»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.
pink_mist Ben_Goldberg: did you watch youtu.be/JpqnNCx7wVY yet? I'd recommend to watch it if you're looking into these kinds of things 00:00
Zoffix Is there an easy way to upgrade? I've got rakudobrew installed and an who-knows-how-old version of P6... How do I get the latest and greatest? rakudo HEAD if possible? 00:01
00:01 ilbelkyr_ joined
timotimo if it's very old, you may want to rakudobrew self-update (or is it self-upgrade?) 00:01
and then just rakudobrew build moar should give you HEAD
Zoffix Thanks.
(it's rakudobrew self-upgrade) 00:02
timotimo mhm
00:04 TEttinger joined 00:06 laouji joined
Zoffix pink_mist++ That's a good vid I'll definitely watch. What nick is Jonathan Worthington BTW? (to put the face to the name) 00:09
pink_mist jnthn 00:10
Zoffix Thanks :)
00:10 laouji left 00:11 [Tux] left, khw left
dalek oblem_solver_tutorial: 0bb860d | (Herbert Breunung)++ | chapter/text0.md:
reedit chapter 0 paragraph 2
 00:19
oblem_solver_tutorial: be15397 | (Herbert Breunung)++ | .gitignore:
ignoring backup files
oblem_solver_tutorial: 8e2fb0f | (Herbert Breunung)++ | / (7 files):
linking all chapter
00:19 ilbelkyr left, ilbelkyr_ is now known as ilbelkyr 00:20 [Tux] joined
Ben_Goldberg Unless I'm mistaken, 'whenever' is un-cancellable. The closest one could do, would be to store the Supply returned by IO::Async::Socket.listen into a variable, and make a second Supply; .tap the first supply and have it emit into the second supply; then use 'whenever' on the second supply, and later call .done on the second supply. 00:23
unless of course whenever returns some sort of cancellation object which I'm unaware of. 00:24
pink_mist Ben_Goldberg: doc.perl6.org/type/IO::Socket::Async <-- both of the two large code blocks in the beginning of this seem to show how to close/stop a whenever 00:34
00:34 aborazmeh joined, aborazmeh left, aborazmeh joined
pink_mist the first with $conn.close, the second with done; 00:34
maybe the $conn.close doesn't close it though 00:36
but the done one certainly seems to
00:43 laouji joined, laouji left 00:44 tony___ left
timotimo whenever returns a supply, IIRC 00:49
um 00:50
i just went to doc.perl6.org and the search bar was gone
then i ctr-r'd and it appeared
i wonder if it works fine for first-time visitors
oh, wait 00:51
the supply block returns the supply; whenever can only be inside a supply (or a react) block
anyway. the whenever is cancellable from the inside by last-ing 00:52
pink_mist oh. so I was /almost/ right when I said LAST earlier :P 00:53
... is react/whenever properly documented somewhere? I can't seem to find good docs for them
timotimo yeah, except LAST is the phaser that fires when you last out of it or the last iteration happened
pink_mist just the incidental mention of them in IO::Socket::Async
yeah I know
00:54 thou left
timotimo hm, whatever happened to the S17 draft thing jnthn had 00:54
gist.github.com/jnthn/a56fd4a22e7c43080078
no mention of "react" yet 00:55
but react is basically a shorthand for "await supply { ... }" or something similar
00:57 llfourn joined
Zoffix Mind. Blown. 01:04
01:04 TingPing left
Zoffix jnthn++ very nice talk. Like your accent too. 01:04
Wow. Perl 6 really blows my mind. It's pretty revolutionary. 01:05
01:06 Kcop left
timotimo peh. everything perl 6 can do, lisp could already do ~40 years ago 01:07
Zoffix Dunno... the whole await stuff (youtu.be/JpqnNCx7wVY) really makes questions like this one utterly ridiculous: www.quora.com/Perl-programming-lan...body-care/ 01:09
Ben_Goldberg pink_mist, In the first of those code blocks in IO::Socket::Async, the $conn.close does not stop the listening socket, but rather closes one accept()ed connection. 01:15
pink_mist Ben_Goldberg: yes, I realised. I already said I was wrong. 01:16
ShimmerFairy Am I wrong, or is Perl 6 one of the only languages to provide built-in async support that goes beyond what other languages provide? (That is, beyond just primitive stuff, and/or having to resort to various libraries?) 01:25
01:27 yqt left, nuc_ joined, nuc left
awwaiid mmm. "libraries" can get pretty fuzzy. OTP languages (Erlang, Elixir) have a whole lot of async for example 01:28
pink_mist I think Erlang might still be better at async/concurrency/paralellism than p6, but not by much
ShimmerFairy Yeah, I really hope I'm wrong about that, since it would seem like other languages have had some time to focus on just async :) 01:29
awwaiid I'm not sure how important it is to be in the language vs in a library, to a large extent. Though I guess if you do like clojure where most memory access is some form of software-transactional-memory then that helps... but more for consistency 01:31
Zoffix I find panda' usage example is a bit too verbose with the whole $PATH_TO_PANDA business: fpaste.scsys.co.uk/paste 01:33
Looks messier in the terminal than the pastebin: i.imgur.com/DVd4BjX.png 01:34
ugexe because you are using a rakudobrew installed panda 01:35
Zoffix Ah
ShimmerFairy Zoffix: your fpaste link doesn't seem to link to the thing you meant to
Zoffix Sorry, it should've been this: fpaste.scsys.co.uk/500065 01:36
ShimmerFairy Zoffix: I agree though, it should work off $*EXECUTABLE_NAME (or whatever it's called) instead. Call 'panda' what the user just did, imo :) 01:37
timotimo especially the ".." in between ... *shudders*
Zoffix But it looks neat in the pastebin ('cause my browser is wide), but in the terminal it's a mess
ugexe add that full path/to/site/bin/panda to your path before rakudobrew 01:38
otherwise it literally is showing you what the user just did... run a shim that runs *that* command 01:39
timotimo oh, is that so?
01:40 cibs_ joined 01:44 laouji joined, cibs left 01:45 ilbot3 left 01:46 ilbot3 joined 01:49 laouji_ joined, laouji left 02:05 sufrostico left 02:08 tokuhirom joined 02:09 eliasr left 02:13 kaare_ joined, chenryn joined 02:14 FireFly left, cibs joined 02:15 FireFly joined 02:18 FireFly left, kid51 left
M-tadzik It seems to.me that using *WHATEVER_NAME here is more trouble than it's worth compared to just hardcoding “panda” 02:19
02:19 cibs_ left
pink_mist I can't say I know how shims work, but maybe the shim could set an env var with how it was called and that could be used instead? 02:20
02:23 vendethiel joined
ShimmerFairy M-tadzik: I can see the appeal in giving the correct path (esp. if you were typing out concrete, copy-paste style examples in usage, which isn't happening here incidentally), but I also think most people are clever enough to know to substitute 'panda' for the Right Thing :) 02:24
02:25 noganex joined
MilkmanDan Googling for "perl6" and "list" seems to have a lot more of the mailing- variety than it does data types. :) 02:26
Can someone familiar with the internals tell me if lists are a full fledged type now? 02:27
ShimmerFairy m: say (1,2, "foo", True).WHAT
camelia rakudo-moar 94d457: OUTPUT«(List)␤»
[Coke] thinks he fixed a doc search bug.
ShimmerFairy MilkmanDan: we do in fact have a List type, if that's what you're asking :)
[Coke] MilkmanDan: doc.perl6.org/type/List
MilkmanDan ShimmerFairy: Is it the same kind as perl5?
ShimmerFairy MilkmanDan: safe bet is no, though I don't know P5 so I can't answer that question helpfully. 02:28
02:28 noganex_ left
[Coke] MilkmanDan: if you are coming from perl 5, you might want to read the 5-6 differences on the doc site. 02:28
doc.perl6.org/language.html - first few options. 02:29
ARGLEBARGLE. I just got moar to crash on malloc building the docs.
MilkmanDan [Coke]: My knowledge of perl5 is minimal. 02:30
[Coke] ... then don't worry if Perl 6 lists are like perl 5 lists. ;)
MilkmanDan [Coke]: I'm afraid I very much will worry about that. :) www.youtube.com/watch?v=zDJxgLe7JQo 02:31
M-tadzik ShimmerFairy (IRC): yeah, good point
02:31 cibs_ joined
M-tadzik Pull request welcome, I think :) 02:31
[Coke] MilkmanDan: I'm not going to watch a 30m video; can you tl;dr it for me?
pink_mist MilkmanDan: that's just him not understanding perl5 02:32
[Coke] if it's "are the internals the same", the answer is a definite no.
pink_mist MilkmanDan: and no, perl6 lists don't work like that unless you tell them to
MilkmanDan [Coke]: Using perl5 lists considered hazardous.
[Coke] ok. this is perl6. You have a completely different set of things to worry about.
ShimmerFairy M-tadzik: ok, I've also got another fix in mind too :) ("panda help" being interpreted as trying to install package "help")
MilkmanDan pink_mist: Eh? What do you mean, not understanding? He successfully exploited list behavior.
[Coke] there are no internals in common. 02:33
pink_mist MilkmanDan: yes, but all his complaints show that he don't understand how lists work in perl5. neither did the authors of the code he managed to break. that's why he managed to break it.
M-tadzik ShimmerFairy (IRC): oh yeah, that's a silly thing :) 02:34
MilkmanDan pink_mist: I think it's a bit of a stretch to Blame The Programmer when we're talking about DBI, don't you?
pink_mist MilkmanDan: no, he was using utterly stupid interfaces to DBI.
[Coke] Anyway, please let us know if you have any Perl 6 questions.
MilkmanDan I see. 02:35
ShimmerFairy M-tadzik: my fix for that would be to add the 'help' command, and to throw away 'panda [pkgname]' as a shortcut for the 'install' verb; if you have unadorned verbs for a tool like that, it's a bad idea to have that shortcut
02:35 Ben_Goldberg left 02:36 cibs left
ShimmerFairy (if the verbs were instead --install, --info, etc., then naturally there wouldn't be an issue with that shortcut ☺) 02:36
pink_mist usually does `command -?` if he doesn't know how to use a command 02:37
02:38 thou joined
ShimmerFairy My favorite is when 'command -h' responds with "unrecognized option" and tells you to use --help for the help :P 02:38
02:39 FireFly joined
ShimmerFairy M-tadzik: since panda uses the default usage, I'll fix it by adding a sub USAGE for it to use. 02:40
M-tadzik Oh, right, it's inherently USAGE's problem 02:41
ShimmerFairy M-tadzik: Not a worry, I figure panda is complex/important enough to deserve a more nicely written USAGE anyway :) 02:42
02:43 thou left, leedo joined 02:44 leedo left 02:59 Cryptoknight joined
Cryptoknight Hello p6ers 02:59
trying to get a linux install going; stuck on a rakudobrew command - 'rakudobrew build-panda' - the rakudobrew I'm using expects a version number, and I have no idea what to use. 03:01
03:04 Cryptoknight left 03:05 Cryptoknight joined
dalek c: b306073 | coke++ | htmlify.p6:
Don't stop processing too early.
 03:05
03:07 z8 left 03:14 cibs joined
MilkmanDan pink_mist: Well, being a bear of little brain myself I'm hoping to avoid making those same mistakes, as I expect my code to have to touch the scary Internet. 03:18
03:18 cibs_ left
MilkmanDan And I don't kid myself that I'm going to be smarter than the authors of DBI or Bugzilla or whatever. 03:18
03:22 Cryptoknight left
pink_mist MilkmanDan: it's not about being smarter, it's about using good APIs ... for example, DBI's ->quote is absolutely horrible and should be avoided at all costs. that the authors of some projects didn't do that is their own fault. if I see a codebase that use DBI's ->quote except in very very controlled conditions, I'd never use that software. and about CGI.pm, that's been discouraged for literally more than a decade. better alternatives have been available for 03:22
years and years
03:23 Cryptoknight joined
MilkmanDan pink_mist: But the list behavior is still present, yes? 03:23
So I'd need to avoid stepping on that particular landmine.
pink_mist MilkmanDan: if you understood how perl5 does lists, it wouldn't be surprising at all, and actually very useful if used properly
MilkmanDan: but this really isn't a discussion to have in #perl6 03:24
MilkmanDan Well, I'd like to know if this particular landmine is still present in perl6. 03:25
Since I think we're still going to be waiting a bit for the new llama and camel books I don't expect to have a guide telling me where not to step for a little while. 03:26
M-tadzik Cryptoknight (IRC): what do you mean by “expects a version number?” 03:27
03:28 vendethiel left
ShimmerFairy MilkmanDan: without having any P5 knowledge, it's safe to say that Perl 5 security/etc. concerns aren't relevant to Perl 6. Any issues P6 has are independent of anything in Perl 5, and vice versa. 03:29
pink_mist 04:32 <pink_mist> MilkmanDan: and no, perl6 lists don't work like that unless you tell them to 03:30
MilkmanDan: I already answered that question of yours a long time ago'
-'
MilkmanDan pink_mist: Ok. Thanks.
pink_mist (for those who may be wondering what attribute of perl5 lists we're talking about, it's the auto-flattening behaviour of them) 03:32
(combined with the fact that parameters to functions are by default lists) 03:33
ShimmerFairy pink_mist: that second one shouldn't be a concern with P6's "real" signatures, right? Or is my lack of P5 knowledge really showing? :P 03:35
pink_mist ShimmerFairy: that's right 03:36
Cryptoknight MilkmanDan: when I run './rakudobrew build-panda' it responds 'No version set' - the documentation does mention a version argument, but not much else. 03:37
sorry, I mean to reply to M-tadzik ! 03:38
M-tadzik Cryptoknight (IRC): you may need a “rakudobrew switch”, or “rakudobrew global”, to tell it which installation of rakudo you mean
Cryptoknight even if I just have one ? 03:39
M-tadzik It should perhaps be made so that the first rakudo that gets installed becomes tge default
Cryptoknight that could be a nice touch!
M-tadzik Yeah, to my defense though you only need to do this once :) 03:40
Feel free to open a bug, that'll minimize the chance of me forgetting about it:)
Cryptoknight theoretically
I'm not sure if it is a bug, or rather my unfamiliarity 03:41
M-tadzik I think ideally it should just work for you and you need not seek help :) 03:42
Cryptoknight yes, but you know how these things so often go 03:43
ShimmerFairy M-tadzik: added pull request, when you have time to look it over :) 03:44
M-tadzik Cryptoknight (IRC): thanks, will do :) 03:45
pink_mist . o O ( that was ShimmerFairy )
Cryptoknight does build-panda require perl6 ? 03:48
M-tadzik Yes 03:50
03:54 chenryn left
Cryptoknight ok, part of my problem was that I wasnt in bash, so I'm starting over 04:02
personally, I prefer zsh. 04:04
m-tadzik, did you write the installation instructions by any chance ? 04:05
04:13 adu joined
Cryptoknight anyway, to whoever did: thanks! good job, really. 04:13
04:19 laouji_ left
tokuhirom Why is .^add_method lower_case? 04:19
04:19 laouji joined, laouji left, laouji joined
M-tadzik Cryptoknight (IRC): I might have, but it could be one of the awesome contributors :) 04:20
04:20 binjoy joined 04:25 chenryn joined
Cryptoknight M-t: for all that you do, this post-increment is for you: ++ 04:26
04:26 thou joined
M-tadzik :) 04:28
04:31 thou left 04:32 TingPing joined, binjoy left 04:34 thowe joined
skids tokuhirom: ^add_method is part of the MOP, which looks different than normal Perl 6 naming conventions, perhaps for the best. 04:41
tokuhirom skids: ah. i see! 04:42
skids The ALLCAPS stuff is more for strange things you might normally do in the middle of normal code -- MOP things hopefuly will be more and more isolated as time goes on. 04:44
thowe is "use v6;" still something that is done? 04:56
ShimmerFairy You don't have to, but it's a good idea nonetheless :) 04:57
04:57 laouji left
thowe Is the explanation for why it's a good idea brief? 04:58
04:58 laouji joined
ShimmerFairy thowe: among other things, it's just a nice way of saying "this is definitely Perl 6", and it makes for a better error message if you try running it with Perl 5 by accident. 05:00
thowe I'll buy that.
ShimmerFairy (and in the future, when we have newer versions of Perl 6, it'd of course serve as a way of saying which version of P6 you support)
thowe nice
05:00 chenryn left
Cryptoknight I dunno. I never preface any conversations by saying I'm speaking English, or what regional version thereof, so it always seemed rather extraneous to me to say 'use v6' in my code 05:01
but the version supported thing is an excellent point. 05:02
ShimmerFairy It's really just like Perl 5's 'use v5.20.2' statement, but for Perl 6 :) 05:03
05:03 adu left
thowe I'm missing something about named arguments... 05:04
I get this... :named(35)
but how can this work? :35named 05:05
can't the 35 be anything?
Also, I'm not sure if this doc is showing me the definition or the calling form...
05:08 chenryn joined
thowe oh, there's a whole other doc on signatures... 05:08
there's a lot of language here...
05:11 skids left 05:14 aborazmeh left
[Coke] :65foo is syntax that is to enable things like :3rd , and was genericized. 05:20
m: say :35.named.perl
camelia rakudo-moar 94d457: OUTPUT«5===SORRY!5=== Error while compiling /tmp/_bqT7gSVIt␤Malformed radix number␤at /tmp/_bqT7gSVIt:1␤------> 3say :357⏏5.named.perl␤ expecting any of:␤ number in radix notation␤»
[Coke] m: say :35named.perl
camelia rakudo-moar 94d457: OUTPUT«:named(35)␤»
[Coke] m: say :666beast.perl 05:21
camelia rakudo-moar 94d457: OUTPUT«:beast(666)␤»
05:30 Cryptoknight left, Cryptoknight joined 05:40 [TuxCM] joined
[TuxCM] test 50000 34.462 34.352 05:51
test-t 50000 35.271 35.162
\o/ way under 36 :)
tux.nl/Talks/CSV6/speed4.html 05:52
best I measured - ever
05:58 bjz joined 06:04 bjz left, diana_olhovik_ joined 06:05 itwalter joined 06:08 FROGGS left 06:09 FROGGS[mobile] joined
FROGGS[mobile] morning 06:10
.tell jnthn the MOP is fun btw
yoleaux FROGGS[mobile]: I'll pass your message to jnthn.
06:15 thou joined 06:20 thou left
dalek ast: e361a0e | FROGGS++ | S02-types/subset.t:
fix typo in type name
 06:21
06:24 bjz joined, FROGGS joined, FROGGS[mobile] left 06:31 Cryptoknight left 06:32 Cryptoknight joined 06:35 domidumont joined, firstdayonthejob joined 06:38 vytas left 06:39 Ven joined 06:44 vytas joined 07:00 z8 joined, z8 left 07:01 [TuxCM] left
lizmat .botsnack 07:05
yoleaux :D
synbot6 om nom nom
yoleaux 8 Oct 2015 21:21Z <FROGGS> lizmat: is it possible that you wait working on smileys (implementation wise, not test wise) until I've merged the MOP stuff?
lizmat I'll gladly postpone that :-) 07:06
FROGGS :o)
07:10 nuc_ left
lizmat .tell jdv79 you probably want to check with _itz_ about inclusion in the pl6anet.org feed 07:17
yoleaux lizmat: I'll pass your message to jdv79.
07:19 [TuxCM] joined
lizmat [TuxCM]: great to hear it's better, but I don't see exactly *why* that happened :-( 07:20
meanwhile, I see bare startup at 115 ~ 120 msecs again :-( 07:21
07:25 firstdayonthejob left, zakharyas joined 07:34 ely-se joined, sivoais left
masak morning, #perl6 07:35
lizmat masak o/ 07:36
ely-se hello world
masak hely-se :)
ahoj, lizmat 07:37
I would just like to say that 007 is moving forward quite nicely
(and then be all mysterious about it and divulge nothing more) :P
lizmat
.oO( I hear the movie is coming out soon :-)
07:39 Ven left 07:41 Ven joined 07:43 sivoais joined 07:46 z8 joined 07:48 g5 joined
FROGGS *g* 07:49
ely-se The next Perl should be Perl 007. 07:51
masak actually, Perl 006 is already so advanced it's going to backport all features from 007 :P 07:52
07:53 sivoais left 07:55 patrickz joined, sivoais joined
ely-se Wrote my first parser in Perl 5 yesterday. 07:58
07:59 RabidGravy joined, jkva joined
patrickz .tell jdv79 (wrt your Blog post) Do you know about testers.perl6.org/ ? 08:00
yoleaux patrickz: I'll pass your message to jdv79.
Ven ely-se: that's not the first one. Mill's was the first
masak ely-se: how was it? I find it a similar-but-different experience to writing a Perl 6 parser.
08:03 thou joined
chenryn learning NativeCall, how to push a NULL into CArray.new? I don't know what's the NULL in Perl6? 08:03
08:04 sivoais left 08:05 sivoais joined
ely-se Ven: oh, hmm. :( 08:06
indeed :v
Ven ely-se: I remember :-).
ely-se masak: I used Parser::MGC which is very nice.
Ven ely-se: shoulda used faux-combinator/perl, told ya
08:06 jkva left, [Sno] left 08:07 thou left, jkva joined
ely-se no 08:08
it's not on CPAN; CBA to figure out how to install it :P
Ven aw, true
dalek oblem_solver_tutorial: 3f2cb79 | lichtkind++ | menu.md:
TOC test
ely-se also it passes a reference to push, which is experimental 08:09
masak .oO( CPAN or it didn't happen )
ely-se Your code's bad and you should feel bad! 08:10
Ven feels bad 08:11
ely-se yay!
Ven even does it thrice :-).
it felt good. 08:12
dalek oblem_solver_tutorial: 4bae7d7 | lichtkind++ | menu.md:
TOC indent fix
08:13 darutoko joined 08:14 sivoais left
Ven ely-se: am I supposed to @foo[@#foo] = $ref instead, btw? 08:15
08:15 sivoais joined
Ven err, $foo and $#foo. sigil variance... 08:15
masak (this is #perl6) :P 08:16
08:16 antiatom left 08:17 Celelibi_ joined
ely-se no, push @$foo, $x; 08:17
08:18 Celelibi left, patrickz left
Ven ely-se: fixeth 08:21
masak: sorry :-).
08:22 antiatom joined
jnthn morning, #perl6 08:23
yoleaux 8 Oct 2015 21:25Z <FROGGS> jnthn: does nqp::settypecheckmode affect nqp::istype?
06:10Z <FROGGS[mobile]> jnthn: the MOP is fun btw
Ven waves
what should be the namespace for MOP stuff? simply MOP? `use MOP::Ruby` *g*
jnthn chenryn: You push a type object of some kind (like if it's a CArray[SomeStruct] then just push (literally) SomeStruct)
woolfy o/ jnthn
08:24 domidumont left, sivoais left
masak jnthn++ # knowing the answer to that 08:24
jnthn: is it documented somewhere that type objects are the way to do NULL in NativeCall? 08:25
Ven \o woolfy
08:25 domidumont joined
jnthn .tell FROGGS Yes, that's exactly what settypecheckmode controls the behavior of. And you probably want to set the type check mode to the same thing SubsetHOW does, and implement and accepts_type method that just does nqp::isconcrete 08:25
yoleaux jnthn: I'll pass your message to FROGGS.
FROGGS jnthn: funnily I'm doing something wrong in a way that nqp::istype(Int:D, Int:D) is true in the Actions, but rightfully not true in user scripts
yoleaux 08:25Z <jnthn> FROGGS: Yes, that's exactly what settypecheckmode controls the behavior of. And you probably want to set the type check mode to the same thing SubsetHOW does, and implement and accepts_type method that just does nqp::isconcrete
08:25 Ven left 08:26 sivoais joined
jnthn FROGGS: Did you (as CoercionHOW does) set up type canonicalization, so there's only ever One True Int:D? 08:26
FROGGS jnthn: aye, after cloning the behaviour of CoercionHOW I came to the conclusion that I really want SubsetHow, and did that
jnthn No, you need to breed them!
FROGGS I... WAT?
jnthn So you get the type check behavior of SubsetHOW combined with the interning behavior of CoercionHOW 08:27
FROGGS hmmmm
okay
jnthn o/ woolfy
.tell Ven In Rakudo it's exposed under Metamodel:: but in modules I've written that are MOP modules I've not tended to call that out in the module name. 08:28
yoleaux jnthn: I'll pass your message to Ven.
jnthn Who cares that Grammar::Tracer and OO::Monitors use the MOP? It's just how they do what it says on the tin :) 08:29
masak: Yes. 08:30
In doc.perl6.org/language/nativecall both for strings and for other C-ish types
jnthn wonders if we shouldn't just inline the win32 example linked at the bottom of that article 08:31
It's probably the only real reason for the jnthn/zavolaj repo to continue to exist...
masak +1 on not calling out things as MOP just because they MOP
jnthn oh, there's a few more examples in there too
Could maybe migrate to examples.perl6.org in the future 08:32
dalek ast: 52c04f4 | lizmat++ | S12-attributes/smiley.t:
Add some more attribute smiley tests
jnthn Time for my Greek final...
masak I have another philosophical question, also prompted by 007 work: in `my $b = { ... };`, I would say `{ ... }` is a block literal. but is it in `if 2 + 2 == 4 { ... }` ?
jnthn: "My Big Fat Greek Final" 08:33
lizmat masak: I would say, yes
jnthn Sigmas do look a bit on the chubby side...
08:33 sivoais left
jnthn masak: Yes; as another data point both (preliminarily) compile into a QAST::Block also 08:33
masak don't literals occur only in expressions, though?
jnthn Though they are a bit different 08:34
masak yes, that's what I'm feeling
jnthn The first is blocktype declaration, the second blocktype immediate
Because you have to distinguish
if foo { bar } else { baz }
from
masak oh, and `{ ... }` as a standalone statement is also blocktype immediate, right?
jnthn foo ?? { bar } !! { baz }
masak: yes
08:35 laouji left, laouji joined, espadrine joined 08:36 sivoais joined
masak ok. in 007, immediate blocks are simply Q::Statement::Block. so maybe that's what should go in an `if` statement, then, and not Q::Literal::Block... 08:36
08:37 rurban joined
jnthn m: say Uni.new(0x03A3, 0x03C3, 0x03C2).Str 08:38
camelia rakudo-moar 94d457: OUTPUT«Σσς␤»
masak what makes it extra weird is that once you've said "no, `if` is a special form so it can't contain a literal", someone else goes, "ah, but what if it were defined as a macro?"
jnthn I guess the difference here is that at QAST level we don't care if you wrote if, unless, and, or, &&, ||, a ternary...they're all QAST::Op if/unless 08:39
But at the leavel you're working you do care
masak yeah
I want to care in a way that feels consistent and makes sense 08:40
"literals can only go in expressions" makes sense to me, but it's not the only possible way to make the distinction
*structurally*, it feels slightly better to say that an `if` statement is (Q::Expr, Q::Statement::Block) than (Q::Expr, Q::Literal::Block) 08:42
though I can't really put my finger on why
08:44 sivoais left 08:46 sivoais joined
dalek ast: 382fa02 | jnthn++ | S32-str/lc.t:
More complete final sigma tests.
 08:47
RabidGravy morning! 08:49
lizmat RabidGravy o/ 08:51
jnthn o/ RabidGravy 08:52
08:53 Cryptoknight left 08:54 Cryptoknight joined, sivoais left 08:56 sivoais joined
RabidGravy are we all winning this fine morning? 08:57
lizmat so far :-) 08:58
09:00 FireFly left
RabidGravy fabulous! 09:02
09:04 sivoais left 09:05 FireFly joined 09:06 sivoais joined, spider-mario joined 09:08 jkva left 09:11 domidumont left 09:12 domidumont joined 09:14 sivoais left
masak Wikipedia says "In computer science, a literal is a notation for representing a fixed value in source code." 09:14
jnthn uses up his monthly goto 09:15
masak there's a certain sense in which the `if` code block is not "a fixed value in source code", the way a block literal assigned to a variable is
in the sense of "a first-class value".
jnthn masak: And neither is a fixed value
masak: Thanks to closure semantics. 09:16
09:16 Cryptoknight left, sivoais joined
masak arguably the thing that is fixed is the reference to the code block. 09:17
09:17 Cryptoknight joined, ely-se left
masak as in, you could declare the block literal as a `const` (in ES6, say), and changing a value that the block closed over would still be allowed. 09:19
(because the block "itself" didn't change)
very similar to how you can declare a `const obj = { x: 42 };` and then go `obj.x = 5;`
09:21 ely-se joined 09:22 ecocode joined
masak (and if that seems wrong to you, on an intuitive level, you're not alone. it kind of sucks. fixing it is still an open research topic.) 09:22
09:24 sivoais left 09:25 TEttinger left
jnthn masak: That's one of those "all the solutions make things worse" areas at the moment... 09:25
masak aye 09:26
09:27 sivoais joined
masak I believe I have seem glimpses of the final answer to it -- immutability, borrowed pointers, linear types -- but the tradeoffs currently seem too harsh. 09:27
09:28 antiatom left
jnthn I think Clojure does something interestingish there (pervasive use of persistent data structures) 09:29
09:29 Psyche^_ joined, Cryptoknight left, domidumont left 09:30 Cryptoknight joined
masak aye 09:30
09:31 domidumont joined
masak but even with persistent data structures, you have to have a really good notion of "who owns what" and "what contains what" 09:31
jnthn *nod*
We have a couple of xmas RTs on readonlyness that I suspect are going to be rejected along these lines. 09:32
(In the list 'cus we need to make a decision that we're not taking the problem on.)
09:32 antiatom joined 09:33 Psyche^ left 09:34 kipd left, olinkl left, SmokeMachine_ left, sivoais left
btyler timotimo: re "no need to go onto seperate thread" in gist.github.com/kanatohodets/04c34...5db983d6dc -- I'm less interested in having a CPU core, and more interested in CSP-style/golang style concurrency, where I can set up a goroutine/promise to hang out and react to what it is interested in independent of others 09:36
what I wrote there is very close to a go socket server pattern: 'go handle_client(connection)' and then a 'select' in handle_client
09:37 riatre joined, sivoais joined
masak jnthn: aye. to me it feels like the spec overpromises with things like "readonly one level down", without realizing that that's a tall order. 09:38
jnthn Thankfully, they're only design docs, not a spec :P 09:39
Now, if we have spectests that want this behavior... :)
09:39 telex left
btyler that said, maybe I'm doin' it wrong in that gist, since Moar gives me a 'I unwound the entire stack and there's no handler left!' when I disconnect clients, and the JVM doesn't want to consume data from the socket in handle-client 09:40
dalek ast: 2b70f97 | jnthn++ | S32-str/uc.t:
Add test for uc of precomposed lower -> NFG upper.

Yet another fun case!
09:40 telex joined
dalek p: be87a41 | jnthn++ | tools/build/MOAR_REVISION:
Bump MOAR_REVISION for SpecialCasing handling.
 09:40
kudo/nom: e97ba11 | jnthn++ | tools/build/NQP_REVISION:
Bump for SpecialCasing support.
 09:41
kudo/nom: 2c1af83 | lizmat++ | src/core/Any-iterable-methods.pm:
We don't need no scopes to grep
 09:42
lizmat jnthn: do you have some juicy examples for the P6W ? 09:43
dalek ast: cb062b4 | jnthn++ | S32-str/ (2 files):
We now get uc of German Sharp S correct.
ast: 57b8aa5 | jnthn++ | S32-str/lc.t:
Unfudge tests for final sigma semantics.
jnthn lizmat: S32-str/fc.t, plus the tests I just unfudged/added 09:44
lizmat okidoki
09:45 sivoais left
lizmat so, could we argue we have the best unicode support in the world now ? 09:45
jnthn You can argue anything you want, though you might not win :P 09:46
We've still the "what is a grapheme" NFG issue to go
And I'm sure we'll discover more weird cases in the future
We can certainly say we're doing really pretty good on it, though :) 09:47
ShimmerFairy jnthn: is there any reason to think "extended grapheme cluster" won't cover everything?
jnthn ShimmerFairy: I meant weird cases in Unicode support more generally, not just NFG
ShimmerFairy ah :)
09:47 sivoais joined
jnthn ShimmerFairy: Though we'll have things there too, just perhaps not around "what is a grapheme" 09:48
ShimmerFairy: I didn't yet fix the interaction of SpecialCasing and NFG, for example...
ShimmerFairy jnthn: one thing that's on my mind there is the CLDR. I really need to look into it sometime, but I suspect it's definitely worth a module :)
jnthn And while *that* issue I do know about and will take care of...there'll probably be things I didn't think of
ShimmerFairy: Yeah, that'll go module space
chansen_ I was pleasantly surprised when I read about Strings in Swift 2, <developer.apple.com/swift/blog/?id=30>
09:51 ecocode left, thou joined 09:55 thou left
jnthn chansen_: Good they're also going in the "by character people mean grapheme" direction 09:56
09:56 sivoais left, kipd joined, olinkl joined
jnthn Not doing the synthetics thing we are, alas. 09:56
But still an improvement.
chansen_ True
09:56 SmokeMachine_ joined 09:57 kipd left, olinkl left, SmokeMachine_ left, rindolf joined
literal the link at the bottom of this page (and similar ones) is wrong: doc.perl6.org/type/Cool 09:57
"github.com/perl6/doc/raw/master/li.../Cool.pod" =~ s{raw/}{} to fix it
09:58 sivoais joined 09:59 Cryptoknight left 10:00 Cryptoknight joined 10:01 jkva joined, Ven joined
jnthn Anyways, that's the SpecialCasing ticket knocked off the xmas RT list :) 10:03
10:04 itwalter left, SmokeMachine_ joined 10:05 antiatom left, laouji left, jkva left
ely-se special cases must die 10:06
10:06 sivoais left
dalek kudo/nom: 911cf4a | lizmat++ | src/core/native_array.pm:
Don't need a my int $i to keep track
 10:06
10:07 laouji joined
jnthn ely-se: When it comes to language, humanity seems hugely unconvinced about this :) 10:07
10:08 sivoais joined
lizmat
.oO( I think we can make an exception for that :-)
 10:08
jnthn
.oO( X::HumansSuck )
 10:10
10:12 antiatom joined
gfldex i would even go so far to say that mammals suck 10:13
ely-se sucking is my hobby
RabidGravy :-O 10:17
btyler jnthn: if you have spare brane for a concurrency question -- is something like this gist.github.com/kanatohodets/04c34...5db983d6dc in reasonable alignment with the concurrency design goals?
10:17 sivoais left, jkva joined
btyler e.g. handle a socket, toss it off to a promise, have that socket do it's own 'react' with a central supply to interact with other sockets 10:17
*its
10:18 pippo joined, sivoais joined
btyler I'm glad to hunt and golf breakage in async bits, just want to make sure I'm holding it the right way, and that the breakage isn't just LTA error handling on my broken design 10:18
jnthn btyler: It's fragile because if there's a death in the Promise, nothing gets it
Also, since the message bus is shared, I'd maybe put it's whenever in the top-level react block 10:20
oh wait
I misunderstood. It's fine :)
You're doing a chat server like thing :) 10:21
btyler well, it's just a silly little example for go-style communicating processes
jnthn Oh, and that return is wrong
btyler go/erlang/whatever
jnthn Should be "done"
btyler oh, nice, that sounds much better
jnthn But yeah, I'd stick a CATCH in handle-socket just in case 10:22
btyler re "nothing gets the failure" -- that sounds quite a bit like 'unwound the stack and nothing caught it, kaboom'
awesome, I'll do that and then poke around to see if things are still fragile. thanks!
jnthn If so lemme know 10:24
10:24 jkva left
dalek kudo/nom: dca64b9 | jnthn++ | src/Perl6/Grammar.nqp:
Toss <colonpair> case of trait.

We're not doing this in 6.christmas. It may return in the future, but it caused some crappy errors while unimplemented, so for now we just kill it off entirely.		 10:24
jnthn There's at lesat one known cause of fragility in Moar that's on my todo list
Though you have to be a little unlucky to hit it 10:25
btyler some of the ones I've bumped into are simple enough, usually just needs a little more robustness in the interaction with libuv
dalek ast: e5a0358 | jnthn++ | S32-exceptions/misc.t:
Tests for RT #119763 and RT #117417.
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=119763
btyler double closing an async socket, for example
10:26 sivoais left 10:27 andreoss joined
andreoss m: say [+] <3, 1, 4> x ** 2 10:28
camelia rakudo-moar 911cf4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/dh6ferwlig␤Two terms in a row␤at /tmp/dh6ferwlig:1␤------> 3say [+] <3, 1, 4> x **7⏏5 2␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ …»
Zoffix I recall someone mentioning dropping any performance stuff as a gist here... so... here: gist.github.com/zoffixznet/68c6b3bc85c70e291fd4
It goes up by about .5s more if I run the code from file instead of -e 10:29
10:29 sivoais joined
FROGGS Zoffix: is Inline::Perl5 precompiled? 10:29
(I guess not, which might explain things) 10:30
Zoffix ¯\_(ツ)_/¯
FROGGS *g*
Zoffix I just installed it with panda last night.
FROGGS and panda does not precompile anymore
Zoffix Ah
jnthn dca64b9 picked off one more xmas issue :)
Zoffix \o/
jnthn So, which next.. 10:31
FROGGS Zoffix: how long does this take? time perl6 -MInline::Perl5 -e ''
dalek kudo/nom: bcc4d34 | lizmat++ | src/core/native_array.pm:
Make iterating whole native arrays ~2x as fast
FROGGS jnthn++
jnthn: gist.github.com/FROGGS/cf2a24d89361e760e7bc
10:31 Woodi left
Zoffix FROGGS, real0m3.268s 10:31
FROGGS Zoffix: see
Zoffix yeah
FROGGS we should add precomp support to rakudo soonish, really 10:32
RabidGravy right off to the seaside to take advantage of the clement weather
have fun
10:33 chenryn left, Ven left
FROGGS have fun RabidGravy 10:33
jnthn FROGGS: Does the module databae need to survive a rebuild of Rakudo though/ 10:35
?
10:35 kipd joined
lizmat jnthn: :-( 10:35
did you read my CURLI gist ??? 10:36
jnthn Yes, but I need to again, and the comments on it, and... :)
lizmat gist.github.com/lizmat/f3807956c354c14902a3
in my opinion, it should
jnthn lizmat: OK, in which case we can't use the VM-level serializer for it 10:37
FROGGS jnthn: no, we could habe a json backup for that case
10:37 sivoais left, RabidGravy left
lizmat or any other, dedicated format, backup 10:37
FROGGS jnthn: but of course it would be better if we'd not need to use that backup
lizmat it's all in the gist
which also addresses distribution of info needed at different stages 10:38
jnthn OK, let me do some reading
lizmat jnthn: thank you :-)
jnthn And probably then some designing :)
ShimmerFairy FROGGS: I personally don't think that much of JSON, doesn't even allow comments. :) .oO(I know! Let's create yet another markup language... wait...)
FROGGS ShimmerFairy: what do we need comments for?
and yes, I dont want to design a language, that's why json 10:39
10:39 sivoais joined
ShimmerFairy FROGGS: nothin', just writing my own META6.json the other day and wanted to leave a comment. Turns out JSON doesn't want you make your configuration/metadata files explained :) 10:39
lizmat jnthn: that would be brill
jnthn The other two big items on my todo list at the moment are correcting the NFG algo and refactoring Moar's sync IO...neither of which are much less hairy :) 10:40
FROGGS lizmat: your gist does not tell anything about serializing a databaseish thing, right?
ShimmerFairy FROGGS: I'd be more than happy to come up with a YAML-ish thing (my understanding is that YAML is a Perl thing, right? If so, YAML6! :P), if only my head wasn't in my Pod6 work... :)
FROGGS ShimmerFairy: I... don't like YAML much 10:41
ShimmerFairy FROGGS: yeah, I've not looked too much into it, but it seems too much for configuration/metadata/etc.
FROGGS ShimmerFairy: JSON is limited, yes, but that helps me remembering how to write it
lizmat FROGGS: it does, it's the installed modules meta info in the .precomp dir
and the runtime meta info for each compunit, also under .precomp 10:42
ShimmerFairy FROGGS: what didn't help was that emacs goes to Javascript mode for JSON, and it told me // comments would work just fine, by coloring them as normal :P
FROGGS lizmat: yes, you state that the meta info should be precompiled for performance reasons, but that's not quite detailed :o) 10:43
lizmat well, since it lives in the .precomp dir, which is compilation bound, it sort of implies it's precomped MoarVM style to me :-)
FROGGS lizmat: and there is nothing I can see about the livetime of the precomped thing, which is what the topic was about
andreoss www.pigdog.org/auto/software_jihad/.../3138.html they probably meant say [+] <3 1 4> X** 2 10:44
lizmat FROGGS: well, that's really a meta-issue
FROGGS lizmat: yes, I understand that your were talking about the moarvm serialization format, though that does not solve the issues we're seeing
andreoss what 'say [+] <3, 1, 4> x ** 2' could do?
lizmat do we want to keep different compilations around when switching versions with e.g. rakudobrew ?
FROGGS lizmat: we are trying to solve a bug
jnthn andreoss: looks like infix where term expected to me 10:45
m: say [+] <3, 1, 4> x ** 2
camelia rakudo-moar 911cf4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/_O7T_HFemQ␤Two terms in a row␤at /tmp/_O7T_HFemQ:1␤------> 3say [+] <3, 1, 4> x **7⏏5 2␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ …»
jnthn Exactly
FROGGS m: say [+] <3, 1, 4>, x ** 2
camelia rakudo-moar 911cf4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Hi5owelEs9␤Two terms in a row␤at /tmp/Hi5owelEs9:1␤------> 3say [+] <3, 1, 4>, x **7⏏5 2␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ …»
FROGGS m: say [+] <3, 1, 4>, 'x' ** 2
camelia rakudo-moar 911cf4: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '7⏏5x' (indicated by 7⏏)␤ in block <unit> at /tmp/lFZu4MQh8A:1␤␤Actually thrown at:␤ in block <unit> at /tmp/lFZu4MQh8A:1␤␤»
FROGGS m: say [+] <3, 1, 4>, 42 ** 2
camelia rakudo-moar 911cf4: OUTPUT«1767␤»
lizmat FROGGS: agree fixing the seriali/deserial any data bug is a prerequisite for further work 10:46
10:46 mprelude joined, ely-se left
FROGGS lizmat: it would be handy to support switching, but that's one step too much right now me thinks... at least implementation wise (not design wise) 10:46
lizmat FROGGS: let us at least set it up in such a way that it *might* work in the future 10:47
FROGGS lizmat: read my gist to understand what the showstopper wrt serialization is: gist.github.com/FROGGS/cf2a24d89361e760e7bc
10:47 sivoais left
lizmat otherwise we get in all sorts of issues with installed base after Xmas 10:47
FROGGS lizmat: yes, I very much agree
andreoss FROGGS: isn't <3, 1, 3> coerced to 3 here? 10:48
lizmat FROGGS: just thinking: might using an EVAL with a fresh :context help ?
andreoss m: say [+] 3, 42 ** 2
camelia rakudo-moar 911cf4: OUTPUT«1767␤»
lizmat or with :context(CORE::) ?
FROGGS m: say [+] |<3, 1, 4>, 42 ** 2
camelia rakudo-moar 911cf4: OUTPUT«Cannot convert string to number: trailing characters after number in '0337⏏5,' (indicated by 7⏏)␤ in block <unit> at /tmp/gADypNwf5_:1␤␤Actually thrown at:␤ in block <unit> at /tmp/gADypNwf5_:1␤␤»
lizmat FROGGS: wouldn't that fix the dependency issues ?
FROGGS ohh 10:49
andreoss: either use (3, 1, 3) or <3 1 3>
10:49 olinkl joined
andreoss somebody should tell them to fix that in their article 10:49
FROGGS lizmat: no, EVAL at that stage failed too, tried that (also with :context)
andreoss also sub infix: ($a, $b) {}; doesn't make sense
10:50 sivoais joined
FROGGS andreoss: who are "they"? 10:50
andreoss FROGGS: www.pigdog.org/auto/software_jihad/.../3138.html
ilmari should commas in <>? like they do in qw() in perl5?
should commas *warn*
10:51 ely-se joined
FROGGS ilmari: I don't like warnings... it is either legal and should be quiet or it is an error and should explode... an "in between" only exists in rare exceptions to me 10:51
10:52 kid51 joined
ShimmerFairy m: say <> # that reminds me; I think this should obviously work as () (empty list) 10:52
camelia rakudo-moar 911cf4: OUTPUT«5===SORRY!5=== Error while compiling /tmp/zEda3ZMPZU␤Unsupported use of <>; in Perl 6 please use lines() to read input, ('') to represent a null string or () to represent an empty list␤at /tmp/zEda3ZMPZU:1␤------> 3say <7⏏5> # that reminds …»
10:57 sivoais left 10:58 antiatom left 10:59 chenryn joined 11:00 chenryn left, sivoais joined 11:01 chenryn joined
chenryn one more question about nativecall @jnthn: what's __int128? 11:04
FROGGS chenryn: nothing we support right now
chenryn so, is there some other way? 11:05
11:05 antiatom joined
FROGGS chenryn: you can read two int64 11:06
if this is about unpacking memory
jnthn has never seen __int128 :) 11:07
ilmari postgres uses it when aggregating 64bit integers 11:09
11:09 sivoais left
jnthn Ah, interesting. 11:09
jnthn so far didn't use postgres much
jnthn finally read all the gist and comments 11:10
11:10 laouji left 11:11 sivoais joined 11:12 laouji joined 11:13 antiatom left
ilmari git.postgresql.org/gitweb/?p=postgr...0ee76e488a 11:14
probably not very portable
jdv79 oh i think i mistook dha's usage as being example usage not abstract 11:16
yoleaux 07:17Z <lizmat> jdv79: you probably want to check with _itz_ about inclusion in the pl6anet.org feed
08:00Z <patrickz> jdv79: (wrt your Blog post) Do you know about testers.perl6.org/ ?
Zoffix oooo shiny. 11:17
Zoffix didn't know about that site
jdv79 .tell lizmat i did and i saw my inclusion PR get merged but still don't see my post on the site
yoleaux jdv79: I'll pass your message to lizmat.
Zoffix jdv79, I'm pretty sure I saw it announced on Twitter last night. 11:18
I recall reading something other than gist.
jdv79 .tell patrickz yes. i'd like to try to get cpants support though since that's more mature
yoleaux jdv79: I'll pass your message to patrickz.
Zoffix (and noticing a misspelt "Its" that I've missed originally)
11:19 sivoais left
jdv79 .tell patrickz less work for all:) 11:19
yoleaux jdv79: I'll pass your message to patrickz.
chenryn and CStruct don't support bool now?
Zoffix Ah. This is what I've read. jdv79.blogspot.ca/2015/10/perl6-and...um=twitter 11:20
jdv79 .tell patrickz oh, seems cpants and cpantesters aren't the same thing. I mean the latter.
yoleaux jdv79: I'll pass your message to patrickz.
jdv79 that is what i made, yes 11:21
11:21 sivoais joined
FROGGS m: use NativeCall; class Foo is repr<CStruct> { has Bool $.a } 11:25
camelia rakudo-moar bcc4d3: OUTPUT«===SORRY!===␤CStruct representation only handles int, num, CArray, CPointer, CStruct, CPPStruct and CUnion␤»
FROGGS m: use NativeCall; class Foo is repr<CPPStruct> { has Bool $.a }
camelia rakudo-moar bcc4d3: OUTPUT«===SORRY!===␤CPPStruct representation only handles int, num, CArray, CPointer, CStruct, CPPStruct and CUnion␤»
FROGGS hmmm, I thought that at least CPPStruct would support it
11:25 antiatom joined
jnthn also re-read S11 and S22 11:27
11:29 Ven joined 11:30 sivoais left
jnthn Lunch, and got a few other things to do this afternoon. But will be doing some thinking on the module installation stuffs 11:31
lizmat jnthn++
yoleaux 11:17Z <jdv79> lizmat: i did and i saw my inclusion PR get merged but still don't see my post on the site
11:31 sivoais joined
jnthn bbiab 11:32
dalek kudo/nom: 9371b58 | lizmat++ | src/core/native_array.pm:
Made for int @ -> \i, \j { } about 15% faster
11:33 jkva joined
grondilu anyone uses lua? I'm wondering what would be the best way to implement Lua's tables. 11:37
that is what would be the Perl 6 equivalent.
11:37 rotcpy left
Ven grondilu: a new MOP? I'm not sure we have a prototypal structure in 6 11:37
yoleaux 08:28Z <jnthn> Ven: In Rakudo it's exposed under Metamodel:: but in modules I've written that are MOP modules I've not tended to call that out in the module name.
rurban p2 has the lua tables 11:38
Ven jnthn: yeah, I know you used OO:: for monitors/..
(which is somewhat "MOP"-like)
11:39 thou joined 11:40 sivoais left, Peter_R joined, Cryptoknight left 11:41 Cryptoknight joined 11:42 sivoais joined
lizmat jdv79: I have no idea how pl6anet.org work... 11:43
11:43 thou left
dalek kudo/nom: bb5283c | lizmat++ | src/core/metaops.pm:
Remove unneccesary scope in the DwimIterator
 11:45
lizmat *works
nine Our CMS at work now requires File::Slurp, File::Slurp::Unicode and File::Slurper. Oh Perl 6, I can hardly await using you there...
moritz timotwdi :/ 11:46
lizmat There Is More Other Teas With Diluted Ingredients ? 11:47
ShimmerFairy
.oO( File::Slurp::Unicoder )
11:48 ely-se left
nine
.oO( File::Slurp::Unicodest )
 11:48
ilmari Filest::Surpest::Unicodest
grondilu m: class Table { has @!array is rw; handles AT-POS }; my Table $a .= new; $a[4] = pi; say $a[4]; 11:49
camelia rakudo-moar bcc4d3: OUTPUT«5===SORRY!5===␤Undeclared name:␤ AT-POS used at line 1␤Undeclared routine:␤ handles used at line 1␤␤Other potential difficulties:␤ useless use of 'is rw' on @!array␤ at /tmp/_CBhGwmEwv:1␤ ------> 3class Table { has @!a…»
grondilu m: class Table { has @!array is rw handles AT-POS }; my Table $a .= new; $a[4] = pi; say $a[4];
camelia rakudo-moar bcc4d3: OUTPUT«5===SORRY!5=== Error while compiling /tmp/5psZ6AmjkA␤Cannot invoke this object␤at /tmp/5psZ6AmjkA:1␤»
11:50 sivoais left
chenryn @FROGGS and the error report need modify too, at least can handle Str. 11:50
grondilu m: class Table { has @!array is rw handles <AT-POS> }; my Table $a .= new; $a[4] = pi; say $a[4];
camelia rakudo-moar bcc4d3: OUTPUT«Potential difficulties:␤ useless use of 'is rw' on @!array␤ at /tmp/HYe6qck6yK:1␤ ------> 3le { has @!array is rw handles <AT-POS> 7⏏5}; my Table $a .= new; $a[4] = pi; say $␤3.14159265358979␤»
grondilu m: class Table { has @!array handles <AT-POS> }; my Table $a .= new; $a[4] = pi; say $a[4]; 11:51
camelia rakudo-moar bcc4d3: OUTPUT«3.14159265358979␤»
11:51 ely-se joined
grondilu m: class Table { has @!array handles <AT-POS>; has %!hash handles <AT-KEY> }; my Table $a .= new; $a<foo> = pi; say $a<foo>; 11:51
camelia rakudo-moar bcc4d3: OUTPUT«3.14159265358979␤»
grondilu cool
grondilu wonders if he could make those anonymous. 11:52
m: class Table { has @! handles <AT-POS>; has %! handles <AT-KEY> }; my Table $a .= new; $a<foo> = pi; say $a<foo>;
camelia rakudo-moar bcc4d3: OUTPUT«5===SORRY!5=== Error while compiling /tmp/KIY6AvcZs5␤Cannot declare an anonymous attribute␤at /tmp/KIY6AvcZs5:1␤------> 3class Table { has @7⏏5! handles <AT-POS>; has %! handles <AT-K␤ expecting any of:␤ constraint␤»
grondilu anonymous attributes anyone?
11:52 sivoais joined, diana_olhovik_ left 11:53 diana_olhovik joined
grondilu I mean it's annoying to pick a variable name where an anonymous var would do 11:54
m: my $a; $a<foo> = pi; $a[4] = 42; 11:56
camelia rakudo-moar 9371b5: OUTPUT«Index out of range. Is: 4, should be in 0..0␤ in block <unit> at /tmp/d7V__fx9DW:1␤␤Actually thrown at:␤ in any at gen/moar/m-Metamodel.nqp:2860␤ in block <unit> at /tmp/d7V__fx9DW:1␤␤»
lizmat grondilu: and how you refer to it later ? only by using AT-POS/KEY ?
grondilu lizmat: yes
which in a way is a way to protect the attribute
11:57 kid51 left
daxim lizmat, this programmer wants a p6 equivalent of the "as" module: stackoverflow.com/questions/330294...short-name 11:57
11:58 rindolf left, rindolf joined
lizmat daxim: sadly, that's still NYI (but hopefully soon when we unblock on CURLI and related precomp stuff) 11:59
ShimmerFairy has yet to see the benefit of anonymous vars, preferring to just pick a name and go with it
grondilu m: my $a; $a<foo> = pi; say $a.WHAT 12:00
camelia rakudo-moar 9371b5: OUTPUT«(Hash)␤»
12:00 sivoais left 12:02 antiatom left 12:03 sivoais joined
nine I still don't get why the module database is such an extreme performance problem for rakudo. It's a couple of kilobytes of text. A Perl ought to read that in microseconds. 12:03
lizmat nine: the problem is really JSON and use of a grammar in there ? 12:04
that's why I would be in favour of a combination of precomp and dedicated data format
ShimmerFairy could module Foo::Bar { class Bar is export { ... } } accomplish the shorter name desire, at least on the module dev's side?
lizmat *not* for the distro info, but for the installed modules / runtime info 12:05
not sure if "is export" is valid on a class ? 12:06
ShimmerFairy sure it is, just tried it :)
moritz daxim, lizmat: I thought it was implemented, as use Shortname:someattr<Long::Name>; though I forgto what :someattr is
dalek kudo/nom: 28c3836 | peschwa++ | src/Perl6/World.nqp:
Clarify comment and stop caring about the number of parts.

The latter is handled later in add_categorical, which throws a proper typed exception.
kudo/nom: 6c01dcd | peschwa++ | src/Perl6/World.nqp:
Catch errors from unboxing in non-Op nibble_to_str case.

Otherwise we get cases that leak "Cannot unbox...", which is undesirable.
jdv79 nine: maybe if p6 had something like p5's JSON::XS
nine lizmat: I understand the technical issues. It's just the glaring irony that some of the brightest heads in the Perl universe have been struggling with that for months. It's Perl! It got large because of its text processing capabilities.
dalek ast: dee7a12 | peschwa++ | S06-operator-overloading/sub.t:
Add another test for constant circumfix declaration.
 12:07
lizmat well, with grammars we have a swiss army knife: not sure that beats a hammer for hammering though :-)
nine lizmat: that's a wonderful description of the problem :)
ShimmerFairy lizmat: just use a null-separated file of all the fields, that'd be the best solution :P 12:08
lizmat ShimmerFairy: something like that has been on my mind, but probably using something I could use .lines on though :-) 12:09
nine Do we need to support embedded newlines?
ShimmerFairy "my\0string".cstrs, clearly :)
12:10 ab6tract joined
lizmat nine: I don't think we need that for the modules meta-data or the runtime info 12:10
nine lizmat: sounds like a job for an .ini like format
lizmat for the distribution introspection, we clearly would need to allow for embedded newlines
ab6tract m: class Long::Foo { has $.t = "test" }; constant F = Long::Foo; my $x = F.new; say $x.t; 12:11
camelia rakudo-moar 9371b5: OUTPUT«test␤»
lizmat nine: even simpler than that, because it's not really supposed to be humar-readable anyway
*human
12:11 sivoais left
jdv79 how abuot dbm or sqlite then;) 12:11
ab6tract seems like I should be able to munge OUTER:: context to define the constant in the importing scope?
12:12 rurban left
ab6tract not sure if that 100% solves that SX question, but I posted it there anyway 12:12
nine dbm actually doesn't sound that bad. I don't know who long sqlite takes to initialize
ShimmerFairy lizmat: my first thought is one null for separating fields, two nulls for separating distributions :)
jdv79 i can guarantee faster than p6 and json:(
ab6tract lizmat: is there some obvious horror to the constant approach that I'm not seeing?
masak I think I've decided that a Q::CompUnit does have a Q::Block, even though there are no braces. 12:13
because the "intent"/"spirit" is still there, that of a block/lexpad that can hold variable values
12:13 sivoais joined
moritz sounds like Q::Scoped 12:13
12:14 chenryn left
nine time sqlite3 foo.sql 'select * from foo' 12:14
real 0m0.003s
moritz which is a role that both Q::CompUnit and Q::Block do
12:14 antiatom joined
nine That's quite ok, I'd say 12:14
lizmat m: class Foo::Bar::Baz { }; constant Zap := Foo::Bar::Baz; dd Zap.new # looks like that works, jdv79
camelia rakudo-moar 9371b5: OUTPUT«Foo::Bar::Baz.new␤»
jdv79 BUT wouldn't it be nice if p6 could parse stuff faster... 12:15
ab6tract lizmat: what is the bind providing there? it works fine with normal assignment
nine It's simple, extensible, scalable, fast. and has a compatible license. I'd go with sqlite.
12:15 weihan joined
ab6tract FWIW I support nine here 12:15
lizmat m: class Foo::Bar::Baz { }; my Zap := Foo::Bar::Baz; dd Zap.new # was thinking of this
camelia rakudo-moar 9371b5: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YmOuhfF3fJ␤Malformed my (did you mean to declare a sigilless \Zap or $Zap?)␤at /tmp/YmOuhfF3fJ:1␤------> 3class Foo::Bar::Baz { }; my Zap7⏏5 := Foo::Bar::Baz; dd Zap.new # was th␤»
jdv79 i was only kidding around but okk
lizmat m: class Foo::Bar::Baz { }; my \Zap := Foo::Bar::Baz; dd Zap.new # was thinking of this 12:16
camelia rakudo-moar 9371b5: OUTPUT«Foo::Bar::Baz.new␤»
timotimo jdv79: wanna make JSON::Fast faster? :)
jdv79 maybe
i'd perger if grammars adn the lang runtime were faster though
timotimo yes
very much so
nine jdv79: I have thought about sqlite for that job before your comment. I just didn't try and see how fast it would be.
lizmat nine ab6tract jdv79 ShimmerFairy : before we can deal with all those suggestions, we need to agree on the layout
masak moritz: Q::Scoped doesn't really have the right connotations for me...
jdv79 my only p6 module is terrible compared to it's p5 twin becuase of this.
ab6tract lizmat: fair enough
timotimo there's lots of things that could be improved in the grammar engine still
lizmat and jnthn now has his thinking cap on to see how we should move forward 12:17
timotimo but a few things that i can recall off-hand aren't helpful for json parsing at all
not helpful for rakudo's core setting either
or code in general
jdv79 lizmat: is there a summary of that?
i don't recall much talk in here about it
just curious
moritz masak: maybe s/Scoped/Scope/ ? 12:18
timotimo also, nobody seems to realize we DO have a JSON::XS-alike for rakudo already
12:18 antiatom left
lizmat gist.github.com/lizmat/f3807956c354c14902a3 # jdv79 12:18
ab6tract masak: Q::Pad ?
timotimo jdv79: one notable/noticable piece of overhead for json::tiny is how we construct match/result objects
12:18 Erendis42 joined
timotimo TT has already poured a bit of time into that with some success, but i expect there's more to be had 12:19
masak moritz: better. it feels like it trades concreteness for correctness, or something. not sure I want to make that trade.
ab6tract: in my mind, a "[lexical] pad" is something a block or a scope *has*
moritz a block *has* a scope
and that scope is implemented as a lexpad 12:20
ab6tract masak: yeah fair,
moritz except when it's not a lexical scope
timotimo there's also the thing about having $*ACTIONS as a dynamic variable; having fewer dynamic variables to look up may help performance, but a measurement is very required to verify that hunch
moritz dunno if a compunit is or has a scope
ab6tract masak: so what does it use rather than braces?
timotimo when you dump a piece of code with --target=ast, i believe you'll see that a CompUnit has a Block inside it as one of the earliest children 12:21
lizmat moritz: has a scope, I think
12:21 weihan_ joined
ab6tract is that really the only difference vs a block? 12:21
masak yes, agree with lizmat. has a scope.
timotimo aye, has a scope 12:22
masak I think CompUnit is practically a block. for convenience we've decided not to surround the whole program with braces.
but besides that syntactic difference, it has all the properties of an ordinary block.
ab6tract masak: ah, I though Q::Block was macro-related, sorry
timotimo the compunit rakudo builds has a var for the __args__ (param), then an empty Stmts (perhaps for any kind of var declarations, which it never seems to do) and then the next thing is Op(call) on a QAST::Block that has $!,$/, ..., EXPORT, GLOBALish, ...
so that's a clear indication that a compunit has a scope, rather than is a scope 12:23
12:23 Ven left
lizmat masak: that's basically what UNIT-OUTER is 12:24
the missing braces around the compunit
12:24 sivoais left
masak ab6tract: *everything* is macro-related :P 12:24
timotimo a macro shall allow you to replace Moar with jvm on the fly! :)
12:24 weihan left
masak timotimo: [patent pending] 12:25
lizmat
.oO( isn't that what Perl 5 is? :-)
12:25 sivoais joined
timotimo perl 5 is a macro? 12:25
lizmat you only need to configure the macro :-)
hmmm... maybe I got "has" and "is" confused :-) 12:26
12:26 andreoss left, antiatom joined
psch runtime backend replacing sounds interesting 12:30
moritz s/interesting/insane/
pink_mist moritz++ #couldn't have said it better myself
psch "make easy things easy, hard things possible, and insane things interesting"...? :P 12:31
pink_mist hah
12:32 Loren joined
ab6tract lizmat: you mean that you maybe feel a CURLI is more of a scope than a thing which has a scope? 12:33
lizmat no?
this was re the p5 is a macro meme
12:34 sivoais left
moritz imagines MoarVM sneakily overtaking the JVM's runloop, deleting JVM-specific from the registers, transmogrifying the JVM stack into a Moar stack 12:34
timotimo write that fanfiction, moritz :)
jnthn I'm *so* not implementing that :P 12:35
timotimo i'm not so implementing that
12:36 sivoais joined
moritz sounds like a good April 1st topic :-) 12:36
12:37 Ven joined
[Coke] RT: 1036; nom: 8; glr: 4; lta: 88; weird: 12; tests: 9; xmas: 84 12:38
ShimmerFairy
.oO(Today we proudly announce Mocha, a program that modifies your JVM process into a MoarVM process in real time...)
timotimo oh lord, the name fits %) 12:39
psch that's a js testing framework though 12:40
timotimo pff, nobody cares if names are already taken
ShimmerFairy psch: the alternative would be to have people scout various coffee shops, longest drink name wins :P
psch moarchaccino? :P 12:41
masak moritz: "The Perl community announces HovercraftFullOfEels, a handy utility to hotswap your VM at runtime." :P
timotimo :D 12:42
masak .oO( <python_community> stop using Monty Pyhon names for your projects! )
12:42 pmurias joined
timotimo "Introducing the GiantFootFromTheSky compression algorithm" 12:42
pmurias hi
12:42 llfourn left
masak ahoj, pmurias 12:42
ShimmerFairy stack transmogrification is done with an innovative SillyWalk memory traversal algorithm 12:43
pmurias moritz: there is a research project that allows turning the JVM into a Perl 6 VM 12:44
moritz: or anyother language VM
timotimo "perl6 ported to quadrocopter firmwares thanks to FlyingLessons"
12:45 sivoais left
pmurias moritz: it's called the graalvm 12:45
timotimo "for your mailing list subscription please choose from the following options:
spam, eggs, development news and spam
sausage, spam, puns, spam, spam, cabbage and spam, 12:46
..."
12:46 sivoais joined
nine I get the distinct impression that Friday has arrived at #perl6 ;) 12:46
ShimmerFairy "This Mocha project is dead!" "No it's not, it's pining for the fjords."
lizmat
.oO( nobody expects the singing Vikings to fly in )
 12:47
dalek p: 2104888 | (Pawel Murias)++ | src/vm/js/bin/run_tests.pl:
[js] reorder the tests
p: 6d79697 | (Pawel Murias)++ | src/vm/js/ (2 files):
[js] move the way closures are handle closer to the way the MoarVM backends does things

noop explicit takeclosure, and insert closures in a more declarative manner
masak .oO( nobody expects the submethod invocation! )
ShimmerFairy
.oO(nobody expects the <span> jQuery!)
12:48 vytas left
ely-se <center> 12:48
timotimo <marquee> 12:49
ely-se omg that's an awesome nickname, as it actually shows up as <center> :O
timotimo not on my watch^Wsmartphone^Wdesktop computer
12:49 vytas joined 12:50 weihan_ left, weihan joined
masak .oO( oh NULL you didn't! ) 12:53
12:54 rurban joined, sivoais left
pippo masak: how you do that thing ^^. I.e. your nick name appears as a '*' ? 12:54
grondilu std: constant ∞ = Inf; 12:55
camelia std 28329a7: OUTPUT«5===SORRY!5===␤Missing symbol in constant declaration at /tmp/nJO9AeVG4w line 1:␤------> 3constant 7⏏5∞ = Inf;␤Missing initializer on constant declaration at /tmp/nJO9AeVG4w line 1:␤------> 3constant 7⏏5∞ = Inf;␤Confused at /tmp/nJO9…»
grondilu can't I use ∞ as an identifier?
masak pippo: /me
jnthn m: say '∞' ~~ /<ident>/
camelia rakudo-moar 6c01dc: OUTPUT«Nil␤» 12:56
jnthn No :)
grondilu ok
jnthn m: constant \∞ = Inf;
camelia rakudo-moar 6c01dc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/gkAIGW5inZ␤Missing initializer on constant declaration␤at /tmp/gkAIGW5inZ:1␤------> 3constant7⏏5 \∞ = Inf;␤»
masak just wrote "/me just wrote ..."
ShimmerFairy m: constant term:<∞> = Inf; say ∞
camelia rakudo-moar 6c01dc: OUTPUT«Inf␤»
pippo OK. Thank.
jnthn m: constant \term:<∞> = Inf;
camelia ( no output )
12:56 sivoais joined
pippo :-)) 12:56
jnthn m: constant \term:<∞> = Inf; say ∞
camelia rakudo-moar 6c01dc: OUTPUT«Inf␤»
jnthn ^^
ShimmerFairy jnthn: do you need the \ ?
jnthn m: constant term:<∞> = Inf; say ∞
camelia rakudo-moar 6c01dc: OUTPUT«Inf␤»
jnthn No :)
grondilu cool
thanks
masak m: constant \\\\\\\\\\term:<∞> = Inf; say ∞ 12:57
camelia rakudo-moar 6c01dc: OUTPUT«5===SORRY!5=== Error while compiling /tmp/YsB4Rxg9AC␤Missing initializer on constant declaration␤at /tmp/YsB4Rxg9AC:1␤------> 3constant7⏏5 \\\\\\\\\\term:<∞> = Inf; say ∞␤»
jnthn One \ enough!
masak :P
12:57 ely-se left
masak some would say too much 12:57
pink_mist what did the \ actually accomplish there? 0_o
12:58 Cryptoknight left
moritz m: constant &term:<∞> = Inf; say ∞ 12:58
camelia rakudo-moar 6c01dc: OUTPUT«5===SORRY!5===␤Argument to "say" seems to be malformed␤at /tmp/rhVRj6DNFr:1␤------> 3constant &term:<∞> = Inf; say7⏏5 ∞␤Bogus postfix␤at /tmp/rhVRj6DNFr:1␤------> 3constant &term:<∞> = Inf; say 7⏏5∞␤ expecting any o…»
ShimmerFairy jnthn: btw, do you know off-hand if constant &infix:<A> := &infix:<B> is sufficient to create a 'synonym' ? When I've done it before I could create new multis on either name, and call from the other name, so I'm curious.
12:58 ely-se joined
ShimmerFairy (if it does work for making synonyms, it would be a better way of defining texas variants methinks) 12:59
jnthn ShimmerFairy: If it calls add_categorical it'd work 13:01
Dunno if it does
13:01 sufrostico joined, aborazmeh joined, aborazmeh left, aborazmeh joined
ShimmerFairy jnthn: ah, I'll check then :) . If not, guess I'll have to dream up a SynonymHOW or something :P 13:02
jnthn ShimmerFairy: Or just patch the constant declarator
jdv79 _itz_: is your thing broken or is mine? 13:03
jnthn m: constant &infix:<+++> = &infix:<+>; say 1 +++ 2
camelia rakudo-moar 6c01dc: OUTPUT«Parameter '$a' expected a writable container, but got Int value␤ in block <unit> at /tmp/PAoMArOSC2:1␤␤»
13:03 sufrosti1o joined, Ven left
jnthn m: my &infix:<+++> = &infix:<+>; say 1 +++ 2 13:03
camelia rakudo-moar 6c01dc: OUTPUT«3␤»
jnthn ShimmerFairy: Seems not
13:03 jkva left 13:04 sivoais left
ShimmerFairy jnthn: oh, so 'constant' supposed to work in making synonyms then? (In the sense of "both refer to the same thing", aka aliases or pointers or whatever-you-call-em) 13:04
jnthn ShimmerFairy: I don't see why not right off, though I've got my head in a heap profiler at the moment for $other-job so... :)
lizmat toons.gotblah.com/archive/glasberge...9-1501.gif :-) 13:05
jnthn :P
ShimmerFairy jnthn: no worries, that's why I said "off-hand", not "after a thorough investigation" :)
jnthn :)
13:06 sivoais joined
ShimmerFairy I did try naively implementing texas variants with 'constant' the other day, but it broke compilation :P 13:06
gfldex they are now making CPUs for perl6, see: www.qualcomm.com/news/releases/201...-ecosystem 13:07
takadonet morning all 13:08
lizmat takadonet o/ 13:11
dalek kudo/nom: b76e4ab | lizmat++ | src/core/Range.pm:
for ^10 -> i,j { } and @a = ^10 about 10% faster

Turns out there *is* a significant difference between updating an int attribute or an lexical int in a loop.		 13:12
kudo/nom: fd403a8 | lizmat++ | src/core/native_array.pm:
Make iterating over an entire int array 5% faster
13:14 sivoais left
masak another advantage of preferring $!foo to $.foo in methods: the former fails at compile time if you typo or change your attribute 13:14
13:16 sivoais joined
pmurias blogs.perl.org/users/pawel_murias/2...world.html # nqp-js blog post 13:16
lizmat Pawel++ but the post is a bit bare ?
Maybe something like: Proud to present NQP running on a Javascript backend! 13:17
pmurias lizmat: it's just the backend
ilmari huh, why is $i = $i + 1 10x faster than ++$i?
lizmat "a small, but necessary step towards supporting rakudo on a javascript backend" 13:18
pmurias I'll write a bigger post once I bootstrap the full NQP
pink_mist so how far are we from having perl6 running in the javascript of a browser? :D
lizmat ilmari: because $i is a native int, and $i++ is not implemented for natives yet
ilmari lizmat: ah
jnthn It...is
lizmat so it will first box the native, then ++ it, then unbox it again
jnthn lizmat: bs
It's that we can't inline the native ++ yet
lizmat well, there you go 13:19
jnthn So, just not so well optimized
lizmat hence my $i = $i + 1 when using natives atm
jnthn *nod*
13:19 yuppie joined
lizmat once we have it, we'll just apply a macro on the core settings and have a faster $i++ :-) 13:20
ilmari lizmat: that's what made me wonder and test
yuppie i have arrived
jnthn It is implemented for natives, though, and it's not boxing to an Int. It's just the overhead of passing a reference and really doing the call is pretty high.
Compared to the cost of actually incrementing a native integer, which is nearly free
lizmat ilmari: what jnthn said :-)
yuppie is there going to be something like cpan for perl6? 13:21
pmurias pink_mist: nqp can already run in the browser
pink_mist pmurias: 0_o whoah
[TuxCM] yuppie, that is what jdv79 is working on
lizmat pmurias: so the next steps are: 1. making nqp on javascript self-hosting ? 13:22
pmurias yes
jnthn yuppie: modules.perl6.org/ is where the collection of Perl 6 modules so far can be found
lizmat 2. write the necessary interface bits for rakudo
3. profit ?
yuppie such "foo" is "bar" wow
pmurias making nqp on javascript self-hosting is the current step, I need to make the NQP grammar compile (and work correctly) on js 13:23
jnthn pmurias++ # that's an important step towards bootstrapping...nice :)
13:23 huf is now known as plnk_mlst, Ven joined
pmurias and then I can move onto rakudo 13:23
lizmat pmurias++
13:24 plnk_mlst is now known as huf, pink_mist is now known as hnf, hnf is now known as pink_mist, sivoais left
Loren Nick blackcat 13:25
13:25 Loren is now known as blackcat, blackcat is now known as blackcat_ 13:26 blackcat_ is now known as Loren_, sivoais joined 13:27 thou joined 13:29 pecastro left 13:31 Loren_ left 13:32 thou left 13:33 blackcat joined, blackcat is now known as Guest5909 13:34 blackcat_ joined 13:35 sivoais left 13:36 sivoais joined 13:37 mr_ron left 13:38 Guest5909 left 13:41 n0tjack joined 13:42 pecastro joined, andreoss_ joined
andreoss_ m: my &f = [+] 1..*; say f(5) 13:42
camelia rakudo-moar 6c01dc: OUTPUT«(timeout)»
jnthn 1..* on a Range isn't a currying, it's an infinite range
13:44 ely-se left 13:45 chenryn joined, sivoais left
ilmari but why does it hang rather than throwing an exception when assigning it to &f? 13:46
ah, no, it's the reduce that's hanging 13:47
jnthn It's...right
13:47 sivoais joined
grondilu wasn't there a rule that [+] 1..* == -1/3 or something :) 13:48
13:48 pecastro left
grondilu Ramanudjan sum or somthing 13:48
[Coke] that was just bad maths.
13:48 chenryn_ joined
[Coke] if it's the one I'm remembering that went around the youtubes about six months ago. 13:49
n0tjack where should I start if I want to create a meta-operator which consumes (binary) operators as inputs and produces a (binary) operator as an output? 13:50
andreoss_ m: my &f := my $x = { ... }; $x = 1; say f()
camelia rakudo-moar 6c01dc: OUTPUT«Cannot find method 'CALL-ME'␤ in block <unit> at /tmp/F6rlRg9Q_h:1␤␤»
13:50 Ven left, ely-se joined
andreoss_ shouldn't it say Cannot invoke this object instead? 13:50
m: my &f; say f(); 13:51
camelia rakudo-moar 6c01dc: OUTPUT«Cannot invoke this object␤ in block <unit> at /tmp/JM3xVPi79C:1␤␤»
grondilu creating meta-operators is not easy.
13:51 vytas left
n0tjack grondilu: can it be done in-language, or would I have to hack the source? 13:51
grondilu I think you'll have to wait for parsed macros or something. 13:52
n0tjack ok, thanks
(I was secretly hoping I could do something like sub infix:<dot> (&reduce, &map) {[&reduce] @^x >>&map<< @^y} ) 13:54
psch m: sub infix:<XX> ( $lhs, *@ [ $op, $rhs ] ) is looser(&infix:<,>) { say $op($lhs, $rhs) }; 5 XX &infix:<+>, 10 # ehhh.. 13:55
camelia rakudo-moar fd403a: OUTPUT«15␤»
psch n0tjack: that's what's possible
13:55 sivoais left
psch n0tjack: not it's not really the same, and pretty hackish 13:55
s:1st/not/note/
13:56 aborazmeh left
psch m: sub infix:<XX> ( $lhs, *@ [ $op, $rhs ] ) is looser(&infix:<,>) { say $op($lhs, $rhs) }; 5 XX &[+], 10 # ehhh.. # a bit shorter invocation 13:56
n0tjack psch: let me go play with that
camelia rakudo-moar fd403a: OUTPUT«15␤»
13:56 thou joined 13:57 sivoais joined, WizJin joined
n0tjack the slurpy is in case the rhs has a bunch of stuff past the op? 13:57
dalek kudo/nom: 693ca01 | lizmat++ | src/core/Range.pm:
Oops, fix an off-by-one error introduced just now
jdv79 cna anyone review a post about metacpan?
psch n0tjack: the slurpy is probably not neccessary like that, i think. i just couldn't get the easy unpacking to work :) 13:58
s/easy/plain/
n0tjack haha
psch m: sub infix:<XX> ( $lhs, [ $op, $rhs ] ) is looser(&infix:<,>) { say $op($lhs, $rhs) }; 5 XX &[+], 10 # ehhh.. # a bit shorter invocation
camelia rakudo-moar fd403a: OUTPUT«15␤»
lizmat jdv79: when I'm back from cycling
cycling&
psch n0tjack: that also works, dunno what i did wrong before :)
n0tjack thanks, at least I know I can hack my way out if I ever need it
jdv79 sooner anyone? my battery is about to die and i'd like to go check out a blegian themede oktoberfest thing across town for a few hours 13:59
n0tjack I was gonna write sub infix:<linear_combination> (@x, @y) { @x + map_reduce * @y }
PerlJam jdv79: just paste the link and see who bites. :)
n0tjack so then you could say (1,2,3,4) linear_combination (5,6,7,8) to get the sum of their product
jdv79 jdv79.blogspot.com/b/post-preview?...;type=POST 14:00
someone doesn't care about url brevity there
nine jdv79: s/handle, these/handle these/
psch n0tjack: isn't that easier as "[+] @lhs X* @rhs"? 14:01
n0tjack: maybe call that linear_combination(@lhs, @rhs)
[Coke] jdv79: All of thse - *these 14:02
n0tjack psch: that's definitely easier, but I have more uses for map_reduce than just linear_combination
psch oh, no, linear combination wouldn't be X but Z
jdv79 oops
psch m: say &reduce
camelia rakudo-moar fd403a: OUTPUT«sub reduce (&with, + is raw) { #`(Sub|50017240) ... }␤»
n0tjack is there a diff btwn >>foo<< and foo Z?
psch n0tjack: hyper doesn't guarantee order of execution, only of result 14:03
[Coke] jdv79: nifty. good work
jdv79 cool, thanks
PerlJam jdv79: what Coke said.
nine jdv79: s/its been suggested/it's been suggested
PerlJam jdv79++
n0tjack psch: good, that'll teach those recaltrant side-effectors!
jdv79 oh, forgot title
d'oh
n0tjack I'm gonna build me a little Grammar in p6 today 14:04
nine n0tjack: have fun!
n0tjack funny how the first thing I want to do in a brand spankign new language is built an even newer spankier language
TimToady morning from PPW hackathon
14:05 skids joined
[Coke] TimToady: oh shit, that's today. 14:05
jdv79 published. thanks guys!
FROGGS morning TimToady
jnthn o/ TimToady
n0tjack can I rely on the Grammar stuff in the butterfly book to be up to date?
[Coke] I had hoped to go to the PPW. dammit.
psch m: sub map-reduce(&mapsub, &redsub, *@args) { redsub mapsub @args }; say map-reduce &[*], &[+], ((1,2,3), (1,2,3))
camelia rakudo-moar fd403a: OUTPUT«6␤»
[Coke] asks, "what year is it?" "WHAT YEAR IS IT!?"
psch huh
14:06 sivoais left
PerlJam jdv79: if that metacpan instance on hack is only for P6, you might want to change the "No result" message to not mention Task::Kensho 14:06
psch m: sub map-reduce(&mapsub, &redsub, *@args) { [[&redsub]] [[&mapsub]] @args }; say map-reduce &[*], &[+], ((1,2,3), (1,2,3)) # there 14:07
camelia rakudo-moar fd403a: OUTPUT«36␤»
14:07 sivoais joined
nine jdv79: I'm a bit sad that we're gonna have two distinct metacpan sites. Will it be possible to merge them later on? 14:07
n0tjack psch++
psch n0tjack: maybe make &mapsub and &redsub named, if that fits your api. i'd say redsub probably also could have &infix:<,> as default value... 14:08
14:09 chrstphrhrt joined
psch eh, or just call (map { ... }, @args).reduce(:with({...})) i guess... :) 14:09
n0tjack psch: that seems like a natural choice. in APL, map-reduce is named dot, and its default (when passed jot) is append
[Coke] I am fine with them being separate. Each is going to care about different aspects of their modules. 14:10
n0tjack alright, I'm gonna take a risk and follow Using Perl6 for the learnins 14:11
jdv79 nine: how would you do that ?
n0tjack you guys better not have made any progress since it was published in 2012 *shakes fist
14:12 ely-se left
jdv79 a lot of things refer to a dist by name and such 14:12
all that would ahve to be bifurcated ^H
14:12 rurban left
jdv79 that's more work than i' prepare to do right now. maybe later. but what's the benefit anyway? is it worth it? 14:12
PerlJam "one metacpan to rule them all" seems to be the main benefit. 14:13
ShimmerFairy I'm also fine with two separate sites; they are two distinct languages, after all :) (in the same way wikipedia's subdomain-based division of languages is still the nicest solution for i18n in wikis I've seen)
jdv79 PerlJam: :)
let's just get something up and then we'll talk about making it harder:)
lunch & 14:14
nine jdv79: well, maybe separate sites is not that bad at all. I'd just love to see it suggest Perl 5 modules with an automatically added Inline::Perl5 dependency in addition to the Perl 6 modules. That could fight the preconception that Perl 6 has a module problem.
jdv79: of course. That's why I asked about possibilities of the future :)
PerlJam nine: that /would/ be nice
14:15 sivoais left, andreoss_ left
ShimmerFairy nine: Being the purely-P6 user that I am, I'm not quite comfortable with the idea of so readily suggesting Perl 5 modules. That may well be just me though :P . 14:16
FROGGS I also think that these two should be separate
otherwise searching for dists can be quite confusing
jnthn ShimmerFairy: Many people are just interested in solving their problem. :) 14:17
14:17 sivoais joined
PerlJam jnthn++ I was just about to say exactly that 14:17
14:17 antiatom left 14:18 antiatom joined, Kogurr joined
PerlJam (though, taken reductio ad absurdum, why shouldn't it also suggest Python modules or Lua modules too?) 14:18
ShimmerFairy jnthn: of course, that's just a personal thing, I can't help but feel like it kinda says "oh hey we don't have that why don't you try Perl 5 instead why bother with P6 anyway!?"
14:19 chenryn left, chenryn_ is now known as chenryn
PerlJam ShimmerFairy: I've almost always said TMTOWTDI includes ways that do not involve Perl . That's part of the "Perl Spirit" as far as I'm concerned. 14:19
concerned yeah, I'm concerned too 14:20
lol
nine PerlJam: because we do have easy access to Perl 5 modules and I don't know of a real centralized Python module database
PerlJam concerned: :P
jnthn ShimmerFairy: Not really. If someone is already wanting to write in P6 (and so are looking at the P6 modules site), and then finds they have most of what they need and one missing thing, and we tell them "oh hey, you can get at it through Inline::Perl5", they're better equipped to use Perl 6.
PerlJam nine: an excellent point :)
jnthn ShimmerFairy: Rather than say "oh...it doesn't have something I need, I'd better use something else for the whole thing" 14:21
nine PerlJam: I arrived at the same question a couple of minutes ago :)
ShimmerFairy I'd prefer if the response was "I'd better go write it!", but I see the point.
PerlJam ShimmerFairy: see "some people just want to get their work done" :)
ShimmerFairy I guess I'm just worried that this early on, being too quick to suggest FFIs would stagnate the development of P6 modules. 14:22
14:23 Erendis42 left, laouji left, cognominal joined 14:25 sivoais left
PerlJam ShimmerFairy: maybe. 14:25
pmurias PerlJam: once we get :from<python> fully working having metacpan6 suggest python modules too seems like a good idea
PerlJam ShimmerFairy: but maybe it also gives people time to figure out how the P5 module isn't well suited for P6 and then write a *better* P6 module. 14:26
TimToady would be funny if the best python infrastructure was written in p6
14:26 cognominal_ joined 14:27 sivoais joined
ShimmerFairy So I guess the reason why I'm not comfortable with the idea (even though I don't think it's bad) is because it might lead too many people into feeling like P6 is in fact some kind of successor to P5, 'cos look at all these P5 modules I can still use. 14:27
14:27 WizJin left
nine ShimmerFairy: I could live with people believing that if they also believe it's Python's successor :) 14:28
PerlJam heh
14:28 cognominal left
PerlJam ShimmerFairy: I think we have a good story there because "sister languages" 14:28
TimToady but the fact remains that the younger sister has the longer life expectancy :) 14:29
nine ShimmerFairy: also, people can use Perl 6 modules in Perl 5. And I could live very well with P5's metacpan suggesting Perl 6 modules as well. Bridges go both ways after all. 14:30
14:31 pippo left
n0tjack given how long it took her to get from conception to birth, I'd say p6 has a long life ahead of her indeed 14:31
ShimmerFairy nine: hm. I like the sound of it more if it went both ways :) . I think my concerns boils down to "no wait, you just got here!" :)
n0tjack I think p6 years are like inverse dog years
14:32 tokuhirom left
PerlJam I don't think there's a correlation between gestation period and life expectancy 14:32
14:32 Ven joined 14:34 s_kilk joined 14:35 sivoais left 14:36 ely-se joined
n0tjack PerlJam: oregonstate.edu/instruct/st352/koll...tation.htm 14:36
14:37 sivoais joined, WizJin joined 14:38 llfourn joined
PerlJam n0tjack: are you suggesting that I Do The Math? :) 14:41
TimToady decided that the user-visible method to do triangular reduce is likely gonna be called "produce", to avoid collision with actual triangles in any geometry package, and because "triangle" is not a verb 14:42
and the "triangle" mnemonic is just for the [\op] form
14:42 llfourn left
n0tjack PerlJam: nah, do what everyone else does, scroll to the bottom line: " The interpretation: we are 95% confident that the median gestation period will be between 1.05 to 1.12 times longer for every increase of one year in life expectancy." 14:43
jnthn TimToady: triangulate? ;)
PerlJam TimToady: triduce (make up your own words :) 14:44
ilmari it would be nice if panda did $*PROGRAM.basename in its usage message
jnthn We could do with a method on Supply that maps to Rx's .scan
Which is basically a triangle reduce
14:45 sivoais left
PerlJam (though, "triduce" could be confused with "traduce" which could cause people to get the wrong impression about things) 14:45
14:45 reneeb joined
ilmari I don't need to know that it's /home/ilmari/.rakudobrew/bin/../moar-nom/install/share/perl6/site/bin/panda 14:45
PerlJam ilmari: isn't that an issue in the panda repo already?
ilmari PerlJam: possibly, I didn't look
nope 14:46
14:46 g5 left
PerlJam huh. Well, it's come up before so I guess I thought it had been "issued" 14:47
TimToady I already thought about all those possibilities before settling on "produce"
ilmari creates an issue
n0tjack in other contexts, tri-duce is known as "running"
"running sum"
"running product"
jnthn TimToady: Will that exist as a method form too, like .reduce does?
14:47 sivoais joined
TimToady hmm, didn't think about "running" 14:48
n0tjack in J, you spell it +/\ and it does have a little triangle. Which I always picture as pointing to the spaces where the op will be inserted
TimToady it's not a verb though
pink_mist 0_o
TimToady and it's a little confusing with run()
jnthn TimToady: Also known as a cumulative product I guess, so accumulate or so may work 14:49
n0tjack "running" isn't a verb in "running sum", either
it's modifying sum
TimToady accumulators don't usually advertise their intermediate products
jnthn uh, cumulative sum
Yeah, I just realized that after I wrote it :)
moritz cummulous cloud sum 14:50
dalek p: fc8dffe | (Pawel Murias)++ | t/nqp/59-nqpop.t:
Add tests for nqp::isge_s, nqp::isle_s.
p: 5ba3ba3 | (Pawel Murias)++ | / (3 files):
[js] Implement nqp::is{gt,ge,le,lt}_s
TimToady anyway, if we pick a -duce word, produce is the best one
14:50 [Sno] joined
TimToady idea of a production line is there to 14:50
ilmari PerlJam: oh, that's perl's usage thing, not something in panda itself 14:51
jnthn It just makes me think a bit of producer/consumer a bit much I guess.
PerlJam works as far as "growing things" is concerned too
TimToady well, it is a producer :)
PerlJam ilmari: oh.
jnthn True, but many other things also are :)
concerned growing things isn't concerned, I'm concerned!
jnthn I've nothing against produce, just wonder if there's something more vivid... 14:52
PerlJam concerned: if you keep this up, I may never use that word again :)
TimToady well, traduce is more vivid :)
concerned looks like you're in a PerlJam
epic nickname puns
jnthn .Scan from Rx didn't really tell me what it did and I had to go read the docs... :)
PerlJam or ... maybe I'll just /ignore ;) 14:53
concerned sorry I derailed the chat
ilmari PerlJam: ah, it's because the rakudobrew shim invokes it by full path
_and_ that path has a .. in it
14:54 concerned left
ab6tract triduce is kinda cool, tbh :) 14:54
ilmari fpaste.scsys.co.uk/500127 14:55
ab6tract is there an op form of it as well?
jnthn [\+]
m: say [\+] 1..10 14:56
camelia rakudo-moar 693ca0: OUTPUT«(1 3 6 10 15 21 28 36 45 55)␤»
14:56 sivoais left
TimToady also considered treduce and induce, but produce seemed better 14:57
FROGGS m: use variables :D; my $a 14:58
camelia rakudo-moar 693ca0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/b6Gjfk0cx7␤Variable definition of type (with implicit :D) requires an initializer␤at /tmp/b6Gjfk0cx7:1␤------> 3use variables :D; my $a7⏏5<EOL>␤ expecting any of:␤ constraint␤»
moritz preduce
the best of produce and reduce
TimToady rreduce
pink_mist triangle-produce?
TimToady for running reduce
14:58 sivoais joined
n0tjack partial products? 14:59
pink_mist sees where triduce comes from ... how about proangle? :P
PerlJam "produce" looks better and better (though, I'd still favor "triduce" :) 15:00
15:00 [TuxCM] left 15:02 [Sno] left
n0tjack is there a built-in for sliding windows of size N? 15:03
15:03 [Sno] joined
PerlJam n0tjack: rotor 15:03
n0tjack nice!
ilmari ah, canonpath doesn't strip .. segments
FROGGS ilmari: try .resolve 15:04
n0tjack we need a perlfunc
PerlJam n0tjack: doc.perl6.org/routine/rotor
n0tjack too much work to go to perl6.org every time
PerlJam agreed
thowe do named arguments like this ":35thing" only work with integers or whatever? I mean, how would you use a string as an arg there? 15:05
PerlJam thowe: :thing<value> 15:06
TimToady only integers
thowe OK
n0tjack m: :35thing 15:07
camelia ( no output )
n0tjack m: say :35thing
camelia rakudo-moar 693ca0: OUTPUT«Unexpected named parameter 'thing' passed␤ in block <unit> at /tmp/fNS4g90x_c:1␤␤»
n0tjack hmm, thought that just create a Pair
m: say :!f;
camelia rakudo-moar 693ca0: OUTPUT«Unexpected named parameter 'f' passed␤ in block <unit> at /tmp/3Ni79PgFZm:1␤␤»
PerlJam aye, it does
psch m: say (:35thing)
camelia rakudo-moar 693ca0: OUTPUT«thing => 35␤»
n0tjack parens are extra magical in p6 15:08
psch m: :35things.say
camelia rakudo-moar 693ca0: OUTPUT«things => 35␤»
psch n0tjack: no, bare Pairs are named arguments
n0tjack m: say Pair("foo","bar");
camelia rakudo-moar 693ca0: OUTPUT«Cannot find method 'Pair'␤ in block <unit> at /tmp/eTH7WZt1ly:1␤␤»
n0tjack m: say Pair.new("foo","bar");
camelia rakudo-moar 693ca0: OUTPUT«foo => bar␤»
n0tjack why should that be different 15:09
thowe this must be wrong... :name => <arg>
15:09 sivoais left, diana_olhovik left
PerlJam thowe: eh? 15:09
psch m: say (:name => <arg>) 15:10
camelia rakudo-moar 693ca0: OUTPUT«(name => True) => arg␤»
15:10 sivoais joined
psch n0tjack: "bare" was probably a bad choice of words. literal colonpairs are parsed as named arguments, as are fatarrow pairs with implicitely quoted key 15:10
m: say foo => bar
camelia rakudo-moar 693ca0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/1AMsaMjm1d␤Undeclared routine:␤ bar used at line 1. Did you mean 'bag'?␤␤»
psch m: say foo => "bar"
camelia rakudo-moar 693ca0: OUTPUT«Unexpected named parameter 'foo' passed␤ in block <unit> at /tmp/HFpg5d40pR:1␤␤»
psch m: say "foo" => "bar"
camelia rakudo-moar 693ca0: OUTPUT«foo => bar␤»
n0tjack psch: ah, ok. makes sense. 15:11
synatically, how is the convenient sugar :$named_param ~~ :$named_param($named_param) justified? 15:12
uh, I mean :named_param($named_param) 15:13
thowe I'm not quite sure what I've gotten when I've gotten "(name => True) => arg"
n0tjack thowe: you've got a Pair whose key is a Pair
psch n0tjack: i always understood it as "mirroring the signature for easy invocation"
n0tjack psch: Absolutely, and I have already made great use of it. But I'm not sure how to explain that grammatically. 15:14
psch sub f(:name($name)) { } => sub f(:$name) { }
thowe n0tjack, so if that was given calling a function, how would I get "arg"?
FROGGS DRY
psch but that's just moving the goalpost, i realize
n0tjack thowe: I'm new to p6 myself, but I'd expect if you had that object as a scalar named $blah, you'd say something like $blah.value 15:15
psch m: sub f(:$name) { $name.perl.say }; f :name => <arg> 15:16
camelia rakudo-moar 693ca0: OUTPUT«Too many positionals passed; expected 0 arguments but got 1␤ in sub f at /tmp/jmJKprK9YS:1␤ in block <unit> at /tmp/jmJKprK9YS:1␤␤»
n0tjack psch: I ask because the lights went on for me yesterday when someone told me :blah is just sugar for creating a pair
pmurias jnthn: what should the nqp::for op evaluate to?
jnthn: the comments in the implementation imply the input list
psch n0tjack: i guess FROGGS++ response is pretty good. DRY is an established principle 15:17
thowe n0tjack, but the intention was to have a param named "name". Do I need to get that from the Collection or something?
jnthn pmurias: That's OK I guess. It certainly isn't expected to produce a new list
pmurias: And null is maybe a bit harsh
n0tjack psch: No, I get the "why" of it. I'm asking about the "how" of it.
pmurias jnthn: but when I run it's the last element
jnthn hm
psch thowe: you're building the Pair wrongly. :name<arg> or name => <arg> is what you're looking for
PerlJam perhaps :name => <foo> should have some sort of warning
jnthn pmurias: I'm pretty sure we don't actually rely on what it evaluates to anywhere
15:18 reneeb left
psch n0tjack: i'm not sure i follow. implementation? 15:18
thowe psch, yeah, I started by saying "this must be wrong". But if it was given, how would it be accessed? Can it?
jnthn pmurias: I can believe it's "last element" too, but...it's not too useful :)
pmurias jnthn: paste.debian.net/315122
n0tjack psch: the "how" of :stuff(42) is the colon creates a Pair
15:18 sivoais left
n0tjack psch: I'm looking for similar insight into :$name 15:18
FROGGS n0tjack: if the thing after the colon (which is a Pair of dots btw) looks like a variable, we take its name (without sigil/twigel) and its value to produce a Pair 15:19
jnthn wonders if nqp-j does that too
n0tjack psch: before I knew the colon created a Pair, the syntax looked like a bolted-on special case; after the explanation, I can see the elegance much more clearly
FROGGS twigil*
jnthn pmurias: If it's not a pain you could do similar. Alterntively the thing we iterated over may also be fine, but then we have to keep that around, which we needn't do
thowe anyway... thanks. gotta go to work...
15:19 thowe left
n0tjack FROGGS: Yeah, I guess it is just that straightforward. I thought maybe there was some hidden justification. 15:19
jnthn pmurias: Which is least annoying from a JS perspective, ooc? 15:20
pmurias: Most of the time the thing is in a void context anyway
15:20 sivoais joined
pmurias jnthn: the least about of pain would be a null 15:20
jnthn: on the jvm it's an arrayiterator 15:21
jnthn pmurias: OK, if it's already ont consistent then...just do what's the least pain 15:22
15:22 Dom__ joined
pmurias jnthn: is there a reason why we seem to have (none) very little ops that don't return anything? 15:23
jnthn pmurias: Largely a reflection of there being very little in Perl 6 that doesn't return something 15:24
pmurias: For bindattr, which could be a candidate, for example, it evaluating to the bound value makes $a := $!b := 42; easy to code-gen 15:25
15:26 momoko joined
psch .tell thowe you can access it with .value for one: my $PoP = :name => <arg>; say $PoP.value 15:27
yoleaux psch: I'll pass your message to thowe.
15:27 momoko is now known as Guest3379
Guest3379 m: say 10, 9, 8, { $_ - 1 || last } ... *; 15:27
camelia rakudo-moar 693ca0: OUTPUT«(...)␤»
TimToady pmurias: could have something to do with Perl 6 being a functional programming language
15:28 Dom__ left
Guest3379 hi perl6, is this output correct? 15:28
pmurias TimToady: it's even with things such as nqp::setscdesc
15:28 sftp joined, ely-se left
psch m: say eager 10, 9, 8, { $_ - 1 || last } ... *; 15:29
camelia rakudo-moar 693ca0: OUTPUT«(10 9 8 7 6 5 4 3 2 1)␤»
grondilu Guest3379: yes, it's correct. The list is not evaluated by a say
(not eagerly evaluated that is)
15:29 sivoais left
Guest3379 thanks! 15:29
15:30 sivoais joined
psch r: sub circumfix:["@", "@"] ($a) { say $a }; @ 5 @ # o.o 15:31
15:31 s_kilk left
camelia rakudo-moar 693ca0: OUTPUT«5␤» 15:31
..rakudo-jvm 693ca0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/tmpfile␤Two terms in a row␤at /tmp/tmpfile:1␤------> 3 circumfix:["@", "@"] ($a) { say $a }; @7⏏5 5 @ # o.o␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ …»
psch jvm breakage reached a new WAT vOv
jnthn :( 15:32
psch jnthn: could that be related to the literal LTM patch?
15:32 tokuhirom joined
jnthn psch: My one that made constants interpolate? 15:32
psch: No, that was a Perl 6 compiler patch for Perl 6 constant decls, while the code that adds custom ops is all in NQP 15:33
psch jnthn: oh, i guess i lacked contextual information. interpolated constants shouldn't do anything there
15:33 blackcat_ left, Guest3379 left
psch r: sub prefix:<$> { $^a.say }; my $a = "5"; $$a 15:35
camelia ( no output )
psch r: sub prefix:<$> { $^a.say }; my $a = "5"; $ $a
camelia rakudo-{moar,jvm} 693ca0: OUTPUT«5===SORRY!5=== Error while compiling /tmp/tmpfile␤Undeclared routine:␤ a used at line 1␤␤»
psch uh, yeah DIHWIDT :p 15:36
although i remember something along those lines that worked some time back...
n0tjack holy crap
m: say 'h̟̖̻̻e􏿽xCD􏿽xA1􏿽xCD􏿽x87􏿽xCC􏿽x9C􏿽xCC􏿽xA9􏿽xCC􏿽xB1􏿽xCC􏿽xB3􏿽xCC􏿽x9F􏿽xCC􏿽xACl̘l̠̩͚o̰̘̮̬̹̪'.chars;
camelia rakudo-moar 693ca0: OUTPUT«5␤»
n0tjack that's amazing
15:36 tokuhirom left
n0tjack "making unicode go away" is like a top 5 reason to switch to p6 15:37
m: say uc 'h̟̖̻̻e􏿽xCD􏿽xA1􏿽xCD􏿽x87􏿽xCC􏿽x9C􏿽xCC􏿽xA9􏿽xCC􏿽xB1􏿽xCC􏿽xB3􏿽xCC􏿽x9F􏿽xCC􏿽xACl̘l̠̩͚o̰̘̮̬̹̪';
camelia rakudo-moar 693ca0: OUTPUT«H̟̖̻̻E͇̜̩̱̳̟̬͡L̘L̠̩͚O̰̘̮̬̹̪␤»
jnthn I'm glad Perl 6 is less confused about that string than my terminal... :)
psch s/go away/silently work reasonably/
n0tjack hahahaha
psch m: say fc 'h̟̖̻̻e􏿽xCD􏿽xA1􏿽xCD􏿽x87􏿽xCC􏿽x9C􏿽xCC􏿽xA9􏿽xCC􏿽xB1􏿽xCC􏿽xB3􏿽xCC􏿽x9F􏿽xCC􏿽xACl̘l̠̩͚o̰̘̮̬̹̪'; 15:38
camelia rakudo-moar 693ca0: OUTPUT«(signal ABRT)»
psch ooohh
jnthn ABRT is...a bit unfortunate
jnthn only added fc yesterday, and didn't do the NFG handling for it yet
But I thought it should hit an NYI exception 15:39
n0tjack m: fc 'hello';
camelia ( no output )
15:39 sivoais left
psch yeah, i was thinking it was probably still on a specific TODO list somewhere... 15:39
n0tjack m: say fc 'hello';
camelia rakudo-moar 693ca0: OUTPUT«hello␤»
psch m: say fc 'hÄllö'
camelia rakudo-moar 693ca0: OUTPUT«hällö␤»
Ven jnthn: wow, nodejs gave me 27 15:40
15:41 sivoais joined
Ven ruby same. I guess most will 15:41
n0tjack Ven: like I said, top 5 reason.
anyone know why the REPL doesn't grok arrow keys / line recall on my Mac? 15:42
15:43 RabidGravy joined
psch n0tjack: you need to install LineNoise with panda 15:43
n0tjack thanks 15:44
jnthn Ven: Wow!
timotimo n0tjack: wit a lowercase n i believe 15:45
psch hmm, anyone have an idea where i should BP for implicit return on jvm..? :)
r: sub f { sub g { 42 } }; f()() # had to pin this NPE down
*hard 15:46
camelia ( no output )
..rakudo-jvm 693ca0: OUTPUT«java.lang.NullPointerException␤ in block <unit> at /tmp/tmpfile:1␤␤»
dalek kudo/nom: 7b54492 | TimToady++ | src/core/Any-iterable-methods.pm:
Add produce method to do triangle reduce
15:46 domidumont left
dalek ast: d93ef20 | TimToady++ | S32-list/produce.t:
add tests for produce
 15:47
psch oh, there was advice already re: that bug
something about compile_all_the_stmts iirc
15:48 patrickz joined
dalek kudo/nom: d05cb80 | TimToady++ | t/spectest.data:
add S32-list/produce.t
 15:48
psch right, i got pretty much nowhere trying to compare the moar logic to jvm for void handling
15:49 sivoais left
n0tjack oh god that's so much better; psch++ 15:50
15:51 sivoais joined 15:52 laouji joined
psch n0tjack: iirc you also get tab completion for CORE functions with linenoise 15:53
and methods too
15:53 xfix joined, xfix left, xfix joined
ab6tract is windows still the blocker on getting Linenoise distributed with rakudo? 15:53
n0tjack I have vim setup with p6 syntax highlighting, but haven't done much more than that for tab completion etc 15:54
15:54 FROGGS left 15:55 xfix left, xfix joined
timotimo we have ctags for perl6, but i'm not sure that gives you tab completion 15:55
but at least the perl6 vim thing sets up - and ' as valid inside-identifier-characters 15:56
psch was tab completion in the REPL pulled from linenoise?
15:57 Kogurr left
timotimo in the repl! 15:57
i thought n0tjack was talking about vim
yeah, linenoise can tab-complete
15:57 Ven left
psch yeah, the sudden talk of vim confused me as well, tbh 15:57
15:59 patrickz left, sivoais left
tony-o :wq 16:00
timotimo memfrobs tony-o's buffer
16:00 patrickz joined
tony-o sounds hot 16:00
16:01 sivoais joined 16:02 patrickz left
n0tjack it was a train of thought where you guys only saw the caboose (linenoise -> tab completion -> need to get that set up in vim too) 16:06
tony-o doesn't ctrl+r or something in vim do something similar? 16:07
ctrl+n
n0tjack: this looks cool too github.com/Valloric/YouCompleteMe 16:08
timotimo yeah, ctrl-n and ctrl-p is the opposite direction
vim has like 8 different kinds of completion 16:09
tony-o ah
16:10 sivoais left, FROGGS joined
pmurias compiler/editor integration for Perl 6 would be awesome 16:10
eiro a vi clone written in perl6 would be awesome 16:11
hello everyone
16:11 sivoais joined
tony-o lol. 16:11
timotimo first build something that makes terminal emulators less finnicky
pmurias convincing people to switch editors is really hard 16:14
timotimo yeah, because the only good editor is vim (or vim-like editors) 16:16
so switching editors means switching from a vim to a not-vim :p
psch is reminded of the "A Whole New World" talk, re: finnicky terminal emulators
pmurias for some weird reason getting people to switch to vim is also hard
16:16 zakharyas left
timotimo haha 16:16
yeah, because then they're switching away from notepad, which does exactly what they expect
press a button, get a character 16:17
psch 'what do you mean ci" is not intuitive?'
tony-o hah 16:18
psch actually usually uses F"ct" instead of ci", though
timotimo oh?
why is that?
psch muscle memory, mostly
timotimo ci" will also search for a " to thel eft or right
psch well, it's probably T"ct" 16:19
yeah, ci" is probably the better sentence, i'm just not used to it :)
timotimo :)
ok, afk for a bit
16:20 sivoais left 16:22 Celelibi_ is now known as Celelibi, sivoais joined
n0tjack tony-o: that does look cool 16:23
16:28 Ven joined
FROGGS jnthn / lizmat: the MOPy stuff passes all tests (even some of the new TODOs), but fails the exploding :D attribute tests (which were under discussion anyway) 16:29
16:30 sivoais left
FROGGS jnthn: do you also think that 'has Any:D $.foo' should not explode until object instanciation time? so that :D kinda implies 'is required' ? 16:30
16:31 antiatom left, ^elyse^ joined
jnthn FROGGS: We could make it imply "is required" for attrs I guess... 16:32
16:32 sivoais joined
FROGGS k 16:32
jnthn FROGGS: For lexicals we probably should continue as we already are and demand an initializer
FROGGS: Making it imply "is required" feels a little dirty but people didn't seem to like being forced to provide a default... 16:33
FROGGS jnthn: yes, when there is no initializer, I just check if the default value matches the bind constraint
jnthn +1
Oh
Actually
FROGGS to allow:
m: my Int:D $a is default(0)
camelia rakudo-moar d05cb8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/mUAf92dtdZ␤Default value '0' will never bind to a parameter of type Int:D (with implicit :D)␤at /tmp/mUAf92dtdZ:1␤------> 3my Int:D $a is default(0)7⏏5<EOL>␤ expecting any of:␤ constra…»
jnthn Maybe don't make it imply "is required", just check if it's marked "is default" or "is required"
(for attributes, I mean) 16:34
FROGGS hmm? I dont get that
jnthn has Int:D $.x is default(1); # ok
has Int:D $.x is required; # ok
has Int:D $.x = 1; # ok
has Int:D $.x; # error like now
16:35 laouji left
FROGGS well, that's almost what I got now 16:35
jnthn What bit is different? The last one automatically adds "is required"?
FROGGS no, I planned to make it so, but it is not yet in 16:36
jnthn is a tiny bit reluctant
FROGGS I check for the default trait, but not the required trait
jnthn Ah
Let's maybe try adding the check for required (which also makes it OK)
FROGGS well, I check for the default trait indirectly, by checking the value after running the trait
jnthn ah 16:37
Yeah, it's going to be a special case for attributes whatever we do
FROGGS yeah
I was also not very fond of making :D meaning 'is required later'
because that implies that you always want defined values ever, which is not true 16:38
especially because we have that many undefined things
16:38 pmurias left
n0tjack in a regex, is * only the "whatever star" when it's used in a range (i.e. adjacent to a ..)? 16:38
psch grrr object.asm 16:39
java.lang.RuntimeException: java.lang.ArrayIndexOutOfBoundsException: 0
n0tjack the example a ** 3..* threw me
psch that's not helpful for why codegen fails!
n0tjack: right, * as a quantifier is still "zero or more times", but on the rhs of a range it's same as in Perl 6 itself 16:40
16:40 sivoais left
psch n0tjack: ** is the new way of writing {m,n}, where m,n turns into m..n as rhs to ** 16:40
(given you have perl5 background and that's a useful translation) 16:41
n0tjack yes, it is, thank you
16:41 Vivek joined 16:42 Vivek is now known as Guest9682
psch m: say "aaaa" ~~ m:g/(a ** 1..*)/ 16:42
camelia rakudo-moar d05cb8: OUTPUT«(｢aaaa｣␤ 0 => ｢aaaa｣)␤»
psch m: say "aaaa" ~~ m:g/(a) ** 1..*/
camelia rakudo-moar d05cb8: OUTPUT«(｢aaaa｣␤ 0 => ｢a｣␤ 0 => ｢a｣␤ 0 => ｢a｣␤ 0 => ｢a｣)␤»
16:42 sivoais joined
psch oh 16:42
m: say "aaaa" ~~ m:o/(a) ** 1..*/
camelia rakudo-moar d05cb8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/fmoMXwldCs␤Adverb o not allowed on m␤at /tmp/fmoMXwldCs:1␤------> 3say "aaaa" ~~ m:o/(a) ** 1..*/7⏏5<EOL>␤»
n0tjack makes sense, but my brain is *insisting* that that's got a /.*/ in it
psch m: say "aaaa" ~~ m:ov/(a) ** 1..*/
camelia rakudo-moar d05cb8: OUTPUT«(｢aaaa｣␤ 0 => ｢a｣␤ 0 => ｢a｣␤ 0 => ｢a｣␤ 0 => ｢a｣ ｢aaa｣␤ 0 => ｢a｣␤ 0 => ｢a｣␤ 0 => ｢a｣ ｢aa｣␤ 0 => ｢a｣␤ 0 => ｢a｣ ｢a｣␤ 0 => ｢a｣)␤»
psch m: say "aaaa" ~~ m:ov/(a ** 1..*)/ 16:43
camelia rakudo-moar d05cb8: OUTPUT«(｢aaaa｣␤ 0 => ｢aaaa｣ ｢aaa｣␤ 0 => ｢aaa｣ ｢aa｣␤ 0 => ｢aa｣ ｢a｣␤ 0 => ｢a｣)␤»
psch *that's* what i wanted... :)
Guest9682 hi
Ven \o
psch hi Guest9682
Guest9682 when is general release of perl6? 16:44
PerlJam Guest9682: every month 16:45
Ven Guest9682: 6.0 is planned for christmas
but fwiw, I've been using rakudo in production for more than a year now
n0tjack Ven: what does your prod system do 16:46
Ven n0tjack: tons of stuff. it used to just process a few files together when it got ping'd
n0tjack I guess what I'm asking is, how bad would it hurt if it went down
Guest9682 ok thanks
Ven at that $internship, I don't know what it does now, since my internship ended several months ago... but I expect it's still there running in the background :P 16:47
n0tjack haha
FROGGS huh, checking for required attributes (or even setting it) seems to be a one-line patch
jnthn Guest9682: To clarify the conflicting info: the Perl 6.0 language definition and a comforming implementation are planend for Christmas, but as we converge on it we make compiler releases every month.
(And will continue to make them every month after it, to quickly push out perf improvements, fixes, etc. How fast distributions choose to follow that is up to them.) 16:48
Guest9682 I wanted to learn a scripting language, which should I choose, Python 3 or Perl 6?
[Coke] FROGGS: required attributes already work, you should be able to just set whatever "is required" is setting there.
n0tjack Guest9682: check the titleof the room ;)
PerlJam Guest9682: you realize your audience will be biased? 16:49
[Coke] Guest9682: well, we're going to tell you #perl6, of course.
n0tjack Guest9682: more seriously, perl6 is newer and has a ton of really cool, never seen before features
16:49 krunen_ left
n0tjack if you have any in programming languages qua languages, perl6 is the way to go 16:50
16:50 sivoais left
n0tjack *any interest 16:50
Juerd Guest9682: What, in your opinion, is a "scripting" language?
I'm not sure Perl 6 qualifies
[Coke] Juerd: I would say it does, sure. 16:51
n0tjack imo, a "scripting language" is "one which doesn't get in your way", and p6 definitely qualifies for that
Juerd [Coke]: Depends on one's definition I think.
[Coke] but the whole scripting vs. not classification doesn't mean as much these days.
Juerd [Coke]: I've seen "scripting language" defined as a language that's used to customize or automate parts of a larger program 16:52
16:52 sivoais joined
Juerd Like how Javascript is supposed to extend the browser's functionality, or how Perl can be used to automate irssi, or the difference between VB and VBA. 16:53
psch and like C compiled to dlls for overriding directx hooks to change a games functionality..? 16:54
Juerd As far as I know, Perl 6 is not (yet) used write scripts for programs in, but the code written in Perl 6 is typically a program by itself or a module.
psch: Heh, possibly.
Guest9682 :) thanks. Everyday I read news related to perl6 and I am happy that it is finally releasing this christmas.
dalek ast: 2f6fd89 | jnthn++ | S15-nfg/case-change.t:
A bunch more tests covering NFG casing issues.

Hopefully correct, and covering various edge cases; there's no doubt some further ones.		 16:55
Juerd I wouldn't write CGI scripts in Perl 6 either, because of the startup overhead. But same goes for Perl 5 with Moose :P
16:56 paulmr joined
PerlJam Juerd: no one writes CGI scripts these days anyway ;) 16:58
Juerd PerlJam: Actually, I do. And I've been asked to extend existing ones. 16:59
In fact, I've been hacking on a 24 ksloc CGI script yesterday. One file :(
16:59 weihan left 17:00 gonz_ left
FROGGS [Coke]: exactly 17:00
17:01 firstdayonthejob joined, sivoais left, Guest9682 left 17:03 sivoais joined, Ven left
[Coke] Juerd: for me scripting was always as opposed to compiled. e.g. C is compiled, Tcl is scripting, and things were on the spectrum between. These days, it doesn't mean so much to me. maybe your meaning is the new one in the wild, I hadn't heard that one. 17:03
Juerd I thought that was interpreted vs compiled 17:04
psch mostly agrees with [Coke]
the distinction is in "what do i distribute" for me
if it's a script it's in a scripting language
Juerd Are dynamic, interpreted, scripting the same thing now? :D
psch if it's a binary it's a compiled language
but that's not a useful distinction because exceptions...
so i'm not using the distinction 17:05
Juerd So if you distribute C source code for a program, C's a scripting language?
psch Juerd: that's one of the exceptions that makes the distinction useless, in my eyes :)
n0tjack I stand by "if the language gets out of my way, it's a scripting language" 17:06
a good test is the existence of a (useful) REPL
psch fwiw, wikipedia does have an article on "Scripting language"
17:06 larion left
psch n0tjack: java has at least one pretty good REPL, nowadays... :) 17:06
n0tjack psch: and how much can you express in one line of Java? ;) 17:07
psch bash.org/?946461
scnr
n0tjack hahaha
skids The other layer of "scripting" other than compilation is that "scripts" make other components from "external" software do things. 17:08
n0tjack I've written very many "scripts" in my life, almost none of them invoked external programs 17:09
I do a lot of statistical analysis
17:09 ab6tract left
skids Not necessarily an invoke, just not a whole-cloth "built-here" thing. 17:09
In that respect nci, use :from<> are key to "scripting" capabilities. 17:10
Juerd On startup overhead: 17:11
1;0 juerd@cxie:~$ strace perl -E'say "Hello, world!"' 2>&1 | wc -l
206
1;0 juerd@cxie:~$ strace perl6 -e'say "Hello, world!"' 2>&1 | wc -l
994
17:11 sivoais left
Juerd That's a huge amount of system calls. Haven't looked at them yet. 17:11
17:14 sivoais joined
pink_mist huh, I get 294 for ther perl5 one 17:14
*the
psch throws 216 in as another number 17:15
pink_mist strace perl -e'print "Hello, world!\n"' 2>&1 | wc -l <-- this version gives me 278 17:16
n0tjack I just got the *weirdest* error message 17:17
then I realized I typed perl -e, instead of perl6 -e ...
dalek ast: 7d0f82e | jnthn++ | S15-nfg/case-change.t:
Correct NFD; those things decompose.
17:22 coetry joined 17:23 ovidnis left, sivoais left
coetry Hello folks, i installed rakudo-start via homebrew on osx, and it came coupled with panda, but everytime I try to install a panda module, I get an error similar to this:gist.github.com/coetry/2e05d107115015aad5ca 17:24
I also tried going on panda's github page and installing it manually, but when i run `perl6 boostrap.pl`, I get a compilation error
psch coetry: what's the output of perl6 --version? 17:25
17:25 sivoais joined
coetry This is perl6 version 2015.03 built on MoarVM version 2015.03 17:25
psch
n0tjack I installed mine a couple days ago; mine says 2015.09
psch n0tjack: with homebrew on osx? 17:26
n0tjack yeah
psch coetry: that version is about half a year out of date, according to n0tjack you need to update your homebrew (or something like that, i don't use OSX :) )
coetry ok thanks
psch does it have something equivalent to apt-get update
coetry: and of course afterwards rebuild rakudo
17:27 espadrine left
coetry correct, thanks 17:27
17:27 ^elyse^ left
psch right, "brew update" it is, according to the docs 17:27
17:28 antiatom joined
coetry yup yup, thanks bud 17:30
17:34 tokuhirom joined, sivoais left 17:35 sivoais joined, laouji joined 17:36 llfourn joined 17:39 tokuhirom left 17:40 lichtkind joined, laouji left 17:41 llfourn left 17:43 adu joined, sivoais left 17:45 sivoais joined 17:49 ^elyse^ joined 17:50 gonz_ joined, gonz_ left 17:51 gonz_ joined 17:53 sivoais left 17:55 sivoais joined
jdv79 _itz_: ping 17:59
18:04 sivoais left 18:05 sivoais joined
n0tjack the butterfly book has an example m/(\w+)% % [\,\s*]/ is that first % a typo? Or some weird kind of quantifier I'm unfamiliar with? (the second % I get) 18:10
m: 'eggs, milk, sugar, and flour' ~~ m/(\w+)% % [\,\s*] \s* 'and' \s* (\w+)/; say $/[0]; 18:11
camelia rakudo-moar d05cb8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/7xB_teJA2H␤Missing quantifier on the left argument of %␤at /tmp/7xB_teJA2H:1␤------> 3ggs, milk, sugar, and flour' ~~ m/(\w+)%7⏏5 % [\,\s*] \s* 'and' \s* (\w+)/; say $/[␤»
psch that looks all kinds of wrong to my eyes
timotimo yes, % only works after a regular quantifier 18:12
n0tjack yeah, I want something more long-form to read up on Grammars and Rules and such, but I'm a little leery of depending on an outdated book
timotimo what's the butterfly book? github:perl6/book ?
the docs were too short for you?
n0tjack yeah, Using Perl6, with the butterfly on the cover
timotimo OK
n0tjack rather than long form, I should say "discursive"
psch what i think it could mean is m/(\w+)+ % [\,\s*] \s* 'and' \s* (\w+)/
18:13 sivoais left
timotimo i always found S05 pretty good to read 18:13
n0tjack yeah, that aligns better with the discussion after the pattern too
psch m: 'eggs, milk, sugar, and flour' ~~ /(\w+)+ %% [\,\s*] \s* 'and' \s* (\w+)/; say $/ 18:14
camelia rakudo-moar d05cb8: OUTPUT«｢eggs, milk, sugar, and flour｣␤ 0 => ｢eggs｣␤ 0 => ｢milk｣␤ 0 => ｢sugar｣␤ 1 => ｢flour｣␤»
psch that, actually
i don't remember the distinction between % and %% in regex off-hand
n0tjack what's the %% all about?
I thought % meant "with separator"
timotimo the %% allows the thing to appear after the end again
n0tjack ah, ok
timotimo so "foo,bar,baz" vs "foo,bar,baz,"
%% allows both, % only the first kind
psch ah
%% is for the oxford comma :)
timotimo i read "%%" as "one extra %" 18:15
18:15 sivoais joined
timotimo should we put "unexpected routine 'unit', did you mean 'uniq'?" into the FAQs? 18:15
because that's an extremely common sign for "your rakudo is way out of date"
moritz timotimo: +1 18:18
psch does sound rather useful, considering we can't guarantee we'll have 6.christmas in every package manager when it's out
18:20 vendethiel joined
masak +1 18:21
18:23 sivoais left, leont joined
[Coke] timotimo: no. 18:24
.. but I am overruled. ok. :)
n0tjack huh, tokens and rules are simpler than I thought they'd be
18:25 sivoais joined 18:27 coetry left
timotimo we can't put "i see a 'unit' here so i must be out of date" checks into the old rakudos now :P 18:28
DrForr n0tjack: I'm writing a talk on those for Cluj.pm :)
timotimo n0tjack: perl6 grammars have been intentionally made simpler than perl5 regexes wherever possible 18:29
so it makes sense :)
n0tjack DrForr: preprint? :)
DrForr Not at the moment, I'm looking over perl6-ANTLR so I remember what I thought were the sticking points. 18:31
n0tjack timotimo: "a token is just a non-backtracking regex" and "a rule is a token with sigspace", and here i was thinking they'd be some complex and scary recursive-descent LL(k) something something
18:33 mattprelude joined
psch n0tjack: all of Perl 6 is full of "gee, that's it? i thought this was gonna be *hard*..?" 18:33
at least that was my impression when learning it :) 18:34
18:34 sivoais left
psch s/learning/starting to learn/ # probably fitting, not like i know all of it 18:35
FROGGS n0tjack: reducing the "lists to remember" was one of the big design principles (and still is)
timotimo my biggest problem with learning perl6 at the beginning was that "has $.foo" had a sigil, but i didn't put the sigil anywhere when i accessed $instance.foo
"where do i put the dollar sign?! how do i even ... gaah!"
FROGGS so, when you implemented something, and needed a matrix to describe its behaviour, you did your job wrong
psch .oO( and what better way to learn perl6 than by hacking the backend layers or the language itself... )
FROGGS :P 18:36
18:36 sivoais joined
timotimo well, personally, my descent into the madness that is backend hacking started when i noticed junctions were ridiculously slow even when the static optimizer could completely bypass junctions 18:37
and so i built an optimization
psch hmm, i guess my actual first backend contribution was the jvm interop improvements 18:38
things like tr/// and throwing typed exceptions from the optimizer don't really count i guess
moritz started really boring, with tests
psch: sure it counts; it's compiler hacking after al
l
psch moritz: well, yes, it's compiler hacking, but does that make it backend hacking? 18:39
tadzik it's in the backend of the compiler even :)
TimToady m: sub infix:<op>($a,$b,$c) { $a + $b * $c }; say (1..9).reduce({ $^a + $^b * $^c })
camelia rakudo-moar d05cb8: OUTPUT«141␤»
moritz psch: traditionally, the frontend of the compiler is the parser
TimToady m: sub infix:<op>($a,$b,$c) { $a + $b * $c }; say (1..9).reduce(&infix:<op>)
camelia rakudo-moar d05cb8: OUTPUT«Use of uninitialized value <element> of type Any in string context␤Any of .^name, .perl, .gist, or .say can stringify undefined things, if needed. in block <unit> at /tmp/jVefGwmYu_:1␤141␤»
moritz psch: so anything behind it is backend, ish
flussence idea: have fatal parse error messages check whether your rakudo is more than a year old, and print a suggestion to upgrade if it is 18:40
TimToady m: sub infix:<op>($a,$b,$c) { $a + $b * $c }; say [op] 1..10
camelia rakudo-moar d05cb8: OUTPUT«Too few positionals passed; expected 3 arguments but got 2␤ in sub infix:<op> at /tmp/hVYzsGwaGs:1␤ in block <unit> at /tmp/hVYzsGwaGs:1␤␤»
moritz flussence: -1
TimToady m: sub infix:<op>($a,$b,$c) { $a + $b * $c }; say [op] 1..9
camelia rakudo-moar d05cb8: OUTPUT«Too few positionals passed; expected 3 arguments but got 2␤ in sub infix:<op> at /tmp/imzAPrqlmi:1␤ in block <unit> at /tmp/imzAPrqlmi:1␤␤»
flussence yeah, figured :)
moritz flussence: I want my old code to work on my old compiler :-)
flussence but your old code wouldn't be generating parse errors on that compiler...
timotimo in the beginning i did a bunch of work with "testneeded" tickets in RT ... and forgetting to update the "plan" line at the top of the file like 9 times out of 10 :D 18:41
causing everybody grief
FROGGS my first contribution was fixing a bug when merging symbols wenn importing traits from two different modules... took two months
when*
moritz remembers FROGGS++'s fun 18:42
18:44 sivoais left
timotimo oh hey TimToady 18:44
TimToady: do we know if using $*ACTIONS rather than hanging the actions object off of the ParseShared costs us much? 18:45
oh, dinner
18:46 sivoais joined
FROGGS moritz: and the fun thing was that jnthn explain something to me which I did not understand well enough; so I tried my ways through it, to finally understand what he said to solve the issue that way 18:46
which still happens today
18:46 ^elyse^ left 18:48 Ven joined 18:49 ^elyse^ joined
n0tjack ack! 18:50
m: sub infix:<#>($a, $b) {say $b;}; 'hi' # 'there';
camelia rakudo-moar d05cb8: OUTPUT«WARNINGS:␤Useless use of constant string "hi" in sink context (line 1)␤»
n0tjack wait, that worked in my REPL 18:51
well, really best that it doesn't. I was worried.
18:52 ienh left 18:54 sivoais left 18:55 sufrosti1o left
moritz the REPL doesn't put the last statement in sink context, to be able to print it 18:55
psch n0tjack: well, it's TODO'd in S06-operator-overloading/sub.t
18:55 sufrostico left
psch i.e. it's tenetively (sp?) supposed to work, but doesn't yet 18:56
18:56 sivoais joined
n0tjack why is that a feature request? just because Perl? 18:56
psch it's not a feature request, it's a part of the specification
unless it gets removed vOv
...or maybe delayed for after 6.christmas 18:57
n0tjack yeah, sorry, terminology. "why was that considered desirable sufficient to include it in the language spec"
psch n0tjack: because CORE ops aren't special
n0tjack is # an op?
it's meta-linguistic
18:58 ab6tract joined
psch if "comment starts here" is not an op, how can we support slangs that want other characters for their comments? 18:58
n0tjack I'm not familiar with "slangs", but in other programming languages, the text to parse is an input, and the parser can handle it however it wants, irrespective of the syntax of the host language 18:59
tony-o m: sub infix:<***> { @_.join(' ').say; }; 'hi' *** 'there';
n0tjack is a slang a mini-language properly embedded in a perl script?
camelia rakudo-moar d05cb8: OUTPUT«hi there␤»
psch n0tjack: a slang is a sub language, yes 19:00
tony-o can also be used to augment the way things directly inside the script work
n0tjack ... you guys have already created a brainf*ck slang, haven't you?
psch n0tjack: although slangs are probably not the best argument there; every new foofix declaration effectively creates something like a slang
n0tjack I do like foofix declarations
tony-o n0tjack: github.com/tony-o/perl6-slang-sql 19:01
n0tjack I redefined log to be binary, because I don't like parens
psch n0tjack: not that i'm aware. i wrote a (bad and limited) ASM parser though
cognominal_ n0tjack, also regex and quotes are parsed by specialized slangs in Perl 6
psch n0tjack: there's v5, which is (parts of, because stalled) perl5 implemented as Perl 6 grammar
n0tjack yeah, I knew v5 compat was a goal of p6 from the getgo 19:02
this is an interesting and general way to achieve that
psch well, the working and recommend path currently is Inline::Perl5, which is Nativecall to perl5.so (or whatever that so is called...)
TimToady timotimo: I'm more worried about hanging the actions on the right peg, which is the cursor, since you can pass cursors outside your current dynamic scope
though as dynvars go, $*ACTIONS is one of the more expensive ones, along with $*W 19:03
psch n0tjack: github.com/masak/ipso might be interesting for you 19:04
19:04 sivoais left
n0tjack psch: oh, actually that's quite topical to what I'm reading up on right now. thanks brother. 19:05
(or sister)
19:05 Ven left 19:06 sivoais joined 19:09 Ven joined
ab6tract so I'm curious as to how far I can extend my answer to stackoverflow.com/questions/3302943...9#33037839 19:09
the obvious next step would be creating a constant inside of the OUTER or ... IMPORTER ? do we already have a "large lexical" for accessing an imported class? or is there an even more direct troute that Perl 6 provides that I'm not thinking of right now 19:11
*an OUTER
psch ab6tract: your current answer is missing two colons, isn't it? 19:12
ab6tract psch: ah yes! thanks for pointing it out. fixed. 19:13
19:14 sivoais left
n0tjack is there a cheaper way to ask "how many newlines" than .split("\n").elems? 19:15
psch m: "a\nb\nc".split("\n").elems.say 19:16
camelia rakudo-moar d05cb8: OUTPUT«3␤»
psch m: "a\nb\nc".comb("\n").elems.say
camelia rakudo-moar d05cb8: OUTPUT«2␤»
psch n0tjack: not sure about "cheaper" but "correcter" is with .comb
19:16 sivoais joined
n0tjack psch: ah, good catch 19:17
psch m: "a\nb\nc".comb("\n").say; say now - BEGIN now
camelia rakudo-moar d05cb8: OUTPUT«(␤ ␤)␤0.01508313␤»
psch m: "a\nb\nc".comb("\n").elems.say; say now - BEGIN now
camelia rakudo-moar d05cb8: OUTPUT«2␤0.0113984␤»
19:17 telex left
psch m: say +("a\nb\nc" ~~ tr/\n//); say now - BEGIN now 19:18
camelia rakudo-moar d05cb8: OUTPUT«Cannot modify an immutable Str␤ in block <unit> at /tmp/bvUoW7smEj:1␤␤»
n0tjack you'd have to test a much huger string
and average out the timings
psch m: my $_ = ("a\nb\nc");say +(tr/\n//); say now - BEGIN now
camelia rakudo-moar d05cb8: OUTPUT«Potential difficulties:␤ Redeclaration of symbol $_␤ at /tmp/f_YCUJu_59:1␤ ------> 3my $_7⏏5 = ("a\nb\nc");say +(tr/\n//); say now -␤2␤0.0766607␤»
psch m: $_ = ("a\nb\nc");say +(tr/\n//); say now - BEGIN now
camelia rakudo-moar d05cb8: OUTPUT«2␤0.025041736␤»
19:18 telex joined
psch yeah, no need, i know enough :) 19:18
tr/// is slow
luckily i didn't implement .trans, which is the slow bit :P
n0tjack puts "phasers" on the TOREAD list 19:19
psch S99:blast
synbot6 Link: design.perl6.org/S99.html#blast
pink_mist so where's the doc about photon torpedoes? 19:20
FROGGS hmmm, seems my Type:D patch is not sufficient yet... it complains about an abstract type T 19:22
masak in all fairness, the abstract type T is such a pain in the backside :P 19:23
FROGGS :P
Ven n0tjack: I explain phasers a bit in the learnxinyminutes
masak <-- quipping all the way to Christmas
lizmat
.oO( shifty types, those abstract ones )
Ven (if you have feedback for that, I'd very gladly take it)
psch "Cannot invoke this object" as an nqp compilation error on a line that does a bind..? o.o 19:25
19:25 sivoais left, Kogurr joined
psch oh, no, misread the trace 19:25
it's the call to QAST::OperationsJAST.box two lines further down 19:26
n0tjack Ven: LearnXinYMinutes is why I'm here :) S
that's what finally pushed me over the edge to download the thing
Ven n0tjack: ohhh, that's good to know :-).
I've read a few people say it was far too big...
n0tjack not for me, I wanted more
psch ohh 19:27
19:27 sivoais joined
cognominal_ Ven++ 19:27
19:27 Hor|zon left
Ven n0tjack: that's great to hear :-). I probably won't add *more* quite yet though (though I do have a list of what I'd like to add) 19:27
maybe I should just add everything I want, and add a big disclaimer "you'll learn the whole language here" 19:28
n0tjack I'm reading the butterfly book now
also a very good, accessible treatment
19:29 ienh joined 19:30 ab6tract left
Ven n0tjack: well, if you have any critiques, I'd gladly listen 19:31
n0tjack Ven: When I get a little more experience using the language, I'll be in a better position to go back and give informed feedback 19:34
Ven n0tjack: the position you're in is also very, very interesting
19:34 firstdayonthejob left 19:35 sivoais left 19:36 sufrostico joined 19:37 sivoais joined
n0tjack ok, so to the best of my recollection, my experience was (a) "hey, this is all code, awesome", because if a language can't sell itself, I don't want to hear your sales pitch 19:37
and enticed by that (and my previous investigations of p6), it was something like
cool. cool. cool. I want that. Oh, and that! I want that! Hey, I wish I could do that in $current_favorite_language 19:38
and on and on
19:38 yqt joined
FROGGS aye :o) 19:38
Ven n0tjack: well, there's still a good chunk to learn, so that's good as well 19:39
n0tjack there is a hell of a lot to learn
Ven still hasn't decided if he wants to make a "comprehensive learnxiny", just add stuff to said tutorial, or make a "blog" with separated articles and stuff 19:40
n0tjack but because p6 invites noodling around, it doesn't feel like studying, it feels like playing
garu wow, that should be a logo somewhere :)
n0tjack but then I don't know if my experience will scale. I'm an old Perl guy and an APL (J) guy, and p6 is like "Hey, do you want Perl, but never to have to write a loop again?". I mean, seriously. 19:41
Ven n0tjack: ooh, you're an APL guy?
APL is amazing :D
n0tjack yeah
I love APL. J is my primary language. K and R are also up there.
[Coke] n0tjack: you'll fit in just fine.
Ven hit some limitations in GNU APL, and hasn't found how to launch Dyalog APL in the command-line for just a script (not a workspace)
n0tjack I know Dyalog has script support in recent years, but I have no experience with their CLI 19:42
The Dyalog IDE is an integral part of the experience.
It's sublime.
Ven n0tjack: I don't want to leave my terminal :[. I have everything set up.. 19:43
n0tjack I know the feeling.
Ven but, say, GNU APL's monadic uparrow doesn't work, so oh well. Anyway, welcome along :D 19:44
n0tjack Thanks. You guys have all been great the last few days. I remember burning out on IRC many years ago because people were all so unfriendly. I was not expecting this.
grondilu a R slang will be pretty cool 19:45
19:45 sivoais left
cognominal_ dyalog, is that open source? 19:45
n0tjack no, vendor product
there are OSS APLs, but not as good.
the personal version is free-as-in-beer though 19:46
Ven vendor product, but as a student, they gave me a free license (and very rapidly at that)
19:46 coetry joined 19:47 sivoais joined
Ven (being a student is at times amazing for software stuff.. IntelliJ, dyalog, microsoft stuff... I know github also has stuff...) 19:47
dalek kudo/smile: 6643eb5 | FROGGS++ | / (11 files):
add DefiniteHOW to make Type:D and Type:U first class types

That mean that we can now smartmatch against these types, and also that the smiley does not just vanish, when these types are used stand-alone. Parameters and signatures still need to be adjusted to make use of this new DefiniteHow.
FROGGS .tell jnthn please review github.com/rakudo/rakudo/commit/6643eb58c6 19:48
yoleaux FROGGS: I'll pass your message to jnthn.
FROGGS lizmat: ^^
lizmat will look at it in a mo
Ven FROGGS: is there a perf hit? that sounds amazing :-)
19:48 pfhork joined, Hor|zon joined
FROGGS Ven: no, there isnt 19:48
19:48 ZoffixWork joined
Ven then it might just be *g* 19:48
FROGGS *g* 19:49
19:49 bbarker joined
ZoffixWork Hey. What's a usecase for ENTER phaser? Wound't code still run "every time you enter the block"? 19:49
19:51 ZoffixWork is now known as ZoffixW
[Coke] well, for one, you can put it wherever you want in the block. 19:51
psch my $db-conn; ... { ENTER { $db-conn = db-conn-init; } ... ; LEAVE { $db-conn.close } } # something like this springs to mind
[Coke] m: sub a($a) { say "$a"; ENTER { say "HI" } } ; a 19:52
camelia rakudo-moar d05cb8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/UhlZp3XsRY␤Calling a() will never work with declared signature ($a)␤at /tmp/UhlZp3XsRY:1␤------> 3($a) { say "$a"; ENTER { say "HI" } } ; 7⏏5a␤»
[Coke] m: sub a($a) { say "$a"; ENTER { say "HI" } } ; a("me");
camelia rakudo-moar d05cb8: OUTPUT«HI␤me␤»
ZoffixW psch, yeah, that's pretty much the example I'm looking at, which made me raise an eyebrow since ENTER was the first thing on the line.
psch although the out-of-order placement would still be the stronger argument
s/stronger/strongest/
ZoffixW Is COMPOSE phaser still unimplemented?
lizmat FROGGS: am I correct in seeing there is no support for parameters yet?
[Coke] ZoffixW: no tests for it, anyway. 19:53
lizmat ZoffixW: we don't really know what the COMPOSE phaser is supposed to do, afaik
ZoffixW Ah. Alright :)
Thanks.
moritz lizmat: run at the compose time of a class?
lizmat yes, but you have the mainline of the class already for that, basically 19:54
FROGGS lizmat: there is still what we got before
lizmat FROGGS: ah, okl
moritz maybe something funny with inheritance? like it's also run during child classes compose time?
lizmat m: class A { say "hello" }; BEGIN say "after" 19:55
camelia rakudo-moar d05cb8: OUTPUT«after␤hello␤»
lizmat hmmm...
m: class A { say "hello" }; INIT say "after"
camelia rakudo-moar d05cb8: OUTPUT«after␤hello␤»
19:55 sivoais left
lizmat that is.... unexpected ? 19:55
Ven m: say "foo"; class A { say "bar"; } 19:56
camelia rakudo-moar d05cb8: OUTPUT«foo␤bar␤»
lizmat m: INIT say "before"; class A { say "hello" }; INIT say "after"
camelia rakudo-moar d05cb8: OUTPUT«before␤after␤hello␤»
Ven lizmat: not really?
19:57 sivoais joined 19:59 pfhork_ joined
FROGGS m: sub foo(Int:D $a is copy) { $a = Int }; foo 42 # lizmat: should be disallowed eventually 19:59
camelia ( no output )
FROGGS maybe we can remove the special casing of :D/:U in the signature binder (right?) at some point, but jnthn might know if its worthwhile 20:00
and also when we wanna do it... because I guess that can very well wait
cognominal_ in src/core/Match.pm is there a reason that :%hash and :@list are left out from .BUILD and .new parameters? That breaks EVAL/perl round triping
20:01 pfhork left, pfhork_ left
cognominal_ also, $ast, @list and %hash can be left out from .perl output when ANY/empty 20:01
I can propose a patch 20:02
20:02 sivoais left
moritz cognominal_: +1 to patch 20:02
20:02 sivoais joined
FROGGS jnthn / lizmat: I forgot to mention, my patch causes regressions in t/spec/S14-roles/parameterized-type.t and t/spec/S32-exceptions/misc.rakudo.moar 20:04
lizmat FROGGS: ok 20:05
FROGGS most likely something simple, but still
20:08 [Sno] left
ZoffixW Is 'splat' a proper term that means something? In '*@' is the '@' the "splat"? 20:09
FROGGS I thought the * was 20:10
Ven ZoffixW: I took it from ruby for the learnx
iirc
FROGGS or not?
20:10 [Sno] joined
ZoffixW Ah ok, yeah it's the asterisk. 20:11
Thanks.
Ven FROGGS: yeah, should be
cognominal_ also I don't understand why in someplaces BUILD is a method and in some other a submethod
20:11 adu left
lizmat cognominal_: yeah, also, sometimes use $!attr instead of $.attr 20:12
FROGGS consistency is unevenly distributed
lizmat the latter being much more expensive
jnthn FROGGS: Bit tired tonight... First thing that stands out though is that you may want to declare constants (just a knowhow or something) for CONCRETE/TYPE to use with nqp::parameterizetype
yoleaux 19:48Z <FROGGS> jnthn: please review github.com/rakudo/rakudo/commit/6643eb58c6
FROGGS jnthn: k
lizmat
.oO( inconsistency however is very evenly distributed )
jnthn FROGGS: As the interner doesn't understand value types, and only cares for pointer equiv
FROGGS jnthn: no need to review it today :o) 20:13
jnthn FROGGS: otoh if you're using Rakudo's True/False for that you're already covered I guess
n0tjack if /^^/ is start of line, is /$$/ end of line?
FROGGS n0tjack: yes
jnthn: will add constants 20:14
n0tjack and if I write a grammar Foo { rule TOP {^^ stuff $$} ... }; what happens when I apply it to a multi-line string?
does the grammar get invoked for every line/
jnthn FROGGS: The MOP additons looks as I'd expect, anyway
FROGGS n0tjack: you can use the grammar as a token in another regex to quantify the regex I think 20:15
\o/
jnthn: I take that as a "well done" and a clap on my shoulder :D
cognominal_ FROGSS, I think I tried a smartch with a grammar to no avail. This different but similar. 20:16
jnthn FROGGS: Yes, nicely figured out :)
FROGGS n0tjack: though keep in mind that Grammar.parse implicitly aligns to the entire string... so use subparse here instead
psch grammars don't smartmatch, grammars .parse
jnthn FROGGS++ # daring to hack MOP stuff :)
FROGGS *g*
jnthn Like many parts of Rakudo, once you work out what's going on it's not *that* scary.
FROGGS nods 20:17
jnthn (Yes, there are some genuinely scary parts too... :))
FROGGS it helps to understand the MOP bits though
psch cognominal_: you can add a 'method ACCEPTS($str) { self.parse($str) }' to a grammar to make it support smartmatching
nine The MOP is fun :)
psch (maybe want a type on the parameter...) 20:18
FROGGS m: Grammar G { token TOP { . } }; say "foo" ~~ /<G>/
camelia rakudo-moar d05cb8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Jn3S4bAhZj␤Two terms in a row␤at /tmp/Jn3S4bAhZj:1␤------> 3Grammar7⏏5 G { token TOP { . } }; say "foo" ~~ /<G␤ expecting any of:␤ infix␤ infix stopper␤ statement end…»
FROGGS m: grammar G { token TOP { . } }; say "foo" ~~ /<G>/
camelia rakudo-moar d05cb8: OUTPUT«Method 'G' not found for invocant of class 'Cursor'␤ in block <unit> at /tmp/ZdgbmgD7EI:1␤␤»
FROGGS m: grammar G { token TOP { . } }; say "foo" ~~ /<::G>/
camelia rakudo-moar d05cb8: OUTPUT«Method 'G' not found for invocant of class 'Cursor'␤ in block <unit> at /tmp/iwqsy4bflN:1␤␤»
FROGGS m: grammar G { token TOP { . } }; say "foo" ~~ /<::('G')>/
camelia rakudo-moar d05cb8: OUTPUT«Method 'G' not found for invocant of class 'Cursor'␤ in block <unit> at /tmp/b44mZ0xwww:1␤␤»
FROGGS :/
m: grammar G { token TOP { . } }; say "foo" ~~ /<G::TOP>/
camelia rakudo-moar d05cb8: OUTPUT«｢f｣␤ G::TOP => ｢f｣␤»
psch oh, that's neat
FROGGS m: grammar G { token TOP { . } }; say "foo" ~~ /<.G::TOP> +/ 20:19
camelia rakudo-moar d05cb8: OUTPUT«｢foo｣␤»
FROGGS m: grammar G { token TOP { . } }; say "k-e-b-a-b" ~~ /<.G::TOP>+ % '-'/
camelia rakudo-moar d05cb8: OUTPUT«｢k-e-b-a-b｣␤»
FROGGS m: grammar G { token TOP { . } }; say "k-e-b-a-b" ~~ /<yummi=.G::TOP>+ % '-'/
camelia rakudo-moar d05cb8: OUTPUT«｢k-e-b-a-b｣␤ yummi => ｢k｣␤ yummi => ｢e｣␤ yummi => ｢b｣␤ yummi => ｢a｣␤ yummi => ｢b｣␤»
FROGGS well, you get the idea :o)
n0tjack yeah, makes sense 20:20
cognominal_ psch, would not that belong to Grammar.pm?
psch cognominal_: not in general, i don't think
cognominal_: there was some discussion about this a few weeks ago, but i don't remember when nor the exact reason that convinced me...
FROGGS btw, TimToady++ or whoever invented the regex assertion and (non)capturing syntax - I *love* it
20:21 Ven left, gonz_ left
psch cognominal_: also, there is the syntax FROGGS showed, which probably makes it clearer what exactly happens 20:21
ZoffixW Heh: "Meta operators ! Oh boy, get ready. Get ready, because we're delving deep into the rabbit's hole, and you probably won't want to go back to other languages after reading that." 20:22
ZoffixW gets excited
(it's from learnxiny
FROGGS I guessed so :o) 20:23
psch m: sub f($a, $b) { $a ** 2 + $b }; say [[&f]] ^5 # wee \o/
camelia rakudo-moar d05cb8: OUTPUT«148␤»
psch that's not even metaop, actually
just symbol-y reduce
m: say ^2 R[Z-] 2..4 20:24
camelia rakudo-moar d05cb8: OUTPUT«(2 2)␤»
psch m: say ^3 R[Z-] 2..4
camelia rakudo-moar d05cb8: OUTPUT«(2 2 2)␤»
lizmat m: my Int @a; @a[5] = 42; .say for @a # should be (Int) xx 5, 42
camelia rakudo-moar d05cb8: OUTPUT«(Any)␤(Any)␤(Any)␤(Any)␤(Any)␤42␤»
FROGGS lizmat: aye
lizmat fix found already 20:25
FROGGS lizmat++
it is in $*W.create_container_descriptor I guess?
dalek kudo/nom: 5b2b247 | lizmat++ | src/core/List.pm:
Fix default value of iterated List
FROGGS ohh 20:26
lizmat $ 6 'my Int @a; @a[5] = 42; .gist.print for @a'
(Int)(Int)(Int)(Int)(Int)42
it was in the iterator logic
FROGGS ahh
jnthn lizmat: ooh, I think that might be one of the xmas RTs? :) 20:27
FROGGS that'd be great
lizmat m: $ 6 'my Int @a; @a[5] = 42; $_ = 666 for @a
camelia rakudo-moar d05cb8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/_9eEiomsVM␤Two terms in a row␤at /tmp/_9eEiomsVM:1␤------> 3$7⏏5 6 'my Int @a; @a[5] = 42; $_ = 666 for ␤ expecting any of:␤ infix␤ infix stopper␤ statement end␤ …»
lizmat m: my Int @a; @a[5] = 42; $_ = 666 for @a
camelia rakudo-moar d05cb8: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/FOUi6_KTBs:1␤␤»
lizmat another one... :-)
sufrostico screen
20:28 darutoko left, Hor|zon left
FROGGS ohh, my patch actually only regresses t/spec/S14-roles/parameterized-type.t it seems 20:29
20:30 xfix left
FROGGS Variable definition of type T (implicit : by pragma) requires an initializer 20:30
------> has T $.data is rw⏏;
this will be interesting tomorrow
20:31 ZoffixW left 20:32 TEttinger joined
TimToady okay, I think I've managed to move the code over from method reduce into metaops that handles N-ary operations, so that all reductions/productions now handle N-ary ops 20:33
oops, just broke it again :) 20:34
lizmat given a type in $foo, how can I make a container of that type ? 20:35
nqp::p6scalarfromdesc doesn't help, as I can't get at the descriptor ?
20:36 Skarsnik joined
lizmat FROGGS jnthn ?? ^^ 20:37
jnthn lizmat: You...need to have a descriptor to do that
lizmat: Where can't you get at it, ooc?
lizmat well, I have just a type object 20:38
jnthn Where're you writing the code?
FROGGS m: say Int.^descriptor
camelia rakudo-moar d05cb8: OUTPUT«Method 'descriptor' not found for invocant of class 'Perl6::Metamodel::ClassHOW'␤ in block <unit> at /tmp/vGFukqeq46:1␤␤»
jnthn Array?
lizmat m: my Int @a; @a[5] = 42; $_ = 666 for @a # trying to fix this
camelia rakudo-moar d05cb8: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/jlvjmQSx5F:1␤␤»
lizmat List
Skarsnik Hi there, is that normal that rakudo take over a minute to compile a perl script (there is like 300-400 lines)? or did I fuck up my install 20:40
n0tjack wait, WHAT?
You can assign _to_ $_ ?
lizmat jnthn: e.g. line 379
n0tjack that's incredible!
moritz n0tjack: why not?
it's just a variable :-)
TimToady and it's bound rw by default
FROGGS Skarsnik: can you show us the script?
n0tjack why has no one ever done this before?
20:41 ^elyse^ left
lizmat n0tjack: pretty common in p5, for the past 25+ years afaik 20:41
FROGGS n0tjack: well, I did... I also do that in P5
n0tjack I never saw that in p5
maybe I wasn't paying attention
or maybe doing it just never occured to me
jnthn lizmat: But there should be a $!descriptor in the Array being iterated over
n0tjack I don't mean I never wrote $_ = something 20:42
FROGGS elsif ($_ = foobarbaz) { # do something with $_
jnthn lizmat: Maybe that needs passing into the ArrayIterator?
Skarsnik FROGGS, is that public? www.nyo.fr:1080/svn/fimstuff/
lizmat m: my %h = a => 42, b => 666; $_++ for %h.values; dd %h
camelia rakudo-moar d05cb8: OUTPUT«Hash %h = {:a(43), :b(667)}␤»
20:42 kfo joined
n0tjack I mean I never thought about using it to update an array in place 20:42
FROGGS Skarsnik: no, it prompts for user/pass
moritz n0tjack: s/^\s+// for @array; # is pretty common, afaict
FROGGS Skarsnik: you can use nopaste.linux-dev.org/
n0tjack moritz: that I did, but somehow $_ = 48 for @array seems different, and cooler, to me 20:43
though of course it's the same thing only more explicit
lizmat n0tjack: same applies for values of hashes
even things that look like hashes, like sets 20:44
_itz_ .tell jdv79 sorry for the deplay in adding your RSS feed, I've been away from the computer but it's added now
yoleaux _itz_: I'll pass your message to jdv79.
FROGGS m: my @a = ^10; .++ for @a; say @a
camelia rakudo-moar d05cb8: OUTPUT«[1 2 3 4 5 6 7 8 9 10]␤»
FROGGS m: my @a = ^10; .-- for @a; say @a
camelia rakudo-moar d05cb8: OUTPUT«[-1 0 1 2 3 4 5 6 7 8]␤»
n0tjack now that's cute
FROGGS *g*
lizmat m: my $s = set(<a b c d e>; dd $s; $_ = False for $s.values; dd $s
jnthn OK, heading afk for l'evening
camelia rakudo-moar d05cb8: OUTPUT«5===SORRY!5=== Error while compiling /tmp/O20x8rY1qK␤Cannot use variable $s in declaration to initialize itself␤at /tmp/O20x8rY1qK:1␤------> 3my $s = set(<a b c d e>; dd $7⏏5s; $_ = False for $s.values; dd $s␤ expecting any of:␤ …»
jnthn 'night, #perl6
FROGGS yes, we can do linenoisish things
gnight jnthn
lizmat m: my $s = set(<a b c d e>); dd $s; $_ = False for $s.values; dd $s
camelia rakudo-moar d05cb8: OUTPUT«Set $s = set("a","c","b","e","d")␤Cannot assign to an immutable value␤ in block <unit> at /tmp/jDlFHVI_CJ:1␤␤»
lizmat huh? 20:45
m: my $s = <a b c d e>.SetHash; dd $s; $_ = False for $s.values; dd $s # sets are immutable :-)
camelia rakudo-moar d05cb8: OUTPUT«SetHash $s = SetHash.new("a","c","b","e","d")␤SetHash $s = SetHash.new()␤»
n0tjack wait, so a set is just pairs where the values are True (exists) and False (doesn't)? 20:46
what do you know, it's 2015 and we have a rational perl 20:47
lizmat n0tjack: indeed
and it takes *objects* as keys
so you can have sets of sets :-)
psch m: $_ = <a b c>.Set; say .<d>
camelia rakudo-moar d05cb8: OUTPUT«False␤»
20:48 ^elyse^ joined
psch m: $_ = <a a b c>.BagHash; .grabpairs.say; .perl.say 20:50
camelia rakudo-moar d05cb8: OUTPUT«c => 1␤("a"=>2,"b"=>1).BagHash␤»
20:50 ZoffixW joined
Skarsnik FROGGS, www.nyo.fr/~skarsnik/tmp/fimstuff/ the extractsomthing.pl, it took forever to start on the 'shitty' CPU of my dedicated server 20:51
ZoffixW How come this doesn't produce all letter codes:
m: say { 'A'.ord }, { 'B'.ord } ... { 'Z'.ord };
camelia rakudo-moar d05cb8: OUTPUT«(65)␤»
psch m: say { 'A'.ord }(), { 'B'.ord }() ... { 'Z'.ord }(); 20:52
camelia rakudo-moar d05cb8: OUTPUT«(65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90)␤»
ZoffixW Ah. Thanks.
lizmat m: say 'A'.ord .. 'Z'.ord
camelia rakudo-moar d05cb8: OUTPUT«65..90␤»
psch or that... :)
ZoffixW :)
psch not sure why it only prints the first though
m: say ({ 'A'.ord }, { 'B'.ord } ... { 'Z'.ord });
camelia rakudo-moar d05cb8: OUTPUT«(65)␤»
20:53 adu joined 20:54 laouji joined
psch ...something about single arg rule? o.o 20:54
i have no idea
FROGGS Skarsnik: ahh, I thought you were saying it took a minute to *parse* the script... seems more like it would take a minute to *run* it
Skarsnik No no, it took a minute to parse 20:55
FROGGS hmmmmm
I mean, it also needs to parse the dependencies but still
20:55 rindolf left
FROGGS we're parsing (and partially running) 25_000 lines of code of the setting in about 48s on my laptop 20:56
20:56 lichtkind_ joined
Skarsnik (skardev)skardev@ks364334:~/fimstuff$ time /opt/bin/perl6 extractbookshelf.pl 149291 20:57
===SORRY!===
Unable to parse expression in parenthesized expression; couldn't find final ')'
real 1m2.600s 20:58
user 1m0.500s
ZoffixW Hm.... What would be a good example in response? twitter.com/cpan_pevans/status/652...6457304064
Skarsnik I put an error or purpose
*on
ZoffixW is too n00b to think of anything good
20:58 laouji left
lizmat
.oO( it's wierd alright :-)
 20:59
psch ZoffixW: for Whatever? well, it's the most light-weight closure syntax i've seen yet
(barring a few gotchas where we don't allow it to autocurry)
20:59 lichtkind left
Skarsnik I should ask for an upgrade of this server x) model name : Intel(R) Celeron(R) CPU 220 @ 1.20GHz 20:59
FROGGS Skarsnik: you can try running it as perl6 --parse-compile ... 21:00
Skarsnik: maybe that tells us something
21:01 lichtkind_ is now known as lichtkind
Skarsnik also the vim syntax file slow the hell of vim on the StoryFac.pm6 x) 21:01
does it output a file?
21:02 adu left
Skarsnik hm no option --parse-compile 21:03
psch maybe -c? that's syntax check and BEGIN and CHECK 21:04
n0tjack Skarsnik: perl6 --help offers a --stagestats
maybe run that to pin down what stage is slow
psch .oO( on a 1.2ghz celeron it's likely all stages will be slow... ) 21:05
pink_mist . o O ( might it be possible to use more than one core to speed up initial compilation? :P ) 21:08
21:08 n0tjack left
psch m: my &add2 = * + 2; say ^5 .map: &add2 21:08
camelia rakudo-moar 5b2b24: OUTPUT«(2 3 4 5 6)␤»
dalek kudo/nom: 2275f64 | TimToady++ | src/core/ (2 files):
generalize N-ary reduce/produce to [op] & [\op]

The N-ary code was only implemented in method reduce (and by extension, sub reduce). This is now moved over to the REDUCE generators in metaop.pm so those also do N-ary correctly.
Additionally, the code should now work for functions that return lists and not just for scalars. (Not tested, but same transformation we used in the sequence operator to use push (or unshift) that is overridable with a slip, so a list is treated as a unit by default, but can be overridden.)		 21:09
psch ZoffixW: is that an example for paul evans..?
lizmat pink_mist: not before Xmas
dalek ast: 6f2a9f0 | TimToady++ | S32-list/ (2 files):
updated reduce/produce tests for general N-ary
Skarsnik There is probably only one core x)
ZoffixW psch, "that" is what?
pink_mist Skarsnik: haha, yeah, but I meant for other people :P with saner CPUs :P 21:10
ZoffixW psch, timotimo gave him some crazy stuff :D I think he'll be satisfied (Paul Evans == Leonerd from Perl 5, BTW)
Skarsnik Does not seem very helpful 21:11
Stage start : 0.000
Stage parse : 62.914
Stage syntaxcheck: Syntax OK
lizmat .tell jnthn my Int @a; @a[5] = 42; $_ = 666 for @a fails because it gets an IterationBuffer as target, not an ArrayReificationTarget
yoleaux lizmat: I'll pass your message to jnthn.
psch ZoffixW: before daleks outburst i let camelia run a snippet 21:12
lizmat .tell the IterationBuffer doesn't have a descriptor
yoleaux lizmat: I'll pass your message to the.
psch ZoffixW: but if timotimo++ gave example already those are probably better... :)
ZoffixW :)
psch, Ah. I'll relay your example as well. Thanks :) 21:13
psch ZoffixW: it's somewhat academic, i guess :P 21:14
ZoffixW: the important bit, in my opinion, is that Whatever is strongly multi-purpose
ZoffixW nods 21:15
21:16 chrstphrhrt left
FROGGS Skarsnik: the option is called --profile-compile 21:19
Skarsnik It take 14 sec on my desktop Oo (it's a recent I5) 21:20
FROGGS that
err
gnight
lizmat gnight FROGGS
alpha123 anyone happen to know if rakudobrew works on freebsd? 21:23
21:25 Hor|zon joined, skids left
Skarsnik --profile-compile does not create of a file or display info, hm 21:26
21:27 chrstphrhrt joined
timotimo i was really quite certain PEvans was refering to the ... syntax 21:28
21:29 Hor|zon left 21:30 heisenberg joined
ZoffixW I think he was referring to the whole thing, because it's unclear that there are Whatever Stars as well as ... 21:30
lizmat .tell jnthn I put the descriptor issue on RT #126312 , I think it needs deeper thoughts than I currently can muster 21:31
yoleaux lizmat: I'll pass your message to jnthn.
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=126312
21:31 coetry left 21:32 coetry joined
psch timotimo, ZoffixW: yeah it's completly unclear which syntactic part he's wondering about, but perl5 has ... (although somewhat different), so i thought it's about Whatever 21:33
timotimo oh 21:34
psch additionally, Whatever is mentioned in the original tweet, which is another hint that's what it's about
timotimo i think he probably meant the "my @FIB" part
"i've only seen this used to demonstrate fib"
makes more sense, huh? :)
psch i'm not saying i'm not overinterpreting... :P
maybe it's about *+*, which is obviously a specific token for fib 21:35
&term:<*+*>
21:35 tokuhirom joined
psch as a side note: progress! the ArrayIndexOutOfBoundsException now complains about index -1, not index 0! 21:36
ZoffixW timotimo, he did say "...this syntax..." as a Perl coder whose modules are in core Perl 5, I'm sure he knows what 'my @fib' means :)
psch .oO( steps in the wrong direction are still progress, right..? )
timotimo well, it's perl 5 vs perl 6; maybe he wanted to point out how foreign perl6's syntax feels?
or how perl6 is a one-trick-pony that can only do fibonacci?
(and not even fast)
ZoffixW ¯\_(ツ)_/¯ 21:37
timotimo m: sub prefix:<¯\_(ツ)_/¯>($a) { say "pff." }; ¯\_(ツ)_/¯1
camelia rakudo-moar 5b2b24: OUTPUT«pff.␤»
21:37 coetry left
timotimo m: sub term:<¯\_(ツ)_/¯>($a) { say "pff." }; ¯\_(ツ)_/¯ 21:38
camelia rakudo-moar 5b2b24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/3O2CDun_n6␤Calling term:<¯\_(ツ)_/¯>() will never work with declared signature ($a)␤at /tmp/3O2CDun_n6:1␤------> 3ub term:<¯\_(ツ)_/¯>($a) { say "pff." }; 7⏏5¯\_(ツ)_/¯␤»
timotimo m: sub term:<¯\_(ツ)_/¯>() { say "pff." }; ¯\_(ツ)_/¯
camelia rakudo-moar 5b2b24: OUTPUT«pff.␤»
ZoffixW nerdgasms
21:38 llfourn joined
timotimo that's just nerdgames 21:39
vendethiel just won a game, and is going to sleep
have fun, #perl6!
21:39 tokuhirom left
ZoffixW \o 21:39
timotimo gnite vendethiel! 21:40
what did you play? monkey jousting? :)
ZoffixW imagines Perl 6 will eventually have a quite amusing Acme:: collection
timotimo did you see ACME::AddSlashes? 21:41
masak slashes. such safe.
ZoffixW lols
timotimo is now safe from slashes 21:42
jonadab Ah, when there's an ACME:: module that outputs code that looks like a NetHack dumpfile, then I'll be impressed.
timotimo it'd output code?
mst timotimo: if it doesn't make an HTTP connection to AO3 it's not trying hard enough
jonadab timotimo: Or accept code. 21:43
timotimo hmm
mst: i don't follow; what's AO3?
jonadab (I'm thinking like Eyedrops or Bleach.)
21:43 llfourn left
timotimo i don't know either of those, jonadab :( 21:43
ZoffixW How do I make that warn message refer to where my shiny new operator is at:
m: sub prefix:<¯\_(ツ)_/¯>(Str $warning) { warn $warning }; ¯\_(ツ)_/¯ "dunno what's going on here"
camelia rakudo-moar 5b2b24: OUTPUT«dunno what's going on here in sub prefix:<¯\_(ツ)_/¯> at /tmp/KyvmA9Hy6J:1␤»
jonadab timotimo: Old Perl5 ACME:: modules.
mst timotimo: archiveofourown.org
alpha123 jonadab: was there a nethack dump Acme for perl5? 21:44
jonadab alpha123: Not to my knowledge.
timotimo ah, i haven't done any perl 5 programming so far
jonadab Well, at this point, it's better to go ahead and learn 6. 21:45
timotimo hah
not for everybody, surely
alpha123 Don't learn perl6, you won't be able to use any language except perl6
jonadab If you're starting, I mean.
timotimo too late for me now :)
masak alpha123: I find the opposite to be the case.
jonadab alpha123: I've been having THAT problem with 5. 21:46
masak alpha123: I use a lot of other languages, but I use them better because I know Perl 6.
alpha123 masak: Mostly tongue-in-cheek, I use C++ and ruby and javascript for work all the time
masak nod
timotimo we don't have Inline::JavaScript yet
alpha123 but if I could write everything in perl 6, rust, and common lisp I probably would :) 21:47
timotimo but Inline::C should handle C++ as well; or it will at some point
so far, Inline::C is only for individual functions, though, IIRC
alpha123 unfortunately ruby has started to feel like a half-assed perl 6 ;(
jonadab timotimo: What, you haven't got the Mozilla suite running in Inline::C yet? 21:48
timotimo the mozilla suite? now there's a blast from the past
jonadab alpha123: Ruby is intermediate between Perl5 and Perl6, both chronologically and in terms of the design.
timotimo: Well, Seamonkey then, same thing.
ZoffixW m: sub prefix:<¯\_(ツ)_/¯>($msg) { $msg.say }; ¯\_(ツ)_/¯ “Programming is hard! Let's go shopping!”
camelia rakudo-moar 5b2b24: OUTPUT«Programming is hard! Let's go shopping!␤»
ZoffixW This is amazing :) Perl6++ 21:49
jonadab Hehe.
21:49 sufrostico left
masak m: say "Programming is ", <hard easy weird>.pick, ", let's go ", <shopping hunting programming>.pick, "!" 21:50
camelia rakudo-moar 5b2b24: OUTPUT«Programming is hard, let's go hunting!␤»
ZoffixW lol
psch weird:programming is my favorite combination
timotimo m: say "Programming is {<hard easy weird>.pick}, let's go {<shopping hunting programming gathering>.pick}!" 21:51
camelia rakudo-moar 5b2b24: OUTPUT«Programming is hard, let's go programming!␤»
21:51 ^elyse^ left
ZoffixW 0o 21:51
masak psch: good news! you'll get it one time out of nine :P
mst jonadab: I've not seen you on #perl asking for help, so I'm not convinced you're trying to learn perl5 yet :P
lizmat m: say "Programming is {<hard easy weird>.pick}, let's go {<shopping hunting programming>.pick}!"
camelia rakudo-moar 5b2b24: OUTPUT«Programming is hard, let's go shopping!␤»
mst also the correct answer at the moment is probably "learn both"
masak the correct answer is almost always "both", modulo time constraints and sanity
jonadab mst: No, no, I mean the problem I have with Perl5 is that I learned it, and now I can't use any other language because they're all so cumbersome.
I want to learn 6 though. 21:52
ZoffixW Yeah. Especially since P5 has a huge number of modules that can be used in P6 with Inline::Perl5
mst jonadab: oh, sorry
jonadab: I basically have that problem given Moo(se)
21:52 rurban joined
mst jonadab: nobody seems to believe me ;) 21:52
jonadab Oh, man, I've seen Moose, and it makes my head hurt.
masak would like to learn Perl б
jonadab: Moose have antlers, and if you approach them in the wrong way, they can make your head hurt. 21:53
alpha123 every time I learn a little bit more about perl 6 I always go "HOLY COW THAT'S ACTUALLY AMAZING". today it was with Actions and multi MAIN()
TimToady m: sub infix:<op>($a,$b,$c) { $a + $b * $c }; say [op] 1..9
camelia rakudo-moar 2275f6: OUTPUT«141␤»
masak Meese* 21:54
TimToady m: sub infix:<op>($a,$b,$c) { $a + $b * $c }; say [\op] 1..9
camelia rakudo-moar 2275f6: OUTPUT«(1 7 27 69 141)␤»
21:54 n0tjack joined
TimToady m: sub infix:<op>($a,$b,$c) is assoc<right> { $a + $b * $c }; say [\op] 1..9 21:54
camelia rakudo-moar 2275f6: OUTPUT«(9 79 479 1919 3839)␤»
jonadab masak: If you actually understand Moose, perhaps you can figure out why TAEB won't work lately. 21:55
masak I don't claim any extraordinary understanding of Moose, no.
jonadab Ah.
masak I've watched people implement MOPs, though.
dalek kudo/nom: f254cd4 | lizmat++ | src/core/Hash.pm:
Remove copy-pasto
kudo/nom: 51d2c07 | lizmat++ | src/core/Iterable.pm:
Don't use a scope if we don't need one
kudo/nom: ec2cad9 | lizmat++ | src/core/Str.pm:
Make .comb.elems faster
lizmat good night, #perl6! 21:56
psch g'night lizmat o/
masak 'night, lizmat 21:57
psch m: my Int @a; @a[4]++; @a.map:{say .WHAT.perl};
camelia rakudo-moar 2275f6: OUTPUT«Int␤Int␤Int␤Int␤Int␤»
psch that's #123037, which iirc lizmat fixed today, i'll note that in the ticket 21:58
synbot6 Link: rt.perl.org/rt3/Public/Bug/Display...?id=123037
mst jonadab: huh? Moo(se) is basically 'we stole the perl6 metamodel, ish' 21:59
jonadab: it's the only way to write OO in perl5
timotimo that's not right 22:00
you can build your own object orientation
masak as many did, pre-Moose 22:01
and post-Moose too, I'm sure
timotimo oh!
sorry, i read perl6 where you wrote perl5
i'% silly
masak also "we stole the Perl 6 metamodel" is a truth with qualifications
mst masak: did you miss the 'ish' ? 22:02
ZoffixW m: 'fooABCABCbar' ~~ / foo ( AFOO BC+ ) + bar /;
camelia rakudo-moar 2275f6: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at /tmp/hfSzcPl9Ja:1␤ ------> 3'fooABCABCbar' ~~ / foo ( A…»
ZoffixW ^^ isn't that annoying?
masak what actually happened was more like "we stole *Pugs's* Perl 6 metamodel", and then Rakudo stole Moose's back
ZoffixW I'd even call it a bug.
mst right, I spent quite a while running around to get people to actually talk to each other ;)
TimToady well, both were developed in the same cabin at the same time
22:03 vendethiel left
masak at some point yes. after both had existed for quite some time :) 22:03
oh, or was there an earlier cabin thing, around 2006?
psch ZoffixW: what's supposed to be buggy there?
(might be clear if i could see where the eject is...) 22:04
masak I really like how Moose puts almost everything worthwhile into the methods. that's necessary, because the blessed-hashrefs substrate it's building on makes the attributes quite dumb and un-private 22:05
with Perl 6's MOP, the balance is a bit different
ZoffixW psch, why is it warning me? It's a capture and I want to space out the terms.
psch, here: fpaste.scsys.co.uk/500157
n0tjack what's the right doc to understand all the rx modifiers, like sigspace, i, 1st, etc?
ZoffixW n0tjack, docs.perl6.org/language/regexes#Adverbs
n0tjack merci 22:06
ZoffixW n0tjack, I don't remember if I've seen 1st in there. It might be elsewhere
psch ZoffixW: i think i agree with you. i don't see that warning adding anything of value 22:08
ZoffixW nods
psch m: say "foo" ~~ / f o o /
camelia rakudo-moar 2275f6: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at /tmp/bjD588wdro:1␤ ------> 3say "foo" ~~ / f7⏏5 o o…»
timotimo consider this:
m: say "here be dragons" ~~ /here be dragons/
camelia rakudo-moar 2275f6: OUTPUT«Potential difficulties:␤ Space is not significant here; please use quotes or :s (:sigspace) modifier (or, to suppress this warning, omit the space, or otherwise change the spacing)␤ at /tmp/hVqDOXMXIy:1␤ ------> 3say "here be dragons" ~~ /h…»
timotimo that will be a match fail.
m: say "here be dragons" ~~ /"here" "be" "dragons"/
camelia rakudo-moar 2275f6: OUTPUT«Nil␤»
timotimo ^- you can always make explicit that you don't expect spaces to match there 22:09
m: say "here be dragons" ~~ /"here" be "dragons"/
camelia rakudo-moar 2275f6: OUTPUT«Nil␤»
timotimo ^- it only triggers when spaces show up between non-quoted literals
psch yeah, between atoms
that's what i saw in the code :s
ZoffixW Alright. I guess I need to use this in real life code to see if it's annoying or not :)
I just spotted it from an example in the learnxiny
ZoffixW will fix the example 22:10
psch with the explanation i can readily accept it being there - it's apparently not something that would come up in my own code, cause i haven't seen it before... :)
which probably means it's might be useful
ZoffixW Yeah
psch -'s 22:11
22:17 cognominal_ left, n0tjack left
timotimo i'd kind of be interested to see each error message get its own little page on the 'web 22:18
with a comments function
"like this exception on facebook"
psch haha
timotimo angular has something like that 22:19
when it spits errors at you, it also has a link to a sort of explainer page on their 'site
psch well, we already try to have awesome errors
but i guess having it explained more accessible is always possible 22:20
s/accessible/thoroughly/ # maybe? 22:21
timotimo mhm 22:22
22:26 Hor|zon joined
ShimmerFairy I think a link in error messages would be weird, but naming the exception type would be very nice for looking up more info :) 22:26
(that is, sticking a self.^name somewhere appropriate in error messages) 22:27
22:28 lichtkind left
ZoffixW W00t! I'm not the final stretch: "APPENDIX A: ... It's considered by now you know the Perl6 basics." 22:28
22:30 Hor|zon left
timotimo ShimmerFairy: hm. on the one hand, i agree, on the other hand, that could be a bit weird? 22:30
not sure
ShimmerFairy timotimo: just thinking about when you want to find more info on an error, I think it'd be more helpful to have 'X::Foo::Bar' to look up, than the more usual "copy-paste part of the error message" you do in other languages. 22:32
timotimo yeah
ShimmerFairy Maybe not printed by default, but an easy way to get the name of the error type wouldn't hurt :)
timotimo aye, like --ll-exception 22:33
but it's not quite easy to get --ll-exception "in there"
i'd prefer an env var for that
or perhaps an env var that just has the same effect as putting --ll-exception
ShimmerFairy --verbose-exception :P
timotimo "--i'm-trying-to-fix-this-so-please-tell-me-what's-up" 22:35
22:35 cognominal_ joined
ShimmerFairy
.oO( --VERBOSE / --BRIEF / --SUPERBRIEF )
 22:36
pink_mist --BOXER
timotimo i've never worn a verbose down there; is it comfortable? 22:37
pink_mist 0_o I was thinking the breed of dog 22:38
timotimo haha
isn't "boxer briefs" a common kind of underwear?
ZoffixW yeah 22:40
pink_mist I'm not sure .. I know of boxer shorts, and those are usually 'opposed' as to briefs 22:41
-as
timotimo oh
Skarsnik hm, now DBIish do method always give me 0E0
pink_mist seems like one of those perl5 '0 but true' type values 22:42
timotimo nope
0e0 is just a Num
pink_mist m: say so 0e0
camelia rakudo-moar ec2cad: OUTPUT«False␤»
Skarsnik 0E0 is no row
oh it's a real false value? 22:43
pink_mist apparently
timotimo i have a hard time coming up with the right english words to explain what my understanding of boxer shorts vs boxer briefs is
pink_mist I've never heard of boxer briefs before
(of course I may just have lived a very sheltered life 0_o) 22:44
ShimmerFairy timotimo: "boxer briefs" is a thing, but I've no clue how it compares to the other types :P
timotimo this won't surprise you, but if you google-image-search for "boxer briefs", it'll show you a bunch of crotches :)
pink_mist hah
Skarsnik The issue. I do a SELECT * request. doing prepare + execute and a fetchrow_array.elem give me 17, doing the same with do give me E0E, it should give me an object ~~
timotimo so, this isn't a good description, but i think boxer briefs is the type that offers a sort-of pouch for the male physique
22:44 spider-mario left
timotimo whereas the boxer shorts just have a seam from the front to th eback 22:45
god damn it, all these dudes in the images for "boxer briefs" are *super* ripped
pink_mist comparing google image search results for boxer briefs vs boxer shorts, I'd say boxer shorts are loose, and boxer briefs are tight
but I've never heard the distinction before
timotimo mhm 22:46
ZoffixW Boom. Done! I finished learnyinx! I now officially know a useful amount of Perl 6 :) 22:47
pink_mist Zoffix++
ZoffixW And I even get to score a PR for #Hacktoberfest by submitting some corrections :)
timotimo Zoffix++ 22:48
i think i need me some sleepy times
pink_mist ep.yimg.com/ay/stylinonline/domo-ku...iefs-5.jpg hah, this was a fun one
(it's safe for work)
timotimo heh 22:49
i don't actually know what domo kun actually is
i've seen 'em show up in a bunch of gifs, though 22:50
waving their arms up and down
pink_mist aye I don't know either, I just found the strategic placement of the mouth funny =)
22:51 localJoe joined
timotimo of course :) 22:51
localJoe Is there a Perl module for use on Windows with WPF applications that is similar to Win32::GuiTest ? 22:52
... namely, to do Windows UI automation.
ZoffixW localJoe, I had some success using NativeCall to call functions from AutoIt3 (www.autoitscript.com/site/) DLLs. I'd go that route.... but... have you tried seeing if Win32::GuiTest works with Perl 6's Inline::Perl5 ? 22:53
Oh, it's got XS. Never mind. I don't think Inline::Perl5 can handle XS stuff 22:54
localJoe ZoffixW, thanks for your reply. I am familiar with AutoIt, which is nice but not robust for my needs (lots of automated tests needed). Win32::GuiTest is for WQin32 applications, does not fly well with WPF applications. 22:55
*Win32
ZoffixW is lucky enough to not know what "WPF applications" are :D
localJoe I would like to do automate Windows WPF applications using Perl. For Win32 applications, Win32::GuiTest was good 22:56
Well, my work thing is on Windows... 22:57
22:57 RabidGravy left 22:59 Skarsnik left
masak 'night, #perl6 23:03
ZoffixW night
localJoe hmm, if irc doesnt know the answer, then it doesnt exist, ya? 23:04
ZoffixW localJoe, all Perl 6 modules are here and I see no Win32 in there: modules.perl6.org/ 23:05
23:06 skids joined
ZoffixW As for IRC answers: keep in mind it's Friday night :D Also, try asking in #perl too, maybe they know. 23:06
colomon Zoffix: I’m pretty sure Inline::Perl5 can indeed handle XS stuff. # disclaimer: have not tried this myself 23:07
ZoffixW Oh yeah, the docs do say "Supports Perl 5 modules including XS modules" 23:09
ZoffixW wonders where they heard "no XS" bit
localJoe Thanks for the hints. I wonder even if Wind32::GuiTest in Inline::Perl5, even if it supports XS stuff, could handle WPF applications. I need to look around more. And try in #perl too. Thanks 23:12
23:14 kid51 joined, captain-adequate left 23:15 bjz left
timotimo you can totally use XS modules with Inline::Perl5. 23:17
"no XS" is for the v5 module
because that's basically a grammar that parses perl 5 and generates the same kind of AST perl6 uses 23:18
23:18 bjz joined
ZoffixW Ahhh. Thanks for clarifying. 23:18
For a second I thought I will have to quit drinking :)
23:19 kid51_ joined, kid51 left, kid51_ is now known as kid51, n0tjack joined 23:21 Peter_R left 23:23 kid51_ joined 23:24 kid51 left, kid51_ is now known as kid51 23:27 Hor|zon joined 23:30 n0tjack left 23:31 tokuhirom joined, Hor|zon left, n0tjack joined
alpha123 Anyone have any idea how to get perl6 to compile on FreeBSD? I've determined it seems to want GCC (which I ahve installed) but I can't actually get it to use my GCC 23:32
23:32 localJoe left, localJoe joined
alpha123 (rakudo, specifically) 23:34
timotimo alpha123: it can handle gcc as well as clang 23:36
you may have to do moarvm individually from the rest so that it picks up the right configuration?
all that's important for building everything correctly is that you give all Configure.pl scripts the same --prefix=
in my case that's /home/timo/perl6/install
and then you build MoarVM, nqp and rakudo in that order, with "make install" steps in between 23:37
alpha123 timotimo: indeed, I seem to have trouble building moarvm with --backend=moar --gen-moar from rakuod's configure.pl script. I'll try building it separately. thanks
23:38 chrstphrhrt left, localJoe left, tokuhirom left
timotimo good luck! 23:42
i'll go to bed now
alpha123 had to fiddle around with --cc, --compiler, and --toolchain but it seems to work, though it wasn't super happy building with clang (I probably messed up my environment somehow while trying to get it to use gcc) 23:45
thanks
23:46 yqt left 23:48 BenGoldberg joined
ZoffixW If anyone wants to have a whack for the few items (top) that I was not able to fix myself: github.com/adambard/learnxinyminut...ssues/1390 23:48
And now it's... Beer O'Clock! :D 23:49
23:49 ZoffixW left 23:56 n0tjack left 23:57 rurban left 23:59 adu 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!