Please test pre0: pugs.blogs.com/pre0/Perl6-Pugs-6.2.12.tar.gz | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org
Set by audreyt on 12 June 2006.
00:16 ajs_home joined 00:24 lisppaste3 joined
christopher fglock/wolverian: Why is the CPAN version of Pugs::Compiler::Rule fresher than that in the pugs tree? To whom should I send tests or patches, if not to the pugs tree? 00:25
00:27 hcarty joined
trym does anyone know how to pass a maxmessage paramter to RPC::PlClient or ::PlServer when using DBI::ProxyServer (or dbd::Proxy) 00:43
00:45 rlb3 joined 00:48 hikozaemon joined
trym ah wrong chan 00:49
sorry
pasteling "TimToady" at 64.0.7.8 pasted "really minimal basic.t version" (1 line, 190B) at sial.org/pbot/17863 00:51
TimToady audreyt: the previous is the result of running a minimizer that scans for the same assertion failure. 00:52
01:01 hcarty left 01:11 weinig|dinner is now known as weinig 01:13 buetow joined 01:16 christopher joined 01:26 mako132_ joined 01:31 Gnurdux joined, Gnurdux left 01:38 shachaf joined 01:41 weinig is now known as weinig|sleep
dduncan this is off-topic, but can anyone recommend DVD-R brands to use or stay away from? ... looking online shows a wide range of opinions for any particular brand 01:53
02:00 mjk joined
dduncan or never mind 02:33
02:57 shachaf joined 03:33 pdcawley joined, _SamB_ joined 03:41 justatheory joined 03:46 pdcawley joined 04:38 dduncan left
svnbot6 r10753 | audreyt++ | * perl2.t: TODO a couple of regex .perl tests. 04:38
04:43 pdcawley joined 04:50 pdcawley joined 04:52 vel joined 04:57 pdcawley joined
gaal so, emit_yaml :) 04:58
05:03 pdcawley joined 05:13 pdcawley joined
ingy hi gaal 05:23
audreyt has been safely delivered to her Seattle hotel 05:24
(in Bellevue :p)
05:26 cmarcelo left
svnbot6 r10754 | audreyt++ | * reduce-metaop.t - unTODO one surprisingly succeeding 05:26
r10754 | audreyt++ | test involving [=]. That's all for tonight...
TimToady yay. 05:29
audreyt gaal, TimToady: sorry, sleep has priority :)
TimToady np
take care of yourself.
audreyt (and then wakeup, and then write some 100 more slide, and then deliver this 90min talk at MS)
audreyt is driving the FPS up
(frames per second, in this context)
TimToady good luck 05:30
audreyt thanks, I'll need it :)
*wave* *thud* &
gaal heheh, at some point her slides will achieve persistence of vision 06:00
TimToady if she's gonna be an animator, she'll need to hire some 'tweeners... 06:06
06:09 amnesiac joined 06:16 iblechbot joined 06:18 vel joined
wolverian how does one control the AST creation in grammars/rules? <foo> { return $<foo> as Foo }, to get the type I want? 06:18
makes me want to add return types to rules 06:21
I'm trying to port j.mongers.org/pub/haskell/darcs/con...BEncode.hs 06:23
in short, I want to construct the data structure directly with the grammar, or hand the match object to a constructor, whichever is saner 06:24
or some other way, I'm kind of still learning this... :) 06:25
06:28 pdcawley joined 06:33 lumi joined
svnbot6 r10755 | audreyt++ | * INSTALL: Bump our parrot requirement to 0.4.5 for the shiny 06:35
r10755 | audreyt++ | regex/token/rule :ratchet/:sigspace support.
06:48 Aankhen`` joined 07:27 marmic joined 08:05 pdcawley joined 08:09 xerox joined 08:10 xerox joined 08:37 iblechbot joined 08:40 nothingmuch joined 09:04 lisppaste3 joined 09:10 buetow joined 09:18 SM_ax joined 09:20 bjoern_ joined 09:51 SM_ax joined
audreyt wakes up and resumes slidemaking 10:03
gaal is slidemaking for yapc 10:18
that's days ahead 10:19
which only makes it harder
10:24 arguile joined 10:37 chris2 joined 11:05 ludan joined 11:23 jsiracusa joined
Aankhen`` Yay, managed to build Pugs. 11:31
(After manually updating Module::Install in perl5/Data-Bind and perl5/Module-Compile.)
11:32 iood joined, christopher joined
Aankhen`` *** Error: no ICU lib loaded 11:40
What does that mean in the context of a rule that PGE didn't understand?
cognominal_ sounds like the ICU library, the IBM library to handle Unicode. is not loaded 11:43
nothingmuch ships a module with compiled .pmcs
and no filter in deps
*pray*
Aankhen`` Ah.
Hiya nothingmuch. :-D
nothingmuch hola
Aankhen`` I'd ask what's up, but that seems to have been answered already. 11:44
In lieu of that, let me ask if your refrigerator is running.
nothingmuch in that case, i'll ask what's up
yes, it is =)
thanks for asking
Aankhen`` Then you better catch it!
Aankhen`` ducks and runs.
nothingmuch it's a freerange refrigerator
we let it eat and poop in the yard 11:45
Aankhen`` Sorry, that's a really lame joke. :-P
Heh.
nothingmuch and it gives us organic coldness
M::C note:
Aankhen`` I remember when I was in Class II or III... my friends and I thought it was hilarious to call up people and pull this gag on them.
11:45 ko1_away is now known as ko1
nothingmuch if you use Module::Build etc you must touch the .pmc 11:45
so that perl will be nice to it
if they are the same mtime it will go to the .pm 11:46
Aankhen`` I never did get to pull it myself though. :-\
nothingmuch my best remembered phone prank is in the 8th grade
Aankhen`` Do tell!
nothingmuch we called our loathed literature teacher on a holiday
and told her that she had one a "Wella Cosmetics Holiday Package" for elderly women 11:47
11:47 weinig|sleep is now known as weinig|work
nothingmuch we told her that we were in the marketting dept, and that she needs to contact customer relations with a voucher code to get the package shipped 11:47
s/one/won/;
and the number we gave her was to some sex line
Aankhen`` LMAO. 11:48
ko1 i have a question about parrot
nothingmuch was grand orchestrator
friend #1 was actor
Aankhen`` Heh.
nothingmuch and friend #2 provided an el-cheapo magazine with some sex ads at the ened
meppl guten morgen 11:49
nothingmuch ko1: if it's a very technical question then maybe #parrot on irc.perl.org can help
QtPlatypus ko1: What is the question?
nothingmuch but in general we should also know
ko1 parrot's register number is infinite?
nothingmuch hola meppl
ko1 nothingmuch: ah, thanks
meppl good morning nothingmuch
ko1 at ver. 0.2 or so, i can only 32 registers on parrot. but now, i can use 500 registers (i0 ... i500) in pasm script 11:51
nothingmuch ko1: at some point parrot got a variable register allocator, i think 11:54
you had to predeclare the number
and it would compute a static call frame based on that
PIR otoh has a register spilling thingy that is less "hard coded" 11:55
that's background info
as fcor your question:
it's not really infinite
but it pretends to be 11:56
ko1 nothingmuch: i see
nothingmuch that said, using a very big number is probably not as efficient
because parrot tries to map registers to hardware registeres
urf
so try to keep your lexical scopes slim at the pasm level
if you want better support for infinite stuff, use PIR 11:57
does that answer your question/ 11:58
?
ko1 www.atdot.net/sp/view/zkq51j/ (pir) works :)
nothingmuch: yes! thanks 11:59
nothingmuch search audrey's blog for it 12:00
she mentioned working out with leo on lexical scopes
and how the calling semantics and register decls/limits were changed to better support Perl 6
ko1 i only want to know that "register number is logically infinite or not" 12:01
i'm writing parrot introduction article in Japanese :) 12:02
nothingmuch well 12:03
that article should be shared with #parrot
so you still have to go there ;-)
good luck, and ko1++
ko1 nothingmuch: ok. thanks 12:04
QtPlatypus watches ko1 mutate into ko2
12:05 bjoern_ is now known as DEpwnsEC 12:17 ko1 is now known as ko1_away 12:23 frederico joined 12:30 visq joined
gaal visq: hi 12:39
what platform are you on? I'm not getting color in the .html files, and I'm wondering if it's some silly path problem. 12:41
12:43 hugues23 joined 12:45 jsiracusa joined
gaal hmm, yes! it works on linux but fails on windows. checking. 12:52
Juerd webmind: What do you need root access for on feather? 12:56
13:01 buetow joined 13:07 amnesiac joined, _SamB_ is now known as SamB 13:23 iblechbot joined 13:27 amnesiac is now known as amnes1ac 13:28 amnes1ac is now known as amnesiac 13:40 renormalist joined 13:41 fglock joined 13:42 justatheory joined
fglock my first blog ever - use.perl.org/~fglock/journal/ 13:45
kolibrie fglock: short and sweet 13:47
svnbot6 r10756 | fglock++ | Pugs::Grammar::Perl6 - added @{exp}, exp[], s///, 13:50
r10756 | fglock++ | - added modules Pugs::Compiler::Perl6, Pugs::Emitter::Perl6::Perl5, Pugs::Runtime::Perl6
r10756 | fglock++ | - added stub module: v6-pugs
r10756 | fglock++ | Pugs::Compiler::Precedence - fixed postcircumfix to allow an empty list
fglock kolibrie: I'm not much into writing (human language)
svnbot6 r10757 | fglock++ | renamed Pugs-Grammar-Perl6 to Pugs-Compiler-Perl6 13:53
Juerd fglock: That comes naturally when you've been doing it a while :) 13:55
fglock I hope so :)
svnbot6 r10758 | fglock++ | Pugs-Compiler-Perl6 - fixed test.pl 13:56
Juerd It's just like IRC, but without the immediate response. 13:57
fglock but it is also a different public 13:59
14:01 arguile joined
Juerd Ignore that :) 14:02
Juerd tries to ignore the audience
Otherwise I'm constantly distracted
And start to question if I phrased things right
14:10 rzoz joined 14:20 rlb3 joined
audreyt ?eval [+] 1..100 14:26
svnbot6 r10760 | fglock++ | Pugs-Compiler-Perl6 - updated t/02-expression.t
evalbot_10746 5050
audreyt ?eval { $_ ?? $_ * &?BLOCK($_-1) !! 1 }.(10)
evalbot_10746 3628800
audreyt @djinn (a -> b) -> (c -> b) -> Either a c -> b 14:30
lambdabot f a b c =
case c of
Left d -> a d
Right e -> b e
audreyt @pl f x = x
lambdabot f = id
audreyt @pl @pl f h = do { x <- hGetContents h; return (lines x) } 14:32
lambdabot (line 1, column 1):
unexpected "@"
expecting white space, natural, identifier, lambda abstraction or expression
audreyt @pl f h = do { x <- hGetContents h; return (lines x) }
lambdabot (line 1, column 10):
unexpected "{"
expecting variable, "(", operator or end of input
audreyt @pl @pl f h = hGetContents h >>= \x -> return (lines x)
lambdabot (line 1, column 1):
unexpected "@"
expecting white space, natural, identifier, lambda abstraction or expression
audreyt @pl f h = hGetContents h >>= \x -> return (lines x)
lambdabot f = (lines `fmap`) . hGetContents
audreyt (sorry for the spam, doing screenshots for $talk)
(done now)
14:51 FurnaceBoy joined
fglock audreyt: how about - would you present the haskell talk in YAPC::SA? We could arrange for it to be presented just before cmarcelo's talk on his SoC work 14:51
lunch & 14:53
gaal hey audreyt, if you got a minute, how could (yaml ::String) <- liftIO $ toYamlNode something work? 15:00
15:02 vel joined
xerox Hey gaal, how're arrows? (: 15:03
gaal hey xerox! dangerous as intended
xerox Special relativity is more dangerous! (My mind is boggling.) 15:04
gaal General relativity's a lot worse!
xerox Just don't accelerate, for the sake of my mental stability, please. 15:05
gaal xerox: maybe you can help with my typing problem? 15:07
take a look at Pugs.CodeGen.YAML
genParseYAML
lumi Can I help too?
gaal the question is how could the line with liftIO typecheck
I hope so :) 15:08
lumi It does typecheck, and you don't know why?
xerox Do you have a url? I can take a look but I really gotta study for tomorrow exam. And the day after, and the one after that one too.
gaal svn.openfounrdy.org/pugs/src/Pugs/C...en/YAML.hs
xerox What's toYamlNode type?
xerox goes to see
gaal toYamlNode :: YAML a => a -> IO YamlNode 15:09
data YamlNode .... deriving (Show, Ord, Eq, Typeable, Data)
xerox liftIO :: (MonadIO m) => IO a -> m a
gaal so maybe it's a SYB thing I don' understand?
sure, but look at the next line 15:10
it's VStr yaml
xerox It's return (VStr yaml)
gaal VStr :: String -> Val
xerox return (VStr yaml) :: (Monad m) => m Val
gaal so how can that work?
xerox return (VStr yaml) :: Eval Val -- In that context.
return is NOT a keyword, it is a function! 15:11
return :: (Monad m) => a -> m a
gaal sure, forget the monad thogh
the expression (VStr yaml)
xerox What is showYaml type?
gaal implies that yaml :: String
xerox ... -> String, I would say, since it is called show*
gaal showYaml :: YAML a => a -> IO String
which is more sane
xerox It is very straighforward, do you know what <- does in a do block? 15:12
gaal so I fully understand the function genParseYAML, but not genParseHsYAML
xerox Ah. 15:13
yaml is a YamlNode in the former function.
yaml is a String in the latter one.
Thanks to, respectively, toYamlNode, and showYaml.
gaal indeed, and I don't see how it could be anything else in the first one. 15:14
xerox So?
gaal so how can it typecheck?
15:14 Southen_ joined
xerox show . toYamlNode = showYaml 15:14
More or less.
showYaml = fmap show . toYamlNode 15:15
lumi Aren't they the same?
xerox lumi, no, because VStr takes a string, and in the former function the `yaml' binding is passed to `show', you see?
gaal aaaahhhhh it's show . toYamlNode!
lumi Yeah.. but show is pure 15:16
gaal looks for a nearby wall
yes yes yes
I didn't see the "show" :)
xerox toYamlNode :: YAML a => a -> IO YamlNode
toYamlNode x = runReaderT (asYAML x) IntSet.empty
showYaml :: YAML a => a -> IO String
showYaml x = do
node <- toYamlNode x
emitYaml node
gaal I'm stupid, no more questions
xerox Hmmm. 15:17
Let me show you a thing, now.
gaal please show me a thing.
xerox video.google.com/videoplay?docid=34...6217914779 15:18
gaal nice clip 15:23
15:30 Southen_ is now known as Southen
gaal why does make unoptimized ; make ghci need to recompile a lot of .hs? 15:36
15:39 visq joined
visq gaal: you were talking about highlighting in the htmlified tests ? 15:58
Works for me Windows/IE, Linux/Firefox, MacosX/Firefox
gaal visq: it's generating everything, just not coloring lines in the tests html files. 15:59
visq Did you look at the page source ?
Should have three inline CSS definitions
And <span class=test_status>... for highlighting 16:00
gaal sure, but all the tests are classified as non_test.
visq I see
16:01 pdcawley joined
gaal :g/ pos:/s,\\,/,g on the yaml file fixes this. 16:01
but I couldn't find the erroneous line in the code yet.
svnbot6 r10762 | gaal++ | * refactor Pugs.CodeGen.YAML a bit
r10762 | gaal++ | * introduce filterPrim, which is supposed to weed out things like handles
r10762 | gaal++ | in AST dumps
r10762 | gaal++ | * unfortunately, it doesn't in Pugs::Internals::emit_yaml (though it does
r10762 | gaal++ | work in -CParse::YAML).
gaal afk& 16:03
16:06 penk joined 16:09 marmic joined 16:12 lollan joined
visq util/catalog_tests.pl:283 : ($start, $end) = $test->{pos} =~ /line (\d+)(?:.*line (\d+))?/s; 16:13
maybe 16:14
16:14 gaal joined 16:21 buetow joined 16:31 bpederse left 16:36 bernhard joined
visq gaal: I found 2 places where '/' was used as a to split resp. join path(s) and rewrote them 16:38
svnbot6 r10763 | benedikth++ | util/catalog_tests.pl: Path delimiter issues 16:41
16:42 bpederse joined, lumi joined 16:44 bpederse left 16:51 xerox joined
gaal visq: nope, they're still miscataloged 16:53
visq++ for working on this though
16:54 shachaf joined
svnbot6 r10764 | fglock++ | Pugs-Compiler-Perl6 - added hello_world.pl (parses ok) 16:59
16:59 Ymmv joined, frederico joined 17:05 neoesque joined 17:11 myavuzselim joined 17:18 ruoso joined 17:19 nxu7 joined 17:22 DEpwnsEC is now known as bjoern_ 17:35 lumi joined 17:39 vel joined 17:41 chris2 joined 17:43 rlb3 joined
cognominal_ the class does not need to declare has $.foo as a property # ... in syn12, s/property/attribute/ ? 18:01
svnbot6 r10766 | fglock++ | Pugs-Compiler-Perl6 - "hello, world" works 18:14
18:16 justatheory joined
svnbot6 r10768 | fglock++ | Pugs-Compiler-Perl6 - v6-pugs renamed to P6-pugs (perl 5 ignores 'v6') 18:35
r10769 | fglock++ | Pugs-Compiler-Perl6 - renamed again: P6-pugs.pm to P6.pm 18:38
TimToady fglock: in r10756 you commented, "added @{exp}", but @{exp} is not legal syntax any more. That's @(exp) these days. 18:41
fglock TimToady: fixing - thanks
18:45 hexmode joined
cognominal_ TimToady, am I confused about the s/property/attribute/ in syn12? 18:46
18:56 frobnitz joined
TimToady cognominal_: no, I checked in a fix for it about 20 minutes ago. 19:08
svnbot6 r10770 | fglock++ | Pugs-Compiler-Perl6 - 'use P6' creates a .pmc and exec() it
fglock "proof-of-concept" - misc/pX/Common/Pugs-Compiler-Perl6$ perl -Ilib hello_world.pl 19:09
(ignore warnings)
19:19 chris2 joined 19:20 stennie joined 19:21 stennie left, pmurias joined
pmurias hi 19:21
is getting all revisions with svk acceptable?
fglock pmurias: hi 19:22
svnbot6 r10772 | fglock++ | Pugs-Compiler-Perl6 - recompile .pmc only if needed
fglock re svk: I think it is ok, but it takes a lot of time 19:23
pmurias i'll eat dinner and propably just the last 1000 or so 19:26
fglock P6.pm now has very low overhead for precompiled .pl files 19:30
svnbot6 r10773 | fglock++ | P6.pm - much faster cache access
19:36 chris2_ joined 19:42 larsen joined 19:55 prefiks joined 20:01 kane_ joined 20:09 amnesiac joined
sahadev is there "hyper" version of the unary operator ! 20:12
?
i wanna toggle (boolean-wise) all the elements in a list
xerox map not ? 20:19
sahadev map works, of course. but i wanted to see if i could use the hyperoperator. 20:20
it seems to work: !<< @a # returns the list of toggled values
xerox @a >>! # too ? 20:21
lambdabot Maybe you meant: all-dicts arr ask . v
sahadev no
pugs> @a >>!
Internal error while running expression:
***
unexpected ">"
expecting comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
xerox Aw. 20:22
>>! @a # ? :)
sahadev "expecting program"
gaal well, it's a prefix operator, so why should >>! work?
xerox Dunno.
gaal ?eval perl !<< <1 0 0 1> 20:23
works.
evalbot_10746 "(Bool::False, Bool::True, Bool::True, Bool::False)"
20:23 larsen_ joined 20:24 iood joined 20:26 Bleadof joined
gaal on the other hand: 20:27
?eval sub postfix:<!> ($n) { [*] 1 .. $n } say <1 2 0 3> >>!
evalbot_10746 OUTPUT[126 ] Bool::True
gaal (please exuse the lousy factorial)
s/x/xc/ 20:28
sahadev nice!
so, if you define any new list operators, the hyper operator version comes free? 20:29
s/list //
gaal sure, because they aren't hardcoded, they're a metaoperator. 20:30
sahadev how come this gives me an error? 20:31
sub postfix:<!> ($n) { $n ?? [*] 1 .. $n :: 1 }
gaal "!!" instead of "::"
sahadev ternary ?: is spelled ?? !! now?
gaal When laziness is fixed, try this out:
my @facs := [\*] 1..*; sub fac ($n) { @facs[$n] }; say fac(4)'
yes, for a while now 20:32
s/'$//
the above is a port of one of the nice Evolution of a Haskell Programmer functions
sahadev cool!
gaal [\*] is a product scan, but here too you can put any user defined op 20:33
ingy seen audreyt 20:39
jabbot ingy: audreyt was seen 6 hours 6 minutes 34 seconds ago
gaal hmm:
seen jabbot
jabbot gaal: jabbot was seen 373 days 3 hours 52 minutes 34 seconds ago
gaal heh.
this guy must thrive near the Ravenous Bugblatter Beast of Traal 20:40
more packing & 20:41
sahadev S29 says: multi Int|List Array::keys ( @array : MatchTest *@indextests ) 20:50
how can i use the *@indextests parameter here?
?eval my @a = 1..5; @a.keys { $_ % 2 } 20:52
evalbot_10746 Error: unexpected "{" expecting comment, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
TimToady sahadev: that's not implemented. (And even if it was implemented, that syntax doesn't work. You'd need @a.keys({ $_ % 2}) or @a.keys:{ $_ % 2 } or some such.) 21:28
.foo by itself is assumed to be argumentless. 21:29
sahadev TimToady: thanks for the clarification. when it is implemented though, what would this return? 21:33
my @a = (1..3); for @a.kv: { $_ % 2 } -> $k, $v { say "$k: $v" }
"1: 2" ? 21:34
or, "0: 2" ?
PerlJam What's $_ bound to there? 21:35
sahadev oops. s/$_/$k/ 21:36
hmm. i dunno actually :)
in the first example: @a.keys { $_ % 2 } it's obvious
PerlJam sahadev: well, there's an "obvious" interpretation with .kv too, $_ would be bound to the keys ;-) 21:40
sahadev and, if I want to test the "value" in my block, what would I use? 21:41
the test block, i mean
21:46 FurnaceBoy joined
TimToady I tend to think that the list passed as the selector would be multidimensional, and if there was an extra dimension on the end, it would be applied to the values. 21:50
pmurias good night 21:53
cognominal_ does the Match object contains the name of the matched rule? I can't see anything like that in PGE::Match 21:56
22:00 fglock_ joined 22:02 jsiracusa joined
ingy cognominal_: ping 22:04
or TimToady 22:05
or anyone here :P
fglock_ ingy: pong 22:06
cognominal_ pong
ingy fglock_: lemme test msg you
fglock_: did you get my msg?
fglock_ not yet 22:07
ingy sigh
nothingmuch ingy: how do i see MC output of a .t file? 22:08
rgs ingy: you should be a registered user to msg people on freenode
nothingmuch btw, i released a module with some .pmcs in it
and no deps for the filter or MC
wish me luck ;-)
search.cpan.org/dist/Cache-Cascade/
ingy rgs: I know :(
rgs and /whois tells me you aren't
ingy I know :( 22:09
fglock_ nothingmuch: it should work without problems - Pugs::Compiler::Rule in CPAN depends on .pmc too
nothingmuch: did you see Pugs::Compiler::Perl6?
ingy rgs: nickserv says I'm already registered
nothingmuch fglock_: yes, but there's lots more to screw up ;-)
rgs nothingmuch: interesting module, I suppose that might be usable in large deployments of catalyst things
nothingmuch (than a normal dist) 22:10
rgs: that's precisely the niche market
nothingmuch is rewriting Catalyst::Plugin::Cache
rgs nothingmuch: but why pmcs on that ?
nothingmuch rgs: use tt
rgs OMG it's intt
nothingmuch hehe
rgs [% FOREACH method IN methods %]
you are a sick, sick man !! :)) 22:11
nothingmuch wow, you called me a man
most people stick to "boy" ;-)
i think i will add 'use tt for => qw/[ ]/' which will automatically do the [% FOREACH %] thing for you 22:12
rgs is there some tt frontend to M::I ?
nothingmuch all my usage so far has been loopy
i don't understand the question
fglock_ nothingmuch: I'm interested in making .pmc at use-time - not sure how to do that yet
nothingmuch fglock_: explain?
rgs Module::Compile, I mean
nothingmuch search.cpan.org/perldoc?tt 22:13
fglock_ you deploy a perl6 module that 'use P6', and it compiles to p5 and then run - no pre-compilation, if possible
22:13 weinig|work is now known as weinig
fglock_ but it gets cached 22:13
nothingmuch fglock_: doesn't it work like that anyway?
rgs nothingmuch: ok
ingy nothingmuch: what was it you wanted?
nothingmuch ingy: i have a 'use tt' in a .t file 22:14
ingy ok
nothingmuch and I have a syntax error
i want to view the filtered code
ingy ok
lemme check
nothingmuch <3
fglock_ nothingmuch: not yet, it calls the p6 version first, and compiles it - but it doesn't transfer control to the compiled version yet
nothingmuch ohhhh 22:15
i dont think that's possible for the lines up to the 'use p6'
but it should be identical, shouldn't it? 22:16
or does the p6 compiler twiddle the .pmc file post factum?
ingy nothingmuch: subclass pmc_template
XXX @_;
or somesuch dump
nothingmuch !<3
rgs nothingmuch: that said, to generate methods, p5 does it quite well already. But to generate tailored code that would be more useful
like, lib/lib_pm.PL in the core perl distor
distro
ingy dist0rz!! 22:17
nothingmuch rgs: perlmonks.org/?node_id=556341
fglock_ maybe require'ing the .pmc from P6.pm works - I'll try
bbiab &
nothingmuch i find TT more easy to read structurally than BEGIN { eval/closure + glob assign }
22:18 jiing joined
rgs because you're used to it 22:18
nothingmuch yeah
well
i'm not more used to it
i've been doing pretty sick BEGIN { eval + glob assign } for a while now
and i was like "hmm... this should stop
rgs actually to generate test suites that could be extremely neat
ingy :)
nothingmuch because if i want code that i can easily explain to my army of minions but that doesn't DRY i suspect this will be easier 22:19
rgs wonders whether there is an rss feed for nothingmuch's cpan uploads
ingy rgs: there is
rgs woot !!
ingy ask miyagawa
nothingmuch <dc:creator>Michael R. Davis</dc:creator> 22:20
ingy there is for any author or combo of authors
nothingmuch you can plagger + filter
miyagawa > hand_rolled
you can view an rss feeds on the bugs for my modules ;-) 22:21
ingy I saw the demo of this at yapc::asia
nothingmuch uh, i think
wolverian plagger is GREAT.
ingy it wasn't plagger iirc 22:22
just a feed
wolverian a cpan.org feed?
ingy looking for it.... 22:23
wolverian there's the overall feed, at least
I don't see per-user feeds 22:24
i.e. plagger to the rescue :)
22:24 hcchien joined
ingy wolverian: yeah not that one 22:25
grrr
hi hcchien!
nothingmuch: did that work? 22:26
nothingmuch ingy: no, because i didn't try it 22:27
a temp mv t/foo.t t/foo.pm is simpler
=)
ingy I'll make a note about making it easy to find errors 22:28
nothingmuch: where is your test code? 22:29
I want to see what it looks like :)
nothingmuch not yet commitited to public facing repo 22:32
i'll /msg you when it is
ughghgghghghghghgg
nothingmuch doesn't understand how comments can make a syntax error
oi vey 22:33
perl -c foo.pmc *does* pass
nopaste time
ingy :)
pasteling "nothingmuch" at 84.108.53.74 pasted "output from perl -c" (6 lines, 338B) at sial.org/pbot/17883
ingy looks 22:34
pasteling "nothingmuch" at 84.108.53.74 pasted "troublesome block" (12 lines, 184B) at sial.org/pbot/17884
nothingmuch without the #moose it passes
ingy can I check this out from somewhere? 22:35
nothingmuch darcs get nothingmuch.woobling.org/Class-Workflow for the code
but uh
i'm still debugging this test
it's not supposed to work yet
Class-Workflow->state != $ready
i can commit a partial patch 22:36
and then unpull
but that == conflicts when i commit the real one
and you'l have to unpull it too
ingy can't you just delete the #moose line 22:37
and check it in?
and I'll add it back
nothingmuch ingy: no
because there's supposed to be actual code there
pasteling "nothingmuch" at 84.108.53.74 pasted "also fails, what i really want" (15 lines, 334B) at sial.org/pbot/17885
ingy can you just nopaste the whole test, and let me checkout the rest? 22:39
nothingmuch i did
huh
must have done it to another channel
pasteling "nothingmuch" at 84.108.53.74 pasted "whole .t" (268 lines, 4.4K) at sial.org/pbot/17886
ingy ok 22:40
what's the file called?
nothingmuch t/example_pop.t 22:41
t/example_pop3.t
darpull 22:42
darcs pull
i just checked in all my working stuff
i can also put a tarball of my working dir
but my webserver doesn't implement garbage collection, so i don't like doing that often
ingy so... 22:43
how do I run these tests
ok.pm? 22:44
nothingmuch osrry, back
err, like normal tests
ingy prove -l t
nothingmuch yep 22:45
thanks, i always used prove -Ilib
i guess that'll change now =)
ingy can't believe he's installing Moose ;)
nothingmuch oh wait
you need trunk moose
this code ironed out some bugs in the role composition stuff
ingy url? 22:46
nothingmuch in /msg
it's not a 100% public repo yet
the sysadmin is slacking off ;-)
ingy um Module::Install! 22:48
nothingmuch where?
ingy doesn't even have Module::Build installed
it's not in Moose
nothingmuch blame stevan
personally i dislike the inc stuff
that's why i don't use it for my code 22:49
and it's got a bad rep because people use auto install needlessly
ingy get off my motorcycle
do I need a custom Class::MOP? 22:50
nothingmuch dunnno
i have them both setup in my devlib alias
ingy cpan is 0.26
it wants 0.29_01
your custom Moose does... 22:51
nothingmuch i think 029_02 is on the CPAN
ingy Isee
nothingmuch install with distpath
but if you like
Moose/trunk/../../Class-MOP/trunk
fglock is someone against removing perl5/Pugs-Compiler-Rule? it is no longer the devel version 22:54
ingy you need a Moose dep on Test-LongString 22:55
nothingmuch svk up -s 22:57
svnbot6 r10775 | fglock++ | Pugs-Compiler-Perl6 - directory cleanup
22:59 fglock left
svnbot6 r10776 | fglock++ | perl5/Pugs-Compiler-Rule is gone - use misc/pX/Common version instead 22:59
ingy nothingmuch: still in dep hell 23:02
nothingmuch: somebody needs Set::Object
nothingmuch yes, Class::Workflow's deps aren't yet in a build tool file 23:03
ingy but I guess you don't have a Makefile.PL yet
use Module::Install
nothingmuch can i use it without bundling in "inc" ?
ingy pansy
NOOOOOOOOOOOOOOO 23:04
nothingmuch it's simple actually
oh, not it isn't
you can't inject the deps later
can you force CPAN.pm to rerun makefile.pl?
ingy is tt.pm on cpan? 23:05
is 0.01 ok?
nothingmuch yes
btw, the .pmc doesn't delete itself if the .pm doesn't use MC anymore 23:06
23:06 justatheory joined
ingy hi David 23:07
svnbot6 r10777 | fglock++ | Pugs-Compiler-Perl6 - fixed hello_world.pl comment 23:14
ingy nothingmuch: how does the test fail for you? 23:19
nothingmuch perl -c fails
the test doesn't do anything yet
ingy ok, wtf was whole .t then? 23:20
nothingmuch setting up a test case
that is also a complete example
ingy - darcs pull - this is a working version with no source filtering 23:22
ingy nothingmuch: rafb.net/paste/results/7Qosd630.html 23:23
buubot The paste 7Qosd630 has been moved to erxz.com/pb/1637
ingy seems to work fine here
except for the warnings 23:24
am I missing something?
nothingmuch i dunno 23:25
maybe you have a more recent MC?
ingy no
nothingmuch oh wait sorry
ingy hold on
nothingmuch perl -c t/example_pop3.t
Workflow.pm is OK
ingy ok
why are you compiling a script? 23:26
you can only compile a module 23:27
example_pop3 looks like a script
nothingmuch: are you peeing or combing your hair or something 23:28
(peeing *is* a euphamism)
nothingmuch urf
you weren't listening all along
ingy oh 23:29
nothingmuch i moved it from .t to .pm to get the compiled output
i can't cache
but i can filter
it was only temporary anyway
but...
ingy well I moved it too :p
nothingmuch when I do that it generates a weird syntax error when perl -c t/example_pop3.pm
but not when perl -c t/example_pop3.pmc
ingy still, it seems like you were going down a weird path
but I'll keep debugging 23:30
don't get uppity
nothingmuch uppity?
ingy don't get your panties in a bunch 23:31
or your knickers in a twist, etc
nothingmuch well, i'm no longer your bitch so i don't have to be nice
ingy you're still *a* bitch 23:32
you're ingy privledge is merely on suspension
so be nice please
nothingmuch bitches are, as a general rule not nice at all 23:33
except to their owner ;-)
ingy noted
ok, this at least points out a bug in M::C 23:34
I need to do 3 hours of stuffs
nothingmuch have fun
the bug is not urgent
ingy I will fix this promptly following
nothingmuch if it's important to you take care of it
ingy it is now bitch!
nothingmuch yessir
ingy :) 23:35
nothingmuch anyway
ingy have fun too
nothingmuch i'm going to work out so that i'm nice and lean for you
wow i love darcs dependency graphs
so purrrrdy
ingy I'm on a diet too, so you'll need to work reeeeally hard
what are you doing after yapc, btw?
nothingmuch hackathon, obrathon, hike-a-thon, workathon 23:36
ingy is everyone going to Boston?
nothingmuch i know of audreyt, stevan & $?SELF
ingy wasn't invited. *sobs*
nothingmuch yeah, it's just for cool folks 23:37
you're too old to be cool
ingy ouch
TimToady OUCH!
ingy hit em where it hurts
nothingmuch oh, and gaal is also coming
TimToady I have too many obligations to take on any new obragations... 23:38
nothingmuch =( 23:39
btw, i wasn't really invited either
i invited myself and it turned out to be OK afterwords
audreyt implied that it would be OK before I did
but did not explicitly say so
23:43 FurnaceBoy is now known as unnoun
obra TimToady: aw ;) 23:45
everyone is welcome
nothingmuch but some ones are more welcome than others ;-) 23:46
(probably depending on the cardinality of the set of ones, and the available physical space)
obra Only my girlfriend is more welcome than everyone else 23:47
we've run out of flat surfaces for sleeping, though
nothingmuch well, we can have your bed then
obra ...
nothingmuch just became less welcome ;-)
anyway, i'm off, i've got bouncing to do 23:48
btw, the more welcome thing was a paraphrase of animal farm, not a cynical comment
audreyt the talk went extremely well :D 23:49
obra yeah :)
audreyt (and we did some VB design work afterwards, and also going hack on PIL<->MSIL in the next couple days with emeijer)
because they want VB to be compiled to JavaScript through Pugs, heh
obra !
23:51 unnoun is now known as FurnaceBoy
cognominal_ split /<sep>/, $string # is this a split that returns the separators as well? 23:52
audreyt I think split doesn't respect named separators
s/separator/capture/ 23:53
split /(<sep>)/ should do it
cognominal_ I though that
split /<?sep>/, $string # would give the old sematntic
audreyt it would 23:55
but not sure about splitting with named captures
cognominal_ I did not realize that the regexp may not be just a rule
split /(<sep>)/ makes more sense because to be homoegeneous with split /(,)/ 23:56
s/because/so as/ 23:57
audreyt nod

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!