Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2006-06-19

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:14 tup joined perl6
00:16 TimToady putter: nested atomically's maybe?
00:16 TimToady putter: re method vs submethod
00:16 TimToady you may not declare a method and submethod of the same name in a scope.
00:19 TimToady leaving aside the subcall visibility, a submethod behaves just like a method with a first statement that says:
00:20 TimToady next METHOD unless self.type === $?CLASS;  or some such...
00:31 weinig|dinner is now known as weinig
00:53 cmarcelo left perl6
00:55 hikozaemon joined perl6
01:05 meppl gute nacht
01:17 glasser joined perl6
01:18 glasser left perl6
01:33 mako132_ joined perl6
01:36 b00t joined perl6
01:41 frederico joined perl6
01:42 dduncan joined perl6
01:44 putter TimToady: thanks.
01:46 putter re "may not... in a scope", it might be worth clarifying S12's "You may mix methods and submethods of the same name within the class hierarchy, but only the methods are visible to derived classes via inheritance."
01:47 putter Hmm.  Does a submethod affect inheritance in any way?  Ie, does it prevent a parent's method from being inherited by a child (parent's grandchild)?
01:49 putter leaving aside signature problems, the next METHOD analogy above suggests not.
01:51 putter Can some clauses of a multi method be multi submethods?
01:56 putter I assume the bodies of a class decl and any class "is extended" decls are the same "scope" re submethods.  or not?  lexical scope or..., hmm, "class" scope?
01:59 putter If a role is included which defines a method, what is the effect of the class defining a submethod?
02:04 justatheory joined perl6
02:07 stevan putter!!!!!!!!!!
02:07 stevan putter: are you coming to the post YAPC::NA Boston hackathon?
02:08 justatheory joined perl6
02:08 cmarcelo joined perl6
02:13 putter hi stevan! :)
02:14 * putter goes to look for YAPC::NA Bos hackathon details...
02:14 stevan I dont know them officially
02:14 stevan ping obra about it
02:14 stevan July 2-6 I think
02:16 putter grumble grumble kwiki grumble.  google links to http://www.yapcchicago.org/wiki/in​dex.cgi?action=revisions&page_​name=HackAThon&revision_id=38
02:16 lambdabot Title: "HackAThon - Kwiki Revisions - Kwiki"
02:17 putter So... is that the most recent version?  url doesnt work without revision_id field.  doesnt work with =HEAD.  doesnt work with =<blank>.  =39 works. =139 doesnt.  Blech.
02:17 weinig is now known as weinig|zzz
02:18 putter Ah, http://www.yapcchicago.org​/wiki/index.cgi?HackAThon
02:18 lambdabot Title: "HackAThon - Kwiki"
02:18 stevan July 1 (and then continue in Boston until July 6)
02:18 stevan :)
02:18 stevan ping obra or audreyt ,.. he will give you details
02:19 stevan s/he/they/
02:24 putter ah.  removed some front page pugs.kwiki.org spam.  but some subpages are taking forever to obtain, suggesting they've been spamified.
02:25 putter PugsCommunity, ...
02:27 putter ok, just PugsCommunity of the links from the front page.
02:27 putter ah, got it.  solid spam.  trying to fix...
02:28 putter re ping, ok.  thanks.  modulo the jobsearch I should start rsn, I should be around, and that would be neat.
02:29 putter Any interest in talking to Boston.pm?  Had 20ish(?) people for this months inside-out object talk.
02:31 putter PugsCommunity fixed.
02:32 putter How goes moose?
02:37 putter audreyt, obra, stevean, etal: any interest in giving a talk to the boston perl mongers group?  http://boston.pm.org/kwiki/  July meeting isnt scheduled yet.  Normally second tuesday.  But have switched in the past when needed.
02:37 lambdabot Title: "- Boston.pm"
02:38 putter s/stevean/stevan/  sorry.
02:42 justatheory joined perl6
02:44 nothingmuch putter: i can regive any previous talk, or a new lightning talk
02:44 nothingmuch but too busy to do something "real"
02:45 putter real enough :)
02:45 putter could load spread and have a couple of people do lightning talks
02:45 nothingmuch okies
02:45 nothingmuch lightning talk topics of possible interest:
02:45 nothingmuch Test::TAP::Harness
02:46 nothingmuch Roles in the real world
02:46 nothingmuch Class::Workflow
02:46 nothingmuch Monju
02:46 nothingmuch (two interesting projects of the past while)
02:46 nothingmuch s/Harness/HTMLMatrix/;
02:48 putter neat
02:48 nothingmuch s/::/-/ and you can find darcs repos for both under nothingmuch.woobling.org
02:49 putter maybe I'll send the Boston.pm scheduler a note that folks will be around the first week of July for a hackathon, and there's the potential for someone to give a PM talk.  plausible?
02:49 nothingmuch okies
02:49 nothingmuch if you like off topic stuff too, my most well received talk was the darcs one
02:49 nothingmuch i'd like to spread that love
02:49 putter :)
03:07 Khisanth joined perl6
03:08 mako132_ joined perl6
03:14 putter audreyt, etal: is there a description of the argument-parameter distance metric for multiple dispatch somewhere?
03:15 putter subtyped parameters are said to be epsilon shorter than unconstrained parameters.  is there any gradation among subtypes, or just one epsilon?
03:17 putter how do inheritance, role inclusion, subseting, etc, affect distance?
03:18 putter junctions?
03:21 putter container types with a specified element type, in the face of large or infinite containers?
03:21 putter (that last one isn't distance, so much as... how?)
03:25 putter tnx.  night &
03:26 lambdabot joined perl6
03:30 topot joined perl6
03:44 wrath joined perl6
03:52 _SamB_ joined perl6
03:55 b00t joined perl6
03:57 tup joined perl6
04:13 _SamB_ is now known as SamB
04:16 mako132_ joined perl6
04:28 justatheory joined perl6
04:39 mtve joined perl6
04:39 Southen joined perl6
04:39 shachaf joined perl6
04:47 wrath left perl6
04:52 cmarcelo audreyt: hi. have some time to hsjudy stuff?
04:52 audreyt just woke up but sure
04:55 cmarcelo do you mind running a test (actually two) in your computer? it's (make testmap; ./testmap), twice, but in the second time, uncomment "show-last-take" line in tests/TestMap.hs
04:55 audreyt k
04:56 cmarcelo (if you don't have || want to install spinclds Judy, just ./configure;make in Judy-1.0.3 and change HsJudy/Makefile to static link)
04:58 audreyt all ok on testmap
04:58 cmarcelo on both ?
04:59 audreyt segfault on last
04:59 audreyt works if 100000 is chg to 10000
04:59 cmarcelo yep.. here too
05:00 audreyt tried gdb yet?
05:00 cmarcelo oh, and try removing "Dummy" stuff from Judy/Map.hs (was a hack to make Refeable work here), see if you get "illegal instruction"..
05:03 audreyt    No instance for (Judy.Map.Dummy [Char])
05:03 audreyt      arising from use of `lookup' at tests/TestMap.hs:107:9-19
05:03 audreyt    Possible fix:
05:03 audreyt      add an instance declaration for (Judy.Map.Dummy [Char])
05:03 cmarcelo i'm recompiling GHC 6.5 now.. gdb doesnt helped much (too much ???, probably i'm missing some compiler flag).
05:03 cmarcelo audreyt: remove Dummy from Refeable context
05:04 audreyt I got it too.
05:04 audreyt how curious. ask the ghc list?
05:05 cmarcelo audreyt: I tried to create a simpler version of the problem to send (not depending on all judy stuff) but couldn't =P..
05:06 audreyt nah, it's okay to send it as a whole :)
05:06 cmarcelo i'll see first if it works with latest GHC 6.5 from darcs..
05:11 cmarcelo besides that illegal instruction, type classing (for Collection stuff) worked fine. (sorry droping so much data on you =o) ) And, you told me about freeing StablePtrs in finalizer, but, this isn't safe as StablePtrs sometimes are the same for "different" objects.. picture two Maps, both with 1 pointing to a Big Fat Datatype called x. It is possible that GHC give same StablePtr for two insert calls (for the two maps), if we free one of
05:11 cmarcelo them and try to use the other => bad bad.. =P any ideas on this?
05:12 dduncan joined perl6
05:14 nothingmuch audreyt: hola
05:14 nothingmuch how was the flight?
05:17 dduncan fyi, I have uploaded a smoke of r10738 ... showing 17 errors
05:17 dduncan darwin / normal runcore
05:17 audreyt k
05:17 audreyt nothingmuch: exhausting
05:19 audreyt cmarcelo: uh, we can certainly use the maps themselves as keys
05:19 audreyt and only free the ptr when all uses of it is gone
05:19 audreyt I think
05:19 audreyt dduncan: will check. thanks!
05:19 audreyt but it's sleepy-brain talking
05:20 dduncan specifically, the errors seem to be with 2 files:
05:20 dduncan 1. t/operators/assign.t has a few failed tests in the middle
05:20 audreyt that's intentional 8
05:20 dduncan 2. t/pugsrun/05-unknown-option.t seems to have died after test 3
05:21 dduncan that's all I saw in a scan
05:22 dduncan if that is unexpected, I will now try to run the second by itself
05:23 TimToady ext/Perl6-Grammar/t/basic.t also seems to be failing for me.  Haven't looked at it though...
05:23 dduncan that one doesn't fail for me
05:24 audreyt dduncan: please run 05 by itself
05:24 audreyt TimToady: you'd need a new parrot for that to pass
05:24 audreyt new as in something in the last 24 hours
05:24 dduncan in progress ...
05:26 TimToady it's pretty darn new...
05:26 TimToady getting pugs: src/string.c:932: string_substr: Assertion `src->encoding && src->charset && !(((src)->obj.flags) & b_PObj_on_free_list_FLAG)' failed.
05:26 dduncan I am seeing not-okays ... will nopaste output
05:27 TimToady in fact, it's the most recent, 12969.
05:27 audreyt oh. embparrot.
05:27 audreyt I'll retest with embparrot.
05:27 audreyt TimToady: only basic.t triggers that?
05:28 audreyt none of t/rules/*?
05:28 TimToady yes, the only other is the 8 assign.t tests
05:28 cmarcelo_ joined perl6
05:28 pasteling "dduncan" at 24.69.53.198 pasted "detailed output of t/pugsrun/05-unknown-option.t on r10738" (89 lines, 3.8K) at http://sial.org/pbot/17843
05:29 TimToady http://m19s28.vlinux.de/iblech/stuff/pugs-s​mokes/pugs-smoke-6.2.12-r10738-linux-normal​--1150683199-2738--11550-11517-33-1591-1161​-0--82fde14b343f46790ec837e0ed1586a0.html is my smoke
05:30 audreyt k
05:30 cmarcelo_ audreyt: i wasn't clear enough. imagine m1 and m2, and object x.. when insert 1 x m1 and insert 1 x m2 happens, it's possible that GHC give the same StablePtr for insert function in both cases. then, after that, if it GC m1 but we are still using m2, the StablePtr for x isnt valid anymore...
05:30 cmarcelo_ is now known as cmarcelo
05:30 dduncan seems to be no-such-file stuff ... one might think there be an inter-test dependency, though I recall that some previous commit was meant to address that by running them in order, though ...
05:31 audreyt cmarcelo: yes I'm saying do a refcount-like scheme
05:31 audreyt to register that m1 and m2 both uses the SP
05:31 audreyt and only when both are gone should we remove the SP
05:31 audreyt Mem.Weak may have something that doe this
05:31 audreyt not sure
05:33 audreyt dduncan: it seems your manual run worked actually
05:33 dduncan looking at the log when the smoke was running, I don't see any order dependency problems ... all the other pugsrun were run in order first that needed to be
05:33 audreyt it ran 12 with all todos
05:33 dduncan I thought todos registered as 'ok'
05:34 dduncan or maybe its just skips that do
05:34 audreyt yeah
05:34 TimToady oddly, I don't get the assertion failure when I run basic.t from the top-level pugs directory.  I just get that when I make test in the Perl6-
05:34 dduncan maybe this is reminiscent of my last reported problem, where if I rerun it the problem may go away
05:35 TimToady Grammar dir
05:35 dduncan but like the last time, such a rerun will have to wait until tomorrow
05:35 TimToady it does fail all the tests after 6 though when run at the top.
05:36 cmarcelo audreyt:  i see.. will check that. also, in Pugs code [re including Judy.Map in IHash], some places that use IHash have in its context only MonadSTM m, but Judy.Map operations are IO dependent. are these functions used outside IO context indeed?
05:36 TimToady mine actually is a rerun--I noticed the basic.t problem earlier today.
05:37 TimToady so what I've got is after a complete rebuild of both parrot and pugs
05:37 TimToady I can fire off just a resmoke though and see if anything changes.
05:38 TimToady need to head off to bed anyway--still fighting off some kind of something from last week.
05:39 TimToady apparently a lot of it going around here--you feel like you're coming down with mono for about a week, and then you get a sore throat and mild cough for a week, then you get better.
05:39 TimToady still working on the last part of that...
05:46 TimToady actually, I can see I'm gonna get the same smoke result, since the same assertion has already come out on my screen log.
05:48 TimToady whole thing seems to be running about 30% slower too.
05:49 audreyt cmarcelo: we need to talk about that too, sometime later :)
05:49 audreyt cmarcelo: for now, use unsafeIOtoSTM.
05:50 audreyt TimToady: sleep tight
05:50 TimToady well, my computer just turned off the lights in my living room, so maybe it's trying to tell me something...
05:50 TimToady night all &
05:50 audreyt cmarcelo: in STM, operations need to know how to undo themselves
05:50 audreyt which Judy actually is probably capable of doing
05:50 audreyt but for now just use unsafeIOtoSTM
05:50 cmarcelo fine.
05:50 audreyt which means it may fail within atomically{} block
05:51 audreyt but we can manually collect just them later
05:53 cmarcelo ok =), i'll try to solve the too much memory segfault... actually, changing the values in TestMap.hs here gave me something like "Internal Error: GC'ed main thread".
06:05 audreyt can someone help me on unTODOing?
06:07 audreyt http://m19s28.vlinux.de/iblech/stuff/pugs-s​mokes/pugs-smoke-6.2.12-r10738-darwin-norma​l--1150697008-798--11550-11540-10-1589-1193​-18--9f158e3531a7af1a305cc2cc176d7797.html
06:07 svnbot6 r10739 | audreyt++ | * Pugs.Eval: the infereed SIMPLE context does not affect runtime
06:07 svnbot6 r10739 | audreyt++ |   anymore; it's now exclusively a parsetime thing.  So this:
06:07 svnbot6 r10739 | audreyt++ |     sub f { 1 }
06:07 svnbot6 r10739 | audreyt++ |     my @a = ({1=>2}, {3=>4})
06:07 audreyt thanks in advance :)
06:07 svnbot6 r10739 | audreyt++ |     @a[f()]{1}
06:07 svnbot6 r10739 | audreyt++ |   now works as you'd expect instead of failing with hash-derefing
06:07 svnbot6 r10739 | audreyt++ |   a slice.
06:07 cmarcelo unTODO = remove todo mark from todo tests that work OR make them work and then remove? =P
06:07 audreyt unTODO = look at the url above
06:07 audreyt look for yellow boxes
06:08 audreyt remove the :todo from tests in the corresponding places
06:08 audreyt and commit
06:08 audreyt that's all :)
06:08 audreyt lunch, bbiab
06:08 * spinclad can testify that it's easy
06:09 spinclad i'll help in a bit if someone else hasn't grabbed it
06:09 cmarcelo well. i'll do it..
06:11 spinclad my system's too sluggish for its browser atm
06:13 audreyt cmarcelo++
06:15 spinclad cmarcelo: when i get a chance here i'll look at your testmap case.  some of the HSIter code could step outside its memory footprints if i still have bugs in there.
06:15 christopher joined perl6
06:16 cmarcelo spinclad: ok. tks. which GHC are you running? right now HsJudy depends on GHC 6.5... =|
06:16 spinclad ah.  6.4.1 ( = debian latest)
06:17 spinclad i could try building 6.5 and get back to you tomorrow or so...
06:17 mjk joined perl6
06:18 cmarcelo spinclad++ # =)
06:25 svnbot6 r10740 | cmarcelo++ | * just unTodo'ing
06:25 spinclad oh, don't ++ me yet, building it could take me days here
06:26 cmarcelo spinclad: there are prebuilt snapshots i think..
06:26 spinclad will look...
06:26 cmarcelo http://www.haskell.org/ghc/dist/current/dist/
06:26 lambdabot Title: "Index of /ghc/dist/current/dist"
06:45 visq joined perl6
06:51 kane_ joined perl6
06:54 Ymmv joined perl6
06:54 svnbot6 r10741 | audreyt++ | * array_ref.t: reword obsolete test
07:14 iblechbot joined perl6
07:15 Noodle joined perl6
07:17 Noodle anyone here good with natural language parsing?
07:18 spinclad algorithmic or particular?
07:19 Noodle hmm.. not sure exactly you know more than I about them at this point, remember the old dos game ZORK? something like that
07:20 dduncan left perl6
07:20 spinclad fondly i do
07:20 spinclad never looked at its method tho
07:21 kane_ hmm, there's a module on cpan that does something like that
07:21 kane_ i read about it on use.perl
07:21 Noodle im accually doing my thing in C++ however I was told that you guys like to talk parsers :)
07:21 kane_ i think a blog post by ovid
07:22 Noodle so I figured I would hop on over and pick your brains about how the simplest way todo it would be. im not talking long commands at most maybe 4 -5 words
07:23 visq Noodle: maybe look at http://lambda-the-ultimate.org/node/1554
07:23 lambdabot Title: "Inform 7: A relational DSL for interactive fiction with natural language syntax  ..."
07:23 Noodle I know about inform
07:23 spinclad i imagine you could go a long way parsing zork-style english with perl6 regexes
07:24 spinclad doesn't really help you directly in c++ of course
07:24 Noodle true,
07:25 Noodle and yes regexes would help im sure, right now I figure I have the strings tokenized delimited by spaces. so that should make things a bit easier
07:27 Noodle but as it stands now i figure I have to check the first token for either the word go,move,put,get,take, and any variation of a direction. then if its go, or move, I have to check the second token for a direction, if its get I have to check the 2nd token for an item name, if its put I have to check the 2nd token for item and the 3rd will be probably an ignored word (on, with, in, etc) the 4th token will be the location to put the item.
07:27 Noodle as you can clearly see I have my work cut out for me.
07:28 Noodle and thats not counting all the others, read, open, etc
07:32 spinclad you could put some of the machinery in the lexicon: a token (word) has a bunch of syntactic/semantic features, like <verb>, <direction>, <takable>; a <verb> has a parser that says what objects it takes, like <direction> or <takable>
07:33 Noodle good idea..
07:34 spinclad a fancier english parser would build up phrase structure for syntactic/semantic categories, but you may be able to get by with verbs and single-word phrases: if you never have two bottles, you don't have to make sense of <the green bottle>
07:35 spinclad but it would be nice to be able to make sense of <pick up the green bottle and drink it>
07:35 Noodle so I would need verbs, adjectives, and nouns...
07:35 spinclad <pick up the bottle> <which one?> <the green one>
07:36 spinclad keeping track of what <it> could refer to...
07:36 Noodle this "little" exercise project of mine has turned into much much more....
07:36 spinclad this could get you a fair way
07:36 spinclad yes, parsing english is as big as you want it
07:38 Noodle normally I would say that if I made something like this in C++ that parsed the english language in a natural way I would be a hero to all the text adventure programmers, however Interactive fiction in C++ is well non existant. so its almost not worth it.
07:39 Aankhen`` joined perl6
07:47 Noodle left perl6
08:10 bjoern_ joined perl6
08:18 audreyt weird, embparrot worksforme without the assertion error larry encountered
08:19 audreyt resmoking just to be sure
08:21 `` joined perl6
08:31 lisppaste3 joined perl6
08:39 audreyt still can't duplicate it. oh well...
08:40 szbalint joined perl6
08:40 audreyt TimToady: I'll check in a revert-fix to the finalizing sequence; see if it works after you wake up
08:42 svnbot6 r10742 | audreyt++ | * Pugs.Embed.Parrot: Revert the finalization sequence to simply
08:42 svnbot6 r10742 | audreyt++ |   call parrot_exit, which may or may not fix the segfault
08:42 svnbot6 r10742 | audreyt++ |   TimToady reported...
08:45 marmic joined perl6
09:03 xerox joined perl6
09:04 xerox joined perl6
10:07 svnbot6 r10743 | audreyt++ | * assign.t: Correct TimToady++'s tests:
10:07 svnbot6 r10743 | audreyt++ |   In this:
10:07 svnbot6 r10743 | audreyt++ |     @a[0==1] = 1,2,3
10:07 svnbot6 r10743 | audreyt++ |   @a[0] ends up as holding a single (1,2,3) object, instead of
10:07 svnbot6 r10743 | audreyt++ |   just getting 1, according to the current S03 wording.
10:07 svnbot6 r10743 | audreyt++ |   We're down to two fail case now, both involving ($a)=L.
10:10 froh-doh joined perl6
10:41 kane-xs joined perl6
10:45 Aankhen`` What are haddock, happy and alex used for?
10:45 svnbot6 r10744 | benedikth++ | * util/catalog_tests.pl: use YAML::Syck if available, make apo/exe/syn optional, fix GetOptions
10:52 Aankhen`` Blargh... Test::TAP::Model seems to use Method::Alias, but doesn't list it as a prerequisite.
11:15 chris2 joined perl6
11:38 pasteling "Aankhen``" at 203.101.7.186 pasted "Error building Pugs on Win32" (12 lines, 517B) at http://sial.org/pbot/17848
11:38 Aankhen`` I'm not sure whether this an error with my config or something to do with Pugs...
11:38 Aankhen`` (In short, it's unable to find inc::Module::Install::Admin.)
11:44 Aankhen`` Gah.  All the distros using Module::Install are using an outdated version of it.
11:44 Aankhen`` (Which means it's nigh unusable on Windows.)
11:47 weinig|zzz is now known as weinig|work
11:51 lollan joined perl6
11:51 frederico joined perl6
11:56 jsiracusa joined perl6
12:08 christopher joined perl6
12:09 shachaf joined perl6
12:11 audreyt oh uhm. hm.
12:11 audreyt upgrading M::I fixes it?
12:11 audreyt weird, it should've used the version bundled with pugs.
12:11 iblechbot joined perl6
12:11 audreyt what's your M::I version?
12:12 tifo joined perl6
12:36 iblechbot_ joined perl6
12:41 gaal rhetorical gripe: all-caps suggests SIMPLE isn't.
12:43 gaal Aankhen``: (hi!!) haddock produces nice docs for .hs
12:43 gaal make haddock
12:44 gaal do we recommend happy and alex? I don't know why if we do. (they're a parser and "a lex"er)
12:59 lumi By rhetorical gripe, do you mean you gripe rhetorically, or that it is a gripe of rhetorics?
13:00 christopher Can I use Perl 6 rules today?  My pugs segfaults on them, and the smokes I've scanned skip the rule tests.  Even the CPAN module Perl6::Rules has its tests die on me.
13:01 wolverian christopher, you can use them in perl5 with http://search.cpan.org/author/FGLOCK/Pugs-C​ompiler-Rule-0.05/lib/Pugs/Compiler/Rule.pm
13:01 lambdabot Title: "Pugs::Compiler::Rule - Compiler for Perl 6 Rules - search.cpan.org"
13:01 gaal lumi: 1. it's a gripe on the rhetorical aspect of the name
13:01 gaal 2. :-P
13:01 christopher wolverian: Thanks!
13:16 svnbot6 r10745 | audreyt++ | * Update PIL Compiler to deal with this:
13:16 svnbot6 r10745 | audreyt++ |     ($x) = 1,2,3
13:16 svnbot6 r10745 | audreyt++ |   to parse as list.  Currently use a horrible (Sym SMy "") form
13:16 svnbot6 r10745 | audreyt++ |   for the lack of a better Noop annotation node.
13:18 audreyt going to catch plane now. more unTODO is prolly needed
13:18 audreyt will continue from airport
13:21 sahadev joined perl6
13:21 audreyt please unTODO the 4 yellow boxes in
13:21 audreyt http://m19s28.vlinux.de/iblech/stuff/pugs-s​mokes/pugs-smoke-6.2.12-r10738-darwin-norma​l--1150723262-747--11550-11534-16-1571-1193​-4--ab16a191337bee06398d1022b3a2f832.html
13:21 audreyt and maybe triage the other red ones
13:21 audreyt I'll be back in ~1hr
13:21 audreyt &
13:24 Lorn joined perl6
13:25 shachaf joined perl6
13:29 wolverian (untodoed, unfortunately I don't understand the red tests well enough to triage)
13:29 svnbot6 r10746 | wolverian++ | untodo some tests
13:33 Limbic_Region joined perl6
13:38 Lorn left perl6
13:48 alch` joined perl6
13:54 shachaf_ joined perl6
13:58 rzoz joined perl6
14:02 shachaf_ is now known as Shachaf
14:02 shachaf_ joined perl6
14:02 Shachaf is now known as shachaf
14:05 hexmode joined perl6
14:16 vel joined perl6
14:35 buetow joined perl6
14:54 audreyt TimToady: this is a super special case:
14:54 svnbot6 r10747 | audreyt++ | * Don't use the klugy Sym"" form; instead use the new Annotation
14:54 svnbot6 r10747 | audreyt++ |   type "Parens".
14:54 audreyt ($x) = 1,2,3
14:54 audreyt according to naive S03, it will parse as list
14:54 audreyt then put (1,2,3) into $x
14:54 audreyt but your test has it take 1 only
14:54 audreyt asif the user has written
14:54 audreyt ($x,) = 1,2,3
14:54 audreyt I assume it's correct
14:55 rlb3 joined perl6
14:58 TimToady Yes, it's super special.
14:59 audreyt what about
14:59 audreyt (lv()) = 1,2,3
14:59 audreyt wwieh sub lv is rw { $x }
14:59 audreyt s/wwieh/with/
14:59 audreyt same thing?
14:59 audreyt (easier if same thing)
14:59 TimToady still get the assertion error with the reversion, by the way.
14:59 audreyt i.e. () on lhs without , has an implicit ,
14:59 audreyt ok.
15:00 TimToady yes, same thing.  "Useless use of parens" around lvalue is always list.
15:00 audreyt ok, implemented. need to go boarding now
15:00 svnbot6 r10748 | audreyt++ | * assign.t: update according to latest discussion on #perl6
15:00 TimToady peddle hard.
15:01 audreyt :D
15:01 audreyt &
15:03 svnbot6 r10749 | audreyt++ | * according to TimToady++'s
15:03 svnbot6 r10749 | audreyt++ |   super-special rule,
15:03 svnbot6 r10749 | audreyt++ |   (...) = 1,2,3
15:03 svnbot6 r10749 | audreyt++ |   where ... has no , in it
15:03 svnbot6 r10749 | audreyt++ |   will have an implicit (...,) trailing comma
15:03 svnbot6 r10749 | audreyt++ |   added to it.
15:06 nnunley Ugh. yadayada (...) makes such commit messages ambugous (sic)
15:09 Aankhen`` gaal: Hi back at you. ^_^
15:09 svnbot6 r10750 | audreyt++ | * Finalize parrot again, now we're sure it's not the change
15:09 svnbot6 r10750 | audreyt++ |   that broken TimToady++'s build.  Now I wonder what is...
15:09 Aankhen`` audreyt: My M::I version is 0.63.
15:09 Aankhen`` audreyt: It fixes a known bug on case-insensitive filesystems, apparently.
15:09 Aankhen`` gaal: I'm not sure Pugs recommends happy or alex; they're mentioned in Makefile.PL and made use of if present, apparently.
15:10 Aankhen`` I like to say "apparently" a lot, apparently.
15:10 audreyt they are not used and not used
15:10 audreyt it's just standard Cabal probing
15:10 Aankhen`` audreyt: Ah, I see.
15:10 audreyt Aankhen``: can you trace a bit to see why the system M::I gets use
15:10 audreyt instead of our bundled inc/Module/Install.pm
15:10 audreyt ?
15:10 Aankhen`` audreyt: The system M::I isn't used, the one bundled with the distros is used.
15:10 Aankhen`` That's the problem.
15:10 audreyt oh.
15:10 Aankhen`` The one bundled is an older version.
15:11 Aankhen`` The one in the pugs/inc seems to be 0.34 or something like that.
15:11 audreyt 0.36
15:11 Aankhen`` Yeh.
15:11 audreyt but but, the case-insensitive @INC fix has been merged in
15:11 Aankhen`` Was this recent?
15:12 audreyt ...and not committed
15:12 Aankhen`` Ah.
15:12 audreyt sorry about this local branch stuff.
15:12 Aankhen`` How dare you? :-P
15:12 audreyt fixed. thanks!
15:12 Aankhen`` No, thank *you*. ;-)
15:13 audreyt I... blame Gisle for uncovering my bug
15:13 audreyt (ActiveState::Config was new and triggering this)
15:13 svnbot6 r10751 | audreyt++ | * Don't break ActivePerl 5.8.8 on case-insensitive filesystems.
15:13 svnbot6 r10751 | audreyt++ |   Aankhen++ for catching this age-old bug in inc/Module/Install.pm.
15:13 Aankhen`` Let's burn Gisele at the stake!
15:13 audreyt otoh, it is my bug ,and Gisle provided the said patch, so maybe not.
15:13 Aankhen`` Hrm.
15:13 Aankhen`` We can't burn you at the stake, we need you for Pugs.
15:13 audreyt (plus he's been kind to me in Oslo.)
15:14 Aankhen`` How about we burn evalbot_10709 instead?
15:15 Aankhen`` BTW, just for reference, I wasn't the one who caught the bug; it's in the ActivePerl bug tracker.
15:15 Aankhen`` The suggested workaround is to delete the inc/ directory, but that doesn't work too well in this case.
15:16 audreyt ?eval 1+1
15:17 evalbot_10709 is now known as evalbot_10746
15:17 evalbot_10746 (no output)
15:17 audreyt I didn't know that 1+1 can diverge
15:18 * audreyt investigates
15:18 * Limbic_Region thought audreyt was supposed to be boarding?
15:19 audreyt yes the queue is long
15:19 audreyt so I'm slowly moving forward
15:19 * theorbtwo hopes audreyt's laptop is durable.
15:22 svnbot6 r10752 | audreyt++ | * remove a case of diverge in Pugs.AST.
15:23 audreyt yay, assign.t all passes.
15:25 audreyt it's my turn now. see ya!
15:25 audreyt *wave* &
15:26 spinclad safe flight
15:26 theorbtwo Have fun, whereever it is you're going.
15:27 justatheory joined perl6
15:28 shachaf joined perl6
15:31 * spinclad ponders the fashion prospects for the style known as The Wearable Borg
15:33 lisppaste3 joined perl6
15:35 * spinclad ponders the consequences for a borg of having to switch off all external devices during takeoff and landing
15:37 lumi Maybe they have a low emissions mode
15:37 spinclad separation from the hive mind and the greater sensorium -- could lead to serious cravings
15:37 lumi They switch to tightbeam
15:38 spinclad aha.  i know i feel some of that when my isp is flaking out for a day.  56k is just too narrow a beam for me these days.
15:39 spinclad but somehow i think what i feel really can't compare
15:39 spinclad to the real experience
15:42 lumi 56k.. *shudder*
15:42 lumi Better nothing than that
15:42 lumi Less exasperating
15:43 spinclad overnight downloads of the latest updates
15:43 spinclad night after night
15:44 spinclad /me is *very* glad to be quit of that
15:45 spinclad now /me can spend /my overnights on more important things
15:46 spinclad like the funnies
15:46 spinclad  and pugs
15:47 Qiang joined perl6
16:08 shachaf joined perl6
16:30 ludan joined perl6
16:32 shachaf joined perl6
16:34 amnesiac joined perl6
16:35 neoesque joined perl6
16:45 TimToady audreyt: for your BLing pleasure, one hint may be that I get the following two warnings:
16:45 TimToady Compiling Pugs.Embed.Parrot ( src/Pugs/Embed/Parrot.hs, dist/build/src/Pugs/Embed/Parrot.o )
16:45 TimToady Warning: retaining unknown function `Pugs_callSubSSS' in output from C compiler
16:45 TimToady Warning: retaining unknown function `Pugs_trap_exit' in output from C compiler
16:47 ludan someone knows the name of the first browser developed in the Pisa University?
16:52 fglock joined perl6
16:54 dudleyf joined perl6
16:57 fglock drbean: re http://colabti.de/irclogger/irclogger_log​/perl6?date=2006-06-18,Sun&amp;sel=11#l20 - I've noticed overloading doesn't work if you don't 'use' the match module - I don't know why
16:57 lambdabot Title: "#perl6 2006-06-18,Sun"
16:59 ruz joined perl6
16:59 fglock cmarcelo: ping
17:02 cmarcelo fglock: hey
17:02 pdcawley joined perl6
17:03 fglock cmarcelo: would you give a talk in YAPC::SouthAmerica::2006 - the site to register is http://www.conisli.org/
17:03 lambdabot Title: ":: CONISLI ::"
17:10 cmarcelo fglock: well, i'm not sure about what... (got my pvts?)
17:11 cmarcelo since i'm new to both haskell and perl6 worlds.
17:13 shachaf joined perl6
17:17 fglock cmarcelo: (got)
17:34 bernhard joined perl6
17:35 ruoso joined perl6
17:40 trym joined perl6
17:47 fglock cmarcelo: please send me a mail - fglock at gmail com
17:59 cmarcelo fglock: sent
18:17 bpederse joined perl6
18:20 Steve_p joined perl6
18:31 Steve_p Hmmm...re0
18:31 Steve_p err...pre0
18:32 FurnaceBoy joined perl6
18:33 wilx joined perl6
18:57 Ymmv joined perl6
19:25 bene joined perl6
19:27 visq joined perl6
19:41 xerox joined perl6
19:41 xerox joined perl6
20:14 masak joined perl6
20:15 vel joined perl6
20:32 dduncan joined perl6
20:41 larsen joined perl6
21:05 dudleyf left perl6
21:08 buetow joined perl6
21:16 frederico joined perl6
21:35 weinig|work is now known as weinig
21:39 dolmen joined perl6
21:41 pasteling "TimToady" at 64.0.7.8 pasted "stack trace of ext/Perl6-Grammar/t/basic.t failure" (55 lines, 2.4K) at http://sial.org/pbot/17857
21:42 TimToady audreyt: see above.  looks like a GC failure of some sort.
21:42 TimToady the saneify_string macro is failing the PObj_on_free_list_TEST
21:44 TimToady I can trim the test case down to about 1000 bytes, but changing almost anything in that makes it not fail.
21:45 fglock joined perl6
21:45 TimToady this is with parrot r12975
21:49 pasteling "TimToady" at 64.0.7.8 pasted "minimal-ish test case of basic.t failure" (23 lines, 886B) at http://sial.org/pbot/17858
21:50 TimToady audreyt: actually, looks like I got it down to 886 bytes...
21:53 jsiracusa joined perl6
22:12 fglock left perl6
22:17 Limbic_Region joined perl6
22:32 ruoso joined perl6
22:48 dduncan joined perl6
23:24 khisanth_ joined perl6
23:25 khisanth_ is now known as Khisanth
23:25 weinig is now known as weinig|dinner
23:47 TimToady audreyt: may or may not be related, but I ran a test case reducer that spat out this:
23:47 TimToady regex ck{<x }/a/
23:47 TimToady spits out:
23:48 TimToady error:imcc:syntax error, unexpected ']'
23:48 TimToady in file 'EVAL_1' line 67
23:48 TimToady pugs: compilers/imcc/pbc.c:1086: verify_signature: Assertion `(((sig_arr)->obj.flags) & PObj_is_PMC_FLAG)' failed.
23:48 TimToady Aborted

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

Perl 6 | Reference Documentation | Rakudo