Camelia, the Perl 6 bug

IRC log for #parrot, 2008-06-24

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:09 AndyA joined #parrot
00:17 contingencyplan joined #parrot
00:17 * Tene debates between napping and aimlessly hacking on rakudo with coroutines and lazy lists and gather/take.
00:19 * Whiteknight votes for aimless hacking
00:20 Limbic_Region Tene - as long as you don't take after audreyt and hack while your napping
00:20 Limbic_Region it leads to burn out
00:22 Whiteknight oh, true. we don't need people getting burnt out\
00:29 Tene I do my best hacking in my sleep.
00:29 Tene Of course, then I have to try to work while sleeping, and I'm dramatically less successful at that.
00:30 cotto_work also, avoid the bathtub
00:30 cotto_work it is not an ideal hacking environment
00:32 Tene I did the first draft of chitchat in the bathtub.
00:37 Whiteknight laptop + bathtub = segfault
00:39 cotto_work Tene, is that a language you're using to teach yourself the PCT or are you porting an existing language?
00:39 Tene cotto_work: smalltalk
00:56 Whiteknight I've always wanted to learn smalltalk. This might be my excuse
00:56 * Auzon thinks chitchat is a brilliant name for a smalltalk implementation
00:57 Whiteknight agreed
00:57 TiMBuS joined #parrot
01:04 davidfetter mmm...chitchat
01:16 cotto_work going home &
01:31 Andy joined #parrot
01:40 Limbic_Region Tene - Randal and I shared a cab back to our respective hotels and we discussed what it would take to port smalltalk to Parrot - he seemed quite interested
01:46 davidfetter he's over at irc://irc.freenode.net/postgresql
01:46 davidfetter as RandalSchwartz
02:04 kid51 joined #parrot
02:04 jjore joined #parrot
02:18 DietCoke I have cygwin, but have crappy luck building parrot.
02:20 Limbic_Region DietCoke - why so?
02:20 * Limbic_Region builds Parrot on cygwin regularly
02:20 DietCoke if I knew that, I wouldn't have crappy luck building parrot. =-)
02:21 DietCoke Yes. I for the longest time, on my work machine, I was never able to get a successful build, and no one was ever able to figure out why.
02:21 DietCoke I haven't tried on my windows laptop in some time, as strawberry perl JFW.
02:21 Limbic_Region well, interestingly enough - there are instances where the only fix is a fresh checkout
02:21 DietCoke didn't help. =-)
02:21 Limbic_Region but I assume you really don't want a parrot cygwin
02:21 DietCoke it was something about my cygwin install.
02:21 DietCoke I don't, but don't mind testing it occasionally.
02:22 Limbic_Region I tested japhb's OpenGL patch tonight - freaking cool
02:22 Limbic_Region but now it is bed time
02:22 Limbic_Region TTFN
02:24 kid51 DietCoke:  Have you had a chance to look at patch re ops_renum.mak?
02:33 DietCoke No, sorry, not yet.
03:08 Andy left #parrot
03:10 gryphon joined #parrot
03:10 apeiron joined #parrot
03:15 tetragon joined #parrot
05:05 Psyche^ joined #parrot
06:48 Andy joined #parrot
06:57 iblechbot joined #parrot
07:00 cotto_home earth status?
07:00 cotto_home purl, earth status is <reply>not destroyed
07:00 cotto_home that bot's getting fickle lately
07:09 _Maghnus joined #parrot
07:13 masak joined #parrot
08:00 _Maghnus joined #parrot
08:43 lichtkind joined #parrot
08:50 apeiron_ joined #parrot
08:59 apeiron joined #parrot
09:02 bacek joined #parrot
09:56 Zaba joined #parrot
10:05 dalek r28673 | fperrad++ | libs4php:
10:05 dalek : [php] some math
10:05 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28673
10:07 bacek "#44211: pi and e behaving strangely when used as a bareword in math"
10:07 bacek I think this one can be closed now.
10:07 bacek evening everyone.
10:07 bacek pugs?
10:12 moritz good localtime() bacek ;)
10:12 moritz bacek: no purl around
10:13 bacek lazy bot...
10:13 bacek Looks like we need another one...
10:13 moritz OH NOEZ
10:15 bacek perl6: my $halfPi = pi()/2; say sin $halfPi;
10:15 polyglotbot OUTPUT[1␤]
10:21 bacek Bot::BasicBot::Pluggable?
10:22 moritz Bot::BasicBot, yes
10:23 moritz I guess it's a modified version of p6eval, which lives in the pugs repo in misc/evalbot/
10:23 bacek ::Pluggable have almost all modules. E.g. Karma, etc
10:23 moritz oh, I thought you were referring to polyglotbot
10:23 moritz bacek: or look at infobot on cpan
10:24 bacek 336 results for 'IRC' on cpan :)
10:24 moritz and 250 for 'bot' ;)
10:24 spinclad perl6: say e**(i*pi/2)
10:24 polyglotbot OUTPUT[Could not find non-existent sub i␤current instr.: '_block11' pc 31 (EVAL_12:16)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 806 (src/PCT/HLLCompiler.pir:469)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1088 (src/PCT/HLLCompiler.pir:598)␤called from Sub
10:24 polyglotbot ..'parrot;PCT::HLLCompiler;command_line' pc 1267 (src/PCT/HLLCompiler.pi...
10:25 spinclad perl6: say e**(pi/2)
10:25 polyglotbot OUTPUT[4.81048␤]
10:25 spinclad perl6: say e**(1i*pi/2)
10:25 polyglotbot OUTPUT[4.81048␤]
10:25 moritz ouch
10:25 moritz that's just a bit wrong ;)
10:26 moritz e^(i pi / 2) == i, right?
10:26 spinclad yes
10:26 spinclad e^(2pi i / 4)
10:26 bacek perl6: say 1i*pi
10:26 polyglotbot OUTPUT[0+3.14159i␤]
10:26 bacek perl6: say 1i*pi/2
10:26 polyglotbot OUTPUT[0+1.5708i␤]
10:26 bacek perl6: say 1i**2
10:26 polyglotbot OUTPUT[1␤]
10:27 spinclad perl6: say e**(1.5708i)
10:27 polyglotbot OUTPUT[4.8105␤]
10:27 moritz ok, complex ** isn't implemented
10:27 moritz or not correctly
10:27 bacek yet another multi-method required function...
10:29 spinclad perl6: say pi / 2; 1;
10:29 polyglotbot OUTPUT[Statement not terminated properly at line 1, near "1;"␤current instr.: 'parrot;PGE::Util;die' pc 120 (runtime/parrot/library/PGE/Util.pir:82)␤called from Sub 'parrot;Perl6::Grammar;eat_terminator' pc 20609 (src/gen_grammar.pir:2813)␤called from Sub 'parrot;Perl6::Grammar;statementlist' pc
10:29 polyglotbot ..19622 (src/gen_grammar.pir:2450)␤called from Sub '...
10:29 spinclad perl6: say pi / 2;
10:29 polyglotbot OUTPUT[3.14159␤]
10:29 bacek src/builtins/op.pir
10:29 bacek line 48
10:29 spinclad perl6 seems old; pi / 2; should be working today
10:30 spinclad polyglotbot control restart
10:30 bacek at least 3 days old...
10:30 moritz spinclad: restarting won't rebuild
10:31 spinclad (control restart not even recognized, right?)
10:31 moritz evalbot control restart
10:31 moritz right
10:39 bacek > say e**(1i*pi/2);
10:39 bacek 4.81048+0i
10:39 bacek yak. Parrot's complex's pow...
10:41 Patterner > say e**(1i*pi*2)+1
10:42 spinclad 2.0
10:45 Zaba joined #parrot
10:50 bacek >  say e**(1i*pi*2)+1
10:50 bacek 536.492+0i
10:50 bacek o-ho-ho...
10:51 bacek what is 'magick spell' to auto-install dependencies with cpan?
10:51 moritz o conf prerequisites_policy follow
10:52 bacek moritz++ # just waste some karma :)
10:54 moritz waste implies loss
10:55 bacek it doesn't work... ==> Auto-install the 5 mandatory module(s) from CPAN? [y]
10:56 bacek it was magick environment var IIRC
10:56 moritz yes | cpan -i $module # /me ducks
10:57 * moritz is confused - 'ask' should ask, 'follow' shouldn't ask
10:58 bacek PERL_EXTUTILS_AUTOINSTALL
10:59 bacek moritz, +1
10:59 moritz shouldn't cpan walk the dependency tree and first install the leaves?
11:00 bacek it should. But it asks...
11:01 bacek and still asks...
11:02 bacek probably sudo clean env...
11:06 bacek installing something on old PPC/G4 is slooooow...
11:39 tetragon joined #parrot
11:39 Zaba_ joined #parrot
11:39 stupidbot joined #parrot
11:40 bacek stupidbot?
11:40 bacek seen moritz
11:40 clunker3 moritz was last seen on #parrot 40 minutes and 45 seconds ago, saying: shouldn't cpan walk the dependency tree and first install the leaves?
11:40 stupidbot Sorry, I haven't seen moritz
11:40 bacek a-ha
11:40 bacek I werks!
11:40 bacek It works
11:41 moritz seen bacek
11:41 clunker3 bacek was last seen on #parrot 7 seconds ago, saying: It works
11:41 stupidbot bacek was last seen in #parrot 7 seconds ago saying 'It works'
11:41 bacek karma moritz
11:41 stupidbot moritz has karma of 0
11:41 bacek moritz++
11:41 bacek karma moritz
11:41 stupidbot moritz has karma of 1
11:41 bacek nice
11:41 moritz bacek: which approach did you take? infobot or basicbot?
11:41 bacek seen moritz
11:41 clunker3 moritz was last seen on #parrot 4 seconds ago, saying: bacek: which approach did you take? infobot or basicbot?
11:41 bacek BasicBot
11:42 bacek ::Pluggable
11:44 bacek stupidbot, perl is rulez
11:44 stupidbot bacek: okay.
11:44 bacek perl?
11:44 bacek stupidbot, perl?
11:44 stupidbot bacek: perl is rulez
11:44 bacek Modules: Auth, Karma, Join, Infobot
11:46 bacek RT #44211
11:46 stupidbot Error calling said() for rt: URL must be absolute
11:46 bacek stupidbot...
11:47 bacek RT #44211
11:47 stupidbot Error calling said() for rt: URL must be absolute
11:49 bacek RT #44211
11:49 stupidbot Error calling said() for rt: URL must be absolute
11:50 bacek RT #44211
11:50 stupidbot Error calling said() for rt: URL must be absolute
11:55 bacek http://rt.perl.org/rt3/Tic​ket/Display.html?id=49085
11:55 stupidbot [ #49085: [PATCH] perl6 add pick method to List ]
11:55 bacek nice.
11:56 tetragon I can certainly understand stupidbot's name after seeing this
11:57 bacek tetragon, I can unload RT module this module.
11:57 bacek But bot is actually stupid.
12:05 kj joined #parrot
12:06 Tene So, the docs say that the 'throw' opcode should pass its current continuation to use for resuming, but when I look in throw-exception in exceptions.c, I can't see any evidence that it does anything with the continuation argument passed to it.
12:07 Tene Anyone awake can show me how to resume from an exception?
12:09 Zaba joined #parrot
12:09 bacek Tene, returncc with continuation from exception?
12:10 Tene bacek: any idea how I can get one from the exception without manually making one?
12:11 Tene and stuffing it in the xception before throwing?
12:11 Tene I tried invoking it, but that behaved badly.  I'll try returncc in 30 minutes.
12:12 * Tene applies clothes and goes to work.
12:12 bacek Tene, you probably should use throwcc
12:14 Tene bacek: it's in DEPRECATED.pod, and several docs say that the one-argument version of throwcc is identical to throw, and src/ops/core.ops confirms that.
12:15 bacek Tene, than just call returncc. Everything should 'just works'
12:15 bacek IIRC
12:27 * bacek wanders how to get REST access to rt.perl.org...
12:29 bacek RTFM rule...
12:29 bacek rt#56004
12:29 stupidbot RT 56004d: [TODO] implement .method (with $_ as implicit invocant) - new
12:37 Tene Nope, doesn't work.
12:38 Tene that returns from the sub that the exception handler is in.
12:41 masak the bug in `sub say { }; say "is this printed?"`, has it been reported?
12:41 Tene masak: what is it?
12:41 Tene perl6: sub say { }; say "is this printed?"
12:41 polyglotbot RESULT[Null PMC access in find_method()␤current instr.: '_block11' pc 38 (EVAL_12:18)␤called from Sub 'parrot;PCT::HLLCompiler;eval' pc 806 (src/PCT/HLLCompiler.pir:469)␤called from Sub 'parrot;PCT::HLLCompiler;evalfiles' pc 1088 (src/PCT/HLLCompiler.pir:598)␤called from Sub
12:41 polyglotbot ..'parrot;PCT::HLLCompiler;command_line' pc 1267 (src/PCT/HLLCompiler.pir...
12:42 masak it's that, yes.
12:43 masak I _suppose_ that's an error.
12:43 masak I mean, that this result is unintended.
12:43 masak just wondering if it's been reported as a bug yet
12:44 masak actually, what should perl 6 do if someone tries to override &say et al.?
12:45 avar use the overriden version, the spec supports it
12:45 avar builtins are just like normal subs
12:45 diakopter masak: here's one for you
12:45 diakopter perl6: sub say2 { sub say1 { 3.say2; }; say1(4); }; 5.say;
12:45 polyglotbot OUTPUT[5␤]
12:45 avar there is no builtin in perl6, just a list of stuff that's loaded at startup
12:46 DietCoke sounds like tcl. =-)
12:47 moritz there are no builtins in tcl, just strings ;)
12:47 Tene DietCoke: can you tell me what the SHIM() marker in ops files means?
12:48 DietCoke moritz: that's only a half-truth.
12:48 avar DietCoke: or CL or Haskell
12:48 DietCoke Tene: not off the top of my head.
12:48 DietCoke ISTR that's one of Andy Lester's inventions.
12:48 Tene Okay.
12:48 Tene Yeah, ARGIN, ARGOUT, NOTNULL, I just don't know about SHIM specifically.
12:49 kj SHIM is about an argument that is not used I think
12:49 Tene That's what I thought.
12:49 kj ehm, paramter I should say
12:49 Tene Well, what I feared.
12:49 Tene So, is allison still reworking exceptions, or is that done and merged?
12:49 kj but I think if you check out the headerizer script (perldoc) there's some documentation on that
12:52 DietCoke Tene: it's the pdd25cx branch; it's still unmerged, sfaik.
12:52 Whiteknight joined #parrot
12:52 masak diakopter: not sure I got the point of that line of code
12:53 Tene Lemme check to see if this works properly in that branch.
12:53 diakopter masak: sorry; see #perl6 - I was playin around
12:55 masak diakopter: ah, ok
12:55 dalek r28674 | Whiteknight++ | gsoc_pdd09:
12:55 dalek : [gsoc_pdd09] update to trunk r28673
12:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28674
12:55 diakopter rakudo doesn't detect indirect infinite recursion (depth)
12:57 Tene Looks like this is implemented a lot better in pdd25cx... let's see if it runs like I think it should...
12:57 Tene Ack, timeforclass.
12:58 diakopter nm; rakudo *does* detect indirect infinite recursion
12:58 diakopter it's pugs that doesn't
13:07 masak . o O(halting problem...)
13:09 moritz masak: it's actually possible in 99% of all real-world (non-perl) code to decide if it halts
13:09 masak moritz: yes
13:11 Zaba_ joined #parrot
13:11 moritz it's typical for mathematics to encounter problems as soon as a construct is powerful enough to reference to itself
13:12 masak moritz: the authors of 'Winning Ways' make a similar point about NP-complete subproblems in games
13:12 masak they argue that they have shown interesting solutions to subproblems of problems known to be in NP
13:12 moritz Turing's proof ussed an automaton which counts other automatons. Likewise you can easily get into troubles when you have sets that can consist of arbitrary sets as items
13:12 moritz masak: sounds interesting ;)
13:13 masak moritz: yes. the Java Collections framework actually has a small note about avoiding Collections containing themselves
13:13 masak but that falls out of a wider faux pas, namely having Collections of mutable objects
13:14 gryphon joined #parrot
13:14 masak moritz: it's a good pair of books. you should take a look
13:14 masak they basically redefine the real numbers to better suit analysis of two player games
13:14 masak ending up with the "surreal numbers"
13:14 moritz ;)
13:15 masak ...of Donald Knuth fame
13:15 masak perhaps the only fictional book Knuth wrote, I don't know
13:15 masak http://www-cs-staff.stanford.edu/~uno/sn.html
13:15 stupidbot [ Knuth: Surreal Numbers ]
13:16 masak but I'm OT-ing. I'll stop now :)
13:17 Zaba joined #parrot
13:18 moritz we need #perl6-ot or #parrot-ot
13:18 moritz ;-)
13:19 masak #perl6 has often filled that purpose
13:19 masak at least in the pugs glory days
13:20 masak smart people who got together and threw smart ideas at each other
13:21 moritz now it's only us ;)
13:21 masak *lol*
13:22 Zaba_ joined #parrot
13:37 rdice joined #parrot
13:38 DietCoke hey, richard.
13:40 tewk joined #parrot
13:43 Tene No, returncc doesn't work in that situation, but it looks like parrot *does* pass the continuation off to the error handler in the pdd25cx branch.
13:43 Tene I just need to figure out how to get it.
13:44 Tene VTABLE_invoke(interp, handler, dest);
13:48 tewk nqp seems to install an implicit excpetion handler in each parser action.  Is there any way I can explicitly install an exception hander as a debugging aid to see whats going wrong.
13:53 dalek r28675 | fperrad++ | libs4php:
13:53 dalek : [php] some string
13:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28675
13:53 stupidbot [ revision: /parrot (Rev: 28675, via SVN::Web) ]
13:54 moritz bacek: could you teach stupidbot not to act on URLs printed by dalek?
14:03 stupidbot joined #parrot
14:03 GeJ joined #parrot
14:03 bacek moritz, done
14:03 moritz bacek++
14:05 GeJ joined #parrot
14:06 pmichaud tewk:  the exception handler that nqp installs is just to handle return exceptions.
14:14 bacek karma moritz
14:14 stupidbot moritz has karma of 1
14:15 bacek karma bacek
14:15 stupidbot bacek has karma of 1
14:15 bacek heh. We now have equal karma :)
14:15 moritz that's so cheating ;)
14:15 bacek it is not Olympic Games :)
14:16 DietCoke what happened to purl?
14:16 DietCoke opbots, trust me
14:16 clunker3 But I do not trust you DietCoke
14:16 slavorg But I don't trust you there, DietCoke
14:16 Coke opbots, names
14:17 bacek DietCoke, purl is drinking with Bender...
14:28 moritz !leave #parrot
14:28 moritz stupidbot: !leave #parrot
14:28 stupidbot moritz: Admin commands in privmsg only, please
14:28 Zaba joined #parrot
14:30 stupidbot left #parrot
14:31 stupidbot joined #parrot
14:35 jhorwitz joined #parrot
14:35 davidfetter joined #parrot
14:38 bacek pmichaud, rt#55438 is waiting :)
14:38 stupidbot RT 55438d: [PATCH] Patched Test.pm and test for Test.pm - open
14:40 bacek test3.diff especially
14:40 bacek afk # sleep
14:40 bacek have a good night everyone
14:41 Tene g'night
14:42 Tene If I can manage to resume after an exception, I should have a workable gather/take.
14:43 moritz Tene: in rakudo?
14:43 Zaba_ joined #parrot
14:43 Tene Yeah.
14:43 pmichaud I'm not sure if the current Parrot supports resumable exceptions.
14:44 pmichaud I know that the pdd25cx branch is supposed to do that.
14:44 DietCoke does "resumable" mean something special in this case?
14:44 pmichaud it means that I can have execution continue from the point where the exception was thrown.
14:44 Tene I tried manually stuffing a continuation into the exception, but that behaved poorly.
14:45 Tene eh, maybe I could get that to work.
14:45 Tene Or I could just wait for pdd25cx
14:45 pmichaud I've been kinda waiting for pdd25cx to think about it too heavily.  :-)
14:46 pmichaud we could ask for a pdd25cx eta at today's #parrotsketch
14:46 Tene That's an excellent plan.
14:46 Tene (clever plans)++
14:46 moritz speaking of ETAs, any idea if (and if yes, when) I can get a commit bit?
14:47 DietCoke moritz: you submitted a CLA?
14:47 moritz DietCoke: yes
14:47 DietCoke Might be my fault, then.
14:48 moritz well, I don't know how long the snail mail takes over the atlantic , but it should be there now
14:49 Tene That means I need to find something different to hack on.
14:51 pmichaud both particle and I agreed to be mentors for moritz on commitbits  (http://irclog.perlgeek.de/par​rotsketch/2008-06-03#i_327702)
14:51 stupidbot [ IRC log for #parrotsketch, 2008-06-03 ]
14:52 moritz Tene: anything particular you're interested in?
14:52 NotFound moritz: according the post office I consulted here in Spain, 4-6 days.
14:52 moritz NotFound: the post office here said "usually one week, but it can be up to 3, and sometimes even 6"
14:53 moritz ie "Deutsche Post" sucks
14:53 Tene moritz: HLLish stuff.
14:53 dalek r28676 | pmichaud++ | trunk:
14:53 dalek : [rakudo]:
14:53 dalek : * Update spectest-progress.csv
14:53 dalek : * 66 files, 849 passing tests
14:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28676
14:53 NotFound moritz: i think yours are more sincere than mine.
14:53 * pmichaud starts preparing his parrotsketch reports.
14:53 stupidbot [ revision: /parrot (Rev: 28676, via SVN::Web) ]
14:53 DietCoke moritz: Yup, I see it in the list of contributor agreements. Did you follow the steps in docs/project/*commit* (and if so, give me your id)
14:54 moritz Tene: there's some string interpolation stuff that needs doing, for example - should be rather HLLish
14:55 moritz DietCoke: will do now
14:57 moritz DietCoke: my id is moritz
14:57 moritz (how surprising)
14:59 pmichaud actually, I was a bit surprised when I couldn't get PMICHAUD as my pause id.  :-)
14:59 moritz why not? taken? or too long?
14:59 pmichaud taken by the other "Patrick Michaud"  :-)
15:00 pmichaud which also confused the housing folks at YAPC::NA again this year, too.  :-P
15:00 tewk pmichaud: No result object
15:00 moritz pmichaud: you need a second name ;)
15:00 Tene like 'tene'.  'tene' isn't taken in too many places.
15:00 pmichaud when Patrick (A) Michaud checked out on Wednesday, they accidentally checked him out of my room.
15:01 jhorwitz LOL
15:01 moritz (I actually have a second name, but never use it. They are mostly ignored in Germany, except on legal documents)
15:01 pmichaud then when I checked out on Thursday, they tried to tell me that I had overstayed by a day and owed $150
15:01 moritz nice
15:01 jhorwitz Tene: where are you?
15:02 pmichaud my middle initial is (R).  Normally I use it on everything (always have) but for some reason it didn't show up in the YAPC::NA registrations.
15:02 Tene jhorwitz: http://maps.google.com/maps?f=q&amp;hl=en&amp;ge​ocode=&amp;q=1201+first+street,+king+of+prussia,​+PA&amp;sll=40.659853,-111.922492&amp;sspn=0.006​763,0.006652&amp;ie=UTF8&amp;z=17&amp;iwloc=addr
15:02 stupidbot [ Google Maps ]
15:02 shorten Tene's url is at http://xrl.us/fm62a
15:02 stupidbot [ Google Maps ]
15:02 jhorwitz 2 miles away.  :)
15:03 moritz the other day somebody with the same first and last name than me asked me if he could have my gmail address ;)
15:04 Tene MOAR UUIDZ
15:04 moritz ;)
15:06 jhorwitz Tene: what's your schedule?  philly.pm is having a dinner in center city with Alias tonight if you'd like to join us, and i'll be around for lunch tomorrow.
15:06 Tene lunch won't work for me.  Tonight might be good.
15:06 Tene Where?
15:09 DietCoke pmichaud: did they eventually straighten that out?
15:10 pmichaud yes, it was pretty easy for me to make my case that I was the one in rm 324, since I had the key.  :-)
15:11 pmichaud also, my driver's license does show my middle initial as 'R' instead of 'A' :_)
15:18 jhorwitz Tene: sorry, got distracted.  we're at nodding head in center city.  it's at 7, but i'm not leaving til about 8.
15:19 jhorwitz if you want a ride down there, i can provide.
15:22 pmichaud "Nodding Head"?  "King of Prussia"?  What's up with those names in that part of the world?
15:22 pmichaud not to mention "Pennsylvania".  :-P
15:22 pmichaud me, I'm in "Plano".  "Texas".  It's just north of "Dallas".
15:23 * moritz thinks of the village called "Niederdodeleben"
15:24 pmichaud our capital is "Austin" -- not something like "Phil-a-del-phi-a"
15:25 jhorwitz pmichaud: nodding head is a bar
15:25 pmichaud "Gilley's"  :-)
15:26 jhorwitz and philadelphia is greek.  it means "beat you senseless"
15:26 pmichaud (although technically I guess that's in Houston)
15:26 jhorwitz er, i mean "city of brotherly love"
15:26 pmichaud I'm not sure where Dallas comes from, but I think it means "filthy rich"  :-)
15:27 Khisanth joined #parrot
15:33 Tene jhorwitz: looks like I'll be done with work at about 4:00
15:33 jhorwitz Tene: are you staying in KOP?
15:34 Tene A co-worker is talking about places called "Llandudno", "Denbigh", and "Llanelwy"
15:34 Tene jhorwitz: The training center and hotel are both at the same place.
15:34 jhorwitz ok
15:34 Tene jhorwitz: also, I have a car and won't need a ride.
15:35 jhorwitz ok, cool.  i'll forward the details to you.  msg me your e-mail?
15:37 Tene I've got a friend to visit tonight too, so I might not come.
15:38 jjore joined #parrot
15:44 Theory joined #parrot
15:46 Zaba joined #parrot
15:53 DietCoke hurm. technically my city's name is "Castleton on Hudson" which is a little more verbose, but not too bad. And patrick, you live in a state with a town named "Corpus Christi", so you're not completely off the hook. =-)
15:59 DietCoke moritz: you've been added.
15:59 DietCoke pmichaud: would you can to send out the email to the list?
16:00 DietCoke moritz: try a small change to verify. We occasionally have issues, nice to make sure it works.
16:03 pmichaud moritz: can add yourself to CREDITS
16:07 sjansen joined #parrot
16:11 pmichaud afk: lunch, errands
16:11 pmichaud I'll be back for #parrotsketch  (in 139 minutes)
16:13 pmichaud (Corpus Christi)  Oh, yeah.  I even work for a place called "Texas A&M University-Corpus Christi", which tends to be too long for various web forms to handle.  Sometimes I have to abbreviate it down to "TAMU-Corpus" or something weird like that.  :-P
16:17 iblechbot joined #parrot
16:22 pmichaud at least none of these are as bad as "St. John the Baptist" parish in Louisiana.  :-)
16:22 pmichaud really gone.
16:34 Zaba_ joined #parrot
16:43 smash joined #parrot
16:43 smash hello everyone
16:49 tewk I'm trying to figure out parse actions, are they called in post order traversal?
16:52 particle joined #parrot
16:56 mj41 hi, I adder compare feature to TapTinder ... see http://tt.ro.vutbr.cz/report/pr-Parrot/​diff?trun-708=on&amp;trun-707=on&amp;tr​un-706=on&amp;Submit=Compare+selected
16:56 stupidbot [ Results diff for Parrot - TapTinder reports ]
16:56 shorten mj41's url is at http://xrl.us/fns8o
16:56 stupidbot [ Results diff for Parrot - TapTinder reports ]
16:58 davidfetter boa noite, smash
17:00 smash davidfetter: boa noite
17:01 davidfetter what's new & good?
17:03 tewk pmichaud:  I'm trying to figure out parse actions, are they called in post order traversal?
17:04 PerlJam "post order traversal"?
17:04 tewk I'm getting "No result object" current instr.: 'parrot;C99;Grammar;Actions;cast_expression' pc 97665 (src/gen_actions.pir:198)
17:04 PerlJam They are called in parse order as the {*} are encountered  :-)
17:04 tewk PerlJam: in other words coming up the tree from the leaves.
17:05 Tene Mmm... iptables lecture.  :)
17:06 tewk PerlJam: http://en.wikipedia.org/wiki/Tree_traversal
17:06 stupidbot [ Tree traversal - Wikipedia, the free encyclopedia ]
17:06 PerlJam tewk: I know all about tree traversal, but read my second statement.
17:07 tewk Ok got it.
17:09 Tene So, if anyone's looking for work, this job is a lot of fun.
17:14 cognominal joined #parrot
17:19 barney joined #parrot
17:21 DietCoke parrotsketch in 69m
17:22 cotto_work barney, any thoughts on "Pipp"?
17:22 tewk I'm an idiot, I had #= keys on some grammar alternatives, but no {*}
17:25 tewk pmichaud: Could pct generate a warning in a case like that?
17:26 barney Both Pipp and Pippit are fine. Pipp is a bit closer to PHP, which is nice. Pipit is a bird so it fits better with Parrot.
17:27 tewk pmichaud++ nqp is very nice.
17:27 * particle wonders if "[ 'a' #= a n | 'b' #= b n ] {*}" would pass a or b to the action method when {*} is called
17:27 particle hi all, back from travel
17:28 DietCoke hio, particle
17:29 particle tewk: if my above example is true, than the warning case is rather complex
17:29 PerlJam particle: I think the #= foo has to come after the {*}
17:30 particle yeah, i need to search pge source
17:30 PerlJam particle: so, in that case, nothing would be passed to the action but $?
17:30 PerlJam er, $/
17:30 cotto_work barney, I'll see what it'd take to get pipp.org or .com.
17:30 tewk particle: yeah, I thought the same thing.
17:30 cotto_work although there is some charm in having the project named after a bird that eats bugs
17:31 particle however, it should be a warning to have {*} in an alternation without #=
17:31 Zaba joined #parrot
17:31 particle err
17:31 tewk What does separating {*} and #= buy us.
17:31 particle i mean, in an alternation with multiple items, where at least one has #=
17:31 particle separating? how?
17:31 particle {*} is valid perl 6, and #= is a comment
17:32 PerlJam tewk: the ability to modulate the action
17:32 particle however, the nature of the magical #= comment makes it more or less *not* a comment
17:34 PerlJam tewk: If you only had {*} then the actions would need to inspect $/ each time to figure out what it was supposed to do (and that's not always an easy proposition).  With #= you can do a simple string comparison and know exactly where you are in the parse.
17:36 tewk PerlJam: I understand that.  #= seems like a hack because {*} doesn't have sufficient features for the common use cases.  But I don't understand what {*} means in perl6.
17:37 PerlJam tewk: What features would you like it to have?
17:37 particle it's a closure with a Whatever inside
17:37 Whiteknight joined #parrot
17:37 particle meaning, "i'll define the parse action later"
17:38 tewk It appears that calling parse actions should have arguments, so why not put the arguments at the call site instead of in a comment.
17:39 PerlJam tewk: you should troll the mailing list archives for that discussion  :)
17:40 * DietCoke feels a little wierd knowing what some of you look like now. :P
17:40 * DietCoke summons purl
17:40 Whiteknight You don't know what I look like yet, that should be comforting
17:41 DietCoke yay!
17:41 DietCoke anonymous internet, the way compuserve intended.
17:41 DietCoke ah, qlink, we miss you.
17:41 Ivatar joined #parrot
17:41 tewk PerlJam: I vaguely remember seeing that discussion, was it on perl6-language, perl6-compiler, or parrot.
17:42 * DietCoke guesses not parrot
17:42 dalek r28677 | bernhard++ | trunk:
17:42 dalek : [Plumhead (Pipp or Pipit)]
17:42 dalek : Plumhead_gen.pir now in gen_grammar.pir.
17:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28677
17:43 cotto_work (picking an OSS project name)--
17:43 TimToady the closure already gets the complete match state--what more can it want? :)
17:43 cotto_work (finding a good one)++
17:43 smash eta for ps ?
17:44 smash 45m ?
17:44 cotto_work smash, yes
17:44 smash cotto_work: ok, thank you
17:45 Zaba_ joined #parrot
17:46 dalek will@coleda.com | Field Guide to Parrot Porters:
17:46 dalek link: http://www.perlfoundation.org/parrot/i​ndex.cgi?field_guide_to_parrot_porters
17:46 shorten dalek's url is at http://xrl.us/fn5dg
17:46 stupidbot [ Field Guide to Parrot Porters / Parrot ]
17:46 stupidbot [ Field Guide to Parrot Porters / Parrot ]
17:47 tewk TimToady: Then why have #=?  Maybe we shouldn
17:47 tewk t
17:47 TimToady tewk: the point of separating the the #= is to separate out the semantic info from the syntactic
17:47 dalek will@coleda.com | Field Guide to Parrot Porters:
17:47 dalek link: http://www.perlfoundation.org/parrot/i​ndex.cgi?field_guide_to_parrot_porters
17:47 shorten dalek's url is at http://xrl.us/fn5dg
17:47 TimToady otherwise they obscure each other by intermixing
17:47 stupidbot [ Field Guide to Parrot Porters / Parrot ]
17:47 stupidbot [ Field Guide to Parrot Porters / Parrot ]
17:47 * DietCoke wonders if dalek is smart enough to not tell us about the 4 times I did that.
17:47 pmichaud also, {*} is visually nicer.
17:47 DietCoke uh, stupid bot? you're incredibly stupid.
17:48 TimToady just like stupidbots obscure the discussion by intermixing
17:49 Zaba joined #parrot
17:49 pmichaud but if someone dislikes {*} and #=, there's always the possibility of doing something like  <action('key')>
17:49 pmichaud which, if you want separate actions for each rule, probably needs to be
17:49 pmichaud <rulename_action('key')>
17:49 TimToady and the {*}...#= format can easily be preprocessed into <action()>
17:49 pmichaud correct.
17:50 TimToady but the point is that the {*} is independent of the use of the grammar, while the #= isn't
17:50 Whiteknight I personally like that form, it's more consistent with the rest of the P6 grammar, I think
17:50 pmichaud anyway, to answer the earlier question about "what #= applies in PGE" -- only any #= that is on the same line as the {*}
17:51 TimToady yes, the line-orientation is also to be construed as preprocessor-friendly
17:51 tewk so in particle's example, the first #= wins?
17:51 pmichaud in particle's example, the {*} doesn't have a #=
17:52 TimToady all #= assume the name of the rule on the front of the rest of the key
17:52 TimToady so a bare #= would assume a key that is the same as the rule name
17:52 TimToady and would probably be erroneous in an alt
17:52 TimToady unless your action is smart enough to figure out which alt it's in
17:53 TimToady anyway, STD should have all alts labeled with #=
17:53 TimToady if not, it's a buglet
17:53 TimToady er, not "all alts", but "all alts with {*}"
17:54 TimToady I tend to leave off the {*} on alts that are going to be part of another token, since {*} is supposed to be equiv to :: and terminate the longest token pattern
17:54 dalek will@coleda.com | Field Guide to Parrot Porters:
17:54 dalek link: http://www.perlfoundation.org/parrot/i​ndex.cgi?field_guide_to_parrot_porters
17:55 stupidbot [ Field Guide to Parrot Porters / Parrot ]
17:55 shorten dalek's url is at http://xrl.us/fn5dg
17:55 tewk pmichaud: ahh, I overlooked the \n in particles example.
17:55 stupidbot [ Field Guide to Parrot Porters / Parrot ]
17:55 DietCoke particle: added you to that page, since I was trivially able to find your gravatar.
17:55 TimToady one of my todos is to figure out when it's appropriate to override a subrule's {*} to incorporate one token in another
17:55 TimToady (if ever)
17:55 dalek r28678 | bernhard++ | trunk:
17:55 dalek : [Plumhead PHC]
17:55 dalek : Reformat some long lines in plumhead/src/phc/phc_xml_to_past_xml.xsl
17:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28678
17:56 stupidbot [ revision: /parrot (Rev: 28678, via SVN::Web) ]
17:56 DietCoke who is running stupidbot?
17:56 DietCoke win 7
17:57 pmichaud I think stupidbot is bacek's
17:58 DietCoke bacek: I have a ton of feature requests if you're going to leave him in here. =-)
17:59 pmichaud feature request:  bot should not speak unless spoken to.
18:00 NotFound First law of botics.
18:02 wknight8111 joined #parrot
18:14 chromatic joined #parrot
18:16 dalek r28679 | bernhard++ | trunk:
18:16 dalek : [Plumhead PHC]
18:16 dalek : Add and use rule <statement_delimiter>.
18:16 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28679
18:16 stupidbot [ revision: /parrot (Rev: 28679, via SVN::Web) ]
18:17 * Tene sends decietful pictures to DietCoke.
18:18 cjfields joined #parrot
18:20 dalek r28680 | Whiteknight++ | trunk:
18:20 dalek : [src/gc] Improve function-level documentation with imperative mood and better POD. Suggestion by chromatic++
18:20 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28680
18:20 DietCoke there are some things I don't mind, but having the bots feeding off each other is a pita.
18:20 stupidbot [ revision: /parrot (Rev: 28680, via SVN::Web) ]
18:20 was kicked by DietCoke: stupidbot
18:21 Whiteknight Agreed, inter-bot relations are disgusting and amoral
18:21 DietCoke there, that'll do for a first approximation. =-)
18:21 Whiteknight bots should stick away from their own kind
18:23 Tene Perhaps the bots should send those as notices to the channel instead of msgs.
18:23 Tene That's the standard way to send a "NO AUTOMATED RESPONSES TO THIS" message.
18:27 pmichaud parrotsketch in 3
18:34 DietCoke PS in progress. the bird has left the cage.
18:36 Zaba_ joined #parrot
18:43 smash chromatic: what would be the first approach on completly rebuild IMCC ?
18:43 chromatic I'd start with PIRC, which already parses PIR.
18:44 chromatic PIRC right now doesn't know how to make PBC, either in memory or on disk.
18:44 * smash nods
18:44 smash but PIRC parser is done from scratch
18:44 jonathan I think we need some generic back-end that we can feed POST to as well, so it doesn't have to be taken to text form.
18:45 chromatic Like bytecode PMCs?
18:45 jonathan Not really, just a generic code-generation interface of some kind.
18:45 jonathan There was some work in compilers/bcg on that.
18:46 chromatic *something* has to create PBC somewhere.
18:47 * cjfields picturing Doug Henning: "It's magic!"
18:48 * cjfields realizing that last post dates me
18:49 davidfetter heh
18:51 dalek r28681 | moritz++ | trunk:
18:51 dalek : added myself to CREDITS as a commit bit test
18:51 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28681
18:53 Tene Hmm.  Perhaps I should add myself to CREDITS someday.
18:54 chromatic Or update mine.
18:55 smash my 'make test' hangs for branch pdd25cx/ in t/src/extend, is this normal ?
18:56 slightlyoff joined #parrot
18:58 magnachef_ joined #parrot
19:02 barney Tene: I already did
19:02 Tene oh, you used my given name
19:07 cognominal joined #parrot
19:10 dalek r28682 | tene++ | trunk:
19:10 dalek : Update my CREDITS entry
19:10 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28682
19:12 pmichaud chromatic: I think jonathan's idea (mine as well) is to go directly from POST to bytecode.
19:12 particle compilers/bcg++
19:12 DietCoke ... wasn't jonathan going to be working on bytecode at some point? =-)
19:12 * particle pokes Infinoid
19:12 chromatic Yeah, we need to generate bytecode somehow from something.
19:12 particle perl 5 works fine without bytecode! :P
19:13 chromatic FSVO "works" and "bytecode"
19:13 pmichaud heh.  Let's re-write imcc in P5.  :-)
19:13 particle i wonder how many cpan modules we can use
19:13 chromatic It'd have fewer memory leaks, that's for sure.
19:14 pmichaud even better -- let's rewrite imcc in rakudo.  Then we can say that rakudo is self-hosting, even though parts of it are written in PIR.  :-P
19:16 * particle has left #sanity
19:17 pmichaud (I'm only half-joking about rewriting imcc in rakudo, btw. :-)
19:17 slightlyoff left #parrot
19:17 avar I knew self-hosting was all a big conspiracy!:)
19:19 Whiteknight If we had a bytecode file that parsed PIR, we could use that instead of IMCC, right?
19:19 chromatic Yes.
19:19 Whiteknight use it one last time, and kiss IMCC goodbye forever
19:21 pmichaud and hope we don't lose the bytecode chain :-P
19:22 japhb pmichaud: That's what SVN history is for
19:24 tewk So is PIRC hand rolled recursive decent?
19:24 wknight8111 from what I've seen, PIRC seems to be pretty standard lex/yacc
19:25 chromatic Yes.
19:26 wknight8111 I want to take a stab at improving it, once I get this GC project out of the way
19:26 wknight8111 assuming I still have the time, once I get a $real_job
19:27 wknight8111 and once $new_wife.wants_children == true
19:28 nopaste "Coke" at 72.228.52.192 pasted "backtrace for allison" (234 lines) at http://nopaste.snit.ch/13383
19:29 dalek r28683 | bernhard++ | trunk:
19:29 dalek : [Plumhead PCT]
19:29 dalek : Use rule 'function_call' for 'php_sapi_name' and 'var_dump'.
19:29 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28683
19:30 smash wknight8111: if i remember correctly PIRC doesn't uses lex/yacc
19:32 barney The compilers/pirc README says the flex/bison is used
19:32 tewk I think smash is write, pct/src looks like hand coded recusive decent. pct/new looks like flex/bison
19:32 wknight8111 flex/bison ~~ lex/yacc, as far as I am concerned
19:32 tewk s/write/right/
19:33 wknight8111 oh, maybe that's what I'm looking at, pirc/new
19:33 barney Why isn't there a PCT based PIR parser?
19:34 wknight8111 That would be necessary if we were seriously considering extreme bootstrapping
19:34 tewk PCT is slow, but that doesn't mean we couldn't write one.
19:34 smash C is faster
19:35 wknight8111 not for the developers :)
19:35 barney Ah there is languages/PIR .
19:35 tewk C based pir parsing is probably going to be the fastest for the near future, faster than pbc too.
19:36 nopaste "tene" at 160.79.186.34 pasted "failures in pdd25cx" (10 lines) at http://nopaste.snit.ch/13384
19:36 wknight8111 I'm interested in the lex/yacc PIRC implementation, once I have time to work on it
19:36 wknight8111 that should make everything much better
19:37 Tene allison: you said 6 failures.  can you tell me which of the failures in my paste here you're not also seeing?
19:37 particle errands &
19:37 cognominal joined #parrot
19:37 DietCoke pmichaud: I think the HLLCompiler stuff is coming in via PGE, based on a -t4
19:38 barney There are probably better parser generators than lex/yacc. I could whip up an ANTLR prototype.
19:38 pmichaud hrm.
19:39 pmichaud $ ack HLL compilers/pge
19:39 pmichaud $ ack PCT compilers/pge
19:39 pmichaud $
19:39 DietCoke I'm just looking at the -t4 output. I can nopaste it, hang on.
19:40 pmichaud after this past weekend, I really don't trust the -t4 output.
19:40 nopaste "Coke" at 193.200.132.135 pasted "pir backtrace with -t4" (2386 lines) at http://nopaste.snit.ch/13385
19:42 DietCoke # current instr.: 'parrot;Tcl::Glob;__onload' pc 136 (runtime/parrot/library/Tcl/Glob.pir:86)
19:42 DietCoke that's where PCT comes in.
19:42 nopaste "pmichaud" at 76.183.97.54 pasted "-t4 annoyances" (31 lines) at http://nopaste.snit.ch/13386
19:43 pmichaud oh.  Glob might use HLLCompiler.
19:43 DietCoke ayup
19:44 DietCoke it registers itself as a Tcl::Glob::Compiler or somesuch.
19:46 slightlyoff joined #parrot
19:46 DietCoke Looking at the 'init' sub in HLLCompile
19:46 slightlyoff left #parrot
19:47 dalek r28684 | bernhard++ | trunk:
19:47 dalek : [Plumhead PCT]
19:47 dalek : Align rule and token names.
19:47 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28684
19:54 dalek r28685 | allison++ | pdd25cx:
19:54 dalek : [pdd25cx] For backward compatibility, delete exception handler when it's first used.
19:54 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28685
19:55 DietCoke allison's patch unsticks tcl, but now I'm failing a LOT of tests.
19:59 dalek r28686 | bernhard++ | trunk:
19:59 dalek : [codingstd]
19:59 dalek : Get rid of a hard tab.
20:00 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28686
20:01 jhorwitz I CAN HAS MOD_LOLCODE? http://www.smashing.org/sandbox/cheezburger.lol
20:02 davidfetter w00t!
20:02 DietCoke I IS FULL
20:03 jhorwitz he gets full after 100 cheezburgers
20:03 Auzon jhorwitz++
20:04 barney the cat can uneat
20:04 jhorwitz you made him eat -1 cheezburgers?  LOL
20:05 DietCoke heh. add a hairball pic if he drops below 0. =-)
20:06 DietCoke jhorwitz: you scare me.
20:06 jhorwitz that's what i'm here for
20:06 davidfetter heh
20:06 barney Can he ead 'DELETE FROM PUPILS' cheeseburgers?
20:06 barney s/ead/eat/
20:06 davidfetter http://xkcd.com/327/
20:07 jhorwitz LOL
20:08 jonathan jhorwitz++
20:10 apeiron http://www.smashing.org/sand​box/cheezburger.lol?eated=a # heh
20:11 apeiron http://www.smashing.org/sandbo​x/cheezburger.lol?eated=eated # 50
20:11 apeiron er, 500
20:12 jhorwitz LOLCODE doesn't quite have the right features for input validation.  ;-)
20:12 moritz ISTHISREALLY string regex ? ;-)
20:13 moritz sorry, IZRLY perhaps
20:13 barney Argh, no LOLCODE user group in Munich
20:13 * jhorwitz should reset the globals after every invocation
20:13 davidfetter itym YARN
20:14 * DietCoke bets everyone reading allison's blog post is wondering who the hell the 5th member is. =-)
20:16 jhorwitz you mean the 4th?
20:17 pmichaud that "Shane" guy.  :-)
20:19 * tewk needs a nqp syntax file for vim.
20:20 moritz tewk: take the perl 6 one
20:20 moritz http://svn.pugscode.org/pugs/util/perl6.vim
20:20 moritz not perfect, but better than nothing
20:21 mj41 http://www.gossamer-threads.c​om/lists/python/python/658656
20:21 mj41 Mailing List Archive: Python - Python 3000 vs Perl 6
20:22 jhorwitz my name was 5th, so i was confused.  then i saw "shane".  :)
20:22 dalek r28687 | chromatic++ | trunk:
20:22 dalek : [config] Made config::auto::perldoc set the proper path to perldoc based on the
20:22 dalek : version of Perl used to configure Parrot.  The generated docs Makefile now uses
20:22 dalek : that value, rather than relying on the user's PATH to find the right version.
20:22 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28687
20:26 tewk moritz++
20:28 * apeiron notes the several jibes at perl and parrot on that page, sighs
20:30 chromatic Oh, how easy for the Pythonistas to forget how long Guido has been talking about Python 3000....
20:34 cognominal the big point for the programmer in the long term is how flexible is the lanuguage so as to incorporate now unthouht abstractions.
20:34 NotFound Will be funny if someone write python 3000 on parrot before any other implementation X-)
20:35 cognominal python and ruby for me are mostly perl 5 without wart and a better object model.
20:37 NotFound It's amusing that people wants to make speed comparaisons between two still non existent language implementations.
20:37 apeiron Well of course, speed is everything! </sarcasm>
20:38 Auzon Odd, I thought Python's site said that they preferred clarity of the interpreter's code rather than having the interpreter be faster
20:38 NotFound Even better is the point: I don't care about speed, provided that will be fast.
20:39 DietCoke *blink*
20:39 DietCoke *blink*
20:40 cognominal well, the question will be about speed on multicores.  in 15 years, people will not speak of MHz but of kilocores.
20:41 cognominal I can't imagine what hell that will be
20:41 cotto_work a good reason to work on concurrent GC, if nothing else
20:42 NotFound Following some recent trends, maybe the better way will be to develop some way to ray-tracing bytecode execution.
20:43 chromatic Ray tracing?
20:43 chromatic bounce photons off of each opcode?
20:43 NotFound Something like that, yes.
20:44 cognominal or ADN : slow but if you get enough strands...
20:44 NotFound Stop using pesky electrons to run code, photons are lighter.
20:45 wknight8111 is there a language spec yet for Python 3000?
20:46 wknight8111 we should seriously court the python people, try to get them using Parrot for their future needs
20:47 chromatic Oh yeah, NotFound, how's your Valgrind fu?
20:48 DietCoke chromatic: no tracing my wife.
20:49 NotFound chromatic: I know very little about valgrind, a few time I try to use it don't get other thing that fake positives.
20:52 DietCoke wknight8111: there is a history with the python people that might make it easier to have 1.0 under our belt before continuing that.
20:53 chromatic NotFound, there are a couple of memory leaks in places like IMCC that I think you might be able to track down.
20:53 chromatic $ alias vgp
20:53 chromatic alias vgp='valgrind --suppressions=/home/chromatic/d​ev/parrot/tools/dev/parrot.supp --num-callers=500 --leak-check=full --leak-resolution=high --show-reachable=yes parrot --leak-test'
20:57 wknight8111 DietCoke, you tell the python people that wknight8111 said so. They'll listen to me
20:58 wknight8111 actually, they probably never heard of me :(
20:58 DietCoke I think we oversold initially.
20:58 purl joined #parrot
20:58 DietCoke and by we I don't mean me. =-)
20:58 DietCoke purl: you shiny bot!
20:58 purl DietCoke: sorry...
20:58 chromatic See "A Comprehensive History of TPF Press Releases, Volumes I-IV: 2001-2003"
20:58 wknight8111 You mean parrot doesn't end war, cure the common cold, and lower the price of gas?
20:59 * wknight8111 has become disillusioned
20:59 DietCoke it does seem to encourage me to produce more gas.
20:59 DietCoke ask chromatic.
20:59 * DietCoke drinks a diet peach snapple
20:59 * DietCoke is kind of creeped out that chromatic is here.
21:00 NotFound chromatic: first thing I see is that there are a lot of malloc calls in the compilers/imcc directory. It's not supposed to use parrot allocation functions?
21:00 chromatic Switching those over to mem_sys_* would be better.
21:00 chromatic (eventually we can use sized header pools to manage them)
21:01 chromatic (now someone revive wknight8111, please)
21:01 wknight8111 oh, i'm revived
21:01 wknight8111 and I'm slowly adding items on my TODO list
21:01 NotFound There is also a hand written malloc declaration ! =:o
21:04 NotFound Ah, it's a lex generated one.
21:04 NotFound Bison, not lex.
21:06 * DietCoke gets the smelling salts.
21:15 davidfetter free() the malloc()s!
21:26 NotFound By the way, other problem I noted some days ago is the debug.h file name.
21:26 NotFound We have include/parrot/debug.h and compilers/imcc/debug.h
21:27 NotFound And I'm almost sure that the included one is not always the desired.
21:28 moritz rename them to parrot_debug.h and imcc_debug.h?
21:29 NotFound Given his usage, include/parrot/debug.h shall be pdb.h
21:30 NotFound Or if a longer name is wanted, parrot_debugger.h will be most clear.
21:34 dalek r28688 | Whiteknight++ | trunk:
21:34 dalek : [src/gc] Improve function-level documentation for src/gc/dod.c with imperative mood and better POD. Suggestion by chromatic++
21:34 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28688
21:41 zarchne joined #parrot
21:56 NotFound Not so easy, part of pdb functionallity is inside src/ops/debug.ops
21:58 tewk ll
22:09 Limbic_Region joined #parrot
22:18 Whiteknight joined #parrot
22:36 dalek r28689 | moritz++ | trunk:
22:36 dalek : [rakudo] add passing test num.t to spectest_regression
22:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28689
22:37 moritz =kwid
22:37 moritz basic Boolean tests
22:37 moritz =cut
22:37 moritz sorry
22:40 lichtkind_ joined #parrot
22:45 dalek r28690 | chromatic++ | pdd25cx:
22:45 dalek : [t] Fixed failing Sub PMC tests -- they relied on a wild, stab-in-the-dark
22:45 dalek : exception message thrown from the deprecated and removed real_exception().  Now
22:45 dalek : they check for a clearer exception message.
22:45 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28690
22:45 chromatic That's two failing tests either explained away or fixed.
22:46 Limbic_Region chromatic - did you happen to catch my YAPC recap?
22:46 Tene Limbic_Region: add it to here: http://www.perlfoundation.org​/perl5/index.cgi?yapc_na_2008
22:46 Limbic_Region if not, see http://www.gatcomb.org/joshua/​2008/06/yapcna-2008-day-3.html
22:46 shorten Limbic_Region's url is at http://xrl.us/frs39
22:47 Limbic_Region specifically, a comment I wanted to make on your talk that I didn't get to make in the hallway
22:51 chromatic I saw that, thank you.
22:51 chromatic I figured that if people wanted to rescue a project, they hadn't yet decided to jump.
22:51 Limbic_Region I agree
22:51 Limbic_Region but on the other hand
22:52 Limbic_Region some people don't know when is the right time
22:52 Limbic_Region and it hurts them
22:53 chromatic I could add a disclaimer to my notes.  "Sometimes you have to leave."
22:53 Limbic_Region I have a friend - a former co-worker that just can't see that she needs to get out - but then again, she isn't listening to me so having it on some slides probably won't help either
22:54 chromatic I pound the podium better than you do though.
22:54 Limbic_Region heh
23:03 tetragon joined #parrot
23:05 pmichaud I've jumped ship on a couple of projects.  Best move I could've made -- especially when the projects failed and the sponsors pleaded with me to come back and save the projects (basically:  "Name your terms--what would it take for you to come back to work on this project?")
23:08 dalek r28691 | moritz++ | trunk:
23:08 dalek : [rakudo] more tests for spectest_regression
23:08 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28691
23:26 Zaba joined #parrot
23:34 diakopter pmichaud: congrats
23:36 japhb joined #parrot
23:37 TiMBuS joined #parrot
23:39 tewk my $a := $<something><something> stringifies the match into a, how do I get it to assign the actual match?
23:40 chromatic That's half of the pdd25cx test failures fixed.
23:41 dalek r28692 | chromatic++ | pdd25cx:
23:41 dalek : [src] Made Parrot_ex_throw_from_c() call exception handlers defined in C; this
23:42 dalek : fixes the crash in t/src/extend.t.
23:42 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=28692
23:43 Whiteknight chromatic++
23:44 chromatic What's left is tricky.
23:45 Whiteknight that's why I'm leaving it to you :)
23:45 chromatic Okay, but you get the 64-bit stack scanning problem.
23:46 tewk pmichaud: ping
23:46 Whiteknight yeah, do you have the ticket #?
23:46 dolmen joined #parrot
23:47 chromatic I don't think there was a ticket.
23:47 chromatic particle may know though.
23:49 dolmen in rakudo src/classes/Array.pir, sud delete, what does mean :slurpy applied to the .param ?
23:49 dolmen s/sud/sub/
23:49 Whiteknight slurpy means the array pmc absorbs all the rest of the passed arguments to the function
23:50 dolmen so what is the type of the pmc? a list?
23:50 dolmen an iterator?
23:50 purl an iterator is in the libdb.  Has to be.
23:50 chromatic ResizableArrayPMC, I believe.
23:50 dolmen libdb?
23:50 purl libdb is berkeley, not gnu (I think)
23:50 Whiteknight I assume it's an array
23:51 tewk dolmen: I believe it is a RPA, ResizablePMCArray.
23:52 dolmen where is the source of RPA?
23:52 chromatic Right, RPA
23:52 chromatic src/pmc/resizablepmcarray.pmc
23:53 tewk dolmen: RPA is parrot's generic array, you can call push, pop, shift, unshift etc on it.
23:55 dolmen So...
23:55 dolmen .param pmc indices :slurpy
23:55 dolmen $I0 = shift indices
23:55 dolmen ... will call shift_integer in RPA?
23:55 chromatic Yes.
23:57 dolmen if the first element is not an integer, but an object, how can I make shift_integer work for the object?
23:57 dolmen or how can I fix the oject so that it implements what is needed
23:59 chromatic $P0 = shift indices

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

Parrot | source cross referenced