Camelia, the Perl 6 bug

IRC log for #parrot, 2008-12-24

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:04 kid51 joined #parrot
00:06 ask_ joined #parrot
00:08 kid51 Does anyone know what I have to do to get any of the tests in languages/PIR/t/ to pass?
00:09 kid51 According to the README, once I've switched to langauges/PIR and run 'make', I should be able to run 'make test'.  But all tests fail.
00:09 AndyA joined #parrot
00:10 nopaste "kid51" at 68.237.9.143 pasted "Results of 'make test' in languages/PIR/" (35 lines) at http://nopaste.snit.ch/15083
00:11 chromatic You have to convince kj to fix the code to run from that directory.
00:12 nopaste "kid51" at 68.237.9.143 pasted "Some output from a single test file in languages/PIR/t/" (22 lines) at http://nopaste.snit.ch/15084
00:12 kid51 So they're currently inoperable?
00:12 dalek r34302 | pmichaud++ | branches/pctloop2/compilers/pct/src/PAST:
00:12 dalek : [pct]:  Add the exception handler to the loop_gen code.  Now we just test.
00:12 dalek review: http://xrl.us/7o4d8
00:12 chromatic what does the error message say?
00:13 kid51 See 2nd nopaste
00:14 chromatic It was a rhetorical question.  I'm asking you if you understand why it's failing.
00:14 kid51 No, I don't.
00:15 chromatic Okay.  Here's the relevant part of the error message:
00:15 kid51 I was simply following the instructions in languages/PIR/README.
00:15 chromatic Parrot VM: Can't stat languages/PIR/pir.pbc
00:15 chromatic That's Parrot's version of:
00:15 chromatic Can't locate Not/A/Module.pm in @INC
00:15 chromatic Make sense?
00:15 kid51 ok.  So do the tests have to run from the top-level Parrot directory ?
00:16 chromatic It looks like it, but I'm not sure that's right either.
00:16 kid51 Isn't working either.
00:17 chromatic Something in the PIR tests thinks it should be in the parrot root directory.  If we can find that, we might be able to run the tests.
00:18 chromatic lib/Parrot/Test/PIR_PGE.pm
00:18 nopaste "kid51" at 68.237.9.143 pasted "Output of a languages/PIR/t/*.t test run from top-level parrot directory" (10 lines) at http://nopaste.snit.ch/15085
00:22 chromatic Looks like it has other problems as well.
00:22 chromatic If you can't run a language's tests from its own directory, skip it and move on.
00:23 chromatic File a bug that the language isn't relocatable maybe, but there are plenty of other things you can fix without having to debug this mess.
00:23 chromatic That's one of the reasons I had so much trouble keeping up with Lua.  I could never figure out how to run a single test file individually.
00:24 kid51 Ironically, the only reason why I was poking my nose in there was to see if I could cop some tests to add to t/perl/Parrot_Test.t in order to boost the coverage of Parrot::Test.
00:37 dalek r34303 | pmichaud++ | branches/pctloop2/languages/perl6/src/builtins:
00:37 dalek : [rakudo]:  Add simple 'last' and 'redo' functions.
00:37 dalek review: http://xrl.us/7rkaa
00:52 tetragon joined #parrot
01:12 dalek r34304 | pmichaud++ | trunk/compilers/pct/src/PAST:
01:12 dalek : [pct]:  Merge pctloop2 branch into trunk.
01:12 dalek : This allows all loops to handle last/next/redo exceptions.
01:12 dalek review: http://xrl.us/7wgst
01:12 dalek r34305 | pmichaud++ | trunk/languages/perl6/src/builtins:
01:12 dalek : [rakudo]:  Merge 'last' and 'redo' functions from pctloop2 branch.
01:12 dalek review: http://xrl.us/7wjqo
01:13 dalek r34306 | pmichaud++ | branches:
01:13 dalek : Remove obsolete branch.
01:13 dalek review: http://xrl.us/7wntm
01:26 dalek r34307 | chromatic++ | trunk/compilers/imcc (8 files):
01:26 dalek : [IMCC] Replaced imcc_fprintf() with Parrot_io_fprintf() calls.
01:26 dalek : Fixed a couple of compilation warnings.
01:26 dalek review: http://xrl.us/7xit7
01:27 kid51 chromatic:  https://trac.parrot.org/parrot/ticket/87
01:28 chromatic Good work.
01:31 dalek r34308 | pmichaud++ | trunk/compilers/pct/src/PAST:
01:31 dalek : [pct]:  Add 'next' component to PAST::Op  while/until/repeat_*  nodes
01:31 dalek review: http://xrl.us/7x6cw
01:32 tuxdna joined #parrot
01:32 dalek r34309 | pmichaud++ | trunk/languages/perl6/src/parser:
01:32 dalek : [rakudo]:  Correct loop_statement to execute iteration expression on 'next'.
01:32 dalek review: http://xrl.us/7yeps
01:34 Coke chromatic: added a PIC comment. votes welcome.
01:34 Coke particle1: you may also wish to comment on that ticket (60048)
01:35 galf left #parrot
01:38 dalek r34310 | chromatic++ | trunk/compilers/imcc (3 files):
01:38 dalek : [IMCC] Made imcc_vfprintf() a thin wrapper around Parrot_vsprintf_c().  This
01:38 dalek : removes yet another sprintf() implementation.
01:38 dalek review: http://xrl.us/7zbfx
01:39 Whiteknight you guys are on a rampage tonight!
01:43 particle joined #parrot
01:49 dalek r34311 | chromatic++ | trunk/compilers/imcc (2 files):
01:49 dalek : [IMCC] Cleaned up some compilation warnings, making sure never to use an
01:49 dalek : uninitialized variable.
01:49 dalek review: http://xrl.us/72zhi
02:01 dalek r34312 | chromatic++ | trunk/compilers/imcc:
02:01 dalek : [IMCC] Added a check for the return value from a system call.
02:01 dalek review: http://xrl.us/74i9b
02:22 MariachiElf joined #parrot
02:32 * Coke watches _Commentary: The Musical_
02:40 tetragon_ joined #parrot
02:44 kid51_at_dinner nick kid51
02:55 * Coke rotfls at his dvd.
02:56 * Coke finds the root cause of these failing tcl spec tess.
02:56 Coke "tests"
02:57 chromatic They're haunted?
02:58 dalek r34313 | Whiteknight++ | branches/pdd09gc_part1/src (3 files):
02:58 dalek : [pdd09gc_part1] fix a failed assertion, fix a segfault, and keep pmc_ext managed by the old ms collector. This solves A LOT of problems
02:58 dalek review: http://xrl.us/79rvk
02:59 Whiteknight chromatic, I'm finding so many obvious problems that I was missing over the summer
02:59 Whiteknight contained_in_pool fails for objects in newly created arenas, which is why objects I was newly creating were failiing those sanity check assertions
03:00 GeJ Is anyone (else, obviously) experiencing segfaults while `make html`-ing?
03:00 Whiteknight but, I'm out of there for the night. later.
03:02 Coke chromatic: http://code.google.com/p/partcl/source/​diff?spec=svn243&r=243&format=s​ide&path=/trunk/t/cmd_info.t&ol​d_path=/trunk/t/cmd_info.t&old=25
03:02 shorten Coke's url is at http://xrl.us/8adms
03:03 chromatic You removed a test, and then everything started working?
03:03 Coke heh. no, I added a failing test that shows why about a dozen of the spec tests are borked.
03:05 Coke info level 0 should always report the sub and the args of your current sub. but it seems to get stuck.
03:06 Coke (so in the test there, once I call bar check it's info level, the calling information is reported incorrectly for the rest of foo.
03:08 Zaba joined #parrot
03:11 chromatic I see.
03:11 Coke are you just humoring me? =-)
03:12 Coke basically, I'm managing my own call chain; I think I'm forgetting to pop something.
03:20 chromatic I've seen something similar in Lua, due to the way Parrot produces backtraces.
03:20 chromatic I just don't have the brainpower to switch between language calling styles at the moment.
03:21 chromatic_dinner ... or the glucose levels.
03:31 * Coke makes the bad horse gleeful.
03:31 * chromatic_dinner thought it was a bad wolf.
03:34 Coke wrong franchise.
03:34 Tene Now that would be an entertaining crossover.
03:34 Coke Dr. Horrible Who.
03:35 chromatic_dinner There's fanfic of it.
03:35 Coke you say that without even checking.
03:35 Coke ... and you're right, no doubt.
03:35 Coke slashfic, even.
03:36 Tene Coke: if he wrote it he wouldn't have to check...
03:40 chromatic_dinner I've only written Star Trek fanfic, and that was in response to Steve Yegge.
03:44 * kid51 must sleep
03:44 purl $kid51->sleep(8 * 3600);
03:48 rurban_ joined #parrot
04:02 Coke ... Wow, of course this never worked.
04:02 elmex_ joined #parrot
04:19 Coke yay.
04:20 Limbic_Region joined #parrot
04:20 Limbic_Region salutations all
04:23 Coke hio
04:25 tetragon joined #parrot
04:26 Limbic_Region I was just reading whiteknight's use.perl post
04:26 Limbic_Region interesting stuff
04:26 purl interesting is relative
04:27 Limbic_Region is any of the stuff getting pulled out being planned to be re-integrated?
04:27 Limbic_Region specifically STM
04:27 Limbic_Region ?
04:27 chromatic_dinner Hard to say.
04:28 Limbic_Region well, presumably a smaller code base will be easier to maintain, optimize, debug, etc
04:29 bacek joined #parrot
04:29 Limbic_Region looking in from the outside and hoping babies don't get thrown out with the bath water is all
04:29 Coke too many mouths to feed
04:30 Limbic_Region and now I need to close out chatzilla so I can upgrade adobe flash so I can watch Adam Sandler's Happy Chanuka on youtube - not quite yak shaving but close
04:30 * Limbic_Region brb &
04:30 chromatic_dinner Just between you, me, and posterity stalking me by reading the logs, Piethon may have been the worst thing ever to happen to Parrot.
04:31 Limbic_Region wow, that was a long time ago
04:31 chromatic_dinner Yeah, and we're still ripping out some of the code from it.
04:31 Limbic_Region well, you will have to tell me when I get back how much of that was Dan and how much was Leo
04:31 Limbic_Region but really, Adam Sandler waits for no one
04:31 chromatic_dinner Not even in private will I speculate about that.
04:31 Limbic_Region TTFN
04:33 Limbic_Region joined #parrot
04:41 Tene What was piethon?
04:42 Coke an attempt to get parrot to run python faster than python.
04:44 Limbic_Region http://www.sidhe.org/~dan/bl​og/archives/cat_piethon.html
04:45 Limbic_Region read bottom to top
04:46 lathos chromatic_dinner: Any thoughts on the NCI rewrite?
04:54 Coke pmichaud: TT#90 is a meta ticket; all deprecations should be tracked in their own tickets.
04:55 pmichaud okay, it can be a meta ticket.  I was just making tickets for critical/important items in the roadmap
04:56 bacek good afternoon
04:57 Coke it's also unclosable by that release. =-)
04:59 pmichaud ...unclosable?
04:59 Coke yes. not all deprecations are removable by the january release.
05:00 pmichaud I'm simply putting down what was on the roadmap page.  :-)
05:00 Coke hokay.
05:00 pmichaud it had it as a 0.9.0 milestone, and marked "critical"
05:00 pmichaud I don't mind at all if it rolls over to 0.9.1, but someone else gets to make that call.  :-)
05:01 pmichaud I'm just the ticket maker.  :-)
05:01 Coke k. I'm only saying something because my name is on it.
05:01 pmichaud you can fix that if you wish.  :-)
05:01 pmichaud I only stuck your name on it because your name was listed first in the roadmap.
05:04 particle joined #parrot
05:05 pmichaud I'm not sure how to break out the pdd30 stuff into tickets, though, so I guess I'll stop here.
05:06 Coke (listed first) I blame particle.
05:06 Coke (added my comments to the ticket.)
05:06 Tene Coke: deprecations should depend on the release, not the other way around.
05:06 Coke tene: huh?
05:07 pmichaud the goal for this ticket/task was to have all of the items in DEPRECATED resolved before 1.0 (0.9.0 in this case)
05:07 pmichaud perhaps that's an unrealistic goal.
05:08 pmichaud but that's what was discussed at PDS.
05:08 Coke it's definitely a stretch.
05:09 Tene Coke: if that didn't make sense, I misunderstood what was being discussed.  Just ignore me.  I'll read scrollback later and feel silly.
05:13 geof joined #parrot
05:44 tewk pmichaud: is there a way to print a backtrace in PGE to see where I am?
05:48 Coke this is so depressing. my work has resulted in no new spectests passing.
05:48 tewk I'll try panic
05:53 Coke panic?
05:53 purl i heard panic was a Good Bunch Of Folks based in Portland, Oregon  They write software for Macs, including Audion, Transmit, Desktastic, CandyBar and others.  Buy their stuff!
05:53 Coke don't panic?
05:53 purl PANIC!!!
05:53 tewk <panic>
05:54 * bacek passing a towel to tewk
05:57 Coke Anyone able to hack on https://trac.parrot.org/parrot/ticket/66 ?
05:58 Coke oh. pmichaud ; were you able to generate a small test case for that?
06:00 tewk so is it a gc problem or a mod problem?
06:01 Coke mod?
06:01 purl mod is defined such that x mod y = x - y * floor(x/y) according to Knuth.  But this is not always what people expect when x or y are negative as floor() would turn -4.5 into -5 while most people expect the result as if it were -4.  This latter behavior is the remainder operator.
06:02 Coke oh, right.
06:02 Coke no clue.
06:10 dalek r34314 | pmichaud++ | trunk/languages/perl6/src/builtins:
06:10 dalek : [rakudo]:  Make sure infix:<~> and prefix:<~> return Str.
06:10 dalek : We hopefully can get rid of this when we have HLL support, but
06:10 dalek : let's at least get them right for now.
06:10 dalek review: http://xrl.us/8rmvr
06:11 Coke I need to set a variable as unix or windows for partcl. Anyone care to point at a parrot config that'll give me that easily?
06:11 Theory joined #parrot
06:12 Coke I am tempted to cheat and just key off 'slash'
06:16 masak joined #parrot
06:18 Coke done.
06:22 Ademan joined #parrot
06:31 * bacek have stupid question.
06:31 bacek How I can read from stdin in Rakudo?
06:35 dalek r34315 | petdance++ | trunk/compilers/imcc:
06:35 dalek : fix function header and do some consting
06:35 dalek review: http://xrl.us/8szvc
06:37 masak bacek: =$*IN
06:40 TiMBuS joined #parrot
06:52 bacek masak: thanks
06:54 chromatic lathos, Allison posted some thoughts to the list, but your rewrite looked sane to me.
06:54 lathos I can't understand how Allison's schema actually fits with NCI.
06:55 chromatic There is that too.
06:55 lathos I'll check in the refactoring, and I have some more refactoring to do, both of which will make it easier for us to implement whatever we end up with.
06:55 tetragon joined #parrot
07:02 bacek masak: http://use.perl.org/~bacek/journal/38146 http://use.perl.org/~bacek/journal/38147 :)
07:03 masak nice.
07:05 masak I was actually on my way through the latest use.perl posts to see if anyone responded to pmichaud++'s challenge. would have found your posis in a few minutes. :)
07:06 dalek r34316 | pmichaud++ | trunk/languages/perl6/src (2 files):
07:06 dalek : [rakudo]:  List methods are already in Any, no need to have them in List also.
07:06 dalek review: http://xrl.us/8veu9
07:08 dalek r34317 | simon++ | trunk/tools/build:
07:08 dalek : Refactor all the little tables into one big table.
07:08 dalek review: http://xrl.us/8vman
07:08 masak dpuu++ also contributed: http://use.perl.org/~dpuu/journal/38142 http://use.perl.org/~dpuu/journal/38145
07:15 masak bacek: in your &fact sub, you assign 1 to $n instead of comparing it. I doubt your actual code does that, because that would generate a runtime error (because params are readonly by default) -- but you might want to change it in the blog post.
07:15 bacek masak: yes... I code it '<='. I'll fix it.
07:17 bacek masak: '<' consumed by use.perl.org... It still in post when I try to edit it...
07:18 masak bacek: use.perl has very strange rules for filtering stuff.
07:18 masak bacek: I wish I could give you specific advice, but I've found that different conditions lead to very different behaviour.
07:19 bacek masak: yeah... No problem, anyway.
07:19 bacek fixed
07:19 masak bacek++
07:31 bacek masak: http://use.perl.org/~bacek/journal/38148 last one for today
07:34 pmichaud I'm currently working on fixes to the issues raised in http://www.rakudo.org/2008/12/perl-6​-scripting-games.html#comment-11058
07:34 shorten pmichaud's url is at http://xrl.us/8w7v9
07:38 masak sounds like a worthy cause.
07:39 pmichaud I've already fixed
07:39 pmichaud rakudo:  say ('foo' ~ 'bar').reverse
07:39 polyglotbot OUTPUT[raboof␤]
07:40 masak raboof!
07:40 pmichaud I'm about to fix Match objects.
07:41 dalek r34318 | pmichaud++ | trunk/compilers/pge/PGE (2 files):
07:41 dalek : [pge]:  Temporary fix to allow callers to decide the Match object
07:41 dalek : (since we can't always provide a 'grammar' named argument).
07:41 dalek review: http://xrl.us/8xrqz
07:48 japhb joined #parrot
07:49 Tene rakudo.paste: my @a = { $^a.perl.say; shift $^a }, {$^a.perl.say; pop $^a }, <a b c d e>; @a[0](@a) while @a;
07:49 nopaste "polyglotbot" at 193.200.132.146 pasted "perl6 paste" (7 lines) at http://nopaste.snit.ch/15086
07:51 masak self-modifying programs. :)
08:02 dalek r34319 | pmichaud++ | trunk/languages/perl6/src (3 files):
08:02 dalek : [rakudo]:  Matches now produce Rakudo Match objects instead of PGE::Match.
08:02 dalek review: http://xrl.us/8zskx
08:09 dalek r34320 | tewk++ | trunk/languages/ecmascript (17 files):
08:09 dalek : [js] the beginning of core types
08:09 dalek review: http://xrl.us/82ata
08:30 Tene btw, using globals in rakudo's interactive mode makes it nices.
08:30 Tene nicer.
08:30 masak in what way?
08:30 Tene they persist, where lexicals don't.
08:30 Tene from line to line
08:31 masak that _is_ useful.
08:31 Tene workaround for broken lexicals in interactive mode
08:31 masak so, er, how do I declare a global?
08:31 Tene just use it.
08:31 Tene $*foo
08:31 Tene no declaration
08:31 Tene afaik
08:31 masak even nicer :)
08:32 pmichaud note that nowhere is it specified that lexicals in interactive mode should be preserved :-)
08:32 * masak has been seriously considering diving into the interactive mode source code and fix the lexicals thing
08:32 pmichaud good luck.
08:32 purl You'll need it.
08:32 masak I suspect it's not an easy fix.
08:32 pmichaud not even close.
08:32 pmichaud how do you preserve a lexical from a block you just exited?
08:33 masak why would I want to do that?
08:33 Tene I looked into that once before.  I'm pretty sure I came up with something that would work, but it was a bit evil, and I didn't try it.
08:33 pmichaud because the lexicals are local to the block?
08:33 masak pmichaud: yes, so why preserve it afterward? I don't understand.
08:33 pmichaud remember that each line of input acts like its own block.
08:33 masak pmichaud: well, that's just wrong.
08:33 purl Fire and brimstone coming down from the skies. Rivers and seas boiling! Forty years of darkness, earthquakes, volcanos. The dead rising from the grave! Human sacrifice, dogs and cats living together... mass hysteria!
08:34 Tene It involved maintaining a separate lexpad and passing it in to a function to run some block using the given lexpad.
08:34 masak Tene: sounds evil-ish, yes.
08:34 pmichaud rakudo.pir:  my $i = 4;
08:34 pmichaud (how do I get rakudo to paste pir output?)
08:35 nopaste "polyglotbot" at 193.200.132.146 pasted "perl6 pir" (53 lines) at http://nopaste.snit.ch/15087
08:35 Tene perl6.pir: my $i = 1
08:35 pmichaud oh, there it is.
08:35 Tene ah
08:35 nopaste "polyglotbot" at 193.200.132.146 pasted "perl6 pir" (53 lines) at http://nopaste.snit.ch/15088
08:35 pmichaud see, when we compile an interactive line of code, it results in a Parrot sub.
08:35 pmichaud Any lexicals we declare are local to that Parrot sub.
08:35 masak I see.
08:36 masak yes, that's a problem.
08:36 pmichaud So, in order for interactive mode to work, we have to have a way to tell the compiler to look in some other place for lexicals, and not declare the top-most lexicals using .lex
08:36 pmichaud -or-
08:36 masak :)
08:37 pmichaud at the end of the block we have to magically find out what lexicals it had and preserve them somewhere.
08:37 pmichaud so that we can restore them for the next line of input we get.
08:37 masak aye.
08:38 masak the first method is dangerous because we effectively alter the behaviour of Rakudo.
08:38 pmichaud I think "preserving the lexicals" is the most likely scenario at this point.
08:38 masak the second method is tricky because of the 'magically find out' step.
08:38 pmichaud well, when we compile the top-level block we have the symbol table available.
08:39 pmichaud so we know what lexicals we declared, and thus which ones we have to preserve.
08:39 Tene unless we eval.
08:39 masak aie!
08:39 pmichaud eval is no problem -- any lexicals defined in an eval *are* local to that eval, and not preserved.
08:39 Tene in Perl 6, at least. :)
08:40 masak oh, will the changes to the interactive mode necessarily affect all HLLs, or just Rakudo?
08:40 dalek r34321 | pmichaud++ | trunk/languages/perl6/src (4 files):
08:40 dalek : [rakudo]:  Clean up reverse methods and functions.
08:40 dalek review: http://xrl.us/83ru9
08:40 pmichaud I'd like to find a way to make it work for all hlls.
08:41 masak what I meant was, is there only one "interactive mode" thingy in PCT, or is there one for each HLL?
08:41 pmichaud there's only one in PCT
08:41 pmichaud individual HLLs can override it with their own if they wish.
08:41 masak oh good.
08:42 pmichaud (afaik none of them do this yet)
08:45 Tene pmichaud: are there any exception issues left to work out for the Jan. release that you know of?
08:45 pmichaud Tene: not that I know of.
08:45 pmichaud I think we're in really good shape there.
08:45 Tene Okay, so I can swap that out and start on HLL stuff next?
08:45 pmichaud I like the way the loop refactor turned out.  Now if only Perl 6 doesn't mess it all up (which it probably will)
08:45 pmichaud I managed to get rid of a ton of code in the refactor.
08:46 Tene Or is there something specific you want me to work on next?
08:46 pmichaud and clean up the standard C-style "for(a;b;c)" behavior in the process.
08:49 Tene I'm working on monday, probably, but other than that and family on Christmas, I'm free until Jan. 4.
08:49 pmichaud HLL is probably where we should focus next, yes.
08:51 Tene I've seen a few indications that I don't really grok your plan for HLL in PCT.  That's why I keep asking for TT #25 to be filled out.
08:52 pmichaud high level plan is   (1) add ".HLL 'perl6'"  at the beginning of perl6.pir,  (2)  see what breaks,  (3) start fixing that.
08:52 Tene Right.
08:52 Tene There was something with exporting that still needed to happen, right?
08:53 ChrisDavaz joined #parrot
08:53 pmichaud yes, but we're nearing an answer to that on p6l.
08:53 pmichaud if nothing else we'll do the exports manually for now.
08:53 Tene The reason you had me add symbols to ::EXPORT::ALL
08:54 pmichaud right, that tells us which symbols to import -- we just need to import them.
08:54 pmichaud otoh, it might be as simple as taking the ['parrot';'PAST'] namespace and binding it to ['perl6';'PAST']
08:54 pmichaud haven't tried that yet.
08:55 Tene Were you planning to delegate most of this to me, or did you want to be significantly involved?
08:55 Theory joined #parrot
08:56 pmichaud I was expecting I would probably need to be significantly involved.
08:56 Tene Okay.  Do you have a day you'd like to start on this?
08:57 pmichaud almost any day is fine.
08:57 pmichaud Friday is probably not so good, I guess.
08:57 pmichaud (kids come home Friday, we open gifts then, they'll want to play, etc.)
08:57 Tene Okay.  I'll start whenever, then.
08:58 Tene ... oh, right, we also need to get the class registry fixed soon.
08:58 pmichaud sure.  I'd just make a 'rakudohll' branch or something and we'll see how far things progress.
08:58 Tene Did we get a decision from Allison?
08:58 pmichaud I don't think we have to worry too much about staying synchronized with trunk -- the hll stuff isn't likely to interact much with trunk changes.
08:59 pmichaud merging whatever we do back into trunk should be straightforward, so it's okay if the branch falls a bit out of date.
08:59 pmichaud I haven't heard _anything_ on the class registry.  I entered a trac ticket for it this past week hoping it might show up on someone's radar.
08:59 pmichaud However, it has a Jan 2009 milestone on it, so it's supposedly to be fixed this month.  :-|
09:00 pmichaud if you wanted to work on that, that would be really useful also.
09:00 Tene Should mail allison directly...
09:00 Tene you have the TT #?
09:01 pmichaud #71
09:01 Tene I'll mail.
09:01 Tene purl: allison?
09:01 purl allison is Allison Randal <mailto:allison@perl.org>
09:01 pmichaud thanks.
09:07 Tene pmichaud: could we add a <warn> rule to PGE, like <panic> ?
09:08 pmichaud I used to have one there.  :-)
09:08 pmichaud ummmmm, sure.
09:08 pmichaud put it in PCT::Grammar, though.
09:08 pmichaud same as <panic>
09:09 Tene So it needs to go in PGE::Util too?
09:09 pmichaud ummmm.
09:09 * pmichaud checks.
09:10 Tene pctgrammar's panic just addmethod's pgeutil's die
09:10 pmichaud it should probably go directly into PCT::Grammar
09:10 pmichaud (so should panic -- PGE::Util die will probably go away.)
09:10 pmichaud just do warn for now.
09:14 pmichaud omg it's 3:14 in the morning
09:18 Tene any idea what mob.'to'(-3) does in panic?
09:18 pmichaud kills the entire match.
09:18 Tene I'm not sure what I need to do to indicate a successful zero-length match.
09:18 pmichaud set 'to' to the same value as 'from'
09:21 alvar joined #parrot
09:23 Tene That still terminates with a syntax error.
09:24 * Tene looks for other examples.
09:29 Tene ... oh, pgeutil already has a 'warn'
09:33 * Tene sleep.
09:36 dalek r34322 | pmichaud++ | trunk/languages/perl6/docs:
09:36 dalek : [rakudo]: spectest-progress.csv update: 264 files, 5837 passing, 0 failing
09:36 dalek review: http://xrl.us/87b99
09:44 iblechbot joined #parrot
09:52 dalek r34323 | pmichaud++ | trunk/languages/perl6/src/classes:
09:52 dalek : [rakudo]:  Refactor to simplify returning Match object from regexes.
09:52 dalek review: http://xrl.us/88d6n
09:53 dalek r34324 | pmichaud++ | trunk/languages/perl6/src/builtins:
09:53 dalek : [rakudo]:  Clean up some bugs in .subst reported by skids1++ .
09:53 dalek : There's still a lot more that needs fixing here, though.
09:53 dalek review: http://xrl.us/88gdo
09:54 pdcawley joined #parrot
09:59 barney joined #parrot
10:23 dalek r34325 | bernhard++ | trunk/docs (3 files):
10:23 dalek : #61638: [PATCH] Fixed typos in docs
10:23 dalek : Courtesy of tuxdna.
10:23 dalek review: http://xrl.us/9adph
10:25 MariachiElf joined #parrot
10:45 dalek r34326 | bernhard++ | trunk/ext/Parrot-Embed/t:
10:45 dalek : [codingstd] svn:mime-type for pipp.t
10:45 dalek review: http://xrl.us/9bcp6
10:49 dalek r34327 | bernhard++ | trunk (6 files):
10:49 dalek : [codingstd] remove trailing spaces
10:49 dalek review: http://xrl.us/9bpc4
11:12 iblechbot joined #parrot
11:33 barney looks like   squaak is broken
11:33 dalek r34328 | bernhard++ | trunk/languages/lolcode/src/parser:
11:33 dalek : [lolcode] Remove two unneeded declarations of @?BLOCK
11:33 dalek review: http://xrl.us/9e35c
11:33 dalek r34329 | fperrad++ | trunk/languages (3 files):
11:33 dalek : [Lua] smolder
11:33 dalek : - add a target 'smoke'
11:33 dalek : - improve harness
11:33 dalek review: http://xrl.us/9e36s
11:49 rurban_ joined #parrot
11:52 dalek r34330 | bernhard++ | trunk/languages/m4 (2 files):
11:52 dalek : [m4] The ParrotIO PMC is gone.
11:52 dalek review: http://xrl.us/9gndm
12:05 ChrisDavaz joined #parrot
12:08 dalek r34331 | bernhard++ | trunk/languages (2 files):
12:08 dalek : [lazy-k] Make it compile again. But no output is printed.
12:08 dalek review: http://xrl.us/9icar
12:14 ask_ joined #parrot
12:18 iblechbot_ joined #parrot
12:30 dalek r34332 | bernhard++ | trunk/languages/squaak/src/parser:
12:30 dalek : [squaak] isdecl() is more happy with a simple integer.
12:30 dalek review: http://xrl.us/9jw6s
12:57 dalek r34333 | bernhard++ | trunk/languages/pipp/src (3 files):
12:57 dalek : [Pipp] Remove unneeded explicit loading of PGE-libs.
12:57 dalek : Initialize @?BLOCK as ResizablePMCArray.
12:57 dalek : Start with keeping track of blocks and associated symbol tables.
12:57 dalek review: http://xrl.us/9mpaw
13:11 dalek r34334 | bernhard++ | trunk (10 files):
13:11 dalek : [Pipp] Rearrange files in pipp/src
13:11 dalek review: http://xrl.us/9n9u4
13:14 dalek r34335 | bernhard++ | trunk/compilers/pirc/new:
13:14 dalek : [codingstd] wrap args in macro definitions
13:14 dalek review: http://xrl.us/9occ2
14:04 tuxdna joined #parrot
14:05 dalek r34336 | bernhard++ | trunk/languages/pipp/src/pct (2 files):
14:05 dalek : [Pipp] Shift from @?BLOCK in TOP rule.
14:05 dalek : Prepare for class defintion introducing a new scope.
14:05 dalek review: http://xrl.us/9sv96
14:07 dalek r34337 | bernhard++ | trunk/languages/ecmascript (11 files):
14:07 dalek : [codingstd] set svn properties
14:07 dalek review: http://xrl.us/9s2q4
14:15 Theory joined #parrot
14:21 Zaba joined #parrot
14:30 ruoso joined #parrot
14:33 jimmy joined #parrot
14:50 gryphon joined #parrot
14:59 Limbic_Region joined #parrot
15:25 rurban sigh. finished with pdd30_install patching (52 .rej files). now the testing starts.
16:03 rurban cygwin parrot-0.8.2-1: 11678 tests, 4 fail
16:04 rurban t/pmc/eval 10, t/pmc/os 15 t/pmc/packfile 6 t/library/pg ?
16:16 iblechbot joined #parrot
16:38 dalek r34338 | pmichaud++ | trunk/languages/perl6/src/classes:
16:38 dalek : [rakudo]:  Adjust Pair.perl .
16:38 dalek review: http://xrl.us/95z2s
16:47 tewk
16:47 pdcawley joined #parrot
17:20 allison joined #parrot
17:37 ChrisDavaz joined #parrot
17:47 apeiron joined #parrot
18:12 chromatic joined #parrot
18:13 dalek r34339 | rurban++ | trunk/languages/befunge:
18:13 dalek : [befunge] fix syntax error
18:13 dalek : new dot => 'string'
18:13 dalek review: http://xrl.us/baduvy
19:21 PacoLinux joined #parrot
19:33 bacek joined #parrot
19:52 rurban_ joined #parrot
19:57 leto joined #parrot
20:08 MariachiElf joined #parrot
20:11 Eevee joined #parrot
21:21 mberends joined #parrot
21:31 tetragon joined #parrot
21:38 dalek r34340 | chromatic++ | trunk/languages/perl6/src/pmc:
21:38 dalek : [Rakudo] Fixed some compilation warnings in MultiSub PMC.
21:38 dalek review: http://xrl.us/batfyh
21:42 dalek r34341 | chromatic++ | trunk/src:
21:42 dalek : [src] Removed an un-COW from string_chopn_inplace().  It's unnecessary, as the
21:42 dalek : STRING's length is stored in the header.  We can modify that in place without
21:42 dalek : copying the string.  This gives PGE a nice speedup; 13.43% in producing
21:42 dalek : Rakudo's operator precedence parser.
21:42 dalek review: http://xrl.us/batv5d
21:54 TiMBuS joined #parrot
21:55 chromatic Hm, MALLOC_CHECK_=2 costs 10% of execution time.  Shoulda known.
22:22 dalek r34342 | pmichaud++ | trunk/languages/perl6/src/classes:
22:22 dalek : [rakudo]:  Fix $x.map to work as list of 1 element.
22:22 dalek review: http://xrl.us/baw8q5
22:32 dalek r34343 | chromatic++ | trunk/src:
22:32 dalek : [src] Only reallocate destination STRING buffer in string_append() when the
22:32 dalek : resulting size exceeds (not merely equals) the current buffer size.  This
22:32 dalek : produces a 12.32% performance improvement in the Rakudo optable-building
22:32 dalek : benchmark.
22:32 dalek review: http://xrl.us/baxpb7
22:34 chromatic That should make PGE about a third faster.
22:53 pdcawley joined #parrot
23:33 Ademan joined #parrot
23:39 Tene I just had a horrible idea.  Let's hope I fail at implementing it.
23:56 gryphon joined #parrot

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

Parrot | source cross referenced