»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:16 khw left
tadzik oo :0 00:32
that'll help :)
hoelzro tadzik: =) 00:36
I think you still have to enable the repo in Travis 00:37
tadzik hm 00:39
hoelzro tadzik: I can walk you through it if you need help 00:40
00:43 raiph left 00:49 atroxaper joined 00:53 atroxaper left 01:05 amurf joined 01:09 amurf left 01:10 Ben_Goldberg joined 01:19 rmgk_ joined, rmgk_ is now known as rmgk 01:20 apejens left, apejens joined 01:23 skids joined 01:25 zakharyas joined
skids .botsnack 01:25
yoleaux :D
synbot6 om nom nom
tadzik hoelzro: that'd be awesome, I'll try to get in touch with you tomorrow about it 01:29
now etoosleepy
01:34 zakharyas left 01:48 BenGoldberg joined 01:51 Ben_Goldberg left 02:06 noganex_ joined
hoelzro ok, sounds good! I may be busy (visiting in-laws) but I might get up early and sneak away with my laptop =P 02:12
02:15 jack_rabbit joined 02:28 aborazmeh joined, aborazmeh left, aborazmeh joined 02:29 jack_rabbit left 02:37 jack_rabbit joined, atroxaper joined 02:42 atroxaper left 02:44 vendethiel left 02:45 nys left 02:52 apejens left, apejens joined 02:54 amurf joined 02:58 amurf left 03:07 yqt left 03:13 bin_005 left, bin_005_k joined 03:21 Sqirrel left 03:23 Sqirrel joined 03:30 raiph joined 03:31 konsolebox left 03:32 konsolebox joined, khw joined 03:33 gfldex left 03:38 atroxaper joined 03:42 atroxaper left 03:49 raiph left 03:53 llfourn joined 03:58 llfourn left 03:59 raiph joined 04:08 Sqirrel left 04:10 Sqirrel joined 04:28 raiph left 04:51 Sqirrel left 04:52 Sqirrel joined 04:57 aborazmeh left 05:05 khw left 05:15 llfourn joined 05:24 atroxaper joined 05:28 atroxaper left 05:36 llfourn left 05:38 Ben_Goldberg joined 05:40 BenGoldberg left 05:53 rurban joined 06:15 konsolebox left 06:21 aborazmeh joined, aborazmeh left, aborazmeh joined 06:29 konsolebox joined 06:30 Ben_Goldberg left 06:32 amurf joined, smls joined
smls m: my @a = ^10; dd @a / 2; dd @a div 2 06:33
camelia rakudo-moar 4b5e19: OUTPUT«5.0␤Cannot call infix:<div>(Int, Int, Int, Int, Int, Int, Int, Int, Int, Int, Int); none of these signatures match:␤ (Int:D \a, Int:D \b)␤ (int $a, int $b --> int)␤ in block <unit> at /tmp/stDkOpuAmu:1␤␤»
smls shouldn't div coerce its arguments like the other numeric ops do?
m: my @a = ^10; dd @a % 2; # even its little brother 'mod' does... :) 06:35
camelia rakudo-moar 4b5e19: OUTPUT«0␤»
06:36 amurf left 06:37 AlexDaniel left
smls Hm, looks like it's according to spec: design.perl6.org/S03.html#line_785 06:40
But why?
instead of having the two sigs: (Int, Int --> Int) and (int, int --> int) 06:44
couldn't it be given these: (Int(), Int() --> Int) and (int, int --> int)
...without loosing any functionality?
06:46 vendethiel joined 07:09 vendethiel left
skids I'm guessing it has to do with the return type restriction. It would be good if the reason for that were explained in the design docs. 07:14
smls Maybe it's from a time when coercive parameters didn't work like they do today? 07:16
07:21 aborazmeh left
skids well, so is infix %, and it is different. 07:22
07:29 RabidGravy joined, domidumont joined, domidumont left 07:30 domidumont joined 07:32 araujo left 07:33 araujo joined 07:34 domidumont left 07:35 domidumont joined 07:36 vendethiel joined 07:38 darutoko joined 07:42 domidumont left
dalek line-Perl5: 871858d | (Stefan Seifert)++ | / (2 files):
Only create Perl 6 packages for the first interpreter

Inline::Perl5::require creates a Perl 6 wrapper package for loaded Perl5 modules. use Foo:from<Perl5> uses the first created interpreter: $default_perl5. When using multiple Perl 5 interpreters, only this $default_perl5 should create wrapper packages to avoid confusion and conflicts when the same package would be created multiple times.		 07:44
line-Perl5: 07dd4fb | (Stefan Seifert)++ | / (2 files):
Allow use'ing a P5 module more than once on the same interpreter

When loading a P5 module with .require, .use or use :from<Perl5> we create a P6 wrapper package for seamless integration. Obviously we should do this only on the first load. Perl allows loading modules multiple times and we should stick to that.
Fix GH #27.
07:51 skids left 07:53 [Sno] joined 08:02 rindolf joined 08:11 Alina-malina joined 08:16 isBEKaml joined 08:19 vendethiel left 08:29 vendethiel joined 08:52 isBEKaml left 08:56 telex left 08:58 telex joined 09:04 TEttinger left 09:13 rindolf left 09:20 konsolebox left 09:48 rindolf joined 09:50 bjz left, bjz joined 09:51 spider-mario joined 09:55 diana_olhovik joined 10:00 jaffa4 joined
jaffa4 hi all 10:00
How can I compile a Perl6 code into moarvm bytecode?
nine jaffa4: I don't know, but Panda does it on installation. Maybe have a look at that? 10:02
10:03 RabidGravy left
masak there was a time when I used precompilation a lot, because it made a huge difference for startup. 10:05
nowadays, I prefer to take the startup hit, because it makes it easier to do development if I don't precompile.
10:09 amurf joined 10:13 amurf left 10:28 geekosaur joined 10:32 beastd joined 10:33 rindolf left 10:42 Sqirrel left 11:05 cognominal joined
masak thanks to conversations with vendethiel++ on our clandestine macros channel, I've come to realize that much of advanced macro constructions comes down to being able to synthetically creating variables, and then having those variables resolve in the right/Least Surprise way in the ASTs they participate in. 11:14
users want something a little bit contradictory there: (a) hygiene, and (b) convenience. 11:15
vendethiel (clandestine but logged channel :-P)
masak it's top secret because not many people are all that interested :P 11:16
vendethiel
.oO( and we're not all that interesting )
11:16 Psyche^ joined
masak I dunno, I find our conversations riveting. 11:16
vendethiel at least I learned a word :) 11:17
jaffa4 masak, riveting.. where did you learn that word? 11:29
masak I'm a voracious reader. guess I picked it up through reading. 11:30
11:34 yqt joined 11:40 domidumont joined 11:59 amurf joined 12:00 rindolf joined 12:03 amurf left 12:07 JimmyZ_ joined 12:10 gfldex joined 12:19 mr-foobar left 12:24 mr-foobar joined 12:25 BenGoldberg joined 12:30 BenGoldberg left 12:45 vendethiel left 12:59 yqt left, tinyblak joined 13:09 nys joined 13:16 tinyblak_ joined 13:18 JimmyZ__ joined 13:19 raiph joined, tinyblak left 13:20 JimmyZ__ left, JimmyZ__ joined 13:21 JimmyZ_ left 13:22 xfix joined 13:26 khw joined, vendethiel joined
masak today's autopun spotting: twitter.com/SwedishCanary/status/6...5218984960 13:41
(this is the kind of autopun that parasitizes an existing pop culture quote)
(and yes, I've seen the xkcd on Tautology Club, so linking it here would be way redundant) 13:42
13:44 bin_005_k left
jaffa4 Cannot find method 'AT-KEY': no method cache and no .^find_method 13:51
Is this a bug?
13:53 FROGGS_ joined
jnthn Not automatically. 13:53
yoleaux 18 Jul 2015 22:12Z <japhb> jnthn: Happened to see the link to your S17 supply/whenever proposal. I like it in general -- which is to say, I'd probably need to code a bit with it to find the holes if any.
jnthn Depends how you ran into it. 13:54
It's not a likely issue to run into, though.
But just showing an error is too little info.
jaffa4 caused by line my $f = Perl6::Tracer.new(); 13:55
jnthn And what is Perl6::Tracer?
jaffa4 a class
jnthn You don't say!
jaffa4 module you can install it
jnthn Ah, it's on modules.perl6.org? 13:56
jaffa4 yes
hoelzro good *, Perl 6
jnthn Ah, but one of yours... 13:57
13:57 FROGGS left
jaffa4 It cointains tracer.p6 13:57
jnthn I guess you're playing with compiler internals, then.
jaffa4 but it works if I use -I in command line instead of installing the module
jnthn Well, there's no way I've got time to golf down a whole module and its dependencies. 13:59
masak jaffa4: this is your signal to golf the code, to make it easier for people to find what's wrong. :) 14:01
again, just to recap, "golf" means "remove anything non-essential to exposing the problem"
jaffa4 firsdt problem there is no decent error message 14:02
first...
Do you agree?
masak depends what's wrong, I'd say.
jnthn jaffa4: You're doing stuff with the internals of the compiler. 14:03
masak that it can't find a method is clear. in that sense the error message is decent.
question is why.
14:03 cognominal left
jnthn We spend a lot of time on user experience for users doing normal things. You're not doing a normal thing. You shouldn't expect hand-holding. 14:03
14:04 sjn_phone joined
masak jaffa4: if you genuinely think that Rakudo is doing something wrong in this case, the burden of proof is on you in the form of a golfed bit of code. 14:06
jaffa4 ok.. if I say new... if it failed inside new, I should see that linie 14:07
There is no such line
i nthe error message
so it fails in rakudo when doing new
masak as far as I'm concerned, you're jumping to conclusions. 14:08
the only thing that matters at this point is the golfed code.
jaffa4 or if it fails inside new but the rakudo does not show the line number inside new , that is like indecent error message
masak jaffa4: yes, but we *don't know*.
there is not enough data.
14:09 Colonel-X joined
jaffa4 iit is all based on what I see 14:09
masak fantastic.
mst jaffa4: please stop guessing randomly and show complete code that illustrates the problem, plus a mock up of what you think should happen
masak jaffa4: what mst said. 14:10
mst jaffa4: deleting all of the unrelated code will not only make it much easier for other people to look at but will prove that that code isn't the problem
masak finds it's *really* hard to not take an elitist stance and just go "listen to your elders" when someone is behaving unscientifically wrt bug reports
Colonel-X Can I use in my regexes some rules defined in STD ? 14:12
masak Colonel-X: yes, if you manage to inherit from STD.
jaffa4 gist.github.com/jaffa4/5728e29b867807c3cd41 , see this is the line number
masak Colonel-X: class:method::grammar:rule
mst jaffa4: yes, and now keep deleting code until only the problem remains 14:13
jaffa4: you're looking for a -minimum- reproduction case here
Colonel-X masak, thanks
mst jaffa4: i.e. if I delete any one line from the repro case, it should stop failing 14:14
masak jaffa4: three people asked you to please golf the code, and you paste the un-golfed code. also, the line number is clearly there.
it's in your script file, line 51. 14:15
mst right. I'm not even sure what's supposed to be wrong about that output
looks like Perl6::Tracer is dying during its constructor
which indicates ::Tracer is buggy
but is otherwise uninteresting
masak well, the interesting thing is that we're getting the line-and-file outside of the constructor. 14:16
mst oh, that's not normal?
masak and the error message is about `AT-KEY`, which doesn't seem to figure in the script code at all.
jnthn Well, soemthings up
masak so, yeah, this needs golfing.
by jaffa4
mst right, ok
masak jaffa4: by you.
jaffa4: please golf it.
jnthn But in the module chain it's doing a use Perl6::Grammar:from<nqp>
mst jaffa4: have you started deleting code yet? 14:17
jaffa4: basically, your responsibility is to enact the process in shadow.cat/blog/matt-s-trout/sherlock-holmes/
14:17 domidumont left
mst jaffa4: then #perl6 can go to work on the faulty unit once they know what the fault actually is 14:17
masak jnthn: yes, that's also significant. in fact, I'd mentally promote that to the biggest clue.
14:18 lucasb joined 14:19 JimmyZ__ left
jaffa4 gist.github.com/jaffa4/94c0325bddf927294142 14:21
mst right, so now you need to delete code from Perl6::Tracer 14:24
until you can include the module code in the gist too
lucasb So, the syntax "has $foo" will continue to be valid? Any plans for killing it in the future? 14:26
masak lucasb: no plans; it's a sop. 14:29
lucasb sop? 14:30
masak lucasb: if -- like me -- you dislike it, then don't use it.
lucasb: it's a sop to people who don't like twigils.
14:30 shlomif joined
lucasb masak: yeah, I think I don't like it; thanks for the answer 14:31
14:32 diana_olhovik left
lucasb but one can clearly see that it will cause confusion :) 14:32
jnthn lucasb: No worse than lack of twigils cause in C++/Java/C#... :) 14:33
mst jnthn: worse. because I'll open a file and assume the lack of a twigil means it's a lexical then go hunting for it in the wrong place
jaffa4 mst: I do not get his error message if the class is not precompiled
mst jaffa4: so find the smallest pre-compilable class that gives an error
14:33 JimmyZ_ joined
jaffa4 How can I precompile what is the command? 14:34
14:34 rindolf left
jnthn mst: Point. Though I guess we can look to p5-mop to know whether that's actually a pain point in reality... 14:35
mst: Provided my memory of p5 mop's latest incarnation is accurate, anyways... :)
mst jnthn: I dunno if the latest one involves twigils, the last one that did was one of the "this is obviously going to fail because it's fighting the VM instead of using it" ones 14:36
so, since I'd already told stevan how and why it wasn't going to work I never bothered trying to use it :) 14:37
notable though - all my personal attempts at binding lexicals to object internals have used '$_foo'
because mix and match scoping really annoys me
jnthn: at the very least, I think there should be some sort of 'opt in to namespace collisons' line, "no strict 'scoping';" or something, that people can spot, to realise the sop is in use in the class 14:38
but that might stop it being an effective sop
14:38 shlomif left
jnthn mst: I thought the latest one didn't have twigils for the very reason you noted :) 14:39
mst and I think it's clear that I really like twigils compared to mixed namespaces, so clearly I'm partisan
jnthn I've a strong preference for them too, fwiw. And as a compiler writer I'm always happy to have less code to maintain. But I don't think those two give me a reason to advocate getting rid of the feature... 14:40
mst yes, probably, the latest one is still in pieces on the floor though AFAIK
jaffa4 if I delete moarvm file, it works 14:47
Is it not enough for you?
ugexe is it enough for someone to tell you whats wrong? no 14:49
you can compile code with perl6 --target=mbc --out=lib/whatever.pm6.moarvm lib/whatever.pm6 14:50
mst jaffa4: you need to produce the smallest thing that when precompiled breaks 14:51
14:53 Colonel-X left
hoelzro nine: I've been reporting a lot of Inline::Perl5 issues lately (with more to come!); let me know if you want some help with those 14:53
I'm just trying to get things into GH issues first so they're publicly visible, and others can help if they choose 14:54
I don't expect you to do it all =)
jaffa4 msi: gist.github.com/jaffa4/94c0325bddf927294142 15:03
it happens when use Perl6::Parsing 15:04
and the code is compiled
15:07 cognominal joined 15:09 rindolf joined 15:10 inokenty left
mst jnthn: is the above any better? 15:11
15:14 beastd left
jnthn mst: Well, Perl6::Parsing is also one of jaffa4's modules...it's not entirely to me clear if you even need an intermediate module to trigger the issue at the moment. 15:15
*clear to me
mst ah
jnthn So it's helped show where the problem *isn't* to a decent degree.
mst jaffa4: can you please try and create a *small* .pm that reproduces it, rather than requiring modules you wrote that people don't necessarily know the internals of? 15:16
jnthn But mostly pointed the finger a step down the dependency chain.
vendethiel okay, I'm going to through a feature idea/request, because it's been a great one in Scala and al 15:21
jaffa4 Missing or wrong version of dependency ? What does it mean?
vendethiel being able to "use alias". say `use Foo::Bar::{Baz => BarBaz};` <- really, I just want to be able to "import a name"
jaffa4 D has that feature too 15:22
ugexe it means you are missing a dependency. either your build order is incorrect or you are not including the path via -I
leftover build files in your project directory could do that as well 15:23
15:25 atroxaper joined 15:26 inokenty joined
jaffa4 ugexe: the file is not missing 15:28
15:29 atroxaper left
ugexe thats not what is meant. can the interpreter find it? if it does find it, is it finding one that was correctly compiled (including build order)? 15:32
is it an older version of rakudo?
i think RAKUDO_MODULE_DEBUG=1 may tell you the file being loaded 15:33
15:36 xfix left, amurf joined
jaffa4 figured out all modules had to be recompiled... 15:36
not just one I changed 15:37
araujo read "rumors" around here that we might get a beta for Sept ... for what?, rakudo?
vendethiel "could not find A::B in ...." <- where did I "use A::B", please , rakudo?
jaffa4 jnthn: here it is gist.github.com/jaffa4/94c0325bddf927294142 15:38
ugexe rakudo now gives errors you will make in the future
jnthn jaffa4: Hmmm 15:40
15:40 RabidGravy joined
jnthn jaffa4: I suspect it's about pulling in NQP code 15:40
15:41 Ven joined
jaffa4 in compiled code 15:41
Ven m: class Actions{ method outer-expr:expr($/) { } }; # <- that's a valid rule name, but not a valid method one?
camelia rakudo-moar 4b5e19: OUTPUT«5===SORRY!5=== Error while compiling /tmp/MMtun8cVXn␤Colon pair value '$/' too complex to use in name␤at /tmp/MMtun8cVXn:1␤------> 3Actions{ method outer-expr:expr($/) { } 7⏏5}; # <- that's a valid rule name, but no␤»
15:41 amurf left
jnthn jaffa4: If you rename your modules to Perl6X::Parsing and Perl6X::Tracer, do things get better? 15:41
Ven std: class Actions{ method outer-expr:expr($/) { } }; # just trying...
camelia std 28329a7: OUTPUT«ok 00:00 139m␤»
Ven Right, seems to be a bug in rakudo then..?
that's probably already been reported, though. 15:42
jnthn Ven: Not a bug, you need whitespace
Ven jnthn: how does std do it?
jnthn Ven: It's taking your ($/) as the value of the colonpair
Ven: Because it only parses
Ven: Or maybe it cheats and stringifies and actually creates a method literally called 'outer-expr:expr($/)' :) 15:43
Ven m: class Actions{ method outer-expr:expr ($/) { } }; # "need whitespace"?
camelia ( no output )
Ven uh, that doesn't work on my local rakudo
jnthn std: class Actions{ method outer-expr:expr($a) { $a } }
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Variable $a is not predeclared at /tmp/2peN3NspoM line 1:␤------> 3s Actions{ method outer-expr:expr($a) { 7⏏5$a } }␤Check failed␤FAILED 00:00 139m␤»
jnthn That's how you can see STD is parsing it the same way as Rakudo 15:44
15:44 skids joined
Ven indeed. thanks 15:45
jaffa4 jnthn: no
jnthn jaffa4: OK, so you can remove everything from Perl6::Parsing except the two two lines and still have the bug? 15:47
skids jnthn: I left some comments on your S17 gist, not that I've used the async API much yet. 15:51
jaffa4 jnthn: correction.. the problem disappears then
vendethiel grrr. perl6 was showing the next one as an error
Ven I mean, rakudo.
15:52 konsolebox joined
jaffa4 jnthn: ao renaming to Perl6X helped 15:52
15:52 konsolebox left
Ven std: class Actions{ method outer-expr:bad($/) {}; method outer-expr:expr ($/) { } }; 15:52
camelia std 28329a7: OUTPUT«ok 00:00 141m␤»
Ven m: class Actions{ method outer-expr:bad($/) {}; method outer-expr:expr ($/) { } };
camelia rakudo-moar 4b5e19: OUTPUT«5===SORRY!5=== Error while compiling /tmp/cHZZgXHpOr␤Colon pair value '$/' too complex to use in name␤at /tmp/cHZZgXHpOr:1␤------> 3ss Actions{ method outer-expr:bad($/) {}7⏏5; method outer-expr:expr ($/) { } };␤»
Ven ^ if you try that in a multiline file, it'll error out on ":expr ($/)", making it hard to debug
jnthn jaffa4: aha, OK. Then it's probably an issue involving namespaces shared between NQP code and Perl 6 code. 15:53
15:53 BenGoldberg joined
jaffa4 jnthn: how should I work? Is this a bug? 15:54
masak araujo: no, a birthday beta for *Perl 6*. 15:55
araujo: and a 6.0 for Christmas.
jnthn jaffa4: I suggest you don't use the namespace "Perl6::" in your modules.
jaffa4: It may be fixable, but it seems that "use a module from NQP" is a necesary ingredient to see this 15:56
timotimo could very well be that having an nqp module with Perl6:: installs a "Perl6" object in the global namespace that doesn't understand methods 15:57
and later adding another class to Perl6:: and then trying to get it back out could go asplode
jnthn I suspect it's that and also that the linking logic doesn't know how to merge NQP namespaces with Perl 6 ones
timotimo mhm
it's not exactly a common thing to want to have 15:58
jnthn *nod*
Generally, the only reason to do these things is because we didn't setttle on any standard API for them yet 15:59
jaffa4 jnthn: being able to import under another name could help 16:01
16:02 zakharyas joined 16:08 tinyblak_ left 16:14 beastd joined
araujo masak, nice :) 16:15
ugexe at first thought, importing under another name could be a cool thing
16:16 spider-mario left
ugexe should Proc::Async give a better error on jvm other than 'Cannot find symbol &Async'? 16:17
16:17 araujo left
timotimo it might be nice to get "cannot find &async inside Proc" there 16:18
jnthn We could install a stub there that ways it's NYI
16:19 Ven left
jaffa4 jnthn: so are you planning to do something about it? 16:22
16:23 sjn_phone_ joined
jnthn jaffa4: Not right now. I suggest picking a different namespace; I'm not sure Perl6:: is a wise choice anyway. 16:23
jaffa4 jnthn: what is the wise choice?
16:23 sjn_phone_ left
jnthn jaffa4: Given what you're writing is heavily tied to Rakudo's internals, you may want to pick a name along those lines 16:24
masak .oO( "you've got to learn how to find things out for yourself!" --- "...how do I do that?" -- ":/" )
:P
ugexe that would certainly help. 'signal' is another one where i've needed to get around missing symbols between VMs 16:25
jnthn jaffa4: I called the Rakudo debugger frontend Rakudo::Debugger::UI::CommandLine, for example
ugexe i.e. sub signal-jvm($) { Supply.new }; my $signal-handler = &::("signal") ~~ Failure ?? &::("signal-jvm") !! &::("signal"); my $sig-resize = ::("Signal::SIGWINCH"); $signal-handler.($sig-resize).act:
16:25 Alina-malina left
ugexe if i want to use signal code if its running under moar, or not use signal code on jvm 16:25
16:25 Alina-malina joined 16:26 zakharyas left, sjn_phone left
jaffa4 jnthn: will Rakudo::Perl6::Parse wprk? 16:28
16:30 Alina-malina left
jnthn jaffa4: Should do 16:31
JimmyZ_ Perl6X:: 😊 16:33
ugexe or dont separate the namespaces at all: Perl6Parser 16:37
16:39 Alina-malina joined
jnthn away for a bit 16:41
geekosaur first wonders if the emoticon is part of the namespace, then wonders if exception namespace should be 😭 16:43
16:44 jack_rabbit left
geekosaur (there is actually some seriousness behind that last part: I wonder about the wisdom of commandeering X) 16:45
BenGoldberg Maybe the exception namespace should be 💩 16:46
JimmyZ_ Perl6 eXtend 16:47
16:47 xinming_ joined 16:48 espadrine_ joined 16:50 xinming left
skids geekosaur: You'd rather "class X" was available to run afoul of infix X? 17:01
.oO(ALL YOUR ALLCAPS ARE BELONG TO 6)
 17:05
17:08 BenGoldberg left 17:10 BenGoldberg joined 17:13 atroxaper joined 17:17 atroxaper left
timotimo skids: but a class only appears in term position, not infix position :) 17:23
skids timotimo: In well formed code. But if your code is well formed, you won't be confused by it :-) 17:24
timotimo fair enough
17:36 domidumont joined 17:41 jaffa4 left 17:51 domidumont left 18:01 AlexDaniel joined
lizmat good *, #perl6! 18:27
lizmat notices CPU / wallclock time of spectest went up by ~10% sometime in the past week 18:28
masak \o, lizmat
lizmat masak o/ 18:29
before: Files=1024, Tests=45592, 189 wallclock secs (11.67 usr 3.66 sys + 1077.30 cusr 82.90 csys = 1175.53 CPU) 18:30
now: Files=1025, Tests=45628, 211 wallclock secs (12.08 usr 3.91 sys + 1226.55 cusr 101.71 csys = 1344.25 CPU)
18:30 domidumont joined
masak hopes someone has the tuits to bisect that slowdown 18:31
lizmat there was one NQP bump in that period, I guess that would need to be checked first 18:32
masak sounds reasonable. 18:33
lizmat hmmm... there were actually 2 bumps ... one for LTM fixes 18:34
masak that sounds like a very likely suspect.
lizmat hmmm... raw startup seems to have bumped similarly 18:35
.12 now on my machine, it used to be < .1 not too long ago :-(
can't find any obvious reasons that would explain the increase in startup time 18:45
moritz \o 18:46
moritz back from a weekend in Bruxelles
timotimo lizmat: potentially the "add &?ROUTINE to every routine" change made a big impact 18:47
lizmat ah... hmm....
18:47 spider-mario joined
timotimo but 10% seems excessive for that 18:48
lizmat is trying spectest without 3db9326bda8d843eb2e7ad 18:55
masak moritz: welcome back!
moritz: how was Bruxelles?
or, um, "comment etait Bruxelles?", I guess.
18:55 yqt joined
lizmat "hoe was het in Brussel?" 19:00
19:00 yqt left
masak haha, it's like Swedish but spelled funny :P 19:00
"hur var det i Bryssel?"
lizmat yeah... just like Dutch, but spelled funny :-) 19:01
19:01 domidumont left, atroxaper joined
masak the grass is always spelled funny on the other side, I guess. 19:01
lizmat yeah...
one of my funniest moments in Sweden was seeing Holy Grail in a cinema 19:02
with Swedish subtitles in the Møøse section
19:02 raiph left, telex left
masak I'm pretty sure I've seen such subtitles on TV at some point. 19:02
similarly, I own a copy of "Harrius Potter et Philosophi Lapis" 19:03
sometimes I go looking for the spells they use, to see how they were translated from fake Latin into Latin :P
19:04 telex joined 19:05 atroxaper left 19:08 yqt joined 19:09 espadrine_ left
japhb masak: Does Hermione's correction of the pronunciation of Wingardium Leviosa still make sense? 19:09
19:09 tinyblak joined 19:10 dustinm` joined
masak japhb: do you know offhand which chapter it's in? 19:12
moritz masak: nice; not as hot as here :-) 19:13
masak japhb: (I remember it being in the movie, but don't remember if it's in the book, or if it's actual dialogue in the book)
19:13 amurf joined 19:14 tinyblak left
lizmat afk for a bit& 19:15
19:18 amurf left
jnthn lizmat: The main slowdown was the LTM thing, I'm afraid 19:19
lizmat: I don't have much idea what to do about that; we had the wrong semantics before.
lizmat: So we don't really want to undo it.
lizmat: I was hoping Parsing Wizard Larry may have an idea :)
masak jnthn: that's fine for the longer spectest time. but startup was slowed down 1000% :/ 19:20
jnthn masak: um...what?
That doesn't match what I see at all
masak um, sorry. misread what lizmat wrote.
jnthn masak: .1 vs .12 is not 1000% :P
masak from < .1 to .12, which at least seems significant
jnthn Stop with the category theory, and go redo arithmetic :P 19:21
masak yeah, I fail at reading, apparently
jnthn Yeah, I wonder if it's because our NFAs got bigger
BenGoldberg m: ((.12 - .1) / .1).say;
camelia rakudo-moar 4b5e19: OUTPUT«0.2␤»
jnthn And yeah, it's significant.
BenGoldberg 20%
jnthn But again, I dunno what to do about it. It's not really a win to be fast and wrong... 19:22
19:22 raiph joined
jnthn So will have to spend some time digging deeper I guess 19:22
jdv79 im a bit confused on the 6 fund and jnthns req. the req was for 10k usd and didnt liz put up 10k usd? so why the call for more? 19:24
masak jdv79: because the grant is ending? 19:25
jdv79 and how much more?
timotimo 10x as much as before, please 19:26
masak yes, someone please fund jnthn forever :)
jdv79 for real
masak jdv79: haven't seen a sum.
jdv79 perhaps its time to figure that 19:28
jnthn jdv79: The 10k was enough to fund 250 hours of work. I've currently done 205 hours under the grant. 19:29
jdv79 is there a new grant coming? 19:30
jnthn jdv79: 6guts.wordpress.com/ has some more details
jdv79: The goal is to do another grant after this one in the same style, yes. 19:31
jdv79 nice. i look forward to it. 19:32
masak we all do \o/ 19:33
jnthn I basically arranged to drop to ~50% at my $dayjob for the rest of 2015, and have the rest of my time available for Perl 6 work. :)
Also, well past the 100 RT tickets resolved mark on the current grant :) 19:35
jdv79 is pmichaud fed by the 6 fund for the glr? 19:36
jnthn No 19:37
jdv79 weird; ok. that's probably the most important thing to me at this point. 19:39
how generous
19:55 xtreak joined 19:56 smls_ joined, smls left
jnthn I may jump in and help with GLR things once I'm done with multi-dim array stuffs. 20:01
ooh yay, tomorrow will be cooler weather. 20:02
xtreak Any intro material abt graphemes and their advantages? 20:07
masak xtreak: well, unless you've seen www.joelonsoftware.com/articles/Unicode.html -- you should give it a read. 20:09
it somehow doesn't mention the word "grapheme", though. but it's an essential read. 20:10
jnthn When I was implementing the NFG stuff I did some writing on it also
masak xtreak: this seems relevant, and canonical: www.unicode.org/reports/tr29/#Graph...Boundaries
masak is reminded from that page that devanagari is insane 20:12
xtreak Thanks masak i read joel article but need a refresh on it.. I will look into it..
arnsholt xtreak: Basicall, a grapheme is what you would call "a single character"; that is, a base thing plus anything that works like an accent or diacritic
masak arnsholt: a codepoint also has claim on the description "a single character", IMHO. 20:13
arnsholt Yeah, thus the quotes
masak arnsholt: the difference is that the grapheme is what the reader perceives to be a single character on screen/paper.
arnsholt A codepoint is what Unicode calls a character, but grapheme is what a human (who isn't a computer scientist) would call a character 20:14
Yeah, what you said
jnthn A codepoint is a single codepoint :P
masak I said no tautology club!
jnthn
.oO( if you like it shoulda put a combining ring on it :P )
20:16 rindolf left
masak doesn't quite have the same ring to it 20:19
masak .oO( if you like it shoulda made it have the same ring to it )
nwc10 is this the hook from which we hang a series of puns? 20:20
there's beena grave shortage recently
the situation is getting accute
jnthn I'm a dire critic of these puns and they get low marks from me. 20:21
masak your puns are all very sicilicus 20:22
nwc10 we're doing the best we can, given the circumflex we find ourselves in.
masak I wish we could take a shortcut and circumflex them all
dang, nwc10 was slightly more acute than I was 20:23
should be more breve next time
dalek line-Perl5: b4318ce | hoelzro++ | lib/Inline/Perl5.pm6:
Remove .so copying workaround

Rakudo no longer has this problem, so filling up /tmp with .so files is no longer needed
line-Perl5: 2cab928 | niner++ | lib/Inline/Perl5.pm6:
Merge pull request #33 from hoelzro/master

Remove .so copying workaround Gladly! Thanks for the help :)
20:27 xtreak left 20:30 zakharyas joined 20:34 bin_005 joined 20:35 darutoko left 20:44 bin_005 left, bin_005 joined 20:45 cognominal left 20:46 MilkmanDan joined 20:47 cognominal joined 20:49 atroxaper joined 20:51 zakharyas left 20:52 davido___ joined 20:54 atroxaper left 20:58 JimmyZ_ left 21:03 Gardner joined, amurf joined 21:04 inokenty left, Gardner is now known as mjgardner
ggoebel2 jnthn: any status update on your funding? 21:04
21:05 bin_005 left, bin_005 joined
ggoebel2 I would be willing to shoot off some emails to leaders of perl mongers groups asking them to pass a hat, but I'm not sure that would be considered kosher. 21:05
21:08 amurf left
vendethiel ggoebel2: it's been accepted 21:10
ggoebel2 that's good news! where was it reported? 21:11
vendethiel ggoebel2: news.perlfoundation.org/2015/05/per...succe.html 21:12
in may already :-).
jnthn vendethiel: I think ggoebel2 is talking about further funding to keep it going 21:13
vendethiel oh! my bad.
ggoebel2 right... "Funding is being requested for an initial 3 month period, with a view to extending the grant provided the work done under it is successful and funds are available."
21:15 cognominal left, cognominal joined
jnthn ggoebel2: Not a status update as in "I can publicly say who and what", but something good is in the pipeline. :) 21:15
21:17 FROGGS_ left
ggoebel2 good. thank you for everything you do. And everyone else actively, passively, and otherwise working to support perl6 21:17
vendethiel nods in agreement, @perl6-people>>++ :) 21:18
jnthn :) 21:19
jnthn is gonna go rest, hopefully well, so he can do useful stuff tomorrow :) 21:20
'night o/
vendethiel g'night!
21:21 lucasb left 21:22 inokenty joined
masak what ggoebel2 said. 21:22
AlexDaniel yes 21:28
timotimo: 2014.28: Title goes here. Fill out under all circumstances! 21:29
you know, something is wrong with that title!
masak think it's an in-joke. 21:32
AlexDaniel masak: hmm? 21:33
skids Now how to make it an autopun...
AlexDaniel masak: are we talking about the year number?
masak oh! no, I missed that. 21:34
that's certainly wrong. s/2014/2015/ :)
AlexDaniel I tend to do that a lot, especially when someone asks my age 21:35
I just don't know whether it incremented already or not
and how many times
21:37 lizmat left
masak ages are very annoying that way. 21:37
skids When you get older it takes longer to answer what your age is. Because you stopped caring. 21:38
But it does make you look senile. 21:39
I guess as you approach official retirement age your response time might speed up a bit. 21:40
japhb
.oO( 3 years, 4 months, 12 days, 5 hours, and 16 minutes until retirement -- but who's counting? )
 21:42
BenGoldberg If you're really lazy, you can say what your birthday and year are, and let the person you're talking to do the math. 21:43
s/year/birth year/
masak 'night, #perl6
21:47 beastd left 21:48 dj_goku_ joined 21:51 xinming_ left, rurban left 21:52 b2gills joined, xinming joined 21:53 mtj_- joined 22:05 spider-mario left 22:11 vendethiel left 22:18 mjgardner left 22:19 mjgardner joined 22:20 cdc joined, cdc is now known as Guest59246 22:27 BenGoldberg left 22:29 bjz left
timotimo oh damn, did i write the previous year again? 22:43
22:49 TEttinger joined 22:51 LLamaRider joined, LLamaRider left, colomon joined 22:54 colomon left, lizmat joined, colomon joined 22:58 tinyblak joined 23:02 colomon left 23:04 mjgardner left 23:07 nebuchadnezzar joined 23:09 colomon joined 23:10 amurf joined 23:11 smls_ left 23:21 bin_005 left 23:35 colomon left 23:42 dayangkun joined 23:51 nys left, nys joined 23:54 RabidGravy left, cognominal left

Powered by Cro and the Raku® Programming Language.

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

Thank you!