pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
meppl good night 00:06
00:12 meppl left 00:14 Limbic_Region left 00:16 silug left 00:25 devogon left 00:26 Limbic_Region joined 00:33 hugues24 joined 00:37 justatheory joined 00:40 [particle1 joined 00:43 lichtkind left 00:44 [particle2 joined 00:49 aindilis left 00:57 [particle] left 01:01 [particle1 left 01:05 [particle] joined 01:10 aindilis joined 01:14 DaGo left 01:15 [particle2 left 01:22 [particle1 joined 01:35 [particle] left 01:53 silug joined 01:55 lambdabot left 01:56 lambdabot joined 01:57 blindfish left 02:13 justatheory left 02:23 Jedai left, Jedai joined 02:41 tcliou_ left 02:44 literal joined 03:54 Lorn left 03:55 Limbic_Region left 04:18 RayMagini left 04:20 RayMagini joined 04:26 chacha_chaudhry joined 04:28 ahmadz joined 04:30 ahmadz left 04:32 chacha_chaudhry left 04:33 chacha_chaudhry joined 05:02 chacha_chaudhry left 05:55 silug left 05:58 Schwern left 06:16 jwest-- left 06:21 meteorjay joined 07:19 smtms joined 07:24 cj left, cj joined 07:26 devogon joined 07:29 arcady left 07:30 drrho joined 07:45 fridim_ left 07:52 silug joined 08:02 Aankhen`` joined 08:08 cj_ joined 08:13 cj left 08:14 iblechbot joined 08:20 IRSeekBot left 08:23 IRSeekBot joined 08:29 silug left 08:46 BinGOs_ is now known as BinGOs 08:58 drbean_ joined 09:00 drbean_ left 09:08 meppl joined 09:30 barney joined 09:48 njbartlett left, njbartlett joined 09:57 whiteg_ joined, gugod left, gugod joined 10:03 BinGOs_ joined 10:17 BinGOs left, iblechbot left 10:26 silug joined, iblechbot joined 10:37 chris2 joined 10:52 BinGOs_ is now known as BinGOs, Chillance left 10:54 lichtkind joined
lichtkind are min and max arraymethods? 10:55
11:05 silug left 11:36 cls_bsd joined 11:53 iblechbot left 11:55 drrho left 12:06 DarkWolf84 joined 12:22 jrockway left, jrockway joined 12:23 drrho joined
pugs_svnbot r20043 | putter++ | src/perl6/STD5_run_post: a postprocesser for STD5_run, which improves the yaml output (currently by removing the prior field in class Perl). 12:40
r20043 | putter++ | ./STD5_run EXPR '-3+2' | ./STD5_run_post
r20044 | putter++ | src/perl6/STD5_run_post: fix the 'example usage' comment.
diff: dev.pugscode.org/changeset/20044
lambdabot Title: Changeset 20044 - Pugs - Trac
12:44 silug joined 13:00 iblechbot joined 13:18 barney left 13:21 chris2 left 13:45 njbartlett left 13:57 aindilis left, aindilis joined 14:04 jrockway left, jrockway joined, DarkWolf84 left 14:10 wknight8111 joined 14:21 Zygo left, Zygo joined 14:34 jrockway left 14:56 chris2 joined 15:01 turrepurre left 15:07 wknight8111 left
TimToady lichtkind: yes, but min and max are also binary operators. so [min] and [max] are reduce ops. 15:22
15:23 iblechbot_ joined
lichtkind TimToady: yes than one i remembered thanks i currently updating my perl 6 tutorial :) 15:25
TimToady: do you consider perl table as a bad name for an tutorial ?
TimToady: the idea behind it is similiarity to emerald table and also table in a sense that you know where to look for if you look for something specific 15:27
15:29 iblechbot left 16:11 simcop2387 left 16:13 kanru joined, simcop2387 joined 16:25 silug left 16:29 drrho left 16:43 sordith_ joined
sordith_ sits with lichtkind at his perl table and computes 16:44
16:45 mncharity joined 16:50 marmic left 16:51 sordith left 16:54 marmic joined
mncharity Consider "class A { }". 17:04
The STD story is token package_declarator:class { <sym> <package_def> {*} } 17:05
modified by a proto token package_declarator (:$endsym is context = 'nofat') { }
to token package_declarator:class { 'class' <nofat> <package_def> {*} }.
The a parser/compiler takes that {*} or Match, and does something magic.
If the user wishes to make class() do something slightly different, say take an extra argument, it's unclear how they should interact with the magic. 17:06
There's also been a second story, fuzzily in my mind, whose validity or connection to the first I dont' grok. 17:07
That the prelude has a macro package_declarator:class is parsed / <sym> <package_def> / {...}.
Now, the <block> in <package_def> is spec'ed as "executes at compile time as the body of a method of the metaclass, which is responsible for interpreting the keywords of the class definition". 17:08
Suggesting that the {...} above rewrites the block, changing say "method m(){}" to something like
"./method(m,\(),{});", and then eval'ing it A.
17:08 wknight8111 joined
mncharity With the second story I understand what a user does to modify 'class'. 17:09
But... thoughts?
In the second story I don't understand how BEGIN{} works. It is needed to finish parsing the <block> (eg, it might modify the running parser). But it can also see things earlier in the <block>, which the macro hasn't had a chance to rewrite, or eval in the right place. 17:12
s/chance/chance yet/ 17:13
17:15 lisppaste3 left
mncharity It makes me think of having separate PARSE{} and BEGIN{}, one only able to see the parser, the other... dunno...? 17:16
s/and then eval'ing it A./and then eval'ing it in A./ 17:17
17:30 lumi left 17:31 lisppaste3 joined 17:38 whiteg_ is now known as whiteg 17:43 njbartlett joined
mncharity Perhaps 'method' is a macro, which does the rewrite. And perhaps BEGIN{} calls a, not magic but unusual, sub which grovels upward over the ast. An "eval everything which hasn't been, in preparation for BEGIN' sub. Which seems a bit messy. Then the package_declarator:class macro doesn't do much. In the absence of a BEGIN... hmm. 17:47
17:49 Psyche^ joined, Patterner left, Psyche^ is now known as Patterner 17:55 DarkWolf84 joined, njbartlett left
TimToady mncharity: declarations always take effect immediately, so in essence the method reduction is what actually runs the code in the metaclass to install the method; it's not a separate pass from the parser, though that's how the 2nd story makes it seem. 18:03
lichtkind: 'table' can't really be made to mean that in English. It's either something you sit in front of, or it's a rectangular layout of data. 18:05
mncharity: in fact, the way declarations are specced to work is that the name is inserted into the symbol table the moment it is seen, even before we've seen the rest of its declaration. 18:06
the rest of the declaration just modifies the initial entry 18:07
whether by (...) signature or "is" traits or by a {...} body
sub foo ($x) { foo($x-1) if $x > 0 } is legal self-reference 18:08
as is sub foo ($x = foo($x-1)) {...}
likewise any declared item in a signature can refer to earlier items in the same signature 18:09
so as soon as you see "method foo" the metaclass must be told that "foo" is a valid method, though not completely defined yet 18:10
lichtkind TimToady: and tablets?
TimToady well, people will think "pills", unless you say "stone tablets" in which case people will think about the Ten Commandments :) 18:11
lichtkind TimToady: but the emerald tablet were there first and moses did just cheap ripoff 18:12
but true the guy from tpf also had problems with it
TimToady I've never heard of the emerald tablet... 18:13
sounds like something from Oz 18:14
lichtkind haha
TimToady maybe it's called something else in our culture?
lichtkind its at least old as pyramids it from agyptian mystery tradition
written long before moses came to egypt 18:15
18:17 pmurias joined
pmurias lichtkind: tables brings more a referance than a tutorial to mind 18:17
lichtkind pmurias: but thats exactly the direction i heading with that attempt :)
pmurias * reference 18:18
TimToady I seriously doubt it's as old as Moses
feels gnostic to me
lichtkind yes was also part of gnostic traditian
TimToady oldest attestation is about 800ad 18:19
lichtkind tradition
shure
but thats only science pint of view 18:20
TimToady the name Hermes also strikes me as rather derivative of greek culture 18:22
18:22 silug joined
lichtkind TimToady: for good reason but thats whyt it was called in theses days 18:22
mncharity hmm. ok. so the key question is how things provide context downwards. eg, a 'class' macro saying "btw, you're now inside a class, not just a module", for all the code which gets run in it's <block>, before the macro gets to touch the control flow again.
lichtkind TimToady: before these times it was called djehuti
TimToady well, let me know if you have any luck turning lead into gold... 18:23
mncharity Perhaps things groveling *up* the ast is part of the story? but has the flavor of an abstraction violation. perhaps one could constrain it as passing a "blackboard" down, somehow restricting/clarifying what gets scribbled there. 18:24
TimToady mncharity: so far we've been using context vars for things like that, but maybe we need a more official way to pass info down the parse tree
lichtkind TimToady: i never tried but its easier than many think if you know the works of walter russel 18:25
mncharity macro package_declarator:class is parsed / <sym> <package_def> / has blackboard(:package_type<class>) {...} ? 18:26
has gossip 18:27
TimToady is parsed / :my $?CLASS = self.mumble; <sym> <nofat> <package_def> / 18:28
or some such
:my $?CLASS is context = ... 18:29
hmm, :constant $?CLASS is context =.... maybe 18:30
something that looks lexically scoped to a user is often dynamically scoped in the compiler 18:31
pmurias TimToady: is $?CLASS constant?
mncharity is parsed / {do_classy_stuff()} <sym> <nofat> <package_def> / ? to abstract away the details, else I fear / <sym> <nofat> <package_def> / becoming many lines of opaque code.
TimToady it's a constant at run time, but the compiler is in the business of determining the value of all sorts of constants, and limiting their scope dynamically
(dynamically to the compiler, statically to the user)
well, <sym> is just parsing "class", so we haven't even got to the name yet 18:33
package_def parses that part
or something package_def calls...
gotta go play drums 18:34
bbl &
mncharity yes, but in general, one want's to abstract it away into an api, no? eg, it's no just :my $?CLASS is context = ..., but :my $?CLASS is context = my $?MODULE is context = my $?PACKAGE is context = ... ; and that's probably not the end of it all. 18:35
mmm, play drums. neat idea. 18:36
nah. STD_red for a few hours I think.
lichtkind where i can find the std.pm ?
mncharity src/perl6/STD.pm. Run with STD5_run. parses '-42+3'. 18:37
err, ./STD5_run EXPR -e '-42+3'
there's also misc/STD_red, with std.rb, a ruby hand transliteration of STD.pm. 18:38
pugs_svnbot r20045 | pmurias++ | [kp6-mp6like] removed a bit of accidentaly comiited code which didn't compile 18:39
diff: dev.pugscode.org/changeset/20045
lambdabot Title: Changeset 20045 - Pugs - Trac
18:42 Aankhen`` left, peeps[work] joined
lichtkind mncharity: what is std_red? 18:47
thanks
(for the answer)
so pugs seems still alive? 18:48
mncharity TimToady: in STD.pm, the paragraph beginning "Also, some rules are named by syntactic category", seems to suggest a infix:sym<+> would get a default "symbol name for the rule" of '+', not 'infix:+'? which seems odd.
lichtkind: STD_red purpose is to be a rapidly available 'more or less full' parser for p6, which runs useably fast, to serve as a temporary support for development of things further down the pipeline (ast's, compilation, and backends). It differs from the other similar things in: 18:52
18:52 Psyche^ joined 18:53 ikeda joined
mncharity STD5 is more focused on figuring out the true longest-token parse story. STD_red just fudges, accepts backtracking, and hopes not to lose to much of the language. parrot 'parse and emit ast', not yet available, plans on doing an incremental converge on STD. If it existed, and proved to have good coverage, STD_red would probably stay on hold. Next, 18:54
18:55 ikeda left 18:57 ikeda joined
mncharity ./script/kp6 -Cast-yaml, parses only the small kp6 dialect of p6. redsix uses a ruby version of pugs's old parser grammar, which is now somewhat out of date. By itself, redsix would be easier to set up as a temporary support. But STD_red has fringe benefits of helping somewhat with STD.pm, and being closer to spec, which helps make downstream more spec-ish. 18:57
And pugs, which can do -CPIL-mumble, but doesn't include oo information, which is needed as soon as we have a running pipeline, to pour prelude and rx_on_re and other p6 code through. 18:59
maze of twisty little passages. 19:00
19:02 silug left, rindolf joined
mncharity re alive, in the last few days, it seems the pugs build broke. i'm unclear on the current story. if it doesn't converge soon, we'll likely need to revert. or perhaps someone haskelly can take on what would have helped us many times over the years, duplicating pugs into a stable and development copy. We were able to fudge the two together early on because we 19:03
19:03 Patterner left, Psyche^ is now known as Patterner 19:04 ikeda left
mncharity had high-powered hs developers. It's rather more iffy now. 19:04
*who could keep the build working, even while doing major changes 19:05
lichtkind ah thanks but pugs itsel dont evolve these days? 19:07
19:07 Daveman left 19:08 ikeda joined
mncharity mostly just "adapt to new versions of ghc" maintenance. 19:08
i suspect once more of the fuzzy uncertainties about what a p6 implementation should do (which in it's day, pugs was key in pruning back, and now, other implementations are helping with), a hs implementation will wake up again. there's just so much neat stuff one can do with one. 19:09
19:10 njbartlett joined
mncharity but pugs is, by far, still the closest thing we have to a working p6 implementation. 19:11
err, avoiding ambiguity, that's *still the thing we have which most resembles a working p6 implementation 19:13
19:13 Daveman joined
ewilhelm any interest in mentoring pugs projects for summer of code? 19:14
mncharity ewilhelm: pugs as in haskell pugs, or non-parrot perl 6 in general? 19:15
ewilhelm either
19:15 Chillance joined
ewilhelm www.perlfoundation.org/perl5/index.cgi?gsoc2008 19:16
lambdabot Title: gsoc2008 / Perl 5 Wiki
diakopter ewilhelm: leave a message for ruoso and/or fglock 19:18
mncharity lichtkind: re closest thing in development space (rather than resemblance), there are currently three interrelated visions, and lots of commone pieces: parrot's rakudo (fully parrot based impl); my decoupling implementation for bootstrap on a big and fast p6 subset and p5/ruby backends; and kp6's bootstrap on a small p6 subset. 19:19
ewilhelm: looking...
lichtkind mncharity: word about rakudo is well spread but havent herad recently anything abou tkp6 19:20
mncharity kp6 has lately been less active than it once was. though pmurias has been working recently to give it a faster p5 backend, hoping to work towards a bootstrap of kp6 on kp6. 19:23
ewilhelm: ahh, you're looking for mentors as the mentor admin, not as a mentoree. :)
ewilhelm yep, trying anyway 19:24
we still have to write the application and hope google accepts us
mncharity hmm. a task would have to have what characteristics: not critical path; not likely to be rendered obsolete by changing context; cost less time in mentoring than doing it yourself; ... what else? 19:25
ewilhelm well, the student will be working full time for 11 weeks and getting paid 19:27
mncharity concrete deliverables (it would be nice to have people become domain experts on various things, to serve as resources, but that doesn't fit well with the soc model).
ewilhelm yeah, hopefully you gain a long-term participant 19:28
and one that had to compete to get in
19:29 rimad joined, innocuous joined
ewilhelm but the main deliverable is code 19:29
mncharity part of the problem is student time ranges from what passes for software engineers in the world, through people you'd hire at $200K/yr and feel you got a deal. which makes it hard to target.
ewilhelm of course, the ideas page is only for suggestions 19:30
19:30 ikeda left, ikeda_ joined
ewilhelm assume 11 weeks at 40hrs/wk with allowance for learning time 19:30
nevermind that this is $8/hr -- it is an internship 19:31
mncharity perhaps something Perl 6 spec-ish? there's lots and lots of work which needs to be done. some (much?) of it not bottlenecked on deciding what the spec should be. 19:32
*needs to be done to get us to a Perl 6 certification test suite.
ewilhelm as long as it is code -- google won't buy documentation-only projects 19:33
mncharity either manual groveling or generative testing
re doc-only, I'm thinking test cases and infrastructure. but I'm out of date / unfamiliar with what the current t/spec vision/story is.
ewilhelm hmm, actually it is $10/hr 19:34
don't get too detailed thinking of project ideas -- you first need to help sell google the idea of "Perl Foundation" as a mentor organization, then you need to sell the student on the idea of working on Perl 6 19:35
mncharity for p6 coding, so much depends on where we are at a few months from now. we might have a wizzy p6 by end of March, or still be slogging along come summer. 19:36
lichtkind mncharity: i also ask because the things i recognized on pugs werent fixed in the last 6 month therefore, even nevermind to hype perl 6 i bit careful to write something about pugs for larger publisher when its not alive
ewilhelm cannot read the future either
mncharity ah, ok. yes, for that, no, pugs is not under active development. 19:37
ewilhelm so, would any SOC work "re: perl 6" be better aimed at supporting/testing the language on parrot? 19:38
or is there something that a student would want (and be able) to do from the pugs approach?
lichtkind mncharity: i also heard that someone is doing something also called p6doc (i maintain the module perl6::Doc) 19:39
19:39 ikeda_ left
mncharity re 'read the future either', my impression, and thus my focus, is it's not actually a technical problem anymore, but a social one. eg, given a couple of good fte resources, I strongly suspect we could have a pugs-like p6 (on ruby backend, static parser, prelude in p6) by end of month. so the question is how to find the people, and get them moving in the same direction. 19:40
diakopter ewilhelm: ruoso is working on a possible backend for kp6/mp6 (fglock's kindaperl6 Perl 6 bootstrapping implementation) - see www.perlfoundation.org/perl6/index.cgi?smop for details. Also see www.perlfoundation.org/perl6/index....kindaperl6 19:41
lambdabot Title: SMOP / Perl 6
mncharity an old thought was to structure it as a ruby-community hackathon, a few days of low order 10 people. but current vision looks like there's much more p6 than ruby code, so that's looking less viable now. 19:42
lichtkind mncharity: why you think its so much ruby inside of pugs? 19:43
diakopter ewilhelm: along with what mncharity mentioned, a possible project could be to "complete" (haha) the Perl 6 test suite (which is currently officially maintained in the pugs subversion repository)
lichtkind mncharity: my focus is strictly on organising and writing docs, since i have project my own
mncharity clarification: smop will work as a backend for anything. given some of what it requires from it's compiler (eg, making lexical scopes go away), a kp6 dialect implementation actually seems the less likely path. 19:44
diakopter mncharity: oh; clarification taken
mncharity but perhaps not less likely by much :)
ewilhelm is just herding cats
lichtkind which status has SMOP i read the wiki but not learned from it that much
mncharity lichtkind: "why you think its so much ruby inside of pugs?", I don't understand? 19:45
diakopter ewilhelm: I am sure that the "Perl 6 Test Suite" SoC project would
meet the qualifications
ewilhelm volunteer to mentor and add your ideas www.perlfoundation.org/perl5/index....8_projects
mncharity re cats, there are some hiding on top of the cabinet
lambdabot Title: gsoc2008_projects / Perl 5 Wiki
diakopter and I'm sure you could find someone here to nominally mentor, though the whole channel would end up mentoring, really.
ewilhelm or just add your ideas 19:46
well, mentors need to be readily available to guide the student through the shark-infested waters of irc/lists
and to nag them for commits, progress reports 19:47
mncharity sharkbots
diakopter ewilhelm: good point.
lichtkind mncharity: why do you think, there is so much ruby code inside of pugs?
mncharity ok, food. bbiab. re www.perlfoundation.org/perl5/index....8_projects , perhaps highlight the deadlines and what is needed for them? 19:48
lambdabot Title: gsoc2008_projects / Perl 5 Wiki
mncharity lichtkind: there is no ruby inside pugs, the haskell p6 implementation. there is some ruby in pugs, the repository, mostly as misc/pX/Common/redsix , though also as misc/STD_red and a stillborn ruby backend for kp6, well, because I wrote it. :) the motivation being 19:50
lichtkind mncharity: may i ask where you are :) 19:51
mncharity it's traditional, and easiest, to implement a new language, and get it bootstrapped, by building on a similar but existing language. eg, lisp-like languages usually bootstrap from lisp. ruby is probably the most p6-like existing language. and also has metaprogramming as a strength, which makes it even more valuable for the task.
re where, geographically? Boston 19:52
lunch. bbiab. 19:53
lichtkind allright
mncharity: i ment more something else your vision what you done ar so 19:54
19:54 justatheory joined, Rmann joined
diakopter ewilhelm: I added 3 entries to the suggested projects list. 19:59
lichtkind ewilhelm: what cats you are hearding? 20:04
mncharity re where, ah, d'oh. backends are not a problem. direct to p5 (simple, fast, ok but not great p6 coverage). direct to ruby (similar, but uninteresting because). better ruby backend (fast, surprisingly good coverage). 20:05
compilation is largely unneeded. redsix, kp6, we're all basically just emitting from a slightly massaged asts. 20:06
20:07 ikeda joined
mncharity ast's aren't a big deal. it's a day's or so work to roll infrastructure for a new (with existing spec, and running code) one. getting it right is very much harder, but unneeded and explicitly not a goal for the exercise. 20:07
parser's are the gotcha. we don't have one. pugs won't talk about oo. kp6 is very restricted. parrot doesn't yet export/expose it's ast (though that could change any day now). STD5.pm is still limited, and is more research than "be infrastructure *now*. STD_red is a bit less limited, and has the right focus, but needs lots of grunge work. redsix... looks less 20:10
interesting with the prospect of parrot soon, and STD_red maybe.
so the development path looks something like 'make the parser bottleneck go away' (finally, after 3-ish years of blockage). slap an ast on the back. do dumb a dumb but fast emitter/backend. start writing p6 objects. priorities being prelude, and a p6rx_on_p5re engine (probably translated from the p5 yet_another_regex_engine). bootstrap whenever - doesn't really matter because you're already writing and running ok p6. 20:14
It's literally under a person week once there's a parser. the kp6 ast wrapper took a day (albeit without finishing a schema validator). the ruby backend is basically just copying redsix's with some tweaks. 20:18
20:19 ikeda left
ewilhelm lichtkind: mentors and students -- its like herding cats who herd cats 20:19
mncharity lichtkind: so that's where I'm at. thinking of do the next task on the STD_red path, 'sync and check against current STD.pm, and add Match constructors', because I don't have a feel for when parrot rakudo's yaml ast will become available (eg, it's rsn, but does that mean Monday, or mid March), and what it will cover. It's March. Time to get moving. 20:21
lunch,really & 20:22
pmurias ewilhelm: will TPF participate in 2008 SoC? 20:58
21:00 silug joined, cmarcelo joined
ewilhelm pmurias: that's what I'm trying to make happen 21:02
TPF has "not impressed" google in the last few years though
so I need to demonstrate that we have ourselves organized -- need to show a significant pool of willing mentors and project ideas in our application 21:03
(by next friday)
pmurias ewilhelm: i would like to work on kp6 as my SoC project 21:05
shower& 21:07
ewilhelm pmurias: you're a qualifying student?
check back on ~Mar 17th 21:08
21:17 hugues24 left 21:23 wknight8111 left
pmurias ewilhelm: i hope i'll be ;) (i won't if i fail to get accepted to university) 21:27
mncharity :) 21:28
21:29 Psyche^ joined 21:30 ewilhelm left 21:32 rindolf left 21:45 Patterner left, Psyche^ is now known as Patterner 21:48 DarkWolf84 left 21:49 buchetc joined
lichtkind is there a sprintf format rule to get dotty number like 44.444 ? 21:51
pmurias %.3f 22:00
pugs_svnbot r20046 | pmurias++ | [kp6-mp6like]
r20046 | pmurias++ | runtime in KindaPerl6::Runtime::MiniPerl6Like
r20046 | pmurias++ | for and .map for arrays support
r20046 | pmurias++ | autobox is used
diff: dev.pugscode.org/changeset/20046
lambdabot Title: Changeset 20046 - Pugs - Trac
pmurias any interest in a mp6like ROADMAP? 22:02
lichtkind pmurias: thanks a lot 22:12
pmurias your welcome 22:13
pugs_svnbot r20047 | pmurias++ | [kp6-mp6like] &push
diff: dev.pugscode.org/changeset/20047
lambdabot Title: Changeset 20047 - Pugs - Trac
lichtkind pmurias: no thats something different i ment every 3 numbers one dot 22:14
found it in the FAQ :) 22:17
pmurias lichtkind: i'm afraid you have to code it yourself
lichtkind: how do you do it?
i mean the sprintf code
lichtkind pmurias: in the FAQ where local $_ = shift; 1 while s/^([-+]?\d+)(\d{3})/$1,$2/; return $_; 22:18
which works perfectly
pmurias: i reduced it to 1 while s/^(\d+)(\d{3})/$1.$2/; 22:26
good night 22:27
22:27 lichtkind left
pmurias literal: g' night 22:28
22:28 pmurias left 22:31 yahooooo joined 22:39 ikeda joined, ikeda left 22:47 naxxtor left 22:49 ikeda joined 22:50 ikeda left 22:51 peepsalot left, peepsalot joined 22:56 peeps[work] left 22:59 [particle] joined
pugs_svnbot r20048 | buchetc++ | [INSTALL] workaround for ghc 6.6.1/cabal 1.1 22:59
diff: dev.pugscode.org/changeset/20048
lambdabot Title: Changeset 20048 - Pugs - Trac
23:03 ikeda joined 23:06 ikeda left
pugs_svnbot r20049 | buchetc++ | [t] S17 implementation checks 23:14
diff: dev.pugscode.org/changeset/20049
lambdabot Title: Changeset 20049 - Pugs - Trac
23:17 [particle1 left 23:18 buchetc left 23:19 nipotaway is now known as nipotan 23:21 ikeda joined, ikeda left 23:25 iblechbot_ left 23:29 innocuous left 23:32 ikeda joined 23:33 ikeda left 23:37 ikeda joined, ikeda left
pugs_svnbot r20050 | putter++ | misc/STD_red: work-in-progress sync'ing against STD.pm r20027. And adding more Match()s. No testing. 23:41
r20050 | putter++ | Regression: "use x;" no longer works above the statement rule (ie, in statementlist and UNIT).
diff: dev.pugscode.org/changeset/20050
lambdabot Title: Changeset 20050 - Pugs - Trac 23:42
23:44 jrockway joined
mncharity will try the 6.6.1 workaround later (doesn't work for me on 6.8.2 either). buchetc: thanks! 23:47
23:47 eternaleye joined
mncharity btw, that regression is, i'm confident, in my edit,not STD itself. 23:48
good night all &
meppl gute nacht
good night 23:49
23:51 meppl left 23:53 ikeda joined, ikeda left

Powered by Cro and the Raku® Programming Language.

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

Thank you!