»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
Mouq p6: grammar G { regex TOP { (<abc>)+ }; regex abc { abc } }; G.parse("abcabc"); .say for @0 00:00
camelia niecza v24-109-g48a8de3: OUTPUT«「abc」␤ abc => 「abc」␤␤「abc」␤ abc => 「abc」␤␤»
..rakudo-parrot 4d8734, rakudo-jvm 4d8734, rakudo-moar 4d8734: OUTPUT«「abc」␤ abc => 「abc」␤ 「abc」␤ abc => 「abc」␤␤»
Mouq fixes that too 00:01
lue Mouq++ 00:02
I'd much prefer if @/ and friends were distinct things, instead of equivalent to @($/), but I guess it's a solution. (Better than two unnecessary bytes per desire to iterate at least) 00:03
dalek kudo/nom: 456f3aa | Mouq++ | src/Perl6/Actions.nqp:
@<h>, @1 mean @$<h>, @$1 and %<h>, %1 mean %$<h>, %$1
00:04
Mouq Not my least cryptic commit message... 00:05
lue I'll be sure to betterify my syngen code once I feel like throwing another half-hour at rakudo compilation :P 00:06
Mouq someone should grep the ecosystem for "for $<" and "for $"\d... 00:10
timotimo++: could you please mention this change in p6weekly? 00:13
timotimo i'm glad you've given me a change to mention :P 00:14
everything else i remember off the top of my head is TimToady's recent contributions
so @<foo> and %<foo> didn't exist so far, yes?
at least in rakudo 00:15
Mouq timotimo: Technically this is mostly his fault ;p
timotimo: They existed, but were equivalent to $<foo>
timotimo ah
00:16 berekuk left
Mouq r: say Date.today() - 7 00:17
camelia rakudo-parrot 4d8734, rakudo-jvm 4d8734, rakudo-moar 4d8734: OUTPUT«2014-02-24␤»
Mouq moar-support landed in panda! 00:18
pretty sure that was this week
timotimo cool
00:23 kaare_ joined
Mouq lue: github.com/lue/pod_vvi/blob/master...ns.pm6#L45 00:23
lue Mouq: yeah, I had a feeling. :) 00:25
00:27 dmol joined
lue Figures I can't use the debugger on Pod::To::HTML :| 00:32
Mouq lue: are you having a problem with "Unknown compilation input 'optimize'"? 00:33
lue Ah, nevermind, I just have to switch around the pod file and the --doc option. For some reason.
Mouq: Illegal option --doc
Mouq Oh. I was getting that on test files with BEGINs in them
Cool :) 00:34
00:34 Pleiades` left
lue ... and switching them around does nothing. Great. 00:34
$ perl6-debug-m ../specs/S26*pod6 --doc=HTML
>>> LOADING ../specs/S26-documentation.pod6
$
Mouq lue: Doesn't that call S26.pod6 with the option --doc=HTML? 00:35
00:35 dmol left, Pleiades` joined
lue probably would. But -debug-m won't accept a starting --doc 00:35
lue just realized why the debug say he was trying to use (before trying the debugger) probably wasn't printing to screen :P 00:36
00:37 virtualsue joined, virtualsue left
lue I see what's happening, "NEL" isn't actually a unicode name. 00:37
It's only got a unicode 1.0 name, and even then it likely doesn't mean the "\n" you want. 00:39
BenGoldberg rn: say "\c[NEL]" 00:40
lue (U+0085 has the name "NEW LINE (NEL)", the closest match it seems)
camelia rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized character name NEL␤at /tmp/tmpfile:1␤------> say "\c[NEL⏏]"␤»
..niecza v24-109-g48a8de3: OUTPUT«␤»
BenGoldberg rn: say "\c[NEW LINE (NEL)]"
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Unrecognized character name NEW LINE (NEL) at /tmp/tmpfile line 1:␤------> say "\c[NEW LINE (NEL)⏏]"␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting …»
..rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized character name NEW LINE (NEL)␤at /tmp/tmpfile:1␤------> say "\c[NEW LINE (NEL)⏏]"␤»
BenGoldberg rn: say "\c[NEW LINE]" 00:41
camelia niecza v24-109-g48a8de3: OUTPUT«===SORRY!===␤␤Unrecognized character name NEW LINE at /tmp/tmpfile line 1:␤------> say "\c[NEW LINE⏏]"␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1502 (d…»
..rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized character name NEW LINE␤at /tmp/tmpfile:1␤------> say "\c[NEW LINE⏏]"␤»
lue "NEL" and "NEXT LINE" are aliases for U+0085 , not the actual name. "NEXT LINE (NEL)" is the Unicode 1.0 name. It has no current Unicode name. 00:42
Mouq is both happy and annoyed that this super-hyped snow storm has missed us completely 00:43
geekosaur heh. we got the start of our hyped snowstorm and then it petered out early 00:44
Mouq "Don't expect school Monday and Tuesday", "Worst snowstorm of the season!" 00:45
-_-
lue r: say "\c[NL]Hello!" 00:46
camelia rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unrecognized character name NL␤at /tmp/tmpfile:1␤------> say "\c[NL⏏]Hello!"␤»
00:54 pippo left 00:55 rurban left, pippo joined 00:57 pippo left 00:58 spider-mario left 01:01 hoverboard left, colomon left 01:04 rurban joined 01:08 xenoterracide left
dalek ecs: 35e12b4 | lue++ | S26-documentation.pod:
Change E<NEL> to E<0x85>

Currently errors on rakudo. NEL is in reality an alias for a codepoint without a current Unicode name, U+0085. Since it's unclear if such aliases are supposed to work, for now play it safe.
01:17
c: f189ac4 | Mouq++ | htmlify.pl:
Make origin links point to #infix_*, fix search
01:23
Mouq updates doc.perl6.org 01:24
01:25 rurban left
lue wonders how long it'll be before perlcabal.org/syn and feather.perl6.nl/syn have the cool stylings of doc.perl6.org :) 01:28
Mouq github.com/perl6/Pod-To-HTML/commit/41db0 lue++ though I don't feel like this should be Pod::To::HTML's job 01:31
01:31 colomon joined
lue Mouq: there's no standard rule to match any kind of number, and convert it to decimal for you, so far as I know. 01:32
(though it would perhaps be nice if the rakudo side converted E<> things for you...) 01:33
Mouq lue: the latter is what I mean 01:34
lue: Though I'm not sure what to in the case of E<apos> etc
lue Like <somebody> said, at this point Rakudo's side needs to be founded on AST nodes, instead of building the collection of classes itself. E<> stuff is a perfect example of how ASTyness would be helpful. 01:35
Mouq: E<> is supposed to handle HTML entities, regardless of your chosen backend.
s/backend/generator/, to be clearer 01:36
Mouq perl6 --doc -e'=pod E<apos>' 01:37
apos
:|
lue: So what is desired behaviour there?
lue Output an apostrophe, clearly :)
(E<> definitely needs to be processed by rakudo, rather than backends) 01:38
::HTML happens to luck out with HTML entities, but ::Man and ::TeX and etc. would need to handle it similarly to how Rakudo would.
(if Rakudo didn't handle it, that is)
Mouq So rather than Pod::FormattingCode.new(type => "E", content => Array.new("apos")), you think it should be Pod::FormattingCode.new(type => "E", content => Array.new("'")) or Pod::FormattingCode.new(type => "E", content => "'")? 01:40
And then Pod::To::HTML can escape it itself?
lue Mouq: maybe? This is why ASTs would be better perhaps; it would be just another "plain text" node, perhaps with an attribute saying "special char; may need to handle this" 01:41
Mouq (with special casing for Pod::FormattingCode.new(type => "E", content => "\n") -> "<br/>"
)
lue But I definitely think it's detrimental to have every Pod backend implement E<> processing.
Mouq lue: And I'm not sure what you mean by AST 01:47
lue: What I've been saying about AST is that Perl6 should parse the code and generate an AST that builds the Pod at runtime, instead of compiletime 01:48
lue: But I'm not sure if you mean something completely different
lue Well, someone (I think tadzik) mentioned a while ago that perhaps the Pod parser in Rakudo should build a more QAST-ish structure, as opposed to the tree of Pod:: stuff it does now. (essentially, a tree of nodes that builds the Pod tree, instead of building the pod tree outright) 01:49
01:49 hoverboard joined
Mouq But the final output ($=pod and whatever's handed to Pod::To::*) would still be a tree of Pod:: stuff? 01:51
or do I misunderstand?
because I don't see a way around that
lue I forget what the specific issue was that gave rise to the idea, but presumably $=pod would hold the pod. 01:52
01:55 araujo left, araujo joined
lue Mouq: it came up while discussing M<> , and the person making the suggestion was you rather :) 02:02
Mouq: from S26: "By default, once the DOC INIT phase is complete, the interpreter then calls a special default DOC INIT block that walks the AST and generates the documentation, guided by the content of the $?DOC hint, the $=pod tree, and any DOC blocks that have loaded Pod-related handler code." 02:04
Sadly the "How Pod is parsed and processed" section doesn't seem to mention anything by way of processing *another* file's Pod from a Perl 6 script :/ 02:06
Mouq EVAL( $slurp ~ "\n$=pod")[0] :P 02:07
Which should really be like use File; File.pod or something
dalek ast: d429f6e | larry++ | S05-mass/properties-derived.t:
for <:Common> pick char with that property, duh
lue Mouq: too bad we can't have something like DOC import "../docs/file.pod6" :/ 02:14
02:14 JimmyZ left, JimmyZ joined
lue I'd much prefer something like $~Pod.parsefile("foo.pod6") actually 02:14
This is the first time I've run syngen all the way through post-flattening-patches on MoarVM, and it took 2098 seconds, a significantly worse time than the ≈1230s I had before :( Don't know what caused it though. 02:21
02:22 LooneyLemur left 02:24 klapperl_ left 02:27 klapperl joined 02:30 Sqirrel_ left 02:31 Sqirrel joined
TimToady p6: say so "〣" ~~ /<:NumericValue(3)>/ 02:38
camelia niecza v24-109-g48a8de3: OUTPUT«True␤»
..rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Unable to parse expression in metachar:sym<assert>; couldn't find final '>' ␤at /tmp/tmpfile:1␤------> say so "〣" ~~ /<:NumericValue[33…»
TimToady wow, there's nothing in roast about that sort of thing 02:39
02:44 __thou left, alc joined 03:06 xenoterracide joined 03:08 jnap left
skids r: =head1 foo$=pod.say # Hrm evalbot won't take U+0085 as an inline line separator; this works on CLI 03:12
camelia ( no output )
03:24 fridim__ left 03:39 hoverboard left
Mouq r: =head1 foo$=pod.say 03:52
camelia ( no output )
Mouq r: =head1 foo$=pod.say
camelia rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«Pod::Heading("level" => "1").hash␤ Pod::Block::Para␤ foo␤␤»
Mouq skids: ^^
04:00 laidback_01 joined 04:04 rurban joined 04:05 hoverboard joined 04:09 jnap joined
Mouq .ask smls What font is used in imgur.com/a/y1dHo ? It looks like Helvetica or Arial but not quite... 04:09
yoleaux Mouq: I'll pass your message to smls.
Mouq And definitely not either 04:10
04:13 jnap left
Mouq .tell smls And I can't say how much I love the mockup btw. I hope you don't mind if I've done the smallest amount of work getting the docs to look like it. The only thing I note on it is that Routine may need to be broken up into sub-categories 04:16
yoleaux Mouq: I'll pass your message to smls.
04:17 xinming left 04:20 xinming joined 04:35 xinming left 04:39 hoverboard is now known as muchbacon 04:40 xenoterracide left 04:41 xinming joined 04:42 muchbacon is now known as hoverboard
skids Mouq: thanks. Wonder why it needs two. 04:43
04:43 itz joined
Mouq skids: Because =alpha will take keep going until it meets a blank line or another =pod declaration 04:44
skids Was trying to test Module::{$=pod} but that seems not functional. Also no way to test on evalbot without pod in an available module in base install. Maybe Pod::To:Text could use some dogfooding.
Mouq: well on star 2014-01 one works. 04:45
04:46 itz_ joined 04:49 vinian joined, vinian left
Mouq r: say "\x[85]" ~~ /\r\n/ 04:51
camelia rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«Nil␤»
Mouq Pretty sure that's why. 04:52
04:52 vinian joined
Mouq It wanted to parse it such that it needed two, but it messed up with the \x[85] because of its outdated use of \r, I'm pretty sure 04:53
star: =head1 foo␤$=pod.say 04:54
camelia star 2013-09: OUTPUT«Pod::Heading("level" => "1").hash␤ Pod::Block::Para␤ foo␤␤»
Mouq No?
Humbug
Well, I'm 99% certain that its current behavior is spec
Pod thinks in terms of paragraphs, foo␤$=pod.say is a paragraph to it 04:56
lue yeah, $=pod.say should be part of the header in that star: line 04:57
Mouq p6: grammar Foo { token TOP { <a>* }; proto token a { ab {*} }; token a:sym<c> { <sym> }; token a:sym<d> { <sym> }; }; Foo.parse('abcabdabdabc').say # Can this work pretty please? 04:58
camelia niecza v24-109-g48a8de3: OUTPUT«「abcabdabdabc」␤ a => 「abc」␤ a => 「abd」␤ a => 「abd」␤ a => 「abc」␤␤» 04:59
..rakudo-parrot 456f3a, rakudo-jvm 456f3a, rakudo-moar 456f3a: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Proto regex body must be {*} (or <*> or <...>, which are deprecated)␤at /tmp/tmpfile:1␤------> n TOP { <a>* }; proto token a { ab {*} }[3…»
Mouq Niecza++
lue that {*} should rather be <...> I think, at least according to the rakudo grammar
Mouq lue: "(or <*> or <...>, which are deprecated)" :) 05:00
lue Eh. I think {*} looks too much like a useless closure for my tastes.
Mouq lue: Perhaps, but it echos the spirit of {*} in regular routines 05:01
lue yeah, but I think <...> is more to the spirit of regexes :)
05:09 jnap joined
TimToady <...> is supposed to be {*} nowadays 05:11
skids Mouq: I think the current behavior is spec, yes.
TimToady <...> is a fossil
lue TimToady: is <*> fossily too? 05:12
TimToady first I've heard of <*>
<*foo> is supposed to mean something else entirely 05:13
05:14 jnap left
skids
.oO(U+2029 is a "paragraph separator." Bother.)
05:14
TimToady you don't like paragraphs?
skids No just wondering what pod should do with it. 05:15
lue treat it as a paragraph separator, of course :)
skids Kinda hard to test given linux console just puts a space there when you type it. 05:16
05:21 [Sno] left 05:48 rurban left
dalek kudo/nom: 8f8ac8e | Mouq++ | lib/Pod/To/Text.pm:
[Pod::To::Text] support nested items
05:51
Mouq (^because: dogfooding^)
TimToady oh, oh, we have 666 entries on RC 05:53
06:20 darutoko joined 06:42 alc left 06:46 kurahaupo_mobile left 06:51 [Sno] joined 06:53 BenGoldberg left 06:57 pdcawley left 06:58 pdcawley joined, dayangkun joined 07:08 dayangkun left 07:31 xenoterracide joined 07:36 dmol joined 07:47 berekuk joined 07:48 Alina-malina left 07:49 Alina-malina joined 07:54 kaleem joined 07:59 FROGGS joined 08:01 berekuk left 08:05 berekuk joined 08:11 kaleem left, kaleem joined 08:12 zakharyas joined 08:14 kbaker joined 08:17 kaleem left, kaleem joined 08:23 kbaker left 08:25 darutoko- joined, salv0 joined 08:27 darutoko left 08:28 kaleem left, Ven joined 08:29 darutoko- left 08:30 kaleem joined, berekuk left 08:40 pdcawley left, hoverboard left 08:41 pdcawley joined 09:01 vinian left
dalek kudo/nom: 3333a41 | Mouq++ | src/ (3 files):
Parse and store Pod FormattingCode info in @.meta

This works in conjunction with an updating commit to Pod::To::HTML
09:05
Mouq is pretty happy that Just Worked after he finally got the parsing correct. Corresponding ::HTML commit: github.com/perl6/Pod-To-HTML/commit/bbdfc 09:16
moritz Mouq++ 09:17
09:22 fhelmberger joined
Ven Mouq: can't you do something like `my @meta := $<meta>.map: {~*}` or something ? 09:33
Ven is needlessly golfing 09:34
Mouq Ven: Where?
In the NQP?
Ven yes
does nqp not support that ?
Mouq Not so much...
It could be golfed to `my @meta := []; @meta.push($_) for $<meta>`, but if you need to debug that gets annoying 09:35
*push(~$_) 09:36
Actually... why does that work?
moritz why wouldn't it work?
Mouq r: =pod L<thing|perl6.org/>␤␤say $=pod[0].contents[0].meta[0].WHAT 09:39
camelia rakudo-parrot 3333a4, rakudo-jvm 3333a4, rakudo-moar 3333a4: OUTPUT«No such method 'contents' for invocant of type 'Pod::Block::Named'␤ in block at /tmp/tmpfile:3␤␤»
09:40 xinming_ joined
Mouq r: =pod L<thing|perl6.org/>␤␤say $=pod[0].content[0].meta[0].WHAT 09:40
camelia rakudo-parrot 3333a4, rakudo-jvm 3333a4, rakudo-moar 3333a4: OUTPUT«No such method 'meta' for invocant of type 'Pod::Block::Para'␤ in block at /tmp/tmpfile:3␤␤»
Mouq Bluh
r: =pod L<thing|perl6.org/>␤␤say $=pod[0].content[0].content[0].meta[0].WHAT
camelia rakudo-parrot 3333a4, rakudo-jvm 3333a4, rakudo-moar 3333a4: OUTPUT«No such method 'meta' for invocant of type 'Str'␤ in block at /tmp/tmpfile:3␤␤»
Mouq -_- 09:41
r: =pod L<thing|perl6.org/>␤␤say $=pod[0].perl
camelia rakudo-parrot 3333a4, rakudo-jvm 3333a4, rakudo-moar 3333a4: OUTPUT«Pod::Block::Named.new(name => "pod", config => ().hash, content => Array.new(Pod::Block::Para.new(config => ().hash, content => Array.new("", Pod::FormattingCode.new(type => "L", meta => Array.new("perl6.org/…»
Mouq r: =pod L<thing|perl6.org/>␤␤say $=pod[0].content[0].content[1].meta[0].WHAT
camelia rakudo-parrot 3333a4, rakudo-jvm 3333a4, rakudo-moar 3333a4: OUTPUT«(Str)␤»
Mouq Weird
moritz what would you expect it to be?
timotimo o/ 09:42
moritz \o timotimo
Mouq moritz: Not sure, but it's not being serialized like the rest of the code does to strings 09:43
09:43 xinming left
Mouq \o 09:43
dalek kudo/nom: 6479fff | Mouq++ | src/Perl6/Actions.nqp:
Serialize strings being put in Pod::FormattingCode.meta
09:45
09:51 dakkar joined
timotimo i'm now looking through the git commit logs to find things to put into the p6weekly 09:52
feel free to point out anything i might miss
dalek kudo/nom: ed13bb1 | Mouq++ | src/Perl6/Grammar.nqp:
Pod definitions are semicolon delimited
09:53
09:54 kbaker joined 09:57 telex left 09:58 telex joined
dalek c: 9da5a03 | Mouq++ | htmlify.pl:
Fix htmlify.pl for the latest Rakudo
10:05
timotimo r: say [<a b c>] Z=> <1 2 3>; 10:07
camelia rakudo-parrot 3333a4, rakudo-jvm 3333a4, rakudo-moar 3333a4: OUTPUT«["a", "b", "c"] => "1"␤»
timotimo good
r: say [1, 2].WHAT 10:08
camelia rakudo-parrot 3333a4, rakudo-jvm 3333a4, rakudo-moar 3333a4: OUTPUT«(Array)␤»
FROGGS timotimo: TimToady++ has probably fixed >60 spectests for perl6-m about unicode properties, so we are down to 71 failures as of yesterday 10:09
nwc10 where is the JVM?
FROGGS timotimo: dunno if you have mentioned it already, panda now supports moar as well 10:10
4 rakudo.jvm 2014-03-02 100.00% 39214ef 28979 5 610 1505 31099 28579
5 rakudo.moar 2014-03-02 99.83% 39214ef 28931 71 577 1334 30913 28579
6 rakudo.parrot 2014-03-02 99.59% 39214ef 28859 64 595 1395 30913 28579
parrot should be clean in today's run
nwc10 so Moar is 48 behind JVM? 10:11
FROGGS sort of
nwc10 but Parrot has the only star.
FROGGS it is quite possible that tests are fudged on jvm that work on moar and vice versa
masak good antenoon, #perl6
FROGGS hi masak 10:12
timotimo tro
er.
FROGGS: how deep does the moar support go?
can i just go around and perl6-m rebootstrap.pl and panda my::favourite::module? 10:13
FROGGS timotimo: yes
timotimo i'll surely mention that
FROGGS but you might want to mention that there are two known precomp bugs that break modules like URI
maybe it is just one bug, I dunno 10:14
timotimo yeah
FROGGS I am working hard on the S11 stuff if you want to put that in 10:15
timotimo rakudo-parrot should get back up to 99.72%, right?
dalek kudo/nom: f3cf9d5 | Mouq++ | src/Perl6/Grammar.nqp:
Allow any amount of content in C<code> etc
10:16
FROGGS p: say (28859+64) / 28979
camelia rakudo-parrot 3333a4: OUTPUT«0.998068␤»
timotimo i should also mention that if we were to run the moar-conc branches for the daily spectest runs, we would be getting even more passes.
FROGGS yes
and when it is merged we can tick some boxes here: perl6.org/compilers/features 10:17
which is very nice :o)
timotimo \o/
FROGGS: i'd like to link to the "experimental S11", where do i find that? 10:23
FROGGS rakudo/eleven and panda/eleven 10:24
timotimo i meant the spec document :)
FROGGS and perhaps link to the S11 spec as well, it fits almost the implementation I think :o)
timotimo oh, that's in specs/master?
cool.
FROGGS yes
maybe link directly to perlcabal.org/syn/S11.html#Versioning 10:25
timotimo yes
FROGGS did I mention that I love the way you can invoke `panda-m --auth=github:tadzik install Foo` ? 10:26
timotimo oooh :)
FROGGS (you can even invoke scripts from different dists that install binaries under the same name) 10:27
like Foo and Blurb provide a bar2baz script 10:28
you will end up with bar2baz-m bar2baz-p and bar2baz-j, and to invoke the correct script you can do: bar2baz-m --name=Blurb <args for bar2baz here> 10:29
timotimo that already works? o_O
FROGGS I've not tested it for some weeks but it should still work, yes 10:30
timotimo oh wow!
i didn't even realize!
FROGGS only on linux though, I still need to write down the windows wrapper script
timotimo yeah
what a drag ;) 10:31
FROGGS :o)
Mouq FROGGS++
FROGGS thanks :o)
timotimo finally some breakfast \o/ 10:32
Mouq FROGGS: Does =NAME and so on work?
FROGGS ahh, and I need to think about a File::DistDir alike so that a module can load its shared data files
Mouq: you mean how a CU specifies its metadata? 10:33
Mouq FROGGS: mmhm
FROGGS no, I think that is not handled yet 10:34
I just cared about what happens when you install a distribution
Mouq FROGGS: Ok, it just caught my eye :)
FROGGS at that time you already have a META.info where everything should be included
so, some day we need a tool that bundles a bunch of files for release, perhaps creating the META.info on its own 10:35
10:39 darutoko joined
timotimo is that what perl5ers use Dist::Zilla for? 10:40
FROGGS but this tool is probably a combination of a local installed toolchain and a proper dist upload server 10:41
timotimo: sounds like, I only used Module::Build
dalek ecs/pod6ify: f2f6e3b | Mouq++ | S26-documentation.pod6:
[S26] This won't work without explicit code block
10:43
Mouq Oop. Should probably be on master
dalek ecs: 39866b1 | Mouq++ | S26-documentation.pod:
[S26] This won't work without explicit code block
10:45
Mouq Much better. o/ #perl6 10:46
timotimo o/ Mouq 10:47
FROGGS ciao Mouq 10:48
10:52 Mouq left 11:07 xenoterracide left 11:08 denisboyun joined 11:17 xenoterracide joined
colomon masak: ADT started failing overnight, Test::CSV still failing 11:28
CSS::Grammer started failing overnight
timotimo: ADT failed overnight
masak: meant Druid, not ADT, for you. :) 11:29
timotimo colomon: oh noes :( 11:34
it's tough being a module author
even tougher since i don't have any users, at least to my knowledge 11:35
colomon on the bright side, your module is pretty cool
timotimo what a relief :) 11:40
i appreciate the notifications in any case
it's slightly annoying to have a module that needs fixing every now and then, but it'd be much worse to have a module that silently fails as soon as someone tries it out 11:41
tadzik lue: (Pod in QAST) no :) 11:42
11:42 xenoterracide left
tadzik actually, I did extra effort to do it the way it's done now, and I think it's the right way to do that 11:42
timotimo tadzik o/ 11:43
you got a favourite thing i should add to this week's p6weekly?
tadzik \o
colomon o/ 11:44
colomon wonders if anyone else has tried installing File::Find on moar
tadzik timotimo: other than panda on moar? :)
colomon: I usually don't since updating modules that are panda deps usually breaks panda :|
timotimo tadzik: already in :) 11:45
colomon tadzik: emmentaler does it, and emmentaler / panda completely break when it does
tadzik oh, but it installs it to a different location, not updates the existing one 11:47
that should work, let's see
or am I one of those lucky fellas who can't connect to feather:3000?
...yes 11:48
colomon tadzik: I poked around the issue a bunch the other day.
tadzik IO::Socket::INET.new(:host<feather.perl6.nl>, :port(3000)) hangs for me
(on moar)
let's see if the update helps 11:49
colomon tadzik: for what it's worth, I think it's a moar bug and not a panda bug. 11:53
tadzik it is 11:54
lucky for me :D
ok, the bug is still there in the latest moar (for me)
timotimo aaow :( 11:55
aaw*
12:09 dmol left
masak colomon: thank you for the reports. 12:10
I looked into Text::CSV, but I couldn't identify the problem quickly. it seems parsing-related.
colomon tadzik: what it looks like is happening is File::Find installs (this is with pandabrew/bin/pandabrew panda install File::Find, but the same error comes from emmentaler), then when it wants to clean up the install process it calls rm_rf (Shell::Command) which in turn calls find (File::Find) and when that call is made, it bombs out with an illegal opcode. 12:12
tadzik huh 12:14
as if it indeed overwrote the original File::Find, and re-loaded it somehow?
colomon yeah, that's what it seems like, and that's why I suspect a Moar bug 12:16
it's not even supposed to be possible to trigger a re-load, is it?
tadzik to my knowledge, no 12:17
colomon once it happens, panda is completely borked until you rebootstrap it
12:18 colomon left 12:22 bluescreen10 joined 12:24 colomon joined
masak this is pretty big: perltricks.com/article/72/2014/2/24...signatures 12:28
tadzik it is 12:29
nwc10 that article is more accurate now than when it was first published.
but "PerlTricks.com and the authors make no representations or warranties with respect to the accuracy or completeness of the contents of all work on this website and specifically disclaim all warranties, including without limitation warranties of fitness for a particular purpose. No warranty may be created or extended. The advice and strategies published on this website may not be suitable for every situation. All work on this website is provided with
so that's OK. 12:30
FROGGS cool (sub sigs)++ 12:33
12:33 xenoterracide joined
Ven so it took them as much time as perl 6 to mature that much :p 12:33
masak Ven: yes, but don't ridicule them too much for that. they have other constraints than Perl 6 does. 12:35
Ven masak: well, that's the point. It's ridicule to have *that much constraints*. I don't even know how they can move 12:36
FROGGS let's just agree that we are happy to have signatures in Perl 5 as well, because we love Perl 5, okay? :o)
masak what FROGGS said.
nwc10 If you move too far and break their code, your userbase doesn't come with you: alexgaynor.net/2013/dec/30/about-python-3/ 12:37
masak I'm not here to dump on Perl 5. those guys are doing an amazing job.
FROGGS exactly (both of your sentences)
12:37 xenoterracide left
masak the more I'm involved in the Perl community, the more respect I have for Perl 5. 12:37
the fact that Perl 6 is an awesome vision doesn't change the fact that (size-wise), we're the Rebels to Perl 5's Empire. 12:38
12:40 spider-mario joined
Ven maybe perl got the "urge to move" stance a bit better? 12:41
s/perl/the ruby community/ # duh
masak Ven: I'm not following what you're saying. 12:43
Ven masak: from what I've seen in the community, the 1.9 had a nice list of breaking change (certainly not as much as perl 6 and probably less than py3), but people have been prompt to upgrade 12:44
masak Ven: that sounds quite likely to me. 12:45
Ven: maybe there is even some measure of "entrenched"-ness along which Ruby is much less affected than either Python or Perl.
Ven masak: maybe we're just hipsters who try the new things instantly :p 12:46
masak Ven: seriously though, I think it's connected to how much code you already have that'll break. 12:47
Ven: also possibly connected to how good test coverage you have.
Ven masak: that's why php people don't do any breaking change -- they don't have test </troll -- I'm sure that's not true>
masak: but I guess it works much better if, instead of a big list of breaking changes 12:48
masak: you just break little things over time. Less things to fix in one time, you're not "left behind" 12:49
masak yeah. 12:51
or even "of course" :)
12:51 colomon left 12:53 skids left
masak someone please remind me, what is the undocumented grammar debugging feature in Rakudo again? 12:53
tadzik <?DEBUG(1)> I think
masak thank you.
hm, no; doesn't work. 12:54
tadzik hm 12:55
bah, google sucks for searching irc log 12:56
and I can't find it with normal search either :( 12:57
12:58 rindolf joined
masak I tried doing 'git grep' in the Rakudo source, but I came up blank, too. 12:58
moritz that would be an NQP feature 12:59
but also there git grep --word DEBUG comes up empty
masak: there's a --rxtrace command line option, and $grammar.HOW.trace-on($grammer) 13:00
13:03 virtualsue joined
masak moritz: thank you. 13:03
ahaha. --rxtrace traces the parse of my *program*, not my input. :D 13:04
moritz yes :-)
for your own grammars, you have the grammar object available :-)
masak ===SORRY!=== 13:05
No such method 'trace-on' for invocant of type 'Perl6::Metamodel::GrammarHOW'
:/
13:05 BenGoldberg joined
masak guessing this is an NQP feature, too... 13:05
tadzik why not Grammar::Debugger? 13:06
or Grammar::Tracer?
moritz ah yes, it's from src/how/NQPClassHOW.nqp
13:07 vinian joined, vinian left
masak tadzik: because I tried it on Friday and discovered it's busted. 13:07
tadzik: gist.github.com/masak/eccadc58b72cd32f720e
tadzik: I don't know if it's busted gloablly on all Rakudo backends, or just on Moar. 13:08
13:10 DarthGandalf joined, darutoko left, xenoterracide joined 13:11 benabik left
moritz masak: is the tracer also busted? 13:13
masak good question. 13:15
masak checks 13:16