Check your feather email | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com | www.treehugger.com/files/th_images/paradigm.jpg
Set by audreyt on 29 August 2006.
00:05 jferrero joined, radians joined 00:17 DaGo joined 00:22 weinig is now known as weinig|bbl 00:29 nekokak joined
markstos I copied the first post over to pugs.blogs.com: pugs.blogs.com/ I'll continue to copy my Perl6 blogging stuff over there from my use.perl.org journal, and then send stuff directly to the pugs blog after that. 00:33
TreyHarris markstos: so i should take your use perl journal off my newsreader lest i get duplicates? :-) 00:43
audreyt: are you around?
00:44 discordja_ joined 00:47 ludan joined, soisoisoi joined
markstos TreyHarris: it's a compliment to know I'm in someone's newsreader. Once I catch up with a few current things, I'll post Perl6 stuff just to the Pugs blog,and other perl stuff back at use.perl, so you won't have to suffer through duplicates for long. 00:48
01:06 mako132_ joined
markstos How do I define an array of hashes in Perl6? I didn't see an example in S09 01:10
01:16 JarJarBinks joined
TreyHarris markstos: define? my Array @array of Hash, I'd assume 01:17
markstos The "of" is not recognized yet there. 01:18
I'm looking for something like my @a = ({ a => b}, { c => 'd' } );
TreyHarris yeah, the typing tests mostly fail
that should work
that's why i was trying to raise audrey actually 01:19
because:
?eval my @l = ({d => 'a'}, {d => ''}, {d => 'cd'}); @l.perl
01:19 evalbot_12913 is now known as evalbot_12915
evalbot_12915 "[\{(\"d\" => \"a\"),}, \{(\"d\" => \"\"),}, \{(\"d\" => \"cd\"),}]" 01:19
TreyHarris that works on evalbot
but it doesn't work on my pugs build at the same rev 01:20
it hangs for about four minutes, eating up cpu and mem, and then segfaults
markstos ?eval my @a = ({ a => 'b', a2 => 'b3' }, { c => 'd' } ); for @a -> %h { say %h.perl; }; 01:21
evalbot_12915 OUTPUT[{("a\tb" => ("a2" => "b3")),} {("c\td" => undef),} ] undef
markstos Not quite right.
TreyHarris ?eval my @a = ({ a => 'b', a2 => 'b3' }, { c => 'd' } ); @a.perl 01:23
evalbot_12915 "[\{(\"a\" => \"b\"), (\"a2\" => \"b3\")}, \{(\"c\" => \"d\"),}]"
markstos I also tried %() instead of {}, which didn't work either. 01:24
TreyHarris ?eval my @a = [{ a => 'b', a2 => 'b3' }, { c => 'd' } ]; @a.perl
01:24 DaGo joined
evalbot_12915 "[[\{(\"a\" => \"b\"), (\"a2\" => \"b3\")}, \{(\"c\" => \"d\"),}],]" 01:24
TreyHarris huh, that's bizarre
markstos ?eval my @a; push @a, %( a => 'b', a2 => 'b3' ); push @a, %( c => 'd', c2 => 'd2' ); for @a -> %h { say %h.perl; }; 01:25
evalbot_12915 OUTPUT[{("a" => "b"),} {("a2" => "b3"),} {("c" => "d"),} {("c2" => "d2"),} ] undef
markstos So, the hashes are always getting broken up into subhashes. 01:26
I'm not sure how to be more explicit about what I mean. 01:27
?eval my @a; my $b = %( a => 'b', a2 => 'b3' ); push @a, $b; my $c = %( c => 'd', c2 => 'd2' ); push @a, $c; for @a -> %h { say %h.perl; }; 01:28
evalbot_12915 OUTPUT[{("a\tb" => ("a2" => "b3")),} {("c\td" => ("c2" => "d2")),} ] undef
markstos Does Perl6 have a builtin way to read YAML, or just to produce it? 01:31
01:55 weinig|bbl is now known as weinig
audreyt markstos: read yaml? 02:04
?eval eval("1: 2", :lang<yaml>) 02:05
evalbot_12915 {("1" => "2"),}
audreyt like that?
markstos Whoa. Call.
Cool. :)
audreyt =)
yaml is pretty much the B:: of Pugs
I'd like to make SQLite the DBM of Pugs as well...
markstos I was thinking of using yaml as a back door to solve my array of hashes creation problem. 02:06
audreyt just wrote a GC in p5+SQLite for $job... going to run now
02:07 kanru2 joined
audreyt bbiab :) 02:07
markstos: give your blogpost a title? 02:08
markstos Sorry I missed that. Doing now. 02:10
02:10 mako132_ joined
markstos done. published now. 02:11
02:19 mjk joined 02:25 peacock joined 02:40 ajs_ joined
peacock any pugs people around who can help me with a build error? 02:43
markstos I figured out the issue with declaring an array of hashes. The declaration was correct, but it is necessary to iterate over it with 'for each' instead of just 'for', which I don't fully understand yet.
?eval my $quiz = ( { a => 1, b => 2 }, { a => 1, b => 3 },); for each($quiz) -> %h { say %h.perl; }
evalbot_12915 OUTPUT[{("a" => 1), ("b" => 2)} {("a" => 1), ("b" => 3)} ] undef
markstos peacock: I'm here but I doubt I'll be much help with a build error. 02:44
peacock well i'll throw it out anyway
i get an error when trying to build libHSPugs.a 02:45
it says illegal option, then no archive members specified
any ideas?
audreyt nopaste the whole buildlog? 02:46
peacock by it i mean the 'ar' command
audreyt and trace util/build_pugs
peacock ok one sec
markstos Here comes the professional-level support. ;)
audreyt illegal option is most likely due to embperl5
if you build again with env PUGS_EMBED set to "-perl5" 02:47
(need realclean perhaps)
then it's likely to work
though I'd ofcourse like to make embperl work
pasteling "peacock" at 72.225.222.225 pasted "build error" (464 lines, 31.4K) at sial.org/pbot/19393 02:48
peacock i pasted what i had - but no trace from build_pugs
audreyt this is osx?
peacock yeah
its a fresh source tree from yesterday 02:49
audreyt ok. at util/build_pugs.pl
line 288
put a "die @a_file" there
and run "make pugs" 02:50
and see what it prints
actually
die "$ar - @a_file";
and see what your perl thinks the full_ar is 02:51
most likely your $Config{full_ar} contains spaces
do you use the OSX builtin perl?
ppc?
peacock yeah
ok i'm retrying with that modification 02:52
having some trouble with my irc client :( did you see the result? 02:53
audreyt no
peacock "/usr/bin/ar - /Users/rob/proj/pugs/dist/build/libHSPugs-6.2.12.a at util/build_pugs.pl line 288."
there it is
audreyt ok. if you run 02:54
/usr/bin/ar s /Users/rob/proj/pugs/dist/build/libHSPugs-6.2.12.a
does it succeed with no messages?
peacock "ranlib: can't open file: /Users/rob/proj/pugs/dist/build/libHSPugs-6.2.12.a (No such file or directory)" 02:57
audreyt weird.
do a
find dist/build | grep libHSPugs
and see where the .a files is there
audreyt blames Cabal -- this is apparently a safe bet 02:58
peacock hmm...just don't have it it seems
audreyt find . third-party | grep libHSPugs
then
peacock wait
./third-party/installed/lib/Pugs-6.2.12/libHSPugs-6.2.12.a 02:59
i have that
audreyt that's the only libHSPugs? there's none under dist/build/?
find . |grep libHSPugs
peacock nope :\
audreyt and see if it's anywhere else
it has to come to install/ from somewhere...
peacock thats the only one on the system 03:01
audreyt ok
peacock searched in spotlight
could this be the result of earlier builds?
i had some earlier problems that I managed to muddle through
audreyt how large is it? 03:02
ah. that might be the reason too
look at your Pugs.cabal
the last line should begin with ghc-options:
nopaste that line?
peacock its 44.1mb
sial.org/pbot/19394 03:04
lambdabot Title: Paste #19394 from "peacock" at 72.225.222.225
audreyt hmm 03:05
03:06 soisoisoi joined
audreyt perl is build with debug on (-g)? weird 03:06
peacock the darwin stuff is always just different enough to cause me problems 03:07
audreyt what does
perl -V:ranlib 03:08
say on your system?
peacock ranlib='/usr/bin/ar ts';
audreyt aha.
peacock i'm afraid i'm still completely in the dark 03:09
svnbot6 r12916 | markstos++ | A first draft of the Perl6 dispatch quiz! 03:10
r12916 | markstos++ | Interactively use Test.pm to enter your own answers about how
r12916 | markstos++ | you expect dispatching to work.
r12916 | markstos++ | More quiz questions are coming, but they are also easy to add
r12916 | markstos++ | in the YAML in the source. (Which could be broken out into a separate game file...)
markstos The quiz is at examples/games/dispatch_quiz.pl
03:11 weinig is now known as weinig|zZz, lollan joined 03:12 kanru joined
audreyt peacock: up to r12917 03:12
and nopaste the build log again
this time it'll become instrumented with "ar" calls
which should help to diagnose the rune to use 03:13
peacock ok i'm on it 03:14
svnbot6 r12917 | audreyt++ | * add instrumentation to build_pugs.pl
03:17 SubStack joined
peacock sial.org/pbot/19395 03:22
lambdabot Title: Paste #19395 from "peacock" at 72.225.222.225
audreyt ok, I got it. 03:24
it's embperl5 problem
please change util/build_pugs.pl line 303
from
$run_setup->('build');
to
$run_setup->('build', '--verbose');
and nopaste buildlog again 03:25
my taxi is here; I need to run in a bit; will be back in a couple hours
if you need a pugs built now, set
export PUGS_EMBED=-perl5
and realclean makefile.pl make
will get you a nonembed working pugs
but I'll look at the --verbose log and adjust cabal flags accordingly as soon as I'm back on the net
see you :)
*wave* &
peacock ok - thanks a lot for all your help 03:26
!
*wave*
markstos peacock: Yesterday audreyt turned on Perl5 embedding by default. It's a cool feature allowing you do things like use perl5:HTML::Template, and have that module work in Perl6. 03:28
But apparently there are still some kinks to work out with getting it to build everywhere. :)
peacock markstos: looks like i picked the wrong day to try building pugs! 03:29
markstos I suspect it will work if you that out that build option.
There is still a lot to explore and play with without Perl5 embedding.
peacock the build system alone has been keeping me entertained for quiet a while ;) 03:30
markstos Ah.
Well, you always play with the evalbot here in the channel:
?eval say "Hello, Rob"
03:30 evalbot_12915 is now known as evalbot_12917
evalbot_12917 OUTPUT[Hello, Rob ] Bool::True 03:30
peacock yeah thats fun, but i actually do enjoy digging around in the pugs build 03:31
the haskell stuff scares and fascinates me
markstos likewise. 03:32
peacock and i have to say, its pretty cool that people take the time to help a newbie to get set up 03:34
markstos As it turned out, I think you were providing valuable feedback about how Perl5 embedding builds, so it's a two way street. 03:35
pasteling "peacock" at 72.225.222.225 pasted "verbose log of os x build error" (252 lines, 24.6K) at sial.org/pbot/19396 03:36
03:40 markstos left
svnbot6 r12918 | markstos++ | Update the dispatch quiz game a little, including comments 03:40
r12918 | markstos++ | about more scenerios I'd like to model.
peacock *thinks thats a good idea*
03:59 mako132_ joined 04:16 frederico joined 04:29 mako132_ joined 04:34 lambdabot joined
TreyHarris ?eval ({d => 'a'}, {d => ''}, {d => 'cd'}) 04:35
04:35 evalbot_12917 is now known as evalbot_12918
evalbot_12918 ({("d" => "a"),}, {("d" => ""),}, {("d" => "cd"),}) 04:35
TreyHarris ?eval ({d => 'a'}, {d => ''}, {d => 'cd'} :lang<yaml>) 04:36
evalbot_12918 Error: unexpected ":" expecting comment, operator, "," or ")"
TreyHarris ?eval ({d => 'a'}, {d => ''}, {d => 'cd'}).yaml 04:37
evalbot_12918 "--- \n- \n ? d\n : a\n\n- \n ? d\n : \n\n- \n ? d\n : cd\n\n"
04:38 lambdabot joined 05:11 kanru joined 05:18 lambdabot joined, DaGo joined 05:38 chrisdolan joined
svnbot6 r12919 | malon++ | examples/games/dispatch_quiz.pl - print questions in a friendlier format 05:41
05:42 chrisdolan left 05:44 justatheory joined 05:56 lambdabot joined 05:57 nonSmoker joined, nonSmoker left
gaal is there a suggested technique for parsing things that can come out of order and accummulating them somehow classified? 06:11
e.g. :($x? = 42 is rw is moose where { $x % 2 }) 06:12
afaik 'is rw', 'is moose' and 'where...' can come in any order
the way i was thinking of was to wrap them all in the same type, and do [modifiers] <- many choice [...] 06:14
but that's kinda artificial since 'is rw' is by no means an Exp, for example
07:05 Lorn joined 07:23 azr|elious is now known as azr|el
svnbot6 r12920 | trey++ | Adding a smartlink 07:26
r12921 | trey++ | [docs/Perl6/Perl5/Differences.pod] 07:39
r12921 | trey++ | Made the section on twigils actually contain a twigil or two ;-)
gaal ugh. what's the best way to deal with clashing names oldval<->newval? In Pugs.AST so far we were lucky and didn't need to import Pugs.Val qualified; but now I want Param(..)... 07:40
...and Pugs.AST reexports 'module Pugs.Val' 07:41
I think for now I'll do type NewParam = Param and export that, if that allows constructors 07:42
doesn't seem to work :( 07:44
this is ugly, but the fix is maybe to have the data type called NewParam and Param be an (unexported) alias to that?
of just call it SigParam and be done with it :/ 07:45
going with that, as it does make some sort of sense. 07:47
07:48 ezraburgh joined
araujo what are you trying to do gaal ? 07:48
gaal take over the world! :)
araujo ok, well, we all are at it ....
gaal araujo: introduce a parser that deals with :()
and constructs a Signature object with Params in it 07:49
the problem was that the Param symbol was already used in old-val land.
but it's fixed now.
svnbot6 r12922 | trey++ | [Differences.pod] 07:54
r12922 | trey++ | Fixed section on quotes to refer to the
r12922 | trey++ | quotes they were referring to....
08:00 iblechbot joined 08:11 drrho joined
svnbot6 r12923 | trey++ | [Differences.pod] 08:18
r12923 | trey++ | Infinite C<loop>. :P5 regexes.
08:21 soisoisoi joined 08:24 nekokak_ joined
gaal audreyt: in Sig, does s_namedSet include names of positional params too, or only "pure nameds"? 08:28
likewise s_requiredNames
08:50 xerox joined 08:51 elmex joined 09:06 lollan joined 09:11 bernhard joined 09:29 plisk joined 09:37 xerox joined
gaal I'll commit rudimentary Signature parsing later today. Something that can be really helpful to continue with this is tests, but as far as I know an introspection API isn't specced yet for Signatures, so that might be hard to do :) 09:57
can we get at least a conjectural API please? 09:58
if anyone wants to invent one and just commit tests, please go ahead 10:00
e.g.: my $sig = :($x); is $sig.requiredPositionals, 1 10:01
10:02 onsen joined 10:31 DaGo joined
gaal yay! ghc 6.6 release candidate, only a week late! www.haskell.org/ghc/dist/current/dist/ 10:48
lambdabot Title: Index of /ghc/dist/current/dist
nothingmuch can a type class have an instance of another type class? 10:57
audreyt subclassing is allowed, sure 10:59
nothingmuch is going to solve the layout/slot issue by overengineering once and for all 11:00
and then i think i am going to write me a blondie 2 in haskell
but just OO
11:01 ruoso joined
gaal going out for lunch for a few hours 11:01
audreyt: will you be here in say 2h?
nothingmuch sounds like a fun lunch
enjoy =)
gaal yeah, buchari's the best :)
nothingmuch dunnos
www.post-literate.com/gerpunx/archi...ursion.gif 11:02
audreyt sure
11:02 ludan joined
gaal darn, forgot my new feather pw 11:02
ok so i won't mail you my latest patch, we'll talk then 11:03
ludan hi
gaal gotta moose, later! &
nothingmuch we need to implement hasselhoffian recursion in perl 6
12:04 mako132_ joined 12:22 Limbic_Region joined 12:26 weinig|zZz is now known as weinig 12:29 nekokak joined 12:32 araujo joined 12:35 dolmans joined
gaal remoose 12:50
araujo morning 12:54
12:57 lumi joined 13:05 buetow joined 13:10 polettix joined 13:13 frederico joined
svnbot6 r12924 | gaal++ | * Basic support for parsing Signatures with :(). 13:13
r12924 | gaal++ | Only named parameters are parsed, and optionals (incorrectly)
r12924 | gaal++ | require a '?' even if they have a default value. No parameter
r12924 | gaal++ | decoration (type constraints, access traits, etc.) is supported
r12924 | gaal++ | yet. The following forms do work:
r12924 | gaal++ | :($x:)
r12924 | gaal++ | :($x)
r12924 | gaal++ | :($x: $y, $z)
r12924 | gaal++ | :($x: $y, $z? = 42)
r12924 | gaal++ | :($x: $y?, $z) # correctly raises a parse error
gaal tests and golf welcome... 13:14
13:15 takanori joined
gaal s/named/positional/.... 13:16
13:17 [particle] joined
gaal bbiab... 13:20
13:25 cjeris joined
audreyt ooooh. 13:27
13:28 vel joined 13:33 takanori joined 13:34 weinig is now known as weinig|bbl
svnbot6 r12925 | audreyt++ | * Proper support for PCRE submatch counts. 13:34
r12925 | audreyt++ | * Disable explicit imports from RRegex.PCRE, in preparation
r12925 | audreyt++ | for a #ifdef PUGS_HAVE_PERL5 hook in RRegex.hs so that we
r12925 | audreyt++ | can switch into RRegex.Perl5 some day.
13:49 chris2 joined 13:50 gaal_ joined
Limbic_Region audreyt - "some day"? Is the plan to move in that direction now on hold in lieu of something else? 13:50
audreyt as in "not today" 13:51
as my brain is fried
as in "perhaps tomorrow."
Limbic_Region ah
ok
13:52 marmic joined
audreyt peacock: In case you are backlogging, I've got identical setup on powerpc g4 14:02
looking into the build problem now
though not sure if I'll stay awake long enough... 14:03
obra 'morning 14:04
audreyt yo 14:05
clkao audreyt: hey 14:06
audreyt: what happened to the plain that using rule as a rendering template language
s/plain/plan
audreyt well, we didn't have a reasonable PCR ack then... 14:07
*back
now I guess it'd be nice and fun.
clkao ya
so... mmm very interesting.
i will talk to tina later 14:08
audreyt tina?
the notes are still in misc/pX/audreyt/rules-are-templates
clkao audreyt: html::template::compiled 14:09
audreyt nice!
excellent stuff.
clkao ya. it's even faster then tt 14:10
there are more complaints about jifty being mason than i expcected 14:11
audreyt View::HTC then?
mason is more polarizing than tt2.
as it's not a humble framework.
obra HTC?
ah. parsed.
audreyt HTML::Template::Compiled
definitely not MSIE5 HTML Components :) 14:12
obra but yes, TT people Hate mason, while mason people usually prefer mason to TT. ;)
audreyt I prefer mason to TT, and I hate mason.
obra mm. need hackathon to make jifty::view::declarative go 14:13
audreyt I guess that makes me in both camps comfortably
clkao norman ported 50% to tt
he said the other 50% is hard
audreyt yeah. you not going to oopsla...
not sure what other conferences are there between oct and feb
clkao i am proably going to yapc::sa 14:14
audreyt but it's brazil, we are more likely to hack p5p6 or p6p6 or p5p5
obra when is oopsla? 14:15
"p5p5" for real?
audreyt B::Visitor
obra clkao: norman never committed anything
audreyt or some other lathos-isomorphic plans
obra :)
audreyt metaprogramming p5 in p5, that is 14:16
clkao audreyt: oh, seee if we can switch to PCR based templates. then it's relevent
14:16 weinig|bbl is now known as weinig
audreyt that's even more crazy that declarative UI... 14:16
14:16 nothingmuch joined
audreyt autoscraping reversible templating 14:16
obra heh
clkao X
audreyt obra: oopsla is oct22
clkao D
audreyt ~26
clkao where's oopsla?
obra portland
actually. 14:17
audreyt portland convention centre
obra cl: when is google summit?
clkao 16-18
obra so. we could do 19-21 in portland
clkao i guess sfo pdx isn't too far away
obra or after oopsla
clkao . o O (perl travel agent service)
obra also, we have a staff programmer in portland.
who knows where one can find onsen
clkao this is crazy talk
audreyt in portland? wow. 19~21 would be good for me
obra schwern claims there are hotsprings in portland. 14:18
clkao Intternals::SvREADONLY(${\undef}, 0) undef = 42
audreyt orig. planning pair with fglock right after oopsla
clkao: yes, I saw that hack (and SvTRUE = 0 and SvFALSE = 1) in perlmonks a while ago...
obra audreyt: fglock is comming to oopsla?
audreyt no, flying to brazil for preconf hackathon
obra ahh.
19-12 is a little short
audreyt 19~26 is not 14:19
obra and knowing you, you'll be slideshacking
that's true.
clkao mmm
then machupichu before yapcsa...
audreyt well, my keynote is 22
obra keynote?! wow.
audreyt opening day of dynamic language symposium
not oopsla main. side conf. but yeah, invited speaker :) 14:20
which means I don't have to submit a paper.
(yay)
obra ah :)
audreyt so that means 24~26 is perfectly good hacktime for me, as is probably 19 and 20 14:21
obra *nod* 14:22
sounds like we have a potential plan. 14:23
sadly, it doesn't mean I have to go to .tw 14:24
clkao i will be back home tomorrow noonish
obra: winter is not there yet
14:25 fglock joined
gaal_ remoose 14:26
14:26 gaal_ is now known as gaal
audreyt hey gaal. I've looked at your patch 14:30
and :(Int $x) seems doable
and I wonder if we should bump CompUnit version for this
but I need to fade now...
14:31 lollan joined 14:32 weinig is now known as weinig|bbl
audreyt *wave* & 14:32
gaal why need compunit be bumped? we don't use this parser for blocks/subs/methods yet 14:34
anymoose named args and fixing 'is X' look more urgent than Int $x... 14:36
14:36 Eimi joined
gaal was there any additional thought given to the relationship of multis and guards? 14:37
14:41 pasteling joined 14:42 hexmode joined 14:43 [particle] joined 14:47 elmex joined, onsen joined 14:58 lisppaste3 joined 15:02 maard joined 15:07 kanru joined 15:14 prefiks joined 15:17 penk joined 15:19 nothingmuch joined 15:21 weinig|bbl is now known as weinig 15:26 xerox joined
Limbic_Region hrmmm - my make smoke output looks awfully strange 15:32
um - pugs on Win32 is b0rk 15:33
pasteling "Limbic_Region" at 129.33.119.12 pasted "Win32 is currently b0rked" (17 lines, 1.8K) at sial.org/pbot/19411 15:34
Limbic_Region This smoke report isn't going to look too good 15:35
Limbic_Region does a rebuild just in case
anyone else having this problem (Win32 or otherwise)? 15:37
[particle] makes unoptimized pugs 15:38
Limbic_Region particle - you are Win32 exclusively or is it one of multiple environments you build pugs in? 15:42
[particle] i'm win32-msvc only for now
Limbic_Region ok 15:43
what version of ghc?
[particle] 6.4.1
Limbic_Region interesting to know if you get the same problem or not then
AS Perl or from source or other?
[particle] as 5.8.7 15:44
Limbic_Region begins the process of building from scratch and wanders off to get some lunch
bbiab - let me know what your results are
[particle] still compiling, will do 15:45
15:47 xinming joined
[particle] D:\usr\local\pugs>pugs -e"say 'hello'" 15:51
hello
15:55 iblechbot joined
Limbic_Region hopefully rebuilding from scratch will yield the same results 16:01
we do have a number of differences though 16:02
[particle] which perl? 16:04
ooh, actually i'm 5.8.6
Limbic_Region AS 5.8.8 (which I have patched to make Module::Build work with MinGW)
so I am on GHC 6.5, AS 5.8.8 (patched), and MinGW 16:05
[particle] you're all new, i'm pretty old
Limbic_Region heh
particle - the performance difference alone is worth going to GHC 6.5
[particle] sure, if pugs works with it :) 16:06
is compile time slower?
Limbic_Region it does, audreyt is the one who convinced me to upgrade
no - faster
[particle] compile+runtime faster? wow.
Limbic_Region and there really isn't any upgrade - you basically just untar it 16:07
well, the compile time speed difference might not necessarily be related to ghc
changes to pugs internals and the build process might explain that
[particle] wonders if he needs to upgrade to 6.4.2 before patching to 6.5 16:08
Limbic_Region no
the binary build is this simple - pull down the snapshot, untar it, rename your current ghc directory to ghc_6.4.1 or some such and rename the untared dir your old ghc 16:09
at least that's all I had to do
that way if you want to back out - all you have to do is rename some directories 16:10
[particle] hrmm. i suppose i want one ending in -mingw32
Limbic_Region actually - you want the one from the 19th since the Win32 builds are all b0rk after that date
let me get you an exact URL
16:11 elmex joined
Limbic_Region www.haskell.org/ghc/dist/current/di...w32.tar.gz 16:12
[particle] thanks! i'll let you know how the install goes 16:13
gaal try the latest one though, it's a 6.6 release candidate
Limbic_Region ok - since you are using MSVC you shouldn't need to patch AS when/if you upgrade to 5.8.8
gaal - have they fixed the Win32 problem - I don't think so 16:14
gaal fwiw my experience shows pugs compilation to be a little slower on 6.5, but the resulting pugs is *way* faster.
Limbic_Region: if not please submit a report to the ghc team, this is a release candidate!
16:14 Psyche^ joined
[particle] the difference in compile time between msvc 6 and 7.1 for parrot is ~50% 16:14
gaal in favor of? 16:15
[particle] 7.1 is slower, but the parrot exe is definitely faster
Limbic_Region gaal - I only use binary builds and I think that is specifically where the problem is
I am sure the source builds work fine
[particle] is there a link for a binary build of the 6.6 rc?
gaal Limbic_Region: I understand, but seeing as this is a release candidate...
Limbic_Region they just aren't including ghc.exe for instance in the binary build
[particle] if so, i'll stop my current d/l 16:16
Limbic_Region gaal - it is easy enough to tell, just check to see if ghc.exe exists in the tarball
gaal [particle]: try the same dir Limbic_Region linked you to, just not that file.
[particle] k
Limbic_Region gaal - fwiw, I did tell the folks in #haskell - no one seemed too interested
gaal Limbic_Region: oops, the build moose is my bad. please rm blib6/lib/*yml 16:17
[particle] notices there's a ghc hackathon in portland in two weeks
16:17 Eimi joined
Limbic_Region gaal - well, the rebuild from scratch is almost complete 16:17
gaal [particle]: yes, and there's a ICFP soon too
Limbic_Region if realclean didn't do that for me - will be sure to do that next 16:18
gaal realclean kills blib6/, I think (it should!)
Limbic_Region gaal - WRT 6.5 being a little slower - then the overall build process has improved because before 6.5 it used to take me a minimum of 45 mins to rebuild from scratch
it is down to 30 minutes now 16:19
Limbic_Region tests pugs -e "say 'hello'"
gaal the difference being? just the compiler?
[particle] and the revision of pugs, i presume
Limbic_Region no way to isolate the difference
right particle
gaal because there have been improvements in pugs build time. 16:20
Limbic_Region: of course there's a way to isolate the improvement, try same r with different ghcs
Limbic_Region as I said before, the improvements may be completely unrelated to ghc - but I noticed the difference at around the same time
[particle] so i should see the speedups even with 'unoptimized'?
Limbic_Region gaal - no way to tell that I am *that* interested in pursuing
16:20 Eimi joined
Limbic_Region gaal - fixed 16:21
Limbic_Region starts a new make smoke
gaal [particle]: my small benchmark didn't include an unoptimized comparison, sorry :)
[particle] :P
Limbic_Region still can't figure out why t/examples/all_parse.t won't work under the test harness other than the fact that it has to do with $debug being turned on for File::Find 16:23
[particle] +o+
Limbic_Region I can't see where the debug information is coming out on STDERR and I would assume printing to STDOUT would be ok
Limbic_Region wanders off to a cow orker's going away party 16:25
[particle] gaal: are you moosing \u2559\u263b\u255c ? you should
Limbic_Region bbiab
svnbot6 r12926 | gaal++ | * Limbic_Region's breakage report makes me understand audreyt's
r12926 | gaal++ | comment about the CompUnit version needing a bump.
gaal ā•™ā˜»ā•œ! 16:26
[particle] my irc client can't display that :(
gaal ā•™ā˜ŗā•œ 16:27
or the ASCII fallback: }:)
gaal envisions a Perl 6 parsing combinator library called Moosec 16:28
16:30 Psyche^ is now known as Patterner
xerox loool 16:30
Moose :D
gaal hello xerox!
xerox Hi!
[particle] moose is a palindrome, if you turn your head and squint 16:31
16:31 turrepurre joined
[particle] hrmm, the ghc-6.5.20060901 has a gcc.exe, is that expected? 16:32
gaal yes
all win32 ghcs bundle gcc
[particle] well, i don't see ghc.exe
oh, there it is 16:33
gaal see README
16:37 lollan joined
[particle] hrmm, makefile.pl seems to be failing 16:44
Updated to revision 12926. 16:45
D:\usr\local\pugs>makefile.pl
Global symbol "$pcre" requires explicit package name at D:\usr\local\pugs\Makefile.PL line 522.
oh, i'm conflicted 16:46
16:51 xerox joined
Limbic_Region gaal - today 16:52
err
gaal - today's snapshot having the ghc.exe means they fixed the build problem
good stuff
gaal yes indeed.
Limbic_Region might upgrade himself next week at work and this weekend at home 16:53
[particle] rebuilds pugs with latest ghc
16:55 justatheory joined
Limbic_Region gaal - should I expect a major difference between the snapshot from the 19th and today? IOW, was the 6.6 branch merged recently or had it been integrated for a while? 16:55
gaal difference in what sense? 16:56
I'm using one from the 18th myself (linux)
Limbic_Region gaal - just what I said WRT 6.6 16:57
IOW - did the 6.5 and 6.6 branch and then get merged
or has the 6.6 rc always been part of the 6.5 branc
branch
gaal I don't know, but since this is the first release for 6.6 presumably 6.5 is where all the stuff's at. 16:59
but I never even looked at ghc source control, so better ask someone more into it.
[particle] wow, a parrot/perl6 question on p6l
i find nothing in S04 to say whether nested statement modifiers are allowed on blocks 17:01
gaal anyone here know ruleQualifiedIdentifier and ruleTrait etc.? I don't understand who ruleTrait supposedly services things like 'is export', because it ultimately calls 17:02
ruleVerbatimIdentifier `sepBy1` (string "::")
IOW, 'is export' won't parse, but say 'is MyProj::Moose' would
*understand how 17:03
Limbic_Region gaal - it isn't that important to me WRT significant changes being merged in GHC. I just wanted to know if there was something spectacular I should be on the lookout for. Wait and see I guess 17:04
Limbic_Region checks the results of his most recent smoke to see how things are doing 17:05
94.89% ok 17:06
gaal oh, no: sepBy1 just means ruleVerbatimIdentifier must match once or more.. but a trace shows it never does. hmm 17:09
17:09 weinig is now known as weinig|away
[particle] D:\usr\local\pugs>setup configure 17:12
Configuring Pugs-6.2.12...
setup: cannot satisfy dependency HsJudy-any
??
gaal [particle]: you may have to 'make' once w/o 'fast'
17:12 miyagawa_ joined
gaal you can interrupt once deps have been built. 17:13
Limbic_Region setup configure?
svn up; perl Makefile.PL; nmake;
17:15 nothingmuch joined
Limbic_Region how the moose are you nothingmuch? 17:16
[particle] my command was C<nmake clean && Makefile.PL && nmake>
Limbic_Region ah 17:18
the setup.exe invoked by Makefile.PL
gaal [particle]: perl Makefile.PL I hope 17:19
Limbic_Region gaal - why?
windows file association
gaal ahhhhh win32
right
wow, w32 golf 17:20
[particle] win32: 1; linux: 0 ;)
gaal [particle]: by all means keep using windows if you prefer it
Limbic_Region particle - did you get the build working? I don't know about you but I don't trust "clean" and always opt for "realclean" 17:21
[particle] i'm doing realclean now
TreyHarris how do you use a perl5 module in pugs? I successfully used perl5:IO::Prompt, but I can't figure out how to call prompt(). it isn't directly imported, and IO::Prompt::prompt gives me an undefined subroutine err 17:23
gaal TreyHarris: the hack is IO::Prompt.can('prompt'), IIRC 17:24
svnbot6 r12927 | gaal++ | * Signatures: fix access traits, e.g. :($x is rw)
r12927 | gaal++ | * cleanup a little, golf a little
TreyHarris gaal: ah, ok, thanks. unsurprisingly perhaps, prompt screams about not recognizing its context ("Are you trying to call me from inside a tie handler?") 17:26
gaal hm, a hack may be required, wrap it in a p5 closure? 17:27
TreyHarris yes, i think so
but before perl5: worked, I was toying with porting IO::Prompt anyway. maybe i'll just do that 17:28
gaal we had loads of ports once in the pugs repo, then they moved, and I don't know if anyone is still maintaining his any more
TreyHarris oh, i was actually just going to ask you about that, where they went and if there was a place in the pugs repo for such 17:29
as i wanted to write a util/ that would use it
gaal i think.. freepan?
google for memcached and perl6
TreyHarris i found it, i think: tpe.freepan.org/ 17:31
gaal yes that's it
TreyHarris but a util that uses something outside of the pugs repo would probably be bad form
gaal it's actually a good excuse to put your code in ext/ :) 17:32
TreyHarris gaal: good point. somebody ported one of my cpan modules, i see; since it's out of date and also a very literal translation, i think i'd like to do it myself again 17:34
17:34 [particle] left
svnbot6 r12928 | gaal++ | * golf, golf, golf is good, golf is great yeah 17:34
17:34 [particle] joined
Teratogen are perl 6 functions in builtin pseudo-package-classes? 17:35
builtin functions that is
gaal Teratogen: see Prelude.pm
Teratogen ok
thanks
where can I find that, gaal? 17:36
17:36 Daveman joined
gaal Teratogen: svn.openfoundry.org/pugs/src/perl6/Prelude.pm 17:36
Teratogen thanks again
gaal np
if you're asking about things that aren't there, say.... &say, then I think the answer is that "they should be!" but not always are 17:37
see also S29
perlcabal.org/syn/S29.html
lambdabot Title: S29
Limbic_Region which is still mighty confusing
Limbic_Region thinks the dev.perl.org/perl6 link for S29 should not point to what it does now 17:38
oh, nevermind me
it has been updated at least to refer to the one in the pugs repo 17:39
TreyHarris gaal: do you have any insight as to why literal lists of hashes segfault pugs unless used in an eval?
gaal ahem
example?
TreyHarris ?eval ({a => 'a'}, {b => 'b'})
Limbic_Region TreyHarris - example using ?eval would be even better 17:40
nevermind
17:40 evalbot_12918 is now known as evalbot_12927
evalbot_12927 user error (Incompatible version number for compilation unit Consider removing blib6/lib/Prelude.pm.yml and make it again ) ({("a" => "a"),}, {("b" => "b"),}) 17:40
TreyHarris oh... that was working last night on evalbot
gaal works here
evalbot needs cleaning its .yml, sec
TreyHarris ah, maybe audreyt fixed it since last night 17:41
Limbic_Region is a day late and dollar short on everything today
TreyHarris i'll update and make
Limbic_Region she's like that
always going and fixing things
TreyHarris Limbic_Region: yeah, but usually she mentions it on the channel :-)
[particle] the S29 link should actually be a hyperlink
gaal aaaaagh 17:42
Juerd: ping, please add me to sudoers
pasteling "[particle]" at 144.81.84.207 pasted "ghc-6.5 pugs on win32 is all moosed up" (523 lines, 27.1K) at sial.org/pbot/19412 17:43
TreyHarris hmm... only changes are yours, gaal. i'll re-make, but unless your changes fixed it as a side effect, it may be a OS X-specific issue
Limbic_Region disagrees with particle 17:44
or rather - suggests particle indicate which snapshot the 6.5 was taken from
TreyHarris or a ghc6.4.2 issue.... 17:45
gaal Limbic_Region: he's using 6.5.20060901, as mentioned in the paste.
[particle] latest
right.
Limbic_Region gaal - I knew that, I just took issue with the headline 17:46
[particle] }:P
gaal [particle]: hmm, dunno what to suggest except nuke dist/ and try again? 17:47
what's FPS_Dummy anyway?
Limbic_Region hrm - yeah
I seem to recall unregistering fps something or other a while back as it was the only way to get pugs to build
17:47 xerox joined
Limbic_Region after some change - but don't recall which 17:48
gaal I'm not using that snapshot yet. could it be that that ghc comes with a good enough fps, so our fps builds a dummy instead? shooting in the dark here
Limbic_Region: this is compilation, not linkage
brb 17:49
Limbic_Region well, as soon as the current make finishes
I will try it with the latest GHC to see if I get the same problem
that could possibly eliminate the snapshot as being the problem
[particle] that would be quite helpful 17:50
Limbic_Region it could be up to 40 minutes or so before I know for sure it works - will know if it is b0rk faster 17:52
my work download speed is abysmal
[particle] actually, it took me a long time to d/l that file, too 17:53
it may be the haskell.org site that's slow
Limbic_Region well, my top speed is 30kb/sec
and norm is like 28 17:54
cmarcelo gaal: re fps dummy: you're right. Pugs.cabal prefer always depending on pugs-fps and let u/build_pugs.pl choose between the empty (for GHC 6.5, which have fps in the core) and non-empty for GHC 6.4 people.. 18:00
Limbic_Region particle - downloaded and untarred - svn upping now 18:03
18:05 pbuetow joined
Limbic_Region particle - if I am reading your nopaste correctly - you don't even make it past Makefile? 18:07
oh, nevermind 18:08
you must have not nopasted that
[particle] k
right, just the end
Limbic_Region build starting now
getting very close to the point yours blew up 18:10
interesting 18:11
mine failed too
but for different error
Setup.exe: cannot satisfy dependency HsJudy-any 18:12
[particle] did you realclean?
Limbic_Region yes
Limbic_Region tries something
will be another 15 minutes or so 18:13
ok, starting build again 18:20
18:20 soisoisoi joined 18:23 nainef joined
Limbic_Region approaching previous break point 18:24
18:25 xinming_ joined
Limbic_Region particle - confirmed, error is related to latest ghc snapshot 18:25
[particle] but you have a different error, right? 18:26
Limbic_Region not now
[particle] oh, great! 18:27
Limbic_Region you had gone from 6.4.1 to 6.5 which meant audreyt had already done something special for third-party
I went from 6.5 to 6.5 and had to convince third-party to play nice
once I did - I am getting exact same results
[particle] i assume audrey has started her weekend by sleeping 18:28
Limbic_Region assumes so too
18:28 iblechbot joined
Limbic_Region no matter though, just download the link I gave you before and use that directory 18:28
once the fix has been made for the most recent release
just rename directories again
Limbic_Region likes being able to switch between versions of ghc just by changing dir names 18:29
even easier would be setting the env variable but that is not as easy on *this* box
[particle] yep, i just modified my path statement
Limbic_Region well, that was fun - NOT 18:32
gaal cmarcelo: could you loook at Limbic_Region and [particle] build failure? I think you may have an idea how to approach it
cmarcelo I already did but didn't had a clue :(... by the build log from particle looks like ghc.exe doesn't compile anything, maybe testing if it can compile a helloworld, just to check if the binary is sane.. 18:33
Limbic_Region cmarcelo - got a command line 1 liner for me to try? 18:34
[particle] ...or a small nopasted sample 18:36
cmarcelo Limbic_Region: create a file hello.hs with
18:36 Lorn joined
cmarcelo main = putStrLn "haha" 18:36
and then ghc hello.hs -o hello 18:37
18:37 onsen joined
Limbic_Region works fine 18:37
[particle] here too 18:38
D:\usr\local\pugs>ghc hello.hs -o hello.exe
D:\usr\local\pugs>hello
hello, world!
gaal [particle]: that's a bug, it should have printed "haha". I do not know what could have caused it.
[particle] the bug is, my hello.hs is different :)
Limbic_Region reminds particle that it is Friday and not to kick gaal 18:39
cmarcelo ok.. (are you testing with the bad-latest-ghc rifht?)
[particle] licks gaal
Limbic_Region has no idea WTF that means cmarcelo
gaal cmarcelo: they both have today's 6.6 release candidate
Limbic_Region we are both using the latest binary build (from today at around 9AM)
cmarcelo i meant the release candidate :) 18:40
gaal Limbic_Region: good to know there's a day in the week you're not violent in
cmarcelo rifht => right..
Limbic_Region I currently have 6.4.2, 6.5-2006_08_19, and 6.5-2006_00_01 to play with cmarcelo
cmarcelo - interestingly enough, the binary snapshots for Win32 had been b0rk since the 19th
[particle] and i have 6.4.1 and current 18:41
Limbic_Region (missing ghc.exe for instance)
it looks like this was fixed around the 29th
that is based of the tar ball file size change 18:42
cmarcelo Limbic_Region: looks like the binary is sane.. i'm looking for the error msg on the web..
Limbic_Region the difference though is that on the 19th with a working ghc/pugs the file was 61mb - today it is only 41
svnbot6 r12929 | gaal++ | * Parse type constraints in Signatures:
r12929 | gaal++ | :(Moose $x)
r12929 | gaal++ | * Commented out is support for:
r12929 | gaal++ | :($x where { $x % 2 })
r12929 | gaal++ | ...which requires an Exp -> Pugs.Val.Code convertor.
Limbic_Region cmarcelo - so with 20 mb difference, I am still not convinced the build is 100% 18:43
though I could just be throwing stones at the unknown
cmarcelo Limbic_Region: check if the files the pugs build complain missing (Stg.h HsBase.h) are in the RC tree..
s/check/would you mind checking/
Limbic_Region no worries - will then confirm the same (present or missing) from the 19th's build 18:44
18:44 maard left
cmarcelo Limbic_Region: recently they dropped some libs out of the core.. I think they should be distributing a *-extralibs-* together with every snapshot.. 18:45
Limbic_Region cmarcelo - good news, both files are present in both snapshots
cmarcelo - bad news, they aren't in the same paths between snapshots
[particle] erk 18:46
Limbic_Region I think this might be an easy short-term fix
Limbic_Region tries something
cmarcelo Limbic_Region: well, I suppose you're trying linking or copying them around.. 18:47
Limbic_Region for now just copying 18:48
Win32 doesn't support linking
but if this does work
then it will just be a matter of adding an -I
or whatever Pugs/GHC needs to look for header files in additional directories
gaal Limbic_Region: can you paste the two paths here? looks like it might be a reportworthy bug in the distro 18:49
Limbic_Region sure
ghc_root/include
18:49 onsen_ joined
Limbic_Region is the working path where ghc_root should be obvious 18:49
ghc_root/lib/i386-mingw-unknown/include 18:50
gaal did you follow the steps in INSTALL? on linux at least you need 'make in-place'
Limbic_Region is the current broken path
gaal - no need on Win32
gaal ok
Limbic_Region though that may be interesting - it might ultimately be that it is necessary and that the build on the 19th was b0rk 18:51
gaal fwiw: on my system (with the older snapshot), it's in lib/$platform/include too, and it works.
Limbic_Region appears to be working
though still very early in build process
gaal Limbic_Region: could you write your findings up and post them to [email@hidden.address] please? 18:52
Limbic_Region gaal - I have 8 minutes before work gets out
gaal they too are nearing release and will appreciate it
okay, I'll do it then 18:53
Limbic_Region but particle should be able to
gaal - new build error Setup.exe: cannot satisfy dependency mtl-any
going to try one more thing
then going home
ok, gaal - positive the make or make in place not necessary on Win32 as no configure present in the snapshot 18:55
so it basically boils down to root/include vs root/lib/$platform/include
but since we still aren't building (though further along)
posting to bugs is a bit premature
what exactly is mtl-any anyway? 18:56
cmarcelo monad transformer library 18:58
TreyHarris what's the incantation for regenerating the yaml prelude, and is it documented anywhere?
Limbic_Region is that one of the libraries that got dropped?
and if so, where can I get it from to install manually 18:59
cmarcelo Limbic_Region: not sure, they were discussing whether include or not..
get the extralibs for RC then
the error now is for missing mtl ?
(only)
19:00 Yappo joined
Limbic_Region cmarcelo - that's the only *current* error - doesn't mean fixing that won't point to a new one 19:00
TreyHarris i tried this: ./util/gen_prelude.pl --inline src/perl6/Prelude.pm
got errors
Limbic_Region cmarcelo - same download location for the extralibs?
cmarcelo Limbic_Region: same d/l location 19:01
gaal TreyHarris: util/gen_prelude.pl -v -i src/perl6/Prelude.pm -p ./pugs --output blib6/lib/Prelude.pm.yml
Limbic_Region cmarcelo - all that is provided is src (no binary) so I hope all that is needed is ghc.exe to compile them 19:02
cmarcelo yes
Limbic_Region great
gaal - going to build these extra libraries and try again
gaal TreyHarris: this is what build_pugs does. if can spare some tuits, a flag to build_pugs to do just that would be great, as well as a Makefile.PL rule to invoke that
Limbic_Region if it works then it will be two things - neither of which is necessarily a haskell issue 19:03
cmarcelo Limbic_Region: build instructions => www.haskell.org/pipermail/glasgow-h...10968.html
lambdabot Title: ANNOUNCE: GHC 6.6 Release Candidate, tinyurl.com/nh2rm
Limbic_Region thanks cmarcelo
Limbic_Region should be on his way home now
cmarcelo did lambdabot create tinyurl? cool! =) 19:04
Limbic_Region cmarcelo - there are no instructions there - perhaps next in the thread?
cmarcelo "unpack the extralibs tarball on top of the source tree to add them, they will be
included in the build automatically."
Limbic_Region but I am not building 19:05
Win32 is binary distro
[particle] perhaps extralibs should be included with binary distros automatically
cmarcelo hmm;.. maybe.. but I think the binary dist maybe ready to compile this extralibs.. ii'm d/ling it to check.. 19:06
Limbic_Region cmarclo - there is a makefile so perhaps I can still do it 19:07
Limbic_Region tries stuff all willy nilly
TreyHarris gaal: will try to find the tuit. in the meantime, documented it in TASKS.
gaal: thanks
gaal TreyHarris++ 19:08
svnbot6 r12930 | trey++ | [TASKS]
r12930 | trey++ | Documented need to add gen_prelude flag to build_pugs and Makefile.PL
TreyHarris ?eval ({a => 'a'}, {b => 'b'})
19:08 evalbot_12927 is now known as evalbot_12929
evalbot_12929 user error (Incompatible version number for compilation unit Consider removing blib6/lib/Prelude.pm.yml and make it again ) ({("a" => "a"),}, {("b" => "b"),}) 19:08
Limbic_Region cmarcelo - fwiw, they did decide to remove it as it is in the extralibs
gaal TreyHarris: still works :) 19:09
TreyHarris still segfaulting me. dunno how to write a test for it though
cmarcelo Limbic_Region: i just saw it.. well, i gotta go, but people in #haskell can help you building the extralibs...
Limbic_Region k 19:10
will wander over there
gaal bbiab 19:11
svnbot6 r12931 | gaal++ | * Pugs.Parser: parse out sigils for labels correctly, and some cleanups. 19:12
cmarcelo bbl & 19:14
19:17 onsen joined
TreyHarris is there a pugs variable to tell me the GHC version? 19:17
gaal no, but it's worth adding one :) 19:22
it could be done in PugsConfig.pm 19:23
19:30 cm joined
gaal :($x where {...} is elk) -- legal? is the trait on $x or on the where constraint? 19:32
:($x is ro, is rw) -- legal? is $x ro or rw? 19:33
(do commas separate traits?
)
19:38 crem joined 19:50 elmex joined
Limbic_Region gaal - it appears no response in #haskell so I must go for the weekend 19:51
will likely pop back on from home
TTFN
19:54 mdiep joined 19:58 cm_ joined
gaal Limbic_Region: a proper bug report (to the tracking system or via email) is pefered. 20:08
audreyt: Pugs.Val.Code.Param does not, I think, have a way to express the ::T in Dog ::T $fido.
zzZ& 20:09
20:19 meppl joined 20:22 ruz_ joined 20:23 Limbic_Region joined 20:25 arguile joined 20:33 mauke_ joined
Limbic_Region gaal ping 20:39
audreyt ping
svnbot6 r12932 | trey++ | Made a util/test_loh_segfault.pl to try to 20:42
r12932 | trey++ | nail down this list-of-hashes weirdness.
r12932 | trey++ | Please try it.
TreyHarris please try the above and let me know if it's broken... if not, i'm going to post it to p6-u to try to collect some data 20:43
TreyHarris &'s for an hour or so
20:48 mauke_ is now known as mauke 20:52 ezraburgh joined 21:01 justatheory joined 21:02 scw joined
Limbic_Region gives up and decides to build GHC from source 21:06
Limbic_Region shakes his fist in the air first
cm_ on win32?
Limbic_Region yes 21:07
the current binary build is b0rk
and I know why
cm_ awesome, are you gonna share the binaries once you're done? :D
Limbic_Region cm_ - you can download the binaries from the net 21:08
unless you want today's binary build which is broke
I can't get any help on #haskell
cm_ GHC 6.6 / win32? where?
Limbic_Region cm_ - no, that's the release candidate that's broke
PerlJam Limbic_Region: starting work on Strawberry and Vanilla pugs? ;)
Limbic_Region you can get 6.5 though
PerlJam - not exactly
21:09 crem_ joined
cm_ ah, hmm.. 21:10
Limbic_Region problem is that 6.6rc requires happy 1.15 to build 21:12
and there are only Win32 binaries for 1.13
which means I need to build happy from source
and that is proving to be a bit difficult
cm_ is happy using c? 21:13
Limbic_Region no - ghc
circular dependency
cm_ can't you just build happy using an old ghc?
Limbic_Region fortunately that's not a problem
cm_ 6.4
Limbic_Region actually - 6.5
but yeah
21:14 sili-babypuncher joined
Limbic_Region but I am getting a "LALR.lhs:626:34: Not in scope: `bounds'" 21:14
so perhaps I am not going to build from source
sili-babypuncher audreyt: present? i've got a problem trying to install Language::MzScheme: sial.org/pbot/19413 21:16
lambdabot Title: Paste #19413 from "sili" at 65.119.25.146
21:17 dolmen joined
sili-babypuncher maybe it has something to do with me not being able to find a scheme.h 21:18
21:18 lichtkind joined 21:21 fglock_ joined 21:24 DaGo joined 21:36 integral_ joined 21:40 fglock_ left 21:41 integral_ is now known as integral 21:55 markstos joined 21:59 cjeris left 22:08 takanori_ joined 22:10 crem joined
svnbot6 r12933 | markstos++ | refine Regexes and Rules section in Differences.pod 22:12
22:15 weinig|away is now known as weinig
svnbot6 r12934 | markstos++ | minor syntax fixes in Differences.pod 22:15
Limbic_Region gaal ping 22:40
audreyt ping
cm_ ping
22:55 xerox joined 23:13 jferrero joined 23:18 thomason joined 23:19 thomason joined 23:22 diakopter joined 23:25 Lorn joined, thomason is now known as athomason, sili-babypuncher joined 23:26 plisk_ joined 23:27 Alias_ joined
Alias_ seen TimToady? 23:27
(dammit when will you people get something that supports seen) 23:28
wolverian ?seen timtoady
lambdabot timtoady is in #perl6. I don't know when timtoady last spoke.
wolverian well, that was useful
Alias_ ... and works
wolverian he hasn't spoken here in a few days, or so.. not in my backlog anyway. sec
Alias_ Must be caught up at YAPC::EU
ok, thanks 23:29
wolverian 2006-08-31 13:35:38 TimToady : which is how you do a sublanguage...
that's the latest entry in my log
23:30 penk joined
Alias_ is trying to work out how to tell the different between C< { a => 1 } > and C< { print 'Hello World!' } > in Perl 5 23:30
i.e. Telling an anon hash-ref from a block
diakopter my guess is that it's assumed to be a block unless it could be parsed as either an empty hashref or a hashref with at least one pair. 23:34
Limbic_Region audreyt - if you backlog - see paste.lisp.org/display/25124 23:35
Alias_ diakopter: The tricky bit is how does perl tell, and can I simulate it closely enough 23:36
Limbic_Region I worked with #haskell to get Pugs to almost compile with the current binary distro (6.6 rc)
diakopter a hashref will always be part of, or fully define a single expression, right?
whereas a block can't appear inside of an expression, right? unless its marked by sub or something? 23:38
please excuse my naivete
Alias_ Well, more subtle 23:39
Blocks can appear inside statements
However a block on it's own is one variation of a "compound statement"
Specifically, it is the smallest form of LABEL BLOCK continue BLOCK
Where "compound statement" includes if and while and so on
So C< { print 1 } > is quite different to C< do { print 1 } > 23:40
The former, the block is the statement, the latter the block is within the statement
diakopter okay, so control structure markers in addition to subs
Alias_ Because in the latter you can do { block } or do { block } etc
It's more like do( sub { ... } ) 23:41
There's a lot of places something is a block, but not lexically significant 23:42
grep { }
et al 23:43
23:47 plisk__ joined
diakopter Alias_: are all the working perl golf submissions part of the PPI test suite ... heh. 23:48
Alias_ patches welcome
diakopter Alias_: do the cpan's modules have metrics that indicate how much ram they use when loaded into various perls (mod_perl, etc)? 23:50
Alias_ No, and it will vary enormously 23:51
And even with your version of Perl, as various structures change size
diakopter well sure, but I mean with a mere "use" 23:52
Alias_ That requires compilation
diakopter surely there's some kind of rough relative scale
Alias_ And has other issues, like do you count deps etc
I do it by hand for a few of the ::Tiny modules
diakopter i was imagining not counting deps. exclusive usage.
23:56 spoop joined 23:57 azr|elious joined