pugscode.org/ | nopaste: sial.org/pbot/perl6 | pugs: [~] <m oo se> (or rakudo:, kp6:, smop: etc.) || We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by TimToady on 25 January 2008.
00:15 sri_work_ left 00:18 felipe joined 00:19 Chillance left 00:22 nipotan is now known as nipotaway 00:29 cookys left 00:31 elmex left
Auzon perl6: my $foo = 'a'; my $bar = 'b'; say <<$foo $bar>> 00:31
p6eval kp6 r20874: OUTPUT[syntax error at position 46, line 1 column 46:␤my $foo = 'a'; my $bar = 'b'; say <<$foo $bar>␤ ^ HERE␤]
..pugs: OUTPUT[ab␤]
..rakudo r28528: OUTPUT[Statement not terminated properly at line 1, near ">"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
..elf r20874: OUTPUT[$␤]
Auzon elf gets points for creativity. 00:32
00:32 mncharity joined 00:50 jan_____ left 00:54 jan__ joined, ruoso joined
meppl good night 00:58
00:59 meppl left 01:12 pbuetow left
pugs_svnbot r20875 | putter++ | [STD_red] Improve operator name handling. Unbreaks elf. 01:13
r20876 | putter++ | [elf] Cleaning up elf_f_src/ . Moved run-tests to elfish/run-tests/ . Pugs t/ svn revision added to output. 01:16
01:24 lichtkind joined
lichtkind perl6: $a = do { 3; leave; 5; } 01:24
p6eval kp6 r20876: OUTPUT[Can't locate object method "new" via package "Leave" (perhaps you forgot to load "Leave"?) at compiled/perl5-kp6-mp6/lib/KindaPerl6/Grammar/Control.pm line 843, <> line 1.␤]
..pugs: OUTPUT[*** ␤ Unexpected " ="␤ expecting "::"␤ Variable "$a" requires predeclaration or explicit package name␤ at /tmp/tbe8P2Mzgb line 1, column 3␤]
..rakudo r28528: OUTPUT[Scope not found for PAST::Var '$a'␤current instr.: 'parrot;PCT::HLLCompiler;panic' pc 156 (src/PCT/HLLCompiler.pir:103)␤
..elf r20876: OUTPUT[Useless use of a constant in void context at (eval 113) line 3.␤Undefined subroutine &GLOBAL::leave called at (eval 113) line 4.␤ at ./elf_f_faster line 4656␤]
pugs_svnbot r20877 | putter++ | [elf] Cleaning up elf_f_src/ . Moved some experimental files to a new elfish/elf_f_work/, and added README guidance on using them from there. 01:31
01:52 DietCoke left 01:54 SamB left 01:55 SamB joined
lichtkind i thought bare blocks dont execute in perl 6? 02:04
02:13 saladsn left
Coleoid . 02:17
rakudo: my $simple = 'begin'; say; 02:23
p6eval rakudo r28528 OUTPUT[␤]
Coleoid rakudo: .= 'begin'; say;
p6eval rakudo r28528 OUTPUT[Syntax error at line 1, near ".= 'begin'"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
Coleoid rakudo: $_ = 'begin'; say;
p6eval rakudo r28528 OUTPUT[too few arguments passed (0) - 1 params expected␤current instr.: '_block11' pc 0 (EVAL_10:3)␤
Coleoid Er,
rakudo: say 'begin' 02:25
p6eval rakudo r28528 OUTPUT[begin␤]
Coleoid rakudo: $_ = 'begin'; say $_
p6eval rakudo r28528 OUTPUT[too few arguments passed (0) - 1 params expected␤current instr.: '_block11' pc 0 (EVAL_12:3)␤
Coleoid rakudo: $_ = 'begin'; say $_; #perhaps a missing semicolon?
p6eval rakudo r28528 OUTPUT[too few arguments passed (0) - 1 params expected␤current instr.: '_block11' pc 0 (EVAL_10:3)␤
Coleoid ... 02:26
Okay, I'm getting interested in Rakudo given current news, yet this failure is pretty rudimentary. 02:32
And I don't know how to get from here, to seeing if there's a test for something like this already in the spectests... 02:33
...or something like this already in the bug tracking for Rakudo, which is RT? At some URL? 02:34
So I'm glad for any pointers I can get from the folks here. 02:35
Meantime, I'm searching for the pointers. Perhaps my difficulties are characteristic of multiple folks, and somehow this gap I'm yurking around in can be narrowed? 02:38
rakudo: say "Until later, thanks for the improvements!"; 02:39
p6eval rakudo r28528 OUTPUT[Until later, thanks for the improvements!␤]
bacek rakudo: my $a = 'hello'; say $a; 02:42
p6eval rakudo r28528 OUTPUT[hello␤]
bacek rakud: sub hello { say 'hello, ' ~ $_ }; hello('world') 02:43
rakudo: sub hello { say 'hello, ' ~ $_ }; hello('world')
p6eval rakudo r28528 OUTPUT[hello, world␤]
bacek Coleoid: there is no $_ at top-level module (yet) 02:44
Coleoid bacek: Thanks! Top-level module of Rakudo as a whole?
bacek Coleoid: hmm... I didn't quite understand your question... 02:45
Coleoid rakudo: sub inner { $_ = 'inner'; say }; inner(); 02:46
p6eval rakudo r28528 OUTPUT[too few arguments passed (0) - 1 params expected␤current instr.: 'inner' pc 33 (EVAL_11:21)␤
Coleoid I'm not sure what you mean by 'top-level'.
bacek rakudo: sub inner { $_ = 'inner'; say $_; }; inner();
p6eval rakudo r28528 OUTPUT[too few arguments passed (0) - 1 params expected␤current instr.: 'inner' pc 33 (EVAL_10:21)␤
Coleoid rakudo: sub inner { $a = 'inner'; say $a }; inner(); 02:47
p6eval rakudo r28529 OUTPUT[Scope not found for PAST::Var '$a'␤current instr.: 'parrot;PCT::HLLCompiler;panic' pc 156 (src/PCT/HLLCompiler.pir:103)␤
Coleoid rakudo: sub inner { my $a = 'inner'; say $a }; inner();
p6eval rakudo r28529 OUTPUT[inner␤]
Coleoid rakudo: sub inner { my $_ = 'inner'; say }; inner(); 02:48
p6eval rakudo r28529 OUTPUT[too few arguments passed (0) - 1 params expected␤current instr.: 'inner' pc 33 (EVAL_11:21)␤
bacek Coleoid: support for $_ is still rudimentary in rakudo. 02:49
Coleoid Well.
bacek rakudo: say <1..10>:sort { $^b <=> $a } 02:52
p6eval rakudo r28529 OUTPUT[Statement not terminated properly at line 1, near ":sort { $^"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
bacek rakudo: say <1..10>.sort { $^b <=> $a } 02:53
p6eval rakudo r28529 OUTPUT[Statement not terminated properly at line 1, near "{ $^b <=> "␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
Coleoid rakudo: say <1..10>.sort { $b <=> $a }
p6eval rakudo r28529 OUTPUT[Statement not terminated properly at line 1, near "{ $b <=> $"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
bacek rakudo: say <1..10>.sort({ $^b <=> $a })
p6eval rakudo r28529 OUTPUT[Scope not found for PAST::Var '$a'␤current instr.: 'parrot;PCT::HLLCompiler;panic' pc 156 (src/PCT/HLLCompiler.pir:103)␤
bacek rakudo: say <1..10>.sort({ $^b <=> $^a }) 02:54
p6eval rakudo r28529 OUTPUT[Method 'sort' not found for invocant of class 'Perl6Str'␤current instr.: '_block11' pc 45 (EVAL_12:22)␤
bacek what???
rakudo: say (1..10).sort({ $^b <=> $^a })
p6eval rakudo r28529 OUTPUT[Method 'sort' not found for invocant of class 'Range'␤current instr.: '_block11' pc 63 (EVAL_14:25)␤
Coleoid Saner!
bacek someone broke it...
rakudo sya sort { $^b <=> $^a }, 1..10; 02:55
rakudo: say sort { $^b <=> $^a }, 1..10;
p6eval rakudo r28529 OUTPUT[10987654321␤]
bacek rakudo: say ~sort { $^b <=> $^a }, 1..10;
p6eval rakudo r28529 OUTPUT[10 9 8 7 6 5 4 3 2 1␤]
bacek rakudo: say ~grep { $_ %2 }, sort { $^b <=> $^a }, 1..10; 02:56
p6eval rakudo r28529 OUTPUT[9 7 5 3 1␤]
Coleoid Here's a working '$_'...
02:57 Limbic_Region joined
bacek rakudo: say ~grep { $^yak % 2 }, sort { $^b <=> $^a }, 1..10; 02:57
p6eval rakudo r28529 OUTPUT[9 7 5 3 1␤]
Coleoid ...which works the same auto-named as a parameter. 02:58
bacek rakudo: say (grep { $^yak % 2 }, sort { $^b <=> $^a }, 1..10).reduce { $^a+$^b }
p6eval rakudo r28529 OUTPUT[Statement not terminated properly at line 1, near "{ $^a+$^b "␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
bacek pugs: say (grep { $^yak % 2 }, sort { $^b <=> $^a }, 1..10).reduce { $^a+$^b }
p6eval OUTPUT[*** ␤ Unexpected "{"␤ expecting operator, ":" or ","␤ at /tmp/H608OSTsWT line 1, column 63␤] 02:59
bacek rakudo: say (grep { $^yak % 2 }, sort { $^b <=> $^a }, 1..10).reduce{ $^a+$^b }
p6eval rakudo r28529 OUTPUT[too few arguments passed (0) - 2 params expected␤current instr.: '_block11' pc 0 (EVAL_14:3)␤
bacek rakudo: say reduce {$^a+$^b }, grep { $^yak % 2 }, sort { $^b <=> $^a }, 1..10
p6eval rakudo r28529 OUTPUT[25␤]
bacek pugs: say (grep { $^yak % 2 }, sort { $^b <=> $^a }, 1..10).reduce{ $^a+$^b } 03:00
p6eval OUTPUT[*** Cannot cast from VList [] to Pugs.AST.Internals.VCode (VCode)␤ at /tmp/XFMRn0YyIp line 1, column 6 - line 2, column 1␤]
diakopter woo hoo! I got the recursive-descent parser (* that can handle left recursive non-terminals, both direct and indirect) working! :D :D 03:19
03:19 meteorjay left
diakopter so now grammars can have left associate operators specified in the exact same syntax as right-associative ones. 03:20
without transforming
oh, and the packrat/memoization comes with that.
mncharity Auzon, et al: Is there an up-to-date set of regex tests? 03:25
pugs_svnbot r20878 | putter++ | [elf] Elf has had two variants, only very slightly different. The default one used Moose, the faster one not. The two have been merged, with the default one now not. elf_f_faster no longer exists - use the similar elf_f.
r20878 | putter++ | Motivation - the low startup latency is nice, and with a dearth of development resources, the time when Moose would most help is now distant, so supporting both doesn't currently seem worth the complexity cost.
mncharity diakopter: could you remind me... are you working on something which is strictly a p6 parser, or a general regex engine? 03:27
diakopter mncharity: well, it's turned into a general regex engine 03:29
mncharity have a test suite?
diakopter since that's what's needed to translate STD.pm into a readable Perl6in5::Grammar
mncharity: ha
I only *just* got the memoizing working 03:30
and the output format (ast and actions and such) has yet to be formulated fully
so in other words, there still would be nothing to test against, so no. :)
it's still untestable.
mncharity I find it very hard to tell how real, and the related - how usable(ie, fast), regex engine implementations are, without seeing some kind of test suite results. 03:31
re still untestable, ah, ok.
diakopter yeah...
mncharity packrat++ 03:32
diakopter well, my first guess about its speed is that it will be slower than most for most things, but faster than most on a few things... namely situations where naturally left recursive constructs are usually translated into a bottom-up parser system, or right-recursive equivalents
mncharity have you considered targeting p6 rather than p5? :)
diakopter mncharity: no, but I don't see where you're going with that :) 03:33
mncharity long term, it would be nice to have a grammar compiler which analyzed the grammar, and perhaps even path stats, and assembled an optimized implementation tuned for that specific grammar. 03:34
diakopter but that's only a slightly special case of standard compiler optimization steps....
diakopter thinks a bit on "path stats" 03:35
oh
I think what you mean by that is a profile of how interdependent various rules are... and which ones are used all the time. 03:37
mncharity re only slightly special, I suspect not. It's a rather specialized domain, and thus likely to have domain-specific optimizations possible.
diakopter for instance, the unspace rule has to run at nearly every character position :)
mncharity yes
diakopter mncharity: ok, I haven't read about those... urls? 03:38
or keyphrases?
mncharity for some grammars, just compiling to C, or Inline::C, could be a dramatic win. who needs yacc... 03:39
re urls, hmm, there's a very good question I don't have an answer for. Are there any wizzy survey papers on the current state of parsing?? Dunno. My fuzzy impression is aside from localized activity, eg, around packrat, or "attribute grammars" (some meanings), and functional programming folks, the field as a whole is off in book land. 03:42
ie, dead
diakopter heh
my next task is to figure out how to make operator precedence easier to declare in this grammar system 03:43
I think that should come before the tree generation/manipulation 03:44
mncharity on my not-short list of "things I hope p6 will accomplish" is shake up parsing land. there's really no excuse for not having good General Parsers readily available. but the pragmatics of getting adoption and support community has killed the attempts I've seen over the years.
diakopter mostly b/c TimToady kept challenging me on that
TimToady: thanks for challenging me so much 03:45
mncharity with concequences like ruby making language design decisions based on what bison can deal with. which pragmatically makes sense... but is an unfortunate place for us as a field to be.
re challenge, :) 03:46
diakopter also, he challenged me to table tennis
mncharity just to see "yet another way to do it", you might look at the redsix opp if you haven't already.
My fuzzy recollection is it could backtrack through the opp, and have n-part tokens (ala smalltalk), which might contribute interesting ideas. 03:47
re n-part, ie, rather than just x ?? x !! x , you could have arbitrary x A x B x C x D x ... with assorted precedences. 03:48
03:49 Limbic_Region left
diakopter mncharity: ah 03:49
Auzon mncharity: I'm not certain about the status of the regex tests
mncharity tries to remember how to restart the evalbot...
Auzon Most tests so far have been up to date
mncharity elf: say 3
p6eval OUTPUT[sh: ./elf_f_faster: No such file or directory␤]
Auzon evalbot control restart
03:49 p6eval left
mncharity evalbot control restart 03:49
03:49 p6eval joined
mncharity elf: say 3 03:50
diakopter patrick explained to yapc how he implemented ??!! in pge
mncharity elf: say 3
p6eval OUTPUT[3␤]
mncharity *nod*
Auzon: tnx :)
diakopter combination of unless and if) 03:51
but for detection, yeah, that's interesting.
mncharity Auzon: re tests, hmm, ok. so old | have been changed to || and | tests added?
Auzon looks 03:52
diakopter mncharity: ha, so you're saying A,B,C,D... could each have its own precedence.. :P or individually be optional... or be optional in weird groupings...
03:53 DietCoke joined
diakopter looks at DietCoke 03:53
mncharity :) at least for redsix, no optionality or weirdness. just n sub-tokens, each with their own left and right precedences, and some other tweakable constraints which seemed needed given the fuzzy understanding of how to parse p6 which existed at the time. 03:54
ie, whitespace sensitivity.
diakopter ah ok
so basically, all issues that a recursive descent parser (with whitespace sensitivity) understands 03:55
lichtkind perl6: my @a = gather { take 5; leave 5; take 3 } 03:56
mncharity err... only if you encode precedences in the tree. rd can't do opp otherwise, no?
p6eval kp6 r20878: OUTPUT[Can't locate object method "new" via package "Leave" (perhaps you forgot to load "Leave"?) at compiled/perl5-kp6-mp6/lib/KindaPerl6/Grammar/Control.pm line 843, <> line 1.␤]
..pugs: RESULT[[5, 3]]
..rakudo r28535: OUTPUT[Null PMC access in type()␤current instr.: 'parrot;Perl6;Grammar;Actions;statement_prefix' pc 115190 (src/gen_actions.pir:10907)␤
..elf r20878: OUTPUT[Type of arg 1 to Private::gather must be block or sub {} (not null operation) at (eval 117) line 5, near "})"␤ at ./elf_f line 3826␤]
diakopter mncharity: btw, I guess I didn't tell you, the grammar syntax for Perl6in5 (yap6) has whitespace-ness built into the pattern definitions... so you don't need a distinction betweeen rule/token
lichtkind seemes very wrong to me
i mean the result of pugs
diakopter lichtkind: 'leave' is known to be pugsbug 03:57
Auzon: are there many tests on 'leave'?
Auzon Yes.
lichtkind so result should be [5] oder [5,5] ?
Auzon see t/builtins/control_flow/leave.t 03:58
diakopter mncharity: yes, I meant handled by explicit encoding like that, yeah
I think it should still be handled that way, except without having to specify the extra "fall-through" rule at each layer. 03:59
in other words, nest the rules in an "opp()" parser generator, which does all the fall-throughs automatically.
mncharity with packrat to pick up the pieces... hmm... 04:00
diakopter that way, opp() can also generate the flexibility needed to support interlayer insertion
Auzon What's packrat?
diakopter packrat refers to a memoization technique of recursive descent parsers, by which they can achieve linear parse times on general input. except left-recursive stuff. except now it can. 04:01
thanks to those researchers who wrote the paper from which I implemented that algorithm
Auzon Oh, I thought it was some regex library or something. 04:02
mncharity a "packrat parser" basically remembers what happened when rule x was applied at position p... so when some other part of the grammar asks about it for the nth time, you don't have to recompute it for the nth time.
Auzon I remember you talking about such reductions in the past, but I'm not far enough in my CS studies to have seen it
mncharity figuring out how that interacts with a p6 grammar, where rules may be passing dynamically scoped information between themselves... may require a large set of grammars to yield a "I 04:03
'm sorry, don't know how to compile that for my rat"./
diakopter I did some informal tests.... for chains of nested left-associate rules (things handled by left-recusive rules), parse times still increase more quickly than linear... but rarely will you have 90 nested left associative things (where I started noticing the time increase) 04:04
04:04 lichtkind left
mncharity hmm... could still memoize everything which didn't play games, inside the framework of a parser which didn't depend on being able to memoize everywhere. 04:05
diakopter mncharity: but a grammar's generation can't be dependent on input
a modification to an existing grammar could...
but not the initial one
otherwise, please explain... 04:06
mncharity: there is a don't-memoize-me version of parser() called parser2() 04:07
mncharity if rule foo { <{let($x)=$global_mumble++;}> } then foo isn't idempotent, and can't be memoized.
or whatever the syntax for embedded code is. i'm fuzzy. 04:08
diakopter but that introduces what TimToady calls an immediate stopper
essentially ends the compilation unit
mncharity you mean for '|' "longest token match"?
diakopter <{}> 04:09
04:09 zhangsoft joined
mncharity ah, I didn't realize you were only going for linear on |, but not || (backtracking). 04:09
diakopter I'm still not clear on ||
mncharity p6 || is the p5 | 04:10
04:10 zhangsoft left
diakopter right, so, standard recursive descent, right? 04:10
mncharity it got renamed :/
diakopter standard alternation, I mean
?
mncharity re standard rd, modulo that it's past midnight and I'm increasingly fuzzy, I believe so. 04:11
yes
diakopter but | is now what?
mncharity re standard alternation.
re | ... I'm less than entirely clear. 04:12
diakopter I mean... standard alternate chooses "first match"
mncharity yes
diakopter || chooses "hungriest match"
mncharity err, no. 04:13
diakopter I mean
mncharity s/yes//
diakopter least hungry match
diakopter backtracks
er, backpedals
mncharity err, yes. saw your "I"(eye) as an '|' vertical bar. very end-of-dayish just now.
diakopter ok :D 04:14
ok... currently | in Perl6in5 does longest-match
mncharity "I mean" != "| ... <rest is definition of |>" sigh
diakopter in fact, so do '.' and '-' 04:15
there's also "first-match"
I guess there needs to be a symbol for that (since I obviously can't overload ||) 04:16
Auzon mncharity: I won't be handling regexes until the end of July. I can move it up on my schedule if you question the regex tests.
mncharity so to handwave fuzzily, yes, rx || is standard regular expression | alternation. and rx | is longest match... for some fuction "longest match" defined on rx primitives, which I'm quite fuzzy on.
*function
diakopter actually I just realized, without the left-recursive ability, the LTM function wasn't yet complete 04:17
but now it is
mncharity Auzon: re move it up, no need for me. optimize your own productivity. TimToady may need a test suite at some point, but I'm fuzzy on how general/complete his engine is supposed to be, or what the timeframe is. 04:18
Might be "would help now", or might be "unneeded". dunno.
diakopter but I can overload '|=' :P 04:19
mncharity Auzon: re move up, might check with TimToady and pm if that didn't already happen. 04:20
04:20 zhangsoft joined
diakopter oh! there is a postfix + that's overloadable... I wonder how overload returns its arguments to distinguish it from unary prefix... 04:20
mncharity s/pm/pmichaud/
04:20 nekobaka left
Auzon mncharity: I can easily swap S03 and S05 (so S05 would be in 1 week and S03 would be towards the end). I gave S0[2345] 1.5 weeks each because they're more detailed. 04:20
diakopter heh 04:22
I'd give S05 around 20 weeks
mncharity beware that rx's are their own language, of daunting complexity. I always underestimate it, even after applying several levels of "correct for always underestimating it" corrections.
Auzon diakopter: I guess, but I only have 12 weeks total.
diakopter Auzon: :P surely you won't leave us to go back to *school* :P 04:23
mncharity the pugs regex tests are by far the largest group of tests in the suite... and they are woefully incomplete.
Auzon diakopter: It seems that I will ;) but I'll still be around and helping. Just not for 20+ hours a week 04:24
mncharity end of day for me.
good night and have fun :)
Auzon good night, mncharity.
04:25 mncharity left, meteorjay joined
diakopter mncharity: oh, I thought longest match would just be longest number of characters.... 04:25
heh
I guess it could be "# of terminals" 04:26
actually
it has to be, highest number of characters eaten, with *fewest* number of terminals breaking ties. 04:27
04:28 lichtkind joined
diakopter TimToady: correct me where wrong, plz 04:28
lichtkind pugs: print if $_ % 2 for 1..10;
p6eval OUTPUT[*** ␤ Unexpected "for"␤ expecting operator␤ at /tmp/Te0HBAwUAv line 1, column 17␤]
diakopter audreyt: are you around to clarify LTM 04:31
diakopter ->(bed=>1) 04:36
04:38 lichtkind left 04:39 alester joined 05:04 Eevee left, Eevee joined 05:13 stevan_ joined 05:15 iblechbot joined 05:31 alester left 05:51 barney joined 05:52 barney left 06:07 DietCoke left 06:25 iblechbot left 06:29 wknight8111 left 06:50 Psyche^ joined 06:58 mncharity joined 06:59 _Maghnus joined
mncharity a very quick note. I was way too negative about the | to || rename. p5 re was a vast improvement over the diverse crud which went before, setting a new and better standard, and no doubt p6 rx will do it again. I look forward to working with them. 07:00
&
07:00 mncharity left 07:05 Patterner left, Psyche^ is now known as Patterner 07:10 stevan_ left, Maghnus left, _Maghnus is now known as Maghnus 07:17 snail joined 07:19 snail left 07:41 alanhaggai left 07:52 penk joined 07:56 luqui joined 08:01 elmex joined 08:06 alc joined 08:33 sri_work joined 08:34 sri_work left, sri_work joined
moritz_ perl6: say 'x' x 5 08:37
p6eval kp6 r20878: OUTPUT[xxxxx␤]
..pugs: OUTPUT[xxxxx␤]
..rakudo r28536: OUTPUT[xxxxx␤]
..elf r20878: OUTPUT[Undefined subroutine &GLOBAL::infix_x called at (eval 115) line 3.␤ at ./elf_f line 3826␤]
moritz_ elf: say 'x' x 5 08:40
p6eval OUTPUT[Undefined subroutine &GLOBAL::infix_x called at (eval 115) line 3.␤ at ./elf_f line 3826␤]
pugs_svnbot r20879 | moritz++ | [elf] added infix:<x>
moritz_ elf: say 'x' x 5 08:41
p6eval OUTPUT[Undefined subroutine &GLOBAL::infix_x called at (eval 115) line 3.␤ at ./elf_f line 3826␤]
moritz_ do I have to recompile elf_f?
moritz_ tries 'make bootstrap' 08:42
08:42 pmurias joined 08:43 pmurias left
moritz_ elf: say 'x' x 5 08:46
p6eval OUTPUT[xxxxx␤]
pugs_svnbot r20880 | moritz++ | [elf] recompiled after implementing infix:<x>
moritz_ YaY, first successful elf contribution
08:50 luqui left, meppl joined
pugs_svnbot r20881 | moritz++ | [spec] start adding a few regex tests. This first one tests that longest 08:52
r20881 | moritz++ | alternatives wins
r20882 | moritz++ | [spec] more tests for longest alternatives in regexes 09:01
09:06 Auzon left 09:13 bacek left
rakudo_svn r28537 | jonathan++ | [rakudo] Start refactoring 'does' to be more in line with STD.pm. This gets rid of trait_or_does, making does just be a trait_auxiliary. 09:26
pugs_svnbot r20883 | moritz++ | [spec] tests for single quotes (from from_perl6_rules) 09:40
09:43 Lorn left 09:47 pmurias joined
pmurias moritz_: re contribiution, where there any previous unsuccesfull ones? ;) 09:48
moritz_ pmurias: no, just unsucessful tries to build/run it (but mncharity++ was very helpful) 09:49
perl6: say [('a') xx 2].perl
p6eval kp6 r20883: OUTPUT[no method 'APPLY' in Class 'Undef'␤ at compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl5/MOP.pm line 345␤
..pugs: OUTPUT[["a", "a"]␤]
..rakudo r28537: OUTPUT[["a", "a"]␤]
..elf r20883: OUTPUT[Undefined subroutine &GLOBAL::infix_xx called at (eval 117) line 3.␤ at ./elf_f line 3828␤]
pmurias found your text task? ;) 09:50
moritz_ pmurias: perhaps ;) but I don't know if I can add multi infix:<xx> (@a, $b) is p5, because you can't pass an array as first arg in p5 09:52
pmurias moritz_: shouldn't xx be written in p6? 09:54
moritz_ multi sub infix:<xx>(@a, Int $b) { return gather { take @a for ^$b }} 09:55
that would even be lazy as soon as gather...take is implemented correctly 09:56
pmurias yes 09:58
you wrote a perl5 gather,take module if i recall correctly ;)
moritz_ pmurias: yes. Perl6::GatherTake 09:59
the only lazy one on CPAN - the others are all just syntactic sugar 10:00
pmurias infix:<x> couldn't have been implemented in perl6 too 10:12
it would be less efficient likely 10:13
moritz_ ENOSENSE
I think you have a superfluous negation in that sentence ;) 10:14
pmurias moritz_: yes
moritz_ ok, then I know what you mean ;)
pmurias: should stuff like that go into Prelude.pm? 10:16
pmurias yes 10:17
moritz_ tries that
elf: say 2, 3 10:20
p6eval OUTPUT[2␤3␤]
10:20 rindolf joined
moritz_ that's wrong, actually ;) 10:21
perl6: say 2, 3
p6eval kp6 r20883: OUTPUT[23␤]
..pugs: OUTPUT[23␤]
..rakudo r28537: OUTPUT[23␤]
..elf r20883: OUTPUT[2␤3␤]
moritz_ bootstrapping... 10:23
pmurias moritz_: do you set STD_RED_CACHEDIR? 10:25
10:26 ruoso left
pugs_svnbot r20884 | moritz++ | [elf] 10:26
r20884 | moritz++ | * implemented infix:<xx> in Perl 6
r20884 | moritz++ | * fixed say. It should print just one newline, not one after each arg.
r20884 | moritz++ | Could probably be simplified even more
moritz_ elf: say (1, 2) xx 2 10:29
p6eval OUTPUT[Undefined subroutine &GLOBAL::push called at ./elf_f line 638.␤ at ./elf_f line 3835␤] 10:30
moritz_ ouch
maybe I should use the method form instead
10:30 wknight8111 joined
pmurias yes 10:31
10:34 Jedai joined
pugs_svnbot r20885 | moritz++ | [elf] use method form for push() in the Prelude 10:38
moritz_ elf: say (1, 2) xx 2 10:40
p6eval OUTPUT[Undefined subroutine &GLOBAL::push called at ./elf_f line 638.␤ at ./elf_f line 3835␤]
10:40 Ara4n joined, Ara5n left
moritz_ perl6: say 1, 2; 10:41
p6eval kp6 r20885: OUTPUT[12␤]
..pugs: OUTPUT[12␤]
..rakudo r28537: OUTPUT[12␤]
..elf r20885: OUTPUT[12␤]
moritz_ at least that's fixed
don't know why it doesn't like my push method, though
elf: say (1, 2) xx 2
p6eval OUTPUT[Undefined subroutine &GLOBAL::push called at ./elf_f line 638.␤ at ./elf_f line 3835␤]
moritz_ pmurias: any idea? 10:42
10:46 zhangsoft left 10:52 Auzon joined
pmurias moritz_: was afk 10:57
10:58 LCamel left, LCamel joined
pmurias moritz_: looks like elf_f was missing from commit 10:59
moritz_ pmurias: thanks, I forgot to make if_bootstrap_scceeded 11:00
pugs_svnbot r20886 | moritz++ | [elf] update elf_f
moritz_ elf: say (1, 2) xx 2
p6eval OUTPUT[1␤]
moritz_ elf: ((1, 2) xx 2).perl.say 11:01
p6eval OUTPUT[1␤]
moritz_ is stupid
still, even with my stupidity the result should bei 1,2 not 1 11:02
pmurias elf: ([1,2] xx 2).perl.say 11:03
p6eval OUTPUT[[1,2]␤]
pmurias elf: say ([1,2] xx 3)
p6eval OUTPUT[12␤]
moritz_ Maghnus: I'm about to fix that
sorry, pmurias
where did my tab-key go?
11:04 wknight8111 left
pmurias np 11:04
elmex: say 1 xx 6
sorry 11:05
elmex pmurias: 1
pmurias elf: say 1 xx 6
p6eval OUTPUT[1␤]
elmex elf: say sex
pmurias pugs: say 1 xx 6
p6eval OUTPUT[Undefined subroutine &GLOBAL::sex called at (eval 117) line 3.␤ at ./elf_f line 3835␤]
OUTPUT[111111␤]
moritz_ elf: [1,2] xx 2 11:06
pugs_svnbot r20887 | moritz++ | [elf] infix:<xx>: return multiplied list, not the original one ;)
p6eval RESULT[[[1,2],[1,2]]␤]
moritz_ elf: (1,2( xx 2
p6eval OUTPUT[Parse error in: /tmp/Po58p0JI2y␤panic at line 1 column 0 (pos 0): Can't understand next input--giving up␤WHERE: (1,2( xx 2␤WHERE:/\<-- HERE␤ STD_red/prelude.rb:98:in `panic'␤ STD_red/std.rb:76:in `scan_unitstopper'␤ STD_red/std.rb:224:in `comp_unit'␤ STD_red/std.rb:210:in `_U...
moritz_ elf: (1,2 xx 2
p6eval OUTPUT[Parse error in: /tmp/emnPBexGX2␤panic at line 1 column 0 (pos 0): Can't understand next input--giving up␤WHERE: (1,2 xx 2␤WHERE:/\<-- HERE␤ STD_red/prelude.rb:98:in `panic'␤ STD_red/std.rb:76:in `scan_unitstopper'␤ STD_red/std.rb:224:in `comp_unit'␤ STD_red/std.rb:210:in `_UN...
moritz_ d'oh
elf: (1,2) xx 2
p6eval RESULT[[1,1]␤]
pmurias elmex: elmex [1,2] xx 3
ef: [1,2] xx 3 11:07
elf: [1,2] xx 3
p6eval RESULT[[[1,2],[1,2],[1,2]]␤]
moritz_ I'm not the only one with either broken self or keyword ;)
pmurias keyword?
moritz_ s/word/board/ ;-) 11:08
omg
I should really just shut up for today, and avoid IRC like hell
11:08 donaldh joined
pmurias perltidy++ #making generated code readable 11:09
rakudo_svn r28538 | jonathan++ | [rakudo] Big refactor of package declration code. This brings us closer to STD.pm and also breaks one enormous method down into several smaller ones, which should make the code easier to extend and maintain.
11:14 alc left 11:15 mjk joined 11:19 jan__ left 11:22 tcliou_ left, penk left 11:23 c1sung left, tcliou joined, LCamel_ joined 11:24 LCamel left 11:26 meppl left 11:37 iblechbot joined 11:44 a-jing joined, a-jing left, chris2 joined 11:49 DietCoke joined 11:57 Psyche^ joined 12:01 redicaps joined 12:14 Patterner left 12:16 Psyche^ is now known as Patterner 12:20 pbuetow joined
pugs_svnbot r20888 | pmurias++ | [sm0p] actions work 12:21
r20889 | diakopter++ | [yap6] another interim checkin; I should've done this in a branch... 12:24
pmurias TimToady: would it be wise to overload operators on the matches (Cursors realy) returned from STD5? 12:26
12:26 kanru left 12:29 kanru joined
diakopter ponders n-ary operator rule constructors 12:29
12:47 |Jedai| joined
rakudo_svn r28540 | jonathan++ | [rakudo] Get !keyword_does to take a role PMC rather than a name, making it re-usable. Also switch from ident to name in a couple of places, which will help us get namespace things right later on. 12:47
12:50 lichtkind joined
lichtkind is an bare block executed? 12:50
syn says yes but a larry talk once said no 12:51
moritz_ lichtkind: it is context dependent
12:57 Lorn joined 13:05 Jedai left 13:16 lichtkind left 13:26 b_jonas joined 13:37 jan__ joined
pmurias ls 13:39
sorry
13:43 alanhaggai joined 13:50 wknight8111 joined 13:54 wolv joined, mjk left 14:05 wolverian left, wolv is now known as wolverian 14:08 chris2 left
pmichaud Auzon: (S05 tests) -- note that we have a fair number of tests in the parrot repo already (derived from the ones from Pugs) 14:14
the ones in the parrot repo are up-to-date.
at any rate, I'm not in any hurry for S05 tests -- I'd much rather have S03 tests earlier 14:15
14:16 alanhaggai left
DietCoke ~~ to pmichaud. 14:25
DietCoke found a picture on his phone of you, me and allison.
DietCoke guesses it was Kennedy.
pmichaud heh 14:26
must've been
DietCoke perljam? 14:27
pmichaud I doubt it. 14:28
14:28 r0bby left
DietCoke was looking for his email addy. 14:28
pmichaud oh
[email@hidden.address]
b_jonas take care with mail addresses, this channel has a public log on the web 14:29
rindolf b_jonas: so?
pmichaud yes, I know, but I also know his addr is public already :-)
DietCoke spam filters ho!
rindolf b_jonas: spammers also install malware to harvest addresses.
b_jonas: I already got 11 spam messages to my new GMail account.
b_jonas rindolf: oh sure, but it might still increase the volume
it's not like a yes or no thing
rindolf b_jonas: and they subscribe to mailing lists to harvest the addresses from there. 14:30
b_jonas: yes.
b_jonas sure, I use spam filters and stuff too
and still get spam
not surprisingly
rindolf b_jonas: I have a mailto:shlomif@iglu.org.il address on many places, and I can handle my spam using spam assassin.
p3m.org/pfn/1561 - what horrible code. 14:31
lambdabot Title: p3m.org/pfn paste from "udinesh5" at 122.166.34.242...
rakudo_svn r28546 | jonathan++ | [rakudo] Implement basic use of does operator for mixing in a role to an existing object at runtime. This was most neatly done with a dynop, so this check-in also adds a dynops file for Rakudo. make makefile will be required to build after this check-in.
b_jonas rindolf: sure, but you can't really know that about nother person
rindolf b_jonas: OK.
I think GMail's spam filter has many false positives.
Many ham messages are marked as spam there. 14:32
www.shlomifish.org/philosophy/philo...h-century/ - I made this accessible from the rest of the site today.
lambdabot Title: Why Closed Books are So 19th-Century, tinyurl.com/5gvmdl
b_jonas rindolf: probably yes 14:33
it's diffult to tell how many because I don't check all of the mails marked as spam
rindolf b_jonas: yes, neither do I.
14:33 |Jedai| left 14:35 cornelius_ joined 14:36 cornelius_ is now known as c9s
diakopter pmichaud: I'm wondering how much of S05:Longest Token Matching are implementation detail suggestions as opposed to specifications... 14:42
pmichaud I haven't looked at it that closely yet. But to me they have the feel of "specifications"
DietCoke S05? 14:43
... wrong bot.
pmichaud s/yet/lately.
diakopter www.perlcabal.org/syn/S05.html#Long...n_matching
lambdabot Title: S05
moritz_ b_jonas: at least the logger that I maintain obfuscate email addresses 14:45
b_jonas moritz_: ah, I didn't know that
nice
I hope it doesn't have much false positives
moritz_ the @ is replaced an @ image
pugs_svnbot r20890 | jnthn++ | [spectest] Some tests for mixing in roles with 'does'.
b_jonas ah, I see
moritz_ if it's iin something that looks like an email address 14:46
nobody complained about false positives for now
diakopter well, I'm starting to get a clear(er) picture of the goal for which the author is striving, but it seems that much of the description has to do with "how to ensure we don't leave anything hanging on the right side when parsing, when using a predictive parser backed by NFA..."
b_jonas there's a mailing list where I've once sent a shell script, lost the email, and can't get it back because the archives webpage removes something in it irreversibly that's not actually an email address but it thinks it is
diakopter pmichaud: plz see above. in other words, it seems to me that much of the logic/constructs described in this section aren't necessary when using a recursive descent parser. 14:48
14:48 revdiablo joined
diakopter that is, the specified behaviors can be provided in other (probably simpler) ways 14:48
pmichaud except I think the point of the description is to say "here's where regexes aren't doing recursive descent" 14:49
moritz_ b_jonas: don't think I fell into that trap. If somebody needs to have large excerpts from the logs and copy & paste doesn't help, I can just spit out some SQL for them ;)
rakudo_svn r28547 | jonathan++ | [rakudo] Add (passing) mixins test to Rakudo passing list.
diakopter pmichaud: thanks; I'll think on that and re-read with that perspective
pmichaud recursive descent isn't prohibited, it's just that one uses || instead of | when one wants the traditional rec-descent way of looking at things 14:50
but to get DFA longest-token matching semantics, one uses |
and in particular, I think that STD.pm relies on longest-token matching semantics to work properly. 14:51
b_jonas moritz_: good
diakopter pmichaud: but currently in yap6, | checks all the alternatives and returns the one that matches the most characters...? 14:53
pmichaud: though of course there can be another alternator that does first-match 14:54
pmichaud how would yap6 handle "if(2+3)" ? 14:55
diakopter pmichaud: I guess my point is, recursive descent doesn't necessarily imply first-match
14:55 toddr joined
pmichaud (doesn't imply first match) I agree. 14:55
diakopter pmichaud: it doesn't know about "if", but it does parse "say say 3" correctly.
(as say(say(3))) 14:56
pmichaud right... but I'm curious how one would distinguish "if(2)" from "if (2)"
diakopter oh
I've built into the alternation/concatenation operators knowledge about mandatory/disallowed/optional whitespace 14:57
so you just have to encode the rule correctly. is "if(2)" not allowed (must there be a space between if and ()? 14:58
pmichaud yes, but "if(2)" is an expression while "if (2)" is a statement.
i.e., "if(2)" gets parsed by the <EXPR> rule, and "if (2) ..." is statement_control:if 14:59
moritz_ oh my.
diakopter what's the behavior of if(2) - I mean, what does that do?
pmichaud calls a function called 'if'.
diakopter ohhh, ok, now I see what you're asking
pmichaud perl6 doesn't have reserved words.
anyway, I haven't studied it in detail, so "if" might be not an example of what I'm trying to describe. 15:00
rakudo_svn r28548 | jonathan++ | [rakudo] Fix to make $foo does (R1, R2) working.
pmichaud But I do know tha TimToady mentioned at one point that he could simplify STD.pm because of having LTM implemented.
diakopter pmichaud: no, you have a good point. I'll think on it a bit and reply here. 15:01
moritz_: oh my?
moritz_ diakopter: that will become a FAQ 15:02
diakopter pmichaud: oh. this is simpler than I first imagined... Perl6in5's equivalent (before grammar-bootstrapping) statement control rule would require the whitespace, and token {<function_name_equiv> '(' etc..} would be the start of one of the options of function invocation pattern 15:04
pmurias rindolf: do you receive much critisism for your essays? 15:06
pugs_svnbot r20891 | moritz++ | [spec] added test for sequential alternations (||) in regexes
diakopter (where I'm using "token" as Perl6in5's equivalent to don't-allow-whitespace
)
rindolf pmurias: a lot of people comment on them.
diakopter pugs: say; 15:09
p6eval OUTPUT[␤]
diakopter hrm, that's going to be tricky 15:10
pugs: say say; 15:11
p6eval OUTPUT[␤1␤]
pugs_svnbot r20892 | moritz++ | [spec] tests for interpolated sequential alternations ( ||@list )
diakopter pugs: sub if { say 3; }; if() if if(); 15:13
p6eval OUTPUT[*** ␤ Unexpected "if"␤ expecting operator or bare or pointy block construct␤ at /tmp/fcE6kcl6g0 line 1, column 25␤]
15:14 pmurias left
diakopter pugs: sub if { say 3; }; if 1 { if() }; 15:14
p6eval OUTPUT[*** ␤ Unexpected "};"␤ expecting operator or bare or pointy block construct␤ at /tmp/F23g9OK4sk line 1, column 32␤]
diakopter pugs: sub if { say 3; }; if 1 { if() }
p6eval OUTPUT[*** ␤ Unexpected "}"␤ expecting operator or bare or pointy block construct␤ at /tmp/zKjOME5hS2 line 1, column 32␤]
diakopter pmichaud: pugs doesn't like a declaration of a sub named 'if' 15:15
pugs_svnbot r20893 | moritz++ | [spec] test for longest alternation match with interpolation in regex
pmichaud "Dear IIT. Your wireless access sucks." 15:21
15:21 TJCRI joined
pmichaud pugs is likely out of date on "if()" But I know it's a sub call -- TimToady even mentioned it in his keynote. 15:22
It's even in S03. 15:23
diakopter pmichaud: ok
pmichaud if $a == 1 { say "yes" } # preferred syntax
if ($a == 1) { say "yes" } # P5-ish if construct
if($a,$b,$c) # if function call
diakopter pugs: if say { say 3 } { say 4 } 15:24
p6eval OUTPUT[␤3␤4␤]
15:24 udinesh5 joined
udinesh5 will i get help to know where i am going wrong in a perl script ? 15:24
i am newbie
diakopter pugs: if !say { say 3 }
p6eval OUTPUT[␤]
moritz_ udinesh5: only if your script is written in Perl 6
udinesh5: if it's perl 5 (most likely), try #perl or #perlhelp or www.perlmonks.org/ 15:25
lambdabot Title: PerlMonks - The Monastery Gates
15:26 TJCRI left 15:27 b_jonas left
diakopter pugs: say 4 if !say { say 3 } 15:27
p6eval OUTPUT[<SubBlock(<anon>)>␤]
diakopter pmichaud: :) see pugs above... I think this is an ambiguity in the grammar... 15:28
it should've said "3" then <SubBlock(<anon>)>... 15:29
pugs: say { say 3 }
p6eval OUTPUT[<SubBlock(<anon>)>␤]
diakopter pugs: { say 3 }
p6eval OUTPUT[3␤]
diakopter moritz_: (anyone), shouldn't it say 3 then 1 (2 evals above) 15:30
moritz_ diakopter: no, don't think - blocks aren't evaluated if they are in the argument list of a sub 15:31
diakopter: otherwise you couldn't get a map {$code}, @stuff working
diakopter moritz_: oh. thanks :) 15:32
15:32 udinesh5 left
TimToady that's one of the reasons we distinguish statements from expressions 15:33
but note that we can't really add the implicit .() to a bare block until we see how it ends, since {...}.() is an expression 15:35
diakopter ok. so bare blocks are executed inline only in a statement list 15:36
TimToady: (& pmichaud & moritz_) thanks for the clarifications/corrections 15:37
(& pugs)
moritz_ diakopter: you're welcome ;) 15:38
TimToady you too :)
15:40 cjfields joined
rakudo_svn r28549 | jonathan++ | [rakudo] Get initialization of one-attribute roles working ($thingy does Answer(42)). 15:41
pugs_svnbot r20894 | jnthn++ | [spectest] A couple more mixin tests, for does (R1, R2) and does Answer(42) usage. 15:42
diakopter pugs: say { say 4; }.() 15:43
p6eval OUTPUT[<SubBlock(<anon>)>␤*** Cannot cast from VBool True to Pugs.AST.Internals.VCode (VCode)␤ at /tmp/nLpAEV8hW3 line 1, column 1 - line 2, column 1␤]
diakopter pugs: say ({ say 4; }.())
p6eval OUTPUT[4␤1␤]
diakopter that precedence isn't right...?
moritz_ why not? 15:44
diakopter shouldn't the one without the parens have the same output as the last one?
moritz_ uhm, yes
(I thought you were referring to the last eval) 15:45
diakopter right, so the precedence inference is wrong on the first one
ok
pugs: say ({ say 4; }.({ say 3 }.({ say 2 }))) 15:46
p6eval OUTPUT[4␤1␤]
diakopter that's not right either, I think..? it should've said '3' first
or do anonymous sub executions not take arguments? 15:47
moritz_ diakopter: only if you declare them (-> $args, {...}), or if they use $_ or $^a or some such 15:49
pmichaud 50% packet loss and train to airport is here
bbl
diakopter but in that case, I would think they would at least be evaluated before it decides whether to use them as arguments...? 15:50
if not, that's some magic...
15:50 r0bby joined
moritz_ why evaluated? the arity of a block is determined at compile time, I think 15:51
15:51 donaldh left
diakopter pugs: say ({ say 2; $_ }.({ say 3; $_ }.({ say 4 }).())) 15:53
p6eval OUTPUT[3␤4␤2␤1␤]
diakopter that's pretty darn lazy... 15:54
not evaluating the argument before it's accessed
(for the deepest nesting. it does do the first nesting before "say 2") curiouser ** and 15:55
diakopter waits for another rebuke :D 15:56
15:57 alanhaggai joined
diakopter TimToady: any clue about pugs' last? or is it a pugsbug? 15:58
moritz_ diakopter: lists are lazy by default 15:59
diakopter: and argument lists are lists, aren't they?
elf: say 1 xx 4 16:01
p6eval OUTPUT[1111␤]
moritz_ elf: say 2,3 xx 4
p6eval OUTPUT[23333␤]
16:08 scrottie joined 16:31 shlomif joined 16:32 rindolf left, shlomif is now known as rindolf 16:36 alester joined
diakopter moritz_: ok, but my 2nd point... it should've printed 2341. 16:37
rakudo_svn r28551 | jonathan++ | [rakudo] Get lookup of subs with & sigil working again, so we can do &foo.arity. Do this in a way that should work once we add lexical subs.
16:41 araujo left 16:43 alanhaggai left 16:52 cspencer joined 16:54 alanhaggai joined 16:55 shlomif joined 16:59 rindolf left, shlomif is now known as rindolf 17:03 Jedai joined
rakudo_svn r28552 | jonathan++ | [rakudo] Allow declaration of type variables with the :: sigil; Now 'my ::Foo = Int; my Foo $x; say $x.WHAT' gives Int. 17:07
diakopter pmichaud: to more fully answer your question from earlier, because 'if (<cond>)' will always be longer (by the one character) than 'if(<cond>)', it will win the alternation between the function call and the conditional pattern... 17:10
17:45 REPLeffect_ joined 18:10 lichtkind joined 18:13 Lorn_ joined 18:18 sri_work_ joined 18:19 ruoso joined
pugs_svnbot r20895 | diakopter++ | [yap6] added short-circuiting alternation as a special case of longest match alternation. Used the only other overloadable Perl operator with the same precedence as '|'. 18:19
ruoso in Brasil already 18:21
18:24 valkyr0 joined 18:27 sri_work left 18:28 pmurias joined
pmurias ruoso: the sm0p grammar works (returns the c code), i haven't plugged it in yet 18:29
ruoso cool
is it faster?
pmurias haven't benchmarked yet 18:30
Khisanth o_O you can write subs named if,while,until,... ? 18:31
moritz_ Khisanth: yes
pmurias ruoso: but it would be hard for it to be slower
ruoso heh 18:32
if it takes less than 5 seconds for the blocks in the tests, then it's faster ;)
(for each one, actually) 18:33
pmurias ruoso: do you have 5.10? 18:34
ruoso not yet... but I have to do the transition today on my machine
Khisanth sounds like it will make debugging a PITA
ruoso (which means a lot of modules to recompile)
pmurias Khisanth: why? 18:35
Khisanth assuming someone decides to be that foolish :) 18:36
18:37 Lorn left
diakopter pmurias: I'm coining 'lrule' as "lambda rule"... 18:37
(another top-level type declaration in this grammar syntax) 18:38
pmurias diakopter: example please
18:40 REPLeffect_ left, lichtkind_ joined, cspencer left, REPLeffect_ joined
diakopter pmurias: lrule commalist { -( plus( ',' ) ) - opt( $_[1], $_[0] ) }; 18:41
rule blkPrmsList { opt( prmDecl ) }; 18:42
oops
pmurias ruoso: i should have used KP6sm0p.p6 not KP6Grammar.pm?
diakopter rule blkPrmsList { opt( commalist( prmDecl ) ) };
...this just might work! 18:43
ruoso pmurias, the KP6 version was changed to make it work... it isn't the ideal implementation 18:44
diakopter in that example, $p (the self-referent for each rule) gets unshifted into @_
oops, I mean self-reference 18:45
which makes lrule capable of generating both left- and right-recursive parsers
rakudo_svn r28555 | jonathan++ | [rakudo] Implement generic subs/methods. You can now collect the type of a parameter by writing a signature like 'sub foo(::T $x) { ... }', and then use T within that method to refer to that type. 18:46
diakopter ... this will clean up Parser.pm a lot...
Auzon backticks are dead, correct? 18:49
Eevee think so, hope so, not positive
Auzon Larry wrote to the list that they are reserved for user-defined stuff 18:50
pugs_svnbot r20896 | pmurias++ | [sm0p] sm0p_with_actions does dump the match object unless --horizontal or --vertical options are given 18:52
pmurias ruoso: is KP6Grammar.pm enough? 18:53
ruoso: it's misc/sm0p/sm0p_with_actions if you want to plug it in
diakopter is there a term for self-currying? 18:54
ruoso KP6sm0p is the one that is used....
the others shouldn't be that different...
pmurias diakopter: what does self-reference have to do with lambdas?
ruoso but they might have some small things that were only fixed/implemented in the version being used
pmurias fix is used for recursion in the lambda calculus 18:55
18:56 lichtkind left
pmurias diakopter: got confused with the typed lambda calculus 18:58
19:01 tuxdna joined
ruoso pmurias, btw... I'm still finishing that project... 19:01
It should end in late june 19:02
and I'll be able to get back to smop in the start of july
pmurias ruoso: what are you creating? 19:04
ruoso it's a SOAP+XMPP application with web interfaces for public service scheduling 19:05
19:05 REPLeffect_ is now known as REPLeffect
ruoso that places where you get a numbered ticket that is later called on a video screen 19:06
it's a system to handle all that
19:07 Limbic_Region joined
diakopter pmurias: wait 'till you see sub lrule 19:14
19:25 FurnaceBoy joined 19:27 Auzon left 19:30 cookys joined 19:32 FurnaceBoy left
TimToady hi, guys, you're on the screen at Google-Chi, (though no one's here yet...) 19:47
diakopter lrule is a parser generator generator
ohai
PerlJam What's Google-Chi? Sounds ominous. :)
fullermd It's a calming exercise done by groups of people in the park every morning, where they all slowly and serenely search websites. 19:48
PerlJam fullermd: hmm. I was pronouncing it as google-ki rather than google-chee. That might have something to do with it. 19:50
TimToady (the tech-talk here is in 15 minutes)
diakopter information overload warm-up 19:52
yath what tech talk? 19:55
TimToady a google-tech-talk-kinda thing 19:57
Khisanth google expanding into martial arts? :)
TimToady of course, I'll be talking about the part that isn't googletech
(yet)
scrottie hi Google =)
yath TimToady: hmm, what exactly is the topic? ;)
scrottie btw, thanks for sending us Vint Cerf to Google Phoenix. Most enjoyable!
TimToady I kinda thought I might talk about Perl 6 today 19:58
(for a change)
yath ah, how surprising :)
diakopter as opposed to talking about it yesterday?
scrottie Can you stream video? I couldn't take notes fast enough the first time.
TimToady it's a shorter talk, so I took out all the Japanese
scrottie That means less time to write. 19:59
I'm fighting with SWIG... I'm tryin' to make this go.
20:00 zaphar_ps joined
zaphar_ps lol sitting waiting for Larry to talk at google 20:01
DietCoke ~~ at TimToady.
20:01 Auzon joined
wolverian what's the talk about? is it stupid to ask? :) 20:01
scrottie you're laughing and the jokes haven't even started yet?
zaphar_ps well you know 20:02
scrottie good to have an audiance in a good mood!
TimToady you guys should be running cool snippets for the folks coming in
20:02 Alias_ joined
zaphar_ps I had to miss YAPC so I had to come to this one :-( 20:02
family vacation came at the wrong time 20:03
TimToady we all have our faults
DietCoke perl6: 'Hi, google'.say
p6eval kp6 r20896: OUTPUT[Hi, google␤]
..pugs: OUTPUT[Hi, google␤]
..rakudo r28560: OUTPUT[Hi, google␤]
..elf r20896: OUTPUT[Hi, google␤]
DietCoke supposes he should learn some perl6. =-)
Auzon pugs: multi postfix:<!> (Int $a) {[*] 1 .. $a}; say 6!; # One of my favorite P6 snippets 20:04
p6eval OUTPUT[720␤]
pmurias TimToady: what did you talk about previously? 20:05
TimToady okay, talk starting, bye
Auzon Bye TimToady
pmurias good luck ;)
Auzon perl6: say [~] <m oo se> 20:06
20:06 p6eval left
Auzon oh cool. 20:06
20:06 p6eval joined
rhr_ pugs: -> $X { -> $proc { $X.( -> $arg { $proc.($proc).($arg) } ) }.( -> $proc { $X.( -> $arg { $proc.($proc).($arg) } ) }) }.(-> $func { -> $n { $n==0 ?? 1 !! $n*$func.($n-1) } }).(5).say 20:07
p6eval OUTPUT[120␤]
rakudo_svn r28563 | jonathan++ | [rakudo] Improvements/optimizations to rebless_subclass op from chromatic++. 20:09
20:10 Alias_ left
Auzon pugs: say :60[1,0,0] 20:13
p6eval OUTPUT[3600␤]
Auzon pugs: say :60[1,1,1]
p6eval OUTPUT[3661␤]
Auzon perl6: say :60[1,1,1]
p6eval kp6 r20896: OUTPUT[syntax error at position 4, line 1 column 4:␤say :60[1,1,1␤ ^ HERE␤] 20:14
..pugs: OUTPUT[3661␤]
..rakudo r28563: OUTPUT[Method 'pasttype' not found for invocant of class 'PAST::Var'␤current instr.: 'parrot;Perl6;Grammar;Actions;rad_number' pc 99622 (src/gen_actions.pir:3506)␤
..elf r20896: OUTPUT[Unknown rule: rad_number␤It needs to be added to ast_handlers.␤ at ./elf_f line 1900␤]
Auzon perl6: say <<<a> <b>>> 20:15
p6eval kp6 r20896: OUTPUT[syntax error at position 12, line 1 column 12:␤say <<<a> <b>>␤ ^ HERE␤]
..pugs: OUTPUT[*** ␤ Unexpected end of input␤ at /tmp/oyMXBG64q1 line 2, column 1␤]
..rakudo r28563: OUTPUT[Statement not terminated properly at line 1, near ">>>"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤
..elf r20896: OUTPUT[Use of uninitialized value in concatenation (.) or string at ./elf_f line 3425.␤syntax error at (eval 115) line 3, near "> )"␤ at ./elf_f line 3835␤]
mofino is it ready? can I convert my whole being to perl6 yet? 20:17
(no)
DietCoke you go ahead and try, we could use the bug reports. 20:18
20:20 stevan_ joined
diakopter hopes we're still on the screen 20:24
pugs_svnbot r20897 | diakopter++ | [yap6] added a function that generates functions that generate
r20897 | diakopter++ | parser generators that curry themselves. 8]
20:31 pmurias left
rakudo_svn r28564 | jonathan++ | [rakudo] Add ops to CLEANUP. 20:45
20:48 sri_work_ left 20:58 ruoso left 20:59 zaphar_ps left 21:07 FurnaceBoy joined
pugs_svnbot r20898 | Auzon++ | Vim syntax highlighting: Removed backticks. Attempted to add attribute support to qq and q (qq:!s"" and such) 21:08
21:14 FurnaceBoy left 21:20 jferrero joined 21:24 valkyr0 left 21:27 bacek_ left 21:31 donaldh joined 21:32 pmurias joined
pmurias diakopter: why did you add an unused tmp variable? 21:35
diakopter pmurias: same reason as all through Parser.pm ... b/c the things I read said that it's necessary to hold the reference before calling weaken() 21:38
I'm updating heavily currently; last rev was way broken (of course)
pmurias yes but i's held in stub 21:39
* it's 21:40
21:40 lichtkind_ left
diakopter oh 21:40
pmurias: you're up late 21:41
pmurias diakopter: yes
20 to 12
23:41
diakopter: you mean i'm talking nonsense ;) 21:42
diakopter pmurias: no 21:43
21:43 valkyr0 joined
diakopter I've nearly got this.. probably 10 minutes 21:43
pmurias can't sleep, don't have to get up early too 21:44
diakopter pmurias: question
pmurias go on
diakopter where in Grammar.pm did you have rule() defining what each_rule_name() 21:45
I mean, the sub entry should be defined somewhere, right?
pmurias $Perl6in5::Grammar::{$name} = sub() {$stub}; 21:46
diakopter ok.
the source filter causes it to be predeclared.. ah ok.
ahh, I see where I messed up
21:46 iblechbot left
pmurias the lrules have a different prototype 21:47
* should have
diakopter yep :) I had just realized that 21:49
21:52 araujo joined
pmurias diakopter: why do you weaken $stub? 21:53
what other reference keeps it alive?
it = the sub points to 21:56
s/sub/$stub/ 21:57
diakopter I was reading the mailing list archives for MJD's HOP book, and someone there recommended it
and if I remember correctly, said they weren't sure why it worked, but that it did. 21:58
b/c of a bug in perl GC when anonymous subs are used recursively
allegedly
rakudo_svn r28567 | jonathan++ | [rakudo] First cut at implementing anonymous enums. 21:59
pmurias diakopter: i would just throw it out
21:59 rindolf left
moritz_ and use perl 5.10.0, which hopefully fixed it ;) 22:00
pmurias diakopter: yap6 is broken now? 22:01
22:02 valkyr0 left
diakopter pmurias: yeah, like always ;) 22:02
pmurias: what prototype do lrules need if they need to accept a "subroutine entry" 22:03
pmurias none
prototypes are used for achieving stange parsing 22:05
* strange 22:06
diakopter committed
pugs_svnbot r20899 | diakopter++ | [yap6] some progress on lrule()
pmurias should yap6 work? 22:07
diakopter no... lrule() is still slightly broken.
but a lot less so.
pmurias is lrule essential?
diakopter I was going to use it to convert a lot of the constructors in Parser.pm to grammar rules, since they really belong in the grammar section, not the parser. 22:08
and also so that recursive parser generators can be auto-generated 22:09
with an eye toward eventually being able to generate any grammar...
at runtime.
pmurias diakopter: you should definitly remove weaken,as it's gets you nothing and is really just looking for trouble
diakopter I removed one of them in the last commit 22:10
pmurias it might be cleaner to have a this parser combinator (right terminology) instead of passing the currently defined rule as the first parameter 22:11
diakopter oh, I found my error 22:12
pmurias: can you restate your last 22:14
pmurias wait 22:15
lrule commalist { -( plus( ',' ) ) - opt( $_[1], this ) 22:16
}; 22:17
diakopter this? 22:18
ok, I see what you were saying, but I can't imagine how to construct this()
pmurias i'm sleeping now, will do it tomorrow if you have yap6&lrule in a working state 22:19
g'night
22:19 pmurias left
diakopter pmurias: thanks 22:21
decommutation 22:29
22:33 Auzon left, Auzon joined 22:35 Auzon left 22:46 cjfields left, donaldh left 22:54 valkyr0 joined 23:02 rindolf joined 23:04 rindolf left 23:17 nipotaway is now known as nipotan
pugs_svnbot r20900 | diakopter++ | [yap6] incremental progress on lrule(). I *think* the only blocking problem now is a discontinuity in the grammar. 23:19
23:20 meppl joined 23:31 toddr left 23:51 jferrero left 23:54 bacek joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!