»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend! | Rakudo Star Released!
Set by diakopter on 6 September 2010.
diakopter sorear: what kind of textual pattern does that grammar express 00:05
sorear diakopter: the first, same as /bc/, the second, same as /<?>/ 00:07
but the important part is that $/.<X> == 1 after both 00:08
or, if it happens from a subrule, maybe $/.<a><X> == 1
00:28 wamba joined 00:29 Schwern joined 00:39 tylercurtis joined 00:43 nimiezko left 01:04 Eevee left 01:06 Eevee joined 01:09 wamba left
TimToady whoever put up rosettacode.org/wiki/Numerical_inte...ion#Perl_6 the math is wrong at least three different ways, some of which apparently was already wrong in whichever source you copied, which is unattributed 01:17
midpoint is wrong one way, the trapezoidal method is wrong at least two ways
the perpetrator would appear to be someone from Brazil 01:18
01:25 lichtkind left 01:28 azert0x left 01:33 mfollett joined 01:34 hercynium left
TimToady I suspect it was transcribed from the Java example, which would explain both its verbosity and 2/3 of its wrongness 01:37
01:39 ash_ left
allbery_b heh 01:57
02:00 flatwhatson_ joined
flatwhatson_ hi, is there an implementation of perl6 grammars that i can use in perl5? 02:01
perigrin Perl6::Rules, Regexp::Grammar, Parse::RecDescent all come to mind 02:04
02:04 rgrau_ left
allbery_b oh damn. now I'm gonna be tempted to use them :) 02:05
02:10 leprevost joined 02:20 whiteknight left
florz how would I go about accessing a capture from code within a regex following said capture? as in / (fo*) { here } / 02:25
flatwhatson_ Regexp::Grammar and Parse::RecDescent both look quite good! it seems the major difference is that P::RD lets you trigger actions during the parsing process, whereas R::G is more geared towards the perl5-style parse-then-process pattern
would that be an accurate assessment?
02:27 patspam joined, patspam left 02:32 jaldhar joined
perigrin flatwhatson_: well Regexp::Grammar is based on 5.10+ Regex extensions while Parse::RecDescent implement an entire Recursive Descent parser 02:33
so at a high level "yes" but that's partly because the implementations are wildly different
02:35 meppl joined 02:36 ash_ joined
flatwhatson_ sure, so it is the 5.10 engine's lack of parse actions that limits Regex::Grammars in that sense 02:37
but it seems that it would be more performant due to re-using the built-in engine?
also i think the syntax would be less likely to give my co-workers aneurysms :) 02:38
02:39 b0m1ro left 02:41 mfollett left, mfollett joined, ash_ left 02:42 Holy_Cow joined 02:43 Holy_Cow left
sorear florz: just use $0 02:47
02:51 mfollett left
florz sorear: oh, IC, captures are 0-based now, eh? thx :-) 02:58
sorear yeah 03:13
$0 is short for $/[0], it's zero-based like every other array
florz sounds sensible - thx :-) 03:18
03:28 ioio_1 joined 03:34 meppel joined 03:37 meppl left 03:38 sjn joined 03:41 meppel is now known as meppl, ioio_1 left 03:47 tom_tsuruhara joined 03:59 Schwern left 04:06 envi^home joined 04:08 envi^home left 04:10 qw3rty joined
qw3rty hi... i was just introduced to =pod annotations in perl5 today... I found it very interesting, i like its simplicity. The thing is... 04:12
does pod annotations evolved too with perl6?
tylercurtis qw3rty: yes, although I think it's not really nailed down yet due to not being fully implemented. 04:15
qw3rty is there any specs available???
tylercurtis github.com/perl6/specs/raw/master/S...tation.pod It's written in Perl 6 Pod. 04:17
qw3rty cool... thanks tylercurtis!
04:24 mfollett joined
mfollett Is it a bug that you can't call d, f, and such on an object of type Str and instead of to do $str.IO or was it a bug that you could at some point in the past? 04:29
04:33 gfx joined
tylercurtis mfollett: There is another possibility: both were correct. Which is actually the case, I don't know. 04:33
mfollett tylercurtis: that's true. I'm just seeing a lot of code around that does $str ~~ :f which, at least in * 2010-08, does not seem to be a thing you can do. 04:35
It would make sense that you could do either way and that way just isn't working at the moment though.
04:46 Chillance left 05:01 REPLeffect left, qw3rty left
mfollett Is there an English version of the Perl 6 tutorial that is listed on the perl6 wiki? 05:02
sorear mfollett: AIUI, both were correct *at the respective times*. We haven't frozen the spec yet. 05:09
mfollett sorear: oh, ok, thanks for the explanation. 05:10
05:13 jaldhar left 05:15 jaldhar joined 05:17 REPLeffect joined 05:18 meppl left
TimToady strings are generally not assumed to be filenames anymore 05:27
we broken that on purpose 05:28
broke, even 05:30
05:32 Guest23195 joined
mfollett TimToady: good to know, thanks 05:32
I'm kind of playing around with Web.pm, trying to bring it up to speed and it looks like Tags.pm is getting a "Null PMC access in invoke()" with terms like ::Tags{$tag} and ::Tags::EXPORT::DEFAULT{$tag} (for context here is a pastie pastie.org/1151608 lines 29 and 32). What was this originally doing and what is the now way to do it? 05:38
shortcircuit TimToady: rosettacode.org/wiki/Template:Incorrect is useful for flagging such errors, if one doesn't want (or can't) fix the code oneself. 05:42
05:43 Schwern joined
sorear looks like an exploitation of alpha' 05:46
s chronic parrot leakage
tylercurtis What does this mean? 05:47
std: $::
p6eval std 32123: OUTPUT«ok 00:01 114m␤»
sorear niecza: $::
p6eval niecza 10151ff: OUTPUT«===SORRY!===␤Non-simple variables NYI at /tmp/X7a8zGQzYJ line 1:␤------> $::⏏<EOL>␤Use of uninitialized value in concatenation (.) or string at /home/p6eval/niecza/src/Niecza/Actions.pm line 1388.␤Use of uninitialized value in
..concatenation (.) …
sorear huh
I thought I had an 'XXX STD bug workaround' sorry for that
I don't think it means anything at all 05:48
TimToady shortcircuit: I've found a fourth error in the Perl 6 code; it also comes from the Java I suspect 05:51
a fifth error, this one in the Perl transcription 05:59
06:04 Schwern left 06:05 jaldhar left 06:06 envi^home joined, Schwern joined 06:07 jaldhar joined
mfollett How do I dynamically add a sub to a module? 06:18
ingy hi shortcircuit :) 06:26
tylercurtis Good night, #perl6. 06:46
06:46 mfollett left 06:54 icwiener joined 07:09 brxue joined 07:15 rhebus joined 07:16 Ross joined 07:19 stepnem left
moritz_ good morning 07:30
sorear hello 07:35
07:38 icwiener left, tylercurtis left, icwiener joined 07:39 icwiener left, icwiener joined
rhebus morning 07:50
07:57 leprevost left
TimToady shortcircuit: okay, added the incorrections, thanks for the pointer 07:57
07:57 ashleydev left 07:58 ashleydev joined, sunnavy joined
TimToady I believe rosettacode.org/wiki/Numerical_inte...ion#Perl_6 is now correct, but I didn't fix the other languages :) 07:59
07:59 tadzik joined
Quadrescence TimToady: you look like weird al, just sayin' as I know it's relevant to the topic 07:59
TimToady I knew that
actually, he looks like me
Quadrescence very true; i dare you to write a perl program which calculates the similarity between two images by making your own novel metric. 08:00
TimToady mirrors.develooper.com/perl/history...c/al_wall/
Quadrescence dammit 08:01
moritz_ thought that TimToady stayed away from the more mathematical problems :-) 08:03
Quadrescence moritz_: This is a linguistic problem
08:04 cosimo left
moritz_ Quadrescence: I was talking about the numerical integration rosetta code 08:06
Quadrescence oh >_>
well numerical integration is linguistic too
i mean, what /is/ the etymology of 'quadrature' and 'cubature' 08:07
TimToady rosettacode.org/wiki/Talk:Numerical_integration has my carpage
08:08 leprevost joined
rhebus TimToady: won't you get lossage from numerical error in the endpoints? 08:09
TimToady what numerical error? this is all done in Rats
08:09 araujo joined
TimToady I don't think the function was complicated enough to get into Nums 08:10
rhebus huh
rakudo: say 1, *+0.1 ... 2.0
p6eval rakudo ce565f: OUTPUT«11.11.21.31.41.51.61.71.81.92␤»
rhebus rakudo: say ~(1, *+0.1 ... 2.0)
p6eval rakudo ce565f: OUTPUT«1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2␤»
TimToady we like our Rats
rhebus rakudo: say ~1, *+0.1 ... 2.0
tadzik rakudo: say ~(1, 1.1, 1.2 ... 2.0)
p6eval rakudo ce565f: OUTPUT«11.11.21.31.41.51.61.71.81.9␤» 08:11
rakudo ce565f: OUTPUT«1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2␤»
tadzik no need to closure
rhebus hmm, so i'm losing the endpoint because i'm stringifying the first term, not the whole sequence?
tadzik hmm
rhebus rakudo: say ~1, 1.1, 1.2 ... 2 08:12
tadzik rakudo: say (~1, *+0.1 ... 2.0).perl
p6eval rakudo ce565f: OUTPUT«Unable to figure out pattern of series␤ in 'infix:<eq>' at line 1␤ in '_HELPER_generate-series' at line 849:CORE.setting␤ in 'infix:<...>' at line 871:CORE.setting␤ in main program body at line 22:/tmp/cC1eMHAtxS␤»
rakudo ce565f: OUTPUT«("1", 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9)␤»
tadzik funny
rhebus tadzik: yeah, it's doing Nums not Rats
TimToady rakudo: say 1.0e0, 1.1, 1.2 .. 2.0 08:13
p6eval rakudo ce565f: OUTPUT«11.11.2␤»
TimToady rakudo: say 1.0e0, 1.1, 1.2 ... 2.0
p6eval rakudo ce565f: OUTPUT«Unable to figure out pattern of series␤ in 'infix:<eq>' at line 1␤ in '_HELPER_generate-series' at line 849:CORE.setting␤ in 'infix:<...>' at line 871:CORE.setting␤ in main program body at line 22:/tmp/bBI7jTvNeb␤»
Quadrescence arithmetic sequences with floats seems like a pretty bad idea
TimToady rakudo: say 1.0e0, 1.1e0, 1.2e0 ... 2.0
p6eval rakudo ce565f: OUTPUT«Unable to figure out pattern of series␤ in 'infix:<eq>' at line 1␤ in '_HELPER_generate-series' at line 849:CORE.setting␤ in 'infix:<...>' at line 871:CORE.setting␤ in main program body at line 22:/tmp/tt9fvILPsp␤»
TimToady it can't figure it out because the ratios aren't exactly the same 08:14
Quadrescence ratios or differences?
TimToady rakudo: say 1.0e0, 1.1e0 ... 2.0 # should work
p6eval rakudo ce565f: OUTPUT«11.11.21.31.41.51.61.71.81.9␤»
TimToady work better, I mean
rhebus heh
TimToady with three terms, the differences or ratios have to match 08:15
rhebus right
Quadrescence rakudo: say 1.1e0 - 1.0e0, 1.2e0 - 1.1e0
p6eval rakudo ce565f: OUTPUT«0.10.0999999999999999␤»
TimToady it's...probably good that they don't
rhebus rakudo: say ~1.0e0, 1.1e0, 1.2e0 ... 2.0
p6eval rakudo ce565f: OUTPUT«Unable to figure out pattern of series␤ in 'infix:<eq>' at line 1␤ in '_HELPER_generate-series' at line 849:CORE.setting␤ in 'infix:<...>' at line 871:CORE.setting␤ in main program body at line 22:/tmp/CtYku1hYEZ␤»
08:16 gfx left
TimToady rakudo: say 1.1 - 1.0,' ', 1.2 - 1.1 08:16
p6eval rakudo ce565f: OUTPUT«0.1 0.1␤»
Quadrescence rakudo: say 1.1e0 - 1.0e0, ' ', 1.2e0 - 1.1e0
p6eval rakudo ce565f: OUTPUT«0.1 0.0999999999999999␤»
Quadrescence rakudo: say (1.1e0 - 1.0e0)-(1.2e0 - 1.1e0)
p6eval rakudo ce565f: OUTPUT«2.22044604925031e-16␤»
TimToady rakudo: say (1.1 - 1.0) - (1.2 - 1.1)
p6eval rakudo ce565f: OUTPUT«0␤» 08:17
TimToady Rats++
rhebus TimToady: but what if the user passes in a Num for $a or $b? :)
Quadrescence are those rationals internally? or fixed point?
TimToady rakudo: say 1.1.perl
p6eval rakudo ce565f: OUTPUT«11/10␤»
rhebus rakudo: say (1,1.1 ... 2).perl
p6eval rakudo ce565f: OUTPUT«(1, 11/10, 6/5, 13/10, 7/5, 3/2, 8/5, 17/10, 9/5, 19/10, 2/1)␤»
Quadrescence TimToady: have you considered using fixed point arithmetic instead 08:18
TimToady rhebus: we call that DIHWIDT
Quadrescence rationals just get unwieldy quickly; i just wrote a chapter on this in a book i'm writing not too long ago actually
rhebus TimToady: it's reasonable to make these things explicit in example code. sub leftrect (&f, Rat $a, Rat $b, Int $n) 08:19
TimToady Quadrescence: see S02:704 on that subject 08:20
rhebus: yes, wouldn't hurt
rhebus shall i add it in? 08:21
Quadrescence TimToady: "i just wrote a chapter on this in a book i'm writing not too long ago actually" <-- Is this redundant because I said "just" and "not too long ago"?
(I don't know how much of a grammarian you are. :) 08:22
sorear TimToady: e0 forces Num now?!
TimToady also, S02:733 discusses an optimization that makes rats behave a bit like fixed point by being lazy about reducing
08:22 cjk101010 joined
TimToady sorear: it's been that way for months :) 08:22
actually, S02:751 discusses the lazy idea 08:23
Quadrescence: redundancy is good for communication; I'm a linguist, not a grammarian 08:24
Quadrescence TimToady: Did you ever write anything about redundancy in programming (in the linguistic sense, not any datum sense)? 08:25
I very vaguely recall you saying something about this somewhere from years ago.
TimToady not specifically
I've probably mentioned the principle from time to time
I should probably go to bed; have to get up in 4.5 hours or so... 08:27
TimToady --> SleepState 08:29
08:32 masak joined
masak привет, #perl6! 08:32
08:32 rhebus_ joined 08:33 rhebus left 08:34 rhebus_ is now known as rhebus 08:38 tadzik left
rhebus rakudo: say do $_ for 1..10 08:39
p6eval rakudo ce565f: OUTPUT«12345678910␤»
masak rakudo: say 1..10 08:40
rhebus hmm, that doesn't work on my local rakudo*
p6eval rakudo ce565f: OUTPUT«12345678910␤»
masak star: say do $_ for 1..10
p6eval star 2010.07: OUTPUT«␤»
masak confirmed.
08:44 rhebus_ joined, rhebus left 08:45 rhebus_ is now known as rhebus
rhebus goddamn wireless goblins 08:45
08:45 barney joined
masak one would have thought that the big Goblin De-wire-ification of 1996 would have led to something good. but no. 08:45
rhebus :) 08:46
moritz_ \o
masak o/
rhebus moritz_++ # irclog
masak moritz_: I have another exception for our single wtop so far in github.com/moritz/perl6-wtop/blob/m...ctices.pod 08:47
moritz_: I tend to write $a..$b in some circumstances.
(and $a .. $b in others, but still)
moritz_ masak: agreed. It's not a very hard rule 08:48
moritz_ would like to move all S03-operators/series*.t to S03-series/*.t 08:49
rhebus masak, moritz_: could you have a list of pending wtops? it'd be nice to see what you guys are thinking about as well as what you've agreed
masak moritz_: go right ahead.
rhebus RFC-style 08:50
moritz_ rhebus: no objections
masak rhebus: heh :) I probably have some, but I can't think of any right now.
at least I don't think I can.
moritz_ too
rhebus there was one mentioned the other day... don't overuse junctions
moritz_ should be more consequent in writing them down...
rhebus or something like that
sorear wtop? 08:51
moritz_ right, what rhebus++ said
masak sorear: well-thought-out practice
sorear: it's a tongue-in-cheek riff off "best practices"
rhebus i think the best way to arrive at best practices is to write a load of ideas down and criticise the $expletive out of them and see which ones stay standing at the end 08:52
masak hm, maybe. 08:53
08:53 Ross left
moritz_ rhebus: I disagree. At the start of a best practise there's code, lots of code 08:53
masak in some cases it's hard to speak about best practices, because much Real Code out there is still shock-full of workarounds.
rhebus moritz_: i don't understand. at the start of the best practice document, at the start of the best practice design? 08:54
masak making the best practices hard to identify.
rhebus: what moritz_ is getting at is that people talking is worth less than people with lots of experience.
rhebus masak: quite. The criticism must be based on experience
moritz_ rhebus: at the start of a best practise there should be lots of real-world code that accomplishes real-world tasks 08:55
08:55 Ross joined
masak experience with Perl 6 as well as with programming in general. 08:55
rhebus at the start of the design of a best practice? at the start of the best practice document?
masak rhebus: I don't understand that question. 08:56
moritz_ rhebus: the best practise comes before its document
masak I think 'at the start' was meant figuratively.
rhebus what I'm saying is that once a best practice is identified, it must be tested before being accepted generally
that testing must involve real code and real experience 08:57
but, for example, many of damian's pbps were found to be wanting. eg inside-out objects 08:58
08:58 flatwhatson_ left
rhebus (though pbp is an excellent book) 08:58
masak inside-out objects were run over by the Moose train.
sorear best practices should always come from an examination of already-working code
masak and they in turn superseded much of Damian's OO book.
moritz_ well, even the people using inside-out objects don't typically don't use the modules recommended in PBP 08:59
rhebus i'm not sure best practices come from examining code, but rather from examining habits
08:59 cosimo joined
moritz_ rhebus: and they come from making mistakes 08:59
rhebus i see best practices as a behavioural thing rather than a purely code-based thing
moritz_ and spotting mistakes in other people's code 09:00
rhebus i agree
masak nod 09:01
it's really about discipline, and disciplines.
09:01 flatwhatson_ joined
rhebus but you can't know if a proposed best practice is really good until you've tried coding consciously under that best practice 09:01
moritz_ that's why I wrote that code comes first 09:02
or phrased differently, that's another reason
rhebus well I see it like this: code -> identify problems -> identify best practices -> test bp with more code
masak it's not a best practice until it has been practiced a lot by people who consider it best. 09:03
moritz_ code -> identify problems -> find a solution -> write code with the solution -> discuss -> write up
rhebus masak: that's what i mean. but those people won't consider it best unless they've been asked to consider it as a possible best practice
moritz_: i think more code is needed in the discuss stage
masak rhebus: right.
moritz_ probably, yes 09:04
rhebus and this process can be iterative -- find problems with the bp -> refine bp to accomodate those problems -> test new bp
anyway
masak a practice that I consider to be very good, but that is non-specific to Perl 6, is "use vertical alignment to favourably expose meaningful patterns in your code" 09:05
or, put more simply, "put similar things under each other"
rhebus that's why i want to see proposed and hypothetical wtops, to try them out and see how they affect my coding 09:06
masak and that underlies my total avoidance of tab characters, because I value vertical alignment.
rhebus i imagine a number of damian's pbp can be lifted wholesale into perl 6
masak rhebus: yes. and I code under a number of them daily.
moritz_ can anybody please check if S03-operators/nesting.t or S03-metaops/reverse.t loop under current rakudo? 09:08
rhebus ah here's something i've been wondering: is it ok to use ~~ instead of ==?
moritz_ it loops here, but I have some local patches
masak rhebus: I consider it sloppy to use ~~ when == is what's intended. 09:10
rhebus: but it works, yes.
moritz_ rhebus: IMHO == has the advantage of telling exactly what kind of comparison is wanted
so it's better
*if* you know that you want numeric comparison
rhebus yes, supposing that you do
moritz_ if you don't, ~~ is correct 09:11
09:12 mikehh left
masak ~~ casts to the rhs. == casts to numeric. so it depends on what you want. 09:12
in the case where the rhs is numeric, they're the same. but == shows intent better.
rhebus so 1 ~~ "1x" is not the same as 1 == "1x"? 09:13
masak right.
sorear no current compiler is smart enough to handle ~~ quite well
masak the former is string comparison.
09:13 wamba joined
moritz_ rakudo handles ~~ pretty well, just not fast 09:13
masak I don't know if I consider it a "best practice" to prefer == whenever that's what's intended, or just common sense :)
moritz_ pushed to perl6-wtop 09:14
09:14 wamba left, wamba joined
masak for what it's worth, I don't see the value of differentiating practices.pod and rfc.pod. 09:15
one sounds more official than the other, but who determines when something is worthy of moving up to practices.pod? 09:17
moritz_ good point
you're welcome to join them, and remove the comment from README
masak does so
moritz_ masak++
rhebus well i guess you could sort practices.pod by number of champions, such that more-championed practices rise to the top
09:18 Steppenwolf joined
rhebus but that will lead to horrible diffs 09:18
moritz_ aye
I'd rather prefer the presenting software (if any) to do the sorting
rhebus yeah
09:22 sunnavy left, ashleydev left 09:23 frew left 09:24 rgrau_ joined 09:28 sunnavy joined 09:29 ashleydev joined 09:30 frew joined
moritz_ rakudo: say ~(1, 3 ... 9) 09:36
p6eval rakudo ce565f: OUTPUT«1 3 5 7 9␤» 09:37
09:42 kjeldahl joined 09:46 azert0x joined 09:47 flatwhatson_ left, flatwhatson_ joined
sorear S05 specs that regexes return lists of Cursor, but I (and viv) think returning lists of Match makes some more sense 09:48
moritz_ github.com/rakudo/rakudo/pull/4 09:51
09:58 whiteknight joined
masak writes his first given/when in Perl 5 10:03
10:03 barney left 10:08 mikehh joined 10:11 baest left
rhebus moritz_: it seems the spec has changed to remove inexact endpoint dwimmery 10:17
which would make 'AA' ... 'Z' infinite
masak works for me. :) 10:18
I don't think I'd ever want that to dwim to anything anyway... 10:19
rhebus github.com/perl6/specs/commit/69247...2555e17d8d
well, 2,3,*+2 ... sqrt $n would be nice 10:20
or 1,1,{$^a + $^b} ... 1000
masak in general, I think incrementing strings is pretty futile. beyond perhaps the very basics, which can be useful.
10:22 tadzik joined 10:24 Schwern left
rhebus it's not just strings which may want inexact limits 10:32
masak true. 10:33
rhebus but, as the spec shows, you can always use a closure
masak we've seen arrays in series lately.
rhebus 2,3,*+2 ... * >= sqrt $n
rakudo: [1,2,3].succ.say
p6eval rakudo ce565f: OUTPUT«Method 'succ' not found for invocant of class 'Array'␤ in main program body at line 22:/tmp/ZOOCh7wH05␤»
rhebus so it's be something like [1,2,3], [*[1]+1,*[2]+2,*[3]+3] ... *? 10:34
although it would seem more intutive to me to say 2,3,*+2 ... * < sqrt $n 10:36
iow, i think the limit should be a while condition, not an until condition
masak I had the same thought when I read the latest spec changes. 10:41
but then I remembered that the rhs of a series operator actually reads like an until condition.
rhebus right i'm off to the local park fun day 10:45
bouncy castle and face painting :D 10:46
masak have the appropriate amount of fun :)
rhebus heh 10:47
10:47 rhebus left 10:59 mberends joined
tadzik masak: what do you think about ripping the make-substitute part out of pls and rolling out something like Module::Builder? 11:05
masak tadzik: that might be a good idea.
pls would still carry of copy of such a module, of course.
tadzik sure thing 11:06
masak to avoid having dependencies.
tadzik so is neutro
also, I think of using the actual File::Find instead of writing own sub find-files
masak :)
tadzik crap, I can't build parrot due to something that looks like an ld bug 11:07
wklej.org/id/387595/
11:08 mberends left
masak tadzik: I get a 500 error on that paste. 11:10
"Unhandled Exception -- An unhandled exception was thrown by the application."
tadzik oh, I do too
whatever, seems like --as-needed in LDFLAGS was breaking it
11:20 nimiezko joined
jnthn afternoon, 6folk 11:23
masak o/ 11:25
jnthn yayitsmasak!
jnthn makes a bacon sarnie
masak lunch, now there's an idea... 11:26
11:27 flatwhatson_ left
jnthn breakfast for me :P 11:27
Well...brunch :-)
Plan for today: write my talk for Oslo, write a short-term roadmap for my grant, hack some more stuff into 6model. :-) 11:29
masak when's Oslo?
jnthn Tuesday evening
Thus the talk is at the "can't procrastinate it much longer" stage. :-)
masak likes that stage 11:30
but beware, the next stage is not as fun :)
not to mention the one after that... :/
or the one after that... :( 11:31
Plan for today: continue rescuing my use.perl.org content (it's going surprisingly well, much thanks to Perl 5), maybe design the new blog a bit, pick up Web.pm, have a look at Druid. 11:33
11:35 flatwhatson_ joined
tadzik hmm, I think something like Module::Tools would be nicer. There could be Module::Build, Module::Install, maybe Module::Starter 11:36
masak tadzik++ # doing module infrastructure 11:50
tadzik moritz_++ started some Module::Starter I think 11:51
11:53 isBEKaml joined 11:54 sftp left 11:58 am0c joined
masak aye, and mberends++ had a script in the proto repository before that. 11:59
create-new-project
isBEKaml OHHAI, #perl6! :)
masak: I was surprised when I received mails about bug reports from proto on github. :)
masak: not that I mind, though. But I thought proto was so long ago. 12:00
masak isBEKaml: aye, github's default emailing policy is "email contributors". it's possible to turn those emails off, though. I believe I did for my account. 12:01
isBEKaml: some people still seem to use proto. also, the pls branch is still in that repository.
tadzik I alredy msged mfollet about these
masak tadzik++
isBEKaml tadzik++
tadzik although I had no balls to close the issues with a message 'project abandoned' ;) 12:02
isBEKaml I think it'll be less confusing between proto and pls if we swap branches.
masak or put pls into a separate repository. 12:03
tadzik masak++
isBEKaml I think that would be even more confusing. Consider a new guy, what would he think if he saw two repos by the same author that "purportedly" does the same thing?
We have to have some way of pointing people towards pls than proto. 12:04
12:04 nimiezko left
masak isBEKaml: would that be more confusing than pls being in the main branch of a repo called 'proto', or less? :) 12:04
isBEKaml: I don't like to make decisions for people.
isBEKaml masak: ah, proto/pls/decision-making dilemma. ;) 12:05
tadzik well, there are now 3 projects doing purportedly the same thing
masak not when both alternatives still have different benefits over each other, at least.
tadzik . o O ( troth alternatives )
isBEKaml tadzik: not all by the same author. :) 12:06
tadzik well, I'm a neutro author and a pls contributor :)
isBEKaml that's one less confusion. ;)
tadzik and it's not very confusing to me :)
isBEKaml maybe because it's you? :)
tadzik :) 12:07
maybe
hm, Module::Build is actually an ufo-then-pls code stolen and put into a module 12:09
12:12 isBEKaml left 12:14 isBEKaml joined
isBEKaml Module::Build? 12:16
that makes it quad-alternative. =)
masak lunch &
12:16 masak left 12:20 zulon joined 12:22 Mowah left
isBEKaml oh, Module::Builder seems to be only on p5. :| 12:22
12:23 jasonmay joined
tadzik Module::Build can now Module::Build itself 12:25
isBEKaml I think you (intentionally) left out something. 6 on 6, 5 on 5, 5 on 6, 6 on 5? ;) 12:26
tadzik wait, what? :) 12:27
6 on 6 if I understand correctly
isBEKaml perl6 Module::Build can build itself on perl6? ;)
tadzik yes :)
isBEKaml tadzik++ 12:28
tadzik nothing fancy, it's just pls code stolen :)
isBEKaml tadzik++ # init worked! 12:31
tadzik :)
method install($project --> Result) -- now what is this? 12:33
I like this: run $command and die() ;) 12:37
12:37 stepnem joined 12:44 sftp joined 12:45 am0c left
tadzik phenny: tell masak github.com/masak/proto/blob/pls/pro...ncept#L284 -- shouldn't it be /\.pm6?$/ ? 12:45
phenny tadzik: I'll pass that on when masak is around.
12:47 aloha left, bacek left 12:49 zulon left
tadzik rakudo: my $a = List.new(1, 2, 3); for $a -> $b { say $b.perl } # is this a feature? 12:50
p6eval rakudo ce565f: OUTPUT«(1, 2, 3)␤»
isBEKaml rakudo: my $a = List.new(1,2,3); say $a; # does this seem different? 12:51
p6eval rakudo ce565f: OUTPUT«1 2 3␤»
tadzik well, I thought it'd be iterating over a list
isBEKaml rakudo: my $a = List.new(1,2,3); say $a.perl;
p6eval rakudo ce565f: OUTPUT«(1, 2, 3)␤»
gfldex rakudo: my $a = List.new(1,2,3); for $a.list -> $b { say $b.perl };
p6eval rakudo ce565f: OUTPUT«1␤2␤3␤»
tadzik hmm
isBEKaml yes.
tadzik thanks gfldex
gfldex i can even show you why this is a good feature 12:52
imagin you have a list of lists and you want to iterate over that list
then you need to keep the item context in $a -> $b
12:54 sftp left
tadzik I don't really get it 12:55
phenny: tell masak github.com/masak/proto/blob/pls/pro...ncept#L172 -- couldn't it just be try-CATCH? 12:58
phenny tadzik: I'll pass that on when masak is around.
12:59 sftp joined 13:10 orafu left, orafu joined 13:13 Mowah joined, zulon joined, redicaps joined 13:15 Patterner left, redicaps left 13:17 envi^home left
tadzik Nominal type check failed for parameter '$from'; expected Str but got File::Find::Result instead 13:17
13:17 icwiener left
tadzik shouldn't it be ok if File::Find::Result is Cool and has a .Str method? 13:17
13:18 Psyche^ joined, Psyche^ is now known as Patterner 13:19 Mowah left, icwiener joined 13:20 envi^home joined 13:23 Mowah joined
moritz_ no, Cool !~~ Str 13:24
it's the the other way round
jnthn tadzik: If you want to take anything and coerce it to a string, use "$from as Str" instead. 13:25
tadzik sounds safer than Str $from 13:26
more versatile
jnthn Depends what kinda safetly you want. ;-)
tadzik and less strict
jnthn Right. :-)
tadzik true :) 13:27
moritz_ subset Stringable of Any where .can('Str');
tadzik huh, I just did rm -r ./\~ 13:29
I'm still scared stiff :)
moritz_ accidentally create a file with non-globbed filename? 13:30
13:31 jhuni left
tadzik yeah 13:32
well, I created ~ in a directory
13:34 zulon left 13:37 rainerschuster joined, brxue left 13:51 tadzik left 13:57 masak joined
masak tadzik: ooh, reading pls? nice! 13:58
phenny masak: 12:45Z <tadzik> tell masak github.com/masak/proto/blob/pls/pro...ncept#L284 -- shouldn't it be /\.pm6?$/ ?
masak: 12:58Z <tadzik> tell masak github.com/masak/proto/blob/pls/pro...ncept#L172 -- couldn't it just be try-CATCH?
masak first question: seems it should indeed. good catch. fixing.
second question: obviously you don't have the deep mistrust of try/CATCH in Rakudo that I've built up. :) that's good, I guess. 13:59
either it still doesn't work, or my mistrust is out-of-date.
14:03 mberends joined 14:04 Chillance joined
masak mberends: ping 14:04
mberends masak: pong 14:05
phenny mberends: 10 Sep 12:11Z <masak> tell mberends that if he plans to submit OSDC.fr talks: deadline falls today.
masak mberends: what phenny said.
mberends oops
masak maybe not too late still. 14:06
mberends I saw jnthn's talks, and planned to discuss other talk ideas before making a submission, and then got sidetracked by $work :( 14:08
yes, quite possibly
masak hurry hurry
mberends think think, what would complement the existing schedule? 14:09
jnthn Something at 90 degrees to it? 14:10
;-)
masak |o_
dalek kudo: 62e168d | patrickas++ | src/core/operators.pm:
Starting to implement new series spec
 14:11
kudo: 953705a | patrickas++ | src/core/operators.pm:
There is no more need to check of the wrong side
kudo: 6483439 | patrickas++ | src/core/operators.pm:
Simplify the limit-reached code since geometric series don't require use of abs() anymore
14:11 dalek left
masak rakudo: my class Str {}; say 42.Str ~~ Str 14:11
p6eval rakudo ce565f: OUTPUT«0␤»
masak :)
jnthn :-) 14:12
14:12 mberends_ joined, dalek joined, ChanServ sets mode: +v dalek
masak mberends_: are you mberends' evil twin? 14:12
jnthn Or just the complement to mberends| ;)
masak :D
jnthn I'm sure mberends++ can submit a talk that gets the right angle on Perl 6. 14:13
masak groans appreciatively
14:13 mberends left
moritz_ patrikas++ # the big series hacker! 14:13
masak jnthn: sometimes you manage to angle your puns just right!
jnthn masak: Thanks for the complement. 14:14
masak LOL
14:14 mberends_ is now known as mberends
flussence (scrollback from ~17h ago): It'd be nice to have some sort of .POD method on functions for that. 14:16
masak flussence: that's spec'd. not implemented, though. I plan to. 14:17
jnthn It's .WHY speculated to be for that?
dalek kudo: fb4feba | moritz++ | t/spectest.data:
track test file rename
jnthn *Isn't
moritz_ it is
flussence ooh, .WHY looks a lot more useful than I imagined. 14:20
masak \o/ 14:21
flussence now appreciates the value of docs bundled with the language, after trying to write javascript without a net connection 14:25
masak programming without a net connection? curious. 14:26
and javascript of all languages, to boot :)
14:27 wamba left
flussence stuck on a boat for 3 hours, and perl6 isn't much fun on a slow netbook :( 14:28
14:28 xinming left
flussence (yet) 14:28
masak Rakudo has the slow feature built in right now. no need for a netbook to add that feature :) 14:29
14:29 xinming joined
flussence OTOH, the slow hardware gives me a reason to write faster code :) 14:31
14:31 meppl joined
masak my thinking too 14:31
14:32 tom_tsuruhara left
masak phenny: tell tadzik fixed, thanks! github.com/masak/proto/commit/8599d...5f72a2a7aa 14:33
phenny masak: I'll pass that on when tadzik is around.
14:35 takadonet joined, amkrankruleuen joined
amkrankruleuen Hello. 14:35
masak amkrankruleuen! \o/
moritz_ oh hai
14:40 cognominal left 14:41 rainerschuster left
dalek kudo: a93dcb6 | moritz++ | / (2 files):
[docs] update ChangeLog
 14:41
14:42 cognominal joined
moritz_ rakudo: / \# / 14:45
p6eval rakudo ce565f: ( no output )
moritz_ std: / \# /
p6eval std 32123: OUTPUT«===SORRY!===␤Regex not terminated at /tmp/FMr5XK2dYm line 1 (EOF):␤------> / \# /⏏<EOL>␤ expecting quantifier␤Parse failed␤FAILED 00:01 116m␤»
masak huh. 14:46
moritz_ TimToady: LTA parse error message above
masak LTA parse, I'd say.
moritz_ TimToady: should trigger the "No unspace allowed" from line 4725
std: / /
p6eval std 32123: OUTPUT«===SORRY!===␤Null regex not allowed at /tmp/Axk3n03C7Z line 1:␤------> / /⏏<EOL>␤Check failed␤FAILED 00:01 114m␤»
moritz_ rakudo: / /
p6eval rakudo ce565f: OUTPUT«===SORRY!===␤Null regex not allowed at line 22, near ""␤» 14:47
moritz_ rakudo: / & /
p6eval rakudo ce565f: ( no output )
moritz_ bug
rakudo: / | /
p6eval rakudo ce565f: ( no output )
moritz_ same bug
shortcircuit TimToady: I was up later than usual last night. :) 14:48
moritz_ rakudo: / : /
p6eval rakudo ce565f: OUTPUT«===SORRY!===␤Method 'backtrack' not found for invocant of class 'Integer'␤»
moritz_ LTA error 14:49
14:49 ash_ joined
jnthn std: / : / 14:49
p6eval std 32123: OUTPUT«ok 00:01 116m␤»
moritz_ that's broken too
jnthn LTE error 14:50
moritz_ should say "Backtrack control ':' does not seem to have a preceding atom to control"
14:51 Getty left
isBEKaml rakudo: / #`{ / # just curious 14:53
p6eval rakudo ce565f: OUTPUT«(timeout)»
isBEKaml std: / #`{ /
p6eval std 32123: OUTPUT«===SORRY!===␤Couldn't find terminator } at /tmp/n1sjPtydK7 line 1 (EOF):␤------> / #`{ /⏏<EOL>␤Parse failed␤FAILED 00:01 114m␤» 14:54
shortcircuit TimToady: Wow; didn't expect that you'd do that much analysis on each example. Thanks. :) 14:59
shortcircuit also went in added a test output spec consisting of a couple functions, so something like that doesn't happen again on that page. 15:00
15:00 takadonet left
x3nU well 15:00
i have created rakudo star installer for windows
shortcircuit whups. The 'later than usual' was intended for a different channel/person target. 15:01
15:03 macroron joined, amkrankruleuen left, patspam joined 15:08 whiteknight left
moritz_ std: /a{1,3}/ 15:09
p6eval std 32123: OUTPUT«===SORRY!===␤Unsupported use of {1,3} as general quantifier; in Perl 6 please use X**1..3 at /tmp/VUiqpioGUF line 1:␤------> /a{1,3}⏏/␤Check failed␤FAILED 00:01 116m␤»
x3nU xenu.pl/rakudo-star.2010.08.exe
moritz_ rakudo: /a{1,3}/
p6eval rakudo fb4feb: ( no output )
moritz_ x3nU++
masak moritz_: known, RT'd? 15:11
15:12 The_Ghost joined
moritz_ masak: I'm fixing the a{1,3} thing now (as good as currently possiblein nqq-rx), I don't think the others are RT'ed 15:12
masak backlogs and submits what he can sift out 15:13
15:13 The_Ghost left
moritz_ I think a general "LTA regex parsing errors" ticket wouldn't hurt 15:13
masak std: / & / 15:15
p6eval std 32123: OUTPUT«===SORRY!===␤Null pattern not allowed at /tmp/jD4FwUztCd line 1:␤------> / & ⏏/␤ expecting quantifier␤Parse failed␤FAILED 00:02 115m␤»
15:18 b0m1ro joined 15:19 tylercurtis joined 15:21 isBEKaml left
dalek p-rx: 184e828 | moritz++ | src/Regex/P6Regex/Grammar.pm:
catch p5 gneral quantifiers
 15:21
shortcircuit TimToady: Question regarding your P6 code: rosettacode.org/mw/index.php?title=...oldid=prev 15:28
15:30 takadonet joined 15:33 Chillance left 15:37 b0m1ro left 15:38 ruoso joined 15:40 Ross left, Ross joined 15:43 patspam left 15:48 mikehh left 15:51 cjk101010 left 16:05 jettero left, thundergnat joined 16:06 jettero joined, jettero left, jettero joined
thundergnat rakudo: my $x; $x = $x ?& 1; say $x; 16:08
p6eval rakudo fb4feb: OUTPUT«0␤»
thundergnat rakudo: my $x; $x ?&= 1; say $x;
p6eval rakudo fb4feb: OUTPUT«No applicable candidates found to dispatch to for 'infix:<?&>'. Available candidates are:␤:(Any $a, Any $b)␤␤ in main program body at line 1␤» 16:09
thundergnat hmmm.
known bug?
or does that fall under "don't do that"?
tylercurtis rakudo: my $x; $x [?&]= 1; say $x;
p6eval rakudo fb4feb: OUTPUT«===SORRY!===␤Confused at line 22, near "$x [?&]= 1"␤»
16:09 Steppenwolf left
tylercurtis std: my $x; $x [?&]= 1; say $x; 16:10
p6eval std 32123: OUTPUT«ok 00:01 118m␤»
tylercurtis The former is probably a bug. The latter almost certainly is.
thundergnat Its kind of a degenerate case
pmichaud good morning, #perl6 16:11
jnthn morning, pmichaud 16:12
masak morning, pm! 16:14
I'm getting a build fail on latest Parrot and latest Rakudo: "FATAL ERROR: Duplicated VTABLE function: get_pmc_keyed_int at /usr/local/lib/parrot/2.7.0-devel/tools/build/../lib/Parrot/Pmc2c/PMC.pm line 74."
should I downgrade my Parrot?
PARROT_REVISION says 48909, I has 48934. 16:15
(just to be clear, latest Parrot builds fine; latest Rakudo on top of it doesn't.)
pmichaud we might have a PMC that has multiple get_pmc_keyed_int declarations. 16:17
one of the recent Parrot changes declares that an error.
x3nU also i wonder can someone put my installer on rakudo.org? ;f
pmichaud in fact, I'm wondering if that should've required a deprecation cycle 16:18
tylercurtis Compilers detecting bugs in supplied code is a good thing, no? 16:19
16:20 mfollett joined
tylercurtis By the way, github.com/rakudo/rakudo/blob/maste...b.pmc#L926 and github.com/rakudo/rakudo/blob/maste....pmc#L1112 16:20
pmichaud tylercurtis: sure, but making code that previously worked turn into non-working code isn't.
16:23 rhebus joined 16:25 takadonet left
tylercurtis pmichaud: A warning might be better than a fatal error (at least for a deprecation cycle). But I'd consider pmc2c silently ignoring parts of the PMC code a bug. 16:25
pmichaud sure, I consider that a bug also. 16:26
but part of the purpose of Parrot declaring "supported releases" and a deprecation cycle is so that HLL and library authors supposedly won't have to deal with these sorts of failures.
at least, that's what I understand the purpose to be.
16:27 takadonet joined, tadzik joined
tylercurtis True. 16:27
tadzik oh hai
phenny tadzik: 14:33Z <masak> tell tadzik fixed, thanks! github.com/masak/proto/commit/8599d...5f72a2a7aa
masak o/
pmichaud a warning would be great. failing altogether isn't.
tadzik masak: also, Module::Tools are mostly working
masak \o/ 16:28
tadzik except when they don't
16:30 jferrero joined
rhebus rakudo: say %(1,2).WHAT 16:31
p6eval rakudo fb4feb: OUTPUT«Hash()␤» 16:32
rhebus star: say %(1,2).WHAT
p6eval star 2010.07: ( no output )
16:32 mfollett left
tadzik github.com/tadzik/perl6-File-Tools/...le/Copy.pm can you see anything painfully slow in here? 16:33
it takes ages to copy anything
16:34 sjn left
pmichaud tadzik: there are some pathological cases in I/O in general (both in Parrot and in Rakudo) 16:34
I agree that should not be slow. It's probably worth tracking down a bit to see why it's slow.
16:35 mfollett joined
tadzik sometimes it's immediate, sometimes takes around 25 seconds to copy a single file 16:37
decreasing over time
pmichaud I'm thinking it may be GC again.
but I can run a test to find out.
16:38 Axius joined
tadzik oh, maybe it's old .pir lying around confusing Parrot and Rakudo, let's see 16:38
ash_ tadzik: you could cheat with your cp function and call cp (if its a unixy system)
tadzik well, no
ash_: that's what I want to avoid
moreover, I'm worried about my child being so painfully slow 16:39
while being, well, quite straightforward
ash_ ah, well, i'd do some profiling then
tadzik gist.github.com/575338 that's the code 16:40
16:40 kjeldahl_ joined
tadzik gist.github.com/575340 16:40
here's the result ↑ the copying time is approaching to 0 seconds, yet it's horrible at the beginning
ash_: how do I do profiling on Rakudo? 16:41
16:43 kjeldahl left
tadzik alright, it's not the matter of "the first file", it's about the size I think 16:44
...but _that_ slow?
these are just 13 iterations 16:45
16:45 mfollett left
ash_ in theory, you could do parrot -R profiling but that currently doesn't work in rakudo, so.... i am not sure what the best way of detecting the slow part is 16:49
tadzik rakudo: my $t = 5555555; say "{time - $t} seconds passed" # what?
p6eval rakudo fb4feb: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in 'time' at line 5933:CORE.setting␤ in main program body at line 22:/tmp/pbmFAlpOzT␤»
ash_ rakudo: say time; 16:53
p6eval rakudo fb4feb: OUTPUT«1284222665␤»
ash_ rakudo: my $t = time; say $t - 1;
p6eval rakudo fb4feb: OUTPUT«1284222679␤»
ash_ rakudo: say "{time}"
p6eval rakudo fb4feb: OUTPUT«1284222687␤»
ash_ rakudo: say "{time - 1}"
p6eval rakudo fb4feb: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in 'time' at line 5933:CORE.setting␤ in main program body at line 22:/tmp/7ZHhrTveTq␤»
ash_ thats weird
must be a bug with interpolating functions 16:54
tadzik I sense so 16:55
ash_ rakudo: sub f { 'foo' }; sub g($a) { 'g' ~ $a }; say "{f}"; say "{g 1}" 16:56
p6eval rakudo a93dcb: OUTPUT«foo␤g1␤»
ash_ rakudo: sub f { 'foo' }; sub g($a) { 'g' ~ $a }; say "{f ~ 1}"; say "{g 1}"
p6eval rakudo a93dcb: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in 'f' at line 22:/tmp/jxxAhP2YWx␤ in main program body at line 22:/tmp/jxxAhP2YWx␤» 16:57
ash_ weird
i don't get it but /shrug
ash_ submits rakudo bug
rhebus ash_: is it interpreting "f ~ 1" as "f (~1)"? 16:59
16:59 timbunce joined
rhebus and is that correct with no parameter list? 16:59
ash_ std: sub f { ... }; say "{f + 1}"
p6eval std 32123: OUTPUT«ok 00:01 116m␤»
tadzik oh, std
rhebus kk
ash_ rhebus: that seems possible 17:00
tadzik std: my $t = 5555555; say "{time - $t} seconds passed"
p6eval std 32123: OUTPUT«ok 00:01 118m␤»
tadzik lol bug
ash_ that could be how its interpreting it
tadzik rakudo: my $t = 5555555; say "{time() - $t} seconds passed"
p6eval rakudo a93dcb: OUTPUT«1278667546 seconds passed␤»
ash_ yup thats it
rhebus rakudo: sub f () { 'foo' }; sub g($a) { 'g' ~ $a }; say "{f ~ 1}"; say "{g 1}" 17:01
p6eval rakudo a93dcb: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in 'f' at line 22:/tmp/FSQLthDu_t␤ in main program body at line 22:/tmp/FSQLthDu_t␤»
17:01 Axius left
rhebus rakudo: sub f () { 'foo' }; sub g($a) { 'g' ~ $a }; say "{f() ~ 1}"; say "{g 1}" 17:01
p6eval rakudo a93dcb: OUTPUT«foo1␤g1␤»
tadzik wklej.org/id/387712/ -- yep, it's about the size 17:02
17:02 Chillance joined 17:04 Axius joined
ash_ that probably has something to do with parrot's GC system 17:04
tadzik weird 17:05
ash_ parrots GC is a bit weird at times 17:06
sorear good * #perl6
tadzik hello 17:07
17:07 thundergnat left 17:10 patch left 17:24 dju left
jnthn afk for noms 17:25
pmichaud tadzik: it looks to me like it's taking about 1 sec per 4k block 17:28
let's see if I can get something similar here
tadzik I just thought about trying to increase the amount per iteration
pmichaud if it's gc-related, I think that will make it worse, no better. Ironically.
*not
tadzik I'm trying 2048 now... 17:29
no significant difference 17:30
alike with 8192. Bah 17:32
pmichaud rebuilding latest parrot here
try 512, ooc
tadzik building with 128, just ooc
pmichaud wfm
tadzik I like edge cases for comparing
pmichaud try 1. :-P
try 0. :-)
tadzik :D
pmichaud at 128 I think we might get swamped by the method call overhead, though :-| 17:33
tadzik the lack of any difference is making me wonder whether it actually makes the changes in pirs
pmichaud well, I'm writing a baseline PIR version now so we can do a base measurement. 17:34
tadzik great
pmichaud oh, ouch ouch ouch 17:35
github.com/rakudo/rakudo/blob/maste.../IO.pm#L72 # here's why it's slow 17:36
that's.... not going to work in the long run.
tadzik where exactly?
pmichaud or even in the medium run.
lines 72-97
tadzik well, that I can see :) 17:37
dalek kudo: 085920f | pmichaud++ | src/pmc/perl6multisub.pmc:
Remove extra (unused) VTABLE get_pmc_keyed_int from perl6multisub. Discovered by NotFound++ (Parrot r48932) and masak++.
pmichaud after reading a 4k chunk, it pushes them one-byte-at-a-time into a ByteBuffer
that's going to be... slow.
tadzik ouch
(maybe inserting 1 instead of 128 wouldn't be a bad idea after all :P) 17:38
pmichaud oh, it's worse!
17:38 Axius left
pmichaud it pushes them one-byte-at-a-time into an RPA! 17:38
17:38 hudnix left
lue ohai o/ 17:38
masak pmichaud: o.O
17:39 hudnix joined
tadzik RPA? 17:39
pmichaud ResizablePMCArray
which means it creates a new PMC for each element pushed
tadzik aw snap
lue O.o
pmichaud i.e., one PMC per byte read. 17:40
tadzik btw, what's that tool extracting POD from C source? 17:41
masak sobs softly
17:41 macroron left
pmichaud masak: this code was/is working around Parrot's difficulty in reading binary strings, or ... ? 17:42
masak oh! we're talking about *my* code!? :/
pmichaud oh, is it not yours?
I didn't look to see the authorship. 17:43
masak it's Buf-related?
pmichaud I just assume buf-related item..... yes
masak probably mine, then.
pmichaud at least, I'm guessing it is, since read() is returning a buf
masak checks
pmichaud let me be more precise.
github.com/rakudo/rakudo/commit/692...99cc5fdc53 # looks like it
masak aye. 17:44
pmichaud maybe that was that other "masak++" :)
masak ok, so *all* conversion to and from Buf is done byte-by-byte right now.
it's not ideal by far. 17:45
17:45 leprevost left
rhebus grabs a pitchfork 17:45
pmichaud agreed, and no blame for it afaic
17:45 leprevost joined
pmichaud buf in parrot is... tricky 17:45
masak it's -Oworks rather than -Ospeed :)
pmichaud it's following in the footsteps of GGE :)
rhebus optimising is easy if it doesn't have to work :)
masak *lol*
pmichaud it's the GIO subsystem :) 17:46
lue
.oO(* in parrot is tricky)
pmichaud actually, * wasn't too bad :)
masak not sure there are tests for read(). checking.
pmichaud I think I'd like to see Buf implemented using a ResizableIntegerArray 17:47
lue .oO[ * wasn't meant literally :) ]
masak pmichaud: I can look into that. 17:50
rhebus i'm implementing a graph class, which has a method add-edges(), which I want to accept a list of Pair. A Pair 'a' => 'b' represents a line from vertex 'a' to vertex 'b'
masak rhebus: go on. :)
rhebus however, a Pair with a certain key gets interpreted as a named parameter. eg: sub add-edges (*@_ of Pair) { ... } foo(1 => 2) is ok, but foo (a => 'b') breaks with "Unexpected named parameter 'a' passed 17:51
masak rhebus: yes.
rhebus i don't want to interpret lines as named parameters, because that's not what they are
masak rhebus: to pass a pair, you need either 'a' => 'b' or (a => 'b') 17:52
rhebus er, any suggestions on which part of my design needs to give?
masak: i see
masak rhebus: I've also experienced friction with that part of Perl 6 parameter passing.
rhebus :/
masak rhebus: what you can do (which I don't necessarily recommend) is add a slurpy hash to the signature.
rhebus: that would slurp up all the occasional named params, and you can add them post facto to the array. 17:53
rhebus masak: that fails for foo (a => 'b', a => 'c')
masak not per the spec.
maybe per Rakudo.
rhebus really?
masak checks
rhebus what does it mean then? what does %^a<a> give me in that case? 17:54
masak S06:745
rhebus S06:753 confirms lossage 17:58
rakudo: sub foo(*%p) { say %p.perl }; foo (a => 'b', a => 'c') 17:59
p6eval rakudo a93dcb: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in 'foo' at line 22:/tmp/crSDJW6xjj␤ in main program body at line 22:/tmp/crSDJW6xjj␤»
masak hm.
rhebus i'll go with the literal Pair syntax
masak nod 18:00
rakudo: sub foo(*%p) { say %p.perl }; foo(a => 'b', a => 'c')
p6eval rakudo a93dcb: OUTPUT«duplicate named argument in call␤ in main program body at line 2:/tmp/ebJMPKp82_␤»
masak submits rakudobug
rakudo: sub foo(*%p) { say %p.perl }; my %h = a => 'b'; foo(|%h, a => 'c') 18:01
p6eval rakudo a93dcb: OUTPUT«duplicate named argument in call␤ in main program body at line 22:/tmp/N6sXSOf5hR␤»
rhebus i literally can't see the difference between mine and yours there
pmichaud the potential problem with slurping things into a hash is that ordering gets lost
masak rhebus: so here's the thing about function calls and whitespace.
rhebus in my case, ordering doesn't really matter 18:02
masak rhebus: 'foo(' and 'foo (' mean different things
rhebus masak: :O
Ö
pmichaud the other way to avoid named parameters is to use an extra level of parens around the entire list
masak rhebus: general theme through all of Perl 6.
pmichaud foo(a=>1, b=>2) # named arguments
masak rhebus: they snuck a little whitespace significance in there... :)
pmichaud foo((a=>1, b=>2)) # list of pairs
masak right and 'foo (' means the latter. 18:03
18:03 envi^home left
rhebus hmm, my list of things i don't quite like about perl 6 is growing :( 18:04
it's got past the honeymoon period
masak :)
rhebus: I might be able to argue that that particular significant whitespace thing is part of the whole two-terms-in-a-row reasoning. 18:05
pmichaud I've not been happy about the a=>1 named/pair dichotomy either, but so far the cures aren't much better than the disease.
masak and two-terms-in-a-row is the whole *key* to understanding Perl 6 parsing.
rhebus two-terms-in-a-row?
masak pmichaud: agreed.
rhebus: ok, so here's the thing about terms and operators.
rhebus: to a first approximation, code looks like this: term op term op term
rhebus with you so far 18:06
masak that's with infix operators; prefixes and postfixes work a little different. but not much.
sorear TTIAR has been in Perl since 1.0
masak if you find, term term or op op, it means something special is happening.
sorear I'm not sure what significant whitespace has to do with it
masak for example that something is wrong.
sorear: I'm building up to it. 18:07
at some point in the evolution of Perl 6, it was decided that postfix operators really can't have any whitespace before them.
because then they're suddenly ambiguous with infix operators. 18:08
so nowadays, the '+' in 'foo+' is a postifx operator if such an operator is defined in the current scope.
only otherwise it's an infix operator.
18:08 patspam joined
pmichaud or prefix operator 18:09
since prefix also occurs in "op" position.
oh, wait
masak no.
pmichaud in term position
never mind.
masak please ignore pmichaud :)
pmichaud no, don't ignore me entirely
because foo + bar is not an infix
masak it's not?
pmichaud it's foo( + bar() ) 18:10
masak oh, sorry. I didn't mean 'foo' as a listop.
I shoulda said $foo
pmichaud right
masak thanks.
rhebus this is really helping
masak anyway...
pmichaud listops are also a place where we run into ambiguity
masak postcircumfixes ( () [] <> etc ) are a kind of postfix.
well, not really, but it helps if they follow similar rules. 18:11
pmichaud i.e., listops tends ot be "term term"
*tend to
masak listops basically reset the term/op clocking function.
anyway...
pmichaud and what turns something into a listop is a space.
masak now you see 'foo(' is really just a postcircumfix function call.
sorear what matters isn't "term term", it's that the parser almost always knows whether the next token is going to be an infex or a term
in Perl 1-5, this was a literal boolean flag in the lexer 18:12
masak and 'foo (' is a listop followed by a term which happens to be a parenthetical group.
pmichaud and it's the *space* that causes the listop
masak well 'foo;' is a listop too
pmichaud okay, space or terminator :) 18:13
rhebus oh! so, for example, say (1) + 2 is actually say 1 + 2 == say(1 + 2)
masak aye.
rhebus++ # gets it
rhebus: thus we avoid the awkward misparsings of Perl 5.
pmichaud we just get new awkward misparsings :)
masak where say (1) + 2 puts '+ 2' in void context.
pmichaud: but honestly, I find Perl 6's way of parsing function calls so much nicer. 18:14
pmichaud I do too.
very much so.
rhebus masak: what? surely say(1) + 2?
pmichaud I'm just saying it's not completely resolved in Perl 6 :-)
masak rhebus: no, not in Perl 5. 18:15
pmichaud because "foo + 3" is a listop (assuming foo is a sub), but "pi + 3" is an equation.
masak rhebus: in Perl 5, the space doesn't matter.
rhebus masak: ah, hadn't got perl 5 context
masak sorry, might not have been clear on that.
rhebus you need an operator for that
masak++ # slow and patient explanation 18:16
pmichaud rakudo: say pi + 3; say abs + 3;
masak pmichaud: a lint tool might help there, too.
p6eval rakudo a93dcb: OUTPUT«6.14159265358979␤3␤»
masak rhebus: I was curious whether I would be able to connect the dots :)
flussence I have to admit, whitespace after a sub is easier to remember than "put unary plus here, depending on whether the author wrote sub($) or not"
rhebus does one define a nullary function in the same way as in perl 5? ie sub foo () { }
pmichaud rhebus: sub term:<foo>() { ... } # iirc 18:17
rhebus pmichaud: thankie
pmichaud for simple constants, it's just constant xyz = ...
masak right. because subs are always listops, even if they take no parameters.
pmichaud for example, 'self' is defined as a term. 18:19
rhebus right
tadzik masak: github.com/tadzik/perl6-Module-Tools 18:21
seems to be working for me
masak tadzik: great! kudos.
rhebus right 18:22
off to last night of the proms <333
18:23 leprevost left
masak have the appropriate amount of fun! 18:23
tadzik neutro seems to handle this 18:24
I'd say have even more :)
masak .oO( inappropriate amounts of fun? )
tadzik more than appropriate
oh, btw masak 18:26
pmichaud afk, lunch
tadzik method test($project --> Result) -- what's that?
masak a return type...?
tadzik I don't know, that's why I'm asking :) 18:27
pmichaud iir, it's a return constraint
masak right.
basically, the method promises to return something that is or does Result, if it returns normally.
I actually don't know how that interacts with &fail 18:28
tadzik hmm. my $command = "env PERL6LIB=`pwd`/lib prove -e 'perl6' -r t/";
can't it be just "env PERL6LIB=lib prove ..." ?
18:28 synth left
pmichaud undefined types always match, likely. 18:29
(and Failure is undefined)
really afk this time.
masak tadzik: it probably can. never thought of that. 18:30
tadzik well, `pwd` is just .
and `pwd`/bah is just ./bah, which is just bah
at least I think so
well, let's see if it works in Module::Test 18:31
18:31 rhebus left
tadzik btw, why 'env' at the beginning 18:32
?
18:35 leprevost joined
masak becuase people have told me to write it like that :) 18:37
tadzik :)
tadzik provoking people
aaand, we have Module::Test 18:38
masak \o/ 18:39
sorear I think it has more to do with tcsh
masak 'has' scoping is koan-level curious. 18:41
tadzik star: eval 'use File::Find'; say $!.perl; eval 'use Not::Found'; say $!.perl
p6eval star 2010.07: OUTPUT«Any␤undef␤»
tadzik hmm
masak it basically says 'put the variable/method/whatever on *instances*, only not really'
18:42 Schwern joined, rgrau_ left
masak for different levels of 'not really' :) 18:42
i.e. each instance has its own private attribute to a greater extent that it has its own method. 18:43
serious dwimmery at work.
18:50 Axius joined 18:51 wamba joined 18:52 M_o_C joined 18:55 Mowah_ joined, Mowah_ left 18:59 Axius left, Axius joined 19:02 Mowah left 19:03 Axius left
tadzik how do I recursively flatten a list? 19:08
if I can
masak rakudo: my @a = 1, 2, (3, 4, (5, 6, (7, 8))); say @a.perl
p6eval rakudo 085920: OUTPUT«[1, 2, 3, 4, 5, 6, 7, 8]␤»
tadzik well, an array then 19:09
masak should have said array, then :)
I don't know. I would write a sub, I think.
19:12 wamba left 19:17 takadonet left, Axius joined, molaf joined 19:18 takadonet joined
tadzik how do I modify permissions from Perl 6 (do I?) 19:20
like chmod
masak S32/IO specs chmod. 19:21
rakudo: chmod
p6eval rakudo 085920: OUTPUT«Could not find sub &chmod␤ in main program body at line 22:/tmp/29WkCvWp00␤»
19:21 mberends left 19:22 timbunce left
\xF0 last I checked, lots of system functions don't exist yet 19:22
masak that is so.
19:23 takadonet left
tadzik github.com/tadzik/perl6-Module-Tool...98263ba4cc 19:23
any ideas why doesn't it always work as expected? Seems to sometimes compile some files anyway 19:24
19:24 takadonet joined
tadzik . o O ( maybe I'll get myself some civilized gravatar ) 19:25
19:27 M_o_C left 19:28 timbunce joined
moritz_ jnthn++ # blog post 19:30
tadzik oh, moritz_, was just thinking about you :)
so I'm supposed to write some simple grammar example for the book>?
based on the Q&A from advent calendar? 19:31
moritz_ that would be very welcome
tadzik will do
I'll need to get myself an inkscape then :)
moritz_ re that commit, you don't even need the { } for interpolation 19:32
(but that's the line you deleted)
tadzik it used to by copypasted from somebody else's code I think 19:33
moritz_ anyway, in what way doesn't it work as expected, sometimes? 19:34
tadzik well, sometimes the difference in stat is obvious, but it still build stuff anyway
(will paste some exapmle)
ash_ rakudo: sub f { 5 }; say "{f - 5}" # i found a bug, i haven't seen any tickets in the bug tracker like it, i don't think 19:35
p6eval rakudo 085920: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in 'f' at line 22:/tmp/GnQ4gv4xbg␤ in main program body at line 22:/tmp/GnQ4gv4xbg␤»
ash_ actually tadzik++ found it
tadzik wklej.org/id/387765/
moritz_ not a bug
it's parsed as f(-5) 19:36
ash_ rakudo: sub f { 5 }; say "{f() - 5}"
ah
tadzik oh
p6eval rakudo 085920: OUTPUT«0␤»
moritz_ but f doesn't expect a parameter
ash_ well, if thats how its supposed to work, then thats fine
tadzik If we had static dispatch... :P
19:37 rainerschuster joined
moritz_ Perl 6 - still not able to read the programmer's mind 19:38
ash_ Perl 6 - alpha version of skynet 19:39
tadzik Perl 6 – prototype of Perl 7 19:40
masak Perl 6 - reasonable if you are 19:41
19:41 meppel joined
masak ==> $bed 19:41
tadzik Perl 6 – fun as Haskell, but better
19:41 masak left 19:43 Axius left 19:44 meppl left 19:45 Axius joined, icwiener_ joined 19:46 meppel is now known as meppl 19:47 icwiener left, icwiener_ left, icwiener_ joined
moritz_ phenny: ask masak if he already applied some version of his enum patch, and if not, if I can do anything to speed it up or to help 19:49
phenny moritz_: I'll pass that on when masak is around.
19:50 Axius left 19:51 Axius joined 19:53 pjcj left
ash_ rakudo: say "{1 + 3 + 4}" 19:55
p6eval rakudo 085920: OUTPUT«8␤»
19:55 Axius left 19:56 Axius joined
tylercurtis rakudo: 1..5 ==> map * * 2 ==> my $foo; say $foo.perl; # curious 19:56
p6eval rakudo 085920: OUTPUT«Method 'push' not found for invocant of class ''␤ in main program body at line 22:/tmp/0HiSxB59nl␤»
moritz_ rakudo: 1..5 ==> map * * 2 ==> my @foo; say @foo.perl 19:57
p6eval rakudo 085920: OUTPUT«[2, 4, 6, 8, 10]␤»
moritz_ tylercurtis: push doesn't autoviv yet
though it should
tylercurtis moritz_: good to know.
moritz_ rakudo: my %h; %h<a>.push: 3 19:58
same problem
p6eval rakudo 085920: OUTPUT«Method 'push' not found for invocant of class ''␤ in main program body at line 22:/tmp/Svptg5GHeg␤»
dalek kudo: 8669d77 | Util++ | build/PARROT_REVISION:
Bump PARROT_REVISION to include nwellnhof++ fix for non-ICU Parrots. Fixes RT#77778 part 2.
20:00 qw3rty joined
jnthn "S06 explicitly allows passing the same named parameter several times." # Is that a hangover from the insanity over :@foo collecting up all the foo arguments? 20:02
jnthn votes to toss 20:03
Either that or explicit "last one of the name wins"
20:06 Schwern left 20:09 snearch joined
tylercurtis lazy ping: sorear; 20:11
flussence S06 seems to say that already 20:16
at least for not-@ 20:17
sorear lazy? 20:19
20:21 molaf left, araujo left
moritz_ rakudo: sub f { say callframe(1).line }; f 20:24
p6eval rakudo 085920: OUTPUT«22␤»
sorear p6eval lies
moritz_ it just counts differently than you do 20:25
20:25 qw3rty left 20:26 timbunce left, rainerschuster left 20:27 tadzik left, tadzik1 joined, tadzik1 is now known as tadzik, ash_ left
sorear tylercurtis: ??? 20:30
20:30 Axius left 20:31 Ross left, Ross joined 20:32 timbunce joined 20:34 icwiener_ left 20:35 takadonet left 20:36 tadzik left, qw3rty joined, takadonet joined
qw3rty hi... 20:36
"roles" support, is part of perl6 or parrotvm?
sorear yes 20:37
moritz_ parrot has roles, but they are not enough for what Perl 6 needs 20:38
qw3rty so, one can make a COBOL that runs on parrot and add "roles" to that COBOL?
moritz_ so rakudo doesn't use parrot roles to implement the Perl 6 roles
yes
but it wouldn't be COBOL anymore
qw3rty more like pabol or something like that
but still sounds like a good experiment to do
moritz_ doesn't know how 'COBOL' and 'good' go together in the same sentence 20:39
qw3rty hehe,, is good on the paycheck ;)
some of my coleagues work in cobol
they salaries are good, compared to, say, java developers
sorear supply and demand ;) 20:40
moritz_ compensation for suffering, I guess
20:41 synth joined
qw3rty i think a saw a test suite for cobol somewhere on the net... i have a proyect for these holidays now :) 20:41
anyway, thanks for your answers guys 20:42
great job u are doing with perl6
even if not still out (i mean, production) i already had many interesting discussions online and in "real" life about some features
well.. bye all! 20:43
20:43 qw3rty left 20:44 araujo joined 20:52 patspam left 20:56 araujo left
tylercurtis sorear: sorry about the ping and disappear. I cargo-culted the CursorBase::_REDUCE and REDUCE methods from Niecza::Actions for Bennu, and I've noticed that only a certain number(around a dozen) of the "Action method ... not yet implemented" sorrys actually display. Should that be happening? Can I make it display all of the sorrys? 20:56
sorear tylercurtis: take a look at STD.pm6 line 6028 20:57
also the corresponding line in STD.pmc 20:58
21:00 rainerschuster joined 21:01 araujo joined, tom_tsuruhara joined, araujo left
tylercurtis sorear: thanks. 21:02
21:02 jaldhar left
moritz_ rakudo: say { 1 Z=> True xx * }.perl 21:09
p6eval rakudo 8669d7: OUTPUT«{ ... }␤»
moritz_ rakudo: class A::B { has $.x = 3 }; say A::B.new(x => 1).perl 21:13
p6eval rakudo 8669d7: OUTPUT«A;B.new(x => 1)␤»
moritz_ submits rakudobug
colomon A;B is the bug? 21:18
moritz_ yes
21:18 jaldhar joined, patrickas joined
colomon moritz_++ 21:19
patrickas rakudo: say 1..* ... 5;
p6eval rakudo 8669d7: OUTPUT«12345␤»
patrickas moritz_++ indeed!
colomon patrickas++
that's what I should be doing while I watch the football game -- adding series tests.
21:20 jaldhar left
patrickas moritz_ I am reading the comments on github, indeex that tests should have been removed, I had forgotten to push it! 21:20
indeex! That's a new one! 21:21
moritz_ kinda loves the Xxx operator 21:22
patrickas colomon: I am sure you'll find some new edge case I hadn't thought about and manage to send me back to the drawing board :-)
moritz_ rakudo: say 1..3 Xxx 3
p6eval rakudo 8669d7: OUTPUT«111222333␤»
patrickas Oh yea, that's the nsfw operator!
colomon rakudo: say 1..3 Xxx 1..3 21:23
p6eval rakudo 8669d7: OUTPUT«111111222222333333␤»
colomon rakudo: say (1..3 Xxx 1..3).perl 21:24
p6eval rakudo 8669d7: OUTPUT«((1), (1, 1), (1, 1, 1), (2), (2, 2), (2, 2, 2), (3), (3, 3), (3, 3, 3))␤»
21:24 pjcj joined
moritz_ that usage is interesting too, bu I haven't felt the need in real world code yet 21:25
colomon patrickas: given the scope of series, it's pretty inevitable. :)
moritz_ nopaste.snit.ch/23302 :-) 21:26
21:28 araujo joined 21:29 jaldhar joined 21:30 jaldhar left, jaldhar joined, snearch left, guidjos joined 21:32 guidjos left, guidjos joined 21:34 guidjos left 21:35 Guest23195 left
colomon patrickas: have you checked in all your changes to the series tests? (/me moving very slowly) 21:35
Util_ To handle spectests like S19-command-line/dash-e.t, which contains a mix of Unicode and non-Unicode tests, 21:38
I plan to split the Unicode tests out into a separate .t (dash-e-unicode.t), then condition it with '# icu' in spectest.data.
Compared to just marking the original file as '# icu', this will allow more testing to run on non-ICU systems.
Objections? Have I missed anything?
21:41 Mowah joined
moritz_ Util_: it might be better to have one command-line-unicode.t file somewhere 21:41
Util_: instead of having two files for each option
Util_ moritz_: I see. I would add comments to each block of tests in command-line-unicode.t, pointing to the .t that would normally be their home. 21:43
moritz_ Util_: +^
+1
moritz_ should really sleep
sorear why can't we just depend on ICU? 21:45
We already depend on Parrot, which is much bigger and doubtless harder to compile
pmichaud our use of ICU depends on Parrot's use of ICU. 21:46
and at one time (although this may no longer be true), getting ICU to work with Parrot was a Real Pain on windows environments.
jnthn ICU used to be bundled in with Parrot and built with it.
It was a lot of effort to get that to fly on Windows.
pmichaud ICU is still built with Parrot.
jnthn I got it to work.
pmichaud: Used to be in the repo.
pmichaud (although it's not bundled with it) 21:47
Util_ sorear: I expect that on Christmas, we might start requiring ICU. Until then, no-ICU required means more testers come to our parties!
jnthn Anyway, yes, last time I tried to build a Parrot on Windows with ICU support, I didn't manage to get it to work.
patrickas colomon: sorry I was afk, yes I think moritz_++ pulled everythin I had
pmichaud anyway, requiring ICU significantly increases the steps needed from Windows users, and that goes against our larger goal of increasing Perl 6 usage. 21:48
patrickas rakudo: say 1 ... *>5
p6eval rakudo 8669d7: OUTPUT«123456␤»
pmichaud \o/
patrickas yes that's the alst case I worked on 21:49
jnthn Nice
21:49 Ross left
jnthn rakudo: say 1 ... *>* 21:49
p6eval rakudo 8669d7: ( no output )
jnthn :-)
sorear Util_: we can't just hide a copy of ICU in src/dynops? 21:50
*src/ops
patrickas jnthn what was that supposed to do ? 21:51
jnthn patrickas: Be infinite.
patrickas: Wanted to check it didn't blow up with arity 2 closures
(on the RHS)
patrickas jnthn: the RHS closure taking two items was working yesterday but after talking to TimToady and colomon I specifically made it only work for arity 1 21:52
pmichaud I don't understand why dash-e.t requires unicode, btw.
oh, nm.
I'm thinking of something *totally* different.
yes, it needs it because of the \c[...]
jnthn patrickas: Oh...so what did it do above?
Util_ pmichaud: RT#77778
patrickas no idea! kinda confused me! 21:53
jnthn rakudo: say 1 ... *>*; say "alive"
p6eval rakudo 8669d7: OUTPUT«alive␤»
Util_ sorear: Do you mean a pre-compiled library, or the ICU source?
patrickas rakudo: 1,2 .... {$^a+$^b > 5}
jnthn oh, it makes an empty list and fails silently.
p6eval rakudo 8669d7: OUTPUT«===SORRY!===␤Confused at line 22, near "1,2 .... {"␤»
patrickas rakudo: say 1,2 .... {; $^a+$^b > 5} 21:54
p6eval rakudo 8669d7: OUTPUT«===SORRY!===␤Confused at line 22, near "say 1,2 .."␤»
pmichaud why four dots?
jnthn four dots?
patrickas rakudo: say 1,2 ... {; $^a+$^b > 5}
p6eval rakudo 8669d7: ( no output )
pmichaud Util_: right, those are due to the lack of ICU -- I agree.
jnthn Hmm, if it's not meant to work, maybe silent failure is LTA.
pmichaud rakudo: say 2 ~~ { $^a + $^b > 5 } 21:55
p6eval rakudo 8669d7: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in <anon> at line 2:/tmp/N8P9Onzuge␤ in 'Block::ACCEPTS' at line 5799:CORE.setting␤ in main program body at line 22:/tmp/N8P9Onzuge␤»
pmichaud rakudo: say ({ $^a + $^b > 5 }).ACCEPTS(2) 21:56
p6eval rakudo 8669d7: OUTPUT«Not enough positional parameters passed; got 1 but expected 2␤ in <anon> at line 2:/tmp/xhFVrhlvK1␤ in 'Block::ACCEPTS' at line 5799:CORE.setting␤ in main program body at line 22:/tmp/xhFVrhlvK1␤»
pmichaud that's what I'd expect.
jnthn aye
21:56 Kodi joined
patrickas brb 21:57
Kodi So I'm taking a shot at implementing implementation types, at least for hashes. And by that I mean I'm trying to get 'my %h is MyClass' to work. We should allow 'my %h is Set', although Sets are immutable and have boring values, right? 21:58
patrickas rakudo: my $limit={;*>*}; say $limit.count 21:59
p6eval rakudo 8669d7: OUTPUT«1␤»
sorear Util_: The source. 22:00
patrickas rakudo: my $limit= ->$a,$b {$a+$b}; say $limit.count
p6eval rakudo 8669d7: OUTPUT«2␤»
patrickas rakudo: say 1,2 ... ->$a,$b {; $a+$b > 5}
p6eval rakudo 8669d7: ( no output )
patrickas Does this mean anythig? or was i just typing random stuff under influence ? 22:01
github.com/rakudo/rakudo/commit/927...013#L0R410
Util_ sorear: Parrot used to do that, keeping the whole ICU source in the repo. I think it is a bad practice for any project that we another project wants to include, but it *can* be done. 22:02
I certainly support Rakudo Star packagers providing ICU in their binary packages, though. 22:03
patrickas rakudo: fail('foo') 22:04
p6eval rakudo 8669d7: ( no output )
patrickas oh it's just not priting the failure message!
locally
> say 1,2 ... ->$a,$b {$a+$b > 5} 22:05
===SORRY!===\nLimit arity cannot be larger than 1
22:07 ash_ joined
dalek kudo: 9734b29 | chromatic++ | src/pmc/objectref_pmc.template:
[PMC] Made ObjectRef's mark() more accurate.
 22:07
22:09 dalek left
sorear Util_: I'm not convinced it's bad myself... a little amusing, but submodules take most of the pain out 22:09
22:09 dalek joined, ChanServ sets mode: +v dalek 22:12 rainerschuster left
Util_ me is afk for 3 hours; will commit the unicode split on my return. 22:14
s/me/\/me/ 22:15
22:17 araujo left
patrickas is fail the right way to handle such a case ? 22:19
colomon yes seems like a reasonable answer to me... 22:20
patrickas how do i get p6eval to display the fail emssage ? 22:25
rakudo: try { fail 'foo'; CATCH { say "$!" } }
p6eval rakudo 8669d7: ( no output )
22:28 M_o_C joined
lue rakudo: grammar A { token B { <alpha>+ } }; say "hello" ~~ A.B 22:30
p6eval rakudo 8669d7: OUTPUT«Type objects are abstract and have no attributes, but you tried to access &!regex␤ in 'A::B' at line 22:/tmp/hIhidxIVgI␤ in main program body at line 22:/tmp/hIhidxIVgI␤» 22:31
lue rakudo: grammar A { token B { <alpha>+ } }; say "hello" ~~ A::B
p6eval rakudo 8669d7: OUTPUT«Can not find sub A::B␤ in main program body at line 1␤»
lue
.oO(so close)
 22:32
patrickas jnthn: btw TimToady prefered to remove the arity > 1 on the limit (even thought it was working) because it is rarely used and there is another way to achieve it if you really want to 22:34
rakudo: say 1,1,-> $a,$b { last if $a+$b>100 ;$a+$b;} ... *
p6eval rakudo 8669d7: OUTPUT«1123581321345589␤» 22:35
patrickas I don't really like it but at least it's there
pmichaud 22:03 <Util_> I certainly support Rakudo Star packagers providing ICU in their binary packages, though. 22:36
+1, I would agree to that.
(especially on the windows side, if it can be done relatively painlessly)
lue Is there another way to get token B of grammar A besides /<A::B>/ ?
pmichaud I wonder if <&A::B> works 22:37
well, no, not exactly.
because it also has to change the type of the underlying cursor.
22:38 hudnix left
lue I think using A.B [like I did above] should work, by `translating' A.B to the regex (in this case /<alpha>+/). 22:39
rakudo: grammar A { token B { <alpha>+ } }; say "hello" ~~ /{A.B}/ 22:40
pmichaud lue: it doesn't quite work that way.
p6eval rakudo 8669d7: OUTPUT«Type objects are abstract and have no attributes, but you tried to access &!regex␤ in 'A::B' at line 22:/tmp/4rC3qkMnZ3␤ in <anon> at line 22:/tmp/4rC3qkMnZ3␤ in 'Cool::match' at line 2416:CORE.setting␤ in 'Regex::ACCEPTS' at line 5840:CORE.setting␤ in main program body
..at lin…
pmichaud what if token B uses another subrule defined in A?
lue aah.
tylercurtis Or even if it recurses. 22:42
pmichaud afk for a while 22:45
22:46 hatseflats left, hatseflats joined, mjk joined 22:47 mberends joined
lue rakudo: say "hello" ~~ a.B 22:52
p6eval rakudo 8669d7: OUTPUT«Could not find sub &a␤ in main program body at line 22:/tmp/PjeUhGmNUx␤»
22:55 xinming_ joined, timbunce left 22:56 araujo joined 22:57 kjeldahl_ left 22:59 xinming left
lue rakudo: regex A { '(' ~ ')' 'ab' }; say 'x(ab' !~~ /<A>/; 23:01
p6eval rakudo 9734b2: OUTPUT«Useless declaration of has-scoped regex in a module; add our or my to install it in the lexpad or namespace␤Method 'A' not found for invocant of class 'Cursor'␤ in <anon> at line 22:/tmp/INaopm8zX9␤ in 'Cool::match' at line 2416:CORE.setting␤ in 'Regex::ACCEPTS' at line
..5840:COR…
lue Has the spec changed since #62086 concerning 'has-scoped regexes'? 23:02
23:10 Kodi left 23:13 exodist_ is now known as Exodist 23:23 patrickas left 23:24 shelling_lab joined 23:26 M_o_C left 23:27 gottreu joined 23:31 saaki joined 23:33 shelling_lab left, shelling_lab joined 23:35 saaki left 23:36 saaki joined 23:37 araujo left, patspam joined 23:42 Mowah left 23:46 leprevost left 23:47 shelling_lab left 23:48 mjk left, hudnix joined 23:50 mfollett joined 23:51 mfollett left 23:53 shelling_lab joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!