Summer of Perl | 6.2.12 released! | paste: sial.org/pbot/perl6 | pugs.blogs.com | pugscode.org | pugs.kwiki.org | www.treehugger.com/files/th_images/paradigm.jpg
Set by nothingmuch on 12 July 2006.
08:30 ilogger2 joined 08:46 tup joined 08:53 elmex joined 09:03 Revision17 joined 09:22 dmsantam joined 09:30 wesley joined
svnbot6 r11693 | scw++ | Pugs::Emitter::Rule::Parsec 09:39
r11693 | scw++ | * <ws> and <?subrule>
r11693 | scw++ | * test updated, checked by replacing corresponding rules in
r11693 | scw++ | Parser/Literal.hs
10:04 dolmen joined 10:05 froh-doh joined 10:14 dolmen left 10:16 pdcawley_ joined 10:27 elmex joined
svnbot6 r11694 | cmarcelo++ | HsJudy: 10:36
r11694 | cmarcelo++ | * Change internal GC map-type name from Map to GCMap to avoid confusion.
r11694 | cmarcelo++ | * -Ofun in tests directory: Use prove to "make test", output is nicer.
r11694 | cmarcelo++ | Moved all test-support functions to Test.hs. Experimental syntax for tests.
r11694 | cmarcelo++ | * Remove unused Try.hs, k-judy uses new API.
cmarcelo audreyt: finalizer creation code doesn't work as I thought, it doesn't capture the right "needGC" definition, it takes all to be True, and this cause bugs to Maps who doesn't use GC. I'm going to sleep now and try some more type hackery tomorrow. part of my intent is described in paste.lisp.org/display/23024 [which doesnt work because "a" in undefined::a is diferent from "a" in type signature of the function]. 10:41
sleep & 10:42
10:54 BooK joined 10:57 bernhard joined 11:18 z3ndrag0n joined
pmurias someone fixed irc.pugscode.org 11:36
clkao fglock: fnord 11:49
fglock: is t/11-parsec.t supposed to fail?
11:51 iblechbot joined 12:18 xinming__ joined 12:19 xinming joined, chris2 joined
lumi cmarcelo: I think if you have g (x::a) = ... it should be the same 'a' as that 12:27
Oh, it's phantom.. Maybe you can hang the other type annotation on the C then 12:30
12:36 mauke_ joined
gaal pmurias: well, you could use symbol table stuff, if yout tokens aren't anonymous... 12:36
clkao raah! 12:41
12:51 mauke_ is now known as mauke
svnbot6 r11695 | clkao++ | Don't care about PVLV for now, this is about passing $_ in a for to 13:02
r11695 | clkao++ | another sub.
r11696 | clkao++ | Binding array and arrayref should work now. 13:05
r11697 | clkao++ | * Use Data::Bind's bind_op2 instead of tie for := operator.
r11697 | clkao++ | * Various tweaks for array related emit.
r11697 | clkao++ | * Fix capture emit for @array arguments.
13:10 pmurias_ joined
pmurias_ gaal: how would it look like? 13:11
13:12 drbean joined, xdg joined 13:15 ubuntu joined 13:29 larsen joined
clkao splice @array, 1, 1, (); 13:37
svnbot6 r11698 | clkao++ | binding a element of a bound array should work.
clkao does this replace the array element with undef or actually remove the element? i think it should be the latter. then operator/arrays.t is wrong 13:38
svnbot6 r11699 | clkao++ | * Nuke ::Alias in favour of bind_op2, which works for array. 13:41
r11699 | clkao++ | * +@array --> scalar @array.
r11699 | clkao++ | This passes most of t/operators/binding/arrays.t
gaal pmurias: %hash<$_> = Some::Package::{"&" . $_} for <token1 token2 someothertoken> 13:43
13:43 penk joined 13:49 stephanepayrard_ joined
gaal pmurias: but I don't think we parse symtable lookup yet 13:50
nothingmuch clkao: ping 14:09
clkao nothingmuch: pong 14:14
nothingmuch clkao: can you please, off the top of your head, list some App::CLI best practices? 14:15
like how to get a nice --help subcommand for everything
how to write usage
how to make program with no command say something useful 14:16
etc?
clkao nothingmuch: ping glasser ;)
nothingmuch glasser: ping
glasser nothingmuch: ack 14:17
nothingmuch glasser: backlog 10 lines ;-)
glasser I think the answer is mostly "steal from SVK" 14:18
(or jifty)
nothingmuch i was hoping for a shortcut
since SVK uses base classes extensively
glasser yeah
nothingmuch but i guess i'll just do that
glasser right, you have to watch out for the fact that &options is an App::CLIism whereas parse_arg is svkism eg 14:19
nothingmuch hmm 14:22
wanna look at my code, and maybe suggest some quick stuff?
it'll be a while till i actually get around to it
so in the mean while you may be tempted to do it for me ;-)
glasser Sure 14:23
nothingmuch nothingmuch.woobling.org/Mail-Summa...Tools/CLI/
lambdabot Title: Index of /Mail-Summary-Tools/lib/Mail/Summary/Tools/CLI
nothingmuch and CLI.pm, of course
glasser Yeah, we should whip up an app cli quickstart or something 14:24
nothingmuch when ann and I poked through App::CLI we managed to get a reference docpatch
but nothing introductory
glasser bah, i never set up ghc/darcs on the macbook 14:28
(does darwinports install ghc yet?) 14:29
hmm, where's the bin?
ah, script 14:30
hmm, so bin/svk has a couple stanzas at the top that deal with help and version 14:34
and SVK::Command::Help has some special stuff in it
that could both probably be factored out, either into App::CLI, or into some sort of App::CLI::Help module... clkao's taste
clkao: if you think either of those plans are a good idea, might do that on train to tel aviv 14:49
14:50 pdcawley joined 14:51 cognominal joined 15:04 ivanfrey is now known as Dr_Pi 15:11 cdpruden joined
nothingmuch 23 15:25
15:32 szbalint_ joined
clkao glasser: yes, that's a good idea 15:33
15:40 szbalint_ is now known as szbalint 15:42 shachaf__ joined
glasser nothingmuch, clkao: ok i think the next train trip will be "pull help and friends out of svk into app::cli" 15:45
which is right now :) 15:46
nothingmuch =)
awesome
glasser and if you really want to come to a hackathon at um, my girlfriend's cousin hotel in bat yam in a couple hours :) 15:47
though we probably have to sleep instead :)
15:48 shachaf__ left 15:50 jtextori joined 15:54 frederico joined
gaal glasser: does israrail have wifi? 15:55
jtextori Hi, can we ask for help in this channel ? I'm trying to install pugs from CPAN since this morning and ... 15:58
Aankhen`` On what platform? 16:01
16:02 douglashunter joined
jtextori on linux mandriva 2006 (Intel 32bit architecture) 16:02
gaal and what happened?
jtextori I got a undefined reference during this build : *** Building: /usr/bin/ghc -hide-all-packages -package stm -package network -package mtl -package template-haskell -package base -package pugs-fps -package pugs-HsSyck -package unix -package readline -idist/build -Ldist/build -idist/build/src -Ldist/build/src -o pugs src/Main.hs -lHSPugs-6.2.12 -threaded
/root/.cpan/build/Perl6-Pugs-6.2.12/dist/lib/pugs-HsSyck-0.1/libHSpugs-HsSyck-0.1.a(Syck.o):(.text+0xfe4c): undefined reference to `DataziYamlziSyck_da1l'
and other are following
douglashunter jtextori: I had the same problem this morning until I did a `sudo make realclean` 16:03
Then the build went fine.
jtextori Ok i'll try that first, thanks !
douglashunter This was after about 6 months of staleness in my pugs checkout. 16:04
I had a question about comments. S02 says: 16:05
Multiline comments are provided by extending the syntax of POD to
nest "=begin comment"/"=end comment" correctly without the need for
"=cut".
, but it seems that a "=cut" is needed after the "=end comment"
Am I misreading something?
(sorry for the crapflood) 16:06
gaal you mean it seems in pugs? not all that is specced is 'plemented
could you look for tests and possibly add one?
douglashunter gaal: yes, I meant in pugs. I'm just getting around to reading the synopsese and trying things out. 16:07
gaal: sure.
gaal there is t/syntax/POD.t but that doesn't look very comprehensive.
take another look: there may be more tests hiding about this 16:08
if so, please unify
if not, please beef up that test
douglashunter gaal: thanks for the pointer.
gaal and commit away
douglashunter Is it an open repo?
gaal what's your email address?
douglashunter [email@hidden.address]
gaal there's an invitation on it's way to you now. 16:09
douglashunter Ahh, thanks.
gaal when you commit, please add yourself to AUTHORS. welcome aboard :)
douglashunter Heh, thanks ;-)
gaal cpan-upload? util/add-text-file? what the world needs is util/add-pugs-committer, obviously. 16:11
16:11 justatheory joined
gaal this is actually an interesting test. it means use perl5:WWW::Mechanize, or for the ambitious, porting that to Perl 6 16:12
s/test/task/
svnbot6 r11700 | scw++ | Pugs::Compiler::Rule 16:14
r11700 | scw++ | * accepting <!before ...> and <!after ...>
r11700 | scw++ | * emit Parsec code for <!before ...>
jtextori douglashunter: thanks, but it seems it hasn't change anything (compilation problem) the error is hapening in this Compiling Pugs ( src/Pugs.hs, dist/build/src/Pugs.o ) 16:18
/usr/bin/ar: creating dist/build/libHSPugs-6.2.12.a
gaal jtextori: can you try from svn please?
svn://svn.openfoundry.org/pugs
jtextori Yes, non proble, i'll try that., thanks 16:19
gaal if you want to use parrot though, you also need HEAD of that.
but don't worry, parrot doesn't take long to build
jtextori i got parrot from svn this morning, don't really know why i didn't follow for pugs, and went to CPAN, maybe i'm tired ;-) 16:20
thanks for help
gaal let us know how it goes :) 16:21
jtextori gaal: sorry, but is co available for anonymous ? because i'm getting an error : can't connect to host, connection refused ?? 16:28
gaal svn.openfoundry.org/pugs # try the "http" 16:29
lambdabot Title: Revision 11700: /
jtextori OK
16:38 Marsel joined 16:41 Marsel joined
pmurias fglock: ping 16:42
cancel ping 16:44
gaal wonders what is the sound of a ping cancelled 16:45
16:54 Marsel is now known as jtextori 16:55 dolmen joined 16:58 slanning joined, salty-horse joined
clkao something is going really wrong with v6 17:08
it's trying to compile html::entity???
17:10 jabbot joined 17:11 pdcawley_ joined
clkao its gone totally mad 17:14
nothingmuch I CAN"T LET YOU COMPILE THAT CLKAO 17:19
CLKAO? WHAT ARE YOU DOING?
clkao i am trying to smoke v6, andfound it totally went mad 17:20
17:20 _maydayjay_ joined
gaal you are the smoke? 17:21
nothingmuch clkao: dude, that's what you get for smoking weird shit
17:25 baest joined
svnbot6 r11701 | clkao++ | Emit ; for last statement in block. 17:27
17:29 Dr_Pi joined
svnbot6 r11702 | clkao++ | Need params for bind_op2. 17:33
Dr_Pi I'm running make smoke and it appears to be hanging at t/perl5/mothod.t
Aggggh, s/mothod/method/ 17:34
clkao so "grammar YetAnother { rule def { D <eh> F } }" fails parsing 17:39
17:40 penk joined 17:47 cdpruden joined
gaal yay, it compiles! p6 view of new ast coming up... 17:49
17:51 pdcawley_ joined
clkao whoot 17:51
Juerd Does anyone still use feather for other things than irc? :)
gaal slides :) 17:52
and pugscode.org should move there
and repo too
svnbot6 r11703 | gaal++ | * CapInternals now compiles via CITest, which successfully dumps
r11703 | gaal++ | some AST types. Demo output coming up in a nopaste near you...
17:53 spinclad joined
Juerd I put in extra RAM because the old amount was sufficient 17:55
I mean insufficient
Now it has more than enough and it's hardly even used :)
wolverian gaal, should or will? 17:56
Juerd Why should or would it?
gaal would once tuits are obtained
Juerd Do we want a single point of failure that badly? :)
gaal should for community hackability
Juerd That's a good point.
gaal as for repo, I like fast updates+commits :) 17:57
Juerd Is the current repo slow, then?
gaal outside .tw? yeah
Juerd Perhaps two repositories can synch, so it can be fast both in and outside .tw 17:58
I have a rm -rf running for 10 minutes already 17:59
On a single directory with gazillions of small files.
qpsmtpd forgot to clean up after itself.
Zapping directories should be easier 18:00
gaal I'm sure there are filesystems where it's fast
Juerd There are
gaal but maybe those are slower for other things?
Juerd Not really
But they are, in general, less reliable.
Still, though, that doesn't fix my case :) 18:01
I don't even know how many files are in there 18:02
But it's email, mostly small files
16 GB of it.
gaal finding out how many probably takes about as long as deleting 18:03
Juerd It does.
gaal unless there's some clever incidental way of doing it, like looking at the space allocated for the directory itself.
Juerd I'm using rm -rf just because other tools weren't able to do this as fast.
Nope, because I use directory indexes
If I knew any of the file names in this directory, I could retrieve the file very fast :) 18:04
But this does make the directory entry allocation unpredictable
This indexing thing is also why I hadn't noticed anything. There was no performance drop.
It scales well, that much is obvious to me now :)
I only noticed because the hard drive was getting full 18:05
And du got stuck in this directory, eventually segfaulting.
jtextori Hi again, does anyone would know where this error come from, and what it means ? => Building Pugs-6.2.12... 18:07
ghc-6.4.1: Error: module 'Data.Yaml.Syck' is exposed by package pugs-HsSyck-0.1 and package pugs-HsSyck-0.2
Build failed: 256 at util/build_pugs.pl line 156.
i'm still trying to compile pugs
with last svn version
gaal jtextori: please run 'ghc-pkg unregister --user pugs-HsSyck-0.1'
jtextori OK thanks
gaal and for good measure, the same with 0.2 too 18:08
jtextori OK
18:15 penk left
clkao weird. i don't see to be able to export to core::global 18:16
buubot :global will be added by christmas
clkao i mean in p5 18:17
gaal clkao: do you have root? 18:18
clkao gaal: huh? 18:19
on what?
gaal sorry, I'm in a discordian mood, don't mind me :)
ladies, and gentlemen: 18:20
pasteling "gaal" at 192.115.25.249 pasted "Perl 6 Representation of New AST" (788 lines, 11.5K) at sial.org/pbot/18586
xerox prods gaal 18:23
gaal moose 18:24
s
clkao moose
svnbot6 r11704 | gaal++ | * misc/pX/gaal/CITest.hs: demo generation of Perl6Class on new AST.
r11704 | gaal++ | Output is here: sial.org/pbot/18586
lambdabot Title: Paste #18586 from "gaal" at 192.115.25.249
clkao mmm 18:25
src/Pugs/AST/Internals.hs:1987:0:
Duplicate instance declarations:
src/Pugs/AST/Internals.hs:1987:0: instance Typeable Regex
src/RRegex/PCRE.hs:48:8: instance Typeable Regex
gaal grr. fixing.
xerox `instance Typeable Regex' is scary!
gaal also, bogus. we don't support 6.2 any more... 18:26
eh, no, that's the Tree instance
xerox Who uses 6.2 anyway... <chuckles>
gaal you mean it sucks because it's typeOf () which is obviously a lie? 18:27
well, sometimes a hacker's gotta do... never mind :)
fixed. 18:31
xerox %-)
svnbot6 r11705 | gaal++ | * unbreak the build 18:33
r11705 | gaal++ | * remove some old cruft from Pugs.AST.Internals
r11706 | gaal++ | * unbreak CapInternals, for all-around fairness and wellbeing
18:36 aufrank joined
salty-horse according to t/syntax/parsing/sub_calls.t it's still hasn't been decided whether the call foo'bar' is allowed. can i still 'fix' the examples using that method so they could run? or should it wait? 18:48
clkao So once class is declared, the name became bareword. what to do with Str::Foo ? is it a type ? or is it Foo under Str:: ? 18:49
wilx Start using typename! 18:53
Yay!
j/k
:D
19:08 Corion joined
lisppaste3 xinming pasted "Build fail" at paste.lisp.org/display/23040 19:08
xinming no one encountered my error? 19:20
audreyt xinming: try r11707 19:22
svnbot6 r11707 | audreyt++ | * unbreak the build. 19:23
gaal that seems to be a popular commit message today :(
audreyt clkao: once class is declared, it installs a nullary macro function of the same name
clkao: it's a Pugs-ism (unspecced) but the only approach I find actually usable
gaal I don't understand r11707, audreyt. don't we require 6.4? 19:24
audreyt Str::Foo is Foo under Str::.
<= 604
not <604
gaal is that 6.4.0?
audreyt no
hm 19:25
it's 6.4.x
gaal xinming: what ghc do you have?
audreyt 604 is the level of GHC
.0 .1 .2 are patchlevel
governed by another macro
gaal ah
xinming gaal, 6.4.2
audreyt add a comment perhaps?
I need to go to the airporti n a few mins :)
aufrank I'm beginning to question your commitment, audreyt ;) 19:26
audreyt I'll go back from the airport if it has wifi
:)
aufrank slacker
audreyt lol
gaal oscon bound? have fun! 19:27
audreyt yes, oscon bound :)
thanks!
aufrank audreyt: is this the trip that also involves microsoft presentations, or did that already happen?
audreyt it happened quite a while ago
I'm in Boston now
aufrank cheaper than flying from .tw, I guess ;) 19:28
audreyt :)
gaal !? 19:29
./third-party/HsSyck/Data/Yaml/Syck.hs:{-# OPTIONS_GHC -optc-D__GLASGOW_HASKELL__=604 #-}
another one of our white lies? :)
audreyt no
hsc2hs
inserts syms
it's genned 19:30
clkao audreyt: ok
gaal so is src/Pugs/Embed/Parrot_hsc.h probably? it mentions __GLASGOW_HASKELL__ < 409 :-)
audreyt yeaah
svnbot6 r11708 | clkao++ | forget_unlocal injects to actual caller namespace now. 19:32
audreyt gaal: blog bit sent, happy journaling! 19:33
svnbot6 r11709 | gaal++ | * add a comment warning about version sym gotcha
gaal thanks! :) 19:35
svnbot6 r11710 | clkao++ | * bare_block in param is a anonsub.
r11710 | clkao++ | * avoid \(@x) and \(my @x) for bind_op2, as that flattens the arrays.
r11711 | clkao++ | Be more careful when looking for 'my @x :='. 19:41
nothingmuch audreyt: subetha to see://pasta.woobling.org 19:44
xerox nothingmuch: is it open to anybody? 19:48
nothingmuch yes 19:49
but probably not interesting to everybody =)
xerox If it is code, mind if I stare?
nothingmuch it's mailing list summaries 19:50
but sure, you're welcome
xerox Oh thank you.
gaal journal up 19:51
nothingmuch xerox: well? i'm waiting ;-)
gaal that formatting is probably not very good for a narrowish column... 19:52
xerox I am getting see :)
nothingmuch Dan Kogai requested a version of C<slurp()> which would unconditionally 19:56
return a scalar, and proposed C<swallow()>. Damian Conway offered some
ah
nothingmuch.woobling.org/p6_april.html
please proofread
lambdabot Title:
wolverian haha
oh, nice, I hadn't read the "Context and coercion" thread 19:57
nothingmuch++ 19:58
nothingmuch wolverian: kudra++
nothingmuch is just the toolsmith
xerox nothingmuch: oh the evaluation period finished.
xerox sighs 19:59
wolverian oh, kudra++
nothingmuch xerox: =(
wolverian I still kind of want to get rid of context and only talk about return type mmd :-) 20:00
(but only when feeling bad and vengeful against others)
nothingmuch wolverian: i actually agree 20:03
on some level
links changed from gmane to google
salty-horse asking again :/ : according to t/syntax/parsing/sub_calls.t it's still hasn't been decided whether the call foo'bar' is allowed. can i still 'fix' the examples using that method so they could run? or should it wait? 20:11
nothingmuch Written by audreyt and posted by grinder 20:18
oops
wrong window
svnbot6 r11712 | salty_horse++ | fixed typo 20:23
r11713 | nothingmuch++ | Mailing list summary for April 2006 in Mail::Summary::Tools/YAML format 20:26
20:29 jferrero joined
nothingmuch audreyt: summary draft saved pugs.blogs.com... does anything else have to happen? 20:30
robkinyon: ping
gaal: ping
maybe one of you can see it there?
svnbot6 r11714 | fglock++ | v6 - parses 'Native.pm' from v6-Compiler 20:32
xerox gaal: what does DriFT exactly do in your usecase? 20:33
svnbot6 r11715 | fglock++ | v6-Compiler - fixed 'bool' rule in Native.pm
20:37 buubot joined 20:43 buubot joined
gaal nothingmuch: change it to "publish now" and save again 20:45
nothingmuch delegates
gaal xerox: writes "showPerl6TypeDefs" instances for arbitrary Hs data types
and soon<tm>, asPerl6Object instances, too 20:46
nothingmuch gaal: she says she doesn't want to publish without review
and whther audrey can review it in there'
gaal I expect she can
nothingmuch s/and/and the original question was/
gaal but worst case, copy the html and send? 20:47
audrey's on the way to the airport now, may not have wifi there
nothingmuch i guess it can just be posted, right
gaal xerox: src/DrIFT/RulePerl6Class.hs - see also RuleYAML.hs and YAML.hs in that directory for similar stuff 20:48
and look for files called "Instances.hs" in the source tree
that's where the DrIFT output goes.
20:59 larsen joined 21:00 FurnaceBoy joined 21:01 kanru joined 21:05 buubot joined 21:18 ludan joined 21:23 SubStack joined 21:25 buubot joined 21:35 daxim joined
obra heads up: perl.org will be going off the air for the afternoon in about 15 minutes 21:37
21:41 larsen joined
spinclad "should t/perl5/method.t be diverging in time and space?" (r11690) ... i see Dr_Pi hit this too. 21:45
t/perl5/roundtrip, too (test 3) 21:46
something in or around eval(q/.../, :lang<perl5>); is getting unhappy; dunno what yet tho. 21:49
22:09 rbrown_ joined
clkao summon fglock 22:15
22:22 tup joined
nothingmuch hmm 22:27
are email addresses in svn.openfoundry.org/pugs/docs/summa...04-30.yaml a bad thing?
being publicly readable etc?
svnbot6 r11716 | cmarcelo++ | * Fixes GC.freeRef being called for non-GC Refeable types. Still could 22:36
r11716 | cmarcelo++ | be better.
22:47 lichtkind joined
lichtkind TimToady may i ask you something? 22:49
Juerd seen TimToady 22:53
jabbot Juerd: TimToady was seen 2 days 6 hours 16 minutes 28 seconds ago
Juerd Sending an email may work better :)
lichtkind thanks juerd 22:54
22:55 nnunley joined
lichtkind juerd juerd the editor does fine, 3rd release after bochum 22:56
Juerd Good to hear. Does it have many users? 22:57
I'm going to bed. I'll read your reply tomorrow 22:58
Good night
lichtkind it grows finally today began last phase of cpanification
good nighnt
we got some users in forum i know of but rest ist hard to say 23:00
clkao so we are at 415 passes now. a bit less than last week 23:03
Dr_Pi before I get to t/perl5/method.t hanging I'm getting "t/perl5/array.t 23:07
Not an ARRAY reference at (eval 3) line 23."
23:08 FurnaceBoy joined 23:11 bsb joined
svnbot6 r11717 | clkao++ | Resource bound support for v6.pm 23:15
r11718 | clkao++ | emit for 'sub { 42 }()' properly.
clkao so t/operators/binding/arrays.t mostly pass 23:16
jmm 23:42
so if i set $::{X}, i can call X, but not X() ??
ok, 522 pass 23:52

Powered by Cro and the Raku® Programming Language.

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

Thank you!