6.2.7 released! | pugscode.org/ <Overview Journal Logs> | pugs.kwiki.org | paste: sial.org/pbot/perl6
Set by autrijus on 13 June 2005.
meppl gute nacht - good night 00:41
svnbot6 r4907, autrijus++ | * better handling of modules and qualified names in PIR generation. 00:47
autrijus integral: still there? 00:51
nothingmuch greetings from the autrijus & nothingmuch hackathonon 01:35
oh my, autrijus just took his laptop to the shower 01:36
kelan mini-hackathon?
nothingmuch: what happens in the 7th and 8th chapters of harrorth? fun things? 01:37
nothingmuch well, hackathon starts tomorrow, but we arrived a day early
kelan ah
autrijus rehi \camels :)
nothingmuch i'm 40 something hours without sleep, so remind me what the titles of 7 & 8 are
kelan let me check 01:38
chapter 7 starts "Forth for Real"
nothingmuch ah
that tries to model forth more closely
we're going to have another chapter like that
i've tried doing the next one 3-4 times already
kelan and 8 is "Implementing Our New Knowledge"
nothingmuch each time i backed out of my strategy 01:39
oh, i see
so 7 is a forth primer
discussing "true"er forth
regardless of harrorth
autrijus the one-level-up-truth
kelan nods
nothingmuch and chapter 8 tries to take some of the ideas of chapter 7 and impelment them
autrijus "the truth... is up there" or something
nothingmuch to an extend it does it very well... : is a word, ; is an immediate word, etc 01:40
kelan any plans for the hackathon yet?
nothingmuch i guess "we'll see" is probably the best strategy
what autrijus is trying to say is that i just told him that every time I try doing soemthing forthish my forth guru shows me yet another neat trick
a sort of zen slap 01:41
there's no shortage in work
autrijus sees no complimentary towels nothingmuch talked about 01:46
nvm, got mine
nothingmuch wonders if porting Devel::FIXME to perl 6 will be with macros or with hooks 01:52
geoffb reappears 01:58
autrijus.ping('OpenFoundry bug') 02:01
autrijus geoffb: looking into it now 02:04
finally got connection
geoffb thanks!
autrijus blames toronto and dusseldorf airports for not having wireless
nothingmuch budapest had nice wifi 02:05
geoffb bastards. :-)
autrijus something really fishy is going on with openfoundry's nfs
nothingmuch i rsynced 400mb of comic books from my home box
autrijus sec
nothingmuch then i read one
and felt too sleepy for the rest 02:06
it was a good plan for passing the flight, but I just wasn't up to it =(
mugwump (paper books)++
nothingmuch mugwump: indeed 02:11
i think that and the tripod are the bulk of my suitecase
autrijus geoffb: too silly 02:42
the Rationale field is >255bytes
and hence borked the RT thing
underneath
with no good errmsg
geoffb: so I trimmed it to one line 02:43
and lo it worked
;)
geoffb: so, enjoy -- I'll trim things in the code now 02:44
geoffb: fixed as r1173 02:48
svnbot6 r4908, autrijus++ | * remove debug messages 02:55
abez ? eval 1+1 02:57
?eval 1+1
seen Darren_Duncan 02:59
jabbot abez: Darren_Duncan was seen 1 days 7 hours 33 minutes 41 seconds ago
abez ?eval $x=3; $x 03:01
geoffb autrijus, sorry, was AFK putting kiddo to bed 03:02
Lemme check if the project appears on my page
yep, it's there 03:04
Bloody stupid that RT barfed on a too-long field
geoffb chuckles at the random shuffling of Chinese and English 03:06
Bah, anyone know which fonts to select for Debian's Firefox package to get fairly unicode-complete fonts? I'd rather see the nice soothing Chinese characters rather than a bunch of hex codepoints all over the page . . . . 03:09
mugwump try ttf-arphic-* 03:10
revdiablo geoffb: there's a gigantic arialuni.ttf you can find that has a lot of unicode characters in it 03:12
hmm, seems that it comes with MS Office
mugwump all 23MB of it :) 03:13
geoffb Well, MS Office won't work for me, since I don't have (easy) access to it . . . but trying Arial off the menu just in case 03:14
hmmm, nope 03:15
Khisanth hmm I have to zoom to 200% 03:16
mugwump my unicode support ā™™ā™Ÿs !
Khisanth mine too 03:17
geoffb mugwump, what is that supposed to show?
Khisanth but apparently not things like Chinese!
mugwump That's two chess pieces from plane 0x26
pawns
geoffb ah
OK, mugwump, I know I want Unicode instead of GB or Big5 versions of ttf-arphic-*, but what's the difference between ttf-arphic-ukai and ttf-arphic-uming? 03:18
<rant>And why doesn't installation of a desktop not depend on a good set of unicode fonts in this day and age?!?</rant? 03:19
mugwump I think that the difference comes down to the style of the characters
geoffb what westerners would think of as a typeface, or something orthogonal to that? 03:20
mugwump sure
geoffb LOL
Or as my father used to say, "Why yes." 03:21
mugwump ęˆ‘čƒ½åÆ«äø­ę–‡
That should have been 5 characters 03:22
geoffb Well, in xchat I see it as 5 codepoints, but I'm working on apt-get'ing more fonts now. 03:24
:-)
revdiablo I see 3 boxes and 2 characters, I guess that means I have irssi mostly behaving. :) 03:25
mugwump ęˆ‘čƒ½å†™äø­ę–‡ # what about that? 03:27
jql works for me 03:28
mugwump actually only one character is different there, the middle one the simplified character "写" is used instead of traditional "åÆ«" (xÄ­e, to write) 03:30
geoffb selects a few random fonts from the boggling list from `apt-cache search xfonts` and gives his net connection a thrashing 03:33
revdiablo weird. that time I got 4 boxes, one character, then another box
jql those glyphs all work fine for me
jql hugs ubuntu
revdiablo jql: hmm, what font packages do you have installed? 03:34
I wonder if there are Real Debian (tm) counterparts
geoffb And more importantly, *which one are you using*
revdiablo had my thought first 03:35
jql just the base install of kubuntu with the devel kit and firefox (which means +gtk)
geoffb :-)
What font are you using in your chat client?
(or *term, as the case may be)
jql client. utf-8 encoded bitstream vera 03:36
geoffb hmmm, checking
mugwump fixed (6x13) and 9x15 have the best unicode coverage of the fixed fonts, apparently
jql and the client is utf-8 encoded, not the font. :)
geoffb tries Bitstream Vera with firefox while ttf-arphic finishes installing 03:39
jql I have Big5, BT, and Batang fonts installed, too
wicked
mugwump yeah, I tend to install the lot, too.
jql GB, not BT 03:40
bizzare thinko
mugwump batang?
jql forean iirc
s/f/k/
coral bat-tang 03:41
geoffb brb 03:46
mugwump, can you repaste those characters? 03:51
mugwump ęˆ‘čƒ½åÆ«äø­ę–‡, ęˆ‘čƒ½å†™äø­ę–‡ 03:52
geoffb oh, now that's MUCH better 03:53
coral ?????, ?????
geoffb And the difference between the two ttf-arphic fonts is: squarish v. handwritingish 03:54
jql spiffy
geoffb ah-ha! Now I understand why I couldn't see changes between fonts in the extended characters -- when you change fonts in xchat, if the new one doesn't support some glyphs that the old one did, it appears to keep the bitmaps generated for the old font 03:56
-- FOR JUST THOSE CHARACTERS
coral heh!
geoffb a usefullish feature, but damned surprising
jql hah. cute but evil 03:58
Khisanth hmm that is probably pango's doing
mugwump wŏ nĆ©ng xĒe HĆ nyĒš pĆ­nyĆ­n 03:59
wŏ nĆ©ng dă zƬ HĆ nyĒš pĆ­nyĆ­n, perhaps 04:00
Can you see those? All latin characters with various unusual accents, including a u with umlaut *and* upsidedown circumflex 04:01
geoffb Khisanth, it's the kind of thing that makes me want to figure out a "best order" to cycle through fonts to get the western characters *and* the unicode characters good from that point on
mugwump, yep
can now
Khisanth geoffb: I just use "Monospace" whatever font that is suppose to be :) 04:02
geoffb Khisanth, that was my starting point -- because I assumed that the package monkeys would Do The Right Thing. Unfortunately, they are also Only Human 04:03
Khisanth I mean in xchat, and have my charset set to utf8 04:04
geoffb mugwump, what is your natal language?
(I could swear I've seen you use several)
mugwump Geordie 04:05
geoffb blinks
mugwump way-oo man?
geoffb was not expecting an answer that would leave him more confused than before 04:06
mugwump Then I learned cockney, a year of floridan, followed by many years of kiwi
these are all English dialects ;)
geoffb I recognized all but Geordie. All I could think of was "Laforge?" 04:07
Khisanth was unfortunately thinking the same ...
of course the next thought that popped up was "wtf?" 04:08
geoffb NODNOD
jql doesn't recall Cmdr Laforge speaking another language... that universal translator rocks
mugwump yee mean yee divvent understand Geordie?
geoffb Favorite episode of TNG: when aliens with reference-only language beam captain down to random planet to force them to work through the fact that the universal translator couldn't fix their completely different languages 04:11
mugwump sounds like taking babelfish on for not being able to handle east/west translation :) 04:12
geoffb Any SVK gurus about? #svk seems to be quiet (AKA, no answer to question) 05:00
mugwump sadly no, they're all asleep or travelling to YAPC::na I think 05:03
just trollers like me about
revdiablo I know a bit about svk and svn 05:04
hardly would call myself a guru, but maybe I can help
mugwump cool ... you've seen the parrot SDL examples aye geoffb ? 05:19
geoffb Actually, still haven't got off my ass and compiled a parrot recent enough to include parrot SDL, despite having svk co'ed it a few days ago, sigh 05:20
But as I recall, chromatic hadn't done the OpenGL side of it yet. 05:21
Besides, the Perl 5 people deserve an actual working bugfree compatible portable version of SDL_perl . . . .
And as soon as they have it, I intend to mercilessly push for 'use SDL::OpenGL--perl5' to Do The Right Thing 05:22
:-)
mugwump sounds excellent. There could even be a PP6 (pure perl 6) SDL module that wraps whichever backend is available 05:24
mugwump can just keep on dreaming you know ;)
Something along the lines of freevo would be an excellent demo project 05:25
largely because Freevo is slow, inflexible and generally just bites... but looks great on the lounge TV
geoffb So we want to show them that it works better with Perl 6 than Python? 05:26
mugwump no, we want to show them that perl 6 coders are cooler than python coders. priorities. 05:27
geoffb :-)
geoffb is growing bleary-eyed staring at diffs for hours on end 06:40
Methinks the remaining merges will have to wait for morning 06:41
g'night, all
mugwump gah, didn't get a chance to fix up those broken Date::Gregorian tests ... sorry people :( 07:49
mugwump & # full moon party
10:23 Aragone is now known as Arathorn
nothingmuch good morning 10:45
autrijus yo
broquaint How's NA, autrijus? 10:48
clkao hello autrijus
autrijus very nice 10:51
turns out nothingmuch dreamed about parsing forth's higher order functions as bytes, and muttered something about parsing literals in his sleep 10:52
so the rest of the night I dreamed about optimising literals -- that is constant folding in the SSA/ANF dominance tree -- in the lambda calculus 10:53
...and hence started hackathon even before we ake up
;)
s/ake/wake/
nothingmuch hackathonon 10:54
broquaint Now all you need is a brain => haskell dump that doesn't require a keyboard ...
Or should that be, brain ==> haskell ? ;)
autrijus come to think about it, maybe I should get a Ghost in the Shell brain-port tattoo here
always wanted to get one
nothingmuch well, we talked about stuff when we arrived, and I don't think I was awake at the time either
brain >>= haskell
broquaint I figure brain-port tatoos will be available in Taiwan first, if anywhere. 10:55
autrijus www.justplay.nl/media/7/20040506-Neck%20GITS.jpg 10:59
# older model -- I wanted the more sophisticated panel-like one -- but something like that 11:00
broquaint Wireless would be ideal.
autrijus indeed.
nothingmuch nothingmuch.woobling.org/autrijs_wi..._shell.jpg =D 11:19
i kind of screwed up the fake gradient background, because I sampled the wrong color
sorry
autrijus lol, nothingmuch++ 11:20
nothingmuch properly (un)focused version online 11:27
(sort of ;-)
time for breakfast
kungfuftr autrijus++ # brain port tattoo would be fonq 11:45
stef__ salut slanning :) 12:03
slanning salut
autrijus greetings. 12:04
nothingmuch autrijus is "sleeping", he told me so 10 minutes ago 12:05
don't listen to him ;-)
autrijus I am! 12:06
autrijus is even snorting
stef__ who cares, you just sait your are programming in your sleep
autrijus that'd be correct
stef__ just need a plug to make sure the sleep work is indeed saved somewhere
autrijus I use a brain-local branch of svk 12:07
so I can commit to //hippocampus/...
slanning :)
autrijus wakes up 12:29
svnbot6 r4909, autrijus++ | * fix export aliasing and nesting in PIR compilation 13:14
14:40 Maddingue is now known as Maddingue_
dudley Does anybody know if there's a perl6 mode for emacs anywhere? 14:52
autrijus not afaics 14:53
dudley I was afraid of that. I guess I should start learning lisp... 14:54
QtPlatypus has been using the standard cperl that works with perl5, "Unfortunitly it brakes more then it works" 14:55
slanning dudley: excluding the docs at the beginning, cperl-mode.el is only about 7200 lines long 15:05
integral ~ 1/3 of pugs :-/ 15:06
dudley slanning: Yeah, I scrolled through it a couple of days ago. 15:08
I just wanted to make sure there wasn't something out there already.
QtPlatypus I wounder how hard it would be to make pugs dump its parse tree in the form of an s-expression so that emacs has a better chance of working out what everthing means. Otherwise the problem is going to be very hard.
slanning 8) 15:09
autrijus QtPlatypus: -CPugs uses haskell exp form
which should be trivial to s-expressionalise
or you can write a Emit.SExpr...
integral perl6->elisp 8-) 15:10
dudley QtPlatypus: I think the problem is going to be very hard anyway, considering I don't know perl6 or elisp yet ; ) 15:11
QtPlatypus dudley: Lisp isn't that difficult. Personally I just wish that elisp had lexical scoping. 15:13
slanning what do i install to get the "happy preprocessor" ? (needed during build of haskell-src-exts which is required for hs-plugins) 15:26
autrijus slanning: www.haskell.org/happy/ 15:29
integral autrijus: to use a different emitter in Pugs.Compile.PIR do I need to write a different Translate class? 15:35
The one in there atm uses the Trans monad which uses Stmt from Emit.PIR
autrijus integral: you need to factor out Pugs.Compile.PIR to Pugs.Compile, and then factor out Pugs.Compile.PIR into Pugs.Trans.PIR 15:36
and then write your own Translate class in Pugs.Trans.Perl5
making only use of Compile
instead of the Translate defined inPIR
essentially every translator defines its own Translate class.
integral ok, that makes sense :-) 15:37
autrijus glad I'm making sense :)
15:38 nothingmuch_ is now known as nothingmuch
slanning thanks, autrijus - now i have the prereqs installed, but it says "Pugs build should not get here" - is there something special like `make dist` to be done when using the checkout from darcs instead of from CPAN? 15:44
(it says that during `perl Makefile.PL` I should've said) 15:45
hmm, in fact it does the same thing from `cpan`, therefore "no" 15:51
(or not necessarily no, but irrelevant)
autrijus hrm. 15:52
weird 15:53
try checking out from subversion and see if it works?
autrijus wonders if there's a sync problem
slanning it's inc/Module/Install.pm line 29, included from line 8 of Makefile.PL if that says anything - but i will try checking out from subversion 15:54
is subversion the "standard" checkout, then?
autrijus it is. 15:55
but it should be identical from darcs
slanning hm, same thing :/ i must've skipped something somewhere 15:57
autrijus :/ 15:59
you have a shell I can try on? 16:00
never happened before
slanning no, i think the firewall prevents accessing from outside 16:07
cpansign -v also doesn't work (MISSING Signature file!) if that gives any hint 16:09
that's why i thought maybe i need to `make dist` or something
autrijus no... the missing signature is known for development version 16:22
slanning hello, david :) 16:26
nothingmuch ugh. I don't like it when people make things so much more complicated than what they should be 16:33
it's even worse when people make things more complicated than they already are.
broquaint Who needs --long-options -wsowd? 16:36
slanning ew my god, $FindBin::Bin is saying /home/slanning/bric/svn/rev_1_8/bin whereas i'm executing `perl Makefile.PL` from /home/slanning/pugs/pugs-darcs -- i think my debian perl upgrade today might've hosed something :/ 16:43
but perl -MFindBin -le'print $FindBin::Bin' works as expected -- hm.. 16:44
kungfuftr is there a multiline hash style comment available in perl6?
autrijus hash style? 16:47
kungfuftr ie: # This is a comment 16:49
would be nice if you could do something like "#**********\nThis\nis\na\ncomment\n#**********\n" 16:50
or something like a code block that was a comment 16:51
autrijus: if you clock what i mean 16:52
wolverian kungfuftr: =pod?
Khisanth heh
kungfuftr wolverian: pod must start at the start of the line
wolverian kungfuftr: true enough.
kungfuftr wolverian: =0) 16:53
wolverian kungfuftr: can you find a character series that isn't ambiguous in perl6 that could be used for an in-expression comment block?
kungfuftr #{ }
wolverian (I couldn't find one for perl5...)
hmm.
autrijus #{ #}
#{ }#
wolverian yeah, didn't think of using # in it.
sounds like a p6l issue.
autrijus I rather like that, yes.
kungfuftr *blink* 16:54
wolverian I rather like pod, myself.
literate programming and all that..
kungfuftr wolverian: pod is great, but i use pod for usage documentation and hash comments for when i really need to describe what's going on in this part of the code
wolverian kungfuftr: =for hackers 16:55
geoffb Whereas I HATE pod.
slanning i think pod is an abomination
geoffb slanning appears to agree
kungfuftr wolverian: breaks up indenting
wolverian geoffb: yes, me too. I love the concept, though.
kungfuftr: yes.
geoffb Yeah, the concept isn't bad -- pretty good, even. The syntax is awful. Worst syntactical mistake Larry every made, IMHO 16:56
oh dammit, brb
wolverian geoffb: good thing it's being fixed in perl6 then!
nothingmuch i think pod comments still apply
kungfuftr autrijus: i don't have proper net access at the minute... can you8 forward it to p6l?
geoffb: pod is being replaced with Kwid, iirc
nothingmuch needs to make sure to scroll down before answering questions 16:57
oh, whoops
kungfuftr pod comments for describing the ins and outs for a particular peice of code are BAD, m'kay
wolverian kungfuftr: why?
QtPlatypus thinks that there should be a documentation taint to subs. 16:58
wolverian QtPlatypus: the language that really uses that, python, seems woefully underdocumented to me. 16:59
kungfuftr wolverian: breaks indenting, will confuse pod translators and will completely screw auto-indenters
wolverian (in many places.)
kungfuftr: the latter two have never been a problem for me. honestly, never. the first issue is, of course, true.
kungfuftr hash-comments are for code, pod is for describing how to use that code
QtPlatypus wolverian: I've seen it used in Lisps as well.
wolverian QtPlatypus: I meant to say 'the language I know ...'
:)
slanning documentation taint? 17:00
wolverian kungfuftr: I like POD because it lifts comments to an equal standing with code. comments are not embedded in code, the file just alternates between the two! that is literal programming, to an extent.
slanning: I assume he meant 'trait'
QtPlatypus is assuming that I meant 'trait' as well. 17:01
kungfuftr wolverian: there's fairly often i need to describe how a particular regex works in the code for the rest of my team 17:02
bbl
wolverian kungfuftr: sure. but there's nothing forcing you to not use POD for it.
slanning i've still never heard of "documentation trait", but i think you mean like in elisp where you add a docstring as the first line of a function
wolverian slanning: something like that, yes. perhaps using perl6's trait functionality. 17:03
QtPlatypus slanning: Exactly.
wolverian slanning: if you don't know what traits are in perl6, read dev.perl.org/perl6/doc/design/syn/S12.html
slanning oh, it's perl6 vocabulary, that explains it :)
QtPlatypus Would make it possable to implement perl6explain (or whatever its going to be called) open ended. 17:05
svnbot6 r4910, autrijus++ | * detab; use Noop for END blocks
Khisanth nothingmuch: you keep missing that six in the URL :p 17:07
nothingmuch Khisanth: oh crap, that's not in the URL 17:08
that's just pbotutil with nothing
Khisanth ah
nothingmuch i thought it didn't announce anywhere, i'll revise my conf
Khisanth time to fix settings?
nothingmuch yes =) 17:09
Khisanth: ping 17:12
Khisanth hrm?
nothingmuch where is it normally announced? #perl?
11278 should not have been pasted 17:13
Khisanth yes
nothingmuch okay, seems like it didn't
Khisanth err didn't see a 11278 17:14
nothingmuch goody
autrijus nothingmuch: t/subroutines/arg_ref_type_preservation.t 17:21
svnbot6 r4911, autrijus++ | * So much for not compromising... change the := ?? :: line
r4911, autrijus++ | in Test.pm so it can work with PIR.
slanning hm, if i do `perl Makefile.PL` it fails because $FindBin::Bin is wrong, but if i `chmod 755 Makefile.PL` and `./Makefile.PL` it works - but `which perl` says "/usr/bin/perl" which is what the top of Makefile.PL says 17:27
does that say anything to anyone? i'm going crazy 17:28
autrijus slanning: your perl in path is not /usr/bin/perl ? 17:32
and do you have problem installing other cpan modules?
which platforms is this? 17:33
slanning `which perl` says /usr/bin/perl; Pod::Text installed okay from `cpan`, though probably it just doesn't use FindBin; debian, which i think is where the problem is from because the perl package updated today 17:37
hm, maybe `which perl` is lying
no, /usr/bin/perl is the only `perl' in my $PATH 17:39
geoffb sheesh, bak 17:40
autrijus, what did you change in 4911 (while I go `svk pull`)? And why? 17:42
autrijus geoffb: there's this line
$ok := $cond ?? "ok " :: "not ok "
which aliases $ok to a constant string
nothingmuch Khisanth: i saw my name in #perl, and i've been googling for logs for 3 minutes now, and found nothing 17:43
i closed the window before I could react though
autrijus and I gave up, after trying to ~2hr, to simulate it in parrot
nothingmuch so i missed whatever you said, if you said anything
autrijus geoffb: so I changed it to a simple if/unless
that's all.
geoffb gotcha
Khisanth I haven't said anything to you in #perl recently
nothingmuch i saw "nothing: " 17:44
or something like that... perhaps it was around 10 minutes ago, about the pbotutil
geoffb Well, honestly, as crazy as Test.pm does, I think it's got the wrong design ideal -- it should not *itself* be a test of language features. In fact, it should be just plain *boring*, because it's used to test the language features . . . .
Khisanth you sure it wasn't #perl6?
slanning I set PARROT_PATH=/usr/local/parrot-0.2.1 ; in Makefile.PL, `sub parrot_config', it basically chdir($PARROT_PATH) then `./parrot parrot-config.imc ...`; but `parrot` isn't in $PARROT_PATH, it's in $PARROT_PATH/bin (linked to from /usr/local/bin), while parrot-config.imc is in $PARROT_PATH 17:45
nothingmuch quite sure 17:46
geoffb That was part of where I was going with t/01-sanity -- eventually there should be a set of tests in there that include all the necessary functionality for a cleaner, meaner Test.pm.
nothingmuch but not sure it was "nothing: ", could have been just "nothing"
Khisanth ah that was PerlJam :) 17:48
slanning hm, i think i'm going to cut my losses (several hours) and give up - good luck
Khisanth 1119461459<PerlJam> Khisanth: nothing. That's another option.
integral *argh* PIL is type-indexed on Stmt and friends from Emit.PIR :-/ 17:56
nothingmuch Khisanth: yes, that was it 18:00
svnbot6 r4912, autrijus++ | * add Shlomi to AUTHORS for the readline bug report -- no idea 18:04
r4912, autrijus++ | why I didn't added him earlier
nothingmuch well, maybe because he didn't really help ;-) 18:05
svnbot6 r4913, bsmith++ | Moved the Compile class to Pugs.Compile. The interface to the compilers 18:44
r4913, bsmith++ | is in Pugs.Trans (for the moment at least).
geoffb D'oh! /quit != /leave 18:47
nothingmuch_ hackathon time 18:55
ciao everyone!
geoffb go nothingmuch_ !
Qiang hm. they got cable internet in the cottage? 18:59
svnbot6 r4914, autrijus++ | * 01-sanity/07-test.t is now working!
geoffb Go autrijus! 19:00
autrijus alright. I'm going to meel lwall at all 19:01
and stevan and other friends
autrijus waves &
PerlJam If "meeling" is anything like "mealing" I hope no one does that to lwall ;-) 19:05
svnbot6 r4915, autrijus++ | * `make pirsmoke` is now a reality...
integral
.oO( make BACKEND=PIR smoke if we have multiple backends )
19:06
PerlJam surely you mean "PUGSBACKEND=pir make smoke" 19:07
integral maybe :-)
Khisanth hmm 19:12
PerlJam: but don't some cannibals believe that when they eat other people, they are acquiring some traits and such from those they eat? 19:13
PerlJam Khisanth: aye, 'tis true. I still hope no one does it to lwall.
Khisanth Qiang: they don't NEED internet and I think it's something slightly slower 19:15
geoffb AAAAAUUUUUUGGGGGGHHHHH! I HATE WORKING WITH C! 19:19
I believe, deep in my heart of hearts, that gcc error messages were designed specifically to piss me off. 19:20
arcady and C hates working with you
Khisanth but C is fun!
geoffb The C I learned in college was fun.
Khisanth geoffb: they seem to have gotten better over the years
integral segfaults lead to hours of entertainment!
geoffb The C I have to deal with now is NOT
Khisanth wait not, gcc error messages are ok, g++ error messages are just cryptic 19:21
arcady C++ is just cryptic
and leads to such fun things as internal compiler errors
geoffb 'syntax error before "foo"' is driving me nuts.
How far before "foo", jerkoff?!?
arcady, bleah, good to know it could always be worse 19:22
stef_ Knisanth: i would say he would acquire some properties, not traits :)
depends if you consider eating as compilation or runtime... 19:23
geoffb is trying to get glprocs.c to compile in Debian, so that the Wayne's new SDL_perl automatic-GL-extensions-code will work on something other than Win32.
Khisanth don't think I have ever seen an "internal compiler error"
hopefully never will
integral GHC does those occasionally with pugs
geoffb Khisanth, I second that emotion 19:24
Khisanth geoffb: missing brace or semicolon somewhere? :)
geoffb Khisanth, that's what I'm thinking maybe, but I can't figure out where.
And these header files are so damned repetitive, they lull your eyes into missing stuff 19:25
svnbot6 r4916, bsmith++ | Moved Pugs.Compile.PIR to Pugs.Trans.PIR 19:34
geoffb breathes deeply 19:42
must . . . stop . . . fist . . . of . . . death 19:43
I'm thinking it's header file version skew
Khisanth more C?
geoffb yep
Khisanth what are you writing this C for anyway? 19:44
geoffb looks at the long picture:
If I get this to work, then SDL_perl 1.x will be resurrected stronger than before, more powerful, faster . . . .
And I can get back to trying to get SDL_perl and pugs to work nicely together. ;-) 19:45
Khisanth hmm
lightstep fight version skew: www.haskell.org/tmrwiki/EternalComp...tyInTheory 19:46
geoffb And really, I'm not writing the C. Just trying to merge Wayne's work in the Win32 world with current Debian package.
But Wayne unfortunately did some Win32-specific stuff, and my C merging skills are weak. 19:47
arcady I wonder what parrot's NCI would think of the opengl libraries
geoffb arcady, if I get this working, I might throw some energy at that as well. :-)
lightstep sorry, it's my first day, i don't even know what SDL is 19:48
nor NCI
integral there's already a ParNCI SDL binding :-)
arcady parrot's NCI is not very difficult to use, I think
lightstep oh, wait, i recall SDL
geoffb lightstep, SDL is a cross-platform, cross-language toolset for doing simple graphics stuff. NCI is Native Call Interface
ah
integral, yep: without OpenGL, last I heard 19:49
integral oh well :-)
geoffb And OpenGL is what I actually want to have working.
geoffb would really like to just port my 3D engine to Perl 6 and throw away all of the source filtering that gets me just partway to the happy place 19:50
But to do that, I need to get pugs and SDL::OpenGL to talk to each other. And to do that, I need to get recent SDL_perl to build on Debian. Which is where I currently am.
About this far --> <-- from yak shaving
khisanth_ so as I was saying ... 19:53
19:53 khisanth_ is now known as Khisanth
Khisanth so the latest SDL Perl doesn't actually work anywhere besides windows? 19:53
geoffb s/latest/latest 1.x/ 19:54
but otherwise, yes
Because that's all Wayne has . . .
I'm working to bring his work back to the masses
Um, did you see my reply? It's looking like you're having connection issue 19:55
s
19:55 khisanth_ is now known as Khisanth
Khisanth yes I did 19:56
geoffb k
OK, afk for a few to make sandwich and clear head, then back to try banging on this again 19:57
integral gives PIL a nasty look for being type-indexed on Emit.PIR types 20:12
stef Reading package info from stdin... ghc-pkg: Line 1: Invalid syntax (no colon after field name) # does someone get that when trying to compile ghc plugins? 21:14
hum. plugin.conf.inplace which is fed to ghc-pkg containes only "[]\n" 21:16
geoffb stef, I think I saw that when I was first trying to build. 21:29
Um, what was that . . . .
old ghc version?
stef nope, 6.4 21:30
delivered by mandriva thx to rgs
geoffb damn, what *was* that 21:31
sorry stef, can't remember for the life of me. I know I've seen it though, so it's not an unknown issue . . . just not a documented fix, I guess. 21:32
stef I see the same kind of error in the log of #haskell 21:34
geoffb stef, you know, I seem to recall having missed a prereq in the chain . . . check the READMEs again carefully, perhaps? Sorry I'm not giving better help, but I know somewhere in my brain there's the answer, and I'm kinda poking at it. :-) 21:38
Darren_Duncan ?eval 1+1 21:40
?pugs 1+1
mugwump jabbot: seen evalbot6? 21:41
jabbot mugwump: evalbot6 was seen 5 days 15 hours 6 minutes 17 seconds ago
Darren_Duncan I tried it yesterday within my local perl mongers meeting, to demonstrate it, and it didn't work ... alas 21:42
the meetings topic was on IRC etc bots
geoffb my @tasks_stack = < pugs+opengl pugs+SDL_perl SDL_perl-1.20.5 glprocs.c GLsdk-glxdemo >.map:{ "Get $_ to build on Debian" }; @tasks.pop # And there was much rejoicing . . . . 21:58
well, the latter was supposed to be @tasks_stack.pop, but oh well 21:59
Khisanth geoffb: finally got it to work? 22:15
geoffb I figured there might be a problem with Wayne's merge of GLsdk into SDL_perl. So I downloaded the raw GLsdk and went to compile that. After a few fixes, I got its demo to work, and with the knowledge gained there, am hacking on the previous problem again. 22:17
Wish me luck. :-)
BOO-YAH -- @tasks_stack.pop 22:37
ALL RIGHT! @tasks_stack.pop 22:55
@tasks_stack.push('Now that SDL_perl builds and installs on Debian, finish merge with Wayne and try again.') 22:57