Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2007-02-05

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 svnbot6 r15178 | putter++ | yet_another_regex_engine/OPP.pl - created.  Yet another operator precedence parser.  A first, rough, non-running draft.  See notes at top of file.
00:14 Psyche^ joined perl6
00:15 putter end of day for me
00:15 TimToady night
00:15 moritz the package is called libghc6-network-dev (probably)
00:15 moritz night ;)
00:21 REPLeffect_ joined perl6
00:21 putter tene: re current state of p6 rules in pugs, looking at http://m19s28.vlinux.de/cgi-bin/pugs-smokeserv.pl? (specifically the recent liniux repository snapshot second ">>", under t/regex/ , where both dark green and red means failed), I'd say "a great deal not working".
00:21 lambdabot Title: Pugs Smoke Reports
00:22 putter and "but some is".
00:22 putter good night all &
00:27 Schwern joined perl6
00:30 Psyche^ is now known as Patterner
00:32 svnbot6 r15179 | moritz++ | mentioned build depenency on haskell-network aka libghc-network-dev in
00:32 svnbot6 r15179 | moritz++ | PROBLEMS file
00:40 neonse left perl6
00:43 nekokak joined perl6
00:50 charsbar_ joined perl6
01:07 justatheory joined perl6
01:09 PartclCoke audreyt: hello.
01:09 audreyt PartclCoke: greetings.
01:11 * PartclCoke wonders where he can get a lambdabot for #parrot.
01:11 allbery_b @ask dons for it?
01:12 * PartclCoke wonders if he's still a meta-committer in new pugs land.
01:12 audreyt PartclCoke: sure, you still are. http://invite.pugscode.org/
01:12 lambdabot Title: Login
01:13 * PartclCoke says, "Freaky."
01:13 PartclCoke allbery_b: looks like "@ask" means something special?
01:14 allbery_b it's a lambdabot command
01:14 allbery_b @help ask
01:14 lambdabot ask <nick> <message>. When <nick> shows activity, ask them <message>.
01:14 allbery_b dons is the keeper of lambdabot
01:15 PartclCoke @ask dons "Any chance we could get a lambdabot on #parrot on irc.perl.org? thanks -Coke
01:15 lambdabot Consider it noted.
01:15 PartclCoke @ask partclCoke huh?
01:15 lambdabot Consider it noted.
01:15 PartclCoke what?
01:15 allbery_b you could build your own but the source is in an odd state currently
01:15 partclCoke what?
01:15 lambdabot partclCoke: You have 1 new message. '/msg lambdabot @messages' to read it.
01:15 allbery_b yeh, it's cae-dependent
01:17 allbery_b so it told you it recorded the query, next tme dons pops up on IRC he'll get informed like you did
01:17 allbery_b he's in .au so the time shifting capability is rather useful :)
01:18 partclCoke mmm. takes care of one of the worst limitations of irc. =-)
01:23 audreyt errands bbl &
01:36 nekokak joined perl6
02:04 luqui joined perl6
02:14 buetow joined perl6
02:25 agentzh joined perl6
02:25 agentzh oh, smartlinks.pl is misbehaving on feather...
02:26 agentzh i'm getting the error "invalid top directory at /usr/share/perl/5.8/File/Find.pm line 592."
02:27 agentzh i think i'd better look into the recent commits to smartlinks.pl :)
02:30 agentzh err, for those who don't know. one group of tests can be linked to multiple places in the spec by putting the links successively.
02:31 agentzh so it's already supported n-to-m correspondance for months.
02:31 agentzh not sure if i have made it explicit in smartlinks.pl's POD...looking now...
02:37 Limbic_Region joined perl6
02:41 justatheory joined perl6
02:45 dmq joined perl6
02:46 agentzh joined perl6
02:48 agentzh @tell szabgab please recheck your recent commits to smartlinks.pl because i'm getting error report from feather's cron. thank you :)
02:48 lambdabot Consider it noted.
02:50 agentzh the synopses on feather are no longer updated automatically :/
02:50 agentzh "This page was generated at 2007-02-03 10:01:14 GMT"
02:53 agentzh @tell szabgab i think it's r15169 that has introduced the problem :)
02:53 lambdabot Consider it noted.
02:55 agentzh no svn usable on my machine, so i can't fix it myself...windows virus-- :(
02:57 mdiep joined perl6
02:57 agentzh looking forward to my shiney new ubuntu box :)
02:58 mdiep joined perl6
03:05 andrei joined perl6
03:06 dduncan amen
03:12 weinig|away is now known as weinig
03:24 svnbot6 r15180 | agentz++ | [util/smartlinks.pl]
03:24 svnbot6 r15180 | agentz++ | Fixed a bug in the --dir support.
03:24 svnbot6 r15180 | agentz++ | smartlinks.pl should work again on feather++ :)
03:27 svnbot6 r15181 | agentz++ | [util/smartlinks.pl]
03:27 svnbot6 r15181 | agentz++ | added comments on smartlinks.pl's m-to-n correspondance linking support to POD.
03:28 agentzh @tell szabgab the bug is now fixed in r15180. never mind. :) thanks to feather++ and Juerd++
03:28 lambdabot Consider it noted.
03:36 agentzh dduncan: is there a web browser on feather?
03:39 agentzh bye &
03:39 agentzh left perl6
03:43 mako132_ joined perl6
03:47 weinig is now known as weinig|zZz
03:53 lrizo joined perl6
03:54 lrizo hi
03:54 lrizo any can help me?
03:58 Aankhen`` joined perl6
03:58 lrizo hi
04:00 jrockway what's your question?
04:01 lrizo i whould like to make a echo with udp
04:01 lrizo but perl can not generate my udp packets
04:02 jrockway perl6?
04:02 lrizo not, perl 2.2
04:03 partclCoke ... 2.2!?
04:03 partclCoke this channel is for talk about perl6, but your version of perl intrigues me.
04:04 jrockway i'm equally confused :)
04:04 jrockway the production version of perl is perl 5.8.8
04:05 lrizo excuseme
04:05 lrizo im confused too :(, is perl 5.8
04:05 partclCoke oh, that's far less interesting. =-)
04:06 lrizo jejej
04:06 jrockway you should ask on plain-old #perl, then
04:06 lrizo are you developer of perl 6?
04:06 lrizo yes i will ask in a old perl room
04:06 partclCoke I think most of the people here have committed something to one of the... 3 version of perl6, ja.
04:06 allbery_b ths whole channel is perl6 developer folks
04:06 partclCoke Try Net::UDP, too.
04:07 allbery_b some of whom may not do perl5
04:07 lrizo ok
04:07 jrockway there's some other UDP server whose name I can't remember...
04:07 * jrockway searches a bit
04:08 lrizo based on you experencies, can i make a UDP ECHO on perl 6?
04:08 allbery_b just as a wild uess, if your server is running on a port < 1024, it must be run as root (OS restriction)
04:08 jrockway i'm thinking of Net::Server with Net::Server::Proto::UDP
04:08 jrockway docs are at http://search.cpan.org/~rhandom/Net-Server-0.95/
04:08 lambdabot Title: Paul T. Seamons / Net-Server-0.95 - search.cpan.org
04:10 lrizo thanks
04:10 lrizo let me see the reference
04:26 ruz joined perl6
04:36 stevan_ joined perl6
05:31 justatheory joined perl6
05:35 dduncan left perl6
05:37 leed joined perl6
05:50 nipra joined perl6
05:55 lrizo left perl6
05:58 ruz joined perl6
06:01 ruz_ joined perl6
06:14 BooK_ joined perl6
06:42 rfordinal joined perl6
07:02 svnbot6 r15182 | lwall++ | first whack at getting ops into the standard grammar
07:37 iblechbot joined perl6
07:38 marmic joined perl6
08:21 devogon joined perl6
08:53 drrho joined perl6
09:27 BooK_ is now known as BooK
09:41 andara joined perl6
10:06 BooK_ joined perl6
10:15 elmex joined perl6
10:26 Ziggy6 joined perl6
10:39 naught joined perl6
10:39 pfarmer joined perl6
10:42 lyokato_ joined perl6
10:44 IllvilJa joined perl6
11:03 leed joined perl6
11:09 lyokato joined perl6
11:17 ruz__ joined perl6
11:23 pfarmer joined perl6
11:28 fglock joined perl6
11:29 fglock hi
11:29 lambdabot fglock: You have 1 new message. '/msg lambdabot @messages' to read it.
11:31 ruoso joined perl6
11:32 fglock putter: re help: the core of kp6 is done - I'm now starting bugfixing, optimization, docs
11:39 svnbot6 r15183 | fglock++ | kp6 - finished core compiler
11:39 svnbot6 r15183 | fglock++ | see v6/v6-KindaPerl6/docs/ANNOUNCE
11:50 fglock putter: for next phase of the project, an optimized gramar engine written in simple-p6 would be a great help
11:57 fglock putter: re p5 code which enumerates the operators, each with all its associated info - please see v6.pm perl5/Pugs-Compiler-Perl6/lib/Pugs/Grammar/*fix.pm
12:00 fglock putter: re pugs regexes: pugs uses the same as v6.pm = Pugs::Compiler::Rule
12:19 explorer joined perl6
12:41 rfordinal_ joined perl6
12:49 mako132_ joined perl6
13:01 REPLeffect_ joined perl6
13:04 rfordinal_ is now known as rfordinal
13:12 moritz has anybody got an idea what to do about http://bugs.debian.org/cgi-b​in/bugreport.cgi?bug=406081 ?
13:12 lambdabot Title: #406081 - pugs_6.2.13-0.1(ia64/unstable): FTBFS: No such file - Debian Bug repor ...
13:13 moritz the quoted file ext/Math-Basic/lib/Math/Basic.pm is in the svn...
13:23 buetow joined perl6
13:36 iblechbot joined perl6
13:39 kanru joined perl6
13:46 bonesss joined perl6
14:31 BooK joined perl6
14:37 mr_ank joined perl6
14:39 amnesiac joined perl6
14:43 kanru_ joined perl6
14:46 gaal moritz: I think it's maybe just a transient failure, like running out of memory likely. prelude compilation takes lots of ram.
14:47 moritz gaal: maybe. but wouldn' other compilations fail as well? and many of them?
14:51 gaal they might, if they run of of memeory. the question is whether this always happens, and how much free memory did the system have when it happened.
14:52 moritz I don't know if the debian buildds collect such information...
14:53 vel joined perl6
14:54 gaal ok, i'll reply to the bug.
15:06 kanru joined perl6
15:31 xinming joined perl6
15:42 REPLeffect_ joined perl6
15:46 ofer1 joined perl6
15:54 mako132 joined perl6
16:00 mdiep_ joined perl6
16:07 BooK_ joined perl6
16:10 Dr_Pi joined perl6
16:21 Dr_Pi_ joined perl6
16:32 chris2 joined perl6
16:44 ProperNoun joined perl6
16:45 bernhard joined perl6
16:55 thepler joined perl6
17:04 kanru joined perl6
17:05 justatheory joined perl6
17:05 svnbot6 r15184 | moritz++ | Added libghc-network-dev to debian/control
17:05 svnbot6 r15184 | moritz++ | Mentioned project_planning/{TASKS,TALK} in Pugs::Doc::Hack
17:16 nipra joined perl6
17:23 moritz hey, what happened  to the -e (etc.) operators from perl5?
17:23 moritz pugs> if -e "/etc/passwd" { say "Yes"; }
17:23 moritz *** No compatible subroutine found: "&e"
17:23 moritz    at <interactive> line 1, column 4-21
17:24 Gothmog_ They're gone and replaced...
17:25 Gothmog_ See S03
17:25 wolverian moritz, see http://perlcabal.org/syn/S03.ht​ml#Changes_to_Perl_5_operators
17:25 lambdabot Title: S03
17:25 Gothmog_ You can use $filename.:e or $filename ~~ :e
17:26 moritz wolverian, Gothmog_: Thanks
17:26 moritz I skimmed over S03 but seemed to have missed it ;)
17:28 moritz the ~~ version works, the .:e not
17:30 Gothmog_ Might not be implemented in pugs yet.
17:32 explorer joined perl6
17:34 prism joined perl6
17:36 TimToady it's not implemented yet.
17:47 bonesss is now known as bonesss`out
17:57 elmex joined perl6
17:58 GabrielVieira joined perl6
18:13 Southen_ joined perl6
18:13 nipra joined perl6
18:20 mdiep joined perl6
18:29 justatheory joined perl6
18:32 ashleyb joined perl6
18:34 ashleyb joined perl6
18:38 Limbic_Region joined perl6
18:38 drupek12167 joined perl6
18:43 Caelum joined perl6
18:44 mdiep joined perl6
19:08 Qiang joined perl6
19:24 jisom joined perl6
19:31 dduncan joined perl6
19:35 gaal TimToady: "Perl 5's ${...}, @{...}, %{...}, etc. dereferencing forms are now $(...), @(...), %(...), etc." in S03/Changes -- shouldn't we, uh, de-"reference" that?
19:37 PerlJam heh
19:38 PerlJam gaal: If it helps the programmer think of them as references, why not let them continue to do so?  :)
19:38 gaal for the same reason that other conceits that stop working should be given up
19:39 gaal @@@@@@@@x fails in Perl 5 where it wouldn't in Perl 6, for example.
19:40 gaal well, $$$$$$$x
19:40 DHGE joined perl6
19:41 PerlJam Well, those are still "perl 5's dereferencing forms".  The sentence doesn't say what they are in perl 6 land.  That seems like a useful omission to me  :-)
19:43 gaal hmm, I guess if the purpose is pedagogical, not to scare off folks with a new concept right in line one, then there's a reason for it. But who said this has to be in line one? :-)
19:46 gaal &
19:50 Limbic_Region gaal - concept ne conceit
19:54 Khisanth and anyone that writes $$$$$$$x deserves a beating
19:55 PerlJam Khisanth: amen!
19:55 * tene gives Khisanth a beating.
19:56 PerlJam tene: you seem to be suffering from use/mention conflation
19:56 tene PerlJam: he said "write", not "use"
19:56 PerlJam touche
19:57 theorb joined perl6
20:03 hexmode joined perl6
20:11 buetow joined perl6
20:14 partclCoke ... even if someone DOES say "jehovah". </python>
20:14 explorer joined perl6
20:15 LimbicRegion joined perl6
20:26 mr_ank joined perl6
20:29 LimbicRegion is now known as Limbic_Region
20:52 svnbot6 r15185 | lwall++ | added skeleton of precedence parser
20:52 svnbot6 r15185 | lwall++ | syntactic categories look a little more role based now
20:53 IllvilJa joined perl6
21:05 ezraburgh joined perl6
21:16 larsen_ joined perl6
21:21 Gothmog_ Where is the grammar located in the repository? I'm somewhat curious. :-)
21:22 ruz__ is now known as ruz
21:24 wolverian Gothmog_, /mirror/pugs/src/perl6/Perl-6.0.0-STD.pm
21:24 wolverian er, you get the idea.. :)
21:28 partclCoke what's the current status of pugs on parrot?
21:30 audreyt -CPIR still works last I checked
21:31 audreyt still no objects, closures should still work if PIR havn't changed too much
21:31 audreyt wouldn't be surprised if embedded parrot has bitrotten
21:31 audreyt wasn't very stable to begin with
21:32 partclCoke is it worth keeping up to date, you think?
21:32 audreyt is the embparrot API declared stable?
21:33 partclCoke nothing in parrot has been declared stable, i didn't think.
21:33 audreyt having worked with 4 generations of that API, I don't quite envision chasing more :)
21:33 audreyt *nod*
21:33 partclCoke though I think particle is trying tof ix that.
21:33 partclCoke I'm just wondering if having the end version of pugs being able to target the end version of parrot is useful. If so, probably better to chase it now than try to fix it up later.
21:34 audreyt target is -CPIR
21:34 audreyt it's in better shape
21:34 audreyt embed otoh probably doesn't need to work
21:34 audreyt it needed to work when we were relying on PGE for rules
21:34 audreyt (target means "take perl6, compile with pugs, and generate PIR")
21:34 audreyt (embed means "call PIR natively from within Pugs")
21:34 ezraburgh joined perl6
21:35 partclCoke c has done a bit of work on embedded. no clue on current status.
21:35 audreyt in any case, firmly convinced that fix it up later is easier than chase.
21:35 TimToady audreyt: have a look at the most recent src/perl6/Perl-6.0.0-STD.pm
21:35 partclCoke eh. I'll poke later this week.
21:35 partclCoke really. Hurm. then I won't poke too much. =-)
21:35 partclCoke STD is an unfortunate acronym!
21:36 TimToady categories are kinda multi-ish rules
21:36 partclCoke (er, abbreviation, I suspect it really is)
21:36 audreyt Six-Targetting Definitions?
21:37 * audreyt sees lower-cased :prec<z=> etc table and flashes thru nethack monsters
21:37 TimToady I suppose we could spell it out "Standard" or something...
21:37 moritz TimToady: how boring ;)
21:38 partclCoke eh. not everyone has the brane of a 14 year old.
21:40 audreyt TimToady: so instead of "multi rule ..." you just write "rule"?
21:40 audreyt after a category is seen that is
21:41 TimToady I suppose "category" could be "proto rule" but I didn't want to commit to rule/token
21:41 TimToady since you could conceivably mix 'em
21:42 audreyt they are just shorthands for regexes
21:42 TimToady but extensible regexes
21:42 audreyt so "proto rule" just means "proto regex {...:sigspace:ratchet}"
21:42 TimToady can be extended by more declarations at derivation time
21:42 audreyt ayeaye.
21:42 audreyt really not identical to subs
21:43 TimToady they're trying to get a handle on the scoping of grammar reanalysis as it mutates
21:43 jferrero joined perl6
21:44 TimToady also I suspect category needs special handling so we know what's legal in foo:<*> syntax
21:44 TimToady unless we just admit anything of that form as a valid name anywhere...
21:44 audreyt token prefix is Named_unary[] {:<rand>}
21:45 TimToady just sub ($) probably
21:45 audreyt anytime a user defines an in-scope name the metasub need to go register a new token?
21:45 audreyt to the current grammar objec that is
21:45 TimToady I suspect we can special case identifier parsing
21:45 audreyt mm?
21:45 mako132 left perl6
21:46 TimToady on the other hand, there's some benefit to shoving idents through longest token as well
21:46 TimToady presuming an efficient DFA or trie front end
21:47 audreyt yeah. find with that way for now
21:47 Gothmog_ wolverian: Thanks!
21:47 wolverian you're welcome :)
21:47 audreyt *fine
21:48 GabrielVieira2 joined perl6
21:48 moritz anybody interested in failing debian builds via dpkg-buildpackage?
21:48 larsen_ joined perl6
21:49 audreyt TimToady: so the dynamic parsing rules (list vs scalar assignment, immediate binding with ::=, declarator pseudoassignment) belong there too?
21:49 audreyt moritz: I am
21:50 moritz audreyt: http://moritz.faui2k3.org/tmp/pugs-dpkg-buildlog (wait a second until that damn slow upload is finished)
21:51 audreyt TimToady: historically those three are among the more tricky ones to lift from precparser land back into dynamic rulespace
21:51 TimToady any rule that gets ahold of the engine can treat however much of the subsequent text as part of its "token", and call EXPR with a different sub-precedence
21:51 audreyt (I saw that => has been taken care of)
21:51 moritz audreyt: upload finished ;)
21:51 * allbery_b wasn't gonna say anything... :)
21:52 TimToady sorry 'bout that... :)
21:52 allbery_b (nobody gets my name right)
21:52 marmic joined perl6
21:52 TimToady that's how ??...!! is parsed already
21:52 audreyt moritz: clearly 10smoker is a faulty patch
21:53 moritz audreyt: right ;)
21:53 audreyt rafl: thoughts?
21:53 audreyt rafl: seems 10smoker no longer applies (assuming it's yours)
21:54 audreyt TimToady:
21:54 audreyt    { :<??> <EXPR(|%conditional> :<!!> }
21:54 TimToady %conditional contains the max precedence
21:54 audreyt explain the <FOO(|%x> metasyntax?
21:54 TimToady missing right paren
21:54 audreyt oh. easy explanation
21:55 audreyt fixed
21:55 TimToady you see how multiple :<> can work together to form infix:<?? !!> from the user point of view?
21:55 svnbot6 r15186 | audreyt++ | * Fix missing right parens
21:55 audreyt what is prefix | doing?
21:56 audreyt expanding into nameds?
21:56 TimToady yes
21:56 TimToady there's a skeleton "method EXPR" at the bottom
21:56 audreyt amazingly simple
21:57 audreyt oh, "your ad here"
21:57 audreyt ok
21:57 TimToady I didn't put in the reduces.
21:57 larsen__ joined perl6
21:57 audreyt nodnod. got the idea
21:57 audreyt what are those suppose points?
21:58 TimToady places where you parse it by the standard rules, but mark that you suspect they meant something else, so when the parse gets into trouble, you have things to look at to give them better diags
21:58 TimToady "Did you mean..."
21:59 audreyt no examples yet?
21:59 TimToady no, it's just a stub idea, but one I've had for ages.
21:59 audreyt but just string?
21:59 audreyt or try some reparse
21:59 TimToady not strictly necessary to get a good parse, but will be necessary for acceptance by p5 programmers
21:59 audreyt and use reparse to give more info?
22:00 audreyt moritz: drop perl6-compiler@perl.org and [email@hidden.address] a mail?
22:00 TimToady yes, grinding through backtracking to figure out some might-have-beens
22:00 audreyt moritz: rafl may not be around atm, but I suspect he can fix it rather quickly
22:01 audreyt why does that remind of of /:panic//
22:01 audreyt *remind me of
22:01 moritz audreyt: I just delete the /debian/patches/ dir and now it builds (and needs quite a while .. ;)
22:01 TimToady 'cause I used the same word maybe?
22:01 audreyt :)
22:02 moritz audreyt: I'll write the mail as sonn as the package is ready
22:02 audreyt all in all sane
22:02 audreyt moritz++ # thanks!
22:02 audreyt TimToady: you still think
22:02 audreyt when .:e {}
22:02 audreyt when :e {}
22:02 audreyt need to be synonyms?
22:03 audreyt admittedly the latter reads better
22:03 TimToady I think so
22:03 audreyt ok. will go implement .:e
22:03 larsen___ joined perl6
22:03 TimToady I don't see a better way out of the prec problem with ~~
22:03 audreyt no, I mean ditch ~~
22:03 audreyt force .: all the time
22:04 TimToady don't want to force that
22:04 TimToady since I think people want "when :r & :x"
22:04 audreyt btw, how is $foo.Foo::Bar parsed nowadays?
22:05 Azure-BOT joined perl6
22:05 TimToady .<name>
22:05 audreyt $foo.Foo: :Bar?
22:05 TimToady method is name, not ident
22:05 audreyt (this is S03:247)
22:05 audreyt then maybe S03 wants to say that
22:05 TimToady name can include ::
22:07 TimToady fixed
22:07 audreyt thx
22:08 audreyt $file.=:e
22:08 audreyt legal?
22:08 TimToady no
22:08 audreyt so it's dot
22:08 audreyt not methname
22:08 audreyt gotcha
22:08 TimToady .: is special
22:09 audreyt S03:2156
22:09 audreyt    Any       Pair      test object             .STATUS(X) (Str,IO do filetest)
22:09 audreyt oh. the parens means #
22:09 audreyt the last pair that is
22:09 TimToady ???
22:10 audreyt I misparsed it as Signature
22:10 audreyt :(Str, IO does filetest)
22:10 audreyt but it actually means
22:10 audreyt # Str and IO objects do filetest
22:10 TimToady yes
22:10 TimToady but intended for any lightweight status reporting
22:10 TimToady syntax also hijacked for :exists and :delete
22:11 audreyt but that is for construction not slicing syntax no?
22:11 audreyt I mean, a pair literal :foo{...}
22:11 audreyt means
22:11 audreyt foo => {...}
22:11 audreyt which doens't use slice syntax inside
22:11 audreyt unless I'm mistaken
22:12 audreyt i.e. it's not a subscribe, it's a normal hash composer (or Code)
22:12 audreyt *subscript
22:12 TimToady well, $obj.:delete{expr} just gets the pair and can do whatever it likes to expr, I expect
22:12 audreyt except
22:12 audreyt @foo.delete:[1;2;3]
22:12 lambdabot Unknown command, try @list
22:12 audreyt works but
22:12 audreyt only if (@x = [1;2;3]) works
22:12 audreyt I mean .:delete[] of course
22:13 TimToady ; is supposed to work inside brackets
22:13 TimToady statement vs dimensional is left to context
22:14 TimToady might not be lazy enough for {;;} though...
22:14 audreyt so in .:exists{1;2;3}
22:14 audreyt it's turned into Code
22:14 audreyt but in
22:14 audreyt .:exists{1=>2, 3=>4}
22:14 audreyt it's turned into Hash
22:14 audreyt either way the pair STATUS gets is not very useful
22:15 audreyt still not quite grasping the picture
22:15 TimToady unless there's some way to eval Code to return dimensions
22:15 TimToady but I admit it's far-fetched.  will think s'more
22:15 audreyt what I mean is "parsed exactly as a subscript will be"
22:15 audreyt need to be a hard mandate
22:15 audreyt and not (generally) to be labelled as colon pairs
22:16 audreyt cool :)
22:16 audreyt (fwiw, I wonder if we can parse it as postcircum, instead of circum.)
22:17 audreyt and generally just treat all brackets in .: form as slice
22:17 audreyt but that will not work with "when" and ~~
22:17 audreyt which was the motivation behind suggestion of dropping them
22:17 TimToady not too worried about "when :delete{...}
22:17 audreyt when :exists{1} {...}
22:18 TimToady but maybe pairs in general need to be a little lazier about {...}
22:18 TimToady or maybe we always use [...] for such things.
22:19 audreyt mmm lua.
22:19 audreyt the autocomposing thing bites us a bit too much
22:19 audreyt not limited to here
22:19 TimToady I have thought some about [a => 1, b => 2] maybe autohashifying lazily
22:19 audreyt wouldn't oppose codeonly{}
22:19 audreyt exactly.
22:20 wolverian I was just thinking about saying that aloud but was too scared..
22:20 audreyt but everybody else is going that way :)
22:20 TimToady "like the rest of the world uses" is not an argument that flies in certain quarters... :/
22:21 audreyt unless it's the dot...
22:21 audreyt j/k
22:21 wolverian the solution is obviously {[ ]}
22:21 TimToady "...and wrong"
22:21 tene {[(< >)]}
22:22 audreyt .oO([{< >}])Oo.
22:22 wolverian space portal!
22:22 TimToady well, %(...) already presumably works
22:22 moritz perl6 should parse ascii-art ;)
22:22 TimToady but only for bare literal
22:22 audreyt moritz: seen Saturn.pl?
22:22 jabbot audreyt: I havn't seen Saturn.pl, audreyt
22:22 moritz audreyt: nope
22:23 audreyt moritz: http://www.perlmonks.org/index.pl?node_id=397958
22:23 lambdabot Title: Saturn
22:23 wolverian jabbot, seen a spelling guide?
22:23 jabbot wolverian: I havn't seen a, wolverian
22:23 TimToady going with optional %foo['bar'] would allow us to use "my" inside.
22:23 TimToady but I've always felt that [...] should be reserved for numeric subscripts so the optimizer knows things are getting coerced to integer generally.
22:23 audreyt "my"?
22:24 nipra joined perl6
22:24 TimToady a { } subscript defines a lexical scope
22:24 audreyt aha.
22:24 TimToady a [ ] one doesnt
22:24 audreyt "deliberately"
22:24 tene Just let people use whatever grouping operator they want and guess at the meaning from context.
22:24 TimToady rebol
22:24 TimToady everything is [...]
22:25 audreyt actually if we make :meth first class
22:25 audreyt and not worry about STATUS
22:25 audreyt then this problem all goes away
22:25 audreyt no more pairs, generally
22:25 audreyt the :exists method gets a normal subscripting arglist as postcircum[] would
22:26 TimToady evaluated too soon
22:27 audreyt for the .:exists form?
22:27 TimToady .meth{foo} is parsed as .meth.{foo}
22:27 TimToady you'd have to return a proxy
22:27 audreyt sure, I mean keep parserule
22:28 audreyt .:meth is special in that it takes postcircum brackets
22:28 audreyt in addition to parens
22:28 audreyt but make the dispatch go through normal dispatch
22:28 audreyt not named redispatch
22:28 TimToady hmm.  :infix<*>
22:29 audreyt method :exists {...} # also easier to write than STATUS variants
22:29 audreyt dunno. seems simpler
22:29 audreyt quite a bit faster too.
22:29 TimToady I actually thought of that a couple of days ago, but it didn't seem like it worked out, forget why.
22:30 audreyt it wouldn't work out if we allow ~~:exists[]
22:30 audreyt but as you said we're fine with limiting ~~ to simple (nonbracketing, or parens-only) args
22:30 audreyt otherwise seems kosher
22:32 TimToady ah, part of the reason I didn't want to go that route was I didn't want :r etc in Str, but maybe that's just does something or other
22:32 audreyt yeah.
22:32 TimToady or maybe we just shove :r into Str anyway
22:32 audreyt and it nicely summarizes :!e to mean !:e
22:32 audreyt !   .:e
22:32 audreyt and do without the "inspect the bool" game
22:32 weinig|zZz is now known as weinig
22:33 TimToady well, we did have :!x meaning two different things, I guess
22:33 audreyt *nod*
22:33 TimToady does it mean !.:x or does it mean .:x(Bool::False)
22:33 TimToady old pair syntax has the latter
22:33 audreyt the current spec places burden on STATUS implementers
22:33 audreyt to inspect the bool
22:33 audreyt and translate by hand
22:34 audreyt (pugs does that currently)
22:34 audreyt icky.
22:34 TimToady or maybe pairs can carry their !-hood independent of their argument
22:34 TimToady method :!r {...}
22:34 audreyt I think it's easier to leave pairs alone
22:34 audreyt and make :!foo consistent instead of as part of the name
22:35 TimToady but people will want :!exists{...}
22:35 audreyt and that is differnet from !.:exists{} how?
22:35 moritz Saturn.pl is way cool ;)
22:35 audreyt i.e. if we assume a rewrite of all
22:35 audreyt $obj.:!meth{...}
22:35 audreyt into
22:36 audreyt !$obj.:meth{...}
22:36 audreyt don't quite see what's lost
22:36 audreyt rewrite by compiler not by hand
22:36 audreyt moritz: so Perl 5 alredy does ascii art :)
22:36 TimToady :!flag as an option gets ! how?
22:36 audreyt $obj.meth(:!flag) still sees $obj.meth(flag => False)
22:37 audreyt the .: form no longer have anything to do with pairs
22:37 audreyt .:foo{...} has nothing to do with foo=>{...}
22:37 audreyt .:!foo nothing to do with foo=>False
22:37 TimToady but prevents "when :!w"
22:38 audreyt does not
22:38 audreyt still means
22:38 audreyt ! ~~:w
22:38 audreyt the smartmatcher can mandate that easily
22:38 audreyt when :w(False) # special cased
22:38 audreyt always means !:w
22:39 TimToady seems not quite a simple as it should be.
22:39 audreyt but pushing all this to Pair objects seems to make it more complex
22:40 audreyt dunno. not very tied to this particular interpretation. but handwritten STATUS tester feels somehow wrong, plus I'd like subscript over constructor
22:40 TimToady well, I guess we're just saying the meaning of ! is contextual.
22:41 audreyt hm. <1.5hr till OSCON proposal deadline
22:41 * audreyt goes making something random up :) bbiab
22:41 miyagawa audreyt: thanks for reminding me
22:41 audreyt miyagawa: heya :)
22:42 miyagawa :)
22:42 audreyt miyagawa: so seems there'll be YAPC::Beijing right after OSDC.tw
22:42 miyagawa ooooh!
22:42 audreyt jp-tw-cn
22:42 audreyt I'll be to all three
22:42 audreyt they'll post to use.perl soon
22:42 miyagawa awesome
22:42 audreyt hopefully we can find some .jp hackers to join
22:43 audreyt next time it could be an OSDC
22:43 miyagawa should we rename ours to YAPC::Asia::Tokyo? :)
22:43 audreyt but YAPC first this time
22:43 audreyt nah. YAPC::Asia is just fine :) the rest of us are OSDC'ing
22:43 audreyt OSDC.cn
22:43 miyagawa we're talking with Sasada and other RubyKaigi organizers to make OSDC (at least Ruby and Perl) happen in 2008
22:44 audreyt eggsellent
22:48 Schwern joined perl6
23:16 bd_ joined perl6
23:22 Aankhen`` joined perl6
23:27 gnuvince joined perl6
23:28 beppu I've heard crazy rumors that pugs can take Perl 6 code and emit Javascript?  Is this true?
23:30 audreyt yes, ./pugs -CJS -e "say 'hello'"
23:30 beppu audreyt += 2 ** 32
23:31 * audreyt overfloweth
23:32 audreyt it's iblech's work, though.
23:32 beppu whooooaaaa -- that was a lot of javascript
23:33 TimToady probably dumped the prelude with it
23:33 weinig joined perl6
23:34 beppu yeah...  I'm going to try to see if it actually works....
23:34 audreyt with the released pugs, definitely works
23:35 audreyt trunk pugs, conditionals may be broken (known bug) as we switched representation for "if ... CODE" to "if ... BLOCK"
23:35 audreyt known issue and will be fixed before release
23:38 weinig is now known as weinig|away
23:42 Psyche^ joined perl6
23:42 beppu ha!  firefox 1.5.0.8's js interpreter says:  too much recursion
23:43 beppu didn't give me any line numbers, though.
23:43 moritz ;)
23:43 moritz pugs -CJS -e "say 'hello';"|wc -l
23:43 moritz 56085
23:43 moritz that's quite impressive ;)
23:46 Squern joined perl6
23:55 Psyche^ is now known as Patterner

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

Perl 6 | Reference Documentation | Rakudo