Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2006-08-17

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 lumi_ joined perl6
00:05 TreyHarris oh, no.  even after a make clean, it's still linked in a non-existant libparrot.
00:08 Eidolos joined perl6
00:10 jferrero joined perl6
00:11 dduncan anyway, I think I'm starting to get the hang of using propedit+VIM for basic stuff
00:16 kanru joined perl6
00:18 zgh is perl5 embedding in pugs broken? or is it just my compile?
00:18 TreyHarris dduncan: hooray!  i have a working pugs again :-)
00:19 dduncan I'm glad
00:19 TreyHarris thanks for your help dduncan++
00:20 dduncan so now I'll just have to figure out how to use libreadline myself
00:20 dduncan assuming its actually useful
00:20 TreyHarris well, i dunno.  just when you get used to being able to edit and use command-line history at a prompt, it's really hard to go back to a dumb prompt
00:22 macroron joined perl6
00:22 zgh readline is essential to my way of life
00:22 zgh vi-keybindings specifically
00:24 nekokak joined perl6
00:28 markstos joined perl6
00:29 svnbot6 r12340 | Darren_Duncan++ | set svn:ignore props to hide third-party/installed and any third-party/judy files that aren't in the repo
00:38 svnbot6 r12341 | Darren_Duncan++ | third-party/ : set svn:ignore props on appropriate parts of HsJudy, HsSyck, fps
00:41 svnbot6 r12342 | Sage++ | The translator test script is now recursive. Just run it in a directory and it recurses through all .yml files and tries to translate them.
00:41 svnbot6 r12342 | Sage++ | Useage is the same.
00:45 svnbot6 r12343 | Sage++ | The parser for the translation utility now fails much more gracefully if the file is essentially empty. It just returns as empty P5AST, which eventually results in an empty output file.
00:45 svnbot6 r12343 | Sage++ | This makes running tests across the full parsed yml of perl5 t provided by lwall much easier, as anything that's empty no longer causes a failure.
00:45 svnbot6 r12344 | Darren_Duncan++ | perl5/Pugs-Compiler-Rule/ : set svn:ignore props on make-generated files
00:45 macroron_ joined perl6
00:45 SageLT I;ve got a full test of the translator using the new recursive script running right now, I'll post the results for anybody interested.
00:46 SageLT Current feeling based on the early part of the test: as expected when you expand out your test base, more problems crop up. Not as much coverage as I'd like, but not too shabby either.
00:46 svnbot6 r12345 | Darren_Duncan++ | ext/Set-Relation/ : continued split of the 2 test files
00:47 SageLT Ah, the test just finished: out of 1278 files attempted, 142 failed in some catastrophic way. 90 had unknown nodes. Looks like I've got my work cut out for me.
00:47 SageLT That's roughly 90% translating without major errors.
00:48 dduncan This is Pugs Version: 6.2.12 (r12345)
00:48 dduncan now there's a number you don't see every day
00:51 dduncan now, the only non-repo files left without svn:ignore props are various perl5/Pugs-Compiler-Perl6/t/ files and folders plus perl5/Pugs-Grammar-MiniPerl6/​lib/Pugs/Grammar/MiniPerl6.pm ...
00:51 dduncan I'm loathe to hide those at the moment, but if someone actually working on those projects thinks its a good idea, then they can do so themselves
01:16 dolmans joined perl6
01:18 kanru2 joined perl6
01:23 markstos I'm running "pugs -cw CGI/Application.pm my CGI::App port, and I'm getting "unexpected end of input" ... at line 2323. I know the error must be about a 1,000 lines earlier, because most of the end of document is POD. Any tips on narrowing down where the error is ?
01:23 lollan joined perl6
01:24 xdg joined perl6
01:29 dduncan I'm not sure if its a transient on my system, but I'm currently running a smoke and getting lots of errors
01:29 dduncan I think I'll abort it and do a clean run, to see if its related to repo code or not
01:30 dduncan but that will take awhile, so see you later &
01:30 dduncan left perl6
01:35 zgh left perl6
01:50 * audreyt wakes up and sees the topic
01:50 TreyHarris audreyt: you broke it :-)
01:50 audreyt fixing
01:50 TimToady welcome back to the funny farm
01:51 audreyt I see that. should've known better than checkin 5kline changes at midnight 2am
01:51 audreyt who would have thunk?
01:51 TimToady looks like some of the Var/_Var thingies are b0rken, like .keys
01:52 TimToady don't know why that prevents "return if" from parsing though.
01:52 TimToady unless it's looking ahead for the wrong category for statement modifiers somehow
01:52 TimToady maybe trying to figure out "return if =>"
01:53 TimToady or maybe it's the other way around, and the parser is hosing the _Var thingies
01:54 TreyHarris i was reading one of audreyt's old journal entries, and I saw a reference to "multi sub foo (0) { ... } multi sub foo (1..10) { ... }".  has that kind of pattern matching gone away, or just not implemented yet?
01:54 TimToady not implemented yet
01:55 TimToady treating values as subsets is rather important
01:55 TreyHarris ok
01:56 TreyHarris i can't find a test, which was why i asked.  should i write one for multi_sub.t?
01:57 TreyHarris but i admit i haven't read every single test yet :-)
01:57 TreyHarris or maybe it belongs in signature_matching.t
01:58 TimToady there are no tests containing /sub.*\(0\)/, which i would expect.
01:59 TreyHarris yeah, that's what i looked for, but it could be any value, and i couldn't figure out how to write a regex for that :-)
01:59 TimToady implementation is probably waiting on all the Capture/Signature stuff to settle down.
02:00 markstos Perl6 blog entry, powered by 'given', 'when' and the smart match operator. Total LoC reduction: 50%! http://use.perl.org/~markjugg/journal/30663
02:00 lambdabot Title: You have exceeded your Internet quota
02:00 TreyHarris lol
02:01 TimToady ./data_types/subtypes.t:        multi sub my_abs (Num where { $^n <  0 } $n){ -$n }
02:01 TimToady that's about as close as it comes, i think.
02:05 TimToady markstos: the parens around the arg to die are doing no useful work.
02:06 markstos TimToady: Thanks. Removing them now.
02:06 TimToady also does weird line breaks.
02:07 hikozaemon joined perl6
02:07 TimToady reducing my font size seems to fix that though
02:08 TimToady (I tend to run with fairly large fonts for my aging eyes.)
02:11 kanru joined perl6
02:12 TreyHarris perlbot: learn "#perl6 logs" as http://colabti.de/irclogger/irclogger_logs/perl6
02:12 perlbot added "#perl6 logs" to the database
02:12 lambdabot Title: You have exceeded your Internet quota
02:12 svnbot6 r12346 | audreyt++ | * A typo (SArray info SArrayMulti) in Operator.hs rendered
02:12 svnbot6 r12346 | audreyt++ |   &return into unary, instead of listop. :-/
02:12 svnbot6 r12346 | audreyt++ |   The @@ sigil caused us trouble even before it's implemented...
02:12 TreyHarris perlbot: #perl6 logs
02:12 rashakil left perl6
02:12 TreyHarris perlbot: learn #perl6 logs as http://colabti.de/irclogger/irclogger_logs/perl6
02:12 perlbot added #perl6 logs to the database
02:13 lambdabot Title: #perl6 irc log
02:13 TreyHarris perlbot: #perl6 logs
02:13 perlbot http://colabti.de/irclogger/irclogger_logs/perl6
02:13 lambdabot Title: #perl6 irc log
02:17 audreyt should be unb0rked now.
02:17 audreyt Foo::bar.new broken. @foo.map:{...} broken; fixing these two. but smoke loop should be runnable
02:22 mdiep joined perl6
02:29 audreyt Foo:bar unbroken.
02:29 svnbot6 r12347 | audreyt++ | * restore metaobjects for Multi::Part::Class's metaobject
02:29 audreyt .map:{} seems fine now
02:29 audreyt changes topic to: SVK users: http://search.cpan.org/dist/SVN-Mirror/ | paste: http://sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | http://www.treehugger.com/f​iles/th_images/paradigm.jpg
02:30 TreyHarris excuse my ignorance, but what is Foo:bar?  run Foo() with an invocant of bar()?
02:32 amnesiac joined perl6
02:32 markstos "trinary" is really spelled ternary, right ? I just spotted that in the docs.
02:33 TreyHarris markstos: a three-arged operator is ternary.  a SQL boolean that can also be null is trinary
02:33 audreyt "ternary"
02:33 markstos audreyt: thanks. Patching docs for !! ?? now.
02:33 audreyt TreyHarris: it's Foo(bar => 1)
02:34 audreyt surprisingly.
02:35 audreyt userdefined infix broken.
02:35 svnbot6 r12348 | audreyt++ | * Remove debug statements and reorg a bit.
02:35 TreyHarris markstos: oops, i was discerning a semantic difference where there is none.  both ternary and trinary can be used either way, and ternary is preferred for both.
02:36 audreyt hm wait. broken just with funny unicode brackets.
02:36 audreyt fixing that too
02:36 markstos TreyHarris: Thanks for the clarification, though.
02:37 markstos Ok, so subversion is finding no change when I know I just made one. How do I make it less dumb?
02:37 audreyt "svn add"
02:37 audreyt "svn st"
02:37 Bit-Man joined perl6
02:37 audreyt if you are patching docs/Perl6/Spec/
02:37 audreyt then that is not meant to be checked in; send to p6l
02:38 markstos Yes I am. Ah.
02:38 audreyt upstream is http://svn.perl.org/perl6/doc/trunk/design/syn/
02:39 markstos How about I just shout it out here: s/trinary/ternary/i ? I'm not already subscribed/following p6l.
02:39 xinming_ joined perl6
02:39 TreyHarris markstos: you should subscribe; conversations seem to move back and forth between here and there fairly willy-nilly :-)
02:40 TreyHarris it's not high traffic at all (lately at least, it flares up sometimes)
02:43 svnbot6 r12349 | markstos++ | s/trinary/ternary in docs/other/primes_rambling.pod
02:44 markstos Ok. Sent to p6l.
02:48 b_jonas left perl6
02:48 xinming__ joined perl6
02:49 traecer joined perl6
02:50 markstos Oh joy. After at least three evenings of hacking, pugs -cw reports "Syntax OK" on my CGI::Application port. Now I only need to start making the tests pass...
02:51 audreyt ooh.
02:51 audreyt with latest pugs?
02:51 * Khisanth waits for the "N"."O"x100
02:51 audreyt also, which parts of the metaobject protocol do you need again?
02:52 markstos No, 6.2.12 (r12060)
02:52 markstos Just class_precendence_list.
02:52 audreyt ok. once the tests startss passing
02:52 audreyt please check in to ext/CGI-Application/ :)
02:52 markstos Ok. :)
02:52 audreyt (if you'd like it not to be randomly broken next release)
02:52 audreyt markstos++
02:53 markstos When you say "start" passing, I assume you really mean "not completely passing?" :)
02:53 audreyt yup.
02:54 markstos Ok. Is the darcs gateway considered functional? I really feel more productive with it, and haven't been able to feel so warm and fuzzy about svk yet.
02:54 audreyt it's functional but not bidi
02:54 markstos Now I just use svn.
02:55 audreyt so to commit you need to use "svn add"
02:55 audreyt but that's fine
02:55 markstos So patches return by 'darcs send' ?
02:55 audreyt I mean, it's entire directory
02:55 audreyt it's not like you want/need merge history :)
02:55 markstos Right.
02:56 audreyt I agree re: darcs UI is more flow-happy
02:56 markstos BTW, is something like Carp::croak built into Perl6, or does that need to be ported?
02:57 markstos I love the interactiveness of darcs-- the interactive updates and commits.
02:57 audreyt currently our die() is just croak()
02:57 audreyt so use it for now I think
02:57 audreyt oh, that. "svk ci --interactive"
02:57 markstos A better default. :)
02:57 markstos (croak as die)
02:57 markstos Thanks for the svk tip.
02:57 audreyt interactive as default may also be better ;)
02:58 markstos Yes. :)
02:58 audreyt I think "SVK::Command::DarcsLike"
02:58 markstos Good for beginners, and in my opinion, good for flow.
02:58 audreyt svk rec # interactive by deefault
02:58 audreyt etc
02:58 markstos :)
02:58 audreyt will rock
02:59 markstos One part of the CGI::App port may not be fixed for a while: it depends on HTML::Template (which it probably shouldn't). My solution will probably be to push that functionality into a plugin and remove the dependency.
02:59 svnbot6 r12350 | fglock++ | v6 - state variables (need some work)
03:00 markstos So what happens if release time comes around and I haven't got all the tests passing for CGI::App yet? I don't want to mess up the bigger picture for my side project.
03:04 audreyt there's no bigger picture :)
03:04 audreyt your tests will just be marked todo-for-release by some passerby gnome
03:05 markstos Ok.
03:05 * markstos waves to gnomes
03:05 audreyt The gnome waves back! --More--
03:05 audreyt You feel much better.
03:05 buubot joined perl6
03:06 buu Hrm.
03:15 audreyt $job, bbl
03:16 markstos Here's what I mean, but it's a syntax error. What's correct? self.$coderef(@args)    ?
03:16 svnbot6 r12351 | trey++ | Added myself to AUTHORS, and fixed incorrect pathnames in docs/SEEALSO.
03:18 TreyHarris markstos: what are you wanting to do there?  run $coderef with self as an invocant, though self does not actually have that code as a method?
03:19 markstos Let me check what it looked it originally as Perl5. :)
03:19 PerlJam pugs groks self?
03:20 PerlJam I thought it was still using ./
03:20 markstos TreyHarris, in Perl5, this worked: $self->$callback(@args);
03:20 markstos PerlJam: I think "./" is a pugism and not official Perl6.
03:20 markstos Using self.foo works for me, so far.
03:21 PerlJam markstos: yes, ./ is a pugism.  I didn't realize that pugs understood self  (shows how often I write perl6 :)
03:21 markstos TreyHarris: So "Yes", what you said.
03:21 markstos PerlJam, well, from basic tests it seems to grok self.
03:24 PerlJam Grr.  svk says "Incomplete data: Delta source ended unexpectedly"
03:24 PerlJam so much for updating my local pugs
03:26 weinig_ joined perl6
03:26 markstos This isn't a syntax error, so I'll try it now for: $callback(self,@args);
03:27 PerlJam markstos: I wonder if self.$method.(@args) works.
03:28 markstos PerlJam: I thought of that, but it's a syntax error.
03:28 weinig_ is now known as weinig
03:29 weinig is now known as tarzan
03:29 tarzan is now known as weinig
03:32 weinig joined perl6
03:32 TimToady markstos: you will find neither trinary nor ternary in the synopses I'm maintaining.  It's now just the conditional operator.
03:33 markstos TimToady: Interesting. Because "conditional" is more grokable, or that it is meant to be chained, with lots of operators?
03:34 TimToady because it's grokable, and says what it's actually doing.
03:34 markstos OK.
03:34 TimToady looks like ext/Test-Builder/t/010_Test_Builder.t is still in a hard loop...
03:34 PerlJam TimToady++ no more bikeshedding on what to call it!
03:35 markstos So it seems like the website and the pugs repo are behind. They both have Trinary: http://dev.perl.org/perl6/doc/design/syn/S03.html
03:35 PerlJam Also, everyone always calls it "*the* ternary" operator as if there's only one possible.
03:35 PerlJam markstos: the website is brokenish as far as keeping things up to date.
03:35 TimToady okay, I grepped for "trinary", not "Trinary".  it's right in the precedence table though
03:36 PerlJam grep -i is your friend
03:36 markstos PerlJam: true, but in that sense /the/ conditional operator isn't better.
03:36 TimToady on the other hand, it's not using "Trinary" as the name of the operator.
03:36 TimToady it's just using it as one step up from binary.
03:37 TimToady but I'll fix it anyway.
03:37 PerlJam markstos: don't make me imagine some cute non-sensical name to call it
03:38 markstos PerlJam: like the ba-bang, qu-question operator?
03:39 PerlJam markstos: I was thinking something like "quabang"
03:39 markstos The double bang operator...
03:40 PerlJam It's a stuttering quabang
03:40 markstos "conditional" is fine by me, but it would be nice to have the word "ternary" nearby, as that's how Perl5'ers will search for it.
03:41 PerlJam TimToady: Why is it that NEXT blocks fire in reverse order?
03:42 TreyHarris ?eval my $coderef = sub (*@args) { say "{self}: @args[]." }; my $foo = "blahblah" but { sub foo {...}; &foo := $coderef }; $foo.foo(1,2,3)
03:42 evalbot_23570 is now known as evalbot_23612
03:42 evalbot_23612 OUTPUT[<SubMacro(self)>: blahblah 1 2 3. ] Bool::True
03:42 PerlJam markstos: yeah, but they won't recognize it when they find it.  :-)
03:42 TreyHarris markstos: that's *almost* what you want....
03:42 TreyHarris markstos: right?
03:43 * markstos thinks about TreyHarris example
03:43 TreyHarris ?eval my $coderef = sub ($self: *@args) { say "$self: @args[]." }; my $foo = "blahblah" but { sub foo {...}; &foo := $coderef }; $foo.foo(1,2,3)
03:43 evalbot_23612 OUTPUT[blahblah: 1 2 3. ] Bool::True
03:43 TreyHarris there.
03:43 * markstos scratches head
03:43 TreyHarris you have to mixin the coderef before you can call it as a method
03:44 markstos Aha.
03:44 TimToady PerlJam: NEXT fires in reverse order for the same reason LEAVE does, which fires in the same order as END blocks.
03:45 TimToady basically as the code is being elaborated, things get built up that naturally want to be deconstructed in the reverse order.
03:45 PerlJam That seems to make sense in my head except for NEXT for some reason.
03:45 TimToady the loop block is entered and left every time through
03:45 TimToady conceptually
03:45 TimToady NE
03:45 TimToady NEXT and LAST are kinda complementary variants of LEAVE
03:46 TimToady much like KEEP and UNDO
03:48 TimToady dinner &
03:50 * PerlJam re-reads S04 in search of Understanding.  :)
03:51 TreyHarris ?eval class Rx { has $.x = 19 }
03:51 evalbot_23612 Error:  unexpected "=" expecting comment, trait, "handles", ";" or "}"
03:51 svnbot6 r12352 | audreyt++ | * Lexical user-defined operators now works again:
03:51 svnbot6 r12352 | audreyt++ |     my sub infix:<foo> ($x, $y) { ... }
03:51 TreyHarris what's wrong with that?
03:52 audreyt defaulting not yet handled?
03:53 TreyHarris ah, right
03:54 Eidolos svnbot6 is reporting Pugs checkins, right? And I assume 23612 is the revision number of some other Perl6 project.. but which?
03:55 audreyt it's feather's local repo number
03:55 audreyt i.e. it's broken. thanks for the headsup
03:56 Eidolos Er.. heh, okay :)
03:59 audreyt fixed
04:00 svnbot6 r12353 | audreyt++ | * evalbot-loop: Eidolos++ reported that svk was reporting local versnum
04:00 svnbot6 r12353 | audreyt++ |   not pugs versnum.
04:01 audreyt TreyHarris: check for tests
04:01 audreyt see if that's tested
04:01 audreyt a workaround exit:
04:01 TreyHarris audreyt: i will
04:02 Eidolos ?eval 1+3
04:02 markstos I was surprised this was an error:
04:02 audreyt class Point { has $.x; submethod BUILD ($.x = 12) {} }; Point.new.x
04:03 audreyt will eval to 12 when evalbot i s back
04:03 markstos ?eval my %h = ( 1 => [2] ); my $b = [ @{%h{1}} ]; say  $b.perl;
04:04 markstos I get an error at "[". It seems something about casting the hash value back to an array doesn't work.
04:04 TreyHarris must uncommute, 9pm is far too late to be going home....
04:04 * TreyHarris &
04:04 markstos Or rather, I'm getting a syntax error.
04:04 audreyt my @b = %h<1>
04:04 audreyt er
04:04 audreyt my $b = %h<1>
04:04 audreyt but will fix
04:05 markstos audreyt: so it's a Perl6 bug?
04:06 audreyt there's no @{} anymore :)
04:06 neoesque joined perl6
04:06 audreyt there is, hoeever, @()
04:06 audreyt prefix @ is just a function like everyone else
04:07 markstos Ah. I'll test that.
04:07 markstos Yep. That works.
04:08 markstos Should that be mentioned in "Synopsis 9: Data Structures" ?
04:10 markstos Ah, I see light mention of it in S2, "Bits and Pieces".
04:14 audreyt I think I changed somewhere else in tokyo...
04:14 audreyt ah. S03
04:14 audreyt =item * Perl 5's C<${...}>, C<@{...}>, C<%{...}>, etc. dereferencing
04:14 audreyt forms are now C<$(...)>, C<@(...)>, C<%(...)>, etc. instead.
04:14 dduncan joined perl6
04:17 markstos What's the best practice for writing #!/bin/perl in a platform-independent way?  #!/usr/bin/env perl ?
04:18 audreyt either that or #!perl
04:18 audreyt which will portably Not Work everywhere
04:19 Maddingue or you let the EU:: modules change the path
04:19 markstos What? You mean there's More than One Way to Do It ?
04:19 audreyt via install_script(), yes
04:19 audreyt er install_scripts
04:20 SamB so that "#!/usr/bin/env perl" thing works on windows?
04:20 audreyt no
04:20 audreyt win32 is doomed anyway
04:20 SamB heh
04:21 svnbot6 r12354 | markstos++ | Update shebang line for util/prove6 to  be more platform independent.
04:29 dduncan ?eval 'fubar'
04:31 dduncan I have used #!perl myself over the last year for my Perl 5 stuff on CPAN
04:31 dduncan but not in Perl 6 considering that "use v6-alpha;" is the current way to do it
04:33 markstos Another question: I'm trying to make prove6 smarter. How I can check to see if 'pugs' is the  path already in a platform-independent way? A number of hacks come to mind, but nothing clean.
04:34 audreyt can_run('pugs')
04:35 markstos Where does 'can_run' come from?
04:35 audreyt Module::Install::Can
04:35 audreyt failing that, use ExtUtils::MakeMaker()
04:35 SubStack joined perl6
04:35 audreyt and then MM->maybe_command()
04:35 audreyt (see EU::MM manpage)
04:36 audreyt    for my $dir ((split /$Config::Config{path_sep}/, $ENV{PATH}), '.') {
04:36 audreyt        my $abs = File::Spec->catfile($dir, $_[1]);
04:36 audreyt        return $abs if $abs = MM->maybe_command($abs);
04:36 audreyt    }
04:37 markstos Thanks, audreyt: Just what I was looking for.
04:37 audreyt glad to know :)
04:38 miyagawa_ audreyt: I've a question about M::I and build_requires and auto_include_deps
04:38 audreyt miyagawa_: see Alias on #perl :)
04:39 miyagawa_ heh
04:39 svnbot6 r12355 | audreyt++ | * Implement "pugs -d" for printing out a debugging trace
04:39 svnbot6 r12355 | audreyt++ |   for programs.  Full debugger is for later.
04:39 miyagawa_ when you say build_requires('Test::Base') and do auto_include_deps, it'll try to include Algorithm::Diff and Text::Diff, which are actually optional requirement from T::B.
04:40 audreyt yeah. auto_include_deps is not "smart
04:40 miyagawa_ I'd just do auto_include (without _deps) but then build_requires('YAML') fails to include other YAML:: modules
04:40 audreyt manual include_deps() call
04:40 audreyt or put auto_include _after_ YAML b_r
04:40 audreyt and then put T::B's b_r after a__i_d
04:41 miyagawa_ I see. lemme try
04:44 kanru joined perl6
04:46 miyagawa_ build_requires('Test::Base');aut​o_include;include_deps('YAML'); does the trick
04:46 miyagawa_ audreyt++
04:46 audreyt yay
04:52 dduncan I'm in the middle of smoking r12345, but one of the tests is taking 10 minutes and counting
04:52 svnbot6 r12356 | audreyt++ | * Unicode identifiers should now work again.
04:52 svnbot6 r12357 | markstos++ | Improve the chances of 'prove6' working on code outside of the pugs tree.
04:52 svnbot6 r12357 | markstos++ | It will hopefully find pugs now, still not 'Test.pm'.
04:52 svnbot6 r12357 | markstos++ | I'm not sure if this is portable:
04:52 svnbot6 r12357 | markstos++ |  `pugs --version`
04:52 svnbot6 r12357 | markstos++ | (Works on Windows?)
04:52 svnbot6 r12357 | markstos++ |     Mark
04:53 dduncan I suppose that functionality to time-out long-running tests was never made
04:53 markstos sleep &
04:55 dduncan fortunately, I can still manually kill the 'pugs' process in question, which would then have a result like the test died rather than hanging, and so the suite moves on to the next one
04:56 dduncan so the rest of the smoke can continue
04:59 audreyt dduncan: yes, I saw that too
04:59 dduncan note that, I saw a comment on the backlog about a hanging test, but that isn't the same one I saw hang
05:00 audreyt nod
05:00 audreyt I'm fixing that
05:01 dduncan eg, the backlog comment said 010_Test_Builder.t was in a hard loop, but it completed normally for me ... on the other hand #21 or so had hung in my case
05:01 dduncan but no matter
05:02 dduncan on a separate note, do you have an ETA where one can actually declare a Mapping?  Afaik, that's the only base type that I want to use right now which isn't implemented last I saw
05:03 dduncan regardless, thanks for your efforts
05:03 audreyt Mapping(x=>1, y=>2) ?
05:03 dduncan yes
05:03 audreyt ETA is probably end of this month
05:03 audreyt i.e. in a couple weeks
05:03 dduncan okay
05:04 agentzh joined perl6
05:05 agentzh hey, folks, what is the easiest way of uploading files from my localhost to feather?
05:05 dduncan sftp?
05:06 agentzh invoke that from feather?
05:07 gaal morn
05:08 audreyt gaal: Eval.Var has a bug
05:08 audreyt eval "  my $say = -> Str $msg { say $msg }; sub f { $say(1,4,5) }; f(1,2,3)"
05:08 audreyt doesn't work currently. can you help investigating if you got cycles?
05:08 audreyt I need to run now to a meeting; bbl
05:09 gaal will after loads of coffee
05:12 kanru2 joined perl6
05:16 kanru2 is now known as kanru
05:18 agentzh gaal: i'm writing a new utility similar to catalog_tests.pl. err, i've named it "smart_links.pl". :)
05:20 agentzh it's supposed to work with both the old L<Sxx/xxx /.../> and the new L<Sxx/xxx/xx xx xx> magic link notations used in the test suite.
05:20 gaal agentzh: excellent!
05:20 gaal what's it written in?
05:20 agentzh perl 5, of course.
05:21 gaal some more speedups and it can be written in perl 6...
05:21 agentzh gaal: do you think i should offer my own css file?
05:21 agentzh gaal: the perl 5 version is even faster than i expected. :)
05:22 gaal I think you should start with the existing one, and if you see it's much different than what you want it to be, switch
05:22 gaal if the changes are minor, patch
05:22 gaal remember to include an option to embed the css in the output file, because when we developed this stuff originally some browsers for some reason required it.
05:22 agentzh gaal: you mean i should start with catalog_tests.pl?
05:23 gaal I mean with the css that that uses
05:23 agentzh gaal: sure.
05:23 agentzh gaal: k
05:24 cnhackTNT joined perl6
05:24 agentzh currently i'm optimizing smart_links.pl for memory instead of time.
05:24 agentzh because i think the number of tests (and also the number of links) are still increasing.
05:26 gaal were you holding all docs in memory?
05:26 gaal was catalog_tests?
05:26 agentzh gaal: nope.
05:27 agentzh i only save the file name and line numbers in the data structure.
05:28 agentzh process one file at a time.
05:28 gaal wow, i'm really groggy this morning
05:28 agentzh gaal: me too. :)
05:30 gaal if you're still groggy I have five hours of mist ahead of me :)
05:30 agentzh heh
05:31 * agentzh goes back to hack on smart_links.pl
05:33 * agentzh is very happy to see he can still compile pugs on win32 this morning.
05:33 dduncan I also noted another test that was running about 7 minutes, and I was about to kill it, then it ended on its own
05:33 dduncan t/rules/from_perl6_rules/properties.t was the one I think
05:35 TimToady yeah, those naturally tend to run long
05:35 agentzh hmm, it's funny to see catalog_tests.pl parses and modifies the HTML source generated from POD using HTML::TreeBuilder. i don't it's ideal, so i will take another approach.
05:36 agentzh *think
05:36 gaal agentzh: XML::All it :-p
05:36 kanru joined perl6
05:37 agentzh gaal: will take a look at XML::All.
05:37 MacVince joined perl6
05:38 gaal whee, "falseguard trace" is so useful
05:38 gaal findSub var invs args | trace ("> findSub " ++ (show var) ++ " " ++ (show invs) ++ " " ++ (show args)) $ False = undefined
05:38 gaal | ... other guards ...
05:40 agentzh gaal: catalog_tests.pl currently uses http://dev.perl.org/css/perl.css by default, which is rather ugly in both my firefox and ie browsers. :/
05:42 gaal for the synopses?
05:42 gaal I thought you were asking about tests
05:42 gaal by all means feel free to use another :)
05:43 dduncan fyi, my smoke of r12345 is done and uploaded to the smokeserv
05:43 gaal it you maintain scheme compatiblity consider doing the alternate stylesheet thing so that usres can pick
05:43 dduncan it shows a 77% success rate, compared to 98% of r12237
05:44 dduncan I will now pull and try again
05:46 agentzh gaal: okay.
05:46 agentzh dduncan: the url of the smokeserv please?
05:47 dduncan http://smoke.pugscode.org/
05:47 lambdabot Title: Pugs Smoke Reports
05:47 agentzh thanks
05:47 dduncan I suggest bookmarking it
05:47 TimToady in Test::Builder, new is calling itself recursively via $Class.SUPER::new(...)
05:47 agentzh dduncan: the url is broken here.
05:48 dduncan it works for me
05:48 dduncan what do you get?
05:48 agentzh dduncan: it's redirecting to another site, right?
05:48 dduncan yes
05:48 dduncan but the shorter url is easier to remember
05:48 agentzh redirection is killed by the great firewall.
05:48 dduncan just a minute ...
05:48 dduncan try this then: http://m19s28.vlinux.de/cgi-bin/pugs-smokeserv.pl?
05:48 lambdabot Title: Pugs Smoke Reports
05:49 agentzh yay, it works!
05:49 agentzh bookmarking...
05:50 dduncan fyi, I'm sure that some of that 20% drop was due to simple things that were already fixed since then, so it may not be the best set for looking at tests to fix
05:50 TimToady which is odd, because Test::Builder.pm doesn't seem to define a superclass.
05:51 dduncan for example, .new stopped working
05:51 dduncan and "return;" was a parsefail
05:55 MacVince I have a quick, non-Pugs or Perl6 related question: why does Audrey add "++" to the name of everyone she mentions in blog posts?  Like gaal++ and merlyn++?
05:55 MacVince Praise?
05:55 dduncan pretty much
05:55 dduncan also, there's this little bot which pays attention to those things and keeps a karma score
05:55 dduncan for those who care
05:55 MacVince dduncan++
05:55 MacVince ;-)
05:56 * agentzh ++ smiles.
05:56 merlyn the karma chameleon
05:56 dduncan that said, the ++ notation is used in many different places, and the bot is only in some of those ... so more generically its just credit/praise
05:56 gaal running out for some errands, back in a couple of hours I hope...
05:57 MacVince A form of new speak in a sense?
05:57 MacVince audreyt++ *= 2;
05:59 dduncan fyi, most IRC clients will also holler when its user's IRC nickname is mentioned, so its common practice that if you want to reach someone, they're more likely to know if you mention their nick, often in the form of "ping <foo>"
05:59 dduncan suffice it to say, your comments probably just caused a number of notifications
05:59 MacVince I usually just just <nick>: hello?
05:59 MacVince Heh
06:00 dduncan the actual thing you say doesn't matter; just spelling out the nick exactly and in full will cause the notification
06:00 dduncan but that's strictly a client feature, a common one though
06:01 dduncan in my Colloquy client for Mac OS X, for example, a nick mention will cause 3 attention-gathering actions: a beep will sound, the program's dock icon will bounce (very attention grabbing, as humans are optimized to notice movement), and a brief floating text message will appear above all other app windows in the corner
06:02 agentzh wow
06:02 weinig is now known as weinig|sleep
06:02 ayrnieu humans are also optimized to shut that shit off right away.
06:02 dduncan the bouncing lasts for awhile, so even if the person's away at the moment, it may still be happening on their return
06:02 MacVince dduncan: I'd find that annoying.  On OSX, I use MacIrssi which adds a red dot to the dock icon.
06:02 dduncan Colloquy is customizable, but what I describe is its default behaviour, and I haven't changed it
06:03 MacVince And on Linux, I use rcirc which just makes a small mention in Emacs statusbar.  Of course, I could add a hook and start playing music and stuff
06:03 dduncan since I like to respond quickly to people, it is useful
06:03 MacVince dduncan: I wonder if all three are necessary. Like you said, the bouncy icon is probably enough to grab your attention
06:03 MacVince That's just me though.
06:03 dduncan yes, but that and the sound attract different senses
06:04 MacVince TMTOWTGN
06:04 dduncan ears vs eyes
06:04 MacVince (GN -> Get Notified)
06:04 dduncan I figured that out
06:04 MacVince Just in case
06:04 dduncan come to think of it, I think the floating message doesn't always appear ... probably just appears when the computer's sound is turned down, as a substitution
06:04 dduncan maybe
06:06 ayrnieu macvince - you should've said "JIC".
06:06 MacVince ayrnieu: ;)
06:07 dduncan anyway, I've got some $job work to do, so I will sign off, &
06:07 dduncan left perl6
06:08 ayrnieu see, with a less annoying IRC client, he could just silently not pay attention for a while, and then respond four hours from now, to someone who will notice tomorrow and catch dduncan in passing.
06:10 agentzh gaal: i would reuse the ``refresh_specs'' sub in docs/perl6/Makefie.PL. so smart_links.pl can stick to docs/perl6/Spec/*.pod and the user don't have to specify the directory where synopses live.
06:34 crem joined perl6
06:41 HXanadu joined perl6
06:50 MacVince Do I absolutely need parrot to use pugs?
06:50 ayrnieu macvince - no.
06:50 MacVince ayrnieu: cool, thanks.
06:50 MacVince Ubuntu doesn't have the latest version and I can't get it to compile properly.
06:50 TimToady however, you're limited to pcre-style regexen.
06:53 MacVince TimToady: I am more interested in the features such as list manipulation
06:53 MacVince Just to check them out.
06:55 agentzh whee! util/smart_links.pl already parses docs/Perl6/Spec/*.pod and t/*/*.t.
06:56 agentzh the next step is to glue them together...
07:05 TreyHarris how do i run just a single .t file correctly?  i notice that 'pugs ....file.t' doesn't work
07:08 TreyHarris ah, i got it.  util/prove6
07:13 agentzh TreyHarris: anther choice is: "pugs -Iblib6/lib t/statements/do.t" :)
07:14 agentzh *another
07:14 TreyHarris agentzh: ah, will that show me what todos unexpectedly succeeded?
07:14 TreyHarris that's what i want
07:14 agentzh of course, util/prove6 is more handy.
07:15 buubot joined perl6
07:16 * agentzh is excited to see more and more people are using util/prove6.
07:17 TreyHarris agentzh: oh, did you write it?
07:17 agentzh yes
07:17 agentzh the p5 implementation is done by me.
07:17 TreyHarris agentzh: i think i'd like to add a flag to point out succeeded todos.  would you mind if i did so?
07:18 TreyHarris like --show-todos or something
07:18 agentzh TreyHarris: go ahead!
07:26 * agentzh is tired of the long long "===" thread on p6l.
07:31 elmex joined perl6
07:32 TreyHarris agentzh: ah, it's more complicated than i thought, i'd never actually looked at the 'bonus' field before.  i thought it was a list of the tests that had unexpectedly passed.  it's actually a counter.  so to do what i want, i'll have to handle the TAP output myself.  so not tonight :-)
07:32 agentzh hehe
07:33 agentzh no hurry.
07:34 agentzh well, manipulating TAP output directly in prove6 is not quite clean i'm afraid.
07:35 TreyHarris yes, i noticed that.  that's why i said not tonight :-)
07:35 shachaf joined perl6
07:36 TreyHarris are we marking tests :todo that succeed, but for the wrong reason?  like a test that 'frobniz' returns undef under the circumstances it's supposed to, but in fact 'frobniz' is always returning undef?
07:37 agentzh sorry, i don't quite follow you.
07:38 agentzh btw, the :todo notation currently used in the test suite is now deprecated.
07:38 agentzh please see TASKS for more info on this.
07:39 TreyHarris say the behavior we want to test for is that "foo" returns zero when "$bar" is 1.  but actually, "foo" is totally unimplemented such that any use of "foo" returns zero.  so all the other tests involving "foo" need to be marked todo.  should the one that's currently succeeding by accident be marked todo as well?
07:41 agentzh TreyHarris: yes, it's very likely to happen.
07:42 agentzh but i think audreyt tends to unTODO succeeding TODO tests.
07:42 TreyHarris agentzh: ok, i read TASKS... it sounds like it's deprecated, but the successor is not yet decided?
07:42 agentzh TreyHarris: the successor is a call to the &Test::todo function.
07:43 agentzh Test::todo currently is not implemented in Test.pm.
07:45 TreyHarris agentzh: so what is the upshot of that?  don't write any tests that are todo until it is?  write them using the new syntax even though that will cause the test suite to fail?  write the new syntax, but wrap it in eval?  or continue using :todo for the time being? :-)
07:46 agentzh TreyHarris: currently the strategy is simply leaving tests failing until the new todo mechanism gets implemented.
07:47 agentzh well, parsefail in tests is more annoying.
07:47 TreyHarris ok.  comment them as todo? or just write them
07:47 agentzh just write them.
07:47 TreyHarris the two tests i was going to write tonight are both parsefail.
07:47 agentzh if they're basic tests
07:48 agentzh put tests which can't be parsed in eval please.
07:48 agentzh so you can test your tests.
07:48 agentzh *so that
07:48 TreyHarris ok.  but don't mark todo
07:49 TreyHarris got it
07:49 agentzh :)
08:06 agentzh oh no...smart_links.pl has found many broken links in the test suite...
08:06 ruoso joined perl6
08:10 traecer left perl6
08:13 TreyHarris agentzh: take a look at this commit (r12358) and tell me if that's what you meant
08:13 svnbot6 r12358 | trey++ | Added tests for 0 but True to oo/syntax-but.t
08:13 agentzh TreyHarris: looking
08:15 gaal remoose!
08:16 agentzh TreyHarris: i think the smart link in your test file is broken.
08:16 agentzh gaal: a lot of old smart links are broken in the test suite. :/
08:16 agentzh gaal: i'm fixing them by means of smart_links.pl.
08:17 agentzh TreyHarris: L<<S02/Context /objects can override the class definition:/>>
08:17 gaal agentzh: excellent! I'm backlogging now but I see you mentioned manipulating TAP data
08:17 gaal please look at Test::TAP::Model
08:17 agentzh gaal: it's TreyHarris who wants to manipulate TAP data. :)
08:17 TreyHarris agentzh: ah, because it spans lines?  the smart link only works on a single line in the POD?
08:18 agentzh TreyHarris: no, smart links can span two lines.
08:18 gaal agentzh: wanna implement &Test::todo?
08:18 gaal ah
08:18 agentzh the syntax is L<...> instead of L<<...>>.
08:18 gaal TreyHarris: I'm backlogging now but I see you mentioned manipulating TAP data
08:18 gaal please look at Test::TAP::Model
08:19 agentzh gaal: are $?VERSION and $?COMPILER already implemented in pugs?
08:19 gaal ?eval $?VERSION
08:19 gaal I know $?COMPILER isn't
08:19 agentzh once they'
08:19 agentzh re available, i'd have a try.
08:20 agentzh :)
08:20 agentzh TreyHarris: please look at t/syntax/comments.t for the new smart link notation.
08:20 gaal agentzh: nah, go ahead, it's a trivial change from "%?CONFIG<xyz>" to $?xyz :-)
08:20 gaal or, alternatively, you could add them yourself! it's easy :)
08:20 agentzh TreyHarris: L<../.. /.../> is deprecated now.
08:20 gaal agentzh: see Pugs.Run
08:21 agentzh gaal: looking
08:21 TreyHarris agentzh: oh.  ok, then one more time after this commit. :-)
08:21 svnbot6 r12359 | trey++ | Fixed smartlink, and the documentation that led me wrong
08:21 agentzh gaal: next?
08:21 gaal hmm?
08:22 gaal around line 140
08:22 gaal all that block defines globals
08:22 agentzh okay
08:22 agentzh hmm, it reads quite well. :)
08:23 gaal look how $*OS works
08:23 gaal it looks at the config
08:23 gaal just add two symbols like that, that peek at versnum and perl_compiler
08:24 gaal of course these eventually need to be renamed, but they do need to be in Pugs.Config so pugs -V shows them
08:24 TreyHarris agentzh: i'm thoroughly confused.  look at the very first smartlink in the file t/syntax/comments.t you pointed me to.  what in the world does "Embedded comments"
08:24 agentzh gaal: will do!
08:24 gaal agentzh++
08:24 * gaal looks into the OUTER::say problem
08:24 TreyHarris ... does '"Embedded comments" "#" plus any bracket' mean?
08:24 agentzh TreyHarris: looking
08:25 agentzh Embedded comments is a =head in POD.
08:25 TreyHarris no it isn't.
08:25 agentzh TreyHarris: after the second slash, it's a list of "keywords".
08:25 TreyHarris unless it's recently been changed
08:25 TreyHarris ok.  and what do the keywords do?
08:26 TreyHarris so the keywords here are ["Embedded comments", "#", 'plus', 'any', 'bracket']?
08:26 agentzh a sec...
08:26 agentzh sorry, "Whitespace and Comments" is the section name (=head)
08:27 agentzh TreyHarris: eys
08:27 agentzh *yes
08:27 TreyHarris ok.  so regexes are no longer used?
08:27 agentzh but all the keywords should appear within one sentence.
08:27 agentzh and the order is significant.
08:28 agentzh Treyharris: the regex notation is obsolete but will still be supported during the meantime.
08:28 agentzh at last, all the smart links in the suite will switch to the new notation.
08:29 agentzh in fact, i'm implemented util/smart_links.pl which can recognized all the smart links in the test suite.
08:29 agentzh *implementing
08:32 TreyHarris ok, so look at this commit and tell me if i have it right now
08:32 svnbot6 r12360 | trey++ | Fixed smartlink again, and the documentation that led me wrong again.  Thanks to agentzh++ for pointing me at the new syntax
08:33 TreyHarris then i'll go to bed :-)
08:33 dduncan joined perl6
08:33 agentzh TreyHarris: the trailing slash is not needed.
08:34 agentzh the trailing slash is actually a indicator of the good old regex notation. :)
08:34 agentzh *an
08:34 TreyHarris oh, bleh :-)
08:35 dduncan ayrnieu, regarding my IRC client, I made a choice to be signed into this IRC channel when I wanted to be responsive to it or to talk, and I sign out when I'm not or it is sleep time ... its not like I miss part of the conversation from it being closed, thanks to this channel being irc-logged
08:35 * agentzh wonders if the new smart link syntax is so difficult to understand.
08:35 svnbot6 r12361 | trey++ | One last time agentzh++
08:35 agentzh TreyHarris++
08:36 dduncan agentzh, regarding the === thread on p6l, I don't plan to respond to it any more, so that should help to not extend it
08:36 TreyHarris agentzh: no, it's just that the README made no mention of it, and was fairly specific about the wrong syntax
08:36 agentzh TreyHarris: i will tweak t/README shortly.
08:36 agentzh dduncan: cool
08:36 dduncan as mentioned, the matter is well documented in Synopsis 3, so a discussion isn't really needed on the list
08:36 TreyHarris agentzh: oh, you didn't like my tweak?
08:37 agentzh TreyHarris: please do if you like.
08:37 agentzh :)
08:37 agentzh dduncan: indeed
08:37 TreyHarris agentzh: no, i mean the one i already committed
08:37 agentzh TreyHarris: oh, i like it.
08:37 agentzh so i said TreyHarris++
08:38 juju` joined perl6
08:38 agentzh TreyHarris: have a sweet dream!
08:38 TreyHarris ah, ok.  i thought you had missed my change to t/README
08:38 TreyHarris thanks :-)
08:46 kane-xs joined perl6
08:46 svnbot6 r12362 | agentz++ | [t/builtins]
08:46 svnbot6 r12362 | agentz++ | fixed various broken smart links in the tests. (s/Perl6::Str/Str/g)
08:58 svnbot6 r12363 | agentz++ | added t/TASKS which holds tasks for the test suite.
08:59 * agentzh has no idea why catalog_tests.pl didn't help fixing broken links.
09:01 shachaf_ joined perl6
09:01 svnbot6 r12364 | agentz++ | [t/builtins]
09:01 svnbot6 r12364 | agentz++ | fixed more broken smartlinks in the tests (s/Perl6::Hashes/Hash/g, s/Perl6::Arrays/Array/g)
09:03 agentzh suppertime &
09:03 agentzh left perl6
09:24 xinming joined perl6
09:31 TimToady_ joined perl6
09:43 dduncan left perl6
09:53 foo\ joined perl6
10:02 chris2 joined perl6
10:17 pmurias joined perl6
10:17 pmurias gaal: ping
10:20 gaal hi pmurias
10:20 pmurias hi
10:20 pmurias have you seen hasktags?
10:21 pmurias included with ghc
10:22 gaal it didn't work for me when i tried it, don't remember why
10:24 pmurias having problems with it to
10:24 gaal how much RAM does make profiled need? I keep getting OOM on Pugs.AST.Internals in -prof mode
10:28 * pmurias has too little RAM, to compile even unoptimised in sensible time
10:51 pmurias gaal: i had sort(1) the tags file and add !_TAG_FILE_SORTED to it
10:51 gaal hasktags or ctags?
10:52 gaal if the former may I suggest you write a patch :-)
10:53 gaal unlines $ "!_TAG_FILE_SORTED" ++ sort $ lines whateverMakesTags
10:53 pmurias the former
10:54 gaal though it's slightly complicated by the fact that recent versions have an '-a'ppend mode
10:55 gaal whew, with 600m allocated and 128m swap 'make profiled' managed to overcome P.A.I
10:55 audreyt rehi
10:55 gaal hey audreyt!
10:55 pmurias hello
10:56 gaal no progress on the Var bug
10:56 gaal I should look into buying some more ram for this machine, clearly 1gb is not enough :/
10:57 pmurias discovered a python runtime in yhc
10:58 gaal wow, the graph for prime search on http://www-users.cs.york.ac.uk/~ndm/yhc/ looks impossible in favor of ghc
10:58 pmurias i.e. a virtual machine for haskell coded in python
10:58 lambdabot Title: yhc - York Haskell Compiler
10:58 Aankhen`` joined perl6
10:58 gaal of course it can't stay linear forever :)
10:59 agentzh joined perl6
11:00 plisk joined perl6
11:01 audreyt the OUTER problem turns out to be false alarm
11:01 audreyt my $say = -> Str $msg { say $msg }; sub f { $say(1,4,5) }; f(1,2,3)
11:01 audreyt doesn't work because param/arg mismatch
11:01 audreyt but the error msg could be improved
11:01 gaal indeed
11:01 gaal though...
11:02 gaal in absence of another say, shouldn't the list be coerced to a Str?
11:02 gaal audreyt: how long does make profiled take on your box?
11:03 gaal here it's quite prohibitive :(
11:03 audreyt from scratch? 30min or so
11:03 audreyt but "make fast profiled" should be faster
11:03 TimToady joined perl6
11:03 pmurias audreyt: how much RAM do you have?
11:04 audreyt 2g
11:04 gaal oh, I wasn't aware that worked: I thought I saw -O hardcoded into the makefile rule
11:04 audreyt "profiled" uses -O
11:04 audreyt because measuring non-optimised profile make very little sense
11:04 audreyt totally different behaviour
11:04 * gaal nods
11:05 evalbot_12364 joined perl6
11:05 gaal though if all you need is a stacktrace...
11:05 audreyt add a fastprof target?
11:06 gaal * explosion?
11:07 audreyt nah
11:07 audreyt done
11:08 gaal arg, beat me to it
11:08 lisppaste3 joined perl6
11:08 evalbot_12365 joined perl6
11:08 svnbot6 r12365 | audreyt++ | * add "make prof" and "make fastprof" as synonyms to
11:08 svnbot6 r12365 | audreyt++ |   "make profiled" and "make unoptimised-profiled".
11:08 ludan joined perl6
11:11 gaal is the pugs used by evalbot compiled fast or opt?
11:11 audreyt opt
11:13 gaal didn't evalbot have a feature of lazy nick changes?
11:14 audreyt it does. I'm just testing the build
11:15 gaal k. once pugs remakes here I'll try to improve that error message (dump the capture!)
11:15 gaal dump the infered Sig from a concrete Capt
11:16 agentzh gaal, audreyt: how about adding a script file docs/Perl6/Spec/update which can sync the synopses by invoking docs/Perl6/Makefile.PL?
11:16 agentzh i think it's a good shortcut.
11:16 agentzh and also an update.bat for win32 users.
11:16 svnbot6 r12366 | audreyt++ | * Eval.Var: improve the errmsg a bit by pretending it's
11:16 svnbot6 r12366 | audreyt++ |   always a multi match failure (which it is.)
11:17 gaal :-)
11:17 agentzh my util/smart_links.pl will use that. :)
11:17 gaal agentzh: dunno, is sv{k,n} up so bad?
11:17 gaal ah
11:17 gaal then sure :)
11:17 gaal or maybe:
11:17 agentzh hehe
11:18 gaal change that Makefile.PL to do the update at make time, not mf generation time
11:18 gaal and call the rule in your script?
11:18 gaal not sure
11:19 agentzh maybe put the code to Spec/update, and let Makefile.PL invoke ``update'' instead.
11:19 gaal sure
11:19 agentzh okay
11:21 SubStack joined perl6
11:24 gaal audreyt: can you explain how withExternalCall works? how the fmap err works?
11:25 audreyt sure. but dinner. briefly:
11:26 audreyt err :: b -> Maybe a -> Either b a
11:26 gaal ah ok, that explains it :)
11:26 gaal it's like assert, a bit
11:26 audreyt yup
11:27 audreyt bbiab :)
11:29 evalbot_12366_ joined perl6
11:29 xinming_ joined perl6
11:29 mdiep joined perl6
11:33 Aankhen`` joined perl6
11:35 evalbot_12366 joined perl6
11:36 audreyt ?eval 1..10
11:36 evalbot_12366 (1, 2, 3, 4, 5, 6, 7, 8, 9, 10)
11:43 nothingmuch joined perl6
11:43 svnbot6 r12367 | agentz++ | [docs/Perl6]
11:43 svnbot6 r12367 | agentz++ | * added script Spec/update which can sync the synopses from the svn repos
11:43 svnbot6 r12367 | agentz++ | removed the syn-updating code from Makefile.PL and make Makefile.PL call Spec/update instead
11:43 svnbot6 r12367 | agentz++ | added Spec/update.bat which is a Win32 shortcut for the Spc/update script.
11:52 lichtkind joined perl6
11:57 tifo joined perl6
11:59 weinig joined perl6
12:01 svnbot6 r12368 | agentz++ | added util/smart_links.pl which serves a successor of util/catalog_tests.pl. this script is still under development but you can already use it to verify the smartlinks used in a test file or the whole test suit. comments are very welcome.
12:02 agentzh s/suit/suite/
12:02 bpphillips joined perl6
12:06 agentzh s/serves a/serves as a/
12:07 svnbot6 r12369 | agentz++ | [util/smart_links.pl]
12:07 svnbot6 r12369 | agentz++ | added support for the old L<<...>> syntax.
12:07 svnbot6 r12369 | agentz++ | better error messages
12:18 xinming_ is now known as xinming
12:19 Limbic_Region joined perl6
12:23 kane-xs seen audreyt
12:23 jabbot kane-xs: audreyt was seen 46 minutes 48 seconds ago
12:33 agentzh hmm, iblech seems to have written many tests for pugs.
12:35 agentzh seen agentzh
12:35 jabbot agentzh: agentzh was seen 2 minutes 10 seconds ago
12:35 agentzh hehe
12:36 nothingmuch iblech++
12:36 tifo q
12:36 tifo \quit
12:36 agentzh nothingmuch: a lot of legacy smartlinks in the test suite are broken...
12:36 agentzh tifo: "/quit"
12:37 agentzh nothingmuch: the synopses are changing everyday...
12:37 nothingmuch agentzh: yes... there's no really good way to link
12:38 nothingmuch maybe we should use X<> nodes in the synopses
12:38 nothingmuch X<feature_Name>
12:38 agentzh currently i'm fixing them via my util/smart_links.pl.
12:38 agentzh nothingmuch: the problem is even feature names can be changing. :)
12:38 svnbot6 r12370 | agentz++ | fixed many more broken smartlinks in the test suite with help from util/smart_links.pl
12:39 nothingmuch yes
12:39 nothingmuch but that bit of text is more noticable
12:39 nothingmuch which means that if someone changes tie
12:39 nothingmuch the X<>
12:39 nothingmuch they can change the links pointing to it
12:39 agentzh nothingmuch: currently i'm introducing "keywords list" instead of plain regexes in the smartlinks.
12:39 nothingmuch X<> is for keywords
12:39 nothingmuch so i think that's a consistent idea =)
12:40 nothingmuch they simply don't show up in the formatted output
12:40 agentzh i think key words are more reliable and more descriptive than a single regex.
12:40 nothingmuch it's like <a name=..."></a>
12:40 nothingmuch yes, i think you are correct
12:40 * nothingmuch needs a break from the computer
12:40 agentzh nothingmuch: i was actually thinking about your suggestion,
12:40 nothingmuch too much sysadmin for the day
12:41 nothingmuch either way, it looks like you're taking it in the right direction =)
12:41 nothingmuch agentzh++
12:41 agentzh but introducing feature names is adding a new level of abstraction,
12:41 agentzh which may be not a good thing.
12:41 nothingmuch *shrug*
12:41 agentzh nothingmuch: thanks.
12:41 nothingmuch every approach has it's drawbacks
12:41 nothingmuch it's just a tool to be aware of
12:41 agentzh nothingmuch: quite right.
12:42 nothingmuch anymoose, ciao!
12:42 agentzh bye
12:46 FurnaceBoy joined perl6
12:46 mauke_ joined perl6
12:48 svnbot6 r12371 | agentz++ | [t/TASKS]
12:48 svnbot6 r12371 | agentz++ | added fixing broken smartlinks to the test suite's TASKS.
12:51 bpphillips left perl6
12:54 l33t_h4x0r joined perl6
12:54 l33t_h4x0r i kno more about computas than u all im da best hacker eva
12:55 mauke_ great, do you want a committer bit?
12:55 Limbic_Region ok - i'll bite, how about finishing parrot off for us this month so we can all focus on something else?
12:55 l33t_h4x0r i kno more about computas than u all im da best hacker eva
12:56 gaal is p = np?
12:56 Limbic_Region s/parrot/parrot, ponie, perl6, pugs/
12:56 mauke_ l33t_h4x0r: patches welcome
12:56 l33t_h4x0r k
12:57 bpphillips joined perl6
12:58 l33t_h4x0r left perl6
12:59 * agentzh has no idea which language 133t_h4x0r is using.
12:59 FurnaceBoy it's some dialect of Idiot :)
12:59 agentzh hehe
13:01 mauke_ is now known as mauke
13:03 agentzh bedtime &
13:03 agentzh left perl6
13:09 b_jonas joined perl6
13:13 fglock joined perl6
13:16 pmurias hi
13:19 jferrero joined perl6
13:29 weinig is now known as weinig|work
13:30 fglock hi
13:45 vel joined perl6
14:09 b_jonas left perl6
14:12 svnbot6 r12372 | audreyt++ | * Unicode as hash keys now work again.
14:23 hexmode joined perl6
14:34 cjeris joined perl6
14:38 penk joined perl6
14:42 elmex joined perl6
14:57 mdiep joined perl6
14:58 plisk joined perl6
15:13 MacVince joined perl6
15:15 mdiep_ joined perl6
15:24 ajs joined perl6
15:26 kanru joined perl6
15:27 marmic joined perl6
15:31 svnbot6 r12373 | audreyt++ | * Reduce the verbosity level of (show ClassTree) a bit.
15:33 justatheory joined perl6
15:33 svnbot6 r12374 | audreyt++ | * Fix infinite loop in SUPER::new dispatch.
15:33 svnbot6 r12374 | audreyt++ | * Also refactoed Eval.Var with implicit parameters, so we can
15:33 svnbot6 r12374 | audreyt++ |   more easily get a stack trace when something goes wrong, with
15:33 svnbot6 r12374 | audreyt++ |   "make fastprof" and then "./pugs-prof t.pl +RTS -xs".
15:34 TimToady audreyt++
15:35 audreyt :)
15:40 audreyt TimToady: try smoking again?
15:40 TimToady already compiling
15:40 audreyt ooh.
15:41 TimToady and have some hope of leaving it running while I shower, due to audreyt++
15:41 TimToady I used to have to baby sit it for SUPER::new
15:41 TimToady s:6th/<sp>//
15:43 TimToady and the compile takes a while since I always to optimized.
15:43 audreyt *nod*
15:43 TimToady *do
15:44 tifo joined perl6
15:44 discordja joined perl6
15:44 TimToady build failed, paste soon
15:44 audreyt it should be noticeably faster, I think. unless there's some hidden infinite-loop-ish things that I missed...
15:44 audreyt oy. ok
15:45 audreyt feather is still building, so I havn't seen 6.4.x compile log yet
15:45 pasteling "TimToady" at 71.139.13.50 pasted "build fail" (9 lines, 579B) at http://sial.org/pbot/19073
15:46 audreyt hm, 6.4 wants all contextual variables to be mentioned by name. even unused
15:47 audreyt thinking
15:52 cm joined perl6
15:55 audreyt essentially reverted.
15:55 audreyt (infinte loop fix is still there)
15:56 audreyt r12375 should build on 6.4 now
15:56 svnbot6 r12375 | audreyt++ | * So much for contextual variables and dynamic binding --
15:56 svnbot6 r12375 | audreyt++ |   GHC 6.4's mutual recursion limitation forced us to put
15:56 svnbot6 r12375 | audreyt++ |   Eval.Var.findSub into a deeply nested recursive group again.
15:56 svnbot6 r12375 | audreyt++ |   /me pines for GHC 6.6...
15:58 svnbot6 r12376 | fglock++ | v6 - 'state' array and hash
15:59 rlb3 joined perl6
16:00 TimToady hang on, parrot broke its build and I'm working around it...
16:01 * audreyt meanwhile fixes pugs -CJS
16:02 [particle] timtoady: same problem from yesterday?
16:02 TimToady no, 4 missing files from WMLScript
16:02 TimToady MANIFEST check upchucks
16:02 [particle] fixed in 14179
16:02 [particle] (aka HEAD)
16:03 TimToady okay, that's recenter than a couple minutes ago. :)
16:03 [particle] you can also work around by C<perl Configure.pl --nomanicheck>
16:03 TimToady k, thanks.
16:03 TimToady is the fix to delete the lines from MANIFEST or add the files?
16:04 [particle] the fix is to regenerate MANIFEST :)
16:04 svnbot6 r12377 | audreyt++ | * fix JS prelude generation to take type sigil into account
16:04 TimToady ok, then my patch will svn up smoothly, thanks.
16:05 [particle] the files were moved from foo/src/bar to foo/bar
16:06 TimToady np, I still had my vi session, so I undid.
16:06 audreyt TesT::Builder on 020.t still loops. fixing
16:06 [particle] there's always svn revert, too
16:06 TimToady that implies an understanding of svn.  :)
16:09 TimToady smoking, getting a lot of pugs: Internal error: Sigilless var: "linux"
16:12 audreyt hrm.
16:14 TimToady looks like Test-Builder is still looping.
16:14 audreyt just fixed. committing
16:14 audreyt re sigillss var: does it occur with rules?
16:14 audreyt (as I'm not seeing it here, and embparrot is only possibility I can think of)
16:16 TimToady not sure, error message is uninformative, must do divide-n-conquer
16:18 audreyt I'm getting it too. fixing
16:18 TimToady minimal test case in one file: use File::Spec; use Config::Tiny;
16:19 TimToady so down in one of those files
16:19 iblechbot joined perl6
16:19 audreyt $?OS triggers it.
16:19 TimToady hence "linux"
16:19 TimToady (for me)
16:19 jferrero joined perl6
16:20 audreyt aye
16:20 audreyt recompiling to make sure
16:28 svnbot6 r12378 | audreyt++ | * Pugs.hs: "run" in ghci now works again.
16:28 svnbot6 r12379 | audreyt++ | * Pugs.Parser: Allow ::Foo::Bar as well as ::Foo.
16:30 TimToady Test::Builder::Test.new doesn't seem to be finding its constructor.
16:30 svnbot6 r12380 | audreyt++ | * Pugs.Eval: Improve the "pugs -d" stack trace to be strict.
16:30 svnbot6 r12381 | audreyt++ | * Pugs.hs: Enable "pugs -d" again
16:31 audreyt r12382 should be good for resmoke.
16:31 audreyt hand-run TesT::Builder tests now pass
16:32 audreyt bug was that ::Test::Builder wasn't recognized as type literal
16:32 svnbot6 r12382 | audreyt++ | * Pugs.Eval.Var: $?OS and $?MODULE no longer triggers
16:32 svnbot6 r12382 | audreyt++ |   "sigilless var" errors again.
16:32 audreyt (only ::Test was)
16:33 Psyche^ joined perl6
16:35 TimToady smoking, my hand test also passes, so I'm probably good.
16:35 audreyt yay
16:35 audreyt TimToady++
16:35 * audreyt goes sleeping
16:36 TimToady sleep hard.
16:36 SageLT joined perl6
16:36 svnbot6 r12383 | fglock++ | v6 - initial implementation of 'constant'
16:36 svnbot6 r12383 | fglock++ | PCR - token <null>
16:38 Eimi joined perl6
16:41 bradleym joined perl6
16:41 svnbot6 r12384 | audreyt++ | * builtins/perl.t and builtins/perl2.t: "\7" and "\123"
16:41 svnbot6 r12384 | audreyt++ |   forms are now verboton.
16:42 Aankh|Clone joined perl6
16:45 Psyche^ is now known as Patterner
16:45 svnbot6 r12385 | fglock++ | PCR - moved "what's implemented" list from Pod to TODO
16:45 svnbot6 r12386 | lwall++ | \229 -> \d229
16:48 Aankh|Clone joined perl6
16:48 svnbot6 r12387 | audreyt++ | * Remove the "reduce 1st chunk by 1/3" heuristics
16:48 svnbot6 r12387 | audreyt++ |   used by parallel yaml_harness.pl, for recent
16:48 svnbot6 r12387 | audreyt++ |   refactorings has rendered ext/ tests roughly as
16:48 svnbot6 r12387 | audreyt++ |   fast as t/ ones.
16:53 bernhard joined perl6
17:01 weinig|work is now known as weinig
17:10 zgh joined perl6
17:12 amnesiac joined perl6
17:15 svnbot6 r12388 | fglock++ | v6 - 'Expression' code cleanup
17:16 baest joined perl6
17:18 Limbic_Region joined perl6
17:18 svnbot6 r12389 | fglock++ | v6 - code cleanup
17:32 Limbic_Region ?eval my @foo = 1..4; @foo.end
17:32 evalbot_12366 is now known as evalbot_12389
17:32 evalbot_12389 3
17:33 wolverian ?eval (1..4).tail
17:33 evalbot_12389 Error: No compatible subroutine found: "&tail"
17:33 wolverian noo :(
17:33 wolverian ?eval &tail ::= &postfix:<[ ]>.assuming(1..)
17:33 evalbot_12389 Error:  unexpected ")" expecting comment or term
17:34 wolverian erm
17:34 Limbic_Region what's the method to get the count again?
17:34 wolverian ?eval &tail ::= &postfix:<[ ]>.assuming(1...)
17:34 evalbot_12389 pugs: *** Undeclared variable: "&postfix:[ ]"     at <eval> line 1, column 11-40
17:34 Limbic_Region ?eval my @foo = 1..4; @foo.count
17:34 evalbot_12389 Error: No compatible subroutine found: "&count"
17:34 wolverian ?eval ("a".."c").elems
17:34 svnbot6 r12390 | audreyt++ | * from_perl6_rules/*.t: save some time by adding a =begin END
17:34 svnbot6 r12390 | audreyt++ |   before disabled tests.
17:34 evalbot_12389 3
17:34 Limbic_Region thanks
17:34 wolverian np
17:34 Limbic_Region ok - what about this
17:35 Limbic_Region for @array -> $elem { $elem.index }   # anything like that possible?
17:36 TimToady ?eval for @array.kv -> $i, $elem { say $i }
17:36 evalbot_12389 Error: Undeclared variable: "@array"
17:36 Limbic_Region heh
17:36 TimToady ?eval for [1,2,3].kv -> $i, $elem { say $i }
17:36 evalbot_12389 OUTPUT[0 1 2 ] undef
17:37 Limbic_Region TimToady - I was thinking more along the lines of an aliased array element automagically knowing stuff so that
17:37 TimToady too much overhead
17:37 Limbic_Region $elem.delete or $elem.index etc would just work
17:37 ayrnieu limbic - what else would it know?
17:37 SageLT joined perl6
17:37 Limbic_Region TimToady - not thinking in core
17:37 Limbic_Region thinking - possibilities
17:37 TimToady taking into account the possibilities induces overhead
17:38 TimToady better if they're declared upfront
17:38 Limbic_Region so in essense, you would have to resort to @foo being an array of tied $elems
17:38 TimToady on top of which, an object can be a member of more than one array simultaneously
17:39 * Limbic_Region mumbles something about that darn quantum entanglement
17:39 Limbic_Region ;-)
17:39 TimToady if you're going to privilege the array in the for loop, it might as well be explicitly declared there somehow
17:40 * [particle] plays a 10-dimensional violin for Limbic_Region
17:40 TimToady if I recall, this is all gone through in the original apocalypse...
17:40 TimToady since this was one of the RFCs
17:40 * Limbic_Region wasn't around then
17:41 TimToady well, there are probably more important READTHEMs than that...
17:42 TimToady and I don't mind repeating myself, up to a point.
17:42 TimToady might even be a 10-dimensional point.
17:43 Limbic_Region well, I think of #perl6 as being more of a playground and the list as being more serious - if I ask on the list it is because I have done my homework and still don't have a clue
17:43 Limbic_Region here - just musings
17:45 Limbic_Region ?eval my @foo = 1..5; my @bar = @foo.elems; ~@bar
17:45 evalbot_12389 is now known as evalbot_12390
17:45 evalbot_12390 "5"
17:46 PerlJam Limbic_Region: only the Synopses are serious.  Anything else is just play.  :-)
17:46 PerlJam And even then the Synposes are only provisionally serious
17:47 * Limbic_Region mumbles something about still thinking .elems should behave differently in list context
17:48 Limbic_Region hrm - but then again, why
17:48 Limbic_Region ?eval my @foo = 1..5; my @bar = @foo; ~@bar
17:48 evalbot_12390 "1 2 3 4 5"
17:49 Limbic_Region ?eval my $str = "1234"; $str.chars;
17:49 evalbot_12390 4
17:49 TimToady ?eval my @foo = 1..5; say @foo.int;
17:49 evalbot_12390 OUTPUT[5 ] Bool::True
17:49 TimToady .elems is actually redundant.
17:49 Limbic_Region ?eval my $str ="1234"; my @foo = $str.chars; ~@foo
17:49 evalbot_12390 "4"
17:49 Limbic_Region that's the example I was thinking of
17:49 TimToady ?eval my @foo = 1..5; say +@foo;
17:49 evalbot_12390 OUTPUT[5 ] Bool::True
17:49 TimToady doubly so
17:50 TimToady rather, Numly so
17:50 * Limbic_Region agrees that .elems is redundant and was thinking of .chars
17:52 TimToady ?eval my $str = "one small step"; say @str.words.int
17:52 evalbot_12390 Error: Undeclared variable: "@str"
17:52 TimToady ?eval my $str = "one small step"; say $str.words.int
17:52 evalbot_12390 Error: No compatible subroutine found: "&words"
17:52 TimToady unimpl
17:52 TimToady words(/./) should give you .chars
17:53 Limbic_Region which synopsis would I find words in? S29?
17:53 TimToady words being an insided out split
17:53 TimToady it'd be S29
17:53 TimToady .words defaulting to .words(/\S+/)
17:54 Aankh|Clone is now known as Aankhen``
17:54 TimToady but maybe we should pun on .split and call it .splat or some such...
17:54 elmex_ joined perl6
17:54 TimToady .spilt, drive the lysdexics nuts...
17:55 Limbic_Region hrm - there apparently isn't an S29 in the pugs repo - *shrug*
17:55 TimToady Specs/Functions
17:56 Limbic_Region and the S29 in the pugs repo is the most up to date one or is Rod still working it?
17:58 TimToady you mean the Spec/Functions.pod?  Rod hasn't worked on it in a long time.
17:58 TimToady it's semi-official in that I've more or less blessed it as official, but it's in pugs so other people can still hack on it easily.
17:59 * Limbic_Region can't find it in Pugs but was sure he saw it there previously
17:59 TimToady the next major change is probably to do way with protos like push() and make them exported methods.
18:00 TimToady docs/Perl6/Spec/Functions.pod
18:00 Limbic_Region ahh, I found it
18:00 Limbic_Region Spec ne Specs
18:00 TimToady need your specs examined
18:01 Limbic_Region actually I do - I can't remember the last time I had an eye exam
18:01 TimToady well, feel free to examine my specs...
18:05 Limbic_Region I do that occassionally but unless you watch every single patch as it is applied on the list - you really need to re-read the whole thing every few weeks
18:06 PerlJam Limbic_Region: amen!
18:07 PerlJam I feel like perl6 has slipped from under me because I haven't kept up with all of the details of the Synopses.
18:07 shachaf joined perl6
18:07 PerlJam Luckily, the language is designed well enough that broad brush strokes can still paint a good picture.
18:08 TimToady trying to butter me up, are ya? :)
18:08 TimToady you must want something...
18:09 PerlJam Just a perl6 by christmas :)
18:09 TimToady *vigorous nod*
18:10 * SamB wonders if PerlJam is trying to say that Perl6 is fat
18:10 TimToady biab &
18:12 lidden joined perl6
18:16 svnbot6 r12391 | fglock++ | v6 - more cleanup; @a.elems; %a.elems
18:32 TreyHarris can anybody tell me why this works:
18:32 TreyHarris ?eval for 1..* -> $i { say $i; last if $i > 5 }
18:32 evalbot_12390 is now known as evalbot_12391
18:32 evalbot_12391 OUTPUT[1 2 3 4 5 6 ] undef
18:33 TreyHarris but this does not:
18:33 TreyHarris ?eval my @a = 1..5; @a[1..*]
18:33 evalbot_12391 (no output)
18:33 TreyHarris (I know they both should eventually work, I'm just having trouble finding where to look in the pugs source to tell why one does and one doesn't)
18:34 Limbic_Region TreyHarris - I am surprised evalbot didn't whine about being out of memory
18:34 PerlJam Limbic_Region: I don't think it ran out of memory.
18:34 Limbic_Region the first one works because 1..* is acting as an iterator
18:34 PerlJam Limbic_Region: I think it "worked", only the return value got lost
18:34 Limbic_Region PerlJam - I know, which is why I am suprised
18:34 TreyHarris ?eval my @a = 1..5; say @a[1..*]
18:35 evalbot_12391 (no output)
18:35 Limbic_Region odd
18:35 TreyHarris i don't think so.  if i run it on my own, it takes constantly growing memory and crashes
18:35 TreyHarris but evalbot kills off things that are taking to long or growing too large
18:36 shachaf_ joined perl6
18:36 hexmode joined perl6
18:36 TreyHarris no, actually its memory usage plateaus, but it spikes the cpu
18:37 TimToady ?eval [1..*].elems
18:37 evalbot_12391 pugs: out of memory (requested 1048576 bytes)
18:37 [particle] whee!
18:37 TimToady should just say Inf
18:38 TreyHarris But @a[1..*] should be evaluable on a finitely-sized @a, right?
18:38 TimToady presumably.
18:38 TreyHarris oh, wait, i guess not:
18:38 TimToady * is an operator that is given meaning by it's subsequent execution context.
18:38 TreyHarris ?eval my @a = 1..5; @a[1..8]
18:38 evalbot_12391 [2, 3, 4, 5, undef, undef, undef, undef]
18:39 TreyHarris so it's trying to create an infinite number of undefs
18:39 TimToady and 1..* should probably pass the Whateverness on to the subscripting
18:39 TreyHarris for some reason i thought *... yes, what TimToady said.
18:39 TreyHarris is that behavior above correct, though?  i asked for eight elems, so i'm going to get eight elems, padded with undefs?
18:40 TimToady yes
18:40 TreyHarris it's only * that should autosize
18:40 [particle] how about '* but finite'? ;)
18:40 TimToady ?eval my @a = 1..5; @a[1..8]:v
18:40 evalbot_12391 Error:  unexpected ":" expecting term postfix, comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
18:40 TimToady ?eval my @a = 1..5; @a[1..8]:val
18:40 evalbot_12391 Error:  unexpected ":" expecting term postfix, comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
18:41 TimToady notimpl, probably
18:41 TimToady ?eval my @a = 1..5; @a[1..8]:p
18:41 evalbot_12391 Error:  unexpected ":" expecting term postfix, comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
18:41 TimToady ?eval my @a = 1..5; @a[1..8]:kv
18:41 evalbot_12391 Error:  unexpected ":" expecting term postfix, comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
18:41 TimToady the subscript modifiers are supposed to "grep" for existing values
18:42 [particle] colon, not dot?
18:42 [particle] modifier, not method.
18:43 TreyHarris ok.  yes, i see t/operators/subscript_adverbs.t
18:43 svnbot6 r12392 | fglock++ | v6 - added some scalar functions
18:43 TimToady S02.pod:    @array[0,1,2]:kv;   # returns 0, 'A', 1, 'B'
18:43 TimToady S02.pod:    @array[0,1,2]:k;    # returns 0, 1
18:44 TreyHarris i hope Test::todo gets implemented soon, it's hard to know now what's a notimpl versus bug
18:47 Limbic_Region TreyHarris - that's tantamount to volunteering so we will all be waiting for you to finish implementing it
18:48 TreyHarris Limbic_Region: I know. :-)  unfortunately, I have no time to be anything but a dilettante for the next couple weeks.  writing a test here and there will be my only contribs till then
18:50 [particle] TreyHarris: i hope you can wait a couple weeks for Test::todo then ;)
18:51 TreyHarris [particle]: well, agentzh intimated that he was going to work on it this weekend, so maybe i can dodge :-)
18:56 * Limbic_Region needs to find a window of time tomorrow to write a better examples/ tester for the test suite since he "volunteered" to do so sometime this week
18:58 ruz joined perl6
18:58 obra hey ruz.
19:00 shachaf__ joined perl6
19:06 fglock joined perl6
19:11 mdiep_ joined perl6
19:13 hexmode joined perl6
19:22 svnbot6 r12393 | kudra++ | Add summary YAMLs for May, June and July
19:24 * nothingmuch cheers for kudra's first commit
19:44 lichtkind Perl6::Bible is horribly out of date, but i found some quirks there
19:45 lichtkind in http://search.cpan.org/~autrijus/Perl6-Bible-0​.30/lib/Perl6/Bible/E05.pod#Interpolate_ye_not...
19:45 lambdabot Title: Exegesis_05 - Pattern Matching - search.cpan.org
19:45 lichtkind the second src code box is wrong
19:48 TreyHarris lichtkind: AFAIK, the Apocalypses and Exegeses are frozen as historical documents
19:53 Odin- joined perl6
19:53 TimToady if you're going to read As and Es, read the versions on dev.perl.org, since they've been annotated with [Update:] boxes.
20:02 svnbot6 r12394 | audreyt++ | * Thoroughly refactored Pugs.Parser.Operator's
20:02 svnbot6 r12394 | audreyt++ |   currentFunctions and related code, so the logic of
20:02 svnbot6 r12394 | audreyt++ |   which functions are relevant to parsing is made much
20:02 svnbot6 r12394 | audreyt++ |   more apparent.
20:02 svnbot6 r12394 | audreyt++ |   Also use interning, memoizing, sharing etc along the
20:02 svnbot6 r12394 | audreyt++ |   way, so parsetime is yet again reduced by ~10% or so.
20:03 TimToady smokin'
20:03 * Limbic_Region wonders how many tests are still failing that should be passing
20:03 * Limbic_Region wonders how long it takes TimToady to smoke
20:03 TimToady 'bout 28 minutes after compilation finishes.
20:04 Limbic_Region out of curiosity - how many failing tests last smoke?
20:04 TimToady my last two smokes when from 880 errors to 530.
20:05 TimToady early yesterday it was 219
20:05 Limbic_Region ok - bad assumption to think each check-in will result in fewer failures since new fixes may introduce new problems
20:06 * Limbic_Region slaps his wrists
20:06 TimToady compiles
20:07 TimToady so give it 30 minutes and it'll be on pugscode.org
20:07 lichtkind TreyHarris its a synopsis
20:07 TimToady well, actually, m19s28.vlinux.de
20:08 TimToady lichtkind: no, E05 would be an Exegesis.
20:08 TimToady and see my remark above.
20:09 lichtkind äh yes your right :)
20:10 lichtkind TimToady did something earthshakin changed since your talk at yapc israel acording to perl6 design principles?
20:11 Limbic_Region TimToady - I think you are the only one regularly smokin linux
20:11 TimToady that's my impression, though I think there used to be nightlies from feather
20:12 TimToady lichtkind: depends on how you define earthshaking.
20:12 TimToady I don't think we'd invented Captures yet.
20:13 lichtkind yes i heard somthing like that but i dont mean syntax, more the theory behind it since im still on chapter 2 of this perl6 tutorial
20:13 PerlJam "perl6 tutorial"?
20:14 bpphillips left perl6
20:14 PerlJam with chapters no less?
20:14 lichtkind PerlJam what you exactly asking me?
20:15 Limbic_Region he is asking what you mean by Perl 6 tutorial - link or reference
20:15 PerlJam lichtkind: I'm boggling that a perl6 tutorial even exists
20:15 PerlJam lichtkind: This isn't merlyn and bdf's thing is it?
20:15 Limbic_Region does it refer to one of the presentation titles given at YAPC israel, etc
20:16 lichtkind yes it does
20:16 * Limbic_Region stops taking words out of PerlJam's mouth
20:16 lichtkind i wrote it in german and its half ready
20:16 lichtkind and audrey gave me commiter bits
20:16 lichtkind so i will rewrite it in english
20:16 lichtkind its in our wiki
20:17 lichtkind the german version
20:17 lichtkind i wasnt aware of that its merlin domain
20:17 lichtkind i thought skud was master of doku
20:18 lichtkind i dont want to pretend to much but im willing to write it :)
20:18 lichtkind and im not stupid :)
20:19 lichtkind PerlJam did i answer your question ?
20:19 Limbic_Region lichtkind - please, write it
20:19 PerlJam lichtkind: sure ... where's the english version wiki so we can hack it too ?  :)
20:20 lichtkind no in wiki is just german because its the german perl community wiki
20:20 Limbic_Region PerlJam - "[16:16] <lichtkind> so i will rewrite it in english" future tense
20:21 Limbic_Region lichtkind - putting a placeholder english version out there might aid in making it a collaborative effort even if you haven't got the time to work on it yourself - perhaps the german perl community might want to take it on together
20:22 Limbic_Region and then us proper english speaking folks can add, modify, etc ;-)
20:22 lichtkind in german wiki the other only corrected misspellings, i think the main work copy should be english since there are more active people
20:22 PerlJam Limbic_Region: I saw the future tense.  I switched to present tense as a subtle form of encouragement  ;)
20:23 lichtkind but you now my first try was informative
20:23 nothingmuch aaah
20:23 nothingmuch Syncing http://svn.openfoundry.org/pugs
20:23 nothingmuch Retrieving log information from 12102 to 12394
20:23 nothingmuch Incomplete data: Delta source ended unexpectedly
20:23 lambdabot Title: Revision 12394: /
20:23 lichtkind for me at last  how to do it :)
20:23 nothingmuch clkao: HELP HELP HELP
20:23 PerlJam lichtkind: how many chapters in this tutorial?
20:23 PerlJam nothingmuch: try #svk
20:24 _abc_ joined perl6
20:24 nothingmuch i think it's a pugs svn problem
20:24 nothingmuch not really svk
20:25 * PerlJam svk pulls pugs to see what happens
20:25 lichtkind PerlJam: preknoledge and history,, theory, variables, operators, control structure, subroutines, objects, rules,  the first 5 are nearly complete with an index as Appendix A and Apendix B-E:
20:26 PerlJam lichtkind: when you get to subroutines, you'll have to talk about Captures at some point.
20:26 * justatheory feels nearly complete
20:27 cmarcelo joined perl6
20:28 jferrero joined perl6
20:28 lichtkind PerlJam shure where can i read about it ? i will give a Perl6 talk in 3 weeks so maybe i it will even therefore usefull
20:29 Limbic_Region 16:12] <TimToady> I don't think we'd invented Captures yet.
20:29 * Limbic_Region ducks
20:29 lichtkind :)
20:29 PerlJam lichtkind: http://dev.perl.org/perl6/doc/design​/syn/S06.html#Argument_list_binding might be a good place to start
20:29 lambdabot Title: Synopsis 6: Subroutines -
20:29 PerlJam lichtkind: just grep the synopses for "Capture"
20:30 lichtkind k
20:31 cmarcelo nothingmuch: svk sy -s 12103 //mirror/pugs or (i think) update your SVN::Mirror..
20:31 lichtkind PerlJam and can you give me an impression how does parrot going ?
20:31 nothingmuch cmarcelo: yes,
20:31 nothingmuch thats what i did
20:31 nothingmuch i just wanted clkao to rescue me
20:31 nothingmuch i like being a damsel in distress
20:31 nothingmuch and since he's an official mountain ninja, he's a good candidate
20:31 nothingmuch clkao++
20:33 PerlJam lichtkind: Seems to be going well from what I can see. Some core people (leo, chip, Allison) have changes jobs within the parrot project but it all appears to be for the better.
20:33 PerlJam lichtkind: pmichaud is working on a big PGE/TGE/PAST/POST refactor
20:34 lichtkind PerlJam yes that was my feeled impression too but im not on dayly basis involved
20:34 lichtkind who is pmichaud?
20:34 PerlJam lichtkind: perl6 compiler pumpking
20:34 lichtkind and whats his real name?
20:35 PerlJam nothingmuch: what was the fix?  My "svk pull" finally gave me the "Incomplete data" message too
20:35 PerlJam lichtkind: Patrick Michaud
20:35 nothingmuch PerlJam: what cmarcelo  said
20:35 PerlJam oh, I just missed that.
20:35 PerlJam thanks.
20:36 PerlJam lichtkind: also, there's lots of work being done on the tcl compiler and we've got a new person making imcc better (vsoni)
20:36 _abc_ left perl6
20:36 lichtkind PerlJam PerlJam im a bit knew in this whole its petty i cant see you all, often its good, not to see faces but i really bit interested who you all are (except larry :) which seen listen often )
20:37 lichtkind what is imcc?
20:37 PerlJam imcc is what lets us use PIR rather than the low-level PASM
20:39 lichtkind thanks
20:40 lichtkind sub bar ($a,$b,$c,:$mice) { say $mice }, that prints $a $b $c ?
20:40 gaal lichtkind: there's a FAQ. http://svn.openfoundry.org/pu​gs/docs/Perl6/FAQ/Capture.pod
20:41 lichtkind thanks
20:44 lichtkind PerlJam from where you are come from?
20:45 PerlJam what do you mean exactly?  Where do I live? or what part to I play in this whole parrot/perl6 story?
20:46 * [particle] thinks PerlJam is a fly on the wall somewhere in texas
20:46 [particle] ;)
20:46 PerlJam "a fly"?  Hmm.  I'm not sure if that's insulting or not.   Maybe I'm a lizard on rock instead  :)
20:49 lichtkind PerlJam both im interest what is important to  you , just want to know with whom i talk :) sorry but i could not open private channel, i have ICQ or Skype
20:49 PerlJam Although thinking of me as "a fly on the wall" conjurs images of a small camel with wings perched on Larry's shoulder or something.
20:50 PerlJam lichtkind: nah, public is fine.  :-)
20:50 lichtkind :)
20:51 PerlJam lichtkind: I'm a programmer who is interested in parrot/pugs and will occasionally (rarely) submit patches.
20:52 PerlJam lichtkind: I do live in Texas  :)
20:52 pmichaud perljam lives in a remote part of Texas :_)
20:52 pmichaud not the "remotest" part, but pretty remote :-)
20:52 PerlJam lichtkind: and I'm always willing to discuss things on #parrot, #perl6 because that's how I'm going to learn and remember this stuff :)
20:53 PerlJam lichtkind: oh, and I may have an article in the next issue of the Perl Review on parrot PMCs.
20:53 pmichaud PerlJam++
20:54 PerlJam pmichaud: I'm not sure if youre implying that I should move to the sticks or to a real city  ;)
20:54 pmichaud PerlJam: you're in the sticks already.  :-)
20:55 pmichaud but I wasn't implying anything about relocation :-)
20:55 pmichaud at least, not this time
20:55 PerlJam heh
20:55 PerlJam pm: btw, have I mentioned how annoyed I am at computer services?
20:56 svnbot6 r12395 | fglock++ | v6 - s/Regex/Token/ in 'Term' - Test.pm not compiles in <10s
20:56 jferrero joined perl6
20:57 PerlJam pm: james davis got his @tamucc.edu mail forwarded to lighthouse and the message said "this is temporary until we hear from above about folding lighthouse mail into the main campus system"
20:57 pmichaud heh
20:57 pmichaud was this from a worker bee or from someone "in the know"?
20:58 PerlJam pmichaud: I don't quite know how to classify Kay
20:58 PerlJam I guess she's more worker-beeish
20:58 pmichaud my opinion probably doesn't belong on a public channel :-)
20:59 baest_ joined perl6
21:00 fglock joined perl6
21:00 fglock joined perl6
21:02 svnbot6 r12396 | lwall++ | Getting rid of \040 and such
21:03 ruz joined perl6
21:14 bradleym is now known as bradleym|afk
21:29 TreyHarris can anyone remind me of what the difference is in Perl 5 between Foo::Bar->method() and Foo::Bar::->method()? (notice double colons before arrow in second case)
21:32 TimToady I don't believe I've ever seen the latter form.
21:32 xinming TimToady: I think he mean Foo::Bar::method
21:32 TimToady no, he explicitly said "before arrow"
21:32 TreyHarris at morgan stanley, i remember it was considered proper practice to call constructors as Class::->new, but I can't for the life of me remember why
21:33 xinming *_*
21:33 TreyHarris they seem to do precisely the same thing
21:33 lollan joined perl6
21:33 TimToady forces it to be not officially a bareword
21:33 TimToady so is unnecessary on anything that already contains ::
21:34 TimToady well, latest smoke failed 2 more tests :( but passed 278 more :)
21:35 TimToady and pugs is definitely a bit zippier now
21:37 TimToady TreyHarris: in fact, if I recall, the trailing :: is removed by the lexer.
21:37 TimToady print + FOO::, "\n";
21:37 TimToady FOO
21:39 TreyHarris something voodooy, i guess.  or maybe there was one particular rev that had some odd breakage and it turned into cargo culting
21:40 TreyHarris thanks, the answer is "i don't care and don't need to provide for a semantic difference", which is what I wanted it to be :-)
21:40 TreyHarris TimToady++
21:40 PerlJam When will perl6 be ready for production?
21:40 * PerlJam wants the answer to by "tomorrow"
21:40 * PerlJam wants really really hard
21:40 PerlJam ;)
21:41 TreyHarris PerlJam: according to audreyt's yapc talk, it already is in production, for some value of production
21:41 PerlJam "pugs" ne "perl6"
21:41 TreyHarris oh wait, you said perl6, lowercase, no space.  so you mean the parrot compiler?
21:42 PerlJam TreyHarris: no, if pugs is fast enough an officially blessed, pugs could become perl6 and that would be fine by me.
21:42 PerlJam s/an/and/
21:42 [particle] and ran on enough platforms...
21:42 TreyHarris ah. :-)
21:42 TreyHarris & for a meeting...
21:42 PerlJam But you got the answer you wanted, so I thought I'd see if that worked for me too :)
21:42 * [particle] hammers nails &
21:43 PerlJam granted, getting the answer and the answer being correct are different things
21:50 svnbot6 r12397 | Sage++ | In the parser for the p5 -> p6 translator: Uni blocks can now deal with the end of file (eof) much more gracefully.
21:50 svnbot6 r12397 | Sage++ | This was needed to handle many of the files in perl5 t/pod, where long comments often end a file.
21:50 svnbot6 r12397 | Sage++ | With this update, a test across all of p5 t/ has 95%+ coverage.
21:53 mdiep_ joined perl6
21:54 weinig joined perl6
22:11 svnbot6 r12398 | fglock++ | v6 - cleanup, fixes in grammar
22:12 TreyHarris *grin* somebody at $job.emaillist<perl> said "Perl == TIMTOWTDI.  Or should that be Perl eq TIMTOWTDI?"  I responded that Perl eqv TIMTOWTDI, and maybe currently Perl =:= TIMTOWTDI.
22:15 fglock left perl6
22:30 pmurias_ joined perl6
22:38 lichtkind sorry PerlJam got guest here im an east german freelancer makin also editor in perl
23:12 clkao stevan++ # the autobox thing is bloody scary
23:21 TreyHarris == clkao.  stevan++
23:22 clkao we need hot girl hacker rather than stupid handsome spokeperson
23:23 TimToady ethnic would help too.
23:33 svnbot6 r12399 | Sage++ | Better debugging info for the Text Node translator. The text node translator now gives more informative debugging information.
23:47 TreyHarris i've noticed folks syntax-highlighting their Perl 6 code in their blogs.  is that done by hand, or is there a perltidy-ish tool that knows Perl 6?
23:54 cjeris left perl6
23:56 Khisanth well there is a vim syntax highlighting thing in the repo and vim can output HTML of the highlighted version ...
23:56 TreyHarris what's going on here:
23:56 TreyHarris ?eval sub foo (: $a) { say "I got '$a'." }; foo(32)
23:56 evalbot_12391 is now known as evalbot_12399
23:56 evalbot_12399 OUTPUT[I got ''. ] Bool::True
23:57 TreyHarris Khisanth: ahh, ok.  can you tell me how?  i'm not a vim user :-)  If it's hard to explain, don't worry about it
23:58 Khisanth I use gvim :)
23:58 TimToady looks like a bug to me.  a sub shouldn't allow an invocant.
23:59 TimToady maybe it's being parsed as :$a though.
23:59 TimToady ?eval sub foo (: $a) { say "I got '$a'." }; foo( :a(32) )
23:59 evalbot_12399 OUTPUT[I got '32'. ] Bool::True
23:59 TimToady yep.
23:59 TimToady shouldn't allow a space there.
23:59 Khisanth TreyHarris: under the Syntax menu there is a "convert to html" option

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

Perl 6 | Reference Documentation | Rakudo