»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
dalek kudo-star-daily: 7cfed4d | coke++ | log/ (13 files):
today (automated commit)
 00:10
rl6-roast-data: de52682 | coke++ | / (6 files):
today (automated commit)
vendethiel m: for <a b c d e f> { .say if 'a' ff 'e'; }; # shouldn't work, right ? 00:12
camelia rakudo-moar 847bd5: OUTPUT«a␤b␤c␤d␤e␤f␤»
vendethiel m: for <a b c d e f> { .say when 'a' ff 'e'; }; # but does this ... ?
camelia rakudo-moar 847bd5: OUTPUT«===SORRY!===␤Cannot convert string to number: base-10 number must begin with valid digits or '.' in '⏏a' (indicated by ⏏)␤»
vendethiel m: for <a b c d e f> { .say with 'a' ff 'e'; }; # but does this ... ? 00:13
camelia rakudo-moar 847bd5: OUTPUT«===SORRY!=== Error while compiling /tmp/6lLaj0fqyR␤Two terms in a row␤at /tmp/6lLaj0fqyR:1␤------> for <a b c d e f> { .say ⏏with 'a' ff 'e'; }; # but does this ... ␤ expecting any of:␤ method argu…»
masak having `.say when 'a' ff 'e'` work would solve oodles of use cases. 00:18
vendethiel I think so.
But break sooo much consistency :D
masak ...but I can't see how it would, because the things in the 'a' and 'e' positions still don't smartmatch. 00:19
vendethiel masak: mmh ?
well, sleep&. 00:22
masak m: macro infix:<smff>($l, $r) { quasi { $_ ~~ {{{$l}}} ff $_ ~~ {{{$r}}} } }; for 'a'..'f' { .say if 'b' smff 'e' }
camelia ( no output )
masak ...I guess $_ doesn't bind right in that one. 00:23
m: macro infix:<smff>($l, $r) { quasi { $COMPILING::_ ~~ {{{$l}}} ff $COMPILING::_ ~~ {{{$r}}} } }; for 'a'..'f' { .say if 'b' smff 'e' }
camelia rakudo-moar 847bd5: OUTPUT«use of uninitialized value of type Any in string context␤use of uninitialized value of type Any in string context␤use of uninitialized value of type Any in string context␤use of uninitialized value of type Any in string context␤use of uninitialized…»
masak ah, well.
BenGoldberg What's ff?
timotimo flip flop op
BenGoldberg Ok. 00:24
timotimo it carries internal state
masak BenGoldberg: it turns true when the lhs is true, then stays true until the rhs is true.
BenGoldberg Like perl5's scalar .. operator.
Which always confuses me and hurts my brain, 100% of the time I see it. 00:25
masak yes, that one. 00:26
and fff is like the scalar ... operator.
BenGoldberg All I can say is that I'm glad perl6 writes flipflop differently from .. and ... 00:28
masak we believe firmly in not making ops context-dependent along that particular dimension (item/list context). 00:30
or subs/methods, for that matter.
BenGoldberg I once spent a couple of frustrated hours on some perl5 code, where I had written $foo[ $x .. $#foo ], where I should have had @foo[ $x .. $#foo ] 00:31
masak heh :)
'night, #perl6
BenGoldberg 'night masak. 00:32
[Coke] waves from Ohio. 00:43
geekosaur waves at [Coke] en passant 01:02
[Coke] ~~ 01:04
grondilu about prefix:<+>, in S03: « For values that do not already do the "Numeric" role, the narrowest appropriate type of "Int", "Rat", "Num", or "Complex" will be returned; 01:22
»
So for instance:
r: say (+"3.0").WHAT
camelia rakudo-{parrot,jvm,moar} 847bd5: OUTPUT«(Rat)␤»
grondilu but for constistency shouldn't prefix:<+> also returned a narrow value when called with a Numeric argument? 01:23
hang on
shouldn't +"3.0" be an Int according to spec? 01:24
grondilu basically thinks prefix:<+> should always narrow 01:25
grondilu thought of that while thinking of rosettacode.org/wiki/Test_integerness 01:27
grondilu would have liked to test integerness with +$x ~~ Int
grondilu well on second thought that's probably not a good idea. If someone writes a literal value as 3.0, he *wants* a Num an not an Int, and prefix:<+> should not change that I guess. 01:31
Mouq grondilu: Isn't that why .narrow exists? 03:53
moritz: That looks.. like exactly what I want to write but better :) I'll be happy to steal from it :) though maybe not tonight 04:01
(RE:: irclog.perlgeek.de/perl6/2014-06-21#i_8908219) 04:02
dalek rl6-bench: 390b19d | (Geoffrey Broadwell)++ | bench:
Show tag dates in `bench list-tags`
 04:16
rl6-bench: 3d4ae2a | (Geoffrey Broadwell)++ | timeall:
Much less broken commit time detection in timeall

After this commit, you should rerun timings so that upcoming history comparison functionality will work properly.
rl6-bench: e6f852b | (Geoffrey Broadwell)++ | / (2 files):
First (text format) support for history comparison
japhb .tell timotimo New perl6-bench command `bench history` ought to interest you. Ping me if you need help. :-) 04:20
yoleaux japhb: I'll pass your message to timotimo.
japhb .tell timotimo html and html_plot versions coming when I get another good hack session.
yoleaux japhb: I'll pass your message to timotimo.
dalek ast: b3fdfeb | (David Warring [email@hidden.address] | integration/advent2011-day20.t:
adding advent 2011 day 20
 05:38
FROGGS__ m: "foo".encode ~ 42 # why does that what it does? 07:14
camelia rakudo-moar 847bd5: OUTPUT«(signal )»
dalek kudo/nom: 25c1e5d | (Tobias Leich)++ | src/core/Buf.pm:
add subbuf-rw, an lvalue subbuf like we have a substr-rw for strings

The replacement must be a Blob and this sub needs to be specced.		 07:25
masak antenoon, #perl6 08:29
moritz \o masak 08:33
masak so quiet -- guess people are at YAPC::NA and stuff. 08:35
moritz it's Sunday :-) 08:40
vendethiel wrt "writing ff" : it's better than `..`, but it's harder to read because it doesn't stand out 08:44
moritz rename it infix:<FLIPFLOP> 08:45
FROGGS__ my expectation when seeing ff is that is a sub and not an infix 08:47
vendethiel moritz: well, I could. But I don't think it helps that much 08:55
moritz vendethiel: then it would stand out
vendethiel ... unless something else than "FLIPFLOP3 is in caps 08:56
moritz well, Perl 6 takes care to only upper-case things that should stand out 09:06
if your own code use lots of all-caps, you're screwed
but of course you can write code to defeat any technique to make things stand out :-) 09:07
vendethiel moritz: YES INDEED DOT GOOD POINT COLON CLOSING PAREN 09:10
.oO( we've already gone back to CL-style naming ? )
psch hi #perl6 09:38
masak hi psch!
psch masak o/ 09:39
gist.github.com/peschwa/cb2a142e49afce7a260c
there's a few spots im not clear on the spec though
for example, tr/// returning False seems weird, but happens in nom for s/// as well 09:40
on moar and parrot that is
at least locally 09:41
maybe that's a platform thing again? :/
also i think 'whitespace characters are taken literally' and 'The .. range sequence is the only metasyntax' clash for ranges that start or end on a whitespace 09:42
this is all related to RT #62064, fyi 09:43
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...l?id=62064
psch oh, and tr[][] (and tr{}{}, and i guess other paired delims) also work 09:46
masak I have no idea why that `False` would be useful. 09:47
the action method seems to generate the moral equivalent of `$_ = .trans($from, $to);` 09:49
which... shouldn't be False. 09:50
psch masak: i updated the gist; i'm seeing the same behavior for s/// in the repl; my tr/// implementation still gives False when called from -e though 09:54
where the repl for s/// is "upstream" nom 09:55
github.com/peschwa/rakudo/compare/implement_tr # this is my branch, for reference 10:06
mostly oriented myself on how s/// works... :) 10:07
wrt spec: the delimiter for tr shouldn't change behavior i assume? i.e. tr[][] still only accepts exactly one string for search and replacement each 10:11
masak sure. 10:15
punter I want to make a Jabber server (for my site). I'm a Perl 5 programmer. Do you think it's possible to make it in Perl 6? Does Perl 6 have the networking features needed? 10:30
I can do it in AnyEvent of Perl5. Does Perl 6 have something similar?
psch punter: perl6 does have working sockets and you i think you can write event-based programms with supplies 10:36
but i'm no expert and there might be a better answer
moritz punter: rakudo's async and socket support might be a bit too unstable atm 10:43
punter ok
moritz punter: if you want something production ready quickly, p5 is still the safer better
punter thanks
atroxaper Hello, #perl6 ! 11:01
use Test; role Role { multi method gist() {return 'GIST';}}; class Class does Role {}; say Role.new; is Role.new, 'GIST', 'works well'
m: use Test; role Role { multi method gist() {return 'GIST';}}; class Class does Role {}; say Role.new; is Role.new, 'GIST', 'works well'
camelia rakudo-moar 25c1e5: OUTPUT«GIST␤not ok 1 - works well␤# got: 'Role<140591977373584>'␤# expected: 'GIST'␤»
atroxaper Why 'is' doesn't call gist? 11:02
moritz because .gist is only the default stringification for say() 11:04
the rest uses .Str
atroxaper Can I override Str? 11:04
Yes. Thanks. 11:07
vendethiel no problem :) 11:09
masak wonders what the biggest obstacle would be for unifying .gist and .Str 12:34
masak hello from a train, by the way :) 12:35
vendethiel whispers "I like trains"
tadzik WZOOM 13:00
dalek ecs: 03a91ba | (Elizabeth Mattijsen)++ | S32-setting-library/Containers.pod:
Spec Buffer.subbuf-rw
 13:01
Su-Shee *hihi* you're all going to love the ypac staff t-shirts: pbs.twimg.com/media/BqsH2TcCYAEHHlQ.jpg :)
tadzik :D 13:02
Su-Shee love it. want one of those :) 13:03
dalek kudo/nom: 49ab1d8 | (Elizabeth Mattijsen)++ | t/spectest.data:
Add advent2011/day20 tests
 13:07
lizmat masak: simple: make say use .gist, run the spectest, see what breaks
*.Str :-) 13:09
masak lizmat: worth a shot. 13:17
lizmat already running spectest 13:20
TimToady the intent of gist is to be a tl;dr 13:22
the intent of .Str is to be completist
this is a pragmatic distinction, not driven by the fact that there's a large overlap for short values
nor by the fact that the spec tests aren't good at testing pragmatics 13:23
TimToady so I don't see that it'll prove much 13:23
TimToady heads back to his talk-writing corner 13:24
lizmat masak: gist.github.com/lizmat/23162d0031578a2eeb56
lizmat does a git reset --hard 13:25
dalek ast: fedf27b | (Elizabeth Mattijsen)++ | S22-package-format/local (2 files):
Initial testing on CompUnitRepo::Local::File
 13:54
amirite @mxit 14:03
dalek kudo/nom: 9017529 | (Elizabeth Mattijsen)++ | / (4 files):
Add skeleton CompUnit class
 14:20
hoelzro I'm writing a module that allows users to declare methods as commands for a CLI, so I don't want the user to be able to declare duplicates 14:36
is there something I can hook into when the class is "closed" so I can detect this situation? 14:37
masak m: say Int.HOW.^methods 14:40
camelia rakudo-moar 49ab1d: OUTPUT«No such method 'gist' for invocant of type 'NQPRoutine'␤ in method gist at src/gen/m-CORE.setting:7190␤ in sub say at src/gen/m-CORE.setting:13769␤ in block at /tmp/SRwb3KffnC:1␤␤»
masak grrbl. 14:41
anyway, it seems to me you should be able to override .add_method or something in a custom ClassHOW.
hoelzro hmm 14:47
well, that would require the user to be aware of this, right?
this is what I'm going for: gist.github.com/hoelzro/d552753137568ac5285c
I was thinking I could monkey with the HOW of App::Subcommand to achieve my goal
masak yes, something like that was what I had in mind. 14:49
psch well, i got a tr/// that always returns True, instead of the previous False. it's spec'd - by omission - to return the number of replaced or deleted characters though 14:56
i can't seem to keep the original lhs for smart match calls for any comparison 14:57
atroxaper haw can I continue search in string by regex? Something like qq{ while 'abcd 15:12
atroxaper haw can I continue search in string by regex? Something like qq[ while 'abcd' ~~ m/ (.) {say $0;} ] 15:12
psch m: 'foo' ~~ m/ (.)+ { say $0 } / # this at least does your example 15:14
camelia rakudo-moar 49ab1d: OUTPUT«｢f｣␤ ｢o｣␤ ｢o｣␤␤»
psch although i'm pretty sure there's an XY-problem here somewhere
atroxaper psch: I think it's good enought :-) 15:15
psch glad to help :) 15:23
dalek ecs: 2d0d5ac | (Elizabeth Mattijsen)++ | S22-package-format.pod:
Some more elaboration about CompUnit class
 15:40
japhb .ask timotimo Did you write nqp/richards yourself, or does it come from somewhere else? (And if so, where? I want to make sure I have a proper copyright note for it.) 16:44
yoleaux japhb: I'll pass your message to timotimo.
japhb .ask timotimo Wait, did nqp/richards in perl6-bench ever actually work? I get some strange error messages from it on nqp-m .... 17:11
yoleaux japhb: I'll pass your message to timotimo.
masak atroxaper: I would just do `for $str.comb(/pattern/) -> $match { ... }` 17:24
lizmat lunch& 17:28
timotimo japhb: i ported it from the pypy benchmark suite 17:52
yoleaux 04:20Z <japhb> timotimo: New perl6-bench command `bench history` ought to interest you. Ping me if you need help. :-)
04:20Z <japhb> timotimo: html and html_plot versions coming when I get another good hack session. 17:53
16:44Z <japhb> timotimo: Did you write nqp/richards yourself, or does it come from somewhere else? (And if so, where? I want to make sure I have a proper copyright note for it.)
17:11Z <japhb> timotimo: Wait, did nqp/richards in perl6-bench ever actually work? I get some strange error messages from it on nqp-m ....
timotimo i think it worked at at least one point
but i didn't have the patience to port it
timotimo media.ccc.de/browse/conferences/gpn..._timo.html - this is my perl6 talk, in german and quite chaotic 17:57
dalek kudo/nom: 4459c54 | (Tobias Leich)++ | src/core/Buf.pm:
add unpack directives "a" and "Z"
japhb timotimo: Do you remember which nqp backend you used when you got it to work the first time? I'm trying to determine if it's just nqp-m that doesn't like it .... 18:14
timotimo hmm, depends on when i did it, maybe parrot, but possibly moar. 18:17
wait, do we even have mutable bufs? 18:21
colomon shift_pmc() not implemented in class 'Continuation' 18:23
in block at bin/panda:21
in sub MAIN at bin/panda:19
colomon has been having major problems smoking parrakudo the last few days.
timotimo o_O
colomon Tuesday the 17th was the last time I did a successful full smoke run. Took me a few days to notice because my laptop was at Apple for repairs. 18:26
[Coke] waves from I94. 18:32
[Coke] II actually have pretty decent internet, considering I'm on the highway. 18:35
dalek kudo-star-daily: b06f95b | coke++ | log/ (13 files):
today (automated commit)
 18:36
rl6-roast-data: 39dffd3 | coke++ | / (6 files):
today (automated commit)
colomon [Coke]: I94! That's like home!
colomon lives a couple hours away from it now, but grew up 30 minutes' drive from the east end of I94, and then lived only a mile from I94 for most of the Naughts. 18:37
[Coke] colomon: heading SW now, ooping around the lake. 18:38
colomon [Coke]: Lake Michigan, I presume?
[Coke] colomon: we were at lake erie earlier today.
and now we're at that one, yes.
colomon Great Lakes++ 18:39
afk # quick run into town, back very shortly
FROGGS__ .tell lizmat subbuf-rw is just a sub like substr-rw is only a sub, not a method 18:58
yoleaux FROGGS__: I'll pass your message to lizmat.
masak colomon: didn't nqp upgrade its Parrot dep just a few days ago? 19:00
colomon: ah, yes nqp's adf65bfd24f70a05459478397e5f54eb754bbfd1. 19:01
masak colomon: would be interesting to see if the problems go away if you build before that commit. 19:01
[Coke] no rakudo's testing clean today 19:05
*rakudos
dalek href="https://modules.perl6.org:">modules.perl6.org: 1076951 | (Tim Smith)++ | web/ (9 files):
Clearer badge icons
 19:10
colomon [Coke]: was that, "No, the rakudos are testing clean today," or "No rakudos are testing clean today"? 19:14
colomon masak: I'm trying to build bcd1861857db94a74afbd7ff6e43d87cc6878fe7 now (June 18th, before NQP bumps, I think) 19:16
masak just got another spesh fail, this time when running from -e 19:23
building latest and trying again.
colomon okay, bcd1861857db94a74afbd7ff6e43d87cc6878fe7 installs panda correctly. 19:29
vendethiel nice new badges :) 19:32
cognominal indeed 19:36
thou Good, I'm glad they are well-received 19:40
I am *not* a graphic designer, so I won't be offended if the community wants them reverted or tweaked. 19:41
cognominal thou, is there a pointer back from modules.perl6.org/ to the program that generated it? 19:45
thou cognominal: Yes, there's a pointer to github.com/perl6/modules.perl6.org repo 19:47
thou … at the bottom of the page 19:48
cognominal ho, I did not notice that contained the program as well as the data. 19:49
thou++
thou hey, ho! let's go!
colomon hmm, panda build works on parrot-d45fccdbaf74f13207bf9118a6dfd9d158108a3f too 19:52
lucas Hello. Is "my Dog $spot = $spot.new;" right or wrong? S02 says ".=" is rewritten into this but S03 say it's an invalid self-reference. 19:58
vendethiel p6: class A {}; my A $a = $a.new; 19:59
camelia rakudo-jvm 4459c5: OUTPUT«(timeout)»
( no output )
moritz lucas: looks correct to me 20:06
lucas: the variable is declared as soon as it's parsed
vendethiel lucas: where does S03 say it's an invalid ref ? 20:07
lucas S03-operators.pod: 5072: my Dog $fido = $fido.new; # wrong: invalid self-reference 20:09
lucas should it be changed? s/wrong: invalid/okay/ ? 20:10
colomon hmmm 20:11
psch makes sense to me, fwiw; as moritz says variables are declared when parsed, and typed variables get instantiated with the type object 20:12
moritz though caution is required
it might be that the variable on RHS is actually the one from the proto lexpad, not from the runtime lexpad
I'm not sure here
jnthn Good evening from the north o/ 20:13
moritz m: my Int $x = do { say $x; 42 }
camelia rakudo-moar 4459c5: OUTPUT«(Int)␤»
moritz \o jnthn
psch hi jnthn
colomon \o
moritz m: my Int $x = do { BEGIN { $x = 23 }; say $x; 42 }
camelia rakudo-moar 4459c5: OUTPUT«23␤»
moritz isn't sure whether that tells us anything 20:14
lizmat jnthn o/
yoleaux 18:58Z <FROGGS__> lizmat: subbuf-rw is just a sub like substr-rw is only a sub, not a method
lizmat and good * from the south
moritz guesses he is south from the north, but no idea how he compares to lizmat's south 20:15
lizmat moritz: as in Orlando, FL
moritz lizmat: oh. I'm guess I'm noth, then :-)
jnthn is almost in the arctic 20:16
It's still rather light outside. :)
TimToady last week at Key West the sun was pretty much straight up
lizmat TimToady: is there a reason we don't have substr-rw and/or subbuf-rw as a method ? 20:17
moritz lizmat: I think it's because sub{str,buf}-rw are mostly p5isms, and p5 doesn't do methods on value "types" 20:18
lizmat subbuf most definitely is new 20:19
but join is a p5ism, and we have a .join method
TimToady lizmat: the only reason is a foolish inconsistency :)
lizmat ok, so you won't mind if I make it consistent ?
masak jnthn \o
TimToady nope 20:20
vendethiel just learned he wouldn't be in france when YAPC::EU happens
jnthn Well, there is the thing that method calls on a scalar are normally seen as being on the value, not on the container...
lizmat vendethiel: that's because YAPC::EU is in Sofia?
vendethiel lizmat: haha, no, not in sofia either
I'm not here in august
TimToady jnthn: true
lizmat so you mean you can't make it to YAPC::EU :-(
vendethiel (my only holidays in 2 years :(.) 20:21
yea :(
lizmat jnthn, TimToady: $ ack '.substr' src/core | wc -l
103
actually:
$ ack '\.substr' src/core | wc -l
50
TimToady jnthn's point is that it would have to be a VAR-like macro in order to make it an lvalue 20:22
lizmat so substr is being used as a method in the core settings
TimToady .substr can just deal with the value rather than the container
TimToady strings being immutable, we emulate mutation in the container 20:23
lizmat right, so using it as a method, gives the wrong idea
TimToady (buffers are supposed to be mutable, of course) 20:23
so it actually more sense for bufs than for strs
moritz but, we should totally have the 4-arg substr as a method
$str.substr($from, $to, $replacment) 20:24
which just returns the new string
TimToady and you use .= to mutate
colomon is starting to suspect his parrakudo problems have something to do with a borked install 20:25
lucas p6: class A {}; my A $a .= new; say 'alive';
camelia niecza v24-109-g48a8de3: OUTPUT«Potential difficulties:␤ $a is declared but not used at /tmp/tmpfile line 1:␤------> class A {}; my A ⏏$a .= new; say 'alive';␤␤alive␤» 20:25
..rakudo-{parrot,jvm,moar} 4459c5: OUTPUT«alive␤»
lucas p6: class A {}; my A $a = $a.new; say 'alive';
camelia rakudo-{parrot,jvm,moar} 4459c5, niecza v24-109-g48a8de3: OUTPUT«alive␤» 20:26
lizmat TimToady: so yes to .subbuf-rw, but no to .substr-rw ? 20:30
dalek href="https://modules.perl6.org:">modules.perl6.org: 7f0ccaa | (Tim Smith)++ | web/readme.png:
Fix blurry readme.png icon
 20:39
href="https://modules.perl6.org:">modules.perl6.org: 40f0e18 | (Tim Smith)++ | web/ (4 files):
Give badge icons a bit more breathing room
dalek rl6-bench: e99a064 | (Geoffrey Broadwell)++ | / (4 files):
Add rc-self-describing-numbers minibenchmark
 20:48
masak $ perl6 -e 'for ^13 { for ^0 {}; (my $)++ }' 20:51
Spesh: failed to fix up handlers (134, -1, 156)
(locally, on a fresh build)
This is perl6 version 2014.06-11-g4459c54 built on MoarVM version 2014.05-2-gb1add28
nwc10 was about to go to bed so can't add anything to 20:52
nwc10 "doesn't (not) work on my machine" 20:52
This is perl6 version 2014.06-11-g4459c54 built on MoarVM version 2014.06-6-gd2e0732 20:53
./perl6-m -e 'for ^13 { for ^0 {}; (my $)++ }'
nwc10 see no output
vendethiel m: for ^13 { for ^0 {}; (my $)++ } 20:54
camelia ( no output )
jnthn masak: 2014.05-2? That's...strangely old, no?
Also, I'm sure I already fixed bugs giving that error... 20:55
nwc10 masak: my MoarVM seems to be 4 commits newer than yors
jnthn 4 commit *and a month* 20:56
nwc10 oh yes. oops
I am clearly not very useful and shall go to bed.
jnthn :)
nwc10 Good night world.
jnthn should do similar soon
lizmat goodnight nwc10, jnthn 20:58
lucas I'm still a little lost on the specs; can someone tell me where does it says that an our-scoped variable cannot have a type constraint? Example: "our Foo $foo;". Does this makes sense? 21:15
FROGGS__ lucas: I don't see why type constraints on our scoped vars should be forbidden... 21:21
lizmat I'm not sure, but maybe because 'our' variables are more like dynamic variables, and thus basically runtime determined?
FROGGS__ and... I can't find anything about that in the spec
lizmat FROGGS__: good to see you awake: 21:22
FROGGS__ hi lizmat
timotimo lizmat: er, what?
lizmat do we support use Foo:from<nqp> already?
and if so, where in the code is that handled?
FROGGS__ m: use QAST:from<NQP>; my Mu $q := QAST::Op.new(); say $q.HOW.name($q) 21:23
camelia rakudo-moar 4459c5: OUTPUT«QAST::Op␤»
FROGGS__ in src/Perl6/ModuleLoader, at the top of load_module
dalek kudo/nom: a7b4cba | (Elizabeth Mattijsen)++ | src/core/CompUnit.pm:
Some more work on CompUnit, not yet functional
 21:24
lizmat okidoki
lucas FROGGS__: I don't see why also. I don't know if it is specced or not. Rakudo died with a message "Cannot put a type constraint on an 'our'-scoped variable" 21:25
FROGGS__ lucas: right, but std does not
our Foo $foo; # niecza 21:26
n: our Foo $foo; # niecza
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Malformed our at /tmp/SU6kFuPnSj line 1:␤------> our⏏ Foo $foo; # niecza␤␤Parse failed␤␤»
FROGGS__ n: our Int $foo; # niecza
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Common variables are not unique definitions and may not have types at /tmp/Qyac0fM8y4 line 1:␤------> our Int $foo⏏; # niecza␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niec…»
lucas p6: class A {}; our A $a;
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Common variables are not unique definitions and may not have types at /tmp/tmpfile line 1:␤------> class A {}; our A $a⏏;␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/b…»
..rakudo-{parrot,jvm,moar} 4459c5: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Cannot put a type constraint on an 'our'-scoped variable␤at /tmp/tmpfile:1␤------> class A {}; our A $a⏏;␤ expecting any of:␤ scoped dec…»
FROGGS__ either niecza explains it better or both are wrong :o) 21:27
masak agreed; though I don't believe "Common variables" is common terminology :) 21:28
lizmat std: our Int $a = 42
camelia std 0f2049c: OUTPUT«ok 00:01 123m␤»
lizmat std thinks it's ok
also, seems like bfd8508f5ceee565380da201ef25bdfb6feb2b3d was the commit, by FROGGS__ in May last year
timotimo lizmat: i'm pretty sure we know all our-scoped stuff at compile-time, don't we? 21:29
lizmat well, I was clutching at straws, really
as I couldn't think of a reason
jnthn Not sure what's unclear about Rakudo's error. It tells you what you tried to do, and says it's not allowed. 21:30
masak std: sub foo { our Int $x }; sub bar { our Str $x }
camelia std 0f2049c: OUTPUT«ok 00:01 124m␤»
masak the above is the problematic bit.
FROGGS__ O.o
jnthn Also, STD is mostly about syntax, not about semantics. 21:31
So I'd not especially expect it to complain about this one.
masak *nod*
though it certainly could.
lizmat masak: but that could be handled as an error at compile time, no?
lucas isn't "our TYPE $var" a syntax thing? 21:32
jnthn lizmat: The error Rakudo gives today is compile time? Same for Niecza's.
lucas: No; it parses fine, just doesn't work out semantically.
lizmat no, I mean:
lizmat sub a { our Int $foo }; sub b { our Str $foo } 21:33
punter would donating to any Perl 6 body speed up its development?
jnthn m: sub a { our Int $foo }; sub b { our Str $foo }
camelia rakudo-moar 4459c5: OUTPUT«===SORRY!=== Error while compiling /tmp/50lgPw_8eQ␤Cannot put a type constraint on an 'our'-scoped variable␤at /tmp/50lgPw_8eQ:1␤------> sub a { our Int $foo ⏏}; sub b { our Str $foo }␤ expecting any of:␤…»
lizmat jnthn: I'm still not clear as to why we don't have types on "our" variables 21:34
aka:
FROGGS__ punter: I'm not sure... more hands could help though :o) (like module porters...)
lizmat m: our Int $a = 42;
camelia rakudo-moar 4459c5: OUTPUT«===SORRY!=== Error while compiling /tmp/rObavPtiYC␤Cannot put a type constraint on an 'our'-scoped variable␤at /tmp/rObavPtiYC:1␤------> our Int $a ⏏= 42;␤ expecting any of:␤ scoped declarator␤ …»
lucas I'm fine with Rakudo not accepting it. I just wish I had read about it in the spec somewhere.
jnthn lizmat: Too many semantic issues with supporting it.
punter FROGGS__, I ain't good with programming. Just have billions in the bank (j/k)
jnthn lizmat: Especially when you consider the binding side of things too. 21:35
FROGGS__ punter: *g*
punter FROGGS__, can't you hire people with money?
FROGGS__, that way, money would help
no eh?
ok
zengargoyle is there a $INPUT_RECORD_SEPARATOR equivalent? open("path", :nl("...")) doesn't seem to exist yet.
FROGGS__ punter: I just can speak for myself, and I already spend as much time as I can... I mean money is nice but it would not help me helping Perl 6 more 21:36
FROGGS__ zengargoyle: I remember that this is not yet implemented... would be nice to be proven wrong 21:36
lizmat punter: what would help, is paying someone like jnthn or masak for a course in Perl 6 21:37
FROGGS__ gnight #perl6, will read you tomorrow
lizmat like the "introduction to NQP" course
jnthn punter: A lot of people working on Perl 6 have other jobs too, and it's not always easy for them to put that aside. 21:37
lizmat and making sure the materials would be open source :-)
jnthn That is, if I want to turn more money into more Perl 6 time, it'll need some substantial life refactors as a pre-req... 21:38
lizmat (and paying jnthn and/or masak through their $work)
zengargoyle yeah, it doesn't throw NYI but seems to be missing. seems whenever i get an itch to do a little p6 i want something that's not there yet. :/
hoelzro I'm trying to play with the role machinery via Metamodel::ParametricRoleHOW.new_type(:name('MyRole')), but I get an error message when I try this. Is there something I'm missing? 21:39
punter ok I'm convinced. I won't donate, because I don't know how to do it.
there are no receivers here: www.gittip.com/for/perl-6/
neither froggs nor flies or jnthn
timotimo punter: it just means nobody is receiving money yet there 21:40
punter timotimo, are they receiving somewhere else?
TPF perhaps?
jnthn hoelzro: Well, what's the error? :) 21:41
hoelzro: The compiler assembles roles using that mechanism, so it must work *somehow* ;)
punter I'm going back to programming my jabber server
jnthn punter: TPF have managed Perl 6 related grant stuff, certainly.
punter oh ok
jnthn has taken grants in the past through TPF
timotimo punter: also, you can select gittip to not show you in receivers lists at all 21:42
hoelzro jnthn: Unhandled exception: Could not instantiate role 'MyRole':
I'm wondering if I have to set something up before I call new_type
jnthn hoelzro: Hmmm...not something ending up sink-contexting it?
timotimo punter: a good choice for donations is money to help people attend conferences and hackathons
hoelzro jnthn: here's the source: gist.github.com/hoelzro/5bfb326cd980070bacb4 21:43
jnthn m: my $r := Metamodel::ParametricRoleHOW.new_type(:name('MyRole')); say 'ok'
camelia rakudo-moar 4459c5: OUTPUT«ok␤»
hoelzro ah
binding, not assignment, maybe
jnthn hoelzro: Last line of the program is sunk, I think.
m: my $r = Metamodel::ParametricRoleHOW.new_type(:name('MyRole')); say 'ok'
camelia rakudo-moar 4459c5: OUTPUT«ok␤»
hoelzro huh
sure enough 21:44
thanks jnthn =)
jnthn np
OK, I got teaching tomorrow, and despite the fact that night time doesn't happen in this part of the world this time of year, I need some sleep :)
'night, #perl6
lizmat gnight. jnthn
hoelzro 'night jnthn
lucas Is there a command line tool that works just like this IRC bot, where I pass my program (file or -e) and it executes in all implementations and joins the output? Such a program already exist? 21:49
masak lucas: not to my knowledge. but it's a pretty great idea. 21:51
lizmat lucas: the bot code is public, so could be a source of inspiration :-) 22:06
lucas lizmat: yes!! I'll seach its source code in the repos... 22:07
*search
lizmat :-) 22:08
timotimo should be called "evalbot"
lizmat AntiArrivalDinner&
masak 'night, #perl6 22:14
hoelzro night masak 22:25
is it possible to do EXPORTHOW magic without creating a separate file?

Powered by Cro and the Raku® Programming Language.

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

Thank you!