Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2006-08-29

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:01 markstos This feels like a bug. I can't store and execute a code reference in an attribute:
00:01 markstos ?eval class Foo { has $.a is rw; method BUILD { $.a = sub {}; $.a(); } } Foo.new;
00:01 evalbot_12805 is now known as evalbot_12822
00:01 evalbot_12822 Error: No compatible subroutine found: "$.a"
00:05 pasteling "markstos" at 12.176.101.89 pasted "Bug? A method appears to be successfully called on a random string." (16 lines, 376B) at http://sial.org/pbot/19331
00:09 foo\ joined perl6
00:09 araujo hello
00:11 markstos Hello araujo
00:12 araujo hi there markstos :-)
00:18 ingy anyone know or guess whether I can "upgrade" debian unstable to ubuntu dapper?
00:19 markstos ingy: I think not. I don't think they are that compatible.
00:19 markstos But I'm mostly a Mandriva user, so get a second opinion.
00:25 ingy k
00:29 nekokak joined perl6
00:30 TreyHarris markstos:
00:30 * markstos listens
00:30 TreyHarris ?eval class Foo { has $.a is rw; method BUILD { $!a = sub { say "hi!" }; } }; my $a = Foo.new; $a.a.()
00:30 evalbot_12822 OUTPUT[hi! ] Bool::True
00:30 TreyHarris inside BUILD, your object isn't built yet
00:30 TreyHarris so you can't call methods
00:32 hikozaemon joined perl6
00:33 markstos TreyHarris: pugs lets you:
00:33 markstos ?eval class Foo { method BUILD { self.see; } method see { say "saw"; }; } my $a = Foo.new;
00:33 evalbot_12822 OUTPUT[saw ] \Foo.new()
00:35 TreyHarris you're not actually accessing the object in that case except for method dispatch (which only requires an actual instantiated class which has already happened by that time), so that's possible.  but what you want to do is chicken-and-egg, I think
00:36 TreyHarris but this works:
00:37 markstos I don't think so. Adding this line did allow the callback to run in BUILD: instead of call $.a(), I added: $aa = $.a, and then called $aa(), that worked.
00:37 TreyHarris ?eval class Foo { has $.a is rw; method BUILD { self.a = sub { say "hi!" }; } }; my $a = Foo.new; $a.a.()
00:37 evalbot_12822 OUTPUT[hi! ] Bool::True
00:37 TreyHarris so maybe it is a bug
00:37 TreyHarris but the Synopses do mention that inside BUILD you'll probably use $! rather than $. ....
00:37 markstos I tried $!, too. Same result.
00:38 TreyHarris eh?
00:38 TreyHarris ?eval class Foo { has $.a is rw; method BUILD { $!a = sub { say "hi!" }; } }; my $a = Foo.new; $a.a.()
00:38 evalbot_12822 OUTPUT[hi! ] Bool::True
00:38 TreyHarris you're not getting that?
00:40 agentzh joined perl6
00:40 markstos ?eval class Foo { has $!a is rw; method BUILD { $!a = sub { say "hi!" };  $!a() } };
00:40 evalbot_12822 undef
00:40 agentzh TreyHarris: why did you say that the synopses on feather was two weeks behind?
00:41 agentzh TreyHarris: the last update to the syn repos was r11527, made by audreyt according to my p6l mails. :)
00:41 TreyHarris hmmmm
00:42 agentzh TreyHarris: please correct me if i was wrong. :)
00:42 TreyHarris agentzh: one sec
00:44 agentzh the rev number of the syn repos is not increasing one by one.
00:44 agentzh because syn/ is a subdirectory per se.
00:47 mako132_ joined perl6
00:48 agentzh hence the rev number on the page of http://svn.perl.org/perl6/doc/trunk/design/syn/ is somewhat misleading. :)
00:48 lambdabot Title: Revision 11549: /doc/trunk/design/syn
00:49 TreyHarris ah, i see my problem
00:49 agentzh :)
00:49 TreyHarris i thought i saw a change to S03 in the svn logs... it was to S02
00:49 TreyHarris so when I searched S03 for it on feather.... i thought it was out of date
00:49 markstos It would be nice to have a more helpful error messages, when one commits this mistake, as Perl5ers like me are likely to do:
00:49 markstos ?eval class Foo {} Foo.new( 'a' => 'b' );
00:49 TreyHarris since it said "Aug 16"
00:49 evalbot_12822 Error: Must only use named arguments to new() constructor
00:50 TreyHarris that's not helpful?
00:50 TreyHarris ?eval class Foo {} Foo.new( a => 'b' );
00:50 evalbot_12822 Foo.new(("a" => "b"),)
00:51 markstos It would be more helpful to say: "Found quoted key in Pair passed to new(), expected unquoted key."
00:51 markstos When it says "Must used named arguments", I think: I /was/ using named arguments!
00:51 markstos In Perl5, Hash quoted and unquoted hash keys mean the same thing.
00:52 markstos s/Hash //
00:52 TreyHarris but you weren't.  you were using Perl 5 "named arguments".  Perl 6 actually has a first-class citizen called named arguments, and you weren't using one
00:53 f0rth joined perl6
00:53 markstos I know that now, but a more helpful error message would help new folks learn that faster, and with less frustration.
00:54 markstos Alternatively, something like "use warnings" for Perl5 would be nice, which gives more detail about each error, including common pitfalls.
00:54 agentzh markstos: pugs has been optimized for feature experimentation or verification. :)
00:55 agentzh audreyt: trypugs (or webpugs) is sort of beyond my capability (i'm still new to AJAX), so i'd return to the pugs test suite. :)
00:55 markstos That's fine. I'm providing feedback from a users perspective. If you want to more users, you'll have to eventually optimize for them, too!
00:55 agentzh markstos: indeed.
00:55 markstos Otherwise, it  may never be Christmas....
00:55 agentzh the problem is that we're having too few audreys here. :)
00:56 markstos I don't think so. I see a website that could be in better shape, docs that could written and clarified, and tests that could be written in Perl6. Those are are different skill sets, and you don't have to be a uber-hacker to contribute.
00:57 agentzh the christmas is still quite far from us, given the current status of the perl 6 language, the test suite, and the implementation.
00:59 markstos I don't know. As I port CGI::Application, pugs has been working rather well for me, in terms of correctness. For me, it's a bigger problem that I don't find the docs accessible to confirm if I have a problem with my code, or if there is a pugsbug. So far, I haven't found any show stopping bugs in pugs, mostly missing features.
00:59 agentzh markstos: what you're saying is all true. i'm talking about the tasks which require skills of "audrey level" anyway. :)
00:59 markstos Of course, it's particularly slow, but it's generally good enough for people to port real applications to it, at least in terms of the features available, even if the performance lacks now.
01:00 agentzh markstos: oh, you must be very lucky then!
01:00 TreyHarris agentzh: but let's not get too doom-and-gloom.  the progress in the past year has been incredible, and it's been accelerating, especially if you look at it quarter-to-quarter rather than day to day (when there are fits and starts)
01:00 markstos Users can become contributers and contributers can be become uberhackers.
01:00 markstos It's not too early to start paying attention to users.
01:00 agentzh TreyHarris: i was not complaining, or i wouldn't be here. :)
01:02 agentzh markstos: yes, i agree everything you've said.
01:02 markstos Something like the named parameters error message could well become an FAQ on the channel for new users, but a better error message could save everyone that headache.
01:02 agentzh markstos: i'm only telling the reality or something like that. :)
01:03 agentzh what i want to say is that we just need to be patient enough. :)
01:04 agentzh markstos: btw, i'll be very happy if you add that warning message to pugs' src/ :)
01:04 agentzh or get audreyt or someone to do that. :)
01:05 markstos agentzh: Maybe I'll poke around. I have limited Haskell skills, but I think I'm equipped to update a string...
01:05 agentzh good luck then. ;-)
01:08 markstos Uh oh. Tornado warning here. I'll have to check the wifi in the basement...
01:12 xdg joined perl6
01:17 * markstos notes there /is/ wifi in the basement, and hunkers down for the storm to pass.
01:21 SamB markstos: but... if the storm hits, will there still be?
01:22 markstos If the storm hits, my 24/7 hosting facility downtown will be a bigger concern....
01:22 markstos (It's my business, not just where I store files!)
01:26 svnbot6 r12823 | agentz++ | [TASKS]
01:26 svnbot6 r12823 | agentz++ | * added a new task that we need a new smoke client to
01:26 svnbot6 r12823 | agentz++ |   upload tests.yml as well as the HTML matrix, and a new
01:26 svnbot6 r12823 | agentz++ |   smoke server which makes .yml files accessible to all
01:26 svnbot6 r12823 | agentz++ |   (especially to feather). :)
01:26 b00t joined perl6
01:26 * markstos notes the electricity blinks for the first time...
01:26 agentzh markstos: ohhh
01:27 markstos agentz: Are you aware of Smolder ? It's a smoke server written in Perl with YAML support.
01:27 dolmans joined perl6
01:27 agentzh aye ;-)
01:29 agentzh once we have .yml on the smoke server, feather can use the latest one to render the synopsis HTML. hence no need to do autosmoking ourselves. :)
01:35 coumbes joined perl6
01:35 xdg If I want to test the Perl 6 CPAN Stack on win32, what modules should I be installing? just v6.pm?  Or others?
01:36 markstos Install v6, it will hopefully handle the dependencies for you.
01:37 xdg OK.  I'm manually smoking Strawberry Perl alpha 2 prerelease
01:37 xdg So I'm trying to test various strange things that might cause problems.
01:37 markstos Ok. I'm listening to hail through the heating ducts.
01:38 xdg That's funky.  (Remind me where you are again?)
01:38 diakopter joined perl6
01:39 markstos I'm in Richmond, Indiana, taking shelter in my basement while a tornado passes.
01:39 b00t joined perl6
01:39 diakopter hi markstos
01:39 markstos Hello diakopter.
01:39 svnbot6 r12824 | markstos++ | Attempt a more helpful error message for "Must only pass named parameters..."
01:39 svnbot6 r12824 | markstos++ | (Sorry about the long line. My Haskell-fu is weak. )
01:40 xdg We had a tornado watch in NYC last week.  Rained cats and dogs, but no hail (and no spouts, either)
01:40 diakopter anyone going to the Chicago hackathon in November?
01:40 markstos I lived in Manhattan for a college semester. I enjoyed my time there
01:41 agentzh left perl6
01:41 markstos diakopter: I didn't know about, but I'm relatively close by. ( < 5hrs by car ).
01:41 markstos What's the focus ?
01:41 diakopter http://hackathon.info/
01:41 lambdabot Title: Chicago Perl Hackathon: Nov 10-12, 2006, Crystal Lake, IL
01:43 khisanth_ joined perl6
01:49 markstos Here's another error message that could use potential improvement. For multi methods, you can get the message: "No compatible subroutine found". I think it should be split into two messages: 1. (same as above). and 2. "Found 3 potential subroutines, but none with compatible signatures".   That gives you more feedback about exactly where the problem is.
01:51 TreyHarris markstos: i agree, but i'm not sure of the prioritization here.  i've generally kept quiet about confusing error messages, because i'd rather the new object model and captures get finished and stabilized first.  i think getting good error reporting could be a tremendous timesink
01:52 TreyHarris i just pretend that any error is like BASIC's "syntax error", and if it tells me more than that, that's just gravy.
01:54 markstos OTOH, I just commited an improvement to an error message. They are often string changes and light Haskell work, not the domain of an uberhacker.
01:55 markstos I used to improve some of the darcs strings, even though I only learned enough Haskell to do "string\n"++"string" and "make"
01:55 markstos The one I suggested above does involve a little logic change in Haskell, which I'm not ready to dive into yet.
01:58 markstos A good question is: What's the most direct route to the tipping point: When a tremendous amount of energy forms around it, and there are lots of tuits? Is focusing on features the shortest route, or is polishing a smaller feature set a shorter route?  I don't have answers, but I know that sometimes answers are surprising. Sometimes little things like string changes can make a big difference, because they turn frustrated users into interested us
02:11 b00t joined perl6
02:17 coumbes joined perl6
02:19 ajs_home Does anyone know off the top of their heads (without dragging the whole channel OT), what the most efficient YAML parser is out there? I just crushed the performance of a CGI program by converting its data to YAML... :(
02:23 xdg I think it's YAML::Syck
02:23 ajs_home Thanks... yeah, I just spotted that. Checking it out now
02:23 TreyHarris markstos: well, i think the whole pugs project showed that.  initially there were people who were rather hostile to energy being drawn away from parrot
02:24 * TreyHarris must decommute now &
02:25 markstos ...but know I would say that interest in pugs probably interest in parrot...
02:25 xdg For anyone working on v6, it looks like somewhere in the chain there is a dependency on Cache::Cache, which depends on Error, and the latest version of that hangs during testing on Win32.  Not fun.
02:25 ajs_home Every time I see that nomenclature of "foo &" I'm reminded how how simple and powerful the Unix shell was....
02:26 ajs_home xdg: there is nothing that I think has done more to errode the Perl user-base than the fact that tests rot.
02:26 TreyHarris ajs_home: was?  still is!  i'm still typing here in the background, even though the other me is busy decommuting ;-)
02:26 ajs_home heh
02:27 markstos Your joking.
02:27 * TreyHarris &'s for real this time
02:27 markstos We know you are really giving voice commands to your treo while driving.
02:29 ajs_home markstos: If you are talking to me, no I'm serious. I've never seen such anger at a programming language as when an end user is told to type "sudo cpan Foo::Bar" and they get 20 test failures in their fact
02:29 ajs_home er face
02:33 mjk joined perl6
02:36 markstos ajs_home: I wasn't talking to you, but maybe you would be happy to know I removed tests against Google from the WWW::Mechanize suite. They kept breaking because Google changed their pages...
02:37 ajs_home hah! That's got to have made google happy ;)
02:38 markstos My personal favorite failing CPAN test that I've written was one that I (accidently) constructed such that it passed  for me, but when people tested it from other time zones, it failed...
02:38 markstos what a pain to fix.
02:38 ajs_home ouch
02:39 ajs_home I love the idea of having tests run, but in practice it seems like the pain gets shifted to the people who have the least time and ability to track down the problem in most cases.
02:39 ajs_home I've tried to think of a way to solve that for Perl 6.... nothing yet
02:40 markstos I always thought tests were good for exposing bugs triggered by different environments.
02:41 ajs_home They're great for that... but if the trade-off is user pain, that's not always a trade-off you want. There must be a middle-ground
02:45 markstos beta testeres.
02:45 markstos s/e//g
02:46 markstos Also,
02:46 markstos interns.
02:48 ajs_home Good ideas... ideas that cost $....
02:48 markstos Well, there's an automated build farm, but that costs money.
02:49 markstos There's having every environment in your home office, and testing it yourself, but that costs money.
02:49 ajs_home There's throwing money at it, but... well.
03:05 crem joined perl6
03:11 norageek2 joined perl6
03:16 rafl joined perl6
03:18 markstos It would be nice if "Parameters and Arguments" quickly mentioned the difference between parameters and arguments.
03:18 markstos http://feather.perl6.nl/syn/S06​.html#Parameters_and_arguments
03:18 lambdabot Title: S06
04:15 coumbes joined perl6
04:18 audreyt markstos: send patch to p6l?
04:18 * audreyt gets ready for $job (horribly late)
04:20 drbean joined perl6
04:24 xdg joined perl6
04:25 buubot joined perl6
04:26 jferrero joined perl6
04:32 buubot joined perl6
04:35 buu joined perl6
04:41 justatheory joined perl6
05:06 Aankhen`` joined perl6
05:07 onsen joined perl6
05:10 Yobert joined perl6
05:26 svnbot6 r12825 | agentz++ | [t/syntax/comments.t]
05:26 svnbot6 r12825 | agentz++ | - added some more tests. all tests now pass.
05:27 agentzh joined perl6
05:28 agentzh thanks to audreyt++'s recent utf-8 fixes. :)
05:28 xinming_ is now known as xinming
05:30 agentzh many new tests for pugs were lost in the hardware failure this noon. :(
05:30 agentzh now i'm rewriting them. :-/
05:30 agentzh that old nasty machine is really HATEFUL!!!
05:31 christo joined perl6
05:35 Yobert Hey do you know what would be cool in perl 6
05:35 Yobert A special variable for when you do a for (@array) style loop
05:35 Yobert it would always have the index of the array
05:36 araujo ?
05:36 araujo you mean the current iteration index?
05:36 DHGE joined perl6
05:40 gaal ?eval my @array = <moose elk caribou>; for @array.kv -> $i, $val { say "$i. $val" }
05:40 evalbot_12822 is now known as evalbot_12825
05:40 evalbot_12825 OUTPUT[0. moose 1. elk 2. caribou ] undef
05:40 gaal Yobert: good enough? :)
05:41 masak :)
05:41 masak ask, and ye shall find perl6 probably already has it
05:44 buu Bah, way too clumsy.
05:45 * araujo likes it
05:46 onsen_ joined perl6
05:46 gaal buu: what don't you like about it, the need for an explicit variable? for an explicit method call on @array?
05:46 buu gaal: Both.
05:46 agentzh has %=DATA been specified?
05:46 svnbot6 r12826 | agentz++ | [t/syntax/comments.t]
05:46 svnbot6 r12826 | agentz++ | - added tests for "=begin comment/=end comment" from
05:46 svnbot6 r12826 | agentz++ |   t/syntax/POD.t
05:46 svnbot6 r12826 | agentz++ | - added tests for "=for comment"
05:46 agentzh i saw $=DATA and @=DATA in S02, but not %=DATA.
05:47 gaal okay then: { my $i; for @arr -> $k { NEXT { $i++ } ... } }.()
05:48 gaal I don't think an implicit "current iterant index" variable makes sense. how would you use it in nested loops?
05:49 kanru joined perl6
05:49 masak if I wanted, could I throw an extra method each_with_index into Array at runtime?
05:49 masak I think that would be the cleanest way in that case
05:49 buu gaal: It would always be the innermost loop
05:50 buu gaal: Or whatever. Honestly it can do whatever as long as it exists.
05:51 gaal well, the point of having something like this is that you can mention the index if you suddenly remember you need it without refactoring your code.
05:51 Yobert yeah, exactly
05:51 Yobert I just want a quickie way
05:51 gaal but if it's always the innermost, or "whatever", then you can hardly do that
05:52 buu gaal: Sure you can
05:52 Yobert it could have local scope so it'd work in nested lops
05:52 Yobert *loops
05:52 gaal buu: well, by all means, propose something on p6-l
05:52 buu 95% of the time it would be in the right scope anyway
05:52 buu gaal: They're scary =[
05:53 masak buu: if you don't, I will :)
05:53 buu Go for it.
05:53 masak 'kay
05:53 buu Magic special variable, scoped inside the loop.
05:53 masak np
05:53 * masak goes off to write to p6-l
05:54 gaal masak: things to consider:
05:54 svnbot6 r12827 | agentz++ | [t/syntax/comments.t]
05:54 svnbot6 r12827 | agentz++ | - added even more tests and smartlinks.
05:54 gaal - what if you're iterating over something that isn't an array? still set the index?
05:54 Yobert well
05:54 Yobert for a file glob
05:54 Yobert it could be the byte position
05:54 gaal - does 'next' skip an index or not? (surely on arrays it should, not clear in other cases)
05:55 Yobert I'm sure there are a zillion perly things you could do
05:55 gaal byte position? of what?
05:55 masak gaal: I'll throw your questions in verbatim
05:55 gaal masak: go ahead. I'm not convinced this is a good idea though :)
05:55 Yobert well, i guess something like STDIN doesn't have a byte position really.. or does it.  I was thinking in a while (<STDIN>) loop you could have the byte position or something
05:55 buu gaal: It doesn't *really* matter either way
05:56 buu gaal: I'm fine with every iteration increases
05:56 gaal buu: I don't understand what really matters.
05:56 Yobert sorry for my perl5 examples I'm not confident yet, haha
05:56 gaal Yobert: ah, that's not a glob :)
05:56 buu gaal: That it works for array/list iterations =]
05:56 Yobert sorry.. just a filehandle in general i meant
05:57 gaal Yobert: sure, there's a "tell" in Perl 5, and in Perl 6 you'll presumably have both the OS byte position and the character position.
05:58 gaal buu: that's what I'm not convinced about. you have a specific problem you want to make prettier, and you don't care about the rest
05:58 gaal any program can be compressed to 1 byte with a certain compression algorithm
05:58 gaal we're not aiming for that though :)
05:58 buu gaal: I care, but in limited ways.
05:58 Yobert gaal:  Ahh.. I see.  I've never used tell, that's probably my problem ;)
05:58 buu gaal: My point was more than either alternative suggestion was perfectly workable.
05:58 agentzh how can i read a single line from a Perl 6 file handle? my $line = =$fh?
05:59 gaal Yobert: well, tell is a system call, so it's not for free.
05:59 gaal agentzh: yes
05:59 gaal brb
05:59 Yobert gaal:  I see.  So it would still be cool to have a builtin index
05:59 agentzh gaal: thanks
06:03 agentzh ?eval my $fh = undef; my $line = =$fh; $line
06:03 gaal Yobert: presumably $fh.position :chars or something shorter queries Perl's notion
06:04 agentzh gaal: the above line hangs in pugs. :(
06:04 gaal hmm.
06:04 agentzh gaal: could you make it return undef immediately?
06:04 gaal looking
06:04 gaal shouldn't though. should fail""
06:05 agentzh :)
06:05 agentzh okay, fail is also fine. :)
06:05 agentzh *failing
06:05 * agentzh hates segfaulting.
06:06 Yobert left perl6
06:06 agentzh ?eval 1+2
06:06 gaal why does AUTHORS have the BOM again?
06:06 agentzh oh, i've killed evalbot. :(
06:06 agentzh BOM?
06:07 gaal http://en.wikipedia.org/wiki/Byte_Order_Mark
06:08 agentzh k
06:14 marmic joined perl6
06:15 agentzh what should the elems of @=DATA be?
06:16 agentzh each of @=DATA should be a single line or a single DATA block?
06:16 agentzh and should the elems contain leading or trailing newlines?
06:19 pasteling "masak" at 130.238.83.176 pasted "Ok, so here's what I've come up with. Last call for changes. :)" (41 lines, 1.7K) at http://sial.org/pbot/19336
06:21 svnbot6 r12828 | agentz++ | [t/syntax/POD.t]
06:21 svnbot6 r12828 | agentz++ | - added new tests for %=POD{'DATA'}, $=DATA, and @=DATA
06:21 svnbot6 r12828 | agentz++ |   (all tests are failing now)
06:22 masak buu: does the email capture your wishes?
06:22 svnbot6 r12829 | agentz++ | [t/syntax/POD.t]
06:22 svnbot6 r12829 | agentz++ | - fixed broken smartlinks
06:23 agentzh xinming: would you please merge the tests in t/builtins/quote_basic.t and t/operators/quoting.t?
06:23 gaal ah, agentzh, it's not hanging, it's waiting on $*ARGS -- presumably stdin
06:23 agentzh xinming: i think they had better be put in the same place (or same directory at least)
06:24 agentzh gaal: ohh
06:24 masak gaal: sounds like a situation that would arise quite often
06:24 agentzh gaal: so this is expected?
06:24 masak gaal: isn't there a way to always feed $*ARGS to evalbot?
06:24 agentzh the file handle is undef anyway...
06:25 mauke_ joined perl6
06:25 gaal masak: well, it's like what you get when you put on the command line
06:25 gaal % grep moose
06:25 masak anyway, anybody has any last comments on the p6-l mail? http://sial.org/pbot/19336
06:25 lambdabot Title: Paste #19336 from "masak" at 130.238.83.176
06:25 masak in other case, I'll just send it along
06:25 gaal looking
06:26 masak gaal: yes, but evalbot never gets any input from anywhere
06:26 masak gaal: making it wait indefinitely does nobody any good
06:26 masak gaal: better to have some "standard" standard input stream for it
06:26 agentzh i kinda dislike the t/builtins/ directory since too many unrelated tests are living there. :-/
06:26 gaal no, it oughta be caught in the evalbot harness, not hacked into pugs
06:27 masak nod
06:27 onsen joined perl6
06:27 masak 'course
06:28 gaal masak: I don't get the each_with index thing. How is it different from .kv?
06:28 masak gaal: it isn't, really
06:29 gaal (except that it reverses the order? and is longer? :)
06:29 masak maybe it's a bad example
06:29 masak I just wanted to show that ruby has thought of this somehow
06:29 gaal anyone happy with the ruby solution would surely be fine with the Perl 6 spelling?
06:29 masak yes... except...
06:29 gaal uh, it's semantically precisely the same as .kv
06:30 masak they have a _method_ for it
06:30 gaal .kv is a method
06:30 masak which makes the purpose so much clearer
06:30 masak .kv doesn't really do it for me
06:30 gaal ?
06:30 masak it doesn't say "index"
06:30 masak it says "key"
06:31 masak maybe it'll grow on me, but right now it feels every so slightly hackish
06:31 gaal er, so you're just unhappy with the name, not the functionality?
06:31 masak kinda
06:31 masak but I'll go back and review the ruby example
06:31 svnbot6 r12830 | agentz++ | [t/operators/quoting.t]
06:31 svnbot6 r12830 | agentz++ | - fixed one smartlink.
06:31 gaal so you're not certain, like buu is, that there needs to be an implicit magical?
06:32 masak no, not at all
06:32 buu I'm all about the implicit magic.
06:32 masak the more I look at the .kv example, the more I think "this is quite ok"
06:32 gaal "use DWIM"
06:33 * masak sends the mail now
06:35 masak &
06:35 buu Bah. I hate the .kv example.
06:36 buu It's long and useless
06:39 mauke_ is now known as mauke
06:39 clkao env is now temp right?
06:40 gaal no, context
06:41 gaal temp is what was 'local'
06:45 clkao so env is now context?
06:45 clkao i saw some stuff in tests with env
06:46 gaal clkao: I think that's wrong, but better check the synopses
06:46 clkao i can't grep out anything about context variables ecept $+foo
06:47 gaal that's the mention, but there's something about declaration. hold on
06:49 gaal L<S09/Names/"The C<CONTEXT> pseudo-package">
06:50 reZo joined perl6
07:04 araujo joined perl6
07:04 svnbot6 r12831 | agentz++ | [t/statements]
07:04 svnbot6 r12831 | agentz++ | - added repeat.t with the tests from t/statements/loop.t.
07:04 svnbot6 r12831 | agentz++ | - added several smartlinks.
07:04 svnbot6 r12832 | agentz++ | [t/statements/loop.t]
07:04 svnbot6 r12832 | agentz++ | - removed tests specific to the repeat statment. (they
07:04 svnbot6 r12832 | agentz++ |   have gone into t/statements/repeat.t.
07:07 rashakil_ joined perl6
07:07 svnbot6 r12833 | agentz++ | [t/statements/loop.t]
07:07 svnbot6 r12833 | agentz++ | - fixed the ``plan'' number.
07:22 agentzh bbiab
07:22 svnbot6 r12834 | agentz++ | [t/operators/quoting.t]
07:22 svnbot6 r12834 | agentz++ | - added many tests for unicode-delimited q-style strings
07:22 agentzh left perl6
07:29 kane-xs joined perl6
07:40 ofer0 joined perl6
08:07 svnbot6 r12835 | yiyihu++ |  r12866@Laptop:  xinming | 2006-08-29 15:45:57 +0000
08:07 svnbot6 r12835 | yiyihu++ |  agentz++ pointed it t/builtins/quote_basic.t
08:07 svnbot6 r12835 | yiyihu++ |  and t/operators/quoting.t should merge.
08:07 svnbot6 r12835 | yiyihu++ |  r12878@Laptop:  xinming | 2006-08-29 16:01:18 +0000
08:07 svnbot6 r12835 | yiyihu++ |  agentz++ pointed out that t/operators/quoting.t and t/operators/quote_basic.t
08:07 svnbot6 r12835 | yiyihu++ |  should merge.
08:07 xinming :-/
08:08 * xinming thinks, svk sucks sometimes
08:11 wolverian all software sucks. it's just a matter of sucking less.
08:35 masak xinming: anything in particular?
08:35 masak there's always ways to improve things
08:41 penk joined perl6
08:47 gaal hmm, I wonder if agentzh's bug report should in fact be fixed in pugs. to wit:
08:47 gaal $ perl -e 'my $x; print readline $x'
08:47 gaal # immediate nothing
08:48 gaal the kink is that readline is currently defined (?IO) in Prim, so it has no way of distinguishing an actual undef arg and a missing arg
08:49 gaal the fix to that is to split to two readlines: op0 "readline" and op1 "readline" of sig (IO)
08:49 gaal this is so that the legitimate "say item readline" still works (ie reads from stdin)
08:50 gaal if Prim were in newval, perhaps that could have been made nicer by passing a UFailure "missing param" instead of an unsidcriminating VUndef.
08:50 gaal I gotta moose for a while, so if anyone wants an easy Haskell task, go for it..
08:53 lollan joined perl6
08:55 lollan joined perl6
08:57 glasser joined perl6
08:59 glasser joined perl6
09:04 nothingmuch joined perl6
09:11 plisk joined perl6
09:17 elmex joined perl6
09:27 b_jonas joined perl6
09:28 Fuzie joined perl6
09:31 PolettiX joined perl6
09:41 ToTimkopf joined perl6
09:48 xinming_ joined perl6
09:54 orafu joined perl6
09:56 onsen joined perl6
10:20 buetow joined perl6
10:26 penk joined perl6
10:39 agentzh joined perl6
10:39 agentzh xinming++ # unified two quoting-related tests.
10:40 agentzh well, i mean test files. :)
10:44 svnbot6 r12836 | agentz++ | [t/operators/quoting.t]
10:44 svnbot6 r12836 | agentz++ | - fixed a smartlink.
10:48 agentzh i think we had better rename t/subroutines/ to t/blocks/. anyway, blocks are much general then subroutines in Perl 6's termilogy.
10:49 agentzh you know, we have pointy.t and closure.t under t/subroutines, which are not subroutines.
10:50 svnbot6 r12837 | bsmith++ | Added a ranlib step after embedding the judy libraries into HsJudy.
10:53 svnbot6 r12838 | agentz++ | - moved return.t and return_from_inner_block.t from
10:53 svnbot6 r12838 | agentz++ |   t/statements/ to t/subroutines/.
10:54 gaal the genesis for names under t/ is the Perl 5 test suite. but since this is indeed a change from Perl 5 changing it looks reasonable to me.
10:59 agentzh gaal: okay. :)
10:59 svnbot6 r12839 | agentz++ | [t/no_implicit_block.t]
10:59 svnbot6 r12839 | agentz++ | - added this new test file to test the nonexistence of
10:59 svnbot6 r12839 | agentz++ |   implicit blocks around standard control structures.
10:59 svnbot6 r12839 | agentz++ |   currently only if/given/while are tested. more tests
10:59 svnbot6 r12839 | agentz++ |   soon. :)
10:59 svnbot6 r12838 | agentz++ | - moved return.t and return_from_inner_block.t from
10:59 svnbot6 r12838 | agentz++ |   t/statements/ to t/subroutines/.
11:01 gaal agentzh: but I'd be happy to see more discussion on this before svn mv.
11:02 gaal p6-c it?
11:02 agentzh gaal: i'm just about to commit. ;-)
11:02 gaal okay, that works too :)
11:02 nothingmuch gaal: http://code2.0beta.co.uk/m​oose/svn/notes/nuffinmop/
11:02 lambdabot Title: Revision 1855: /notes/nuffinmop
11:03 gaal ooh
11:03 agentzh nothingmuch: are you agree to rename t/subroutines to t/blocks?
11:03 nothingmuch i don't remember what's in t/subroutines
11:03 nothingmuch but in general since every block is a closure... *shrug*
11:03 agentzh hehe
11:03 nothingmuch what's the rationale?
11:04 agentzh we have pointy.t and closure.t and etc etc under t/subroutines. :)
11:04 nothingmuch ah
11:04 nothingmuch heh
11:04 agentzh pointy blocks are not subrountes.
11:04 nothingmuch sure, sounds fair
11:04 agentzh okay :)
11:14 svnbot6 r12840 | agentz++ | - renamed t/subroutines to t/blocks with permission from
11:14 svnbot6 r12840 | agentz++ |   gaal++ and nothingmuch++.
11:15 agentzh oh, i just see S06 is titled "subroutines".
11:15 agentzh following this reasoning, i should revert r12840. :)
11:15 agentzh any idea?
11:16 nothingmuch agentzh: or submit a doc patch =)
11:17 agentzh oh, is there a better name for that?
11:17 nothingmuch blocks?
11:17 agentzh S04 is already named "blocks and statements". :(
11:19 nothingmuch mmm
11:19 nothingmuch well, the tests do cover s04 and s06 simultaneously
11:20 agentzh oh
11:20 agentzh but most of them are for S06.
11:21 agentzh okay, i'll forget the name of S06 and use t/blocks.
11:21 agentzh if anyone else feels t/subroutines is better, feel free to rename back. :)
11:21 diakopter joined perl6
11:22 agentzh (but don't svn revert, since i'm going to commit updates to t/blocks)
11:24 xinming joined perl6
11:26 nothingmuch agentzh++
11:27 agentzh :)
11:27 * nothingmuch tries to get the compile time object protocol down on paper
11:27 nothingmuch what happenned to jabbot, btw?
11:29 svnbot6 r12841 | agentz++ | [t/statements/while.t]
11:29 svnbot6 r12841 | agentz++ | - seriously reviewed all the tests and did some code
11:29 svnbot6 r12841 | agentz++ |   cleanup.
11:29 svnbot6 r12841 | agentz++ | - more smartlinks.
11:35 svnbot6 r12842 | agentz++ | [t/statements/no_implicit_block.t]
11:35 svnbot6 r12842 | agentz++ | - added tests for the unless/for/loop control structures.
11:41 b00t joined perl6
11:44 svnbot6 r12843 | agentz++ | [docs/feather/syn_index.html]
11:44 svnbot6 r12843 | agentz++ | - made the synopses much more remarkable.
11:44 svnbot6 r12844 | agentz++ | [t/README]
11:44 svnbot6 r12844 | agentz++ | - now that we have almost all the features of
11:44 svnbot6 r12844 | agentz++ |   util/catalog_tests.pl implemented in
11:44 svnbot6 r12844 | agentz++ |   util/smartlinks.pl, we now claim smartlinks.pl as
11:45 svnbot6 r12844 | agentz++ |   the main cross-referencer. :)
11:47 * agentzh is happily kicking off items from his todo list.
11:47 svnbot6 r12845 | agentz++ | [util/catalog_tests.pl]
11:47 svnbot6 r12845 | agentz++ | - added note for util/smartlinks.pl.
11:47 svnbot6 r12845 | agentz++ | - fixed a typo in the source.
12:00 svnbot6 r12846 | agentz++ | [t/statements/unless.t]
12:00 svnbot6 r12846 | agentz++ | - reviewed all the tests
12:00 svnbot6 r12846 | agentz++ | - tweaked the code to conform with modern Perl 6
12:00 svnbot6 r12846 | agentz++ | - added more tests
12:00 svnbot6 r12846 | agentz++ | - added smartlinks.
12:00 svnbot6 r12847 | agentz++ | [t/statements/unless.t]
12:00 svnbot6 r12847 | agentz++ | - adjusted the plan number.
12:05 nothingmuch agentzh++ # not that you need any more ;-)
12:06 agentzh nothingmuch: (?)
12:06 bpphillips joined perl6
12:08 nothingmuch agentzh: karma
12:08 svnbot6 r12848 | agentz++ | [t/statements/until.t]
12:08 svnbot6 r12848 | agentz++ | - reviewd all the tests.
12:08 * nothingmuch goes to sweat a bit
12:08 svnbot6 r12848 | agentz++ | - fixed the lexical problem.
12:08 svnbot6 r12848 | agentz++ | - added a smartlink.
12:08 agentzh ah, hehe :)
12:09 * agentzh collects all the ++'s for him.
12:14 svnbot6 r12849 | agentz++ | [t/statements/while.t]
12:14 svnbot6 r12849 | agentz++ | - changed the smartlink to point to a more appropriate
12:14 svnbot6 r12849 | agentz++ |   position in the synopses.
12:26 Limbic_Region joined perl6
12:26 audreyt changes topic to: Check your feather email | spec.pugscode.org | paste: http://sial.org/pbot/perl6 | pugs.blogs.com |  http://www.treehugger.com/f​iles/th_images/paradigm.jpg
12:27 * audreyt goes back from $job and saw huge number of checkins
12:27 audreyt agentzh++
12:27 agentzh :)
12:28 audreyt unfortunately, I need to crash now... very long day today with lots of Haskell hacking
12:28 Limbic_Region sleep well
12:29 audreyt it's really a good tool to deal with rapidly shifting specs ;)
12:29 Limbic_Region hope you got my note that embedded AS Perl works
12:29 audreyt I did. yay!
12:29 audreyt I wonder if we can turn on embp5  by default now
12:29 audreyt but I'll sleep on it
12:31 Limbic_Region well, I have it set as a perm env var so I can certainly let you know when it is b0rk now
12:31 audreyt cool :) ciao!
12:32 agentzh night, audreyt. :)
12:32 svnbot6 r12850 | agentz++ | [t/statements/statement_modifiers.t]
12:32 svnbot6 r12850 | agentz++ | - reviewed all the tests
12:32 svnbot6 r12850 | agentz++ | - simplified most of them
12:32 svnbot6 r12850 | agentz++ | - fixed the smartlink
12:42 agentzh oh, pugs doesn't support postfix given.
12:42 jferrero joined perl6
12:42 agentzh gaal: any quick guide to add support to that under src/?
12:42 agentzh (if such quick guide exists)
12:42 Limbic_Region agentzh - if there isn't an existing failing test case (which I assume there isn't since you would be linking to it), then please write one
12:42 agentzh ;-)
12:42 agentzh LR: i've already written one. :)
12:43 Limbic_Region well, if this client didn't eat plus symbols I would agentz plus plus you
12:43 agentzh hehe
12:43 agentzh thanks
12:44 agentzh i've been reading tests in the whole afternoon, so i can quickly commit many updates tonight. :)
12:45 audreyt where is postfix given specced?
12:45 agentzh audreyt: S04
12:46 agentzh # L<S04/The C<for> statement/"given" "use a private instance of" $_>
12:46 agentzh audreyt: but i think you should be sleeping. ;-)
12:46 audreyt indeed. implemented
12:46 agentzh audreyt++
12:48 agentzh ohhhh, smartlinks.pl is really really helpful while writing tests for pugs. :)
12:48 Limbic_Region well, it is really helpful to the extent it can let us know where we don't have tests (and possibly implementation) for that which is specced
12:49 Limbic_Region and that is really really helpful
12:49 agentzh aye
12:49 Limbic_Region but too bad it didn't have some type of mind probe where it could be pointed at @larry to discover what is not-yet specced
12:49 agentzh hehe.
12:50 agentzh LR: oh, no problem, that is *my* task.
12:52 Limbic_Region agentzh - at the risk of bringing her back to the channel - what prompted all the haskell hacking today?  I didn't see a lot going on the list so I presume it wasn't related to changing specs
12:52 svnbot6 r12851 | agentz++ | [t/statements/statement_modifiers.t]
12:52 svnbot6 r12851 | agentz++ | - added tests for use of private instance of $_ when
12:52 svnbot6 r12851 | agentz++ |   postfix for and given are used.
12:52 svnbot6 r12851 | agentz++ | - more smartlinks.
12:52 Limbic_Region and yet she implied rapidly changing specs were a good thing
12:53 agentzh LR: no clue. :(
12:53 agentzh LR: maybe $job?
12:54 agentzh i think she said Haskell is a good thing to rapidly changing specs.
12:54 audreyt $job
12:54 agentzh not the specs themselves.
12:54 audreyt hacking on 面了?爸媽這邊也有許多æ​œ‰è¶£çš„朋友,歡迎妳參加共同的活動。有時間我們父女兩人一齊用一次午餐,好å¥​½äº«å—深談的樂趣,如何?
12:54 audreyt Vater
12:54 audreyt oops!
12:54 mauke AIIRRR
12:54 agentzh cd audreyt && make realclean
12:54 audreyt wow, mispaste mail from my data ;)
12:54 * audreyt has to remember it's not OSX anymore
12:55 audreyt anyway. it's $job, hacking on http://svn.openfoundry.org/openafp/
12:55 lambdabot Title: Revision 31: /
12:55 mauke weirdly enough I can see all those characters
12:55 agentzh oh, i can't understand those characters even if i speak chinese.
12:56 agentzh mauke: can you grok them?
12:56 mauke hah, no
12:56 * agentzh grins.
12:56 mauke win xp <- putty <- linux <- screen <- irssi, and it just works. I think that's impressive.
12:57 agentzh indeed!
12:57 * audreyt mumbles something about X and Gtk having two distinct clipboards.
12:57 agentzh but i'm using the path freenode <- waysonline proxy <- mIRC <- win xp
12:57 audreyt anyway, postfix given is in.
12:58 agentzh audreyt++++++
12:58 svnbot6 r12852 | audreyt++ | * Pugs.Eval/Pugs.Parser: Implement postfix "given" as
12:58 svnbot6 r12852 | audreyt++ |   requested by agentzh++.
12:59 agentzh (brb)
12:59 audreyt (bbl, g'nite)
13:01 mago joined perl6
13:04 wolverian spec.pugscode.org doesn't respond
13:05 agentzh spleep time already &
13:05 agentzh left perl6
13:05 fglock joined perl6
13:07 svnbot6 r12853 | agentz++ | [t/statements/statement_modifiers.t]
13:07 svnbot6 r12853 | agentz++ | - fixed a failing test.
13:07 svnbot6 r12853 | agentz++ | - with Audrey' last commit, all tests here are
13:07 svnbot6 r12853 | agentz++ |   successful!
13:15 cjeris joined perl6
13:33 vel joined perl6
13:40 svnbot6 r12854 | fglock++ | * v6 - Pugs::Emitter::Perl6::Perl5::Perl5Hash - (experimental)
13:40 svnbot6 r12854 | fglock++ |   a compile-time object that represents an unboxed hash;
13:40 svnbot6 r12854 | fglock++ |   attempting to reduce the combinatorial complexity in the emitter
14:02 mago ping fglock
14:02 eden_c joined perl6
14:02 fglock mago: pong
14:03 mago fglock: /j #perl.br
14:03 fglock ok
14:06 eden_c left perl6
14:06 hexmode joined perl6
14:22 svnbot6 r12855 | fglock++ | - misc/pX/fglock/Perl5.pm - optimized emitter; ' %hash5.isa("X") ' compiles to ' "X" eq "Hash" '
14:25 weinig|bbl is now known as weinig
14:26 DaGo joined perl6
14:28 rashakil_ joined perl6
14:32 svnbot6 r12856 | fglock++ | - mv Emitter/Perl6/Perl5 to misc/pX/fglock/Pugs-Compiler-Perl6/lib...
14:33 prefiks joined perl6
14:38 svnbot6 r12857 | fglock++ | * v6 - added Emitter/Perl6/Perl5/Perl5Array.pm; Perl5Scalar.pm
14:39 onsen joined perl6
14:48 Limbic_Region anyone know how to run a specific test only under the smoke harness?
14:48 Limbic_Region I ask because something that passes from the command line fails under the harness
14:49 Limbic_Region nevermind, found it from the Makefile.PL
15:05 Limbic_Region bah, that didn't work - oh well
15:15 soisoisoi joined perl6
15:25 plisk joined perl6
15:25 kolibrie Limbic_Region: does util/prove6 do what you want?
15:26 kolibrie or you want to make smoke <some way to specify which test>?
15:27 Limbic_Region kolibrie - I don't know about util/prove6
15:27 Limbic_Region what I can say is this
15:27 Limbic_Region pugs -Iblib6/lib t/examples/all_parse.t
15:27 Limbic_Region works
15:28 Limbic_Region but the test fails completely under the harness
15:28 Limbic_Region so I want to see why
15:29 Limbic_Region I have to go to lunch for a meeting in a minute but if you could figure out how to test that under the harness or verify prove6 does the same thing - I would appreciate it
15:30 xerox joined perl6
15:31 justatheory joined perl6
15:36 kanru joined perl6
15:37 gaal Limbic_Region: export HARNESS_PERL=./pugs ; export PERL6LIB=blib6/lib ; prove t/some/test.t
15:38 xinming joined perl6
15:38 gaal or in a pinch, $EDITOR util/yaml_harness.pl , change the default test inclusion list, and make smoke
15:45 rashakil joined perl6
15:48 gugod_ is now known as gugod
15:54 takanori joined perl6
16:10 weinig is now known as weinig|brb
16:15 hcarty joined perl6
16:16 hcarty left perl6
16:17 Psyche^ joined perl6
16:18 Aankh|Clone joined perl6
16:19 amnesiac joined perl6
16:29 nothingmuch lambdabot: @seen luqui?
16:29 lambdabot I haven't seen luqui?.
16:29 nothingmuch lambdabot: @seen luqui
16:29 lambdabot I saw luqui leaving #perl6 13 days, 21 hours, 1 minute and 58 seconds ago, and .
16:29 nothingmuch damnit, every time
16:30 fglock I can't remember how to coerce an array into a hash in perl5
16:31 audreyt {@{$a}}
16:31 [particle] my %hash= %{[ @array ]};
16:31 audreyt er no, no [].
16:31 avar my %hash= %{@{[ @array ]}}; :)
16:32 avar just do my %hash = @array if @array % 2;
16:32 fglock perl -e ' @a="a","b"; print %{[ @a ]}{a}; ' - doesn't work
16:32 avar except without the undef behaviour
16:32 [particle] fglock: change % to $ in that exapmle
16:33 audreyt fglock: you've been brainwashed by p6...
16:33 avar my %hash; %hash = @array unless @array % 2;
16:33 avar I mean
16:33 audreyt $ perl -le 'print ${{"a","b"}}{a}'
16:33 audreyt b
16:33 audreyt $ perl -le 'print {"a","b"}->{a}'
16:33 audreyt b
16:34 audreyt pick one :)
16:34 xerox ->{a} ? :D
16:34 audreyt sure
16:34 avar to access the anon hashref
16:34 audreyt $ perl -le 'print ENV->{LC_ALL}'
16:34 audreyt zh_TW.UTF-8
16:35 Psyche^ is now known as Patterner
16:35 fglock no luck with ${{@a}}{a} or ${@a}{a}
16:35 audreyt since it's rewritten to ${ENV}{LC_ALL} and finally $ENV{LC_ALL}
16:35 avar like my $hr = { a => "b" }; $hr->{a];
16:35 asz joined perl6
16:36 xerox anon?
16:36 xerox Yuck.
16:36 audreyt fglock:
16:36 audreyt $ perl -le 'my @a = qw(a b); print ${{@a}}{a}'
16:36 audreyt b
16:36 avar audreyt: I didn't know of that
16:36 avar (ENV->..)
16:37 audreyt it's not exactly common knowledge :)
16:37 avar Is all of this stuff documented somewhere?
16:37 audreyt another not exactly common knowlege:
16:37 audreyt $ perl -le 'print %ENV->{LC_ALL}'
16:37 audreyt zh_TW.UTF-8
16:37 avar like my @a; push a, "b";
16:38 fglock audreyt: thanks :p
16:38 audreyt it's documented somewhere, I'm sure, but it's very hard to find them...
16:38 gaal audreyt: spread the ev0l :)
16:38 audreyt ...I know the ENV-> trick is documented in Damian's selfgol talk
16:38 gaal it's actually only half documented. it's mentioned in perldiag.
16:38 audreyt ah. right. that's where I learned it
16:39 gaal as a technique to avoid sigils in Makefiles that have to be cross-platform, I think it's best documented in the commit log for a certain Perl 6 implementation...
16:40 audreyt :)
16:41 Aankhen`` joined perl6
16:42 gaal is there a config option to get irssi to display unicode? here I see unicode in other screen(1) windows but not in irssi.
16:43 lumi /set term_type utf-8
16:44 gaal lumi: rssi: Unknown setting term_type
16:44 lumi What's in your term_charset?
16:45 wilx Hmm, google, I don't remember what it was, but one of the options was renamed....
16:45 gaal UTF-8
16:45 lumi Mu. Dunno then, sorry :/
16:46 gaal ah well.
16:46 audreyt hm. embedding perl5 always with pugs, and then demand perl 5.8, can give rise to lots of nice utility functions
16:47 audreyt e.g. full encoding support without thinking about libicu.
16:47 audreyt or rx:P5// that supported embedded closures.
16:47 gaal embedded Perl 6 closures?
16:47 audreyt code inside rx:P5// are taken as p5, iirc.
16:47 audreyt but sure, callback to p6land is already there, if needed.
16:49 gaal does activestate come with all the .h files needed to embed perl by default?
16:49 audreyt yes, and Limbic_Region already confirmed it as working.
16:49 gaal cool
16:50 gaal does ghc's bundled gcc take care of everything? that's great
16:50 audreyt yup.
16:50 gaal (on windows)
16:50 audreyt the earlier buildfail was simply that
16:50 audreyt perl5's dirent.h was loaded
16:50 audreyt instad of mingw's
16:50 audreyt so need a -D
16:50 audreyt that's all
16:51 gaal what does switching rx backend entail though? wouldn't that be quite some work?
16:51 audreyt but that means I need to more properly use ExtUtils::Embed...
16:52 audreyt well, calling into perl5 is already there, so it's not a lot of work
16:52 audreyt but not going to tackle that any time soon
16:52 audreyt it involves writing a RRegex.Perl5
16:52 audreyt that uses the Pugs.Embed.Perl5 callconv
16:52 audreyt and then just use it like RRegex.PCRE.
16:53 gaal that doesn't look too hard
16:53 audreyt it probably isn't
16:53 * gaal shudders at the comment at the top of those files.. "arch-tag:..."
16:53 gaal <- tla ph33r
16:55 * audreyt suddenly recall she had another module that used ExtUtils::Embed extensively and contains lots of useful workarounds
16:56 gaal oh? this is going to be funny :)
16:56 audreyt PuAgRs
16:56 gaal w00t
16:57 ludan joined perl6
16:58 chris2 joined perl6
16:58 svnbot6 r12858 | fglock++ | * v6 - more unboxed types (misc/pX/fglock)
16:59 audreyt perhaps PUGS_EMBED to '-perl5' will disable p5 embedding
16:59 audreyt or '!perl5'
16:59 gaal whee http://search.cpan.org/src/SMUEL​LER/PAR-0.952/myldr/Makefile.PL has OS/2 support :-)
17:00 Limbic_Region audreyt - you don't like 5lrep for that?
17:00 audreyt yeah, contributed by ilyaz
17:00 audreyt Limbic_Region: mmm?
17:00 Limbic_Region perl5 spelled backwards
17:00 * Limbic_Region murmurs redrum redrum
17:00 Limbic_Region sorry - just got back from lunch
17:01 audreyt np :)
17:01 * gaal has to moose for a bit! see y'all :)
17:01 Limbic_Region testing in a sec
17:01 Limbic_Region gaal - thanks for earlier
17:04 Limbic_Region anyone know what "dubious - Test returned status 5 (wstat 1280, 0x500)" means in english?
17:04 fglock audreyt: I'm testing compile-time autoboxing emulation, it seems to work so far - http://svn.openfoundry.org/pugs/perl5/Pugs-C​ompiler-Perl6/lib/Pugs/Emitter/Perl6/Perl5/
17:04 lambdabot Title: Revision 12858: /perl5/Pugs-Compiler-Perl6/l​ib/Pugs/Emitter/Perl6/Perl5
17:06 audreyt Limbic_Region: it segfaulted.
17:06 audreyt fglock: ooh! nice
17:06 svnbot6 r12859 | audreyt++ | * First sketch at using ExtUtils::Embed on non-Win32 platforms
17:06 svnbot6 r12859 | audreyt++ |   for perl5 embedding.  Need to the solve "make ghci" problem
17:06 svnbot6 r12859 | audreyt++ |   (probably using 'ar' to extract DynaLoader.o out) before
17:06 svnbot6 r12859 | audreyt++ |   we can enable this by default.
17:07 Limbic_Region audreyt - ok, question is why - it works fine from the command line
17:07 audreyt and with that I'm back to moose as well... see y'all
17:07 Limbic_Region pugs -Iblib6/lib t/examples/all_parse.t # works but not under harness
17:07 Limbic_Region TTFN audreyt
17:08 audreyt there's many possibilities... but TTFN
17:08 Daveman joined perl6
17:12 * Limbic_Region thinks he has it
17:15 crem_ joined perl6
17:15 Limbic_Region the debugging info that is making it work under the command line is making it fail under the harness
17:18 * Limbic_Region will figure it out another time
17:37 TreyHarris heh.  smoke on r12792: duration 12894 secs, 10400 ok, 679 failed.  r12835: duration 9654, 10547 ok, 617 failed
17:37 TreyHarris I think r12835 was better :-)
17:39 nothingmuch joined perl6
17:49 larsen joined perl6
18:06 TreyHarris ?eval class Foo { has ($.y is rw, @!a); method bar (Str $yish, Array @aish) { $.y = $yish; @!a = @aish } }; my $x = 'hi'; my @z = <<hi $x how are you>>; my Foo $obj .= new; $obj.bar($x, @y); say $obj.perl
18:07 [[mago]] joined perl6
18:07 TreyHarris ?eval 1 + 1
18:08 ruoso joined perl6
18:09 xinming TreyHarris: I think you can join #perlchina for eval the bot currently
18:09 xinming hmm, It seems I have to rerun the p6bot-loop
18:09 xinming bbl
18:09 TreyHarris xinming: thanks
18:10 DaGo joined perl6
18:10 xinming I mean the perlchina one, I don't have right to change this bot. :-P
18:11 TreyHarris xinming: yep, understood.  but i don't need the bot after all
18:11 autark joined perl6
18:25 svnbot6 r12860 | fglock++ | v6 - more tweaks to make [email@hidden.address] generate '${{@array}}{"c"}'
18:26 markymoon joined perl6
18:29 weinig|brb is now known as weinig
18:32 qmole joined perl6
18:36 TreyHarris just noticed r12853.  when will ".foo given XXX" and "XXX.foo" not be interchangeable?
18:37 araujo hello!
18:37 TreyHarris is postfix given just provided for completeness?
18:37 TreyHarris araujo: que tal
18:37 araujo hola TreyHarris :-)
18:39 markymoon left perl6
18:42 Limbic_Region TreyHarris - I received your email but marked it as spam at first
18:42 Limbic_Region something made me decide to go pull it out of the trash though ;-)
18:42 TreyHarris um... ok
18:42 * TreyHarris doesn't know how to interpret that
18:42 pdcawley joined perl6
18:43 TreyHarris i wasn't asking you to donate to anything
18:44 Limbic_Region no no
18:44 Limbic_Region just that I received an email from someone I don't correspond with regularly with a subject line of "answer"
18:44 Limbic_Region I flagged it (and many others at the same time) as spam
18:44 Limbic_Region a few minutes later something in the back of my head told me to actually read that one
18:44 Limbic_Region so I fetched it out of the trash
18:45 TreyHarris oh, i see. :-)
18:45 frederico joined perl6
18:45 TreyHarris i thought you meant you read it and then marked it.  lol
18:47 Limbic_Region not that I want to open up that can of worms again but I think I wasn't clear that the majority of my negative sentiment was directed towards the PerlMonks fund within TPF
18:47 Limbic_Region TPF as a whole, I haven't dealt with enough to have formed an opinion
18:48 TreyHarris ah
18:48 pdcawley left perl6
18:58 weinig is now known as weinig|food
18:58 svnbot6 r12861 | fglock++ | v6 - more unboxed Array methods
19:03 Aankh|Clone joined perl6
19:07 penk joined perl6
19:09 ruz joined perl6
19:11 marmic joined perl6
19:11 perlbot joined perl6
19:12 perlbot joined perl6
19:17 weinig|food is now known as weinig
19:29 DaGo joined perl6
19:40 Lorn joined perl6
19:40 jferrero joined perl6
19:43 svnbot6 r12862 | gaal++ | * Alias for upload-smoke, because half of the time I get it backwards
19:46 iblechbot joined perl6
19:47 Daveman :p
19:52 cognominal milgram++
20:10 avarab joined perl6
20:13 fglock araujo: ahora tenemos la pagina tambien en espanol: http://www.perl.org.br/twiki​/bin/view/YAPC/YAPCSA2006_ES
20:13 lambdabot Title: YAPCSA2006_ES &lt; YAPC &lt; TWiki
20:14 weinig joined perl6
20:14 hexmode joined perl6
20:15 weinig joined perl6
20:17 avarab is now known as avar
20:18 buubot joined perl6
20:19 buu joined perl6
20:24 PolettiX joined perl6
20:32 svnbot6 r12863 | fglock++ | v6 - more emitter tweaks - [email@hidden.address] compiles to
20:32 svnbot6 r12863 | fglock++ |   ( map { ( $_, @array[$_] ) } 0..@array-1 )[4,5]
20:37 frustrato joined perl6
20:45 frederico joined perl6
20:49 [[mago]] joined perl6
21:19 ToTimkopf joined perl6
21:24 ruz joined perl6
21:33 fglock left perl6
21:44 justatheory joined perl6
21:49 xdg joined perl6
21:59 frederico joined perl6
22:11 awwaiid joined perl6
22:14 [[mago]] joined perl6
22:44 cjeris left perl6
22:45 explorer__ joined perl6
22:50 weinig is now known as weinig|work
22:51 foo\ joined perl6
22:57 weinig|work is now known as weinig
23:24 weinig is now known as weinig|bbl
23:29 markstos joined perl6
23:39 markstos We have a Perl6::Perl5::Differences, but I wonder if we should have two documents for Perl5 users: 1, like Differences.pod mostly does now, would focus on how to Perl5ish things in a Perl6ish way.  And 2. more of a "new features" document. These would of course be differences as well, but would be things which /didn't exist/ in Perl5.  Thoughts ?
23:54 * markstos just sent a clarification question about return types to p6l
23:55 miyagawa_ joined perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo