Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2007-07-27

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:23 weinig joined #perl6
00:24 weinig joined #perl6
00:32 pjcj joined #perl6
00:32 demq joined #perl6
00:32 Patterner joined #perl6
00:32 pugs_svnbot joined #perl6
00:32 dduncan joined #perl6
00:32 idiotmax joined #perl6
00:32 ashleyb joined #perl6
00:32 BinGOs joined #perl6
00:32 gaal joined #perl6
00:32 Gothmog_ joined #perl6
00:32 blblack joined #perl6
00:32 args joined #perl6
00:32 sri_ joined #perl6
00:32 rhr joined #perl6
00:32 sysdef joined #perl6
00:32 GeJ joined #perl6
00:32 audreyt_ joined #perl6
00:32 clkao joined #perl6
00:32 takanori joined #perl6
00:32 meppl joined #perl6
00:32 statico joined #perl6
00:32 felipe joined #perl6
00:32 BooK joined #perl6
00:32 mr_ank joined #perl6
00:32 SubStack joined #perl6
00:32 JarJarBinks joined #perl6
00:32 theorbtwo joined #perl6
00:32 integral joined #perl6
00:32 Juerd joined #perl6
00:32 wolverian joined #perl6
00:32 PerlJam joined #perl6
00:32 Yappo joined #perl6
00:32 F7 joined #perl6
00:32 Eidolos joined #perl6
00:35 sunnavy joined #perl6
00:35 pugs_svnbot r17148 | perlDreamer++ | Test for when the list is explicit, and not a variable. This
00:35 pugs_svnbot r17148 | perlDreamer++ | fails on parrot currently.
00:35 pugs_svnbot diff: http://dev.pugscode.org/changeset/17148
00:35 lambdabot Title: Changeset 17148 - Pugs - Trac
00:45 weinig joined #perl6
00:56 Schwern joined #perl6
00:58 nipotaway changed the nick to nipotan
01:25 thoughtpolice joined #perl6
01:36 nekokak joined #perl6
01:44 Jmax joined #perl6
01:44 nekokak joined #perl6
01:49 dolmans joined #perl6
02:02 Caelum joined #perl6
02:18 weinig joined #perl6
02:31 Schwern joined #perl6
02:46 aindilis joined #perl6
02:51 kanru joined #perl6
03:40 dduncan joined #perl6
03:44 Schwern joined #perl6
03:47 thoughtpolice hm. seems to have been pretty quiet lately. then again i've not been around too much I suppose :x
03:49 diakopter thoughtpolice: remind me again, you're good with haskell?
03:55 amnesiac joined #perl6
03:57 jiing joined #perl6
03:59 TheSocialistAven joined #perl6
04:01 Khisanth joined #perl6
04:04 TheSocialistAven left #perl6
04:07 xinming joined #perl6
04:14 thoughtpolice diakopter: i wouldn't exactly call myself 'advanced' but i feel i can manage ;)
04:20 diakopter thoughtpolice: neat.  I'd like to schedule a bit of an irc meeting among the haskellites
04:20 diakopter sometime in the next week or two...
04:24 thoughtpolice diakopter: cool.
04:24 thoughtpolice diakopter: if you can alert me ahead of time; my schedule's been a little crazy lately so a forecast would be appreciated :)
04:26 diakopter yes, there's no hurry (within reason).  I'd like to try to find a time when everyone interested can be involved in realtime.
04:27 diakopter any tips on scheduling logistics for globally synchronized meetings? :)
04:27 Tene thoughtpolice: I bet it would help diakopter if you gave him some idea of when you might be available.
04:29 diakopter Tene: there you go :)
04:29 Tene diakopter: date -d "Sun Jul 10 20:00:00 UTC 2007"
04:29 thoughtpolice in general if you just tell me i can make myself available then
04:29 thoughtpolice but with no warning i have no guarantees.
04:29 Tene Give datetime with an explicit timezone (preferrable UTC) and use date -d to display it in local time.
04:32 BinGOs joined #perl6
04:32 Tene (the datetime I gave as an example was, of course, just an example)
04:38 diakopter Tene: oh, you mean you can't be available two weeks ago?
04:38 diakopter oh, but the 10th wasn't a Sunday...
04:43 BinGOs joined #perl6
04:48 BinGOs_ joined #perl6
04:50 justatheory joined #perl6
05:10 Schwern joined #perl6
05:15 lumi diakopter: What's the meeting for
05:17 lumi s/$/?/
05:34 devogon joined #perl6
05:41 jisom joined #perl6
05:47 nipotan changed the nick to nipotaway
05:56 diakopter lumi: I'm interested in asking some questions of the former-lambdacamels as well as the haskellites who came along recently... about Perl 6 on Haskell.  I have some ideas about different backend/frontend ideas that I'd like to run by folks who know functional programming (and Haskell and GHC in particular)...
05:57 elmex joined #perl6
05:58 diakopter er, different backend/frontend models, I mean.
05:58 diakopter ideas about ideas? ;)
05:58 * diakopter winks at himself, he's so sleepy.
06:01 jisom_ joined #perl6
06:04 kanru joined #perl6
06:10 awwaiid joined #perl6
06:15 Aankhen`` joined #perl6
06:26 iblechbot joined #perl6
06:36 buetow joined #perl6
06:44 tcliou changed the nick to tcliouAway
06:51 masak joined #perl6
06:54 masak joined #perl6
07:09 torz joined #perl6
07:11 torz hey all.
07:12 masak torz: hello
07:16 polettix joined #perl6
07:21 masak joined #perl6
07:25 polettix masak: hello :) yesterday I lost connection but I didn't realise, did you ask me something? I have no log here
07:25 masak joined #perl6
07:25 polettix rotfl
07:25 polettix masak: hello :) yesterday I lost connection but I didn't realise, did you ask me something? I have no log here
07:25 masak polettix: um, yes, but it was a fairly impertinent question. I asked if you could help me debugging something
07:26 masak solely on the grounds that you seemed to have time on your hands and that it's more fun and speedier to debug with someone
07:27 polettix I have some time now, but I fear you'll have to wait until I get my pugs refreshed
07:27 masak no problem
07:27 BinGOs_ changed the nick to BinGOs
07:28 masak when you're ready, I'd like to turn your attention towards ext/DateTime/t/basic.t
07:28 polettix ok, I'll give you a call
07:29 masak great
07:35 polettix uhm, I was a bit behind... and now it seems that I have to upgrade ghc as well
07:35 masak polettix: ah, yes. the 6.6.1 thing
07:36 masak take your time. I'm sitting here studying Latin in the meantime
07:36 polettix Latin? amazing
07:36 masak polettix: Latin is an eye-opener. highly recommended if you like languages
07:36 polettix I Vitelli dei romani sono belli
07:37 masak that sounded very italian to me. was it?
07:37 polettix this sentence has a meaning in Latin, and a completely different meaning in Italian :)
07:37 masak polettix: oh. what meanings?
07:38 polettix In Latin it's: "Go, Vitellio, (now that you can hear) the sound of war of romans gods"
07:38 polettix more or less
07:38 masak ok -- I guess that must be medieval latin
07:39 masak never seen imperative "I" for "go" before
07:39 polettix in Italian it's pretty lame: "romans' calves are handsome"
07:39 polettix eo is ivi itum ire
07:39 masak polettix: :)
07:39 polettix don't really remember the imperative form, anyway :)
07:40 masak so "vitelli" is "calves" in italian?
07:40 polettix only if "calves" is the plural of "calf" :)
07:40 masak :)
07:41 * masak is fairly certain it is
07:41 masak yup, it is
07:41 polettix I guessed right, I hope
07:41 masak interesting sentence
07:41 * masak is reading Augustus' Momentum Ancyranum
07:43 polettix ok, I'm compiling pugs at last
07:44 polettix meanwhile, I think I can go and take a look at the test you pointed me to
07:45 polettix please take into account that I know very little of Perl 6
07:46 masak polettix: no problem. that's the beauty of the tests, they tell you if things are working or not
07:46 masak I'm hardly an expert either, I'm just fervently eager to help triaging
07:48 franck___ joined #perl6
07:49 polettix just for curiosity, why the eval-s in test #2?
07:50 masak hold on
07:50 masak could be a fossil from where that syntax broke pugs in some way
07:51 masak I don't see why it would be needed for #2 and not for #3 and #4
07:51 jsmock joined #perl6
07:51 polettix yep
07:52 * masak would like a (jump-to-test number) function in emacs
07:52 masak anyway
07:52 polettix now... the second block of tests and the third one are a repetition?
07:52 polettix ah no
07:52 polettix new() in the first, now() in the second
07:52 masak yep
07:53 masak it's the now() that screws up, I think
07:53 masak have you run the tests?
07:53 masak or are you still compiling?
07:53 polettix not yet, I'm looking at the smoke results
07:53 polettix http://m19s28.vlinux.de/iblech/stuff/pugs-s​mokes/pugs-smoke-6.2.13-r17147-linux-normal​--1185491801-9496--18817-17553-1264-1965-74​2-46--2a438ce684ce329a6200f7e4e6620cfe.html
07:53 lambdabot Title: TAP Matrix - Thu Jul 26 23:16:16 2007 GMT, http://tinyurl.com/yq9tbq
07:54 masak polettix: that's fine
07:54 dduncan left #perl6
07:55 masak now you see why I was attracted to this test file
07:55 masak it's the first one with red entries
07:56 polettix ahahah starting from the beginning
07:57 masak yes
07:57 masak there are 1264 tests failing :)
07:57 masak many of the probably for the same reason, though
07:58 polettix what does   $!year = $time.year    mean?!?
07:59 masak polettix: I must punt to S12 on that one
07:59 polettix ok
07:59 masak but I asked myself that as well
08:00 masak ok, so what can we say generally?
08:01 masak the Date objects are created
08:01 polettix but I still don't get it fully
08:01 masak which accounts for the green boxes in between the red ones
08:01 polettix I mean, the declaration is   has Int $.year;
08:01 masak yes
08:01 polettix so it does not seem to be a private variable
08:02 masak was the ! for private variables?
08:02 polettix Private attributes use an exclamation to indicate that no public accessor is generated.
08:02 polettix        has $!brain;
08:02 masak ok
08:03 polettix compilation complete
08:03 masak I tried replacing the '!'s with '.'s in my Date.pm but that didn't help much
08:03 pro21ms4 joined #perl6
08:04 polettix if you can help me a tiny bit I'll be operative in a snap
08:04 masak sure
08:04 masak if I can
08:04 polettix how can I quickly run the tests for Date?
08:04 masak ah
08:04 masak I do this
08:04 polettix ensuring that I use the recently compiled pugs, of course
08:05 Schwern joined #perl6
08:05 masak ./pugs -Iblib6/lib ext/DateTime/t/basic.t
08:06 polettix ok, it run
08:06 masak (actually I have made an alias pugs='./pugs -Iblib6/lib')
08:06 masak cool
08:07 polettix the test output does not show the difference between the expected and the got values, this is annoying
08:07 masak polettix: I noticed that too
08:08 masak maybe I should start by fixing that :)
08:08 masak should be Test.pm
08:08 polettix yep. It seems that it thinks we're in 2000 instead of 2007
08:08 masak polettix: it does?
08:09 polettix I added a couple of say()s
08:09 masak I think I got undef as an actual value
08:09 masak ah
08:09 masak now for the $1000 question: why?
08:09 polettix 2000/1/1 to be exact
08:09 polettix which BUILD method/methods is/are triggered?
08:10 masak polettix: that looks like zero fields
08:10 masak polettix: not sure that any are. you could check that too
08:10 polettix I'm trying to add some say()s to the lib too
08:11 masak good.
08:11 masak meanwhile, I'm in Test.pm
08:13 masak seems report_failure is using &?CALLER::CALLER::ROUTINE.name to determine whether it should print the expected value
08:14 polettix the Int|Num BUILD() is called, instead of the bare one
08:14 masak polettix: not both? I thought the latter called the former
08:15 polettix yep, I thought this as well
08:15 masak hm.
08:15 masak still, the instance variables should be initiated
08:17 daxim_ changed the nick to daxim
08:19 masak aha!
08:19 masak &?CALLER::CALLER::ROUTINE.name isn't "&Test::is" anymore, but "&is"
08:20 masak changing Test.pm to reflect this
08:20 masak that worked :)
08:23 bsb left #perl6
08:24 polettix I'm stuck
08:24 polettix Int|Num BUILD() with epoch ''
08:25 polettix the Int|Num version of BUILD() is called
08:25 polettix but $epoch is empty
08:25 masak polettix: does it matter?
08:25 masak polettix: or rather, would you expect it to be empty?
08:26 polettix I expected that another method would be called
08:26 polettix that is, the one with the empty parameter list
08:26 masak yes
08:27 masak but doesn't that one just delegate to the one with $epoch?
08:28 polettix so there is a quick hack to make the whole thing work, but we should probably investigate why it does not call the expected method
08:28 masak polettix: yes
08:28 masak definitely
08:28 masak I looked a bit at the tests for BUILD methods
08:29 masak to see if there were corresponding failing tests for what we see
08:29 masak didn't find much
08:29 masak but maybe you will fare better if you look
08:33 masak aha again: according to the tests (t/magicals/subname.t) the Test.pm as it stands is actually correct, and &?CALLER::CALLER::ROUTINE.name is wrongly implemented
08:33 masak seems like names should always be fully qualified
08:35 polettix wow, there's a real mess in calling those BUILD methods
08:35 polettix the call inside now() actually triggers the year/month/day version
08:35 masak um
08:36 masak sounds uncalled for
08:36 masak if you see an easy way of cleaning up the call chain while passing the tests, don't hesitate
08:39 * masak is lost in Eval.hs now
08:40 polettix I have to go to a presentation now, but I'll think about it and return in the afternoon
08:40 masak polettix: excellent
08:40 polettix thank you for giving me yet another time drain :D
08:40 masak polettix: np
08:40 masak thanks for the help so far
08:40 masak it was fun
08:40 masak :)
08:41 polettix yep, and it was useful too: you corrected Test.pm at last!
08:41 masak polettix: actually, no
08:41 masak Test.pm is correct
08:41 masak I'm looking at where pugs is going wrong now
08:41 masak (I have a simple fix, but I don't want to apply it since it will break as soon as Pugs unbreaks -- better to fix Pugs)
08:42 masak but Haskell... still confuses me
08:42 masak calling all competent Haskell people -- could someone lend a teeny, tiny hand? :/
08:43 masak actually, it's not as much unfamiliarity with Haskell that is the bottleneck in this case, as unfamiliarity with the structure of the Pugs source tree
08:44 spinclad (i vitelli...) 'Go, Vitellius, because of (hearing) the sound of the roman god of war'  (one roman god)
08:44 masak I need to find where &?CALLER::CALLER::ROUTINE.name is evaluated, and make it a qualified name (such as "&Test::is") instead of a local name (such as "&is")
08:44 spinclad yes, an interesting sentence
08:45 masak spinclad: is it "because of" because "sono" is ablative?
08:45 spinclad that's how i read it
08:45 masak makes sense
08:46 masak spinclad: you don't happen to be intimately familiar with the Pugs source, do you?
08:46 spinclad 'from the sound', as it were
08:46 polettix spinclad: absolutely, one god! See, "dei" in italian is also "gods", it was easy to get tricked :)
08:46 masak spinclad: something like that
08:46 spinclad not intimately by a long shot
08:47 masak spinclad: but maybe enough for the above little excursion?
08:47 masak I need to find where .name is evaluated for ROUTINEs
08:47 masak any idea where to start?
08:47 spinclad i can look, without promises
08:48 masak :)
08:48 polettix ttfn, byez
08:49 larsen joined #perl6
08:50 masak anyone know how to invite lambdabot to a freenode channel?
08:51 masak @tell masak something
08:51 spinclad you can tell yourself!
08:51 spinclad :)
08:52 masak :D
08:52 masak spinclad: substituting for lambdabot... now there's an unenviable task
08:53 spinclad i'd probably forget the message, or be asleep, or such
08:54 masak :)
08:56 masak actually, the homepage of lambdabot is down, not sure if that's related
08:56 * masak asks on #haskell
09:00 masak turns out lambdabot is down due to power shutdowns until sunday
09:01 spinclad you'll _have_ to tell yourself, then
09:01 masak ok
09:02 masak masak: something
09:02 masak masak: what?
09:02 masak masak: nvm
09:02 masak masak: ok... :/
09:02 spinclad masak: stop mumbling!
09:02 masak :)
09:02 spinclad ext/Test/lib/Test.pm?
09:02 masak spinclad: well, sort of
09:03 masak it relies on the results of ROUTINE.name being qualified
09:03 spinclad sub report_failure ...
09:03 masak as in "&Test:is"
09:03 masak spinclad: yes
09:03 jsmock left #perl6
09:03 masak right now they are only local
09:03 masak as in "&is"
09:03 masak see also t/magicals/subnames.t
09:04 masak which hints at the same thing
09:06 spinclad right
09:06 masak where to turn in src/ to rectify?
09:10 spinclad 'ack -l ROUTINE src' gives six hits: AST/CapInternals.hs, Prim.hs, Monads.hs, Eval.hs, Val.hs, Parser.hs
09:11 masak yes. already tried that
09:11 Schwern joined #perl6
09:11 masak even looked a bit inside Eval.hs, which seemed a likely candidate
09:12 masak that only yielded a heightened respect for Haskell programmers
09:12 * masak would like to start hacking Pugs, but the learning curve is steep
09:12 flokuehn joined #perl6
09:20 masak spinclad: ok, let's reason this out
09:20 masak I don't know what AST/CapInternals.hs, but it sounds like it has something to do with Captures, which is probably not what I want
09:21 masak Prim.hs actually also sounds like a likely candidate, assuming it actually handles primitive things and ROUTINE is such a thing
09:21 masak by the same token, it could be in Val.hs
09:22 masak Eval.hs felt like a likely candidate because it should handle evaluation of things
09:22 masak Parser.hs feels like too early a stage
09:22 masak I don't know where Monads.hs gets into it
09:23 * masak checks Prim.hs and Val.hs
09:30 zamolxes joined #perl6
09:32 spinclad i'm looking at 'data Var = ... VarDynamic { v_name, v_packageName }', defined in AST/CapInternals, used in ... Eval and Parser that i see
09:32 spinclad (Val's is a comment)
09:32 jisom joined #perl6
09:33 spinclad (Prim is for the prelude, maybe)
09:33 masak spinclad: ok
09:34 * spinclad suspicious of Pa4rser.hs:2100
09:34 masak I've found at Var.hs:88 the place where CALLER is returned
09:34 masak ok, heading there
09:34 spinclad s/4//
09:35 masak spinclad: yep, that's it
09:35 masak now to figure out how to get the package in there
09:36 spinclad v_packageName should already be in the Var
09:36 masak why?
09:38 masak I mean, "hm?"
09:39 spinclad hm... this is converting a $.foo to a method &foo call... i don't think this is it
09:39 masak dang
09:39 masak so close
09:40 masak is it reasonable to think that v_name would be involved in the code we are looking for, though?
09:40 spinclad it's a guess...
09:40 masak spinclad: it's in most .hs files, it seems :)
09:40 masak I'll have a look in Val/Code.hs, that sounds like a likely name
09:41 spinclad thinking that our problem is lack of reporting v_packageName when it should
09:42 masak spinclad: exactly
09:43 masak it returns v_name instead of v_packageName ++ "::" ++ v_name (not sure if that's correct Haskell syntax)
09:43 spinclad i'd like to track down how '.name' is defined; if that's just pulling out v_name, we would never hear a package name
09:43 masak spinclad: sounds like a plan
09:43 masak that's why I homed in on Val/Code.hs, but right now that does not seem to be it
09:45 masak there are also vars called v_package and v_longname which might be of interest
10:00 spinclad the var should constructed by the lexer - parser chain ...
10:01 masak another Maybe Clue: &?ROUTINE returns a value of type Sub, and this word exists in only a few .hs files
10:02 masak Sub, in turn seems closely related to VCode
10:03 masak ...which is basically everywhere :/
10:10 masak & # lunch
10:10 spinclad &?ROUTINE is MScalarRoutine, AST/CapInternals/Instances.hs ...
10:11 SixF9 joined #perl6
10:11 spinclad but that seems to be backend, emitting side stuff
10:13 spinclad s:7th/ /-/
10:24 chris2 joined #perl6
10:47 * spinclad now thinks '.name' is Prim.hs:625: op1 "Code::name"
10:48 meppl gugu
10:48 spinclad gugu tuyutu, meppl
10:48 meppl hello spinclad
10:48 meppl :-P
10:49 spinclad good morning meppl
10:51 spinclad '.name' is op1CodeName in Prim/Code.hs:23
10:53 spinclad which packages up subName, which is everywhere...
10:55 spinclad AST/Internals.hs:1039:    , subName           :: !ByteString
11:01 spinclad a field in VCode
11:01 spinclad set in many places
11:04 larsen_ joined #perl6
11:13 masak meppl: gugu
11:14 masak spinclad: sounds good. cheching Prim.hs:625
11:15 pjcj joined #perl6
11:19 masak spinclad: wonder if anything else will break if I just redefine op1CodeName to return subPackage::subName instead of just subName
11:19 masak it really depends on what calls op1CodeName
11:19 spinclad Parser.hs:314 ruleSubDeclaration: ...
11:20 masak what about it?
11:20 spinclad 348: nameQualified ... | otherwise = name  
11:20 spinclad 366: builds the VCode
11:22 spinclad 351: | otherwise = name  looks like the place to change to   (something pkg) ++ "::" ++ name
11:23 masak I'll give it a shot
11:24 pjcj joined #perl6
11:24 toshinori joined #perl6
11:28 spinclad or your change to op1CodeName might be the thing to do
11:28 masak yes
11:28 masak I seem to have broken things now
11:32 * spinclad doesn't grok 'cast' yet
11:32 * masak struggles against GHC's typechecker
11:32 masak how do I create a ByteString from [Char]?
11:34 spinclad sorry, haven't gone there yet.
11:34 * masak is really close now
11:35 * spinclad not qualified for brain surgery.  YMMV.  objects in mirror are closing fast.
11:36 masak :)
11:36 masak we're in this together
11:36 spinclad but once it typechecks, that's already proof of some kind of correctness
11:36 masak there's got to be a way!
11:36 masak if audreyt++ can do it...
11:38 spinclad indeed.  but i'm still looking at the code and its design through a narrow window
11:39 asksh changed the nick to dwave
11:41 masak spinclad: me too
11:41 spinclad running design inference on the code; dual to code generation
11:44 cognominal_ joined #perl6
11:46 spinclad there should be a haskell module Data::ByteString already available
11:46 spinclad or some such name
11:47 masak spinclad: yes, sure. the question is how to "cast" from String
11:48 masak all I get so far is that both are compatible with List
11:48 rindolf joined #perl6
11:48 masak but I haven't found any single function to convert
11:49 masak the situation is further complicated by dons's site (with the API and all the explanations) being down for maintenance. relying on Google cache
11:49 cmarcelo joined #perl6
11:50 spinclad pack?  (from the name)
11:51 masak spinclad: yes, just found it :)
11:51 * masak tries
11:54 falseep_ joined #perl6
11:55 masak ah! cast does Whatever -> ByteString
11:56 spinclad does Whatever -> ThisType, maybe
11:57 masak I was adding my code before this conversion, instead of after
11:57 masak now it seems to be working much better :)
11:58 spinclad UTF8.hs has pack etc
11:58 masak no need, I think
11:58 masak the whole thing compiles
11:59 * masak tests
11:59 * spinclad sees (cast :: Type1 -> Type2) various places
11:59 masak ...almost
11:59 masak I now get Main::&foo instead of &Main::foo
12:00 masak this is a great part victory, however
12:00 masak I suppose deconstructing subName is no problem
12:00 masak tail should do it
12:00 * masak attempts this
12:01 masak I realize I'm probably not writing the best solution here, and likely implementing the fix on the wrong level, etc etc
12:01 masak but frankly, I don't care
12:01 masak this is fun!
12:03 spinclad (growing new lambdacamels)++
12:03 masak I don't know about that
12:04 masak I still have no innate sense of the difference between `a $ b` and `a (b)`
12:04 masak I basically have to permute between them until I get it right
12:04 spinclad no?  it's a start.  once you've made one patch you can never go back
12:04 wolverian just write foo . bar $ baz and you're fine :)
12:05 rindolf wolverian: what does the "." do in Haskell/
12:05 rindolf ?
12:05 masak wolverian: oh yeah, I forgot `a . b`, I don't grok that either
12:05 wolverian rindolf, composes functions.
12:05 wolverian @type (.)
12:05 wolverian oh, lambdabot is sleeping
12:05 masak yes
12:05 masak for the weekend
12:06 spinclad (a -> b) -> (b -> c) -> (a -> c)
12:06 wolverian masak, it takes the lhs, which must be a function that is missing one parameter whose type is that which the rhs returns, and composes the rhs and lhs together.
12:06 wolverian that's the rough explanation. :)
12:07 wolverian "a function" should be replaced with "a section", I think.
12:07 masak spinclad: still substituting for lambdabot, eh? :)
12:07 spinclad yup
12:07 masak wolverian: I'll have to think about that
12:07 wolverian e.g. map (foo bar) . filter (baz quux) $ list
12:07 masak spinclad: I now pass tests #1 and #2 in subname.t
12:07 spinclad good show
12:07 wolverian (since the 'foo bar' is not composed, it means that bar is an argument to foo)
12:07 masak tests #3 and #4 fail in new interesting ways that make me realise that my fix is naive
12:08 wolverian masak++
12:08 masak :)
12:08 masak however,
12:11 spinclad (bug in spincladbot: the two args to (.) should be switched, i think: (b -> c) -> (a -> b) -> (a -> c) )
12:12 wolverian spinclad, yes.
12:23 * spinclad -> bed &   # sunrise
12:24 masak spinclad++ # hacker rythm
12:30 drrho joined #perl6
12:51 pugs_svnbot r17149 | masak++ | [Code.hs]
12:51 pugs_svnbot r17149 | masak++ | * made subName qualified so that Test.pm reports expected values as intended
12:51 pugs_svnbot r17149 | masak++ | before
12:51 pugs_svnbot diff: http://dev.pugscode.org/changeset/17149
12:52 devogon joined #perl6
12:52 BinGOs joined #perl6
12:52 Caelum joined #perl6
12:52 thoughtpolice joined #perl6
12:52 demq joined #perl6
12:52 Patterner joined #perl6
12:52 pugs_svnbot joined #perl6
12:52 idiotmax joined #perl6
12:52 gaal joined #perl6
12:52 Gothmog_ joined #perl6
12:52 blblack joined #perl6
12:52 args joined #perl6
12:52 sri_ joined #perl6
12:52 rhr joined #perl6
12:52 sysdef joined #perl6
12:52 GeJ joined #perl6
12:52 audreyt_ joined #perl6
12:52 clkao joined #perl6
12:52 takanori joined #perl6
12:52 meppl joined #perl6
12:52 statico joined #perl6
12:52 felipe joined #perl6
12:52 BooK joined #perl6
12:52 SubStack joined #perl6
12:52 JarJarBinks joined #perl6
12:52 theorbtwo joined #perl6
12:52 integral joined #perl6
12:52 Juerd joined #perl6
12:52 wolverian joined #perl6
12:52 PerlJam joined #perl6
12:52 Yappo joined #perl6
12:52 F7 joined #perl6
12:52 Eidolos joined #perl6
12:55 masak oops, confuzzled commit comment
12:55 masak masak-- # thought I removed that last line
13:23 TJCRI joined #perl6
13:24 kanru joined #perl6
13:27 masak @tell spinclad ...
13:27 ruoso joined #perl6
13:27 masak dang
13:28 masak spinclad: changing the other place, in Parser.hs, seems to break a lot of things
13:28 masak so I won't do that
13:29 turrepurre joined #perl6
13:45 xinming joined #perl6
13:48 agentzh joined #perl6
13:49 spinclad masak: good -- since VCode has both subPackage and subName and as long as they stay cleanly split, then op1CodeName is clearly the righter place
13:50 masak yay
13:50 masak now Test.pm works again
13:50 masak try running ext/DateTime/t/basic.t
13:50 masak it now shows expected values again
13:51 masak so, back to the original issue: why do some tests in ext/DateTime/t/basic.t fail?
13:52 spinclad sorry, been talking in my sleep again
13:53 spinclad will look 'tomorrow'
13:53 agentzh yay, auto-smoke works again :)
13:53 agentzh http://perlcabal.org/smoke.html
13:54 agentzh oh, lambdabot has gone off...
13:54 * spinclad &  # really
13:55 agentzh spinclad: sleep well
13:56 polettix joined #perl6
14:00 amnesiac joined #perl6
14:02 agentzh @seen audreyt
14:10 TJCRI joined #perl6
14:13 masak joined #perl6
14:23 daxim_ joined #perl6
14:27 turrepurre joined #perl6
14:28 daxim_ changed the nick to daxim
14:30 sk joined #perl6
14:30 kanru joined #perl6
14:39 buetow joined #perl6
14:43 drrho joined #perl6
14:49 rhr masak: I looked at the ext/DateTime/t/basic.t failures once.  The problem is in pugs' MMD impl, it picks the wrong BUILD when called with no args iirc.
14:50 [particle] joined #perl6
14:53 polettix rhr: it seems to pick the wrong BUILD when called with args as well, at least inside one method
14:56 xinming joined #perl6
15:01 iblechbot joined #perl6
15:09 ruoso joined #perl6
15:10 pasteling "polettix" at 81.174.12.28 pasted "some weirdness about new() and BUILD()" (82 lines, 1.5K) at http://sial.org/pbot/26477
15:11 polettix this is what I mean. Sorry for not putting it as a test, but I don't know how to redirect output to variable quickly.
15:27 rhr yeah, that's pretty strange.  MMD with named args is just broken, I think.  Has the MMD spec been completed yet?
15:37 Eidolos_ joined #perl6
15:39 kcwu joined #perl6
15:49 Eidolos_ changed the nick to Eidolos
15:58 hkb_lab changed the nick to hakobe
16:07 diakopter rhr: it's my understanding that Larry is pretty happy with the MMD spec, until further implementation uncovers more cases.
16:17 jhorwitz joined #perl6
16:20 jhorwitz joined #perl6
16:21 justatheory joined #perl6
16:31 cmarcelo joined #perl6
16:32 iblechbot_ joined #perl6
16:38 polettix joined #perl6
16:42 mr_ank joined #perl6
16:44 IceS joined #perl6
16:46 nothingmuch joined #perl6
16:54 IceSham[a]n joined #perl6
16:59 IceSham[a]n changed the nick to IceShaman
17:02 Psyche^ joined #perl6
17:02 justatheory joined #perl6
17:14 kaether joined #perl6
17:18 Psyche^ changed the nick to Patterner
17:26 chris2 joined #perl6
17:30 dduncan joined #perl6
17:39 rindolf joined #perl6
17:49 solar345 joined #perl6
17:49 solar345 hi all !
17:50 diakopter solar345: hi
17:54 rindolf Hi solar345, diakopter
17:54 rindolf Hi all.
17:57 solar345 hi diakopter
17:57 solar345 hi rindolf
18:01 buubot joined #perl6
18:09 franck____ joined #perl6
18:18 weinig joined #perl6
18:24 iblechbot joined #perl6
18:32 franck___ joined #perl6
18:33 franck___ joined #perl6
18:37 jisom joined #perl6
18:50 the_dormant joined #perl6
18:51 Jmax joined #perl6
19:27 rindolf What does [+] [6,7,8] supposed to do in p6?
19:27 PerlJam rindolf: 21
19:28 PerlJam (6 + 7 + 8)
19:28 rindolf PerlJam:
19:28 rindolf pugs> [+] [6,7,8]
19:28 rindolf (6, 7, 8)
19:28 rindolf pugs> [+] (6,7,8)
19:28 rindolf 21
19:28 rindolf pugs on feather
19:28 PerlJam oh.  I didn't even notice the brackets  Hmm.
19:29 rindolf Where is the pugs bug tracker?
19:31 rindolf pugs> [5,6,[10,20]]
19:31 rindolf (5, 6, (10, 20))
19:31 rindolf What sort of notation is this?
19:31 rindolf With the parentheses?
19:31 rindolf Instead of the sq brackets?
19:34 TimToady should return 3
19:35 TimToady or maybe it should just return the single arg
19:35 TimToady looks like that's kinda what pugs does
19:36 miyagawa joined #perl6
19:36 PerlJam TimToady: and what would  [+] [1,2,3], [4,5,6]  give us ?
19:36 TimToady same as [1,2,3] + [4,5,6]
19:36 TimToady that is, 6
19:37 rindolf perlbot: pugs bugs
19:37 PerlJam TimToady: then I think [+] [6,7,8] should return 3
19:38 TimToady depending on whether the numeric context is implied for a single arg
19:39 franck____ joined #perl6
19:40 TimToady for an operator x that has an identity value returned by [x]() we could define [x]($value) to be [x]() x $value perhaps
19:42 rindolf Hi TimToady
19:43 TimToady howdy
19:43 rindolf I want to try and fix a pugs bug.
19:43 rindolf Where should I go to look for one?
19:43 rindolf I mean a list.
19:43 rindolf TimToady: what's up?
19:44 TimToady http://m19s28.vlinux.de/iblech/stuff/pugs-s​mokes/pugs-smoke-6.2.13-r17147-linux-normal​--1185491801-9496--18817-17553-1264-1965-74​2-46--2a438ce684ce329a6200f7e4e6620cfe.html
19:46 TimToady just wrapping up OSCON here
19:53 rindolf TimToady: nice.
20:09 PerlJam TimToady: Did you give a SotO?
20:11 ashleyb joined #perl6
20:11 Eidolos joined #perl6
20:16 TimToady something like that
20:18 PerlJam I hope it went well.
20:19 dduncan as do I
20:25 weinig joined #perl6
20:32 rlpeacock joined #perl6
20:35 TimToady OSCON over, massive decommute &
20:42 dduncan left #perl6
20:53 Blwood joined #perl6
20:57 weinig_ joined #perl6
21:05 ashleyb joined #perl6
21:19 meppl good night
21:25 weinig joined #perl6
21:34 cognominal_ joined #perl6
21:40 cj joined #perl6
21:46 nperez joined #perl6
21:50 weinig_ joined #perl6
21:52 weinig joined #perl6
21:55 dmq joined #perl6
22:12 cj joined #perl6
22:54 szbalint joined #perl6
23:10 weinig changed the nick to jibot
23:10 jibot changed the nick to weinig
23:16 weinig_ joined #perl6
23:17 weinig joined #perl6
23:30 devogon joined #perl6
23:51 devogon joined #perl6
23:56 justatheory joined #perl6

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

Perl 6 | Reference Documentation | Rakudo