717959␤»
23:00 rindolf left
Zoffix grondilu, now this ^ works because $y is directly tied to $x due to is rw 23:00
geekosaur, well, I assigned it to the key the same way as I assigned π and τ and nothing happens when I press the key :) 23:01
grondilu oh yeah I had thought of using is rw at some point but forgot about it somehow
Zoffix++ thanks
23:01 Sgeo left
geekosaur right, my point is if it also doesn't render when displayed, it may be generated fine by the keyboard and something else in the OS chokes on it 23:02
23:02 |2701 joined
geekosaur (or thinks it's something else, like a multimedia keycode or something silly like that) 23:02
23:03 caymanboy joined
Zoffix grondilu, err... the `is rw` was just an example for explanation -> $y is rw { -> { $y } }($x); is just a convoluted way to write -> { $x } 23:04
grondilu it's not uselessly convoluted. 23:05
Zoffix grondilu, what's the use then?
grondilu my initial intent was to write multi candidate algebraic operators for undefined numeric values.
basically a idea for a polynomials module. 23:06
something like: use Polynomials; my Real $x; my $P = 1 + $x*$x; say $P($_) for rand xx *; 23:07
Zoffix k
𝑒𝑒𝑒𝑒 23:08
geekosaur++ changing from "Typing" to "Copy/pasting" mode fixed it. I guess it doesn't know how to "type" a char it doesn't recognize
grondilu I get an ambiguous call when I try to define multi infix:<*>(Real:U $a, Real:U $b) though. 23:09
geekosaur well, or can type it but the OS won't read it (or something else won't; this comes up a lot with vnc or other remote desktop things, when the remote OS can't figure out what it was just sent_
copy/paste hacks around those os shortcomings 23:10
Zoffix I now have all of these set up as single-key typings: :D πτ𝑒×÷−…∞≅»””’‘¯\_(ツ)_/¯(╯°□°）╯︵ ┻━┻
grondilu, yeah, it'll conflict with the core ones, unfortunately
geekosaur it's even more fun when you type one key and the OS reads it as something else, which also happens a lot with vnc and non-Western keys :/
Zoffix You can either use a different char or shadow the core ones by avoiding the multi
grondilu it works with prefix:<+>(Real:U $) though. Why the difference? 23:11
Zoffix m: &prefix:<+>.^candidates.say 23:12
camelia rakudo-moar ad8265: OUTPUT«Method 'candidates' not found for invocant of class 'Perl6::Metamodel::ClassHOW'␤ in block <unit> at /tmp/LQoLF8h6nF line 1␤␤»
Zoffix m: &prefix:<+>.candidates.say
camelia rakudo-moar ad8265: OUTPUT«(sub prefix:<+> (\a) { #`(Sub|55734360) ... })␤»
Zoffix m: &prefix:<*>.candidates.say
camelia rakudo-moar ad8265: OUTPUT«5===SORRY!5=== Error while compiling /tmp/PPrDBD5Uhe␤Undeclared routine:␤ prefix:<*> used at line 1. Did you mean 'prefix:<~>', 'prefix:<->', 'prefix:<+>', 'prefix:<!>', 'prefix:<|>'?␤␤»
Zoffix m: &infix:<*>.candidates.say
camelia rakudo-moar ad8265: OUTPUT«(sub infix:<*> ($x = 1) { #`(Sub|59403080) ... } sub infix:<*> (\a, \b) { #`(Sub|59405208) ... } sub infix:<*> (Real \a, Real \b) { #`(Sub|59404752) ... } sub infix:<*> (Int:D \a, Int:D \b --> Int) { #`(Sub+{Callable[Int]}|59403536) ... } sub infix:<*> (in…»
Zoffix grondilu, ^ there's no ambiguity with prefix + 23:13
grondilu oh ok
23:13 caymanboy left
Zoffix It would be handy to have a way to explicitly resolve ambiguities in multies.. Like "if something's ambiguous, use this one" 23:13
So you could shadow a specific multi without shadowing all of the multies
grondilu I don't want to shadow anything.
actually I thought I understood but I don't. There is no double undefined candidate for infix:<*> in the core, is there? 23:14
Zoffix sub infix:<*> (Real \a, Real \b) { #`(Sub|59404752) ... } 23:16
grondilu m: multi f(Real $) { rand }; multi f(Real:U $) { pi }; say f(my Real $);'
camelia rakudo-moar ad8265: OUTPUT«5===SORRY!5=== Error while compiling /tmp/QlZZMM1vNP␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/QlZZMM1vNP:1␤------> 3i f(Real:U $) { pi }; say f(my Real $);'7⏏5<EOL>␤ expecting any of:␤ sing…»
grondilu m: multi f(Real $) { rand }; multi f(Real:U $) { pi }; say f(my Real $);
camelia rakudo-moar ad8265: OUTPUT«Ambiguous call to 'f'; these signatures all match:␤:(Real $)␤:(Real:U $)␤ in block <unit> at /tmp/0p3_nkR8Qh line 1␤␤»
grondilu oh I see now
I thought Real:U is narrower than Real 23:17
Zoffix I think resolution by smiley is NIY
grondilu m: multi f(Real $) { rand }; multi f($ where !*.defined) { pi }; say f(my Real $); 23:18
camelia rakudo-moar ad8265: OUTPUT«0.154839716764005␤»
grondilu ^I guess I could do that.
Zoffix neat
grondilu well except it did not work
Zoffix m: multi f(Real $) { rand }; multi f(Real $ where !*.defined) { pi }; say f(my Real $);
camelia rakudo-moar ad8265: OUTPUT«3.14159265358979␤»
Zoffix m: multi f(Real $) { rand }; multi f(Real $ where !*.defined) { pi }; say f(my Real $); say f(my Real $ = 42.0)
camelia rakudo-moar ad8265: OUTPUT«3.14159265358979␤0.782998108723283␤»
grondilu oh yeah forgot the Real 23:19
Zoffix Seems to work fine
grondilu cool
basically I can do: subset Monome of Real where !*.defined
(though it's "Monomial" in proper English) 23:20
23:27 spider-mario left 23:32 cpage_ joined 23:34 gobol left 23:52 sufrostico 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!