Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2008-07-25

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:13 pugs_svn r21523 | lwall++ | [STD] globalize $DEBUG
00:19 alester joined #perl6
00:21 pugs_svn r21524 | Auzon++ | [gsoc_spectest] moved one-pass-parsing test into spec. Added tests for regex creation syntax. (added 11 tests, fudged for rakudo)
00:22 Auzon both S02-whitespace_and_comments/one-pass-parsing.t and S05-metasyntax/regex.t can be added to spectest_regression
00:37 Alias_ joined #perl6
00:59 ruoso joined #perl6
01:13 sirhc joined #perl6
01:15 spencer_ joined #perl6
01:16 stephen joined #perl6
01:17 stephen :D   ^_^~~~
01:18 sirhc left #perl6
01:19 stephen Ingy dot Net is on the left side of the room.   ^_^~  yah for Ingy.
01:21 ua joined #perl6
01:25 ua Hello St. Louis!!  Whoo!  Yeah, Bon Jovi!
01:27 justatheory joined #perl6
01:37 agentzh left #perl6
01:44 meppl good night
01:52 agentzh joined #perl6
01:52 a-jing joined #perl6
01:52 a-jing left #perl6
01:53 ruoso are we live?
02:07 nothingmuch joined #perl6
02:17 alester joined #perl6
02:28 s1n joined #perl6
02:28 s1n rakudo: ^0.1
02:28 p6eval rakudo 29733: RESULT[0..^0.1]
02:29 s1n rakudo: [^0.1]
02:29 p6eval rakudo 29733: RESULT[[0]]
02:33 kanru joined #perl6
02:43 a-jing joined #perl6
02:43 a-jing left #perl6
02:51 s1n rakudo: ("d".."a").perl
02:51 p6eval rakudo 29733: RESULT["\"d\"..\"a\""]
02:51 s1n rakudo: ("d"^.."a").perl
02:51 p6eval rakudo 29733: RESULT["\"d\"^..\"a\""]
03:16 pugs_svn r21525 | s1n++ | [pugs_s02,s03] moved the auto-reversing range tests to the S03-operators section. cleaned up the
03:16 pugs_svn r21525 | s1n++ | S02-builtin_data_types to mostly test the auto-reversing ranges can be created. filled in a few
03:16 pugs_svn r21525 | s1n++ | single boundry tests that seemed missing.
03:37 pbuetow_ joined #perl6
05:03 zch051383471952 joined #perl6
05:08 xinming_ joined #perl6
05:34 xinming joined #perl6
05:43 eternaleye joined #perl6
05:53 Psyche^ joined #perl6
05:55 xinming_ joined #perl6
06:13 stephen joined #perl6
06:35 ashizawa joined #perl6
06:42 pugs_svn r21526 | moritz++ | [spec] range.t: fixed tests for decreasing ranges, fudged again for rakudo
06:53 rakudo_svn r29734 | moritz++ | [rakudo] three more test files for spectest_regression, Auzon++
07:07 moritz_ do we still have \N{CHAR NAME} escapes in Perl 6?
07:11 penk joined #perl6
07:17 rho joined #perl6
07:35 Mr`Michael joined #perl6
07:36 Mr`Michael left #perl6
07:36 renormalist joined #perl6
07:40 timbunce joined #perl6
07:41 _timbunce joined #perl6
07:53 iblechbot joined #perl6
08:03 pugs_svn r21527 | fglock++ | [PCR] added 'isGraphemeLink'
08:05 masak joined #perl6
08:22 timbunce joined #perl6
08:32 pugs_svn r21528 | fglock++ | [PCR] unicode fixes
08:49 explorer__ joined #perl6
08:51 ludan joined #perl6
08:55 pugs_svn r21529 | fglock++ | [PCR] version 0.34
08:58 meppl joined #perl6
09:10 pugs_svn r21530 | fglock++ | [v6.pm] 0.26
09:11 donaldh joined #perl6
09:17 cookys joined #perl6
09:20 agentzh heh, OpenFoundry is upgrading to RoR
09:22 moritz_ from what?
09:22 agentzh maybe Perl
09:22 agentzh not sure
09:22 moritz_ RubyOnSticks? ;-)
09:22 agentzh heh
09:27 chris2 joined #perl6
09:29 lumi joined #perl6
09:31 Patterner Ruby is on rails, so you can run over it with a train.
09:32 * moritz_ is reminded of sun's "data center in a container" which would also fit nicely on a train
09:33 wknight8111 joined #perl6
09:47 bacek joined #perl6
09:57 pmurias joined #perl6
10:02 AzureStone joined #perl6
10:04 bacek joined #perl6
10:09 penk left #perl6
10:11 a-jing joined #perl6
10:12 a-jing left #perl6
10:21 Jedai joined #perl6
10:24 snooper joined #perl6
10:38 iblechbot joined #perl6
10:43 Jedai joined #perl6
10:46 * zamanfou is now away: In the rain my tears are forever lost.
10:50 bacek_ joined #perl6
10:53 elmex joined #perl6
11:15 wknight8111 joined #perl6
11:19 agentzh left #perl6
11:26 rindolf joined #perl6
11:35 a-jing joined #perl6
11:35 a-jing left #perl6
11:43 meppl joined #perl6
11:47 pmurias joined #perl6
11:54 pbuetow_ joined #perl6
12:42 pugs_svn r21531 | pmurias++ | Array.elems and Array.postcircumfix:<[ ]> (called postcircumfix:<{ }> temporarly) work
12:54 jferrero joined #perl6
12:55 pmurias pugs: $*OUT.print("hello world\n");
12:55 p6eval pugs: OUTPUT[hello world␤]
12:59 masak perl6: $*OUT.close; $*OUT.print("hello world")
12:59 p6eval elf 21531: OUTPUT[Can't call method "close" on an undefined value at (eval 115) line 3.␤ at ./elf_f line 3861␤]
12:59 p6eval ..pugs: OUTPUT[*** Unsafe function 'IO::close' called under safe mode␤    at /tmp/7CitqEnr4n line 1, column 1-12␤]
12:59 p6eval ..rakudo 29736: OUTPUT[src/io/io.c:1047: failed assertion 'io'␤Backtrace - Obtained 15 stack frames (max trace depth is 32).␤  (unknown)␤    Parrot_confess␤      PIO_putps␤        Parrot_print_p_s␤          (unknown)␤            (unknown)␤              (unknown)␤                (unknown)␤
12:59 p6eval ..   ...
13:06 orafu joined #perl6
13:10 orafu joined #perl6
13:11 cai joined #perl6
13:12 cai left #perl6
13:19 a-jing joined #perl6
13:19 a-jing left #perl6
13:27 cmarcelo joined #perl6
13:55 xinming_ perl6: "hello, print".say;
13:55 p6eval elf 21531: OUTPUT[hello, print␤]
13:55 p6eval ..pugs: OUTPUT[hello, print␤]
13:55 p6eval ..rakudo 29736: OUTPUT[hello, print␤]
13:56 xinming_ "hello, world!".say.say;
13:56 xinming_ perl6: "hello, world!".say.say;
13:56 p6eval elf 21531: OUTPUT[hello, world!␤1␤]
13:56 p6eval ..pugs: OUTPUT[hello, world!␤1␤]
13:56 p6eval ..rakudo 29736: OUTPUT[hello, world!␤1␤]
13:57 ruoso joined #perl6
14:03 iblechbot joined #perl6
14:04 rdice joined #perl6
14:16 pugs_svn r21532 | ruoso++ | [smop] test/14 passing finally! the object was being held in the namespace.
14:17 pugs_svn r21533 | ruoso++ | [smop] p6opaque mostly done, other features will be added as needed
14:18 Lorn joined #perl6
14:19 peeps[work] joined #perl6
14:24 jferrero joined #perl6
14:41 justatheory joined #perl6
14:43 Chillance joined #perl6
14:44 pugs_svn r21534 | fglock++ | [v6.pm] minor fixed in ranges, error messages
14:54 hercynium joined #perl6
14:56 pugs_svn r21535 | fglock++ | [v6.pm] added error msg
14:59 ebassi joined #perl6
15:12 kanru joined #perl6
15:15 Exodist joined #perl6
15:16 jferrero joined #perl6
15:16 pugs_svn r21536 | fglock++ | [v6.pm] more SKIP Unicode tests from cpan-tester reports
15:17 pugs_svn r21537 | fglock++ | [v6.pm] 0.027
15:52 rakudo_svn r29737 | pmichaud++ | [rakudo]: spectest-progress.csv update: 107 files, 1845 passing, 56 failing
15:56 timbunce joined #perl6
16:02 meppl joined #perl6
16:12 xinming_ TimToady: ping
16:14 xinming_ @tell TimToady hi, Since we can consider [infix_op] <a b c>; and raw version of foldl in haskell, Will there be a foldr version in perl 6? :-)
16:14 lambdabot Consider it noted.
16:16 xinming_ @tell TimToady Will we be able to asign some dynamic attr to the reduce operator? eg. [infix_op]:l <a b c>; or [infix_op]:r <a b c>;
16:16 lambdabot Consider it noted.
16:18 rhr reduce ops take their assocativity from the infix op inside
16:18 xinming_ @tell TimToady Or will there be a postfix version of reduce op?
16:18 lambdabot Consider it noted.
16:18 xinming_ rhr: How do you think of this? [-] (1 .. 3);
16:19 rhr so you'd have to define another infix op to change the assoc
16:19 xinming_ rhr: But In my understanding, That's not the best practise. :-)
16:20 rhr infix:<-> is left assoc
16:21 xinming_ rhr: yes, so, [-] (1 .. 3) will be ((1 - 2) - 3); But how about people wants to express it with (1 - (2 - 3)) ?
16:22 xinming_ I don't know if there will be confliction with current syntax for (1 .. 3) [-]; version
16:23 pmurias joined #perl6
16:23 pmurias ruoso: hi
16:23 rhr something like: multi infix:<-r> is assoc<right> ($a, $b) { $a - $b }; [-r] 1..3
16:25 rhr S06:1760
16:27 xinming_ rhr: In pod source?
16:27 xinming_ I think it is.
16:27 rhr http://irclog.perlgeek.de/perl6/today makes it a link :)
16:27 lambdabot Title: IRC log for #perl6, 2008-07-25
16:35 justatheory joined #perl6
16:40 eternaleye joined #perl6
16:41 alester joined #perl6
16:55 pugs_svn r21538 | lwall++ | [is-rw.t] s/\xa0/\x20/g
16:59 a-jing joined #perl6
16:59 a-jing left #perl6
17:00 TimToady I think such rare cases should me mediated by something like foldr {$^a - $^b}, 1,2,3
17:00 lambdabot TimToady: You have 3 new messages. '/msg lambdabot @messages' to read them.
17:00 TimToady @messages
17:00 lambdabot xinming_ said 45m 37s ago: hi, Since we can consider [infix_op] <a b c>; and raw version of foldl in haskell, Will there be a foldr version in perl 6? :-)
17:00 lambdabot xinming_ said 43m 33s ago: Will we be able to asign some dynamic attr to the reduce operator? eg. [infix_op]:l <a b c>; or [infix_op]:r <a b c>;
17:00 lambdabot xinming_ said 42m 14s ago: Or will there be a postfix version of reduce op?
17:02 TimToady or rreduce, or something human readable
17:03 TimToady either that, or there should be a metaoperator that reverses any infix operator
17:03 TimToady and then you could say [rev-] reverse 1,2,3
17:04 TimToady but I think a higher-order function is much better than complicating the operator syntax for something that will be rarely used
17:06 TimToady the point of huffman coding is not just to make everything short
17:08 xinming_ hmm, Ok, thanks, Why I think about this is because of the consistence. BTW, I think we have an op which toggles the association might be the best sollution to me.
17:09 xinming_ But thanks anyway for your explaination.
17:12 pmurias xinming_: you mean an metaop to toggle associativity?
17:13 xinming_ pmurias: yes.
17:13 xinming_ But I know that we normally don't need to toggle this. So, what TimToady said is correct. It's a rare case.
17:26 timbunce joined #perl6
17:28 alester joined #perl6
17:29 rindolf joined #perl6
17:39 stephen joined #perl6
17:50 justatheory joined #perl6
18:02 xinming_ TimToady: with 'our $.attr' vs 'my $.attr', Can we change the scope of these two? 'our $.attr' will be shared by it's class, and all subclasses. and 'my $.attr' will affect only the current class, when we call Class.attr, we will get the one defined with our, If there is 'my $.attr' for a specified class, then, that 'my $.attr' is accessed which overrides the 'our $.attr'.
18:03 * xinming_ just get some _random_ thoughts while reading the synopsis.
18:07 meppl joined #perl6
18:09 alester joined #perl6
18:14 * zamanfou is now away: In the rain my tears are forever lost.
18:19 pugs_svn r21539 | pmurias++ | [smop] removed io and hash from the boilerplate
18:19 pugs_svn r21539 | pmurias++ | Scalar does method delegation
18:20 pmurias xinming_: what would be it's use?
18:21 ruoso pmurias, hi
18:21 pmurias ruoso: hi
18:21 pmurias ruoso: what should i implement now?
18:21 xinming_ pmurias: hmm, I don't know, But since normally, my and our declarator are different. But with class attributes, they are almost the same which will confuse people IMO. :-)
18:23 ruoso pmurias, now we can take from the ROADMAP directly untill the point marked as 0.2.7
18:24 pmurias attributes then?
18:24 pmurias or Code?
18:26 pmurias the Code.new should take a slime frame?
18:30 araujo joined #perl6
18:42 pugs_svn r21540 | pmurias++ | [elfX] undo a accidently comitted not-working-out-of-the-box merge
18:44 peeps_work joined #perl6
18:52 ruoso pmurias, oh... if Hash and Array are done, could you please move them from the roadmap to the changelog?
18:53 ingy joined #perl6
18:56 pmurias ruoso: Hashes don't do any hashing, and i'm fixing Arrays to use [] now
18:56 rindolf Hi all.
18:57 ruoso pmurias, at this level, they just need to implement .{} correctly... they don't need to be too smart
18:57 ruoso .{} and .[] I mean
18:58 pugs_svn r21541 | pmurias++ | [s1p] arrays use postcicumfix:<[ ]>
18:59 lambdabot joined #perl6
19:01 pmurias ruoso: i update the changelog when i feel like fighting with openid again ;)
19:05 ruoso pmurias, I could update it for you ;)
19:05 pmurias anok
19:05 pmurias * ok
19:05 ruoso so, I'll push Hash and Array down...
19:06 ruoso pmurias, have you ever considered signing into ohloh?
19:07 ruoso (btw, I've just updated the changelog... could you please check if I wrote your name correctly?)
19:09 ruoso pmurias, about the lowlevel code... I'm not sure they can store a slime frame because slime frames are dependant of the C context...
19:10 ruoso I think we're going to need the generation of a callback function specific for that code instance
19:11 ruoso at least in s1p
19:13 * ruoso always feels good seeing "All tests successful"
19:14 pmurias what are the benefits of ohloh?
19:15 ruoso an "open source portfolio"?
19:15 ruoso I usually use it when I send a CV
19:15 ruoso and it's nice to see the "contributor map" of the project ;)
19:19 fridim_ joined #perl6
19:20 * ruoso converting test/14 into a benchmark test of the dispatching mechanism
19:27 ruoso SMOP takes 18.509s to execute 10000 dispatchings like the one in test/14, creating a new frame for each invocation...
19:28 pugs_svn r21542 | ruoso++ | [smop] first smop benchmark...
19:29 pugs_svn r21543 | fglock++ | [v6.pm] emitter hacks
19:30 pmurias ruoso: i registered on ohloh
19:30 ruoso :)
19:30 pmurias re slime frames are dependant on the C context, only at creation time
19:31 ruoso but they can't be re-used...
19:32 pmurias that's a problem
19:33 ruoso yes... that's why they're called SLIME
19:33 ruoso but SLIME is only 100x slower than p5+moose ;)
19:34 pmurias what?
19:34 ruoso test/dispatch_benchmark
19:34 * pmurias misread the number a 100% and was horrified
19:35 ruoso but I'm not quite sure the comparision is fair...
19:36 Alias_ joined #perl6
19:37 pmurias isn't p5+moose as fast as pure p5 for method calls?
19:37 ruoso nope... it's about 10x slower
19:37 justatheory joined #perl6
19:41 ruoso maybe not...
19:41 ruoso I think my benchmark was wrng
19:42 ruoso anyway...
19:43 timbunce joined #perl6
19:43 ruoso SLIME have this name because it's the SMOP Lame Interpreter Implementation
19:45 pmurias ruoso: we can either .clone frames *or* replace SLIME
19:46 pmurias we have only 4 sm0p blocks + 2 in sm0p tests
19:46 ruoso pmurias, that we can promote to v6 blocks
19:47 pmurias that was what i meant ;)
19:49 ruoso ;)
19:49 ruoso and we need a better interpreter implementation
19:49 ruoso ;)
19:51 jkeroes joined #perl6
19:52 ruoso something less lame
19:56 pmurias i hopped to write it in perl6, but if work is to be wasted on sm0p we can replace it sooner
19:57 ruoso maybe we could wait a little longer
19:57 ruoso as we're not creating more sm0p blocks
19:57 ruoso we're not making the problem bigger
19:57 ruoso we can replace the v6 blocks implementation later with a small effor
19:57 ruoso t
19:59 pmurias it's not worth to extend something which will be thrown away
20:02 ua_ joined #perl6
20:03 pmurias would there be an major issues with a llvm jit only interpreter?
20:05 pmurias if we won't get away with it i don't see why shouldn't we do a replacement c one
20:07 ruoso pmurias, I think we can neglect that issue for now... considering we have the eval polymorphism and we can replace the interpreter itself later
20:07 ruoso and also considering we are already writing v6 blocks, which are implementation independent
20:10 pmurias ruoso: do you know to implement Code withough extending slime?
20:11 ruoso pmurias, we can make it using a c callback (just like lowlevel method does)
20:14 pmurias that would mean emitting c from s1p :(
20:14 ruoso pmurias, oh... but s1p will be compiled down to C anyway
20:15 ruoso we'll implement the interpreter in s1p
20:15 ruoso then we don't need to generate C anymore
20:17 jkeroes joined #perl6
20:18 ruoso pmurias, if you take a look in the ROADMAP, you'll see that we have two major milestones before implementing the AST types
20:18 ruoso and only after that we'll have the effective interpreter
20:19 * ruoso will update the roadmap and replace KP6 by elf
20:20 ruoso pmurias, can I just s/KP6/elf/g? or is there any other important difference?
20:24 pmurias elf's grammar is in ruby so compiling it to smop would be a radically different task
20:24 ruoso hmmm...
20:25 ruoso maybe we could think about compiling STD to smop then
20:25 ruoso STD targets bootstrap, doesn't it?
20:25 pmurias yes
20:26 ruoso maybe we just go directly to STD
20:26 ruoso we use elf as the s1p compiler
20:27 ruoso pmurias, do you think that is possible?
20:28 ruoso (of course a lot of features will be needed before we actually get there...)
20:30 pmurias directly you mean instead of doing kp6?
20:30 pmurias or withough elf?
20:31 ruoso by using elf in s1p and port STD using s1p to bootstrap it
20:32 pmurias makes sense
20:33 ruoso so that part of the roadmap turns intp
20:33 ruoso * STD AST types
20:33 ruoso * STD AST -> runtime frames (effective interpreter)
20:34 ruoso * STD5 SMOP backend
20:34 ruoso * STD5 bootstraps STD in SMOP
20:34 ruoso * christmas ;)
20:35 ruoso pmurias, right?
20:35 pmurias those are all very big steps, but the basic order is right
20:36 ruoso cool...
20:36 ruoso pmurias, did I mentioned STD × STD5 correctly/
20:36 ruoso ?
20:36 ruoso I'm not completely familiar with them
20:37 ruoso (the difference between them)
20:37 pmurias STD5.pm was the name of the compiled to perl5 STD.pm
20:38 pmurias it is STD5.pmc
20:38 ruoso "was"?
20:38 pmurias i would just use STD everywhere
20:38 ruoso alright
20:39 pmurias s1p wiki page is totally outdated
20:40 ruoso indeed
20:40 ruoso pmurias, take a look into the updated rodmap "Perl 6 SMOP" milestone
20:43 pmurias seems ok
20:43 jkeroes left #perl6
20:47 smg joined #perl6
20:55 ruoso is it ok for the lowlevel attribute to support only scalar, array and hash as valid containers?
21:00 pugs_svn r21544 | lwall++ | [STD] parses 98% of t/spec
21:01 pugs_svn r21545 | lwall++ | [Cursor.pm] delete completely obsolete P6ish Cursor.pm
21:03 pmurias ruoso: why do you need to special case it?
21:03 ruoso pmurias, to avoid having to look into the string of the attribute name to discover the container type
21:04 ruoso hmm...
21:05 ruoso but thinking again...
21:05 ruoso I don't need to do it..
21:05 ruoso if I store simply the container "type"
21:05 ruoso I can call "new" on the "type"
21:06 jferrero joined #perl6
21:06 ruoso then the compiler sets the container protoobject into the attribute
21:07 * ruoso home &
21:10 hanekomu joined #perl6
21:11 hanekomu joined #perl6
21:25 iblechbot joined #perl6
21:44 Limbic_Region joined #perl6
21:52 pugs_svn r21546 | fglock++ | [v6.pm] emitter fixes
22:05 justatheory joined #perl6
22:20 Limbic_Region joined #perl6
23:11 dduncan joined #perl6
23:15 pmurias joined #perl6
23:17 pmurias @tell ruoso i think it's better to do Code before Attribute as it should be possible to do Attribute in perl6
23:17 lambdabot Consider it noted.
23:29 Alias_ joined #perl6

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

Perl 6 | Reference Documentation | Rakudo