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.
00:06 r0bby joined 00:11 armagad left 00:17 Eevee joined 00:58 pbuetow left 01:02 mncharity joined
pugs_svnbot r20371 | lwall++ | [STD] now does double quotes under new quoting regime 01:05
r20371 | lwall++ | [STD] with debugging turned off can now parse about 2000 chars/sec
diff: dev.pugscode.org/changeset/20371
01:08 justatheory left
mncharity re ruby to python compiler, and "He must have lost a bet, I can think of no other reason", perhaps the same reason there may be a EmitDirectToPython.pm as soon as elf is passing lots-of t/. Sometimes you have a python platform or a python shop, but wish to use a different language/compiler/libraries. 01:11
TimToady I wonder if it can translate Rails to Python :)
mncharity my fuzzy recollection is rubinius's experience suggests Rails is a rather high hurdle. 01:12
TimToady ah well, as soon as I get bored with STD, I can start writing all the OtherLang-to-Perl6 translators, since P6 is a superset of everythign else... :)
did you see I'm parsing about 2000c/s now? 01:13
Auzon including Perl 6, apparently ;)
TimToady I base that on parsing the first 3000 chars of radix.t in about 1.5 seconds 01:14
mncharity for all(@languages) { implement("real implementation","translator for porting","p6 backend","p6 backend where your still working in $_") }
01:15 [particle]ventus left
Auzon What does that "benchmark" mean? 01:15
mncharity 2kc/s :) nifty!
TimToady means I could theoretically parse STD.pm in about 50 seconds now 01:16
except it blows up on the first :foo<bar>
Auzon parse it with Rakudo?
TimToady no, with STD5
Auzon Oh.
Cool.
TimToady I've been refactoring the quoting sublanguages, but only have '' and "" so far, so the <> of adverbials blows up still 01:17
I need to figure out a better way of invoking a subrule of another language though, where that other language may be anonymous 01:18
mncharity re '',"", ah, I was just wondering about my evening. sync EXPR, punt name() since that's going to be refactored(?), and... sync ''/"" or wait?
ah, ok. that sounds like a wait.
TimToady <.cursor_fresh($lang).rule> is a bit long, but that's basically what happens underneath 01:19
cursor_fresh is an internal name, but a lousy one if it's to be outward facing 01:20
basically any subrule <foo> calls .cursor_fresh, it's just that if it's a sublanguage, .cursor_fresh takes that language as an optional argument
so I probably just need some syntactic relief 01:21
mncharity hmm. interesting. I was also wondering if being able to switch between compilers might be a desirable near-term feature. so "featureful but slow" could be mixed with "critical path must be fast" code.
TimToady but <$lang.foo> isn't quite right
well, that's what I'm finally trying to figure out the linkage for 01:22
mncharity compiler != parser, but... looks like tweakability might be important earlier than one might think. might use such a mechanism for sublangs.
TimToady I have to generate a subcursor for a subrule anyway, so that's the logical place to rebless the cursor into a different language for the following lexical scope
mncharity hmm... 01:23
TimToady well, I'm definitely trying to eat my own dogfood here with respect to '' "" rx// etc.
those are all derived grammars now
mncharity <OtherLanguage::foo> ...
TimToady has to be able to take it from $lang
mncharity <$($lang)::foo> 01:24
TimToady which could have several other mixed in languages since the last time it was named anything in particular
I also have to officially keep track of what is a language so I know the granulatity of my lexer cache 01:25
*granularity
mncharity <$($lang.WHAT)::foo> ?
TimToady that doesn't spell "relief"
mncharity lol
TimToady It's something more like the recursive <!...> and <foo=...> syntax 01:26
so <mumble/foo>
mncharity so <Foo::bar> for a grammar "class" Foo, and <$foo/bar> for an instance $foo of a grammar? 01:27
TimToady mumble/ is a placehoder 01:28
mncharity nod
TimToady s:nth//l/
01:30 Ched- left
mncharity why does $lang need to be kept track of? 01:30
01:30 Ched- joined
mncharity it doesn't seem like remembering it could affect dispatch. so that leaves 01:30
TimToady certainly it affects dispatch
mncharity passing information between rules/actions, which would suggest a context var.
some rule's subrule<hee> can't depend on the $lang which was used to reach that rule, no? 01:31
TimToady certainly it can 01:32
mncharity ???!???
TimToady THAT'S WHAT PERL6 IS ALL ABOUT!!!!!
mncharity hmm, let me rephrase...
TimToady when you say <infix> you don't know what language's set of infixes you're getting 01:33
it all depends on what class the cursor is blessed into
mncharity the identity of the rule hee invoked by a rule a's saying <hee> is uniquely determined by the grammar to which hee and a belong. it cannot be affected by a grammar or other variable $lang. no?
s/belong/ļ»æ class the cursor is blessed into/ 01:34
err, no!
um, ļ»æs/belong/belong (the grammar instance)/ 01:35
it's not yet clear to me why the cursor would need blessing.
TimToady to switch to an instance of a different grammar
by default cursor_fresh just propagates the current grammar
mncharity grammar B {...} grammar A { rule aa{<B::foo>} } <-- switching to a different grammar. (though not necessarily an _instance_ of a different grammar?) 01:37
I've never been clear on regex's handling of classes/grammars versus instances of same. 01:39
TimToady that doesn't rebless the cursor, so B::foo will end up running with A's dispatch instead of B's, if it calls any subrules of its own
it's like saying $object->B::meth in Perl 5 01:40
all it does is tell it where to look for meth, but the object is still an A object
to really switch parsers it must be fed $object->mutate_to_B->meth 01:41
which is what .cursor_fresh is doing
which, interestingly, is not the original purpose of .cursor_fresh
it was rather to isolate the parsing of the submatch object so we don't confuse it with the current matcher's Match object 01:42
mncharity apparently I still fundamentally don't understand regex.
TimToady (and _SUBSUME re-incorporates the sub-object's match as a capture into the current outer cursor) 01:43
well, I only just came to this a few days ago, so I'm not surprised. :)
mncharity lol
TimToady you'll notice all the funky .prior stuff is gone from Cursor5 now
that was all just bogus thinking
and why I could never get submatches to install themselves correctly into the parent Match 01:44
it's amazing how much refactoring can be triggered by trying to Get It To Actually Work Right 01:45
mncharity indeed
the discipline of the real
so...
TimToady yes, and that's why I've had to write my own regex engine again, after all 01:46
it's punishment for cargo-culting Henry's regexes into Perl in the first place
mncharity :) I've stopped making myself "I will never, ever, write another regex engine" commitments. Broken too many times. :/ 01:47
re Henry's regexp, lol
TimToady one issue I haven't resolved yet is backtracking control 01:48
at the moment I'm using "lazy" lists
which would be okay if I were actually implementing this in P6
but the lists aren't really lazy in P5
they're just out there somewhere on the stack 01:49
which makes it really hard to wipe out a bunch of options that you decide you never want to use
mncharity hmm. generator with out-of-band tweaking? 01:50
TimToady so I may end up having to switch back from list context everywhere to scalar, and just have lazy list objects, just so I can address backtracking control of :: and :::
mncharity doesn't understand the current execution model well enough to usefully comment
TimToady the current model doesn't actually backtrack anywhere in the sense of "returning to a previous stack frame" 01:51
all previous decision points are represented by cursors in a list
and failure to match is just throwing away parts of some list 01:52
mncharity so then why is :: and ::: hard?
TimToady some of the lists I want to throw away are off in other stack frames
mncharity ah... 01:53
TimToady so I may need to have some mechanism to invalidate cursors in the abstract
and then when the cursor finally gets activated, it wakes up and says, "oops, got trimmed, fail"
mncharity so options are exceptions, or returning a "please discard cursors until you find X" flag value. 01:54
TimToady but in that case, I need a naming scheme to keep track of active/forgotten cursors
something like that, but how do you name X? 01:55
probably can just deal with it in terms of match depth, I think 01:56
mncharity oh, or "oops, trimmed, fail"... though does it really need to wake up. re X, let's see, been here before, what did yare do...
TimToady and the difference between ::, :::, and <commit> is just what depth/shallowness to aim for
assuming I don't have multiple cursors in flight that are triming to different depths at the same time... 01:57
so now I get to write an exception engine from scratch as well :/ 01:58
mncharity re as well, LOL :)
sorry
re name X, 01:59
TimToady well, it's probably just the rule name, "if you aren't ._BRACKET (for ::) or .thisrule (:::) or .outerrule (for <commit>), peel back" 02:01
well, it's not .BRACKET, but whatever contains alternatives... 02:02
mncharity looks like yare just skips up through groups by prepending/removing "jump " prefixes to the exception.
btw, do regex have "code to execute we backtracking past here"? 02:04
*have a "..." construct?
TimToady well, that would be a generalization of this 02:05
I think it suffices to return some kind of "I failed successfully" cursor on the end of the list 02:07
as long as every scope knows to deal with it when it tries to use it to continue
at some level might actually involve trapping with eval {...} in p5, if I don't want to have an explicit test everywhere 02:08
well, ratchet quantifiers turn out to be fairly trivial under this model. it's just taking .[0] on the list of cursors :) 02:10
mncharity oy
TimToady and the difference between greedy and parsimonious is simply whether you do a reverse on the list frist 02:11
*first
where the non-laziness of p5 shows through is that if you use .*? it wants to build cursors for all possibilities to the end of the file before examining the short ones for successful continuation 02:13
with real gather/take it would only generate those extra cursors lazily
mncharity 'list to be reversed' all in one stack frame?
is it possible to generate them shortest first? 02:14
TimToady not only is it possible, it's a requirement 02:15
_PLUSg is defined as reverse $self->_PLUSf($block, @_);
you have to do the _PLUSf algorithm to find out the maximum possible greediness before you can start there 02:16
in essence .* has to always do .*? internally first
that becomes more obvious with <random>* has to do multiple <random>*? first because you don't know how long each random will be 02:17
mncharity ah, ok. hmm.
TimToady with . you can theoretically cheat because you know how each . is going to match without scanning left-to-right 02:18
and, in fact, Henry's package cheats just that way
mncharity nod 02:19
TimToady ah, was trying to remember what the 'f' was for--it was "frugal" 02:21
easier to type than parsimonious
02:23 yewenbin joined
TimToady gotta go take kid to karate... 02:23
mncharity cheers. /me hacks on STD_red... 02:24
02:25 eternaleye left 02:27 eternaleye joined 02:30 nipotan is now known as nipotaway 02:40 azure joined 02:42 azure is now known as azure1st 02:43 wknight8111 is now known as wknight-away
azure1st a 02:46
02:49 azure1st left 02:50 azure1st joined, azure1st is now known as azure
azure a 02:51
a 02:56
a 02:57
02:58 azure is now known as aaa, aaa is now known as azure1st 02:59 azure1st left 03:00 rindolf joined, azure1st joined
azure1st a 03:01
03:01 azure1st left 03:02 wknight-away left 03:04 azure1st joined
azure1st a 03:04
Auzon azure1st: What are you doing? 03:05
03:06 SamB left
mncharity TimToady: (1) in EXPR/reduce/chain, perhaps should check for $<sym> in addition to $<prec>? and cease to push the op on chain? then '3 eq 4 ne 5' parses as eq(3,ne(4,5)) 03:06
azure1st change my nick
03:07 azure1st left 03:08 rdice left, tobeya left 03:09 azure1st joined
TimToady mncharity: that's not what you want 03:12
yo uwant it to mean 3 eq 4 and 4 ne 5 03:13
it's just a variant of 1 < $x == $y < 42 03:14
03:14 alanhaggai joined
azure1st i want to join the perl6 dev but i don't know what to do , I know perl just a little. 03:21
Tene azure1st: what do you know? 03:23
azure1st I used perl process some txt changed, txt trans mysql or sqlite. 03:26
at the most time i use php 03:27
Tene azure1st: some work that needs to be done right now is reorganizing and categorizing and cleaning up the pugs test suite into the official perl 6 test suite
azure1st ok, how to do it. my english is not very well,so i speak slowly. 03:30
Tene azure1st: I think there should be some files talking about it in t/spec in the pugs repo. I'm not entirely sure what needs to be done or what to tell you to do right now. 03:32
I can probably find out tomorrow, if someone hasn't already answered before then.
mncharity re ļ»æ'3 eq 4 and 4 ne 5', ah, ok. but is currently ne(3,eq,4,ne,5) or some such. 03:35
azure1st oh 03:39
it's the time to the lunch i will back in one hour! 03:40
03:43 alanhaggai_ joined 03:55 drbean left, drbean joined 03:57 alanhaggai left
mncharity TimToady: when expect_term is at pos 1 in " {}", stdstopper is true (before { and after \s). should expect_term succeed or fail? 04:08
mncharity struggles with 'do {}'
04:11 SamB joined
TimToady do is not parsed with EXPR at that point 04:11
mncharity statement_prefix:do -> statement -> EXPR. at least that's what a std.rb draft is doing with 'do 3;' 04:13
TimToady but I think we're missing a statement_control:block 04:14
mncharity blocks currently show up as circumfix pblock
TimToady or maybe not
but the whitespace should already be eaten 04:15
so EXPR should only see "{...}"
so shouldn't be treated as a terminator
gotta go pick up kid from karate now... 04:16
mncharity if EXPR sees "{}", then stdstopper called from expect_term in loop is true. thus the question of
whether that should be deadly or no.
re karate, ok. just about end of day. 04:17
sigh. a slow process.
epiphany: all this parser stuff is too difficult, we should just use s-expressions. 04:19
just kidding.
spinclad combinators! SKI \o/ 04:23
mncharity *weary grin* 04:24
spinclad and feed it trees, not strings -- no parsing at all then 04:25
mncharity down with monads! up with arrows!
spinclad rah rah rah
mncharity it could make that grinding noise as you feed it trees and it spews out sawdust 04:26
spinclad ah, that would be the K combinator, aka the chipper 04:27
mncharity p6, the new particleboard
lol
thanks ļ»æspinclad :) 04:29
spinclad glad to provide some chaos on my way to bed
g'night, all 04:30
mncharity indeed.
ditto & :)
04:31 mncharity left 04:35 peepsalot joined 04:48 rindolf left 04:56 meteorjay left 04:57 Psyche^ joined 05:09 Patterner left, Psyche^ is now known as Patterner 05:14 peepsalot left 05:16 rindolf joined 05:17 syle left 05:33 vaughn left 06:04 IllvilJa left 06:06 Psyche^ joined
pugs_svnbot r20372 | lwall++ | [STD] more progress in handling :qq etc. as mixins 06:19
diff: dev.pugscode.org/changeset/20372
06:23 Patterner left, Psyche^ is now known as Patterner 06:50 iblechbot joined 07:10 azure1st left 07:13 azure1st joined 07:18 elmex joined 07:29 alanhaggai_ is now known as alanhaggai 07:41 lambdabot left, lambdabot joined
pugs_svnbot r20373 | moritz++ | [svn-search] catch errors in indexer 08:13
diff: dev.pugscode.org/changeset/20373
08:15 charsbar left 08:18 riffraff joined 08:21 charsbar joined 08:23 meteorjay joined 08:38 charsbar left 08:40 masak joined 08:41 charsbar joined 08:46 rindolf left 09:01 Schwern joined 09:14 masak left 09:17 masak joined 09:32 IllvilJa joined 11:09 lisppaste3 left 11:25 lisppaste3 joined 11:32 wknight8111 joined 11:34 zamolxes_ is now known as zamolxes 11:42 rdice joined 12:25 stevan_ left 12:33 pbuetow joined 12:37 rdice left
obra TimToady: feeds.feedburner.com/~r/boingboing/...ili-m.html 12:39
lambdabot Title: HOWTO make a chili mister - Boing Boing, tinyurl.com/3o4qad
12:46 IllvilJa left 12:50 alanhaggai left 12:57 meppl joined 13:00 iblechbot left 13:09 kanru joined 13:10 alanhaggai joined 13:30 TJCRI joined 13:31 ispy_ joined 13:36 rdice joined 13:43 alanhaggai_ joined 13:45 Lorn joined 13:47 eternaleye left 13:58 alanhaggai left, Lorn_ left 14:04 zostay_ is now known as zostay 14:09 mncharity joined
mncharity could someone who knows how please kick dev.pugscode.org? the svn changesets are all Internal Server Error. thanks! 14:13
TimToady: so what is the "RightThing" story on 'do {}'? should expect_term be successfully matching at the { despite stdstopper returning true there? 14:15
It looks like the same issue occurs with "\n{3;}\n". 14:20
pugs_svnbot r20374 | jnthn++ | [spectest] Fudge S12-class/basic.t for Rakudo, and add an extra instantiation test. 14:28
diff: dev.pugscode.org/changeset/20374
mncharity re "fudge", curious... "To perform in an incomplete but marginally acceptable way,"[FOLDOC] vs "1 present in a vague way, especially to mislead. 2 manipulate (facts or figures) so as to present a desired picture."[OED(c)]. And my default binding is "faking scientific data". 14:36
14:39 alester joined 14:46 chris2 joined 14:48 iblechbot joined 14:51 cls_bsd joined 15:00 cmarcelo joined 15:01 IllvilJa joined 15:03 ilbot2 left 15:04 moritz joined
moritz if somebody misses the irc logs - the server is down :( 15:05
TimToady then I won't say anything quotable. :)
mncharity irclog.perlgeek.de/perl6/today down
colabti.org/irclogger/irclogger_logs/perl6 up
lambdabot Title: #perl6 irc log
TimToady mncharity: I think " {}" should probably only be checked by expect_infix 15:06
mncharity ah, ok. I'll try that.
after disabling the (?<!\s)(?={) stdstopper clause, the next issue i hit is 15:07
TimToady the next issue you hit is total silence? :) 15:08
mncharity given "do 3; 4", do's statement eats the ';', but we got to :do via an (pause, :) 15:09
expect_term, which then continues eating into the next statement.
STD5_run does the right thing however. 15:10
TimToady it does?
mncharity s/ļ»æļ»ædoes the right thing/parses it correctly/
well, it's a little hard to read, but I think so? 15:11
./STD5_run statementlist -e 'do 15; 16;'
oh, nm 15:12
TimToady my guess is that we'll have to force it like we do with }\n 15:13
mncharity parsing it as 'zo 15; 16'. maybe.
15:13 masak left 15:15 alanhaggai__ joined
TimToady statement prefixes are intentionally asymmetric, like listops. they both function as a kind of opening paren without a closing paren, and are automatically closed out at the appropriate spot 15:16
listops quit at a terminator, and statement prefixes terminate at end of statement 15:17
they're a bit like ] in some Lisp dialects
or rather, that ; at the end of one is 15:18
do do do do do do do say "hi"; # <--- terminates all the do statements
not sure how to tell STD that though. :) 15:19
"if you've seen one statement terminator, you've seen them all..."
it's almost like statement_prefix wants to uneat the terminator 15:21
but it would be better not to eat it in the first place
so perhaps statement needs to be split
or maybe make statementlist/semilist eat the terminator instead 15:22
then nested calls to statement could use the same terminator 15:23
mncharity re 'ļ»ædo do do do do do do say "hi";', :) 15:26
15:27 jhorwitz joined
pugs_svnbot r20375 | lwall++ | [STD] like this maybe 15:27
diff: dev.pugscode.org/changeset/20375
mncharity would be so nice to have dev.pugscode.org back... :/ 15:28
moritz aye :/
I have root access to feather, but really now idea what to do 15:29
obra moritz: anything in apache's error log?
15:29 alanhaggai_ left
moritz looking... 15:29
[Tue May 06 17:24:44 2008] [error] an unknown filter was not added: includes
lots of times
and sometimes a few pages of syntax errors 15:30
in python, I guess
mncharity TimToady: re moving eat_terminator, does that cause problems for the || in statement? at least in std.rb, I use the e_t to determine whether I have the right || branch. 15:32
moritz: googling... "It sounds like the SSI module mod_include isn't enabled - do so with sudo a2enmod include." ubuntuforums.org/showthread.php?t=586469 15:33
moritz mncharity: but that seems to be a a different problem than the dev.pugscode.org one
[particle] TimToady: you should have a look at rakudo's grammar wrt statement_end
## Here we deviate from STD.pm a fair bit on the handling of
## statementlist and statement_end, if only because this
## version (first prototyped in NQP) seems a bit more
## straightforward. We may discover that it won't work.
svn.perl.org/parrot/trunk/languages/perl 6/src/parser/grammar.pg 15:34
[particle] fixes his client's auto-correct feature that keeps breaking perl6 into perl 6
mncharity there's www.mail-archive.com/debian-apache@...08494.html 15:35
lambdabot Title: Bug#407307: Apache2 - missing mod_includes causes unknown filter errors, tinyurl.com/6pos2r
mncharity which debian thinks is fixed in the latest apache www.mail-archive.com/debian-apache@...09023.html 15:36
lambdabot Title: Bug#407307: marked as done (Apache2 - missing mod_includes causes unknow, tinyurl.com/65sb9l
moritz mncharity: I enabled the module, but dev.pugscode.org still doesn't work
mncharity :(
moritz mncharity: there are some syntax errors reported in python scripts 15:37
and I don't know how to fix those :/
mncharity re ļ»æļ»æurl "svn.perl.org/parrot/trunk/languages/perl" <-- why space is a bad idea in file names.
[particle] svn.perl.org/parrot/trunk/languages...grammar.pg
lambdabot tinyurl.com/5jh29s
mncharity svn.perl.org/parrot/trunk/languages...grammar.pg
lambdabot tinyurl.com/5jh29s
[particle] my irc client inserted the space
mncharity ah, got it :) 15:38
moritz should I apt-get dist-upgrade feather and hope that it unbreaks things? ;-) 15:39
TimToady mncharity: I don't understand the problem you propose with ||
mncharity ļ»æmoritz: curious. for a (few) days, the problem was intermittent. you could refresh the page a time or few, and get the change set. 15:40
TimToady why would expect_term have anything to do with whether the next thing is "if" or "for"? 15:42
pugs_svnbot r20376 | lwall++ | [STD] split infixstopper out from stdstopper
diff: dev.pugscode.org/changeset/20376
TimToady oh, wait, not that e_t 15:43
duh
but still, the || can't proceed unless it sees a modifier of some sort 15:44
mncharity ļ»æTimToady: re ||, ah, token statment's [...||...] <e_t> becomes [...||...]: <e_t>, so I'm being incorrect in using e_t as part of the || "do we have the right branch" check. i think. 15:45
pugs_svnbot r20377 | moritz++ | [irclog] ignore mncharities BOMs 15:48
diff: dev.pugscode.org/changeset/20377
TimToady seems like it's only likely to fail if someone defines a terminator that can be confused with a statement modifier
though there's something going on here with LTM of terminator sets that I haven't quite got as transitive as I like 15:49
in regex, when you match <foo> ** '||', it should tell foo that '||' is a valid terminator somehow
mncharity hmm. /me puzzles over grammer.pg 15:51
rakudo: do 5; 6
exp_evalbot RESULT[Method 'print' not found for invocant of class 'String'ā¤current instr.: '_block10' pc 44 (EVAL_11:17)ā¤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 785 (src/PCT/HLLCompiler.pir:458)ā¤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1067 (src/PCT/HLLCompiler.pir:587)ā¤called from Sub
..'parrot;PCT::HLLCompiler;command_line' pc 1246 (...
TimToady it would be convenient if the LTM matcher could mix in such dynamic terminators with the more static ones
mncharity 6 15:52
moritz rakudo: do 5; 6; print 1
exp_evalbot OUTPUT[1]
moritz rakudo: print do 5; 6
exp_evalbot OUTPUT[5]
pmichaud rakudo doesn't yet have a .print method
15:53 CMA joined
TimToady rakudo: print do do do 5; 6 15:53
exp_evalbot OUTPUT[5]
mncharity ļ»æmoritz: re BOM, :) tnx
pmichaud I can add .print
moritz exp_evalbot does some evil hackery with do and print if a statement didn't produce output 15:55
pmichaud although I'm not sure why it was reporting "Method 'print' not found for invocant of class 'String'", since 6 is an Int
moritz (I didn't implement it, I don't know the details :/) 15:56
pmichaud do 5; 6 looks like two statements to me. Perhaps I'm parsing it wrong.
15:56 andresambrois joined
pmichaud rakudo: 6 15:56
exp_evalbot RESULT[Method 'print' not found for invocant of class 'String'ā¤current instr.: '_block10' pc 44 (EVAL_11:17)ā¤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 785 (src/PCT/HLLCompiler.pir:458)ā¤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1067 (src/PCT/HLLCompiler.pir:587)ā¤called from Sub
..'parrot;PCT::HLLCompiler;command_line' pc 1246 (...
mncharity re looks like, is.
pmichaud rakudo: say 6.WHAT 15:57
exp_evalbot OUTPUT[Intā¤]
mncharity the issue is getting multiple expect_term's to end in the same place. currently one eats the ';', and the enclosing one then chews onward rather than also stopping there. 15:58
TimToady it would appear that STD5 now correctly parses 'do do do 5; 6'
mncharity looks like grammar.pg is handling it with endstmt. not sure.
pmichaud .print added in r27345
grammar.pg handles it with endstmt, yes.
moritz if (!length $response){ $program = '( ( do { ' . $program . "\n} ).perl ).print";
mncharity svn.pugscode.org pings, but svn can't connect. 15:59
moritz wait a second, the update is currently running 16:00
pmichaud (and we get ".print" not available for String because that's what .perl is returning. Which is also wrong, as it needs to be a Str.) 16:01
TimToady but STD5 seems to be doing it right without using endstmt; not that I've tested it exhaustively or anything...
moritz wohoo, trac is back!
mncharity re 'ļ»æcorrectly parses', at least as of ļ»æ20375, I get identical parses for ./STD5_run statementlist -e 'do do 5; 6;' and 'xo xo 5; 6;' 16:02
pmichaud actually
rakudo doesn't implement 'do' yet.
Auzon rakudo: my Int $i; $i
exp_evalbot RESULT[Method 'perl' not found for invocant of class ''ā¤current instr.: '_block10' pc 35 (EVAL_11:16)ā¤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 785 (src/PCT/HLLCompiler.pir:458)ā¤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1067 (src/PCT/HLLCompiler.pir:587)ā¤called from Sub
..'parrot;PCT::HLLCompiler;command_line' pc 1246 (src/PCT...
Auzon rakudo: my Int $i; say $i
exp_evalbot OUTPUT[Intā¤]
moritz that's wrong, isn't it?
shouldn't it be undef?
Auzon I hope so.
pugs: my Int $i; say $i
exp_evalbot OUTPUT[Intā¤]
TimToady Int is one of the names of undef
Auzon Oh
So it's Int, not "Int" 16:03
TimToady there is no undef in the p5 sense anymore
moritz pugs: my Int $i; say $i ~~ undef
exp_evalbot OUTPUT[ā¤]
moritz pugs: my Int $i; say $i ~~ Undef
exp_evalbot OUTPUT[*** No such subroutine: "&Undef"ā¤ at /tmp/Lf0pubEwa5 line 1, column 16 - line 2, column 1ā¤]
TimToady the most general undef is something like Object
moritz pugs: my Int $i; say $i ~~ Object
exp_evalbot OUTPUT[1ā¤]
mncharity re "ļ»æInt is one of the names of undef", :) Someday I have to learn p6. :)
moritz pugs: my Int $i; say defined $i
exp_evalbot OUTPUT[ā¤]
Auzon Oh, so it's just the generic object name. Makes sense
moritz pugs: my Int $i; say defined $i ?? 1 !! 0; 16:04
exp_evalbot OUTPUT[0ā¤]
[particle] all protoobjects are undefined
moritz debian++ # apt-get upgrade fixes things in unstable ;-)
[particle] when you see Int, you're seeing the stringified protoobject
mncharity stringifies as .WHAT ? 16:05
moritz isn't it confusing that an integer stringifies to something which isn't a Int?
[particle] .WHAT returns the protoobject
TimToady .WHAT isn't supposed to stringify
[particle] which stringifies as 'Int' when it's printed
TimToady stringification is very lazy in P6 :) 16:06
16:06 armagad joined
TimToady unlike in p5, where you almost can't even think about a type except in the stringified form 16:06
mncharity re trac, changeset works, !!! ļ»æmoritz++ 16:07
ļ»æmoritz+=2
TimToady which is why STD5's .mixin method actually generates a longer package name for the new type
mncharity yay. having broken infrastructure is so disturbing. my thanks.
TimToady currently, when you mix Q_w into Q_qq, you end up with Q_qq_w as a package name, which is a cool/vile hack 16:08
mncharity re "ļ»æcool/vile hack", :)
TimToady and probably doesn't generalize right to Regex etc.
but I repeat myself
16:15 yewenbin left
mncharity ponders what to do. was on a "get STD_red parsing t/kp6 and pugs's t/" dev path. "do {}" and "n{3;}n" unresolved. could look for other parsefails, but better to wait. so seem blocked on that. hmm... 16:15
lunch & 16:16
[particle] i was under the impression that "manhattan dispatch" wasn't hardwired, but was default for MMD 16:23
...i still am under that impression. am i wrong? 16:24
16:24 CMA left
TimToady no, we abandoned manhattan years ago 16:29
[particle] ah. ok.
glad to be wrong.
TimToady we don't try to compute distance in type space at all 16:30
either one type encompasses another, or it doesn't
16:32 moritz left 16:34 moritz_ joined
TimToady afk & #errands 16:37
16:40 irclogbot_backup joined 16:41 moritz_ left, irclogbot_backup left 16:45 moritz_ joined 16:53 iblechbot left 16:54 eternaleye joined, ilbot2 joined, ilbot2 left 16:55 ilbot2 joined 17:07 nipotaway is now known as nipotan 17:27 d4l3k_ joined, wolv_ joined, Juerd left, cjfields joined, PerlPilot joined 17:28 dalek left, wolverian left, PerlJam left, wolv_ is now known as wolverian, pmichaud left
meppl good night 17:28
17:28 eternaleye left 17:29 d4l3k_ is now known as dalek, andresambrois left 17:32 meppl left 17:35 pmichaud_ joined 17:38 eternaleye joined 17:47 cjfields left 17:49 cjfields joined 18:29 lidden left 18:35 jhorwitz left 18:48 CMA joined 18:50 cotto joined 18:51 lambdabot left 18:52 lambdabot joined 19:05 iblechbot joined 19:09 syle joined 19:11 alanhaggai__ left 19:16 sscaffidi left
moritz_ [particle]: ping 19:18
19:19 armagad left 19:20 armagad joined, rhr left
[particle] hi-o moritz_ 19:22
ooh, a goldfinch 1m away :)
moritz_ [particle]: do you have some thoughts regarding GSoC and SVN branches?
[particle] spring++ 19:23
moritz_: you mean for collecting Auzon's commits?
moritz_ aye
Auzon is here in case that helps
moritz_ it seems to be the recommended way
[particle] we *could* stick a branch in the pugs repo
however, i believe judicious use of commit message tags will suffice
moritz_ but would that be useful? 19:24
I mean a branch
[particle] well, for example, we could add a makefile target to rakudo for 'gsoctest' that pulls from the branch instead of the trunk
moritz_ so, assuming we take the commit message tags...
how do we deliver the deliverable (aka patches) to google? 19:25
[particle] i like tagging messages better.
we can use svn diff tools
Auzon Same here. If we were to branch, it'd make reorganizing a pain.
moritz_ we can't assume that Auzon's patches alone will be make up a clean diff
[particle] right, so we submit a series of diffs 19:26
moritz_ so should we just make a complete diff of t/ and tell google "90% was Auzon's work, see $list for details"?
[particle] we can get a list via svn log of each commit auzon did, and produce diffs of each 19:28
i'm looking for specific info on what google expects 19:29
do you have a link handy?
moritz_ no
I just read eric's mail
[particle] here in terms of service:code.google.com/opensource/gsoc/2008/tos.html
lambdabot Title: Google Summer of Code - Google Code
moritz_ "Google expects us to deliver "your code" 19:30
at program completion, which might mean that you need your own branch
(depending on whether your project is new code or not) to make it easy
to create a nicely-bounded changeset."
[particle] i see "To provide Google with a copy of the finished project code which may be publicly hosted by Google."
to me, i think a set of diffs qualifies. i'll ask leslie, though. 19:31
moritz_ the "finished project code" is the result in my eyes, not the diff
[particle] sure, could be that, too 19:32
19:32 kyrbe joined, kyrbe left
[particle] Auzon: have you submitted the TPF CLA yet? 19:32
Auzon No, I have not. 19:33
Auzon puts that on his todo list 19:34
[particle] thx
19:37 REPLeffect joined
[particle] i suppose auzon could develop in a private branch and synchronize with trunk frequently 19:40
then his work would be in one place. 19:41
still, there would by synchronization commits interspersed
moritz_ aye, it doesn't solve the problem
19:42 pmurias joined 19:43 meteorjay left
[particle] message sent to google mentors list 19:43
moritz_ did you CC me? 19:46
[particle] ah, no. are you subscribed to google-summer-of-code-mentors-list@ ? 19:47
moritz_ no 19:48
[particle] ah. you should be :)
@googlegroups.com
lambdabot Unknown command, try @list
[particle] time to reboot my xp vm into sp3 19:49
let's see if it makes me happy or sad
pmurias mncharity: re blocked: if you are blocked and can't add any features the other way forward would be to remove a cludge/hack ;) 19:50
20:06 sscaffidi joined 20:10 [particle]ventus joined 20:30 meppl joined 20:32 mj41 joined 20:40 rhr joined
[particle]ventus moritz_: Auzon: groups.google.com/group/google-summ...bb7878e16f 20:40
lambdabot Title: Google Summer of Code Mentors List | Google Groups, tinyurl.com/6xeffu
Auzon I can't see that :-/ 20:41
20:41 aindilis left, aindilis joined
moritz_ only open for members :( 20:43
[particle]ventus ok, i'll forward the mail to you after lunch &
moritz_ no hurry ;)
Auzon thanks
21:08 ispy_ left 21:18 stevan__ joined, [particle]ventus left 21:20 TJCRI left 21:26 larsen_ joined 21:27 xerohg joined 21:36 [particle]ventus joined 21:39 rdice left 21:41 cjfields left 21:44 barney joined 21:50 Eidolos left
meppl good night 21:56
moritz_ meppl: didn't you say that some 4 hours ago already? ;-) 21:57
21:57 cjfields joined 21:58 Eidolos joined
meppl moritz_, yea, but i couldnt sleep 21:58
21:58 meppl left
moritz_ d'oh 21:59
good luck this time ;)
22:12 barney left, cjfields left 22:13 larsen_ left 22:14 pmurias left
CMA esta ausente: at high scholl 22:15
22:20 cmarcelo left 22:21 luqui joined 22:22 alester left 22:24 lidden joined 22:29 c0nspiracy joined 22:31 sscaffidi left 22:32 gbacon left, gbacon joined 22:38 iblechbot left 22:49 Schwern left, [particle]ventus left 22:51 eternaleye left 23:07 chris2 left 23:21 eternaleye joined 23:45 c0nspiracy left

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!