Camelia, the Perl 6 bug

IRC log for #parrot, 2008-10-16

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 Andy joined #parrot
00:00 Tene Coke: can you test with that patch and chromatic's patch?
00:02 Coke tene; i'm not using PCT, so I'm unsure how that would help
00:02 Tene Okay, now I need to consider the possibility that I'm high.
00:02 Tene I don't remember eating anything unusual today...
00:04 chromatic Pheme uses PCT.
00:04 Tene Oh.
00:04 Coke + tge?
00:05 chromatic Yes.
00:05 Tene no, pheme is still hosed
00:05 purl okay, Tene.
00:05 chromatic It doesn't use NQP.
00:05 Tene purl: forget pheme
00:05 purl Tene: I forgot pheme
00:05 Coke perhaps if pheme works, tcl will magically work then, since I just use TGE. =-)
00:05 chromatic Probably.
00:05 purl Really? Probably? Are you Certain it's not certain? Are you sure it's unsure? I think you need to look harder.
00:05 Tene Class 'Pheme::AST::Grammar' not found
00:06 Coke yes, that's the problem.
00:06 Coke whee!
00:06 * Coke yells at his desktop to build parrot faster.
00:07 Coke ... I had stopped doing parrot builds for some time now that partcl is out of the repo. It was nice. =-)
00:09 chromatic ccache + parallel builds + secret Roswell technology == FAST
00:09 chromatic Uh.
00:09 chromatic You should probably elide the third part of that equation.
00:09 AndyA joined #parrot
00:09 * Coke realizes that he was really building on feather.
00:09 * Coke also checks for high.
00:10 Coke t/compilers/imcc/syn/file..................6/14 # Looks like you planned 14 tests but ran 1 extra.
00:11 Tene ^5 Coke++
00:11 Coke is that failing for anyone else?
00:11 Coke (or did I break that locally)
00:12 dalek r31990 | tene++ | trunk:
00:12 dalek : [pheme]: Use exception handler type filter to make debugging not suck
00:12 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31990
00:12 chromatic Coke, run it with -v and see if you have two skips.
00:13 Coke ok 13 # SKIP multiple loading not speced - failing
00:13 Coke ok 14 # SKIP Cannot write temporary file to /tmp/with_slash.pir
00:13 Coke ok 15 # SKIP Cannot write temporary file to /tmp/with_slash.pir
00:13 chromatic the 2 on line 476 should be 1
00:13 chromatic ... but the use of temporary files there isn't very good.
00:14 Coke I am more concerned about the harness output atm.
00:14 Coke mind if I commit that fix?
00:15 chromatic I'm thinking about the code paths, but if you skim it and the fix makes sense to you too, go ahead.\
00:16 chromatic That logic is so unclear, but I'm sure that's the right fix.
00:16 Coke it's twisty, yah.
00:16 Coke leaving it alone for now.
00:16 Whiteknight t/library/pcre.t is failing on my system, is that expected?
00:16 Whiteknight or, is it failing for anybody else?
00:17 Coke ISTR there's a ticket.
00:17 Tene okay... get_class ['Pheme';'AST';'Grammar'] is returning null
00:17 Tene What is supposed to make that class?
00:17 Coke pge, neh?
00:17 chromatic lib/ASTGrammar.pir
00:18 Tene is PGE supposed to make that class for pheme?
00:18 Whiteknight yeah, there is a ticket for it, thanks
00:18 Whiteknight RT#59886
00:19 chromatic Either PGE or TGE; I get that confused.
00:19 Tene Someone.  Hmm.
00:19 chromatic They both take .pg and .tg files respectively and turn them into PIR.
00:19 Coke PGE generates the PIR for the grammar, usually.
00:19 Tene Anyway, pheme is non-stupid to debug after that last commit.
00:20 chromatic Non-stupid?
00:20 Coke if you catch and rethrow the error, I suspect the trace gets fubard.
00:20 Tene Before that commit, the 'exit handler' would grab ALL exceptions and rethrow them, so you couldn't see where they came from.
00:20 Coke which makes me think rethrow is broken.
00:20 chromatic Yeah, I usually comment that out for debugging.
00:21 Tene Now you don't need to!
00:21 Coke (If I meant throw, i'd've said it.)
00:21 Coke the patch to remove __'s is pretty small.
00:21 chromatic Coke, it's trickier than you think.
00:22 Coke chromatic: .. not according to 'make test'.
00:22 dmknopp joined #parrot
00:23 chromatic To excise it, you need to touch IMCC, the Class PMC, src/pmc.c, src/oo.c, and a few other places.
00:23 chromatic Hopefully no languages.
00:23 Whiteknight how do you find out what version of pcre you have installed?
00:24 Coke chromatic: I don't think so. Presuming 'make test' works, I'll post the patch on the ticket.
00:25 Tene Oh, chromatic was right.
00:25 Coke (I think the bits that might have cared might have previously been abstracted out.)
00:25 Tene purl: chromatic
00:25 purl chromatic is <req>a lot of fun.  For years I'd try to play stuff like `peter and the wolf', and then I'd be frustrated because it would use some note I didn't have. or the author of jellybean or mailto:chromatic@wgz.org or http://wgz.org/chromatic/ or the winner of the not-a-contest perl-bugathon. or best reached via email. or the guy who hit me in the eye.
00:25 Coke they might not be giants.
00:25 Coke you never know.
00:25 chromatic What are you going to do unless I are?
00:26 Coke ... I'll call particle man.
00:26 chromatic They have a fight.  TimToady wins.
00:26 * Coke forgot to run coretest. *glower*
00:26 Tene anyone remember what it means then a get_clas op on an RPA throws:
00:26 Tene get_string() not implemented in class 'ResizableStringArray'
00:26 * Coke churns through an insane number of perlcritic tests.
00:26 Tene Yes, RSA, not RPA
00:27 nopaste joined #parrot
00:27 Coke tene: that sounds like something allison just said was fixed.
00:27 chromatic I fixed that in fail_if_type_exists() just yesterday.
00:27 Coke er, was something she thought was fixed but patrick said it wasn't and chromatic said he saw ferris pass out at 31 flavors last night.
00:27 Coke ... I guess it's pretty serious.
00:27 chromatic He needs a new kidney.
00:30 bacek_ joined #parrot
00:31 dalek r31991 | coke++ | trunk:
00:31 dalek : [t] un-todo test for recently implemented feature.
00:31 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31991
00:32 nopaste "tene" at 67.137.148.11 pasted "Updated patch for TGE that failes differently for pheme" (59 lines) at http://nopaste.snit.ch/14301
00:33 Tene Now pheme prints a bunch of 'Cannot find the attribut ...' in the past stage, and then dies with a null pmc access in the post stage
00:34 Coke chromatic: patch added to #45965
00:34 nopaste "tene" at 67.137.148.11 pasted "output from pheme" (24 lines) at http://nopaste.snit.ch/14302
00:34 bennett joined #parrot
00:34 Tene now I REALLY need to get home.
00:34 Coke was verrrry small.
00:34 Coke I think everythign is referring to the slots by their macro definition.
00:34 Coke update the macros, it jfw.
00:35 chromatic <earthworm type="jim">GROOVY</earthworm>
00:38 chromatic Coke, ack for __ in the tree.
00:38 Coke that way lies madness.
00:38 Coke (because of all the function attributes)
00:38 chromatic The Underbars of Madness
00:38 Coke wunderbras?
00:39 chromatic And through strange checkins, even crufty old grotty code may die.
00:39 chromatic HDP Lovecraft.
00:41 Coke is http://rt.perl.org/rt3/Tic​ket/Display.html?id=58410 closable?
00:41 Coke (didn't the N_ stuff get removed already?)
00:44 pmichaud the thing that doesn't work is    $P0 = get_class $P1      where $P1 is an array
00:45 pmichaud so we have  $P0 = get_class $S0,   $P0 = get_class ['key'],   and $P0 = get_class $P1  where $P1 is a NameSpace, but not where $P1 is an array.
00:47 nopaste joined #parrot
00:48 chromatic If you people would only read the minutes I haven't posted yet and won't post until tomorrow, you'd have been knowing it already!
00:52 pmichaud (RT 58410)   most of the n_ opcodes have been removed, but not all of them yet.
00:55 Andy joined #parrot
01:00 pjcj joined #parrot
01:08 Whiteknight how do I call a method on a PMC from C, do I have to use Parrot_PCCINVOKE?
01:09 Whiteknight it's a method defined in C, not PIR
01:14 Zaba joined #parrot
01:23 pjcj joined #parrot
01:25 chromatic I believe so.
01:38 particle1 joined #parrot
01:41 Whiteknight I hate waiting for emails to make it to the p6-internals list, i have questions that I want to address so I can start working right now
01:41 Whiteknight not in about 6-10 hours
02:03 particle1 feh. there are hundreds of test failures on my windows box.
02:05 chromatic local $TODO = 'Windows is, after all of that, just not a great development platform';
02:05 particle1 there must be something about this environment... lemme check the smolder failure logs
02:05 particle1 smolder?
02:05 purl well, smolder is http://sourceforge.net/projects/smolder or web-based smoke test aggregator used by developers and testers to upload (automated or manually) and view smoke/regression tests using the Test Anything Protocol (TAP). or http://smolder.plusthree.com/app​/public_projects/smoke_reports/8
02:21 nopaste "particle" at 98.232.28.49 pasted "make test output (386 failures)... see any patterns?" (5293 lines) at http://nopaste.snit.ch/14303
03:08 TonyC joined #parrot
03:19 dmknopp left #parrot
03:31 cjfields joined #parrot
03:39 ab5tract joined #parrot
03:55 Zaba joined #parrot
03:56 dalek r31992 | particle++ | trunk:
03:56 dalek : [rakudo] better refactoring thanks to pmichaud++
03:56 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31992
04:08 tetragon joined #parrot
04:25 Theory joined #parrot
04:43 particle joined #parrot
05:01 particle1 joined #parrot
05:02 particle joined #parrot
05:07 Bzek joined #parrot
05:35 Theory joined #parrot
05:43 Zaba joined #parrot
05:45 uniejo joined #parrot
05:55 dalek r31993 | particle++ | trunk:
05:55 dalek : [cage] remove some unused variables
05:55 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31993
05:56 particle1 left #parrot
06:00 chromatic joined #parrot
06:57 viklund joined #parrot
07:34 iblechbot joined #parrot
07:37 dalek r31994 | fperrad++ | trunk:
07:37 dalek : [Lua]
07:37 dalek : - overload VTABLE 'cmp' in LuaAny (so for all Lua types)
07:37 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31994
07:38 barney joined #parrot
07:52 cosimo joined #parrot
07:58 Debolaz joined #parrot
08:22 mj41 joined #parrot
08:24 tomyan joined #parrot
08:46 cotto_home joined #parrot
08:47 cotto joined #parrot
08:47 cotto seen cotto
08:47 purl cotto was last seen on #parrot 0 seconds ago, saying: seen cotto
08:47 cotto good to know
08:51 cotto is src/dynpmc/rotest.pmc still relevant?
08:54 particle joined #parrot
09:14 cotto joined #parrot
09:15 cotto irclog
09:15 cotto purl, irclog?
09:15 purl it has been said that irclog is http://irclog.perlgeek.de/parrot/today
09:15 cotto maybe she thought I was talking about an infrared clog or something
09:16 cotto would it be ok to add an exception (as in special case) to t/distro/test_file_coverage.t so the dynpmc coverage test can be un-TODOd?
09:19 kj joined #parrot
09:20 barney cotto: sound sane
09:23 cotto it looks like that file needs some refactoring anyway.
09:32 Zaba_ joined #parrot
09:43 xiaoyafeng joined #parrot
10:18 cotto seen allison
10:18 purl allison was last seen on #parrot 8 days, 18 hours, 30 minutes and 44 seconds ago, saying: pmichaud: previously they snuck through by being invoked on a proxy Complex object stored within the subclass of Complex  [Oct  7 15:48:08 2008]
10:18 cotto seen chromatic
10:18 purl chromatic was last seen on #parrot 8 hours, 13 minutes and 45 seconds ago, saying: local $TODO = 'Windows is, after all of that, just not a great development platform';
11:01 bacek evening everyone
11:34 Ontolog joined #parrot
12:09 Ontolog joined #parrot
12:23 ruoso joined #parrot
12:40 tetragon joined #parrot
12:45 Andy joined #parrot
12:48 ruoso joined #parrot
13:28 Lorn joined #parrot
13:46 Zaba_ moin
13:48 Coke moi
13:59 Andy joined #parrot
14:05 gryphon joined #parrot
14:26 jhorwitz joined #parrot
14:31 jq joined #parrot
14:36 bacek joined #parrot
15:05 masak joined #parrot
15:06 masak I'm considering adding #58392 as a dependency to #58946, because I think it's so important and would like it fixed before release
15:06 masak but I'm not sure that's realistic, so maybe I shouldn't add it after all. thoughts, anyone?
15:08 masak generally, what are the criteria for adding a dep to a ticket such as #58946?
15:08 moritz masak: I'd even say it can't be done before the release, because the required changes are too big, and likely break stuff
15:08 moritz masak: (although I really sympathize with you in this one)
15:08 masak I see.
15:09 masak that's what I suspected
15:09 masak we'll fix it till the next release instead :)
15:09 moritz well, till christmas ;)
15:09 masak *cough*
15:10 masak I'm thinking of having some sort of drive during the month November
15:10 masak blogging on use.perl every day or some such
15:12 masak I might add that naming one of your projects November creates funny effects in one's perception. I got the rent invoice for November the other day, and I just stared at it, thinking "what is this? I have to pay for November?" I'm sure similar things will keep happening during the coming month.
15:13 moritz lol
15:15 NotFound I have a dog called Marzo (march in spanish)
15:15 masak I see a trend here.
15:16 moritz masak: could you check if #57740 still segfaults for you?
15:16 masak but it's all been done -- Abbott and Costello have prior art
15:16 moritz ./perl6 -e 'my $s = "hi"; $s.=uc'
15:16 masak moritz: checking.
15:16 masak moritz: no, s'ok.
15:17 moritz then I'll close the ticket.
15:17 masak rakudo: %*ENV<PWD>.=uc # this doesn't work, however :)
15:17 polyglotbot OUTPUT[Parrot VM: Can't stat languages/perl6/perl6.pbc, code 2.␤main: Packfile loading failed␤]
15:18 masak oh, that's another error than on my box.
15:18 masak I got "Null PMC access in name()"
15:18 moritz that's the error from rebuilding parrot
15:18 masak ah.
15:18 moritz masak: there's a different ticket for modifying %ENV, iirc
15:18 masak moritz: aye. I replenished it today. :)
15:19 masak with lots of strange errors and effects.
15:19 moritz also #57866 looks like it could be rejected
15:19 masak aye, I think it should
15:21 masak moritz++ # ticket triaging
15:22 moritz (165-159)/165
15:22 purl 0.0363636363636364
15:22 particle masak: add the ticket
15:22 masak particle: oki
15:22 moritz 4% in 20 minutes. Not bad.
15:22 particle it's a pointer to me to say it's important
15:22 particle it's not a hard dependency
15:22 masak particle: good.
15:22 masak that's what I wanted to hear.
15:22 * masak adds
15:23 masak it's a bit sad, I know that ticket number by heart...
15:23 masak 58392! 58392! shout it from the rooftops! 58392!
15:25 * moritz shouts "fünf-acht-drei-neun-zwo!" out of his window
15:25 moritz ... and gets weird stares from his room mates
15:25 masak hehe
15:27 masak moritz: since when is it "zwo" rather than "zwei"? I've never studied German, but I thought I knew that one.
15:28 kj probably a dialect :-)
15:29 jhorwitz joined #parrot
15:29 masak note that I'm not complaining. it's closer both to the English and the Scandinavian word
15:30 clunker3 joined #parrot
15:32 moritz masak: it's southern dialekt, but somtimes used elsewhere because it's easier to distinguish from "drei"
15:32 masak ah. convenient.
15:33 NotFound Are you trying to implement languages/german ?
15:33 masak lol
15:33 Juerd_ #!/usr/bin/bratwurst
15:34 * masak is reminded of the programming language Sorted!, which can be written either with English keywords, or German ones, or both
15:34 NotFound I like the definition of german in uncyclopedia: "Talk like Yoda and drop the spaces"
15:35 moritz masak++ #  #57400
15:35 Juerd_ In elementary school they had a BASIC dialect in *Dutch*. I was horrified.
15:35 moritz if the force so strong is in master yoda, why proper english he can't speak?
15:36 Juerd_ If so strong the force in master Yoda is, ...
15:36 masak moritz: some of it is related to November, actually. I was tracking down a bug that sprang up when I tried to put in some .trans calls instead of looped .subst calls
15:37 NotFound The common language in Star Wars is called basic, not english ;)
15:37 masak forth strong the so is if Yoda master in, ...
15:37 particle O_o
15:38 NotFound At least in some books, I don't remember any film that mentions that name.
15:38 Juerd_ The trick to yodaspeak is just to put the predicate first.
15:39 masak Juerd_: the predicate? don't you mean the object?
15:39 NotFound Reverse Yodish Notation?
15:39 masak to me, the predicate in yodaspeak last seems to come.
15:40 masak sort of an SVO word order.
15:40 masak no, SOV, sorry.
15:41 Juerd_ masak: Sorry, I'm not using the right word. Indeed the object.
15:41 masak so, maybe even OSV, then.
15:41 Juerd_ Though it's not an object when there's a copula.
15:41 Juerd_ OSV, yes.
15:42 NotFound VSOP?
15:42 particle :)
15:42 masak Juerd_: agreed. and it can be other things in some cases as well.
15:42 Juerd_ masak: That it can.
15:42 particle i see the beginnings of a derived perl 6 grammar
15:43 particle @list { .method() } map;
15:43 NotFound if 'Hello' say ; $i 0 <
15:43 masak ouch.
15:43 Juerd_ It's just aform of RPN :D
15:43 masak Forth 6
15:43 NotFound Perloth X-)
15:45 masak actually, I wouldn't mind trying to write Forth code on top of the Perl 6 semantics. though I'm not (yet) prepared to sit down and write an interpreter
15:45 masak shouldn't be too difficult.
15:46 NotFound Surely it will be easier to write than to read ;)
15:46 masak I'm sure confusedly written Forth is no different from badly written Perl.
15:47 masak ...and brilliant, easy-to-read code can be written in both languages.
15:48 Theory joined #parrot
15:49 NotFound masak: the question is, is possible to write non confused Forth?
15:50 masak I think so.
15:50 NotFound I've never seen one, but I have seen very few Forth code.
15:51 masak actually, I can't really say if Forth is good for most tasks, or specialized like Prolog
15:52 masak I haven't done enough programming in it to get a feel for it
15:52 masak but nothingmuch tells me it's a _very_ nice language
15:53 moritz like most languages it has its applications
15:53 NotFound A simple interactive forth for parrot will surely be useful for testing.
15:54 * moritz thought about implementing RIES, a simple language that's used for formal proofs in CS
15:54 NotFound Counter Strike? ;)
15:55 masak Computer Science.
15:55 purl computer science is, like, no more about computers than astronomy is about telescopes. --E.W. Dijkstra
15:56 masak strange. no Wikipedia article about RIES.
15:57 NotFound masak: giving the state of the art, is not a big surprise that nobody hs interest in formal proofs ;)
15:58 masak :)
15:58 masak exactly, here I put in some effort, surfing to the Wikipedia page and all, and all I get is a blank page.
15:59 moritz http://theoretische.informatik.u​ni-wuerzburg.de/sonstiges/ries/
15:59 NotFound Maybe is an interpreter for the language written in whitespace.
15:59 moritz don't know if there's an English resource as well
16:00 masak ah -- "a Spartan programming language" :)
16:00 masak nobody can code like the Spartans could!
16:00 moritz in Sparta Russians code you!
16:01 NotFound 300 spartans can finish Hurd.
16:01 masak blindfolded.
16:02 * NotFound thinks about writing a script for 'Asterix and the coders'
16:16 apeiron joined #parrot
16:30 ab5tract joined #parrot
16:57 PacoLinux joined #parrot
17:09 rdice joined #parrot
17:14 cotto joined #parrot
17:22 Ivatar joined #parrot
17:32 davidfetter joined #parrot
17:43 dalek r31995 | Whiteknight++ | trunk:
17:43 dalek : [PDD] Add some more information about :unique_reg as per the discussion at RT#57636.
17:43 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31995
17:51 silug joined #parrot
17:53 dalek r31996 | Whiteknight++ | calling_conventions:
17:53 dalek : [calling_conventions] Add a quick macro flag to this function to turn on and off the "good" and "experimental" versions of it. Beats having to insert/remove comment blocks all over the place.
17:53 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31996
18:10 chromatic joined #parrot
18:21 rdice joined #parrot
18:32 ruoso joined #parrot
18:42 sjansen joined #parrot
18:48 rdice_ joined #parrot
18:59 dalek r31997 | cotto++ | trunk:
18:59 dalek : [pmc] update most read-only tests to catch only read-exceptions
18:59 dalek diff: http://www.parrotvm.org/svn​/parrot/revision?rev=31997
19:08 NordQ joined #parrot
19:20 cosimo_ joined #parrot
19:38 Theory joined #parrot
19:39 bacek joined #parrot
19:41 cosimo_ left #parrot
19:46 cosimo_ joined #parrot
20:06 Tene pmichaud: how's your schedule look for the next few days?
20:08 DietCoke joined #parrot
20:14 hercynium_ joined #parrot
20:18 cosimo_ left #parrot
20:20 dalek Krishna Sethuraman | Parrot Development on Windows:
20:20 dalek link: http://www.perlfoundation.org/parrot/i​ndex.cgi?parrot_development_on_windows
20:55 contingencyplan joined #parrot
20:57 bacek joined #parrot
21:06 Andy is the allison anywhere?
21:06 chromatic She's in this time zone, but I'm not sure she's online.
21:06 moritz seldom here
21:19 Whiteknight joined #parrot
21:20 bacek joined #parrot
21:39 johbar joined #parrot
21:43 chromatic Cosimo's integer test is really confusing.
21:44 chromatic Somehow in multiple passes across the PCC/C boundary, we lose information about the return PC instruction.
21:44 chromatic That might be a good one for Whiteknight or Tene to explore.
21:46 Andy I'm gonna post the ActiveState thing on PB and then ask for questions for Allison
21:50 Whiteknight what am I exploring?
21:50 Whiteknight I've already closed my ticket for the day, I thought I was home free :)
21:50 Tene ohcrap, I haven't even looked at any tickets yet.
21:50 Tene I mean... uh... I'm not here.
21:51 chromatic [PATCH] Add subclass tests on t/pmc/integer.t and RT #52198, from Cosimo Streppone.
21:52 nopaste "chromatic" at 69.64.234.10 pasted "Integer Subclassing Crashy Crashy" (48 lines) at http://nopaste.snit.ch/14310
21:53 chromatic The problem is that ->current_results in the calling context in the return value processing code is invalid.
21:53 chromatic I can't tell why it goes invalid, but my current guess is that crossing the PCC/C boundary multiple times is Mr. Not Healthy.
21:58 silug joined #parrot
22:01 Whiteknight I'll take a look into it a little later
22:06 chromatic pcf_P_JPPP() calls Parrot_Integer_multi_add_Integer_PMC() and then the crash occurs in Parrot_init_ret_nci().
22:17 tetragon joined #parrot
22:24 bacek_ joined #parrot
22:53 TiMBuS joined #parrot
23:13 Tene purl: msg pmichaud I'd really appreciate it if you could tell me what I can work on next on PCT.  I'd really like to get a decision from you on how to handle the namespace exporting and such.
23:13 purl Message for pmichaud stored.
23:15 Tene purl: msg pmichaud Also, should NQP get a .HLL?
23:15 purl Message for pmichaud stored.
23:17 Coke tene: in the meantime, feel free to hack on tcl. =-)
23:47 Ademan joined #parrot

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

Parrot | source cross referenced