»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'p6: say 3;' or rakudo:, std:, or /msg camelia p6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by masak on 12 May 2015.
00:04 dp left
garu El_Che: have you seen github.com/tokuhirom/p6-Getopt-Tiny? I'm not sure it does exactly that, but it could :) 00:09
ShimmerFairy I saw a fleeting mention once somewhere in the docs about being able to replace the MAIN parser. But in any case, I don't think supporting '--option value' as an alternate syntax should be too much of a stretch for the default parser. 00:10
00:17 dp joined 00:24 laouji joined 00:25 Peter_R left 00:27 aborazmeh joined, aborazmeh left, aborazmeh joined 00:36 laouji_ joined 00:39 laouji left 00:40 laouji_ left 00:47 laouji joined 00:49 mr_ron left 00:53 leont left, dp left 00:54 kuzgun left
lichtkind garu++ 00:57
timotimo El_Che: you can also preprocess the @*ARGS before your mainline runs to completion to fix up "--option value" pairs into "--option=value" pairs 01:11
ShimmerFairy timotimo: would it be too much to support both by default? I've seen plenty of programs do so. 01:19
I just noticed that short options require = too in P6, which seems especially weird. (I'd sooner try -ovalue or -o value instead of -o=value) 01:20
01:36 laouji_ joined, tokuhirom joined 01:41 laouji_ left, tokuhirom left 01:44 vendethiel joined
[Coke] tries to convert the profiler html from jquery-bootstrap to angular-bootstrap. boom. 01:47
01:57 mjp__ joined 01:58 laouji left, laouji joined 02:06 laouji_ joined, aborazmeh left 02:08 vendethiel left 02:10 lichtkind left, laouji_ left 02:12 kaare__ joined 02:29 noganex_ joined 02:31 colomon left 02:32 noganex left 02:33 kid51 left 02:54 colomon joined 03:18 diakopter left
dalek p: e522c72 | coke++ | src/vm/moar/profiler/template.html:
use angular-bootstrap, not jquery-bootstrap

We're already using angular - use the ui.bootstrap module for all our bootstrap needs; avoids loading two frameworks simultaneously.
03:29
03:29 diakopter joined
[Coke] un-booms it. 03:30
03:32 travis-ci joined
travis-ci NQP build failed. Will "Coke" Coleda 'use angular-bootstrap, not jquery-bootstrap 03:32
travis-ci.org/perl6/nqp/builds/83817866 github.com/perl6/nqp/compare/72b28...22c72da795
03:32 travis-ci left 03:38 aborazmeh joined, aborazmeh left, aborazmeh joined, BenGoldberg left 03:43 laouji left 03:46 tokuhirom joined 04:17 atweiden left 04:21 laouji joined 04:26 laouji left
zostay_ m: :16<deadbeef> 04:38
camelia ( no output )
zostay_ m: :16<deadbeef>.say
camelia rakudo-moar 68f233: OUTPUT«3735928559␤»
zostay_ m: :16<deadbeef>.perl.say 04:39
camelia rakudo-moar 68f233: OUTPUT«3735928559␤»
zostay_ m: :16<deadbeefdeadbeefdeadbeefdeadbeef>.perl.say
camelia rakudo-moar 68f233: OUTPUT«295990755083049101712519384020072382191␤»
zostay_ m: :16<deadbeefdeadbeefdeadbeefdeadbeef>.WHAT.say
camelia rakudo-moar 68f233: OUTPUT«(Int)␤»
zostay_ m: :16{deadbeef}.say 04:40
camelia rakudo-moar 68f233: OUTPUT«5===SORRY!5=== Error while compiling /tmp/Bhjh0UAGl4␤Malformed radix number␤at /tmp/Bhjh0UAGl4:1␤------> 3:167⏏5{deadbeef}.say␤ expecting any of:␤ number in radix notation␤»
zostay_ are blob literals implemented at all?
ShimmerFairy not currently 04:42
04:43 laouji joined 04:47 laouji left, laouji joined, Jaglor left 04:56 khw left 05:01 mattp__ left, mattp__ joined 05:02 kaare_ joined 05:03 oka_ left, oka_ joined, grondilu left, adhoc left, grondilu joined, kaare__ left, cosimo left 05:04 ssflying left, shmibs left, shmibs joined, sjn joined, ssflying joined 05:05 adhoc joined, skids left, cosimo joined 05:30 ][Sno][ left 05:33 xfix joined, xfix left, xfix joined 05:40 bjz joined 05:44 bjz left 05:51 llfourn joined 05:54 ][Sno][ joined, rurban joined 05:55 aborazmeh left 05:56 domidumont joined 05:57 aborazmeh joined, aborazmeh left, aborazmeh joined 06:00 domidumont left 06:01 domidumont joined 06:02 laouji left 06:03 laouji joined
[Tux] test 50000 38.657 38.547 06:14
test-t 50000 41.195 41.085
and Inline::Perl5 didn't build
tux.nl/Talks/CSV6/speed4.html
06:14 diana_olhovik_ joined 06:20 FROGGS left 06:21 Calibellus left, laouji_ joined 06:25 laouji_ left 06:27 domm_ left 06:29 domm joined 06:31 FROGGS joined 06:33 darutoko joined 06:35 Ven joined 06:36 dylanwh_ left
FROGGS masak: the precedece trait in the example existed in the old example (with the unicode "!") as well... I jusr re-borrowed it 06:36
Ven hey 06:37
examples.perl6.org/categories/best-...ctory.html <- seems like the code doesn't render. maybe someone has a clue? :)
FROGGS Ven: can you file an issue? 06:41
Ven FROGGS: what's the repo for examples.perl6.org? 06:42
ShimmerFairy Ven: not sure, but it's certainly under the perl6/ organization, and probably has a sensible name :) 06:43
FROGGS Ven: it shows it in the yellow box at the bottom: github.com/perl6/perl6-examples
Ven heh? I typed "examples" in the github.com/perl6 "search" box, and it gave me no result... -.-
thanks FROGGS.
moritz \o 06:44
FROGGS Ven: we tend to put a link to the repos at the bottom of all our sites... to invite ppl to make small changes
Ven o/ moritz
FROGGS morning moritz
06:45 domm left 06:46 domm joined
Ven WRT the website: I've seen some critics (here and on several websites) say that the code examples aren't all that telling 06:46
I'd like to say (yeah, I'm kinda repeating myself) that I think racket-lang.org/ has a nice solution here: an "explain" model, and each function(/method) used is a link to its documentation. What would people think about this? 06:47
FROGGS IIRC the perl(5)docs also do that
so, +1
Ven s/model/modal/ 06:48
FROGGS: you're right! That's good. (I admit I didn't spend that much time in the online 5 docs :)..) 06:49
Do we really have an infix:< >? :P irclog.perlgeek.de/perl6/2015-10-05#i_11322893 06:51
06:52 spider-mario joined
FROGGS Ven: maybe we could use it to concat barewords into strings? 06:52
Ven .tell ely-se I'm more likely to receive message under "ven", fwiw 06:54
yoleaux Ven: I'll pass your message to ely-se.
06:54 chenryn joined
FROGGS m: sub infix:< >($a, $b) { "$a\x20$b"} # sad 06:54
camelia rakudo-moar 68f233: OUTPUT«5===SORRY!5=== Error while compiling /tmp/3KxVvtQJ4U␤Null operator is not allowed␤at /tmp/3KxVvtQJ4U:1␤------> 3sub infix:< >7⏏5($a, $b) { "$a\x20$b"} # sad␤»
Ven FROGGS: *g* 06:55
06:56 rindolf joined
Ven .tell ely-se WRT comefrom in perl6 – maybe ask masak about it :) 06:57
yoleaux Ven: I'll pass your message to ely-se.
06:59 chenryn left, chenryn joined 07:02 dylanwh_ joined 07:05 laouji_ joined 07:08 chenryn left 07:10 laouji_ left 07:17 bjz joined 07:18 ely-se joined 07:26 larion left, AlexDaniel left 07:28 pdcawley joined 07:32 CIAvash joined
ely-se morning 07:39
yoleaux 06:54Z <Ven> ely-se: I'm more likely to receive message under "ven", fwiw
06:57Z <Ven> ely-se: WRT comefrom in perl6 – maybe ask masak about it :)
ely-se hmm 07:40
dalek line-Perl5: 38241c4 | (Stefan Seifert)++ | lib/Inline/Perl5.pm6:
"postcircumfix:<( )>" is now spelled "CALL-ME"

Fixes tests on current rakudo
07:44
nine [Tux]: ^^^
[Tux]: thanks for reporting!
moritz sub CALL-ME-MAYBE(\c) 07:45
[Tux] and what about the speed regression?
moritz [Tux]: how big is it? how big is the standard deviation of your runs? 07:46
[Tux] tux.nl/Talks/CSV6/speed4.html
moritz [Tux]: that might be noise, or it might not be
[Tux] 38 -> 41 is over the noise level 07:47
07:47 zakharyas joined, aborazmeh left
[Tux] we've been under 36 07:48
[Tux] is rebuilding
07:50 Ven left
moritz [Tux]: so what's the noise level? 07:50
[Tux] 2%?
39 and up is a warning sign to me 07:51
lizmat fwiw, bare startup time has gone down for me from 115 ms to 108 in the past days 07:54
spectest below 210 secs for the first time in a week 07:55
but still a lot slower than the <100 msecs I've seen before
nine Japanese really are fast with everything. Even slides are visible less than a second :) 07:57
07:57 [Tux] left
lizmat there's a name for that style... but I forgot which one 07:58
DrForr Takahashi, I think.
nine When he showed the Java GUI cod example and the awesomeness came bam, bam, bam, one could feel the wow in the room even in the video :) 08:00
lizmat gist.github.com/lizmat/2bf702c8c08587c3a231 # makes Optimizer infiniloop :-( 08:01
nine lizmat: I wonder if we should keep DEPRECATED as public interface? It seems to be very useful for module authors 08:03
08:03 Ven joined
masak morning, #perl6ers 08:03
lizmat ah, perhaps...
it would solve the problem at hand :-)
08:04 mattprelude left 08:06 cibs left
masak I think the `is tighter` trait shouldn't be part of the postfix:<!> example. if we're aiming for simplicity, it's making things needlessly complicated. also (arguably), it's wrong. 08:06
nine lizmat: in the weekly, shouldn't it be "where an Array would get popped off @b"? 08:07
08:07 cibs joined, sftp left 08:09 chenryn joined, rindolf left 08:10 [Tux] joined
lizmat nine: maybe even "popped off of @b" ? 08:10
Ven o/ masak
lizmat nine: keeping it at "popped off" now
nine++
dalek href="https://perl6.org:">perl6.org: a6a9749 | (Carl Masak)++ | source/index.html:
remove `is tighter` on custom operator

This may be highly subjective, but to me postfixes are naturally tighter than prefixes. In other words, I'd expect `-1!` to mean
  `-(1!)`, not `(-1)!`. Compensated a test accordingly.
08:14
08:14 sftp joined, chenryn left
DrForr It's a pity gamma() isn't a core method :) 08:17
lizmat .tell jnthn is there a runtime downside to a sub called from CORE, and a sub called from Rakudo::Internals:: ? 08:18
yoleaux lizmat: I'll pass your message to jnthn.
08:19 rurban left
lizmat .tell jnthn if so, would it help to make the sub an only ? 08:19
yoleaux lizmat: I'll pass your message to jnthn.
masak Ven: \o 08:20
DrForr: gamma() ?
DrForr factorial for real numbers.
(well, gamma(x) == factorial(x-1) for nonnegative integers.) 08:21
ShimmerFairy surely Γ() though :)
DrForr Yes, but I couldn't be arsed to find the compose-keystroke. 08:22
ShimmerFairy masak: I agree with removing 'is tighter', but not necessarily the reasoning. That seems more a question of associativity than precedence.
08:24 Ven left
dalek kudo/nom: c981530 | lizmat++ | src/core/ (3 files):
Move THE_END to Rakudo::Internals
08:25
08:25 dakkar joined 08:26 rurban joined
dalek Heuristic branch merge: pushed 31 commits to rakudo/curli by lizmat 08:26
08:27 xfix left 08:30 llfourn left 08:33 skarn joined
lizmat afk for a few hours& 08:33
Woodi hallo today :) 08:36
long time ago there was a mantra here that "core do not have any ...". I wonder how edges are placed ATM... 08:37
08:37 Ven joined 08:38 kaare_ left
nine lizmat: moving THE_END to Rakudo::Internals makes a sad day for Perl 6 poetry ;) 08:38
08:41 maziar joined
masak ShimmerFairy: definitely not a question of associativity. 08:42
maziar hi i want to write script to do snmp query to device , is there any example ?
masak ShimmerFairy: recall that associativity only kicks in when two operators are on the same precedence level (and next to each other). there are no such operators here. 08:43
ShimmerFairy: I give you partial points on pattern-matching the situation, though. half of the time, people *do* say "precedence" when they mean "associativity".
ShimmerFairy masak: yeah, it's just that your reasoning in your commit message didn't make much sense to me, since the snippets looked to be showing associativity 08:44
nine maziar: mullagainn did some work in that direction
masak ShimmerFairy: it's not.
ShimmerFairy: if you don't believe me, check the precedence table in S03.
maziar nine what ?
ShimmerFairy masak: sure, I'm just saying that the paren'd stuff seemed more associativity to me. (Maybe because I've thought about rakudo's bogus 'unary' associativity lately?) 08:45
nine maziar: irclog.perlgeek.de/perl6/search/?ni...amp;q=SNMP
08:45 [particle] left
masak ShimmerFairy: then you need to update your perception of what's precedence and what's associativity against the normal use of those terms ;) 08:46
08:47 larion joined
ShimmerFairy masak: I think you're misunderstanding me. I'm saying the -(1!) examples didn't make sense at first because they look like talking about unary associativity to me, because that's been on my mind lately. 08:47
jnthn morning, #perl6
ely-se: Thanks for the link, was a fun read :) 08:48
masak ShimmerFairy: it would only be unary associativity if prefix:<-> and postfix:<!> were on the same precedence level. which they're not.
ShimmerFairy masak: yes, but as I've said a couple times now, it *looked* like something different to me at first, because that kind of snippet has been on my mind as about associativity rather than precedence lately. 08:49
masak ely-se: should be able to implement COMEFROM with visitor macros and goodenough goto, yes.
ely-se coooool 08:50
jnthn .tell lizmat The startup speed improvement was due to a fix I did yesterday (2997e4db98)
ely-se although I think an INTERCAL slang would be a better idea
masak ShimmerFairy: if it looked like something else to you, then you need to adjust your perception so that it doesn't decieve you so easily
you seem to be suffering from associativity hallucinations
08:50 yoleaux left, RabidGravy joined
ShimmerFairy masak: I don't have a perception problem, for christ's sake. I wasn't thinking about precedence differences at first, and the last time I saw syntax like that was looking at the unary associativity table in S03, hence my initial connection and confusion. 08:51
masak *nod*
so this entire tangent we went on, is because something in my commit *reminded* you of something? 08:52
thanks, now I know.
08:52 Ven left
masak your original sentence -- "That seems more a question of associativity than precedence." -- sounded to me like an assertion that it *was* associativity, not that it *felt to you* like associativity. 08:54
08:54 Kcop left
ShimmerFairy It's because my brain initially saw it as a talk about associativity, because that table in S03 was the last place I saw such a comparison. So I was initially confused because the words were talking about precedence, but your example snippets looked like associativity examples to me. 08:54
masak and yet you maintain that you're not confused about what associativity looks like ;) 08:55
mathw o/ masak
masak \o mathw 08:56
jnthn
.oO( false associations about associativity )
08:56 Loren_ joined
ShimmerFairy masak: I'm not, however you seem to believe that the !($a!) syntax could _only_ be about precedence, and never about associativity. 08:57
masak ShimmerFairy: earnest recommendation: implement an expression parser. makes things really really clear.
ShimmerFairy: then we probably wouldn't be having this discussion at all.
ShimmerFairy: no, I'm just saying that in the case of my commit, it *wasn't* about associativity, and never could be. 08:58
ShimmerFairy masak: and my brain saw those snippets as being about assoc because that's the last time I saw that kind of syntax being used, and because "- and ! are inherently of different precedences" wasn't in my mind. For all I knew, the 'is tighter' could've put them on the same level. 09:00
masak from my point of view, you've been *distracted* by what you saw in S03, and you feel compelled to tell me how my commit reminded you of associativity, when it has nothing to do with associativity.
jnthn [Coke]++ # profiler feels faster :)
ShimmerFairy So I was confused because my brain saw what looked like assoc examples in the midst of a precedence discussion, which is why I brought it up in the first place.
jnthn Well, profiler UI does :)
masak ShimmerFairy: got it. 09:01
09:01 Xor_ joined
masak ShimmerFairy: now you need to update your understanding of associativity and precedence ;) 09:01
ShimmerFairy: otherwise, this might happen again.
09:02 larion left, larion joined
ShimmerFairy masak: not really though, the fact that rakudo has been cheating with a 'unary' assoc is no reason my initial assumption was automatically invalid. 09:02
masak: at best I just needed to magically remember that prefix:<-> and a custom postfix:<!> are inherently on different precedence levels. 09:03
masak that's by spec, though
ShimmerFairy "custom" means it's not by spec :)
masak no, the spec tells you where custom ops go in the prec table
ShimmerFairy really? Where?
09:04 espadrine joined
masak S06 09:04
in the discussion about 'is tighter' et al
jnthn Hm, also looking at a latest startup profile tells me that the patch I did yesterday to fix pseudo-package introspection managed to undo some lazy deserialization goodness 09:05
09:05 laouji_ joined
ShimmerFairy masak: I only see one mention on postfixes about "since equiv<++> is the default", but aside from that I don't see any confident statement about default precedences 09:09
ah, I see now, the list above that.
09:10 chenryn joined, laouji_ left
masak aye 09:10
ShimmerFairy That's far too buried for my taste, I don't think I can be reasonably expected to have stumbled upon that :) (since I've never sat down and carefully studied the entire synopses)
It's just an unadorned code block, and the mention of what it is even further buried in the paragraph above it. Not really conducive to someone wondering about default precedences and whatnot. 09:11
masak once you're done fuming about it, please update your understanding about precedence and associativity ;) 09:12
ShimmerFairy masak: my understanding doesn't need changing, it's just a case of the last time I saw the !(a!) vs. (!a)! comparison, in the unary assoc table. 09:14
09:14 chenryn left
masak ShimmerFairy: you have a surface understanding of matters, which makes you look at the *syntax* and go "my! that look slike associativity to me!" 09:16
ShimmerFairy: you can't do that. that's a category error. that's wrong. you need to update your understanding so that you don't do that.
09:16 laouji left
Ulti is Peter Sergeant on here? 09:17
masak ShimmerFairy: whether something is or isn't associativity isn't a detail you find in the syntax. you can't reasonably do what you tried to do. that detail is in the precedence table, not in the syntax.
Ulti was just wondering what the plans are for perl6.io :3
ShimmerFairy masak: my understanding of precedence vs. associativity has nothing to do with it, just the fact that I saw the syntax generally because I didn't know the magic defaults of custom postfix ops versus prefix:<->
masak you made a series of assumptions, yes. they were wrong. 09:18
let's move past this. preferably in a learning direction. 09:19
if you sincerely believe your understanding of prec/assoc has nothing to do with it, then I'm not sure I can be of further assistance to you anyway. 09:20
09:21 laouji joined, laouji_ joined, mprelude left
ShimmerFairy masak: it really isn't, it was just a mistaken first idea of what the syntax was demonstrating (again, due to the assoc table being the last place I saw it), and soon after I realized that it was, in fact, completely coherent with the rest of the commit. 09:24
09:24 larion left
masak *plonk* 09:24
09:25 [particle] joined 09:26 maziar left, laouji left
dalek kudo/nom: 8709352 | jnthn++ | src/vm/moar/ops/perl6_ops.c:
Look at correct frame to see if we're in a PRE.

Fixes incorrect running of POST/LEAVE when a PRE fails.
09:26
ast: e284033 | jnthn++ | S04-phasers/pre-post.t:
Unfudge test for RT #121531.
ShimmerFairy I'm honestly baffled you think I have some deficiency in knowledge (aside from the category defaults; thanks for pointing that out, I've wondered about it sometimes), and don't seem to get that I just made a bad connection with the last time I saw *(a*) vs. (*a)* in text on operator stuff. 09:27
09:28 CIAvash left 09:30 Psyche^ joined 09:32 Psyche^_ left 09:36 laouji_ left, laouji joined 09:40 Xor_ left, Xor_ joined 09:41 laouji left 09:46 larion joined 09:47 Ven joined, Ven left 09:51 larion left
lizmat .botsnack 09:53
ely-se :D
09:53 rurban left
lizmat hmmm... yoleaux is AWOL ? 09:53
09:55 larion joined
grondilu searchs AWOL in S99 and finds nothing 09:56
ely-se away without leave
09:56 pmurias joined
ely-se duckduckgo.com/?q=!ud+awol 09:56
grondilu yeah fpund it eventually 09:57
ely-se ! commands is the only reason I use DuckDuckGo :D
09:57 Ven joined
lizmat jnthn: I asked via yoleaux at irclog.perlgeek.de/perl6/2015-10-06#i_11324705 09:58
masak ely-se: yes, ! commands rock 09:59
lizmat whether there's a runtime penalty for calling our subs in other packages, like Rakudo::Internals
10:00 maziar joined
jnthn lizmat: Will "make it an only" help: almost certainly not 10:00
Runtime penalty? Not sure.
If Rakudo::Interals were a class with methods on it, then the answer is easily "no"
lizmat I'd like to know before turning things like RETURN into a Rakudo::Internals::RETURN
10:01 laouji joined, domm_ joined
lizmat (I mean: RETURN-LIST) 10:02
and things like THROW :-) 10:03
10:04 domm left
Ven notes the hugless discussion and shrugs :( 10:05
10:07 maziar left, maziar joined
maziar is ther any sample script to do snmpwalk to a host ? 10:08
10:08 andreoss joined 10:11 chenryn joined
_itz maziar: there isn't UDP support yet 10:12
10:15 laouji_ joined, chenryn left 10:17 andreoss left
lizmat jnthn: I did some benchmarks: calling an our sub from another package, is about 19x slower :-( 10:17
jnthn lizmat: Ouch
lizmat I think I'll halt my "moving to Rakudo::Internals" actions for now
jnthn lizmat: And if you make Rakudo::Interanls a class and make them methods?
lizmat $ 6 'our sub a { 42 }; for ^100000 { a }'
real0m0.276s
$ 6 'module A::B { our sub a { 42 } }; for ^100000 { A::B::a }' 10:18
real0m3.400s
El_Che garu: timotimo : about my cmd param question. Both your suggestion are certainly valid, but I think the *default* of argument parsing should look UNIX native. Does that make sense?
jnthn lizmat: We have a little bit of an issue with our subs in packages in that we kinda have to respect rebinding, I *think*
lizmat: Which prevents spesh-time resolution/inlining 10:19
lizmat $ 6 'class A::B { method a { 42 } }; for ^100000 { A::B.a }'
real0m0.324s
jnthn And if you mark the class "is hidden"?
10:19 laouji_ left
lizmat no difference 10:19
ilmari huh, don't sub calls get resolved at compile time?
lizmat $ 6 'class A::B is hidden { method a { 42 } }; for ^100000 { A::B.a }' 10:20
jnthn heh, our optimizer is too clever :)
lizmat real0m0.317s
jnthn Yeah, not much
Leixcal sub calls do
lizmat within noise
jnthn *nod*
10:20 bjz left
jnthn m: package A { our sub foo() { 1 } }; A::foo(); A::<&foo> := { 2 }; A::foo(); 10:21
camelia ( no output )
10:21 laouji_ joined
jnthn m: package A { our sub foo() { 1 } }; say A::foo(); A::<&foo> := { 2 }; say A::foo(); 10:21
camelia rakudo-moar 870935: OUTPUT«1␤2␤»
maziar when i run this script i get this error : Use of uninitialized value in concatenation (.) or string at ./x.perl line 12.
DrForr maziar: perl6 doesn't use '.', try one of the other perl channels. 10:22
jnthn lizmat: I think the above needs to work. We can get clarification from TimToady on that I guess
dalek rl6-most-wanted: 063ff07 | (Tokuhiro Matsuno)++ | most-wanted/modules.md:
Update modules.md
ilmari maziar: which script? and that's a perl5 error
DrForr (well, it doesn't use '.' for *that* purpose.)
10:23 rurban joined
lizmat jnthn: that's the other thing: if I make Rakudo::Internals a package, all sorts of weird things happen :-( 10:23
but settings won't compile 10:24
lemme try that again to be sure
10:24 Ven left
jnthn lizmat: Well, I used package, but would be same with module 10:25
lizmat hmmm.. it builds now
10:26 laouji_ left
lizmat yeah, if I change the stub to package as well, I get: 10:28
Stage parse : Error while constructing error object:Could not locate compile-time value for symbol X::Package::Stubbed
Error while compiling, type X::Package::Stubbed
packages: (unstringifiable object)
at line 32407, near ""
which is basically a weird way of saying that it didn't see some package be filled out when stubbed 10:29
weird, because Exception *does* have an X::Package::Stubbed it should now about
so, changing it to a module fixed that
10:29 andreoss joined
dalek p: e8e9b08 | (Pawel Murias)++ | src/vm/js/QAST/Compiler.nqp:
[js] when we don't have an outer context use null
10:31
p: 7ae59c9 | (Pawel Murias)++ | / (3 files):
[js] Rewrite the test running script in Perl 5.

That should allow detecting the node version and skipping some tests that fail on travis.
p: 09ae178 | cygx++ | src/vm/jvm/runtime/org/perl6/nqp/io/FileHandle.java:
Implement extended file open modes on JVM
10:32
p: c2610de | cygx++ | src/vm/js/nqp-runtime/io.js:
[js] Minimal patch to make basic file open modes work again
p: 867e81b | jnthn++ | src/vm/j (2 files):
Merge pull request #245 from cygx/openmodes-jvm

Implement extended file open modes on JVM
10:33 bjz joined 10:36 travis-ci joined
travis-ci NQP build passed. Pawel Murias '[js] Rewrite the test running script in Perl 5. 10:36
travis-ci.org/perl6/nqp/builds/83863172 github.com/perl6/nqp/compare/e522c...e59c97bdb8
10:36 travis-ci left 10:37 ely-se left
jnthn spectests his patch to address one more of the xmas RTs 10:37
10:38 travis-ci joined
travis-ci NQP build passed. Jonathan Worthington 'Merge pull request #245 from cygx/openmodes-jvm 10:38
travis-ci.org/perl6/nqp/builds/83863246 github.com/perl6/nqp/compare/7ae59...7e81b4d68c
10:38 travis-ci left 10:46 pmurias left 10:48 pmurias joined
dalek kudo/nom: 0689733 | lizmat++ | src/core/ (6 files):
Make Rakudo::Internals a class with methods

This should be about 18x as fast then calling our subs
10:49
lizmat bbiab 10:50
masak "xmas RTs" ;)
10:50 laouji_ joined
dalek p: 48faa01 | jnthn++ | tools/build/MOAR_REVISION:
Bump MOAR_REVISION for UTF-8 BOM handling.
10:51
10:51 rindolf joined
dalek kudo/nom: 1ee7a96 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for UTF-8 BOM, JVM I/O bits.
10:53
ast: e2a621c | jnthn++ | S16-io/bom.t:
Add tests for UTF-8 BOM stripping.
10:54
10:56 laouji_ left
dalek kudo/nom: 1f7a564 | cygx++ | src/core/IO/Handle.pm:
Enable extended file open modes on the JVM
10:58
kudo/nom: 462ab13 | jnthn++ | src/core/IO/Handle.pm:
Merge pull request #545 from cygx/openmodes-jvm

Enable extended file open modes on the JVM
osystem: a4af5b9 | moznion++ | META.list:
Add Log::Minimal
osystem: 2ca19ca | (Zoffix Znet)++ | META.list:
Merge pull request #66 from moznion/log-minimal

Add Log::Minimal
kudo/nom: a26ecff | jnthn++ | t/spectest.data:
Run S16-io/bom.t.
10:59
11:04 kid51 joined 11:08 grondilu left 11:09 kid51 left 11:11 Ven joined, chenryn joined 11:16 chenryn left 11:20 laouji_ joined 11:21 TEttinger left, ely-se joined
jnthn .ask TimToady What is the <colonpair> for in github.com/perl6/std/blob/master/STD.pm6#L2032 ? I've been looking through the synopses for some info on it and failed to find anything... Currently Rakudo parses it but has no clue what to do with it, and ends up with an internal compiler error... 11:21
11:23 telex left
jnthn TimToady: Unless you have a strong idea of what it's meant to do, I'd vote for tossing it for 6.chirstmas. 11:24
11:24 telex joined 11:25 laouji_ left
ShimmerFairy jnthn: is that meant to allow something like class Foo :colon<pair> { , wherever trait_mods usually show up? 11:27
11:27 llfourn joined
jnthn ShimmerFairy: yeah 11:27
Well
It parses that
But then it's like...what do we do with this? :)
ShimmerFairy yeah, I'm not sure what it might do either :P
jnthn I mean, it could feasibly be a shortcut for is 11:28
sub foo($a :rw) { }
ShimmerFairy jnthn: blame says that was last touched in 2009, was there some sort of annotation mechanism or anything back then?
jnthn Well, traits certainly existed
It's parsed as part of those
My suspicion is we're going to end up tossing it from Rakudo's grammar, unless TimToady++ has a reason to want us to keep/fix it. 11:29
ShimmerFairy right, for some reason the first thing in my mind was "oh, I wonder if you're were able to annotate stuff", like QAST's associative side (IIRC) :)
nine We can always add it back later without back comp issues 11:31
11:32 llfourn left
ShimmerFairy m: sub foo() :colon<pair> { } 11:32
camelia rakudo-moar a26ecf: OUTPUT«5===SORRY!5=== Error while compiling /tmp/1FEg6MG4bN␤Cannot invoke this object␤at /tmp/1FEg6MG4bN:1␤»
jnthn nine: Aye 11:34
11:38 nuc joined 11:40 Ven left
jnthn is looking into RT #113544, which leads down into the bowels of the regex compiler... 11:40
11:41 Lorphos joined
Lorphos Hi. There was an example for a custom unicode operator on the perl6 homepage yesterday, today it has been replaced. Where can I still find it? 11:42
jnthn Yeah, it doesn't have a clue about nameds or flattens in those arglists...
11:43 ely-se left
Lorphos Got it, google cache still had the old copy 11:43
jnthn Ah, phew. Also in the commit history for the site I guess: github.com/perl6/perl6.org/commits/master
lunch time &
11:44 babe joined
ShimmerFairy .tell jnthn worth pointing that I've noticed @var := nqp::list() fails in main Perl 6 code, but works in regexes ( :my @var := nqp::list(); ). There's a good chance stuff like that is being handled by NQP rather than rakudo 11:45
jnthn m: use nqp; say 'a' ~~ /:my @x := nqp::list(); a/ 11:46
camelia rakudo-moar a26ecf: OUTPUT«Type check failed in binding; expected Positional but got List␤ in block <unit> at /tmp/ZJFfxPALBq:1␤␤»
jnthn Looks pretty failing to me
We delegate to the Perl 6 parsing and code-gen for vars, iirc. 11:47
jnthn -> oven :)
masak Lorphos: we changed it because the thing it was implementing (factorial) doesn't really need a Unicode character. much more approachable with just '!'
Lorphos: having said that, Perl 6 still works very nicely with Unicode operators, and Unicode identifiers in general
btyler approachable++ for frontpage examples 11:48
Lorphos i thought it was pretty cool. Frightening but cool :)
btyler unicode stuff is awesome, but yeah, frightening is the word
doesn't help the 'line noise' meme too much :) 11:49
(the new page looks _awesome_, btw)
Lorphos yup. "Perl 6 now has an infinite new number of unicode characters that will let you create obfuscated code!"
masak Lorphos: we want to be cool when possible and frightening only when necessary ;)
ShimmerFairy m: use nqp; "A" ~~ / :my @*A := nqp::list(); /
camelia ( no output )
Lorphos I like the page as well. It's not the same old boring stuff you see everywhere. 11:50
and it is highly informative while being entertaining
ShimmerFairy m: use nqp; my @*A := nqp::list()
camelia ( no output )
masak camelia: what are you doing?
ShimmerFairy jnhtn: sorry, didn't know it was dynamic variables that let it happen, not the inside of a regex :)
jdv79 how come masak's uni op examples didnt get on the sitr? 11:51
*site
11:52 grondilu joined
masak I didn't get a full picture of that either, except maybe irclog.perlgeek.de/perl6/2015-10-05#i_11321391 11:53
dalek kudo/nom: bad9be0 | lizmat++ | src/core/Setty.pm:
Fix Set/SetHash.kv
11:54
masak (or we could do both)
11:54 rindolf left
dalek ast: 9c5c7a3 | lizmat++ | S02-types/set (2 files):
Make sure we test Set(Hash).kv (correctly)
11:55
jdv79 how come a search in the irc logs for ± doesnt work
masak guesses $answer.contains("encodings") somehow
jdv79 :( 11:56
lizmat m: 'my $s = <a b c>.SetHash; $_ = False for $s.values; dd $s # should this work ? 11:57
camelia rakudo-moar a26ecf: OUTPUT«5===SORRY!5=== Error while compiling /tmp/dwCcxBZFxP␤Unable to parse expression in single quotes; couldn't find final "'" ␤at /tmp/dwCcxBZFxP:1␤------> 3 $s.values; dd $s # should this work ?7⏏5<EOL>␤ expecting any of:␤ sing…»
lizmat m: my $s = <a b c>.SetHash; $_ = False for $s.values; dd $s # should this work ?
camelia rakudo-moar a26ecf: OUTPUT«Cannot assign to an immutable value␤ in block <unit> at /tmp/ikSRAUAC6S:1␤␤»
andreoss m: say 1±2 11:58
camelia rakudo-moar a26ecf: OUTPUT«5===SORRY!5=== Error while compiling /tmp/e2LFGCaska␤Bogus postfix␤at /tmp/e2LFGCaska:1␤------> 3say 17⏏5±2␤ expecting any of:␤ infix␤ infix stopper␤ postfix␤ statement end␤ statement modif…»
lizmat m: my $s = <a b c>.SetHash; $s<a> = False; dd $s # this does
camelia rakudo-moar a26ecf: OUTPUT«SetHash $s = SetHash.new("c","b")␤»
[Coke] jnthn: (profiler) Oh good. 12:01
masak andreoss: irclog.perlgeek.de/perl6/2015-10-05#i_11321117 12:04
12:05 laouji_ joined 12:06 kaare_ joined
domidumont Hello. Do optimize debian packages, I need to know if the moar files produced when building rakudo are the same whatever the architecture (i.e amd64, or arm ...) ? Can anyone confirm ? 12:06
pink_mist why not build on the different architectures and diff? 12:07
seems the best way to be sure
moritz domidumont: I think I recall they are the same 12:08
domidumont: but ask in #moarvm to be sure
12:08 ely-se joined 12:10 laouji_ left, ely-se left 12:12 ponbiki left, chenryn joined
jdv79 i think having a uni op example on the front page may signal to some that p6 takes unicode seriously 12:12
12:13 ponbiki joined
jdv79 esp since none of the other examples have uni 12:13
12:13 ponbiki is now known as Guest74396 12:14 rindolf joined 12:16 rurban left 12:17 chenryn left
dalek q: e56e0c9 | baest++ | answers.md:
Fix spelling of dereferencing
12:19
q: 92cd9c2 | (Zoffix Znet)++ | answers.md:
Merge pull request #4 from baest/patch-1

Fix spelling of dereferencing
domidumont moritz: ok. Thanks 12:20
12:21 Lorphos left, flussence left
dalek p: bd169fa | (Pawel Murias)++ | src/vm/js/ (2 files):
[js] Implement nqp::backendconfig so that we can check the int size
12:25
p: a061c8f | (Pawel Murias)++ | t/serialization/01-basic.t:
In t/serialization/01-basic.t also detect when the js backend is using 32bit integers.
p: 0e78cd7 | (Pawel Murias)++ | t/serialization/01-basic.t:
Check for the correct class instead of a nonexisting one (which ends up being a check against NQPMu).
p: e11ec6e | (Pawel Murias)++ | src/vm/js/bin/run_tests.pl:
[js] Pass t/serialization/01-basic.t
12:35 adhoc left 12:36 laouji_ joined
dalek osystem: 3ba4ab0 | (Steve Mynott)++ | META.list:
Display and cache ecosystem README files for perl6
12:38
p: 7719e1b | (Pawel Murias)++ | src/vm/js/bin/run_tests:
[js] Remove test running script we no longer need.
12:40
p: afa4ca1 | (Pawel Murias)++ | src/vm/js/nqp-runtime/runtime.js:
[js] fix errors found by make js-lint
12:40 laouji_ left
masak pmurias: just curious: do you find js-lint to give needlessly harsh "errors"? (I did.) 12:43
pmurias: for example, the one you just fixed doesn't really look erroneous to me.
jnthn ShimmerFairy: Ah, interesting find (the dynamic binding thing). :) 12:47
ShimmerFairy m: my @*A := 42 12:48
camelia ( no output )
ShimmerFairy jnthn: apparently, dynamics just don't care? :P
m: my %*A := 42
camelia ( no output )
jnthn m: my Int $*A = 'oh noes'
camelia rakudo-moar bad9be: OUTPUT«Type check failed in assignment to $*A; expected Int but got Str␤ in block <unit> at /tmp/qs6ymFxrPv:1␤␤»
jnthn m: my Int $*A := 'oh noes'
camelia ( no output )
ShimmerFairy m: my &*A := 42
jnthn Indeed. Fail.
camelia ( no output )
jnthn That one's RT-able :) 12:49
mathw just flicked past this window and saw "rakudo-badbaby"
ShimmerFairy will do :)
masak submits rakudobug
jnthn The race is on...who will submit to RT first? :) 12:50
ShimmerFairy masak: already on it :)
masak oh, ok. standing down :) 12:51
m: my Int $A := 'oh noes' 12:52
camelia rakudo-moar bad9be: OUTPUT«Type check failed in binding; expected Int but got Str␤ in block <unit> at /tmp/zI9Ehl2Xe5:1␤␤»
12:56 felher left
_itz did that .perl looping on circular data get fixed? 12:57
DrForr m: my $x={}; $x.<a>=$x; say $x.perl 12:58
camelia rakudo-moar bad9be: OUTPUT«(my \Hash_139926166332664 = {:a(Hash_139926166332664)})␤»
andreoss m: my $b = 1 ; my Int $a := $b; $b = "abc"; say $a.WHAT;
camelia rakudo-moar bad9be: OUTPUT«(Str)␤»
lizmat _itz: afaik for .gist and .perl, yes 12:59
12:59 felher joined
andreoss m: my $b = { ... } ; my &a := $b; $b = "abc"; say &a.WHAT; say a 13:00
camelia rakudo-moar bad9be: OUTPUT«(Str)␤Cannot find method 'CALL-ME'␤ in block <unit> at /tmp/9o8jGJneUF:1␤␤»
_itz cool
dalek q: 426307d | util++ | answers.md:
Fix spelling of modifier
13:03
13:06 laouji_ joined
lizmat jnthn: does this ring a bell? Stage mast : Code ref 'BUILD' does not exist in serialization context 13:06
jnthn lizmat: Sort of, but it's been a while and I fixed the source of that error I was aware of 13:07
lizmat: It's deep/hairy enough I'd just not do whatever you did to cause it. 13:08
Otherwise it probably costs me a day of debugging.
RabidGravy reboots 13:09
13:09 RabidGravy left 13:10 laouji_ left
lizmat jnthn: ok, I'll just gist the changes and forget about it for now (and you should as well) 13:11
jnthn lizmat: Yeah, feel free to note it down somewhere, but...we've more pressing things 13:12
13:12 Loren_ left
jnthn lizmat: Last time it was something involving multi-dispatcher cloning 13:12
lizmat gist.github.com/lizmat/f7abe2c542670b0e7b40 13:13
13:13 maziar left, chenryn joined 13:14 aborazmeh joined, aborazmeh left, aborazmeh joined
jnthn hm, but it was count-only it complained about there, not BUILD? 13:15
lizmat it complained about BUILD when I removed count-only
jnthn I wonder if it's about roles nested in roles
dalek p: 5d01ca3 | jnthn++ | src/vm/moar/QAST/QASTRegexCompilerMAST.nqp:
Preparatory refactor for fixing regex args.

No (intended) functional changes; this just gets the code into shape for making subrule call arg lists cope with named and flattening args.
13:16
13:16 sufrostico joined 13:17 maziar joined 13:18 chenryn left 13:22 Ven joined
dalek p: caf4c0f | jnthn++ | src/vm/moar/QAST/QASTRegexCompilerMAST.nqp:
Compile flattening in regex subrule arg lists.
13:22
13:24 skids joined 13:26 yoleaux joined, ChanServ sets mode: +v yoleaux, dpk joined
gfldex m: my @l = (1,(2,3,(4,5)),6); @l.deepmap({|$_}).perl.say; 13:28
yoleaux 3 Oct 2015 10:11Z <AlexDaniel> gfldex: I find it a bit messy, but you are probably right. I've changed it as you recommended
camelia rakudo-moar bad9be: OUTPUT«[1, (2, 3, $(4, 5)), 6]␤»
13:28 maziar left
gfldex why is only the inner most List Seq-ified? 13:28
lizmat Seq-ified? 13:29
gfldex i'm under the impression that $(4,5) is a Seq
masak m: say (4, 5).^name
camelia rakudo-moar bad9be: OUTPUT«List␤»
masak m: say ($(4, 5)).^name
camelia rakudo-moar bad9be: OUTPUT«List␤»
masak nope; it's a List 13:30
jnthn Well, an itemlized list
yoleaux 08:18Z <lizmat> jnthn: is there a runtime downside to a sub called from CORE, and a sub called from Rakudo::Internals:: ?
08:19Z <lizmat> jnthn: if so, would it help to make the sub an only ?
masak .oO( a caramelized List )
jnthn mmmmm 13:31
gfldex then why is only the innermost List itemlisted?
13:31 RabidGravy joined, pecastro left
masak what gfldex asked :) 13:31
jnthn I don't actually know. 13:32
It looks a bit off
jnthn isn't sure what the deepmap impl looks like these days...
gfldex at the moment rakudo is my favoured method of confusion. :)
masak m: my @l = (1,(2,3,(4,(5,6))),7); @l.deepmap({|$_}).perl.say 13:33
camelia rakudo-moar bad9be: OUTPUT«[1, (2, 3, $(4, $(5, 6))), 7]␤»
masak it's not "just the innermost"
it's "level N and below"
masak .oO( only-so-deep-map )
13:33 tokuhirom left
masak m: my @l = (1,[2,3,(4,(5,6))],7); @l.deepmap({|$_}).perl.say 13:34
camelia rakudo-moar bad9be: OUTPUT«[1, [2, 3, (4, $(5, 6))], 7]␤»
gfldex .oO( by lazy we didnt mean ...)
masak a-ha.
.deepmap only talks to Arrays
not Lists
I think that dissolves the mystery
whether it's *intended* behavior... I'm not the one to answer that, I think
moritz it not only solves it, but actually dissolves it. I'm impressed! :-) 13:35
gfldex my next question would be what is the current method of deep-flattening that works both List and Array 13:36
jnthn So the mystery remains...dis-solved? :P
gfldex i'm asking mostly because of rosettacode.org/wiki/Flatten_a_list#Perl_6
13:37 pecastro joined
jnthn m: my @l = (1,[2,3,(4,(5,6))],7); .perl.say given gather @l.deepmap(*.take) # one way 13:38
camelia rakudo-moar bad9be: OUTPUT«(1, 2, 3, 4, 5, 6, 7).Seq␤»
jnthn I had a cuter one but forgot it...
13:38 user3 joined
dalek p: eaa6350 | jnthn++ | src/vm/moar/QAST/QASTRegexCompilerMAST.nqp:
Compile named args in regex subrule arglists.
13:39
gfldex if i post that on rosettacode small children might point fingers at perl6
i can haz List|Array::deepflat plox?
DrForr People are already pointing middle fingers at it, why would some more fingers matter? :) 13:40
gfldex because perl6 is ment to be sexy!
user3 DrForr: I'm puzzled by your statement. why are people pointing middle fingers at perl6? 13:42
jnthn gfldex: While I'd agree we can prolly do better than the one I gave, it's still a bunch better than most of the ones on that page ;) 13:44
DrForr user3: Just forget it, I'm still sore and grumpy over here. 13:45
user3 thats alright doc 13:46
13:46 laouji left
masak .oO( "what's up, doc?" -- "this middle finger here, see..." ) 13:47
DrForr "That's a rather personal question..." 13:48
[Coke] notes belatedly that the profile modal windows when you click "view" have an X that doesn't currently work. You can click away from it or press "ESC" to make it go away. I'll get that fixed shortly. 13:49
masak ++[Coke] 13:50
gtodd installs Stats module with shiny GLR perl6 wheee!! ...
_itz wonders if "spunky spokesbug" is negotiable
gtodd when I do: min @dataset ... everything is grand! but when I use Stats.pm6's "summary" method github.com/MattOates/Stats/blob/ma...s.pm6#L148 13:51
[Coke] would get buyin from TimToady before changing that word. 13:52
(for itz)
gtodd no worky unless I change 'max'=>max($x) to 'max'=>max($x.list) (which is priobably wrong but I just wanted it to work)
jnthn _itz: I suspect the whole intro paragraph wants a re-wording
pmurias re deepmap (not deep-map) is there a naming convention for things in the setting?
_itz github.com/perl6/perl6.org/issues/32 13:53
masak pmurias: only that hyphens are generally avoided when something can be argued to be one word 13:54
13:55 rurban joined, ketadmin left, Ven left 13:56 user3 left
dalek kudo/nom: fbac827 | jnthn++ | tools/build/NQP_REVISION:
Get regex subrule arglist compilation fixes.
13:57
gtodd I assume it's something to do with GLR ... Stats is an extremely simple module I'm sure that after Xmas it won't take long for it or something like it to get R-like
dalek ast: c597341 | jnthn++ | S05-grammar/signatures.t:
Tests to cover RT #113544.
13:58
jnthn gtodd: `max => $x.max` is probably the neater fix 13:59
gfldex m: my @l = [1,[2,3,[4,5]],6]; @l.map({$_.Slip}).perl.say; @l.deepmap({$_.Slip}).perl.say;
camelia rakudo-moar bad9be: OUTPUT«(1, 2, 3, $[4, 5], 6).Seq␤[1, [2, 3, [4, 5]], 6]␤»
masak m: my @l = [1,[2,3,[4,5]],6]; say @l.map(*.Slip).perl; say @l.deepmap(*.Slip).perl 14:06
camelia rakudo-moar bad9be: OUTPUT«(1, 2, 3, $[4, 5], 6).Seq␤[1, [2, 3, [4, 5]], 6]␤»
14:06 Guest74396 is now known as ponbiki
andreoss is there an example of matching a quoted string with consideration of escaped quotes? 14:07
pmurias masak: deepmap is one word? 14:09
masak pmurias: arguably, since we make it a concept. 14:10
pmurias: in the same way "substring" is one word, and hence not "subString"
pmurias do we use camel case too? 14:11
gtodd jnthn: thanks .... I discovered that in the repl and was trying to read up and figure out why it worked on the array but not the "reference" :-)
masak andreoss: github.com/moritz/json/blob/master...mar.pm#L26 14:12
pmurias masak: just looking at S32 there seems to be a lot of inconsistency in naming 14:13
14:13 babe left
pmurias which sucks really badly 14:13
14:14 chenryn joined
masak pmurias: I haven't thought that much about hyphens, but I imagine you're right. 14:15
pmurias: the inconsistency I've been noticing is how some methods have a sub form, and some don't. with no actualy thought of system behind it.
pmurias masak: we have both splitdir and is-hidden 14:17
we need to enforce a naming convention before the christmas release 14:18
jnthn JOY!!! More renaming...
masak yeah, that was my gut reaction, too 14:19
even though I agree with pmurias in principle
[Coke] better now than post christmas
jnthn Well, I can imagine that a bunch of the I/O related stuff does want a pass through naming wise.
moritz and please with deprecations this time :-)
masak the good thing is that we largely do have the infrastructure nowadays to handle the fallout of renaming in the ecosystem
14:19 chenryn left
masak the bad thing is that it's still churn, hassle, and tuits 14:19
moritz++ # deprecations, +1 14:20
jnthn didn't even really know splitdir existed, and wonders where it is
andreoss m: say '"abc \"xyz\" "' ~~ / [ <-["]> | \\<["]> ]+ /;
camelia rakudo-moar bad9be: OUTPUT«「abc \"xyz\" 」␤»
14:20 laouji joined
jnthn But yeah, I'm not really up for much more churn. 14:21
andreoss good
masak jnthn: no-one is "up for" churn :P
14:22 bjz left
masak it's like Eddie Izzard says "he does not suffer fools lightly" -- well, who does? who goes, "hey, come on in, fools! welcome, fools!" :P 14:22
andreoss but if i use \" ~ \" it stops wotk
14:22 Xor_ left
andreoss m: say '"abc \"xyz\" "' ~~ / \" ~ \" [ <-["]> | \\<["]> ]+ /; 14:22
camelia rakudo-moar bad9be: OUTPUT«「"abc \"」␤»
pmurias masak: some of the hassle could be removed if we had a grammar using 'fix-names foo.p6' script
masak pmurias: I'd take a good deprecation warning any day over that, given the choice. 14:23
14:23 Xor_ joined
pmurias masak: isn't the optimal solution is to have both? 14:24
14:24 ely-se joined 14:25 aborazmeh left
pmurias you see a deprecation warnings, run fix-names, run the test suit, look at 'git diff' and commit 14:25
14:25 laouji left, bjz joined
dalek kudo/nom: b6c528d | jnthn++ | src/core/traits.pm:
Routines declaring of/returns T are Callable[T].

So you can pass them to something wanting a `T &foo` parameter.
14:26
14:26 khw joined
dalek kudo/nom: e5612aa | lizmat++ | src/core/Set (3 files):
Make SetHash.pairs/values/kv left-valuable

Which means, that if you assign a False value to it, the key will be removed from the SetHash. Just like setting the value to 0 with BagHash and MixHash.
14:27
lizmat commute to Amsterdam.pm&
14:28 lizmat left
andreoss m: say q/"abc \"xyz\"" .../ ~~ / \" ~ \" [ <-["]> | '\"' ]++ /; 14:31
camelia rakudo-moar fbac82: OUTPUT«「"abc \"xyz\"」␤»
masak pmurias: yes, both is better. but given the exclusive choice, I'd prefer the deprecation which told me what to do manually. 14:32
jnthn m: sub f($x) returns Int { return $x }; say &f ~~ Callable[Int]
camelia rakudo-moar fbac82: OUTPUT«False␤»
14:32 bjz left 14:33 bjz joined
dalek ast: 4e9eb58 | jnthn++ | S06-signature/types.t:
Add a test for RT #121426.
14:34
14:36 laouji joined
grondilu wonders why panda is so slow 14:38
(just launching "panda help" and I could almost make myself a coffee before I get any output :/ ) 14:39
dalek ast: 9d891a0 | jnthn++ | S06-multi/type-based.t:
Further coverage of RT #121426.

Ensuring the fix interacts correctly with multi-dispatch.
tadzik grondilu: many factors :/ 14:40
grondilu: mostly it's just that panda is a big beast
lots of code that gets compiled whenever you type 'panda'
14:40 laouji left
grondilu ok 14:41
[Coke] could also stand to be profiled, maybe?
_itz some suggestions for addition perl6.org home page content at github.com/perl6/perl6.org/pulls 14:42
ermm github.com/perl6/perl6.org/pull/38
dalek href="https://perl6.org:">perl6.org: 21e28de | (Steve Mynott)++ | source/index.html:
Some suggested new content for the home page intended to inspire

Basically 5 bullet points of Perl 6 features which probably needs review.
14:47
href="https://perl6.org:">perl6.org: 97ff8f2 | (Zoffix Znet)++ | source/index.html:
Merge pull request #38 from stmuk/master

Some suggested new content for the home page intended to inspire discussion.
pmurias nqp-m: sub MAIN($arg) {say($arg, nqp::isstr($arg))} 14:48
camelia nqp-moarvm: OUTPUT«/tmp/NgXQoV7ulj1␤»
pmurias and yet in NQP.nqp we have sub MAIN(@ARGS) {...} on the moarvm backend 14:49
dalek href="https://perl6.org:">perl6.org: 2053fab | (Zoffix Znet)++ | source/index.html:
Take out feature list from "lead" paragraph
14:51
gfldex m: my @l = (1,(2,3,(4,5)),6); my @f = @l.deepmap({$_ * -1}); @f.perl.say;
camelia rakudo-moar fbac82: OUTPUT«[-1, (-2, -3, $(-4, -5)), -6]␤»
14:52 mspo left
skids ^^ Every one in a while you have to step back and appreciate how menial it would be to "do that in C" 14:53
tadzik :D
14:53 ChoHag_ left
gfldex m: my constant @l := (1,(2,3,(4,5,(6,7))),8); my $f = @l.deepmap({$_ * -1}); $f.perl.say; 14:54
camelia rakudo-moar e5612a: OUTPUT«$(-1, $(-2, -3, $(-4, -5, $(-6, -7))), -8)␤»
ilmari what's the difference between () [] and $()? 14:55
_itz list, array and slip
skids Though... where's that lone $() in the first example coming from?
list, arrray, and item
slip would be |()
_itz oh sorry 14:56
PerlJam And you don't usually see slips. :)
skids Yeah, they tend to slip by :-)
14:57 pyrimidi_ left
_itz Zoffix: the smaller section looks a lot better but you collapsed the 3rd bullet point into the 2nd I think 14:57
14:58 ely-se left, ZoffixWork joined
ZoffixWork _itz, yeah, I've noticed but my dev box froze up :) I'm on it! :) 14:58
_itz, are you stmuck on GitHub? 14:59
_itz ZoffixWork: yes
jnthn ilmari: () List, Array, $ just means "itemize" (put in an anonymous Scalar) so $() is an itemized list and $[] is an itemized Array
14:59 laouji joined
_itz it's a bullet point slip! 14:59
jnthn oops, "[] Array"
PerlJam m: $(7).perl.say; $[7].perl.say; $(3,4).perl.say; $[3,4].perl.say; # some useful differences
camelia rakudo-moar e5612a: OUTPUT«7␤[7]␤$(3, 4)␤[3, 4]␤»
jnthn masak: Did a past you that was a future you when you lasted updated rt.perl.org/Ticket/Display.html?id=77334 run into any use cases for not doing what the ticket wants? :) 15:01
15:01 andreoss left
gfldex m: say nqp::istype((class :: does SlippyIterator { method pull-one() is raw {}}).new, Iterable); 15:01
camelia rakudo-moar e5612a: OUTPUT«0␤»
gfldex i thinks that's where the lone $() comming from
masak jnthn: let the present me check.
jnthn gfldex: I'd not expect that to be Iterable 15:02
gfldex: It's an Iterator, though
gfldex jnthn: i think github.com/rakudo/rakudo/blob/4f70...ps.pm#L518 does expect it to be Iterable
15:03 larion left 15:04 domidumont left, lichtkind joined
jnthn gfldex: I'm not sure that line is wrong but the .item 2 lines below looks a bit suspect to me 15:05
15:06 ChoHag joined
masak jnthn: the answer to your question is "no". please decont away. 15:06
dalek href="https://perl6.org:">perl6.org: 7e2d6c6 | (Zoffix Znet)++ | source/ (3 files):
Fix up styles and alignment
jnthn masak: OK. I pondered waiting for TimToady approval but I already have a queue of those things already, so... :) 15:07
masak: Only tricky thing is that it may not be as simple as decont
15:07 pierrot left, pierrot joined
jnthn m: say (1, 2, 3).map({ $['a' xx $_] }).perl 15:08
camelia rakudo-moar e5612a: OUTPUT«($["a"], $["a", "a"], $["a", "a", "a"]).Seq␤»
jnthn If I decont then we lose the itemization on those.
masak ...hm. 15:09
jnthn So we more need the return from sub semantics (recont if item)
uh
recont if Iterable
masak I can see how that's related, but it *feels* like completely different thing.
uh, ok
I had no idea subs did that :P
jnthn They did 'cus you filed a ticket when they didn't :P 15:11
PerlJam wonders if this falls in the category of tormenting the implementor ... 15:12
jnthn Back when we had lots of implicit flattening and [...] was an item, it was rather more noticable.
Nowadays I'd probably get through spectest with a patch that just deconts, but... :)
jnthn doesn't really have a great name for the decont if not Iterable, recont if it is... :) 15:13
moritz autocont
jnthn otoh
PerlJam dwimcont
jnthn I wonder if we need the magic any more
(Because so much auto-flattening is gone.) 15:14
15:14 diana_olhovik_ left
jdv79 who else thinks a custom unicode op example on the front page would be coool? 15:15
c'mon;)
ZoffixWork jdv79, we had it! :)
jdv79 the past is great but the future is more interesting ot me 15:16
ZoffixWork jdv79, we had this one: perl6.xtatik.org/
15:17 rurban1 joined
ZoffixWork jdv79, there was one from masak that's cleaner. 15:17
jdv79 yes, something like masak's is cool
15:17 rurban left
jdv79 even if just hte simple plus minus 15:17
jnthn Filed this discussion on the ticket; will have to filter this one through TimToady++ also I guess :) 15:18
masak jnthn: I can assure you that most of the time after filing a ticket, I *forget about the ticket completely*, instead of hanging around for the fix and then reviewing it and nodding sagely :P
I *wish* I did the latter sometimes :P
PerlJam masak: you can nod sagely at any time :)
ZoffixWork jdv79, this one? irclog.perlgeek.de/perl6/2015-10-05#i_11321117
jdv79, I don't like that one, 'cause I've no idea what's going on there :P
15:19 chenryn_ joined
jdv79 the concept of "plus or minus" seems pretty easy to me 15:19
what is complicated about the example?
ZoffixWork jdv79, what is it doing? 15:20
jnthn Quite a few of the xmas tickets are things needing little design rulings :)
jdv79 its just op sugar for a range
15:20 Xor_ left
jdv79 i'm sure masak or someone could write an "explain" for it 15:21
i would but i'm not great at docs
ZoffixWork jdv79, I like masak's example with "approx equals", but I'm miserably failing with searching the logs. It was mentioned yesterday 15:22
jdv79, as a Perl 6 outsider, I found it much clearer than the +/- sample. 15:23
jdv79 sounds perhaps highly subjective but ok
[Coke] ≈ at irclog.perlgeek.de/perl6/2015-10-05#i_11321129
jdv79 either is fine to me ^H
ZoffixWork Oh that one also has the +/- bit 15:24
jdv79 esp since the approx is built on the plus minus:)
ZoffixWork :D
[Coke] +1 on the ≈ - it hides the smartmatch in the sub, making it less necessary to understand ~~ to understand adding an operator.
... oh, I didn't notice that. Hurm.
eh, ship it
15:25 Xor_ joined
jdv79 woohoo 15:25
ZoffixWork [Coke]++
:D
jdv79 now if only that would be used in tests and provide col diagnostics
*cool
PerlJam except, as an example on perl6.org, don't we want "understanding"?
ZoffixWork I would get rid of "epsilon" on that example and the +/- think
PerlJam how would you get rid of epsilon? Fix it to some tiny number? 15:26
jdv79 PerlJam: if there was a prose explanation attached what is not "understanding"?
ZoffixWork m: sub infix:<≈>($l, $r) { $l-0.5 < $r and $l+0.5 > $r }; say 2.25 ≈ 2.3; say 3.25 ≈ 2.3
camelia rakudo-moar e5612a: OUTPUT«True␤False␤»
ZoffixWork Yeah, just have some pre-built thing 15:27
PerlJam jdv79: is there a place for prose on the perl6.org?
jdv79: or do you mean in the comments of the code?
15:27 chenryn_ left
jdv79 huh? i'm talking about the suggestion about an explanation like some oether site 15:27
racket? i forget.
ilmari sub infix:<≈>($l, $r) { $l-0.5 < $r < $l+0.5 }; say 2.25 ≈ 2.3; say 3.25 ≈ 2.3
m: sub infix:<≈>($l, $r) { $l-0.5 < $r < $l+0.5 }; say 2.25 ≈ 2.3; say 3.25 ≈ 2.3
camelia rakudo-moar e5612a: OUTPUT«True␤False␤»
ZoffixWork Oh you can do that in Perl 6? Wow, sweet! 15:28
PerlJam jdv79: oh! yes, that would be excellent!
jdv79 irclog.perlgeek.de/perl6/2015-10-06#i_11324349
[Coke] ilmari++
15:29 llfourn joined
_itz pbs.twimg.com/media/CQm-iaVUcAAERGc.jpg 15:29
jdv79 i still think masak's are better thought out but if it has to be dumbed down that much to get on the page sure
_itz dining philosophers?
PerlJam jdv79: I don't know about "dumbed down", but I think we should have examples that are simple to read and understand. We should also have some examples that are maybe not-so-simple to show off :) 15:31
masak I'm fine with removing the epsilon, though I put it there because I thought it was a nice, non-silly use of operator adverbs
something that's sorely lacking outside of :exists and :delete
15:32 chenryn_ joined
[Coke] There isn't a lot of screen real estate there. I like the idea of each little block showing -a- feature. 15:32
PerlJam masak: maybe it would look better as a ternary op.
masak PerlJam: could you express that in code, please?
[Coke] we have the examples page with more complex examples that we can link to
jdv79 yeah, ilmari's only shows one "interesting" thing. masak's could be used a a jumping off point into many different topics
ZoffixWork jdv79, yeah, but if you don't "dumb down", this will be happening, except without you there to explain it: irclog.perlgeek.de/perl6/2015-10-06#i_11326866 15:33
:)
_itz I think the home page examples should be "ad examples" .. easy to understand and compelling
ZoffixWork Agreed
ilmari jdv79: and it still has an additional neat thing, the comparison chainging
PerlJam masak: just as soon as I figure out how to do ternary ops :)
15:34 llfourn left
ilmari m: say 1 < 3 < 5 15:34
camelia rakudo-moar e5612a: OUTPUT«True␤»
ZoffixWork Or maybe do a carusel(sp?) type of thing and shove more complex examples further down the line, so only the users interested enough in seeing them get to them
jdv79 yeah, as long as there's something showing unicode
ilmari m: say 1 == 1.0 == "1"
camelia rakudo-moar e5612a: OUTPUT«True␤»
jdv79 for now
masak PerlJam: please let me know how that goes.
15:34 tokuhirom joined
ilmari m: say 4 == 4.0 == "4" 15:34
camelia rakudo-moar e5612a: OUTPUT«True␤»
15:34 FROGGS left
ilmari m: say 1 == 1.0 == Ttrue 15:34
jdv79 ilmari: true
camelia rakudo-moar e5612a: OUTPUT«5===SORRY!5=== Error while compiling /tmp/EIBEI9k7tP␤Undeclared name:␤ Ttrue used at line 1␤␤»
_itz there probably should be a grammar as well
ilmari m: say 1 == 1.0 == True
camelia rakudo-moar e5612a: OUTPUT«True␤»
jdv79 yes, ad examples++ 15:35
masak m: say 4 == 4.0 == "four"
camelia rakudo-moar e5612a: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '7⏏5four' (indicated by 7⏏)␤ in block <unit> at /tmp/C8mFLL5fK7:1␤␤Actually thrown at:␤ in block <unit> at /tmp/C8mFLL5fK7:1␤␤»
masak m: multi prefix:<+>("four") { 4 }; say 4 == 4.0 == "four"
camelia rakudo-moar e5612a: OUTPUT«Cannot convert string to number: base-10 number must begin with valid digits or '.' in '7⏏5four' (indicated by 7⏏)␤ in block <unit> at /tmp/cs_soyClJJ:1␤␤Actually thrown at:␤ in block <unit> at /tmp/cs_soyClJJ:1␤␤»
masak m: multi prefix:<+>("four") { 4 }; say 4 == 4.0 == +"four"
camelia rakudo-moar e5612a: OUTPUT«True␤»
masak \o/
ZoffixWork :D
15:35 chenryn_ left 15:37 lichtkind left 15:38 tokuhirom left, tm joined 15:39 chenryn_ joined, chenryn_ left
dalek kudo/nom: f04a24b | jnthn++ | src/core/Array.pm:
Fix BIND-POS type checking in typed arrays.

We accidentally allowed the unchecked candidates to get called due to the missing proto.
15:40
15:40 chenryn_ joined
masak m: multi infix:<+>(2, 2) { 5 }; say "2 + 2 == ", 2 + 2 15:40
camelia rakudo-moar e5612a: OUTPUT«2 + 2 == 5␤»
masak \o/
dalek ast: 2e2b9fe | jnthn++ | S09-typed-arrays/arrays.t:
Tests for RT #123769.
timotimo i'm not 100% +1 on "lazy and eager evaluation", because people from a e.g. haskell background will expect pervasive lazyness throughout the whole language, whereas we only have lists that can be lazy 15:43
15:46 ZoffixWork left 15:49 laouji left
dalek href="https://perl6.org:">perl6.org: 4a11eaf | (Steve Mynott)++ | source/index.html:
only lists are lazy as suggested by timotimo
15:50
15:50 laouji joined
timotimo thanks :) 15:51
15:55 laouji left
jnthn I just wrote up a rejection of rt.perl.org/Ticket/Display.html?id=125762 - not sure if there's any nuggets in it worth nabbing for documentation somewhere. 15:55
Hm. RT #125753. Back to the regex engine for me, I guess... 16:01
16:01 synbot6 joined 16:06 laouji joined
TimToady such backlog...at least I dont have to write any more talks this year... 16:07
yoleaux 5 Oct 2015 14:48Z <grondilu> TimToady: I'm curious to read your thoughts about &[o] being associative. It bugs me a bit that we can't define something like [o] &f xx *
jnthn TimToady: How went the talk? 16:08
TimToady quite well, room was almost full 16:09
[Coke] RT: 1037; xmas: 91; LTA: 88; weird: 12; glr: 4; nom regression: 8; testneeded: 10
TimToady will be on video at some point
use vim as my presenter :) 16:10
jnthn [Coke]: yay, I was looking forward to seeing the xmas count...been attacking that some today :)
[Coke] jnthn++
16:11 laouji left
TimToady found a bug during the talk 16:11
m: say $[1,2,3].perl; say $(1,2,3).perl
camelia rakudo-moar e5612a: OUTPUT«[1, 2, 3]␤$(1, 2, 3)␤»
jnthn oops
TimToady where'd the $ go on the []?
jnthn Beats me!
_itz down the back of the sofa?
TimToady likely
timotimo well, i certainly didn't eat it! 16:12
timotimo chews inconspicuously
TimToady maybe they said "buck" and you thought it was venison?
16:14 Ven joined
gfldex m: my @l = (1,(2,3,(4,5,(6,7))),8); my $f = @l.deepmap({$_ * -1}); $f.perl.say; 16:14
camelia rakudo-moar e5612a: OUTPUT«[-1, (-2, -3, $(-4, -5, $(-6, -7))), -8]␤»
gfldex m: my @l = (1,$(2,3,(4,5,(6,7))),8); my $f = @l.deepmap({$_ * -1}); $f.perl.say;
camelia rakudo-moar e5612a: OUTPUT«[-1, (-2, -3, $(-4, -5, $(-6, -7))), -8]␤»
gfldex that could be related
16:14 ^elyse^ joined
[Coke] tries to git pull -rebase and gets a merge conflict. wtf. 16:15
(with, I thought, no local commits)
16:16 zakharyas left
[Coke] must have been an unpushed one, ah well. 16:17
timotimo is the "can split method calls across multiple lines" thing in the latest release already? 16:18
star-m: say $*PERL.version
camelia star-m 2015.03: OUTPUT«vunknown␤»
timotimo star-m: say $*VM.version
camelia star-m 2015.03: OUTPUT«v2015.3␤»
timotimo oh
it's certainly not in that version!
[Coke] use Test; is_approx (-1) ** -i, 23.1406926327793, "(-1) ** -i is approx 23.1406926327793"; 16:19
m: use Test; is_approx (-1) ** -i, 23.1406926327793, "(-1) ** -i is approx 23.1406926327793";
camelia rakudo-moar e5612a: OUTPUT«not ok 1 - (-1) ** -i is approx 23.1406926327793␤␤# Failed test '(-1) ** -i is approx 23.1406926327793'␤# at /tmp/K70AFpQH35 line 1␤# expected: 23.1406926327793␤# got: NaN+NaN\i␤»
timotimo it was committed before the change to 6.Birthday 16:20
jnthn Pretty sure it was after the last monthly
TimToady yes, it was
timotimo right 16:21
16:21 laouji joined
timotimo i was about to rewrite the answer about "why do i have to use unspace for method calls across line breaks?" 16:21
jnthn Oh grmbl, the charrange bug is 'cus the NFA engine doesn't know anything about charrange in the ignorecase and ignoremark case
timotimo but if it's not yet in a release we've done, now is probably not the right moment to do it
jnthn wonders how to find out if there are places in unicode where CPx..CPy is contiguous but fc(CPx)..fc(CPy) isn't 16:24
There surely are such cases.
TimToady I know! Write a program!
timotimo you mean computers can do that?
jnthn I need to implement fc first :)
TimToady that's a program too
timotimo nativecall into something that can do fc :) 16:25
jnthn Though suspect we already have the data handy to do it...
The other problem is
The NFA engine as we have it today is all about codepoints
TimToady ⴽ shaving
jnthn Meaning that we almost certainly have weird bugs in the interaction of NFG and LTM
ilmari jnthn: fc('ß') eq 'ss', for starters
timotimo .u ⴽ
yoleaux U+2D3D TIFINAGH LETTER YAK [Lo] (ⴽ)
timotimo cool
ilmari jnthn: so using that as a range end point is nonsensical
16:25 laouji left
jnthn ilmari: Indeed 16:26
ilmari: Though we don't actually implement the special casing stuff yet
ilmari m: "BLAH".fc.say 16:27
camelia rakudo-moar f04a24: OUTPUT«Method 'fc' not found for invocant of class 'Str'␤ in block <unit> at /tmp/OWe92tIfD_:1␤␤»
ilmari m: "ß".uc.say
camelia rakudo-moar f04a24: OUTPUT«ß␤»
ilmari $ perl -Mutf8 -C -E 'say uc "ß"' 16:28
SS
[Coke] ilmari: FYI, there are RTs & tests about the ß and other more complicated casing issues.
jnthn [Coke]: Indeed, the special casing one is on the xmas list too :)
timotimo we totally need to turn that silly Ss thing into a grapheme cluster >_> 16:29
16:29 pecastro left
jnthn Only if the Unicode spec does :P 16:29
timotimo pls unicode
TimToady all is fair if you predeclare: use screw Unicode;
jnthn m: say ord('ß') 16:30
camelia rakudo-moar f04a24: OUTPUT«223␤»
jnthn m: say ord('ß').base(16)
camelia rakudo-moar f04a24: OUTPUT«DF␤»
16:36 laouji joined 16:37 Loren joined
nine m: say ($[1,2]).perl 16:38
camelia rakudo-moar f04a24: OUTPUT«[1, 2]␤»
dalek kudo/nom: 02e6198 | (Stefan Seifert)++ | src/core/Array.pm:
Fix ($[1,2]).perl missing the $
16:39
16:39 dakkar left
nine I have to admit however, that I'm not sure why exactly this fixes it. 16:39
[Coke] nine: lc self is used right after that - with SELF: in the signature, what does self mean? 16:40
16:40 laouji left
[Coke] anyone else getting failures in t/spec/S04-exceptions/fail.t, t/spec/S04-statements/try.t, t/spec/S32-io/move.t, t/spec/integration/error-reporting.t ? 16:40
TimToady does 'self' do a decont for some reason/ 16:41
16:41 pecastro joined
TimToady ? 16:41
jnthn [Coke]: yes
16:41 ZoffixWork joined
nine Hah, seems like moritz++ fixed the exact same bug in List back in August 16:41
jnthn TimToady: Yes, we've pretty much always had it that way
TimToady: For one 'cus it means self is useful for attribute access without having to throw decont onto every single one 16:42
ZoffixWork TimToady, timotimo I'm pretty sure the "methods on multiple lines" are in the latest Rakudo Star. I installed it late last week and all the examples in these docs work: github.com/stmuk/p6-Text-VimColour
nine jnthn: regarding github.com/rakudo/rakudo/commit/f04a24bc1b so the takeaway is "when in doubt, add a proto"?
16:43 Ven left
jnthn nine: Only when you know you want to hide all inherited candidates 16:43
nine: You often don't
nine Ah, it's about multi dispatch and inheritance. Thanks for enlightening me :) 16:44
16:44 yqt joined
TimToady same thing for lexical scopes and multi subs 16:45
nine It's quite obvious in hindsight, but this definitely has been a missing piece in my understanding of multi dispatch 16:46
TimToady I wonder if we actually test that...
jnthn TimToady: Oh, I had a question in that area... 16:48
proto foo(|) { * }; sub foo($a) { }; sub foo($a, $b) { } 16:49
Here, the synopses have us promote those two subs to multis
But do I have to distinguish it from:
multi sub foo($a) { }; sub foo($a, $b) { }
(That is, we already installed a generated proto)
TimToady otoh, it's much more likely to be a mistake 16:50
well, but if it Just Works...
but we could mark generated proto as not being quite that powerful 16:51
nine If it Just Works it better always worked like intended...
16:51 laouji joined
jnthn Well, I can probably mark it in the compiler's symbol table rather than having to mark the code object itself 16:51
TimToady "Need explicit proto to promote only sub to multi sub"
jnthn Anyway, I can find a way to distinguish them; it's just a little extra effort, which I only wanted to do it it's needed. :) 16:52
Or more like, didn't want to do and then have you tell me I didn't need to. :)
16:52 ZoffixWork left, atweiden joined, ^elyse^ left
TimToady m: proto foo(|) { * }; sub foo($a) { say "HERE" }; sub foo($a, $b) { say "THERE" }; foo(42) 16:53
camelia rakudo-moar f04a24: OUTPUT«5===SORRY!5=== Error while compiling /tmp/7cbxO3We6a␤Redeclaration of routine foo␤at /tmp/7cbxO3We6a:1␤------> 3foo(|) { * }; sub foo($a) { say "HERE" }7⏏5; sub foo($a, $b) { say "THERE" }; foo(4␤ expecting any of:␤ horizontal…»
jnthn Today we don't promote at all 16:54
TimToady otgh, conservative thing might be to just disallow it for now
jnthn Sure, I can change the design docs and resolve the ticket that way too :)
moritz +1 to disallowing it 16:55
TimToady I think it's that way to go with the design of 'my multi $var' allowing multiple 'my $var' which is motivated by making it easier to write macros
16:56 laouji left
moritz we have multi vars? :-) 16:56
TimToady but we don't do that yet either, so I don't see much point; maybe a use case will come up when we do macros right
jnthn OK
Then I'll update design to say we're not doing that, clear out any tests, and resolve ticket that way 16:57
Thanks :)
16:57 diana_olhovik_ joined
TimToady and it's not really quite the same either, multiple my $var are supposed to refer to the same variable, but multiple subs aren't, so it's kinda smelly anyway 16:58
jnthn TimToady: There's 3 other RTs I wanted input on, with stuff scattered through the backlog, so unless it's more useful to ask it again now I'll leave you to discover 'em while catching up :)
TimToady yes, I saw my name being taken in (hopefully not) vain when I skimmed, so I expect to run into them when I backlog For Reals 16:59
jnthn *nod*
TimToady backlog looks dense as opposed to fluffy this time :)
jnthn I'm probably going to focus on the Unicode casing stuff and NFG related bits over the coming days 17:00
There's a small cluster of xmas RTs in that area
17:00 [TuxCM] joined
TimToady maybe we can do better than linear scans of large sets of chars 17:01
jnthn For char classes?
We already do
We have a charrange
ilmari inversion lists?
TimToady isn't that only a single range?
jnthn But...how that intersects with :m and :i is kinda hairy
Yes, like a..f 17:02
TimToady hence the fc continuity question
I think we need a better way to represent non-contig sets
jnthn *nod*
TimToady then you just switch if you notice your fc isn't contig 17:03
jnthn ilmari: Yes, that's a useful data structure for it
ilmari perl5 uses them heavily for unicode properties
TimToady um, I don't think that's how perl5 does it
it uses lazy hierarchical bitmaps 17:04
ilmari unicode properties are compiled to inversion lists that are just mapped into memory
17:04 diana_olhovik_ left
ilmari the sets of codepoints matching a given unicode property, that is 17:04
jnthn should really get foldcase in place and the support for case changes that change the string length
TimToady not really inversion lists, unless they changed
ilmari TimToady: khw has been doing lots of work on that the last few years
TimToady okay 17:05
ilmari perl5.git.perl.org/perl.git/blob/HE...nvlists.pl
some are compiled into regcomp.o, others into utf8.o 17:06
TimToady thanks for the update 17:08
TimToady hasn't been trakcing p5 very closely, for some reason...
17:10 ^elyse^ joined 17:11 [TuxCM] left
TimToady wonders, though, if the performance on non-Latin scripts suffers from not caching recently used bitmaps 17:12
jnthn Time for a break; bbl & 17:13
17:15 rurban1 left 17:16 n0tjack joined
grondilu has anyone considered allowing functions to be algebric? I mean, &f + &g would mean -> |arg { f(arg) + g(arg) } 17:18
17:18 espadrine left
khw Actually, perl5 uses a hybrid scheme. For properties that have few enough entries in their inversion lists, binary searches are used to look things up; for larger ones, the old scheme 17:18
is transparently used. Keep in mind that these bitmaps are accessed through a hash, so there is that overhead. I did some performance testing and didn't see significant differences
but don't remember the details
timotimo grondilu: you can implement &+& as a custom operator :) 17:19
17:19 [TuxCM] joined
grondilu yeah I guess. 17:19
but the main idea was to shorten some code. Having to define the operator myself would prevent that. 17:20
17:20 laouji joined
grondilu not to mention that I would also have to do it for *, -, / and stuff 17:21
what I had in mind was rosettacode.org/wiki/Horner%27s_rul...ion#Perl_6 for instance
instead of map { $_ + $x / * }, @coeff, I whish I could write @coeff X+ $x / * 17:22
TimToady khw: thanks!
yeah, hashes have a large k 17:23
[Coke] -1 on functions using mathops.
grondilu m: multi infix:<+>(&f, &g) { -> |a { f(a) + g(a) }; say ([o] (-19, 7, -4, 6) X+ 3 / *)(0) 17:24
camelia rakudo-moar 02e619: OUTPUT«5===SORRY!5=== Error while compiling /tmp/axuJJii2l4␤Missing block␤at /tmp/axuJJii2l4:1␤------> 3; say ([o] (-19, 7, -4, 6) X+ 3 / *)(0)7⏏5<EOL>␤ expecting any of:␤ postfix␤ statement end␤ statement modifier…»
grondilu m: multi infix:<+>(&f, &g) { -> |a { f(a) + g(a) } }; say ([o] (-19, 7, -4, 6) X+ 3 / *)(0)
camelia rakudo-moar 02e619: OUTPUT«Cannot call Numeric(WhateverCode: ); none of these signatures match:␤ (Mu:U \v: *%_)␤ in block <unit> at /tmp/jH2S3DZ7wd:1␤␤»
17:24 domidumont joined
grondilu m: multi infix:<+>($x, &g) { -> |a { $x + g(a) } }; say ([o] (-19, 7, -4, 6) X+ 3 / *)(0) 17:25
camelia rakudo-moar 02e619: OUTPUT«-16␤»
17:25 laouji left
grondilu m: multi infix:<+>($x, &g) { -> |a { $x + g(a) } }; say ([o] (-19, 7, -4, 6) X+ 3 * *)(0) 17:25
camelia rakudo-moar 02e619: OUTPUT«-16␤»
grondilu m: multi infix:<+>($x, &g) { -> |a { $x + g(a) } }; say ([o] (-19, 7, -4, 6) X+ (3 * *))(0) 17:26
camelia rakudo-moar 02e619: OUTPUT«-16␤»
17:27 flussence joined
TimToady masak, ven, nine: re irclog.perlgeek.de/perl6/2015-10-05#i_11317759 we could think about forcing $ = 42 to mean 'my', and leave non-assigned $ to be 'state' 17:28
but then we get the question of whether = is the right twigil for pod, since people will want to write $=42 and such 17:29
m: say $=42 17:30
camelia rakudo-moar 02e619: OUTPUT«42␤»
TimToady m: say $=sin 1
camelia rakudo-moar 02e619: OUTPUT«5===SORRY!5=== Error while compiling /tmp/9tXvghvan9␤Pod variable $=sin not yet implemented. Sorry. ␤at /tmp/9tXvghvan9:1␤------> 3say $=sin7⏏5 1␤»
17:33 [TuxCM] left 17:35 kuzgun joined
[Coke] just read rethrown to pun with "retcon". whoops. 17:35
We have a test util to get temp file names for tests ,yes? (see custom logic in io/bom.t) 17:37
17:40 Xor_ left 17:44 Peter_R joined 17:45 xfix joined, xfix left, xfix joined 17:51 Xor_ joined 17:52 diana_olhovik joined, FROGGS joined
FROGGS o/ 17:55
17:56 yqt left
gtodd Can anyone tell me the obviousness I am missing in this function/method: github.com/MattOates/Stats/blob/ma...s.pm6#L153 it always falls through to default because it is returning the wrong type for the type check (I think) ... I don't want to finish it for now I just want htis(@data) to print out $bin-width for each adverb type 18:01
atweiden m: my DateTime $dt .= new(:year(1979), :month(5), :day(27), :hour(0), :minute(32), :second(0.999999), :timezone(-25200)); say $dt.perl; say $dt; say $dt.second; say so '1979-05-27T00:32:00.999999-07:00' ~~ $dt
camelia rakudo-moar 02e619: OUTPUT«DateTime.new(year => 1979, month => 5, day => 27, hour => 0, minute => 32, second => 0.999999, timezone => -25200)␤1979-05-27T00:32:00-07:00␤0.999999␤False␤»
atweiden is there a built-in way to stringify a DateTime with fractional seconds displayed by default?
gtodd so hist (@data, breaks => "Rice") would pruint the result of the given block (i.e. $bin-width = abs($x.max - $x.min) / 2*($x.elems ** 1/3) ... with the sample data that would be 13.333) 18:02
18:03 chenryn_ left
gtodd hist(@data) that is ... 18:05
TimToady grondilu: I think we can make ∘ left-associative, but note that [∘] in irclog.perlgeek.de/perl6/2015-10-05#i_11318069 makes no sense, though [\∘] works if ∘ is left-associative 18:08
gtodd atweiden: $dt.seconds will show seconds ... you want the whole date string to just include that? 18:10
masak since function composition is... associative, does it actually matter if it's left associative or right associative?
TimToady pragmatically, yes 18:11
18:11 abaugher left
TimToady m: say [\o](&cos xx *)[^10] 18:11
camelia rakudo-moar 02e619: OUTPUT«Cannot reverse a lazy list␤ in block <unit> at /tmp/lFjgEgi40c:1␤␤Actually thrown at:␤ in any at gen/moar/m-Metamodel.nqp:2860␤ in block <unit> at /tmp/lFjgEgi40c:1␤␤»
TimToady that works in my version
_itz 100% 3769 3.7KB/s 00:00 18:12
oops
18:13 abaugher joined
ab5tract_ moznion++ ; IO::Blob 18:13
18:14 paulmr joined
gtodd atweiden: the way I understand it there is likely something like $dt.format(seconds => 1 ) and when you do the construction you could specify it ... but IANAPH 18:17
18:21 laouji joined
chansen_ atweiden: IMO, If DateTime accepts any fractional seconds precision, it should also format it in a logical decimal unit: milli, micro, nano, pico ... yocto? 18:21
pink_mist planck instants?
chansen_ heh 18:22
TimToady roast seems to think someone brake backtracing
roast also doesn't like the $[] fix 18:24
dalek kudo/nom: 164953b | FROGGS++ | src/ (3 files):
change some nqp::die's to $*W.throw to get file/line

This also means better testability. There are still a bunch of nqp::die's left.
18:25
18:25 yqt joined, laouji left
dalek kudo/nom: 54746f2 | TimToady++ | src/ (2 files):
make o left associative

This allows [\o] &cos xx * to work.
18:29
masak oh, that's the pragmatic reason. got it. 18:30
moritz FROGGS++ # I like it
masak yes, then infix:<o> should definitely be left associative.
I guess that generalizes, even. "when in doubt, make it left associative"
grondilu oh yeah [o] &f xx * makes no sense indeed
(well, at least no computing sense) 18:31
atweiden m: my DateTime $dt .= new(:year(1979), :month(5), :day(27), :hour(0), :minute(32), :second(0.999999), :timezone(-25200)); say $dt.perl; say $dt; say $dt.second; say so '1979-05-27T00:32:00.999999-07:00' ~~ $dt
camelia rakudo-moar 02e619: OUTPUT«DateTime.new(year => 1979, month => 5, day => 27, hour => 0, minute => 32, second => 0.999999, timezone => -25200)␤1979-05-27T00:32:00-07:00␤0.999999␤False␤»
TimToady well, Ramanujan might have an opinion :) 18:32
chansen_ Why is DateTime is using the named parameter :timezone when it takes a offset from UTC in seconds, thats just plain wrong. An offset from UTC is not a timezone
atweiden m: my DateTime $dt .= new(:year(1979), :month(5), :day(27), :hour(0), :minute(32), :second(0.999999), :timezone(-25200)); say $dt.perl; say $dt; say $dt.second; say so '1979-05-27T00:32:00-07:00' ~~ $dt
camelia rakudo-moar 02e619: OUTPUT«DateTime.new(year => 1979, month => 5, day => 27, hour => 0, minute => 32, second => 0.999999, timezone => -25200)␤1979-05-27T00:32:00-07:00␤0.999999␤False␤»
atweiden gtodd: p6 seems to treat the stringified output as having the fractional second but does not display it
chansen_: i ask because for testing purposes, i had to specify a DateTime.new to compare results github.com/atweiden/config-toml/bl...es.t#L1406
masak chansen_: I think the value changed and the name stayed, something like that
grondilu arguably [o] &f xx * is the limit of [o] &f xx $n for $n -> Inf
masak chansen_: agree that it's wrong
grondilu: but only arguably :P 18:33
TimToady well, hard to find the limits of some functions, like infix:<-> :)
grondilu never said the limit is guaranteed to exist 18:34
grondilu considered [o] 1/* xx * the other day and found this example amusing 18:35
gtodd atweiden: hmm reading the docs it seems maybe you need a ":&formatter" ... the formatter parameter is something that accepts a format_datetime() method
18:35 tokuhirom joined
dalek ast: e5f7adc | (Stefan Seifert)++ | / (3 files):
.perl on an itemized array should print the $
18:36
18:36 laouji joined
chansen_ masak: :timezone should probably renamed to :offset and it should probably only accept a minute duration instead of a second duration. 18:36
nine Those spec test failures really make it harder to spot regressions
TimToady you just know some tinpot dictator is going to want noon exactly when the sun is overhead, though 18:37
18:38 kuzgun left
arnsholt I think pre-standardization times are sometimes represented as second-accuracy timezones 18:39
18:39 lsm-desktop left
TimToady wonders if a minute-accurate longitude runs through Kim Jong Il's house... 18:40
18:40 tokuhirom left 18:41 laouji left 18:44 fling left 18:45 fling joined, telex left 18:46 diana_olhovik left
jnthn Given there's been at least one example of a timezone whose offset to UTC was specified in not just seconds, but fractions of a second...minutes seems a bit short-sighted :) 18:46
grondilu m: say pi*12756e3 / 360 / 60
camelia rakudo-moar 02e619: OUTPUT«1855.28499486997␤»
18:46 telex joined 18:47 yqt left
gtodd TimToady: why are time zones so *stupid* ? why can't they be interrigent !? 18:48
nine jnthn: I got to ask: WHAT?!
jnthn nine: en.wikipedia.org/wiki/UTC%2B00:20 18:49
18:51 laouji joined
nine That's....actually at least logical, if incredibly self centered and short sighted 18:51
FROGGS jnthn: what should 'my Int:D $a' complain about? we already have a 'Term definition requires an initializer' 18:52
gtodd atweiden: eek specifying a :&formatter for a DateTime object has cognitive load
at least for me :-\
FROGGS jnthn: 'To be defined variable definition requires an initializer'? 18:53
no, that's weird and wrong
jnthn FROGGS: That's not quite it. More like "Must initialize a variable of type 'Int:D' with a defined value" or so
FROGGS ahh, that works
thanks :o)
jnthn Maybe without the quotes around the typename given it's already uppercase
FROGGS yeah 18:54
18:55 vendethiel joined
FROGGS I'm attempting to force the initializer now, then we have less problems later 18:55
18:55 laouji left
nine jnthn: oh I just saw that I've never answered your question about Bool's key and value methods in BOOTSTRAP. I'm afraid they have to be there, because EnumHOW.add_enum_value expects the $value object to have them. Well, we could add a version of the method that takes the key and value separately. But I'm not sure it's worth it 18:58
jnthn nine: OK, let's keep it as you have it. 18:59
atweiden gtodd: there's an example in roast/S32-temporal/DateTime-Instant-Duration.t
gtodd atweiden: oh goodee ... would be nice to see a simple eg. in the docs :-) 19:00
atweiden source is rakudo/src/core/Temporal.pm 19:03
19:03 chenryn_ joined
gtodd atweiden: it seems a round about way to say you want seconds in the format ... but I guiess olde-fashioned strftime is round about too 19:06
19:08 chenryn_ left
atweiden i notice the default-formatter has the optional arg Bool :$subseconds that would likely fix this 19:09
19:09 yqt joined
atweiden how do you pass :subseconds to DateTime.new rakudo default-formatter? 19:09
TimToady m: [\o](1/* xx *)[$_](10).say for 0..10 19:10
camelia rakudo-moar 54746f: OUTPUT«0.1␤10␤0.1␤10␤0.1␤10␤0.1␤10␤0.1␤10␤0.1␤»
19:10 pierrot_ joined, domidumont left 19:11 pierrot left 19:15 pierrot_ left, pierrot joined
grondilu well it's not much more amusing than say [\*] -1 xx * but still 19:16
nine m: enum Foo <a b>; say Foo.isa(Foo);
camelia rakudo-moar 54746f: OUTPUT«True␤»
19:19 pierrot left 19:20 pierrot joined, laouji joined 19:24 AlexDaniel joined
jdv79 so, very _very_ prelim metacpan preview/prototype 19:24
hack.p6c.org:5001/release/JDV/Data-Selector-1.02 19:25
19:25 pierrot left, laouji left
jdv79 please don't pound it or expect it to work well cause it where i'm working ^H 19:25
19:25 pierrot joined
jdv79 maybe tomorrow i'll try to load more of the modules from the eco into instance 19:25
btyler I think #perl6 might have death-hugged it :)
jdv79 its super slow ^H 19:26
btyler oh, there it goes, loaded now
jdv79 but should come back in < 5s or so =
i only loaded the one dist so far
and that took a few hours
so dinner &
btyler so is this a straight metacpan install that happens to have p6 in the db?
19:26 sjn_phone joined
btyler or just using the metacpan style/layout? 19:26
jdv79 yeah, its just minimal work to get p6 support 19:27
nine jnthn: are the type objects defined in BOOTSTRAP somehow replaced later on? 19:28
btyler cool, great start :)
19:28 pdcawley left 19:29 larion joined
gtodd atweiden: I think the default formatter for DateTime constuctor doesn't groks subseconds you get iso-8601 so you'd need to have formatter function that does it 19:30
but IANAP
6H
:)
TimToady nine: perhaps the BOOTSTRAP versions are counted in the { ... } category somehow and automatically augment?
19:31 pierrot left, pierrot joined, llfourn joined
jnthn nine: No, type objects are forever. 19:32
TimToady are the boot classes counted as stubbed? 19:33
nine jnthn: that's very odd, because use nqp; say Bool.WHERE; say Bool.^mro[0].WHERE; gives me 2 different WHEREs. But Bool.^mro[0] is definitely Bool
19:33 ][Sno][ left
jnthn o.O 19:33
TimToady: Yeah, we don't compose them in BOOTSTRAP
TimToady: So effectively stubbed 19:34
nine That's why Bool.isa(Bool) gives False with EnumHOW using BaseType. With C3MRO, it worked however.
jnthn nine: Yes, that is weird.
19:36 llfourn left
nine enum Foo <a b>; say Foo.WHERE; say Foo.^mro[0].WHERE; # gives the same address for both 19:36
19:38 brrt joined
jnthn o.O 19:38
Very odd
nine The only difference that comes to mind is that Bool is created in BOOTSTRAP while Foo isn't.
jnthn Yeah, but lots of classes are created in BOOTSTRAP
TimToady jnthn: re irclog.perlgeek.de/perl6/2015-10-05#i_11318366 I think we actually need to think about how to expose the metaop factories as HOP features, not just hide 'em 19:39
19:39 lizmat joined
TimToady at least, those that don't already have exposure 19:39
triangle, cross... 19:40
jnthn TimToady: I'm fine with choosing to expose things; it's the things we don't choose to that I'm worried about :)
pmurias Zoffix: re perl6.xtatik.org, I think there are maybe better examples for someone new to the language then unicode operators and the infinite fibonacii sequence 19:41
atweiden gtodd: github.com/atweiden/config-toml/co...4b2f035bd8 19:42
pmurias Zoffix: I would aim more towards "wow, that's useful"/"sounds sane and good" mix rather then "hmm, how peculiar" 19:43
gtodd atweiden: yaya! or that :-) 19:45
grondilu pmurias: what example would you chose? 19:49
btyler a simple class (maybe consuming a role), a grammar, and a non-blocking socket server (yes, I'm already poking at these :)
jnthn Time for me to rest...be back tomorrow o/ 19:52
lizmat good night, jnthn
m: say .2 - .3 + .1 == 0 19:53
camelia rakudo-moar 54746f: OUTPUT«True␤»
lizmat ^^^ perhaps a small example for the home page ?
gtodd heheh
19:53 patrickz joined
gtodd "we invite you to try the above in your current favourite language" 19:53
btyler I think it is super cool (honestly it never struck me how wrong this was in most languages I used, I just accepted it), but it doesn't seem _inspiring_ (to me, at least) 19:54
and using as a put-down on other langs will just invite bickering
or petty questions about performance and memory usage 19:55
[Coke] I think it's a good example for the front page.
lizmat nine: irclog.perlgeek.de/perl6/2015-10-06#i_11328015 # I guess they fell in the cracks between timotimo and me 19:56
will look at them tomorrow... unless someone beats me to it 19:58
brrt good * #perl6
lizmat brrt o/ 19:59
nine lizmat: if anyone's to blame, it's the people who don't work on Perl 6 :)
brrt ohai lizmat
gtodd btyler: good point 20:00
20:00 ][Sno][ joined
btyler gtodd: don't get me wrong, Ovid used this to _great_ effect in his FOSDEM talk :) 20:00
the whole room went: "....woah, you're right, that's so busted" 20:01
gtodd btyler: maybe just have it presented more as a small (?!) "attenton to detail ..." thing that makes perl6 likeable :-D
TimToady I list some of those in my talk last night
*ted
%% 20:02
now - INIT now
gtodd just a bunch of little things ... sort of how in in astronomy little things like air and gravity can make a planet interesting :)
TimToady m: say „a“ … „zz“ 20:03
camelia rakudo-moar 54746f: OUTPUT«(a b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad ae af ag ah ai aj ak al am an ao ap aq ar as at au av aw ax ay az ba bb bc bd be bf bg bh bi bj bk bl bm bn bo bp bq br bs bt bu bv bw bx by bz ca cb cc cd ce cf cg ch ci cj ck cl cm cn co cp…»
20:03 darutoko left
gtodd :-D 20:03
^elyse^ :-D 20:04
lizmat I would not mention an invitation to try in other languages... it should be implicit :-)
20:04 chenryn_ joined 20:05 adhoc joined
brrt hmmmm 20:05
i'm ambivalent about it
it's typical language boasting 20:06
gtodd lizmat: yes
TimToady last night I blew their mind with a simple 1,2,4 ... 2**32
lizmat brrt: it's a USP of Perl 6
m: say 1,2,4 ... 2**32 20:07
camelia rakudo-moar 54746f: OUTPUT«(1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288 1048576 2097152 4194304 8388608 16777216 33554432 67108864 134217728 268435456 536870912 1073741824 2147483648 4294967296)␤»
brrt i agree. i'm trying to imagine how i'd react to it if i didn't already think perl6 was cool
TimToady "How does it know?!" someone cried out.
brrt i think i read in the synopses how that is done 20:08
it's clever for sure :-)
constant ratios iirc
20:09 chenryn_ left
TimToady simple junctions are also kinda wow 20:09
lizmat m: say now - BEGIN now 20:11
camelia rakudo-moar 54746f: OUTPUT«0.0099340␤»
TimToady simple mixins are also wow, though of course later we have to tell them not to do it :)
brrt ooc, when do you consider a mixin simple? 20:12
TimToady I try to discourage the BEGIN variant, because separate compilation
well, 0 but True *looks* simple
btyler I'd humbly suggest that there's a lot of value in showing things that we don't later need to say "but really that's a poor plan and your future colleagues will not be happy" :)
brrt ah, yes
lizmat m: say now - INIT now 20:13
camelia rakudo-moar 54746f: OUTPUT«0.0020325␤»
brrt btyler: depends on whether we want to continue the perl-is-write-only meme i'd think
and fwiw, i don't consider that true at all. but the cool tricks of perl5 are also the things that have made its reputation 20:14
PerlJam remembers a time when regex were black magic to most people and decried in Perl ... then every other language adopted them in some way. 20:15
TimToady m: say bag(<a b c d c d e >).pairs 20:17
camelia rakudo-moar 54746f: OUTPUT«(a => 1 c => 2 b => 1 e => 1 d => 2)␤»
btyler personally I'd suggest that examples should emphasize concise manipulexity, and if some whipitupitude is also demonstrated, that's mostly a bonus
TimToady that's pretty cool
^elyse^ PerlJam: and they all do it wrong :P
lizmat m: <a b c d c d e>.Bag.pairs.say
camelia rakudo-moar 54746f: OUTPUT«(a => 1 c => 2 b => 1 e => 1 d => 2)␤»
lizmat I like the .Bag better :-)
TimToady I don't
PerlJam ^elyse^: aye, but that's more a symptom of unevenly distributed future 20:18
AlexDaniel ^elyse^: there is fuzzy matching in python's “regex” package. I have no idea whether that's a good idea or not, but it looks very promising. 20:20
at least enough for me to say “hm, wow” 20:21
[Coke] ... Wow, I have to fix a bug that is impacting our IE8 users in production. Wish me luck. 20:23
btyler alert("go download chrome already!") # wishful thinking
PerlJam [Coke]: jetison the IE8 users; problem solved! ;)
AlexDaniel [Coke]: good luck! Yet it's not IE7, say thanks 20:24
[Coke] oh, these people already have chrome -installed-, even.
AlexDaniel
.oO(… just symlink IE to Chrome)
20:26
pmurias at least it's not IE6
20:26 rindolf left
pmurias grondilu: grammars, maybe some OO 20:27
20:30 Kcop joined
pmurias grondilu: if without parens is a small things that I like a lot 20:30
20:31 Xor_ left 20:37 tokuhirom joined
brrt hasn't seen the python regex module 20:38
20:38 dalek left, synbot6 left 20:39 dalek joined, ChanServ sets mode: +v dalek 20:41 tokuhirom left
PerlJam dir() looks like it's becoming File::Find-y. It just needs the recursion bit and File::Find will no longer be needed I think 20:41
20:43 sjn_phone left, pnu left, abaugher left, Loren left, [Tux] left, dylanwh_ left, camelia left, ashleyde1 left, psch left, tadzik left, Gardner left, clkao left, masak left, hobbs left, alpha- left
AlexDaniel brtt: pypi.python.org/pypi/regex this one 20:44
brrt oh thanks 20:45
20:45 pnu joined 20:46 abaugher joined, Loren joined, [Tux] joined, dylanwh_ joined, camelia joined, alpha- joined, ashleyde1 joined, psch joined, tadzik joined, Gardner joined, clkao joined, masak joined, hobbs joined, kst joined, asimov.freenode.net sets mode: +v camelia 20:48 ][Sno][ left 20:50 lichtkind joined, adhoc left, [Sno] joined, leont joined, colomon left 20:51 tokuhirom joined 20:52 simcop2387 left, bjz left 20:53 paulmr left, kaare_ left, brrt left, raiph joined 20:55 simcop2387 joined, rurban joined 20:57 bjz joined, tokuhirom left
lichtkind rurban: ich konnte dir nicht mal ne mail schreiben 20:59
21:01 abaugher left
patrickz .seen smls 21:03
yoleaux I saw smls 14 Sep 2015 18:53Z in #perl6: <smls> and pre-GLR, X treated it that way
21:05 chenryn_ joined 21:06 skids left, xfix left
FROGGS spectests a refactor of Foo:D/Foo:U 21:07
pink_mist is the :U there meant to mean it needs to be undefined? 0_o what's the use of that? 21:08
FROGGS pink_mist: the use to accept only types?
well, you can restrict subs/methods from receiving values 21:09
moritz there are use cases for that in the setting
multi method ACCEPTS(Mu:U) { } implements type checking via smart matching, for example
pink_mist ah, fair point
21:10 chenryn_ left
moritz git grep ':U\b' src/core/|wc -l 21:10
145
FROGGS I won't use :U on variables often I guess
21:10 pmurias left 21:12 n0tjack left
masak m: role Foo[::T] { constant a = T.new } 21:13
camelia rakudo-moar 54746f: OUTPUT«5===SORRY!5=== Error while compiling /tmp/JijrN9FMqs␤An exception occurred while evaluating a constant␤at /tmp/JijrN9FMqs:1␤Exception details:␤ 5===SORRY!5=== Error while compiling ␤ Method 'new' not found for invocant of class 'T'␤ …»
dalek ast: 73fc8c3 | FROGGS++ | S11-modules/importing.t:
adjust test for typed exception
21:17 bakedb left, bakedb joined, colomon joined
dalek kudo/nom: 7f0e6a5 | FROGGS++ | src/ (4 files):
refactor Type:D so that it requires an initializer

Calling .new on the type internally seemed to makes sence when you think of e.g. Int:D, but it will get problematic quickly when the constructor would need arguments... Therefore, an initializer has to be specified.
21:18
FROGGS .tell jnthn the initializer is now required... works on vars, attrs but not in sigs: github.com/rakudo/rakudo/commit/7f0e6a56ae 21:19
yoleaux FROGGS: I'll pass your message to jnthn.
FROGGS .tell lizmat can you adjust the weekly so that it states that 'my Int:D $b' is an compile time error? 21:20
yoleaux FROGGS: I'll pass your message to lizmat.
FROGGS $ perl6 -e 'my Int:D $b'
===SORRY!=== Error while compiling -e
Variable definition of type Int:D requires an initializer
21:20 laouji joined
lizmat FROGGS: will do 21:21
yoleaux 21:20Z <FROGGS> lizmat: can you adjust the weekly so that it states that 'my Int:D $b' is an compile time error?
FROGGS lizmat: thank you :o)
.tell #perl6 gnight
yoleaux FROGGS: What kind of a name is "#perl6"?!
FROGGS pfff
colomon o/
lizmat m: my Int:D $b 21:22
camelia ( no output )
lizmat FROGGS: is that something new?
if it *is* something new, I'd rather mention it next week 21:24
21:25 yqt left, laouji left 21:29 whowut joined, whowut left
Ulti gtodd sorry dont think I finished the histogram stuff in Stats 21:33
if you need it I can get back to it :3
lizmat decommute& 21:34
21:34 lizmat left 21:35 llfourn joined 21:41 lichtkind_ joined 21:45 literal left
masak somewhat confused article with a misleading title, but still somehow nice: www.pigdog.org/auto/software_jihad/.../3138.html 21:45
up on HN front page right now
21:47 lichtkind_ left 21:48 laubster joined 21:49 bjz left 21:51 laouji joined, ribasushi left
Ulti there was a bit of an under current of Perl 6 being the next small iteration of Perl in that article 21:52
Zoffix I didn't feel it. I liked the article except for the title. 21:53
21:53 ribasushi joined 21:55 bjz joined, laouji left 21:56 ^elyse^ left 21:57 bjz left 22:00 colomon left 22:01 spider-mario left
Ulti I liked it too but comparing/contrasting only against Perl 5 to describe Perl 6 feels a little odd to me. In my mind I fit Perl 6 in some weird space between Perl 5, Icon/Unicon and Java 22:03
22:04 bjz joined, patrickz left
_itz www.10zenmonkeys.com/2015/10/06/the...ed-perl-6/ 22:05
22:06 laouji joined, AlexDaniel left, chenryn_ joined
Ulti Icon especially its like Perl 6 is riffing off of some of the ideas there even stuff like $* variables are like & prefix variables in Icon then you have all the generator lazy list and sequence stuff 22:08
string scanning in Icon is a bit like a given block with a grammar etc.
kst _itz: I can't access www.10zenmonkeys.com 22:09
22:09 isacloud left, preyalone left 22:10 chenryn_ left
Ulti Icon could have done with the concept of meta operators though, instead of just having an insane number of multi character operators for each built in type 22:11
22:11 laouji left, bjz left
_itz downforeveryoneorjustme.com/http://...onkeys.com 22:11
22:11 woodruffw joined
_itz developers.slashdot.org/story/15/10...s-perl-600 22:12
Ulti guess thats what you get with >20 year gap in language development, better abstractions
masak Perl 600!
it's so weird to read these outside-the-echo-chamber reports on something you're involved in. 22:13
22:13 Moskon joined 22:14 ex-parrot joined
masak so far, these articles really enjoy making it seem (in the titles) that Perl 6 arrived Just Now. 22:14
_itz www.reddit.com/r/programming/comme..._perl_600/
Ulti heh fully formed from the hands of the gods
22:15 eythian joined
pink_mist to be fair, they've been waiting to use that headline for more than 10 years, so I wouldn't blame them for jumping the gun a bit :P 22:15
chomping at the bits, as it were 22:16
masak it's using up a finite resource, is all. a bit like the boy who cried wolf did.
Ulti also once I get my bioinformatics tests to fully pass again I suspect the runtime has gone from 10s to 3s post GLR... 22:17
22:18 jevin joined
_itz I wonder where the 6.0.0 meme started 22:18
leont spent yesterday back in bioinformatics land (PhD defence of a friend), gave me some ideas for perl 6 22:20
Ulti leont well there are two different bioinfo repos in Perl 6 plenty of room for a third :D 22:21
22:21 TEttinger joined 22:22 PotatoGim left, clkao left, Spot__ left
leont He made some interesting observations regarding big data 22:22
Almost all current high level tools are based on loading data into ram and then processing them in a single process/thread 22:23
This is suboptimal
22:24 thowe joined
Ulti huh... apart from all the big data tools which dont do that 22:24
22:24 ex-parrot left
leont Which aren't really high level tools 22:25
Ulti unless you mean specifically sequence data and stuff like assemblers
leont (I'm probably phrasing things imprecisely)
Ulti but yeah research software in general its rare anyone would write it to not be architected in that way 22:27
I think R now has a Hadoop back end but not sure how it works and what the limitations are?
is there a special name for an elephant race? because that should be the name of method that takes a Perl 6 map and runs it via hadoop 22:28
thowe The book "Footfall" might have that name in it. The aliens were like elephants. 22:29
Ulti well weirdly Germany is famous for elephant racing according to Wikipedia and they love giant weird words 22:30
thowe Oh, that kind of race... I was thinking "race of beings" not "race to the finish" :P
Ulti heh
no as in .race() :) 22:31
to the finish line
.elefantenrennen()
not as catchy as one might hope
masak 'night, #perl6 22:32
Ulti o/
22:33 isacloud joined 22:35 kid51 joined, laouji joined 22:38 preyalone joined 22:39 laouji left 22:40 laouji joined
dalek kudo-star-daily: 258ea01 | coke++ | log/ (9 files):
today (automated commit)
22:40
kudo-star-daily: 5528913 | coke++ | log/ (9 files):
today (automated commit)
kudo-star-daily: d38bb61 | coke++ | log/ (9 files):
today (automated commit)
rl6-roast-data: ad0835f | coke++ | / (9 files):
today (automated commit)
rl6-roast-data: 3343257 | coke++ | / (9 files):
today (automated commit)
22:40 firstdayonthejob joined
thowe Is the latest Rakudobrew a good way to get this new release? Or no? 22:40
Last time I build Rakudo was YAPC::NA 22:41
22:41 Moskon left
thowe Where's the big official announcement? Or is that still waiting for Chistmas? 22:41
[Coke] still waiting. The version you get now is going to be pretty close, though, give it a spin. 22:43
22:43 borkus joined
thowe I want to. I love that every number is rational. 22:43
22:43 mattprelude joined
thowe Is there a Pg module? 22:43
gfldex thowe: list of modules that may or may not work with a newish rakudo modules.perl6.org/ 22:44
it's still pretty beta, expect bugs 22:45
pink_mist there's a DBIish .. no clue if there's a Pg driver for it
gfldex there is: github.com/perl6/DBIish/blob/maste...ish/Pg.pm6
pink_mist github.com/perl6/DBIish/#pg-postgresql yeah, just found this =) 22:46
22:46 borkus left
thowe At this point, for me, it looks like a good place to learn some nifty programming concepts. 22:46
What I've read has blown my mind.
22:47 atweiden left
gfldex thowe: you may want to watch this video then www.youtube.com/watch?v=JpqnNCx7wVY 22:47
thowe That might harm my fragile little head. 22:48
22:48 PasserBy joined
thowe Is Parrot just not gonna be a thing anymore? 22:48
gfldex jnthn++ tends to be fairly gently with heads that are not his own
it looks like Parrot went the way of the Dodo for now
thowe They seem to continue to release... 22:49
I don't know what people are doing with it.
But I assume if it stays alive, that it will be one implemenation of Perl6 eventually(?)
gfldex parrots object system doesnt fit that well to the one perl6 is using 22:50
[Coke] parrot might still be a thing, but its path diverged from Perl 6's several years back.
thowe oh, 7:30 tonight. 22:52
decent audio/video will be available I hope
ilmari pmthium.com/2015/02/suspending-rakudo-parrot/
22:52 cognominal left 22:53 skids joined 22:54 jevin left 22:55 jevin_ joined 22:56 dshahin joined 22:58 pdurbin joined 23:00 RabidGravy left, grondilu left
pdurbin ah. Carl Mäsak on Twitter: "Nope. Projected release date is still this Christmas. #perl6 t.co/sL2PsN9sNM" - twitter.com/carlmasak/status/651528795053268992 23:02
23:03 lichtkind left 23:04 firstdayonthejob left 23:07 chenryn_ joined 23:10 lsm-desktop joined 23:11 BenGoldberg joined, chenryn_ left
[Coke] Perl 6 has been releasing every month for years now. 23:12
23:12 PasserBy left
[Coke] you just got caught up in the excitement. It's OK. :) 23:13
pdurbin well, people on Twitter are excited. I started looking for things tagged as 6.0 but couldn't find anything :) 23:14
ilmari the curren version is 6.Birthday
+t
23:17 gensym left 23:18 moznion_ left 23:20 eroc joined 23:21 eroc left 23:22 colomon joined 23:23 pecastro left 23:24 clkao joined 23:26 doublec joined 23:27 laouji left 23:30 pecastro joined
gtodd Ulti: I was using the Stats modulle's hist function to try and figure out how given/when would work toset and just print the $bin_size first then figure out how to select units etc. ... the rest of the function would be sort of mechanical (printing horizontal/vertical ascii/utf bars etc. 23:34
23:34 Spot__ joined
gtodd Ulti: maybe with an eventual plan to hook into a rendering backend/target like R does ? :-) SVG PDF etc etc 23:35
23:35 laouji joined
gtodd hehe 23:35
23:40 pecastro left, laouji left 23:41 pecastro joined, Praise joined 23:45 leont left 23:50 flussence left 23:51 flussence joined 23:55 heythee joined
heythee hi there 23:55
23:56 heythee left 23:57 kk4vcz joined