pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m | FBSD_4.11:r1207(0/3197) win2k:r1197(105/3116) Linux:r1197(105/3116) MacOSX:r1197(105/3116) **[793 tests TODO]**
Set by kungfuftr on 26 March 2005.
kungfuftr stevan: the output wanted? 00:01
stevan it should duplicate the split version 00:02
kungfuftr bah, can't seem to get it 00:05
stevan kungfuftr: not a big deal 00:06
kungfuftr stevan: iirc, the last time i did it, it required fooling perl, so not hugely easy now that it comes to mind 00:07
stevan this /(?:[A-Z]<\s+|[A-Z]<|\s+>|>|\w+|\s+)/g is capturing all the peices 00:12
I can just glue them back together
kungfuftr bah, nearly 00:18
nothingmuch would like a spare perl6.vim file, if anyone has it 00:46
can't reach luqui's site 00:47
kungfuftr stevan: yar, split is a perl opcode... erple 00:49
stevan kungfuftr: it actually doesnt matter now, because I just realized that rx:perl5{} only works in boolean context anyway 00:51
so I cannot get the split values
kungfuftr ah, okay.. that not a bug? 00:52
stevan it is a bug
or rather an un-implemented feature
regexp support is very limited right now
kungfuftr =0)
stevan bascially another more than simple match and substitution is not support 00:53
Darren_Duncan another SQL::Routine commit -- still a lot of translation to go ... 00:58
stevan hey Darren_Duncan
Darren_Duncan I must leave computer for 1.5 hours, request at least an hour afterwards before release
hey Stevan
FYI, I still have more to do before you have at it 00:59
stevan Darren_Duncan: I figued that
kungfuftr still building ghc 6.4
stevan Darren_Duncan: take your time, I wont be able to work on it until next week anyway
Darren_Duncan sounds good ... and I'll alert you when you no longer have to wait
stevan Darren_Duncan: have you seen all the ports iblech did? 01:00
Darren_Duncan I noticed that someone did a lot of ports, but didn't remember their name
stevan all by iblech
but you have to go,.. dont let me keep you
Darren_Duncan good for iblech
yep, back in 90 mins
jabbot pugs - 1214 - large SQL::Routine update - continuing P 01:07
01:12 theorbtw1 is now known as theorbtwo
stevan hey theorbtwo 01:16
theorbtwo Allo, stevan!
autrijus err, split regex.
I'll look at it after I finish my slides. 01:17
stevan autrijus: :)
jabbot pugs - 1215 - first upload of perl6 module Pod::Stream
stevan autrijus: either that, or make rx:perl5:g{} return an array
which ever is easier
and no rush
autrijus ok.
stevan I am just goofing around with a Pod parser
autrijus the later is easier.
stevan see ext/PodStreamParser
autrijus++ thanks 01:19
autrijus perl6.vim is in. 01:26
jabbot pugs - 1216 - * perl6.vim, courtesy of luqui 01:27
ingy hola 01:28
autrijus Qiang++ # use.perl.org/~Qiang/journal/ 01:29
keeping in touch with community++
kcwu autrijus: perl6.vim is broken, only half part left 01:45
autrijus fixed 01:46
jabbot pugs - 1217 - * add missing half of perl6.vim 01:47
pugs - 1219 - * specify some function signature, then 01:57
pugs - 1218 - * implement ?| and fix contextual logica
Darren_Duncan okay, I'm back from my non-computer activities 02:38
If I remember, autrijus, its about 10:40 where you are now, right? 02:39
Anyway, I'd like another hour at least to get stuff done, will provide updates
kcwu right 02:40
theorbtwo Hmm, anybody remember the URL for the stream from YAPC::Tapei ? 02:44
scw rtsp://140.110.17.6/yapc.sdp
autrijus ok. another hour is ok. 02:58
theorbtwo Oh, the stream is only of the translator. 03:00
scw theorbtwo: Err, no. Now the speaker is using Chinese 03:01
theorbtwo Oh.
gmplayer -rtsp-stream-over-tcp rtsp://140.110.17.6/yapc.sdp -fps 25 ! 03:02
Of course, I can't understand a word he's saying. 03:03
Corion Did Test.pm break? 03:12
kungfuftr Corion: not unless it's happened post 1208 03:13
Corion kungfuftr: I get 'unexpected "s"' in blib6\lib\Test.pm at line 149, column 1 03:14
(I'm on 1219)
(Win2k, but I guess that shouldn't matter. But maybe it matters that I'm on GHC 6.2)
autrijus fixing. 03:15
Corion autrijus: Cool :) 03:17
autrijus Corion: please check 03:21
Corion ... checking 03:25
Looks much better - tests that failed before now pass. 03:26
kungfuftr lol, nice to see sam looking like a garden gnome (flickr.com/photos/obra/7453793/in/p...3643@N00/)
jabbot pugs - 1220 - * fix ?? breakage, thanks Corion 03:27
Corion Oooo - one TODO test passes - t/operators/operator.t - 1/29 unexpectedly succeeded 03:28
Corion bows to miyagawa
Darren_Duncan autrijus, I still need another half-hour or so until the next commit
Corion wtf ? "Failed 1/145 test scripts, 99.31% okay. 0/3198 subtests failed, 100.00% okay." ? Did you go on a mad bugfixing spree? (win2k/r1220) 03:29
autrijus k 03:30
kungfuftr youkoso miyagawa
Corion (or did I miss something and most of Perl6 now already works?) Anyway - really cool !
miyagawa chees
cheers
kungfuftr there's a YAPC in Tokyo? 03:32
clkao in taipei 03:33
kungfuftr supposedly there's one in 2006 in tokyo
theorbtwo I don't know why, but hearing languages that I don't understand a word of, then having suddenly a word or two in English, then continuing on, seems amuzing to me. 03:34
clkao hopefully 03:35
kungfuftr excellent, must get another excuse to visit our yokohama office next year 03:36
Corion falls into bed. Good night all!
theorbtwo Boggle. 03:37
I've never seen him up this late before.
We have a release in an hour? 03:40
autrijus asavige! 03:53
asavige Fore! mad golfer reports Files=144, Tests=3198 all tests successful on WinXP (1 subtest unexpectedly succeeded) 03:54
autrijus yay! 03:56
which is the unexpectedly succeeded one?
kungfuftr still waits on ghc 6.4 to compile
Khisanth unexpectedly succeeding tests are always amusing :) 03:57
asavige t/operators/operator 1/29 unexpectedly succeeded running with prove -v now 03:58
Darren_Duncan another incremental SQL::Routine commit is in ...
autrijus, I have now done the minimum I wanted to have in prior to the 6.0.13 release ...
tell me how much time left until you want to go, in case I were to try another increment commit 03:59
or if you want to go asap, then I'm fine with it
obra he's speaking right now
Darren_Duncan that was quick ... he only posted a few minutes ago 04:00
asavige prove -v on operator.t: not ok 20 - boolean or (?|) returns 0 or 1 # TODO (line 56 Expected: 0 Got: )
Darren_Duncan I'm away from computer for next 10-20 mins 04:01
will run another make/make test in absense
got rev 1222
theorbtwo It's a pannel discussion. 04:03
jabbot pugs - 1222 - r3551@speights: samv | 2005-03-27 12:2 04:07
pugs - 1221 - SQL::Routine perl6ification major increm
asavige I guess that means operator.t todo test on line 54 (2 ?| 3 equals 1) was the one that "unexpectedly succeeded"? 04:08
Darren_Duncan now on 'make test' for 1222
theorbtwo You're already one behind, D_D, but 1223 won't change any tests, but only because testgraph doesn't have any. 04:09
Darren_Duncan that's life 04:10
theorbtwo Next up: Leon Brocard 04:14
Exploiting Other Enterprises' Web Services
Oh, but not for another 2 hours.
Congrats, Ovid!
Ovid Thanks, but it turns out it's not over yet. There's a ringleader out there.
Does anyone here have the capability of doing a reverse cell phone lookup? 04:15
Darren_Duncan result of testing 1222 is 'All tests successful (1 subtest UNEXPECTEDLY SUCCEEDED)'
jabbot pugs - 1223 - testgraph: Fix divide by zero when test 04:17
theorbtwo Interesting; I didn't see any dark red from mine.
Darren_Duncan continuing another segment ... 04:19
autrijus Darren_Duncan: take your time 04:36
I'm still pending ingy's fix
asavige In pugsrun/02-dash-n.t I noticed weird "&>" redirection. what does it mean? why not just use ">" ?
Darren_Duncan I hear you
jabbot pugs - 1224 - * ?| is no longer TODO 04:37
autrijus I have no idea.
asavige 1224 make test all ok on WinXP now (unexpectedly succeeded gone) 04:45
autrijus cool. 04:59
Darren_Duncan, ingy: you have 90 more minutes.
Darren_Duncan thanks for the notice
autrijus then I do preflight check 05:00
then I'll release before my talk.
Darren_Duncan that's until 23:30 my time
sounds good ...
autrijus :)
Darren_Duncan++
ingy++
# literally, eleventh hour 05:01
Darren_Duncan after all, what way to be a master presenter than to say to the audience "and you can get it RIGHT NOW"
better then
immediate gratification and all that
I also meant 22:30 my time 05:02
autrijus :)
LCamel ;)
Darren_Duncan :)
hey, everyones doing it 05:03
obra Hi LCamel
tcliou :P
theorbtwo $$(':')::Crazzy
autrijus o/~ crazy / toys in the attic / we are crazy / truly gone fishing o/~ 05:04
LCamel hi, obra ;)
Khisanth el gamel :) 05:09
autrijus lol :)
LCamel :) 05:10
Darren_Duncan another SRT increment committed ... starting next one ... 05:40
autrijus the talk sequence changed. 05:43
you have _another_ 90 mins.
hcchien autrijus' talk would be startted about 4pm (GMT+8) ha 05:44
autrijus jserv is interested in parrot :D 05:46
Darren_Duncan that's nice to know ... however I probably won't take advantage of it since I have to wake up early tomorrow
autrijus (he's this mad kaffe vm guy in taiwan I like very much)
Darren_Duncan: ok =) let me know before you go sleep
Darren_Duncan will do
jabbot pugs - 1225 - another large SQL::Routine perl6ificatio 05:47
Darren_Duncan BTW, I don't know if anyone is doing this, but in the increment after the next one I expect to use multi-methods a bit
Khisanth autrijus: as in the Java thing?
autrijus Khisanth: yup. 05:48
he doesn't like java
but he likes java vm :)
and we both think that parrot has this strong Right Thing feeling.
Khisanth that ... seems a little odd :)
so he just likes writing VMs?
autrijus yeah. 05:49
Khisanth wonders how long it will be before someone writes a Perl -> Java bytecode thingy
or does one already exist? 05:50
autrijus parrot->jvm?
not sure.
Khisanth I mean just perl->jvm 05:51
autrijus oh. no. there's Inline::Java 05:52
but not compiled
finally writings pugs slides now
(was writing perl6/parrot slides)
I just love this example:
% pugs
pugs> { $_ ?? $_ * &?BLOCK( $_ - 1 ) :: 1 }.(10)
3628800
Khisanth was thinking more along the lines of how pugs can output stuff to run on parrot :)
autrijus src/Compile/Parrot.hs 05:53
it runs mandelbrot, as we know :)
(not much else, but hey, it's improving)
Khisanth because it needs more development or because parrot needs more development? 05:54
ingy autrijus: 1226
autrijus both, but mostly my side 05:55
I need to switch to TH2 entirely
Khisanth there is already a 2? 05:56
05:56 cnhackTNT is now known as cnhackTNT|WC
autrijus TH2 is GHC 6.4's TH 05:56
jabbot pugs - 1226 - Make things install to right places 05:57
gaal morning all 05:59
autrijus yo gaal.
Darren_Duncan mornin 06:00
gaal ooh, preflight checks 06:01
tests broke on mingw again :( 06:04
autrijus !. 06:05
break how.
ingy: your bug?
pasteling "gaal" at 192.115.25.249 pasted "mingw make test fails" (5 lines, 463B) at sial.org/pbot/8662
gaal we had something like this a few days back, Schwern had a few fixes 06:06
autrijus aw!
ingy: you should not put the */*.t etc to the command line
they belong to -e body!
(win32 does not expand htem) 06:07
kungfuftr *blink* okay... now ghc has actually finished building
gaal errrr, doh. tests DO run; somehow my environment forgot HARNESS_PERL and PERL6LIB. 06:09
ingy ?
I made no changes to test stuff 06:10
gaal at first none of my test ran, but that was apparetnly my mistake for not setting up the env right.
autrijus hrmph.
weird
HARNESS_PERL is supposed to be set up as part of make test 06:11
ingy: ok, prolly not your fault
gaal this is weird though, it's the same window i've had open for a few days now and i did set up the env in it
autrijus I'll try to fix.
gaal make test doesn't skip rules.Disabled 06:12
and t/Dialects/perlego/sigilless.t dies
Failed 5/148 test scripts, 95.95% okay (mingw)
pasteling "gaal" at 192.115.25.249 pasted "also, some compilation warnings" (26 lines, 1.1K) at sial.org/pbot/8663 06:13
autrijus the warnings are expected. 06:15
gaal k
sigilless dying is weird, because i treated it with an eval yesterday exactly so it doesn't die (and the treatment worked) 06:16
autrijus LCamel: foundry is broken just now. 06:19
did anyone do anything?
it worked a few mins ago.
Darren_Duncan I didn't do nothin' 06:23
autrijus sure, "anyone" here is limited to foundry folks :-/
LCamel autrijus: i see. 06:25
Darren_Duncan upon further investigation, the problem seems specific to the web server ... 06:27
a 'svn update' still works just now, between 2 failings to load a web page
so it appears I will still be able to commit tonight
autrijus preflight #1
Darren_Duncan am I not allowed to commit first? 06:28
gaal r1226, autrijus?
autrijus wagner.elixus.org/~autrijus/tmp/Per....13.tar.gz
Darren_Duncan: if you can still commit, sure., but chance are you can't.
firzen Why I always be off line? mIRC display "Your connection was lost"
Darren_Duncan I can update
autrijus update requires no auth.
auth requires the web server.
Darren_Duncan I see 06:29
autrijus again, for preflight, platform/ghcversion/the "how many tests passed" line
is encouraged.
Darren_Duncan right, I'll save any commit of what I'm on now for tomorrow
will now preflight check ...
gaal Disabled is still begin tested 06:30
autrijus gaal: tarball?
gaal (as of r)
autrijus tarball please.
gaal sorry, don't have the tarball not yet
autrijus svn is broken. :)
gaal ahh
autrijus I can't commit any of fixes ;) 06:31
gaal one sec
kungfuftr also notes "make test" is broken 06:32
autrijus kungfuftr: tarball?
please ignore svn and check tarball
kungfuftr k 06:33
Darren_Duncan tarball is downloaded ... now running 'make' 06:34
gaal make test is running. it triggered a regular make first, which of course didn't have any work to do 06:36
autrijus that is fine. 06:37
gaal getting some errors in examples; details soon
mingw/6.4/Failed 2/142 test scripts, 98.59% okay. 14/3196 subtests failed 06:38
autrijus ok. details?
pasteling "gaal" at 192.115.25.249 pasted "fails in examples" (156 lines, 5.6K) at sial.org/pbot/8664 06:40
gaal this is still mingw/64; checking on linux/6.4 now 06:41
autrijus ok. 06:42
gaal looks like i have permission problems and temp files don't get written?
autrijus yeah.
disabling those tests for 6.0.13.
(not critical to fix.)
(and golf actually works)
gaal isn't it just find . -type d | xargs chmod +xwr ? 06:43
autrijus purr.
assuming there's find and xargs and chmod.
preflight #2.
autrijus.org/dist/Perl6-Pugs-6.0.13.tar.gz 06:44
gaal just checked.. it isn't just :(
autrijus 17 mins.
autrijus still has 7 slides.
to write, that is.
gaal results from preflight#1 on linux/6.4: All tests successful, 1 subtest skipped.(<<--- basic.t) 06:46
1-basic that is
autrijus ok.
kungfuftr still running tests on freebsd 06:47
gaal compiling #2 on linux and on mingw
#2 still fails in examples :( 06:49
no
my bad
untarred wrong package, gaal-- 06:50
sorry about that :)
Darren_Duncan now on 'make test' ... 06:51
gaal still fails on examples/, new tarball :( 06:55
ingy pugs is released? 06:56
autrijus hrmph.
ingy: no, preflight
gaal: there is no t/examples in new tarball.
kungfuftr autrijus: all pass (1 skip [1-basic]) on freebsd 4.11 06:57
gaal is puzzled
Darren_Duncan tests done, all passed, there were 3196 tests in 142 files, no problems here
the t/1-basic thing results in 1/16 skipped, rest successful 06:58
I'm on Mac OS X 10.3.8 with GHC 6.2 06:59
gaal aw :( i forgot the screwdriver in the engine ESCdd reused the pf1 url. 07:00
Darren_Duncan I'm about to sign off for the night ... any last requests?
good luck with the delivery 07:03
gaal tests are running, and the coffee is high. 07:06
mingw/6.4: All tests successful, 1 subtest skipped. 07:08
autrijus ok, released. 07:10
gaal sorry for the delay earlier :(
kungfuftr yay 07:11
autrijus pause.cpan.org/incoming/Perl6-Pugs-6.0.13.tar.gz 07:12
gaal is off to $work 07:13
mugwump Releasing software while you're translating a talk!?! 07:14
autrijus yup.
kungfuftr offers autrijus some crack 07:15
Khisanth multitasking++ :) 07:16
07:17 cnhackTNT|WC is now known as cnhackTNT
asavige Perl6-Pugs-6.0.13.tar.gz seems to be missing new t/README file and t/examples, t/Dialects, t/kwid, t/Synopsis dirs 07:23
autrijus yes.
sadly.
sorry for that. :-/
asavige cool. Not showstoppers
better than crack is blue sugar cubes from the frig (Larry does that) Anyone know what is blue sugar cubes? Is that an American thing? 07:25
Khisanth has never seen blue sugar cubes 07:27
fritzz www.google.ca/search?hl=en&q=%2...&meta= 07:28
weird
q[acme] plays with pugs 07:29
asavige Larry mentions it here www.nntp.perl.org/group/perl.perl6....uage/19473
Khisanth asavige: I seem to recall a similar line in one of the matrix movies 07:30
cm I believe it was not about sugar cubes there :| 07:32
Khisanth yes but same idea? 07:33
nothingmuch morning 07:34
kungfuftr q[acme]: what happened to giving up programming?! 07:37
q[acme] kungfuftr: shush 07:38
kungfuftr q[acme]: remind me... i owe you beer
Khisanth kungfuftr: there wasn't a Programmers Anonymous to be found ;) 07:39
kungfuftr Khisanth: hhhmmm... there might be money in this... self-help books for programmers
Khisanth kungfuftr: already exist 07:40
kungfuftr =0o
lies!
nothingmuch wonders if anyone can keep with iblech's productivity 07:41
and link all the tests he writes
Khisanth all those "Practical $programming_language [Programming]?" books :) 07:42
link?
nothingmuch Khisanth: the backlinks 07:43
nothingmuch.woobling.org/pugs_test_status
browse S02, for example
you see links to the tests under headers
and also links to the tests as little superscripts in the docs themselves 07:44
Khisanth hmm no tests for rules yet?
nothingmuch i don't know
don't think so
p5 ones were disabled or something, a while ago
then again, i'm about 2 days behind 07:45
Khisanth I guess tests for classes and traits should come first
cm why did you want to give up programming, q[acme] 07:48
Khisanth it was either that or his right foot! 07:49
autrijus my talk starts in 10 minutes.
nothingmuch good luck, autrijus!
Khisanth so did you finish the slides yet? ;) 07:50
autrijus wagner.elixus.org/~autrijus/p6tu/start.html 07:57
Khisanth ptu? 07:58
autrijus perl 6 the ultimate
q[acme] hello world!
autrijus now all you guys are on screen 07:59
I have 2 mins before my talk
obra HI YAPC
autrijus I'll post announcement to p6c during those 2mins
obra miyagawa: yeah. let me load my pix
Khisanth :)
revdiablo quotes from the silmarillion? 08:00
lightstep in every source file
revdiablo ah, I wondered where those came from -- didn't recognize them until now :)
I guess 'The Silmaril as lantern light' gave it away 08:01
kungfuftr EAT THE CHEESE OF DOOM!
hcchien the pugs talk is going to start, and the streaming broadcast is in rtsp://140.110.17.6/yapc.sdp
kungfuftr bah... what supports real? 08:04
hcchien but it's in chinese. :p
quicktime
kungfuftr yay 08:05
that's one hairy perl geek in front of the camera!
Khisanth wonders why slide 18 is full of smileys 08:06
tiw how can i see it in linux? 08:15
kungfuftr tiw: mplayer with extra codecs should work 08:16
jabbot pugs - 1227 - fuller rows when splitting 08:17
tiw kungfuftr: thanks :)
revdiablo how long is autrijus' talk? 08:19
hcchien 90 mins
obra he'll go over ;)
hcchien ha
revdiablo ah, I wondered how he was going to fit in all this stuff =)
kungfuftr i was wondering how he was going to cope with not being able to commit for 90 mins 08:20
cm how can i see it on windows? 08:21
kungfuftr cm: quicktime
Khisanth kungfuftr: he can start demonstrating pugs development in the middle of the talk
kungfuftr Khisanth: touche 08:22
cm kungfuftr: thx :o
Khisanth "he we can see pugs development is going on as I speak"
oops s/he/here/
cm he speaks crazy language
hcchien ha, in chinese 08:23
kungfuftr exactly, crazy language
cm the projector had too many blue sugar cubes
lightstep howcome no test fail on bsd, but 105 fail on other platforms?
kungfuftr lightstep: look at the revision numbers 08:24
lightstep oh
Khisanth hrm
does the Makefile not handle the PCRE stuff?
cm ..so who in here understands this language? :)
Khisanth cm: quite a few :) 08:25
mugwump perl6? what's there not to understand? ;)
cm next question: who wants to translate :o)
mugwump oh han4yu3
kungfuftr mugwump: what's this i hear about you stealing the clothes from a garden gnome? 08:26
cm mugwump: i thought the point with p6 is that you just write code and it all magically works :o)
kungfuftr would be nice to see the slides, i would suppose
mugwump only if you write "retardo" after it :)
Khisanth hcchien: will the mpeg have subtitles? :) 08:27
mugwump You're the first person to say it looks like a gnome kungfuftr ... I've had "hobbit" recently :)
there's some pix on flickr
kungfuftr mugwump: i've seen you sitting at the bottom of my garden! 08:28
hcchien Khisanth: sounds a little difficult
cm someone call me when there's code to see on that stream.. :p
mugwump hanging out with the pixies
hcchien but slides would be, I guess.
maybe we should do a online translation etc. on irc. :D 08:30
Khisanth gah what do you have to do to get PCRE.o to link successfully?
cm yes you should!
Khisanth realtime translations?
hcchien gugod and scw know what he say now. :)
q[acme] we've been doing realtime translations for the rest of the conference, just the other way
whiteg autrijus is introducing the history of Perl 08:31
now hi is talking about the Perl 6 08:32
plasmabal New Genesis Camel XD
hcchien he is talking a bout the "languge tax" in perl5
plasmabal * Faster: Compile into JIT machine code, runnable in embedded system.
clkao make test wants /usr/bin/pugs??
why?
plasmabal * Compatible: Direct call to Perl5/C/Java/PHP/Python/Ruby libarary 08:33
kungfuftr clkao: no idea
lightstep ugh
plasmabal * Cleaner: Shorten often-used syntax.
kungfuftr clkao: autrijus just is "svn's broken"
plasmabal Powerful: Buile-in OO, Function-O, Doc-O, Logical-O
ooops, another slide.
skip skip skip. XD 08:34
Let's see Perl 6
lightstep pugs comes only on slide 61
hcchien live demo now
cm (thx for your translation efforts :)
plasmabal :)
kungfuftr ta
whiteg Perl 6 looks like Perl 08:35
plasmabal THis is a moudle, runabble
it's in test suite.
no trick, just plain code.
Khisanth hmm those slides work well as a mini perl 6 tutorial
kungfuftr Perl 6 looks like larry sneezing on the perl5 design spec
plasmabal vim syntax highlight from perl 5, seems workable on perl 6.
s,from,for,
mugwump perl 5 design spec? 08:36
mjhsieh live demo
Khisanth perl 5 has design specs?
kungfuftr yar, all written afterwards
=0)
obra kungfuftr: pointer?
cm each language construct fits onto a sugar cube ;o
mjhsieh benchmark
kungfuftr *slap* i'm taking the piss
whiteg this module will run tests, print yes, ok or not ok
plasmabal benchmark is like a box of cholocate: you never know where it runs faster. 08:37
it's on perl 5
mjhsieh two mendelbrot
plasmabal let's see it runs on perl 6
mjhsieh s/mendelbrot/mandelbrot/ 08:38
plasmabal see the results
mjhsieh -O2 vs -O0
plasmabal i didn't declare the variable as native type, or it'll run faster.
mjhsieh so it's working
plasmabal let's see the result, they are all the same.
whiteg if you download pugs, there are more examples 08:39
mjhsieh sorting
plasmabal teries.pmc, the test is very important. XD
i can play it until the talk is over. 08:40
mjhsieh no re-compilation needed
plasmabal Thee items:
* Perl 6: Language specifications:
*Parrot: Run-time environment 08:41
* Pugs: Compiler
Three major items, sorry. 08:42
whiteg compile Perl 6 to run on Parrot
nothingmuch how would you link the testgraph tds to the tests?
i know the target URL
kungfuftr plasmabal: no need to worry about being precise, just concise
nothingmuch but i don't know the HTML to back it
numbers look bad
plasmabal kungfuftr: :)
nothingmuch i need a sort of &nbsp; that fills the entire td and is clickable
obra nothingmuch: javascript
plasmabal Why they are called major? they're going to complete the New Genesis Camel 08:43
obra onClick
plasmabal Three minor items:
mjhsieh neon genesis camel
plasmabal * Ponie: Perl 5 interepter
nothingmuch obra: i'll try
kungfuftr nothingmuch: set the css style to "display: block;" and it'll fill the whole area
nothingmuch kungfuftr: what will fill it?
&nbsp;?
kungfuftr it's containing element 08:44
nothingmuch and if it's nothing, will that work?
plasmabal * PPD: Perl5 translator
nothingmuch <td style="display: block"><a href="url"></a></td>
kungfuftr should do
nah, on the link tag itself
yinjieh * Ponie: Perl 5 interpreter 08:45
plasmabal yinjieh: my finger flips. :p
yinjieh plasmabal: :p
mjhsieh ha
whiteg * PPD: Perl 5 translator
nothingmuch tries
plasmabal * PGE: Rules Compiler
whiteg PGE -> P grammer engine 08:46
mjhsieh P grammer engine
oops
kungfuftr too many cooks in the kitchen
plasmabal FAQ:
* Who many perl 6 programs can run now?
mjhsieh shut-up
plasmabal s,Who,How,
"Very many"
nothingmuch kungfuftr++ # works great 08:47
kungfuftr nothingmuch: yay!
plasmabal * Can Modules on CPAN run?
nothingmuch if you want to see, just one second
mugwump plasmabal: ingy's working on that :)
plasmabal "Many people is working on it"
mjhsieh excluding OO
plasmabal * Will Perl 5 stop maintaining?
whiteg of course not 08:48
plasmabal "ABsolute not"
absolutely
will maintain until nobody use perl5
* DO i need to learn perl 6 from scratch?
kungfuftr mugwump: is anyone from tpf out there?
plasmabal "no"
he talks too fast. XD 08:49
* I'd like to help, what can i do?
"Don't walk away, complete this talk." XD
lightstep mplayer doesn't get the stream: it keeps waiting
plasmabal good, this slide is a diagram. XD
mjhsieh lightstep: you might need qt 08:50
nothingmuch nothingmuch.woobling.org/pugs_test_...graph.html
whiteg it's a time table of Perl 6 developing process
2001:perl 6 starts
nothingmuch it's not perfect
plasmabal 4 years passed, no single perl 6 line is runnable.
nothingmuch some line numbers are missing
plasmabal bwahahaha
nothingmuch i'm working on that
kungfuftr nothingmuch: looking good
nothingmuch but in general you can click most blocks, and get to the corresponding line in the test file
which will be vimcolorified, as soon as i manage to find a version of perl6.vim
whiteg everyone bzz 08:51
nothingmuch after that I hope to start coloring the tests themselves
kungfuftr nothingmuch: perl6.vim is in svn
nothingmuch good to know
plasmabal this project is a lie.
nothingmuch i googled for it, but couldn't download
plasmabal: it's working now
plasmabal here comes a nut to implete perl 6.
nothingmuch: i'm translating what au is talking. :)
whiteg Feb 2005, a crazy strated to do something.
nothingmuch autrijus managed to save everyone ;-)
ah
hehe
my problem now is that the (platform x revision) matrix creates a bajilion graphs 08:52
and I don't know how to correlate everything
whiteg 2005-2006:hack hack hack
gaal hey, if plasmabal is translating au's talk live, and au has a #perl6 window open, so we gett hofstadter-style weird loops? 08:53
plasmabal Perl 6: Make world in steps
kungfuftr autrijus: skewed vision, eh?
plasmabal Below is progressing mainly on [email@hidden.address]
1. call for RFC, collecting ideas
2. brain-storm in groups, file complete spec suggests 08:54
gaal like videotaping your display as it displays what the vidcam feed
plasmabal 3. Larry split RFC into chapters, write A one by one
lightstep i think it's "strange loops"
whiteg within 15 days
gaal "strange loops" are weird
plasmabal his slide is in traditional chinese, i'm translating them into english.
gaal of course, if the content of #perl6 also mentions the fact that feedback is happening, this makes it doubly weird. 08:55
mjhsieh plasmabal: I can see the slide, then I'll leave it to you
plasmabal mjhsieh: ōæ½xA9Iōæ½xA9Iōæ½xA9I
whiteg so Larry spent half an year to split RFC
kungfuftr btw, anyone who know trad chinese, simp chinese and japanese AND perl, would be well advised to apply for a job with us! 08:56
plasmabal there're too many rfc so larry in in BSOD.
bwahahaha
whiteg Larry split RFC according to the chapters of llama
plasmabal after endless reboot, larry decided to focus one topic at a time.
Khisanth ah finally ... 08:57
plasmabal 4: design team write corresponding notes, providing examples.
5. communities provides suggest to A and notes.
6. Design provide S according to feedback.
tiw mjhsieh: is the slide in web?
plasmabal wagner.elixus.org/~autrijus/p6tu/sl...g.html#end 08:58
mjhsieh plasmabal: sorry (I can't)
plasmabal 7. Implementation team will do initial implements
mjhsieh: can't want?
can't what?
8: Correct S according implemntaiton results
mjhsieh plasmabal: I can't see the slide through the broadcast
plasmabal 9: implemntation team do the following implementation. 08:59
whiteg we're now at step9,
plasmabal Perl 6: ... (Don't know) XD
* Huffman encoding:
gugod Perl 6: Driver Complexity Easily
kungfuftr chuckles at craziness between video and chan
gugod s/Driver/Driving
plasmabal * Huffman encoding: Easy is easier, hard possible. 09:00
whiteg the more used command should be shorter
plasmabal 'or' is so often, why we use so many 'or'? use any(...)
* Meet balance between extermes 09:01
# Too many operators
# Too little operators
* Accept (words from outside world), respect differences between cultures. 09:02
gugod Accept foriegn languages
foreign, sorry 09:03
plasmabal Perl 6: The theory of waterbed.
kungfuftr ullo ij1iao
plasmabal * Press here
gugod Perl 6: Waterbed Theory
mugwump why use any() when you can use the āˆ€ operator? :)
blast
plasmabal * Pop there
lightstep isn't that forall? 09:04
mugwump er, yes, all()
plasmabal * The complexity is a constant
mugwump probably want āˆƒ (\x{2203})
plasmabal Perl 6: Avoid (gugod, here you go) 09:05
* If "cat" is "tog"
* outstanding visual difference
whiteg Avoid a word with diffrernt meanings 09:06
gugod Avoid the phrases with multiple purposes
ij1iao avoid ambiguous
gugod exaclty
LCamel don't overload too much
plasmabal Perl 6 : Freedom
* Give the most freedom to programmers
jabbot pugs - 1228 - Move SHA1 to modules. 09:07
kungfuftr <<NB: live translation of autrijus' talk in progress>>
plasmabal * By custom, no rules.
ij1iao s/idioms/custom/ ?
plasmabal * By idiom, not rules
ya ya
ij1iao * Keep flexibilities 09:08
whiteg * reflect environmental requirments at any time
plasmabal somebody take over, i need a rest. :p
kungfuftr hehe
ij1iao gugod: go
gugod goes
XD
whiteg ij1iao: go
kungfuftr ij1iao++ # my fbsd hero
plasmabal ij1iao: hero!
ij1iao who ?
whiteg Perl 6: DWM
plasmabal DWIM 09:09
whiteg * Do what I mean
kungfuftr not what i say
whiteg * Make Perl more like a phrase
* do have to use formal english
gugod mc:3rd:w/word/ -> match the 3rd occurence of "word" in a string
plasmabal do not
whiteg s/do/do not
mjhsieh mc:3st:w/word/ also do 09:10
whiteg * don't even have to use english
plasmabal * not necessary in english
gugod and here comes some chinese code
mjhsieh 'engrish'
plasmabal (those examples are in chinese)
mjhsieh: haha
ij1iao i think it's "not always use formal english"
whiteg you can do it with some slang module 09:11
gugod Perl6: Document Object Model
* Whatever inside the progrom document, should be also a object
kungfuftr so now were'
ij1iao * use %=POD to access all kinds of embeded data
kungfuftr bah
whiteg Perl : variable signs 09:12
plasmabal * Sigil is now part of variable names
@array[3] # no longer $array[3]
%hash{'key'} # also in %hash<key>
mugwump It's called a DOM but it will look and work more like Tree:DAG_Node or XML::Twig etc
Whichever Tree algorithm gets ported to Perl6 first probably :)
whiteg rules is not very consistent.
s/is/are 09:13
plasmabal i'm back. :p
whiteg welcome back
plasmabal * array and hash could be used as objects
gugod (can I keep trying to release SVK::Churn now? :P)
ij1iao * therefore, arrays and hashs can be used as objects
plasmabal gugod: go ahead. :p
do things through method calls 09:14
Perl 6: Dot
whiteg Perl 6: .
ij1iao * "->" becomes "."
plasmabal * often-used -> shorten to .
$a_ref.[1] # could be shorten into $ref[1]
* . could be used to call sub 09:15
nothingmuch . o O ( p6 autocompletes the a_ ? ;-) )
plasmabal ... # it works
.... # it works too
Perl 6: method
* Easy to memorise method
@... # replaces scalar(@array) 09:16
mjhsieh @array.elems
miyagawa Orz
plasmabal @array.bytes # ... impossible in Perl 5 Orz
&function.arity #... also impossible XD
ij1iao Perl 6: Named Variables
plasmabal opos new slide
* Old syntax is ok
* New named parameters
jabbot pugs - 1229 - r3555@speights: samv | 2005-03-27 17:4 09:17
plasmabal sub clean... # by addr
sub by_value ... # by value
mugwump bah, why don't me commit messages go in cleanly :(
plasmabal sub some_opt ... # optional
sub modify ... # modifiable
sub typed ... # typed
LCamel Perl6::Subs
plasmabal Pelr 6: Call sub 09:18
* Not flatten
:names<a b c> => pass a b c into parameter named 'names' 09:19
* Flatten 09:20
kungfuftr mugwump: i assume they're truncated by the bot
gugod no... it's truncated in the rss
plasmabal Perl 6: High level sub
LCamel (*@_) 09:21
plasmabal * use .assuming to do local currying
gugod I shall change the bot to optionally report the <description> tag
that'll make it more informative
plasmabal * ...... (blank brain...)
autrijus just said he's going to speak FASTER. :O 09:22
kungfuftr mwuahaha
ij1iao perl6: multi-dispatch sub
plasmabal Perl 6: Multiple dispatch sub
* By type
* Type value
s,Type,By,
Perl 6: Operator 09:23
ij1iao perl6: operators
plasmabal * New operator
say ... # read operators
...; # ... operator
whiteg ...: #lalala operators
plasmabal haha
whiteg lalala
plasmabal lalala
* unicode operators 09:24
whiteg you can define operators with unicode 09:25
plasmabal bwahahaha 09:26
ij1iao zip operator ...
plasmabal the jap-yen looks like a zip
unable to type jap-yen on your kb? use Y instead. 09:27
* Extend existing operators
miyagawa hmm, Yen mark is confusable to backslash, for Japanese users
mjhsieh miyagawa: I agree 09:28
miyagawa we already see/use Yen mark instead of backslash :)
plasmabal Perl 6: Numeric value and true value
* Numeric context
* True context
* Trinary operator
LCamel boolean context
plasmabal hum? 09:29
oh
s,True,Boolean,
ij1iao boolean context, not true context
plasmabal Perl: Strings
* String context and concatation
* String interpolation
* Formatting strings 09:30
Perl 6: Logical operators 09:31
* Short curcuit and open curcuit 09:32
open ?
am i right?
LCamel circuit
plasmabal * Italic (?) || is 'defined or' 09:33
mjhsieh magic of 010
ij1iao //=
plasmabal // 09:34
we're unable to make or italic, so we use err.
... ōæ½xB6Wōæ½xBAōæ½xFB?
ij1iao super scalar ?
gugod hyper-dimention operation
obra what does >>|<< do? 09:35
plasmabal i don't remember i saw this before. :<
Perl 6: Junctio operators
Junction
lightstep obra, runs | elementwise
LCamel >>|<< works like map ?
plasmabal * Use shortcut (?)
# In perl 5, you write:
* Prime test 09:36
ōæ½xBDōæ½xE1ōæ½xA6W?
lightstep no, >><< works like map
plasmabal alias?
Perl 6: Alias operators
Juerd obra: @foo >>|<< @bar returns a list of max(+@foo, +@bar) 2-element disjunctions 09:37
LCamel i think codes written in perl6 will be more difficult to maintain..
plasmabal * Don't need to learn hard glob
ij1iao powerful lazy-evaluation
plasmabal * Powerful lazy evaluation
Juerd LCamel: Code is singular, always.
plasmabal Perl 6: Matching operators
* Smarter than =~
Juerd obra: Don't you like the >>+&=<< operator? ;)
plasmabal * Could process not only scalar 09:38
Perl 6: Switch
LCamel Juerd: thx
plasmabal * We use given/when to match
* given/when use the same function as ~~
ij1iao perl6: loop
plasmabal Perl 6: Loops
gugod LCamel: you should start writing perl6 right now
plasmabal * foreach now use for
* for could retreat many values at once. 09:39
gugod LCamel: then you'll know how easily it is to read
plasmabal s,retreat,get,
Perl 6: Exceptions
* Use $! globally, no $@, $?, and $^E.
LCamel gugod: the language becomes bigger
plasmabal Perl 6: Class
... TOO FAST now~~~ 09:40
gugod LCamel: but code becomes more clear
LCamel gugod: i see.
plasmabal Perl 6: Roles
gugod LCamel: peopel don't write a module that uses all feature of perl6
plasmabal Perl 6: Rules
gugod LCamel: (that's scares)
yay, Perl 6 rules
gugod feels like a noise application 09:41
ij1iao too fast to translate ...
kungfuftr shoots gugod
plasmabal * modifiers are prefix instead of postfix
* anonymous and named rules
* rules includes rules
Perl 6: Syntax
* Give rule a namespace
* Perl 6 is a grammar
Perl 6: In one word 09:42
* easy to learn
* Specific (?)
* Accept others
* Workable now
* Make perl more perl
Parrot
....
Juerd wth is plasmabal :)
cm plasmabal is translating autrijus' talk
plasmabal autrijus is insane now
cm plasmabal: how so? :D 09:43
Juerd ahh
ij1iao April Fools joke become true
gugod -- It starts from 4/1 2001
plasmabal Parrot: Dreams come true in April's day
* What is parrot
* Perl 6's Virtual machine
cm * a pipe dream
plasmabal s,april's day,april fool
gugod A joke that said "Perl6 and Python will be merged" 09:44
plasmabal * also run python, ruby, scheme, forth, basic
gugod That comes true as desining VM for both languages
09:44 bytewolf is now known as cnhackTNT
plasmabal * more suitable for dynamic language than java/.net 09:44
gugod ... while it's also the time .Net starts
plasmabal * college course
* (developers etc)
Parrot: downlad and install
* get stable version
* develop version 09:45
Parrot: assembly langauge
ij1iao parrot: assembly language
gugod parrot has 3 internal asm
plasmabal * high-level assembly langauge
gugod IMC, PASM, PBC
plasmabal * lovel-level
ij1iao parrot : register
gugod it uses registrator
plasmabal blah blah blah
Parrot: registers
* Direct maaping to registers on cpu
gugod Some of the binds to CPU register directly
plasmabal * Parrot magic cookie:
Parrot: variable
* Use registers as variables
gugod s/Some of the/Some of them/ 09:46
plasmabal * named local variable and functions
# save as hello.pmc and you can run in parrot
gugod As a VM, the number of registers is like 2^32 -1, but it depends on your CPU arch
plasmabal parrot: control structure
ij1iao Parrot: Control Flow
plasmabal * simple loop
* lable can be saved in continuation objects 09:47
gugod -- This asm has /continuation/
ij1iao * labels can be stored to continuation objects
gugod (are we sub-etha-irc ?)
plasmabal parrot: sub
ij1iao Parrot: function
plasmabal * simple sub
haha
ij1iao * call sub
plasmabal ij1iao: take over?
gugod -- this asm even has function call
ij1iao plasmabal: no ;p
plasmabal foo
ij1iao bar
cm baz 09:48
kungfuftr cheese
plasmabal parrot: advanced sub
ij1iao Parrot: adv. sub
plasmabal * multi-valued sub
* sub call
gugod -- it doesn't require you to learn parrot asm to write perl6, but it does, when you want to know the reason your perl6 is not fast enough
plasmabal parrot:library
ij1iao Parrot: Library
plasmabal * load library
elmex gugod: rofl
plasmabal * get sub wihtin
* blah
elmex gugod: where does that wise saying come from?
plasmabal Parrot call external library
gugod -- you are allowed to load any of 3 asms, or any system depended libraryes
plasmabal * call .so or .dll on system directly 09:49
gugod elmex: I'm trying to translate autrijus speaking that are not covered by ijliao or plasmabal
plasmabal * call imeediately now
parrot: todo
gugod elmex: so, autrijus :)
plasmabal * abstract syntax tree
* charset translation (?)
ij1iao * encoding translation
plasmabal oh, encoding
ij1iao async IO
plasmabal * non-async IO
* object exception handling 09:50
* blah
ij1iao Pugs:
gugod -- Just release 6.0.13
plasmabal pugs: download and install
LCamel * 6.0.13 just released
plasmabal * 6.0.13 just released
* install GHC first
ij1iao * pre-req GHC
plasmabal * get develop version
elmex what is sthis stuff plasmabal says???
plasmabal pugs: the history of development
* start on Feb 1, 2005.
kungfuftr elmex: keep up...
mjhsieh elmex: YAPC::Taipei
elmex ?? 09:51
ij1iao GHC ? www.haskell.org/ghc/
plasmabal * the homework that I learn haskell and TaPL
LCamel elmex: plasmabal is translating the speach of autrijus
nothingmuch elmex: autrijus' talk
plasmabal * visit pugscode.org
cm elmex: plasmabal is translating autrijus' talk
plasmabal * superfast deveopemnt pace
Pugs: interpreating mode
gugod -- and it all starts from a haskell tutoral execisement : Please choose some familiar small language to write a parser
plasmabal * CLI
elmex: sorry, au is in bersek mode, so am i to follow his slide. :p
elmex ok 09:52
gugod -- JAPH code on screen
-- Perl6 Users Golfing System,
golf-driven development!
plasmabal pugs is not acronym of any words, but it's exapanded anyway.
* interactive interface 09:53
gugod -- The early golf contect programes and accounting scripts are now ported to perl6
plasmabal pugs: compiling mode
* three complie backends
ij1iao * 3 compiler back-end
gugod -- it's also a compiler to compile to haskell or C
plasmabal * JIT
gugod ooops, s/C/Parrot/
plasmabal pugs: develop model
gugod -- pugs : JFDI 09:54
plasmabal * blah
* Wiki-style development team
ij1iao JFDI : Just Fxxxing Do It ?
gugod -- develoeprs(x) =~ x
yes
plasmabal * two condtions to be a committer:
LCamel * 50+ commiters
plasmabal * test goes first
gugod -- Test-driven, Golf-driven
-- >3k tests
plasmabal ** golf-driven
* test-driven 09:55
** golf-driven
* attracts a lot of nuts
haha
gugod -- Golf has lots of un-document features of perl6 that's not even present in some source code , and might be even runnable on full-moon night.
nothingmuch =)
plasmabal Pugs: documents 09:56
* pugs: ex..... (too hard to spell)
* Kwid document format
cm exegesis?
:|
gugod Apocryphon
plasmabal cm: never remember how to spell it. :p
jabbot pugs - 1230 - Refactor SHA1 to external Makefile.PL 09:57
gugod Apo-Cry-Phon :p
plasmabal * reamd and howto
s,reamd,readme,
ij1iao pugs: test
plasmabal Pugs: Tests
cm hehe
ij1iao * 3k+ unit tests
plasmabal * t/ dir: 3000+ unit tests
ij1iao ** how to start writing test
plasmabal * examples/: 40+ integration tests 09:58
gugod -- the way you do bug report, is started from writing failed tests
-- and pasted on irc, so that other committers will be impatient so that you'll have a committer permission in like 20 miniutes
plasmabal Pugs: built-in modules
Pugs: Modules in porting 09:59
* current 22 modules
* coming sson: DI:PurePerl.....
kungfuftr buggery, good bit longer than 90 mins so fare
plasmabal "i have to demo this, this is important."
LCamel inline Haskell
obra is the perl6 YAML plan SYCK?
plasmabal "you can write haskell."
Pugs: Embed other lang
kungfuftr buggery... that's what i mentioned earlier... so i suppose i'll have to do it... damn... DBI's huge
gugod -- ingy is still fighting on dynamic loader of inline 10:00
-- even, embed perl5
plasmabal give +10 sword to ingy
Pugs: Internals
obra kungfuftr: convince tim to do DBI2::PP in perl6 ;)
ij1iao fighting ? struggle ? :p
gugod -- starting from main
-- src/Main.hs 10:01
plasmabal Pugs: plan on 6.2
kungfuftr obra: well, i like a challenge... =0)
plasmabal * basic I/O objects
* basic I/O objects and flow control
ij1iao * to-be-released
plasmabal pugs: plan on 6.28
* OO kernel
pugs: plan on 6.283 10:02
* blah blah blah
pugs: far shore
Pugs: Let's do it together
* port module
* think about examples 10:03
* write examples
* translate documents
* clean up (?) parrot imc
mugwump will copy autrijus' idea for Tangram to go for 2.7182818... never get to 3 :)
plasmabal * learn haskell, improve pugs kernel
gugod -- you don't even need to know haskell to help perl6
plasmabal whoooo
gugod >clap< >clap< >clap<
plasmabal finish finally. XD
ij1iao tired 10:04
kungfuftr plasmabal++
cm plasmabal++ :o)
plasmabal :)
nothingmuch plasmabal++
scw plasmabal++
cm plasmabal & co, thanks
nothingmuch gugod++
LCamel mugwump: looks like metafont..
nothingmuch autrijus++
kungfuftr someone ask autrijus where we can send donations!
gugod yay, YAPC::Taipei fihished
nothingmuch kungfuftr: there's the PAR link
gugod YAPC::Asia for the next!
nothingmuch and i'm subscribed at $20 a month
i cna give you the link to that too
kungfuftr ah! gotcha 10:05
ij1iao YAPC::Asia in Tokyo
nothingmuch pugscode.org/donate.html
that's the subscription link
kungfuftr hhhmmm... bah, will have to wait for now 10:06
10:06 cnhackTNT is now known as cnhackTNT|away
jabbot pugs - 1231 - S11 - modules - pseudocode 10:07
cm good night 10:11
kungfuftr who ported Email-Envelope? 10:33
ah... iblech 10:34
bugger... DBI::PurePerl is a bastard 10:54
Khisanth to port? 10:56
kungfuftr yup 11:02
luqui hi metaperl 11:41
metaperl hi luqui
luqui is make test broken for you? 11:42
it seems to be assuming that I have pugs installed...
metaperl if youre asking me, I can check... 11:43
luqui yeah... you're the only one here, it seems
metaperl i'm doing make right now
after svn update
gaal luqui, are you using the tarball or the latest revision? 11:44
luqui latest rev
jabbot pugs - 1232 - Fixed syntax error in Context.hs. 11:47
gaal worksforme linux/6.4
luqui do you have it installed?
gaal luqui, does setting the environment help?
luqui setting the environment to...? 11:48
gaal pugs.kwiki.org/?FrequentlyAskedQuestions :
export HARNESS_PERL=$PWD/pugs
export PERL6LIB=ext/Test/lib
luqui oh, hey, probably
gaal it's still a bug, because make test shouldn't depend on that
but doing should workaround the problem. 11:49
luqui works
rgs apparently in 6.0.13 Test.pm is installed under perl5 instead of perl6. Known bug ?
gaal never make installed :)
luqui hmmm, I don't know
luqui has been absent from the pugs world for a week 11:50
which is an awful lot of time here...
rgs svn ups
gaal a consolidated grep for #perl and p6-[compiler|language] would be cool. 11:51
i need to go to a meeting. bye.
metaperl its trying to run /usr/bin/pugs instead of pugs in my local directory tree 11:52
so it does'nt even start make test
luqui that's what was happening to me
gaal has a workaround
rgs I can reproduce this as well
luqui something else is wrong...
all the tests pass except for the rules ones
wow... 11:53
I wonder how many have been todoed
rgs is there a commiter here ?
luqui me
rgs in t/syntax/inplace.t the shebang line is wrong
it needs to be pugs instead of perl
luqui okay 11:54
any others?
what's that grep option that prints files which don't match? 11:55
rgs t/oo/roles/properties.t (lists perl6)
-v
luqui no, files. and I just found it. -L
rgs oh ok
luqui committed 11:58
rgs why isn't there any MANIFEST file ?
luqui there's a MANIFEST.SKIP 11:59
so... I don't know
integral MANIFEST can be generated by make manifest
rgs ok 12:00
luqui r936 | ingy | Removing the MANIFEST: It always gets out of date, and is only needed for release.
has work on BEGIN, er, begun? 12:01
cls_bsd_ pugs 6.0.13, good 12:02
rgs make pure_site_install doesn't take DESTDIR into account 12:03
luqui wishes he knew anything about configure anything
rgs sends patch 12:05
jabbot pugs - 1233 - s{#!/usr/bin/perl6?}{#!/usr/bin/pugs} in 12:07
luqui to p6c?
rgs yes 12:08
luqui yep
there it is
does the patch include the End. at the end? 12:09
rgs no. It's to make the life of patch(1) easier
luqui ohhhh....
you can just run raw stuff through patch and it will pull it out?
do I need a command line flag to do that? 12:10
rgs no, just pipe the mail through patch -p0
and commit :)
luqui neato
rgs patch has been designed by someone who knew what "dwimmery" means 12:11
luqui gee, I wonder who that was?
thanks, applied, btw 12:13
rgs woohoo 12:15
kungfuftr MANIFEST is only generated for releases
jabbot pugs - 1236 - Pugscc don't need to be used under pugs 12:17
pugs - 1235 - Support make DESTDIR=/... install, court
pugs - 1234 - r1204@scwnb: scw | 2005-03-26 13:57:43
rgs ok, another Makefile.PL patch 12:21
luqui awaits arrival 12:25
rgs well, I'm trying to fix make test
hmm, makemaker black magic... 12:29
for some reason Pugs::MakeMaker isn't run to fix the Makefile 12:36
jabbot pugs - 1237 - debian/ added. 12:37
luqui beds 13:04
jabbot pugs - 1238 - Fix make test, thanks to rgs. 13:07
rgs yay. 13:08
theorbtwo You know, throwing babelfish on these slides is hilarious. 13:15
theorbtwo wonders if anybody saved the videos. 13:27
We have a debian dir now? 13:31
Cool!
scw theorbtwo: I added it but I'm not really familiar with those stuff. And also, we have to find a DD to accept it :) 13:44
theorbtwo Well, we don't have to; we can build them locally. 13:46
scw yes, and I've do that www.csie.ntu.edu.tw/~b92103/pugs/ 13:48
and also www.csie.ntu.edu.tw/~b92103/parrot/
gaal scw++
scw will do snapshot of svn and cvs HEAD every day 13:49
theorbtwo pings nothingmuch
nothingmuch pongs 13:54
did you see the graph thingy?
theorbtwo I saw the one on woobling, but not all changes seem to be checked in. 14:01
nothingmuch no, they aren't
and i didn't want to let it go till i was content
it was a bit of a crappy change
it caused some breakage
but since i'm at work i haven't looked at it
on the plus side, i'm going home in 2 mins 14:02
theorbtwo nods.
nothingmuch but i can check it in anyway
theorbtwo I did notice some changes that are in, and I'm uncertian of the reason for them.
Oh, I see what you were doing now. 14:03
nothingmuch the ones to the testgraph?
it just makes it look a bit more uniform
theorbtwo Right, I see now.
nothingmuch so, you have 20 seconds 14:04
checkin?
theorbtwo Go for it.
I'll do some cleanup of them while you're on your way home.
nothingmuch ok, sent 14:05
next i'd like to use Text::VimColor for catalog_tests
and also add support for multiple test result consolidation 14:06
jabbot pugs - 1239 - graph -> test line hack 14:07
theorbtwo I was thinking of using CSS mouseovers to properly display the test results -- they cut off and show as one line in everything but IE at present.
nothingmuch evening 14:42
stevan hey nothingmuch, theorbtwo 14:48
nice work on the testgraph stuff
we should package that up for perl5 usage too :)
gotta run,.... easter here,... later :)
nothingmuch yeah, i thought about forking it off pugs
stevan nothingmuch++
nothingmuch hola stevan
and mentioning it on perl-qa
stevan theorbtwo++
nothingmuch now i want to do consolidation
ciao stevan
have fun
stevan nothingmuch: do that
it could be a very valuable tool
obra win 24 15:06
nothingmuch stevan: ping 15:08
autrijus greetings! 15:09
I'm finally back to home.
gaal we all had lots of fun "hearing" plasmabal's translation of your talk :) 15:10
theorbtwo Allo, autrijus! 15:11
theorbtwo is sorry he missed the live version.
autrijus that'st fine :)
nothingmuch hola autrijus
gaal if there's a cap, maybe we can author an srt.
autrijus so, I should've said that 80% from the middle two sections 15:12
that is 40% of total talk
is from allison's talk
url:
www.lohutok.net/talks/p6update.pdf 15:13
clkao autrijus++ # wonderful talk
gugod autrijus++ # compress 50 days of pugs into 2 hours 15:14
jabbot pugs - 1240 - * remove Disabled tests and stuff from m 15:17
autrijus :D
actually I took only about 20 minutes for pugs :) 15:18
gugod so that's 85% higher compression ratio
autrijus so! I'm implementing a new language! 15:19
a new Parser, AST, Compiler, Emitter
gaal bored with perl6, eh
autrijus it's Pugs all over again!
yes!
it's IMC!
the higher-level parrot assembly thing! 15:20
gaal bored with sleep, too
autrijus rgs: committer-bit-p
?
rgs autrijus: er, why not.
autrijus mandrake.com email? 15:21
rgs rgarciasuarez at mandrakesoft.com
how many committers are there now?
autrijus done. welcome aboard!
mugwump I'm number 26 :) 15:22
autrijus you're the 50th I think.
rgs: see wagner.elixus.org/~autrijus/p6tu/slide66a.html 15:23
rgs nicely linear.
resistance is futile.
autrijus another graph: wagner.elixus.org/~autrijus/p6tu/slide63a.html 15:25
SVK::Churn++
gugod++
elmex hi 15:27
gaal hi elmex
elmex i love typed languages :) like perl
autrijus yes! 15:28
elmex my int $i; my string $s; ...
and will be beatiful for people who are using their own types in every module 15:29
autrijus sure.
elmex my MyOwnCoolIntWhichIsIncompatibleToAnyOther
i mena, can i pass my own int's for example to a foreign module, which uses special signatures in it's methods? 15:32
or maybe any object which act's like an int?
autrijus scw: may I move debian/ to util/ ? 15:36
I'm not comfortable with freebsd/ and win32/ etc
gaal i've never seen it anywhere but the top level of a source tree 15:37
autrijus really?
won't work at all in anywhere but toplevel?
gaal maybe it's possible, with some extra configuration
elmex hm, no idea?
autrijus anyone got tuits to check? :) 15:38
gaal the dpkg-deb manpage says "Creates a debian archive from the filesystem tree stored in directory. directory must have a DEBIAN subdirectory". so it doesn't sound possible 15:43
unless the spec files refer to ../paths
but i don't know if that works
autrijus ok. in that case it remains, but I'm not shipping it with MANIFEST 15:44
gaal there doesn't seem to be an option for more flexible locations.
autrijus added it to MANIFEST.SKIP.
gaal I'll add a note to the README saying why this is omitted from the tarball and a very short explanation of how to get it anyway 15:45
if that's okay with you
you tag releases in svn, right?
jabbot pugs - 1241 - * do not ship debian/ with tarball 15:47
rgs make test is broken again. 15:48
autrijus how?
rgs t/var/type.................................sh: /home/rafael/foreign/pugs/blib6/script/pugs: No such file or directory
can't run t/var/type.t. No such file or directory 15:49
gaal forgetting PERL_HARNESS again apparently
rgs, does the ol' workaround work? export PERL_HARNESS=$PWD/pugs 15:50
rgs well, pugs goes in blib instead of blib6. or the reverse 15:51
gaal in my box it just ends up at the tree root too, and i use that one for the tests 15:52
rgs "pugs" is supposed to be built in blib/ right ? not in blib6 ? 15:55
rgs commits 16:03
autrijus rgs++ 16:04
clkao autrijus: want to give me short example and pointer to inlining c with callbacks and invoking function pointers? 16:05
rgs well, make test now runs from a clean build, but I feel like a bear in a porcelaine shop
gaal i think you mean you feel like the ownder of said shop
rgs notices that his name is already correctly spelled in utf-8 in AUTHORS :))
jabbot pugs - 1242 - * pugs (and other scripts) are preinstal 16:07
autrijus clkao: www.haskell.org/pipermail/cvs-ghc/2...06777.html 16:08
clkao: more uptodate examples are possible
ask #haskell
clkao year 2000!
rgs commits again 16:17
jabbot pugs - 1243 - * Remove temporary files left by tests
rgs it's addictive.
Khisanth gooood :) 16:20
jabbot pugs - 1244 - Heavy Magic 16:27
pugs - 1245 - drunky crow 16:37
theorbtwo rgs: The utf8 names I steal from CPAN when I can. 16:43
rgs nice. 16:44
jabbot pugs - 1246 - Get object files to install in right pla 16:47
nothingmuch i think we better not use it's HTML backend, but the perl array or xml ones 16:56
vim color is hard to deal with
that way it's easier to promote the backlinks
gaal ? 16:57
theorbtwo checks in some CSS goodness -- please test with more browsers.
(Recent firefox and opera test OK, haven't tried IE...)
gaal to2, can you give me a url? i'm not near something that can generate tests 16:58
theorbtwo desert-island.dynodns.net/misc/testgraph.html 16:59
gaal wide lines on ff/win32
having the diag info inthe boxes is way cool though! 17:00
theorbtwo Damn... yeah, I loaded too quick...
It's not supposed to -- it's supposed to show them on mouseover.
gaal and the tooltip is incorrectly located 17:01
theorbtwo For some reason, some lines are producing very strange HTML.
gaal this is ff. let me see in msie
theorbtwo It's not incorrectly located; I couldn't figure out how to place it anywhere better without modifying flow as you mouseover, which makes moving the mouse do strange things.
gaal you look like you've been slashdotted :) 17:02
the mouseover should probably be on the td, not the text inside it (if that's even possible) 17:03
theorbtwo It's not.
gaal msie doesn't show any tooltips :(
theorbtwo It doesn't show complete junk, though? 17:04
gaal nope
it looks like the old version of the report
mugwump mmm, junk
theorbtwo mugwump?
gaal could it be the boxen with text in them had unescaped "s ?
theorbtwo Unescaped quotes shouldn't matter; it's HTML now, not an attribute. 17:05
ingy hola
gaal firefox++; # "view selection source"
theorbtwo hola, ingy. 17:06
Back from your Asian tour?
jabbot pugs - 1247 - testgraph: CSS mouseover goodness. 17:07
ingy no, still here
for another 9 days!
theorbtwo Oh, wow! 17:09
Oh: Do you know if the pugs hackfest around YAPC::NA is planned for /before/ or /after/? 17:10
autrijus before. 17:11
gaal to2, looks good now, but there are still a few bugs with ff
autrijus somebody offered a nice place.
need to talk more
gaal t/builtins/strings/append.t - test 4
also, on msie subtests are not clickable 17:12
corection: they *are cliclable*, but the clickable area is tiny 17:13
theorbtwo Hmm, that's the same test I'm seeing it on, but I haven't the foggiest clue why it's generating the bumb HTML for those few tests.
gaal and left-justified
but it's better now than your previous html. what changed? 17:14
Corion theorbtwo: Niiice CSS/html !
theorbtwo t/oo/roles/basic.t 5, t/operators/pipe.t 5, t/rules/perl5.t 133&554...
gaal i mean from five minutes ago, from before your ci 17:15
Corion theorbtwo: But it would be better if the tooltip showed below the mouse cursor instead on the left lower corner, because for the pugsrun/ tests, there are only two tests and I have to scroll too far to see both the test and the output 17:16
theorbtwo Mostly CSS.
Corion (but that'll involve JS)
theorbtwo Doesn't neccessarly need JS, just better CSS hacks.
jabbot pugs - 1248 - * semi-working external haskell; needs t 17:17
theorbtwo But the problem is that there's a few tests producing bum HTML for some odd reason that busts the layout.
gaal, mostly the CSS.
autrijus journal up... 17:26
but not really summarized any of your amazing development 17:27
I'm too drunk :-/
will try to make up tomorrow.
mugwump cheap date ;)
autrijus g'night!
verily.
Corion On the one photo I saw of autrijus, he looked really tired :) 17:30
obra heh 17:31
theorbtwo Eeep, the HTML looks fine in less, but quite messed up in view source.
jabbot pugs - 1249 - New logo? 17:37
autrijus mugwump: can you post a png for the logo? 17:42
autrijus stops downloading gimp 17:43
obra what's it start as?
obra has the gimp here ;) 17:44
autrijus docs/phd.xcf
in pugs
obra https server still dead
url?
autrijus just s/https/http/
obra k. svn catted it 17:45
whoa. gimp 2.2 is different 17:46
fsck.com/~jesse/hd.png 17:47
phd
autrijus er. eh. except I'm not a PhD :p 17:49
mugwump: clever work, tho
mugwump: although I don't think it fits in docs :) 17:50
theorbtwo guesses Larry probably does hold a PHD. 17:54
Khisanth don't think so :p 17:55
jabbot pugs - 1250 - Add TODO test for untangling -c and -e o 17:57
theorbtwo "Before he committed to a lifetime of systems administration and associated hacking, he and his wife were graduate students in the linguistics department at UC-Berkeley. "
...but I don't know if he graduated.
Corion ... untangling -c and -e handling will be ugly until we move to Perl parsing of the command line :)
theorbtwo He is a linguist in human as well as machine languages, reads classical Greek, and studied these as well as Chemistry and Music at Seattle Pacific University, U.C. Berkeley, and U.C.L.A. 17:59
theorbtwo shrugs.
There's a tendency, I've found, to want to split education into well-defined blocks called degrees.
elmex how many levels of intermediat languages will there be between perl6 and parrot? imc, pasm, bytecode? 18:00
theorbtwo Either you have a PHD or you don't; there is an implication that there is no value to going to grad school without getting a graduate degree, which isn't the case...
Depends where you place the somewhat arbitrary lines of what parrot is and what perl6 is.
autrijus <- junior high school dropout. 18:01
elmex parrot is that thing that isn't finished....
autrijus elmex: uh, no. it's pugs that isn't finished :)
Odin- Neither is parrot, really. ;)
elmex ah, i see, parrot is already rocksolid and finished ;)
autrijus parrot is in remarkably bettter shape than pugs (or perl6)
maturity wise. 18:02
it's quite enough for perl 6 to target.
elmex well, and what levels of intermediat languages will there be ?
Odin- Parrot needs something on top of it, to become useful. ;)
autrijus we generate parrot-ast
which is compiled to imc
then to pasm
then to pbc.
elmex that are heck of much levels, how do they want to make that fast ? (eg. like perl5?) 18:03
theorbtwo elmex: By having fast compilers for each step, and by caching when you can.
Corion Pugs is what forces the rest of the "Perl-6-world" to move, at least for now
elmex theorbtwo: wow, cool ;) that sounds 'easy' 18:04
Corion I say, eliminate them all and use Haskell as the base :)))
elmex: I look forward to a "port" of Coro to Pugs ! :)
elmex Corion: the coro author/programmer, marc lehmann sits right next to me ;) 18:05
Corion elmex: Ah, I thought you _were_ him (due to your login name) :)
elmex: So poke him a bit!
elmex lol ;)
Corion: he doesn't know perl6 good enough yet 18:06
Corion elmex: Bah! I don't know Perl6 either - I just program it as if I were programming in Perl5 :)
elmex or: he would know it, if he had time. but we're busy with reallife work atm
Corion elmex: Ah - so you're not at the Breakpoint demo party :) 18:07
elmex what thing?
whereis taht
autrijus elmex: so, pugs is damned fast.
on parrot.
and not exactly that fast on GHC.
elmex autrijus: pugs compiles to parrot already? 18:08
autrijus and it's much easier to optimize for parrot than for C--/GHC.
elmex: err sure. pugs --runparrot examples/mandel.p6
Corion elmex: The Breakpoint party is/was every Easter somewhere on the Rhine river. Marc might know it :)
elmex autrijus: ah
autrijus that's like two or three days ago.
elmex oh
Corion: he says he doesn't know ;) but he can;t remember cleanly 18:09
exactly
autrijus elmex: oh, say hi to marc for me, too.
I'm happy I no longer has to bother him with serializable continuations.
Corion autrijus: :))))
autrijus because it's much easier to do in pugs and parrot.
elmex autrijus: hi back 18:10
autrijus hi schmorp!
schmorp autrijus: hi :)
Corion Hi schmorp!
schmorp .. i join and the discussion dies... i am too late :) 18:11
Corion schmorp: Naah - I was just saying that I looked forward to the Perl6 port of Coro :)) 18:12
theorbtwo autrijus: Would running the test suite under the pugscc --run --foo be useful, for various foos? 18:13
kcwu hi schmorp!
Corion ... half-joking, as Coroutines are built-in in Perl6 I think, but I don't read the docs :)
pjcj autrijus: I have posted about coverage requirements to p6c, but I suspect it is in a moderation queue somewhere. Let me know if you need it before it shows up.
autrijus okay. I'll sleep soonish.
theorbtwo: not yet. need to finish on the fly compilation first 18:14
theorbtwo: then we may need a parrot run and a haskell run.
theorbtwo nods.
schmorp Corion: s/port/rewrite/ would probably be more realistic
Corion: how does parrot handle reentrancy problems?
autrijus Coro is builtin in perl6 as lazy given/take no? 18:15
Corion schmorp: Dunno - I haven't done any internals work
autrijus and is in parrot as ./classes/coroutine.pmc
also see ./t/pmc/coroutine.t
er, I mean gather/take
schmorp autrijus: yes, but how does it handle the c stack switches?
theorbtwo It should be possible to do it just by changing the test harness var. 18:16
autrijus schmorp: see parrot/classes/coroutine.c :D
(which is a funny way to say "I don't know and don't care as long as it works") 18:17
schmorp i don't think parrto will ever do more than "just work" 18:18
autrijus sure.
parrot has this very strong Right Thing design feeling.
elmex i've a bad feling 18:19
schmorp seriously? to me it lookas more as if a few people had their first try at doing a VM by emulating real hardware
autrijus seriously.
I had this kaffe main committer friend looking at parrot yesterday 18:20
he thinks parrot should be taught in schools as the example of how to design a VM right.
(I was surprised how good it is when I first looked at it seriously last week.)
schmorp strange - it is so inefficient iwth regards to JIT and compilation
autrijus *shrug* it worksforme. :)
schmorp for example, you need to dor egister allocation at least twice
autrijus and is faster than perl5 even if unannotated.
<- doesn't want much. 18:21
schmorp well, that is relatively easy :->
autrijus mmm maybe it is, but perl5 is actually the faster of the scripting languages.
schmorp i wonder how fast it would be if designed better 18:22
autrijus it can always use more hack :) 18:23
rgs perl 5.9 is faster than 5.8.
schmorp it is not question of mroe hack, it is a question of less hack
autrijus hack, as in improvement
not as in kluge.
schmorp yes
autrijus but really, it's fast enough for me, and I think I like how easy it is to generate imc, so I've got no complaints. 18:24
if you can find ways to improve its performance or redo something that runs imc but faster
I'm all ears :D
schmorp i just would have hoped that the parrot people would learn from the existing results we know about jit compilers and such
pjcj I suspect chip will consider even radical design changes
autrijus pjcj: right.
theorbtwo That was a large part of the point of making it register-based.
schmorp yes, so to jit compile parrot code one has to re-do register allocation another time 18:25
also, a register is not faster than memory in a vm
etc..
(but this is the wrong channel for parrot rants :)
autrijus right. 18:26
and I hope rants results in code :)
jabbot pugs - 1251 - One more permutation of '-c -e' gets now 18:27
schmorp unfortunately, not likely with me and parrot (i'd like, but i have too many things going :()
elmex hehe, if i would be mighty and wise enough to do a VM and JIT and allthat i would od it ;) *G*
autrijus elmex: "learn" :) 18:28
schmorp hmm, it seems parrot doesn'T handle reentrancy when coroutines are active
autrijus it's actually quite learnable. I learned a lot just by reading docs.
schmorp: oh, and I'm going to write a imcc implementation this week.
imc, rather.
elmex autrijus: i need many many years to do that 18:29
and then, perl7 will be out
autrijus and then you can help!
schmorp yes, elme xis very gifted :)= 18:30
autrijus: imc is a good target, it's not that efficient due to the extra layer, but it makes it possible to replace aprrot by a completely different design 18:32
autrijus yes.
or have switchable runcores or something. 18:33
it's very easy to implement imc :)
hi oylenshpeegul. 18:36
schmorp maybe in a yaer or so i'll implement my long-term-plan, using gcc to compile imc "just in time"
oylenshpeegul hi
autrijus that would be so nice.
schmorp++ # in advance
elmex lol
schmorp i am htinking about that for some years now, no matter how long startup tiems get
schmorp++ # need no such thing 18:37
perlbot What kind of idiot karmas himself? Your kind of idiot!
schmorp was it reste at leats?
elmex perlbot: karma schmorp 18:38
jabbot elmex: schmorp has karma of 2
perlbot schmorp doesn't have any karma
elmex ?Q
!
elmex++
perlbot What kind of idiot karmas himself? Your kind of idiot!
elmex perlbot: karma elmex
jabbot elmex: elmex has karma of 3
perlbot elmex doesn't have any karma
schmorp aahh 18:39
elmex lol, i rock
autrijus IMCC lexer translated 18:42
stub IMCC.hs too:
All that is gold does not glitter, 18:43
Not all those who wander are lost;
The old that is strong does not wither,
Deep roots are not reached by the frost.
now sleep. see you tomorrow. :) &
elmex who does come up with thos cites? 18:44
autrijus me.
elmex why not cite bible?
schmorp cu
autrijus because it is a perl cultural tradition to cite tolkien. 18:45
see perl5's source code.
elmex why
autrijus ask larry.
theorbtwo Interesting side note: Larry is a wacko bible freak, and chose to use Tolkien quotes.
autrijus he is also a wacko middle earth freak. 18:46
(as is jhi, as am I)
rgs jhi is a wacko discworld freak.
autrijus rgs: what kind of wacko freak are you?
jabbot pugs - 1252 - * transcribed IMCC lexer 18:47
rgs I'm not a freak !
Corion ook
rgs well, let's assume I'm a wacko jazz freak. Difficult to quote. 18:48
Corion I could add some wacko quotes from X-Men and Sandman comics. Hmmm ;-)
rgs: Ah, bababadudbababdububababa # freejazz
elmex i never liked lord of the ringens much
rgs neither did I, to be honest. But quotes are cool. 18:49
elmex childisch :)
rgs that's how the perl subculture comes into existence.
elmex but ... i'm a child still myself
rgs elmex: are you german ? 18:50
elmex lol, yes
my english isn't the best
rgs you make german mispellings :)
elmex those where only typos
because i wrote it too quick 18:51
rgs i make french mispellings too. exemple, connexion.
elmex heh 18:52
theorbtwo Odd, I wouldn't think ringens would be a german typo -- the german title is Herr der Ringe, so I should think it'd get mistylped ringes. 18:53
elmex lol, indeed 18:54
rgs theorbtwo: "childisch"
elmex that was misspelling on purpose
pjcj Lord of the Rings is on Swiss, Austrian and German TV at the moment. Spookily enough.
rgs _and_ on French TV.
theorbtwo shrugs -- I have all three on extended ed DVD. 18:55
autrijus so, if leo shows up on #perl6 here later and I'm asleep, please keep him comfortable and happy :)
autrijus waves &
theorbtwo Sleep well, autrijus.
pjcj I obviously don't get that French channel, but three showing the same film is enough.
theorbtwo We only get the Austrian and German channels here, and ORF1 barely comes in. 18:57
Corion Heh. Mad props to GHC / Haskell - it detects edge cases of your function definitions ;-) 18:58
pjcj leo on irc?! 18:59
Corion Hi gaal! 19:09
I'm adding/cleaning the Haskell command line handling, as an exercise :) 19:10
gaal Corion++
If you hadn't beat me to it, I'd try it myself 19:12
theorbtwo ponders... the interface with the mouseover thingy directly below the cell is somewhat easier to read, but makes navigation harder. 19:13
(Because the tip obscures the next one down, so you have to mouse out and then back over to see the tip of that one. 19:14
Corion gaal: Feel free to also work on it :)
theorbtwo Also, I have no clue how to work around this firefox bug.
(And I'm fairly certian that's what it is.)
Corion theorbtwo: Hmm - maybe make it appear above instead - as most people will navigate downwards :))
theorbtwo That would imply that I can figure out how to do that. 19:15
CSS positioning seems to behave more like a black art then a well-specified language.
Corion theorbtwo: Aiee :-) Well - consider simply using the overlib JS package that creates JS tooltips that appear right-down of the mouse 19:16
theorbtwo JS--
Corion theorbtwo: Ah, yes - I've sold part of my soul to the gods of JS, true ... 19:17
theorbtwo: But overlib most likely takes care of the ugly parts. On the other hand, that doesn't degrade nicely for non-JS browsers. 19:18
... maybe see what Eric Meyer has to offer in his ComplexSpiral demos ?
theorbtwo Ah-ha!
Some of my HTMLification regexes were missing /g options. 19:19
gaal Corion: Autrijus mentioned GetOpt, which has docs at www.haskell.org/ghc/docs/latest/htm...etOpt.html
Corion gaal: Will look at these :) 19:20
gaal it it's GNUish, I think, so it certainly doesn't do what we want, but if like me you need plenty of exaples, it might help
..doesn't do *all* we want
Corion gaal: I'll look if it does proper collecting of multiple -e fragments, which makes all the rest ugly :) 19:21
gaal i doubt it. 19:23
Corion gaal: Ugh :) 19:25
gaal actually he was talking about System.Console.GetOpt <www.haskell.org/ghc/docs/latest/htm...t.html> but at a glance it looks pretty much the same thing.
well, p5's command line is pretty funky. 19:26
you need potentially a specifyable dispatch table.
well, if you want to write a library for it
if you don't, you just need a dispatch table :)
vladtz hello, I was looking at the goto tests and noticed that you can't use return without a value 19:28
gaal i wrote them, let me see 19:29
Corion gaal: Hmmm. Currently I'm thinking of collecting all the "special" stuff in Named Field or whatever. But first, I want to flesh out -ce and -c -e and -e print -c (which doesn't work)
gaal btw they're all commented out because of a hard parsefail
vladtz I have an adjusted gotos test, which run's 6/8 tests IÄŗl paste 'm
gaal Corion: do declustering in one swoop maybe, instead of trying to parse every combination separately. 19:30
pasteling "vladtz" at 80.127.84.50 pasted "goto return (void)" (42 lines, 825B) at sial.org/pbot/8670
Corion gaal: Yep - that's what I'm currently doing :) Maybe I should simply sort the options before parsing them :))) 19:31
gaal no no, don't do that
Corion gaal: But first, I'll have to get Pugs into a compilable state again before committing :)
gaal: Why not?
gaal some opts have very short effect which can then be changed
man perlrun and read -l and -0 carefully
vladtz gaal: the missing value on return caused the parsefail 19:32
gaal vladtz++, good catch, thanks.
though the error coordinates were messed up
vladtz your welcome
Corion gaal: Hmm - I thought that C<-l> and C<-0> merely inserted Perl code ?
gaal and also, putting this in eval didn't help 19:33
Corion (via perl -MO=Deparse -le "print")
gaal no, see perl.c in p5
even if they did
you couldn't rearrange its order and maintain correctness 19:34
Corion gaal: Ah, C source diving :( I had hoped to (later) just replace C<-l> by C< -e "BEGIN{...}" >
gaal s/though// somewhere up there
Corion gaal: No - I plan to keep all C<-e>, C<-l>, C<-0> arguments in their relative order (and what other cases there are).
gaal: Well, "plan" is a big word for it :) 19:35
gaal: But stripping out the C<-c>, C<-w> and C<-d> stuff would make that check much more convenient, true!
gaal well, arguably this (declerative -l, -0) isn't a very important feature so perhaps you could bring it up for retirement on p6l? the command line isn't specced yet 19:36
vladtz, did you commit? 19:37
or shall i?
vladtz please do, I have no comit access
gaal ah. that should definitely change :)
any metacommiters around?
now that you've made it parse, there are plenty of more things to test with goto :) 19:38
theorbtwo Hmm? 19:39
Corion gaal: Never! I want -l and -0 as I use them often. In fact, my super-secret plan is the C<-g> option, which will make Pugs act mostly like grep.
theorbtwo vladtz, email? 19:40
gaal oh, -l and -0 of course, but maybe remove -0nnn -e 'code' -0 -mmm -e 'more'
that example may not in fact be a good one 19:41
Corion gaal: I doubt that this works under P5 currently, but it would be nice to make it work under P6 :) 19:42
gaal: (a bit overboard, maybe :) )
vladtz [email@hidden.address]
jabbot pugs - 1253 - Added TODO test for unknown command line 19:47
gaal reading it again, it looks like the main issue is that the order of -l and -0 matters, and some weird interactions may occur between the actual command line and the shebang line.
Corion gaal: Ah well - those fine points can be cleaned up later :)
gaal we need to remember the shebang line, because the kernel doesn't do that for us on windows
Corion Is there anything like the "part()" function for Haskell, to part one list into two (or more) buckets ?
gaal: Ah, but the shebang "line" is available via getArgs() anyway :) 19:48
Corion goes looking in the Prelude docs
gaal zvon.org/other/haskell/Outputprelude/ - the other day i saw exactly what you're looking for 19:49
but i forgot what it was :)
Corion gaal: I forgot as well :)
gaal: But I find no inverse function of "zip" or a "grep" into multiple lists there ... 19:50
gaal no multiple return values in haskell
you can return a list of course 19:51
but not two
Corion gaal: Bah. Stupid language :(
gaal so you can return [[a], [b]]
:)
Corion gaal: Yeah,
theorbtwo Corion, you can only return one list in perl, too.
Corion gaal: part() will be a good exercise in generic/general Haskell, and it'll make the later porting to Perl6 easier
gaal heh 19:52
Corion theorbtwo: True, but at least I have the choice of returning a list or a reference to a list, while in Haskell, I have to return a reference to a list :)
gaal he's hot a point there, to2 does
heh, Corion, you (and I) were looking for partition :) 19:54
in List
Corion gaal: Aaah ! One less exercise to do for me :)
gaal and if you want two tuples, use break 19:55
hey you can still write it yourself :)
vladtz theorbtwo: can you remove my e-mail from the log (silly me)? 19:58
Corion gaal: I'll be starting with sortBy(), which implements a stable sorting mechanism, which is what I'll want/need :)
theorbtwo No, I don't have access to that.
vladtz pitty
theorbtwo vladtz, you should be able to checkin now. Please add yourself to AUTHORS as well. 20:00
Sorry I was so slow; Bond is distracting. 20:01
gaal but please svn up before you do more work on goto, i just checked in your fix. 20:02
jabbot pugs - 1254 - fixed syntax error causing parsefail (vl 20:07
vladtz gaal: just missed your message, backed out, did you add the test for empty return somewhere? 20:12
gaal as a parsefail? no, good idea 20:13
can you add it?
prolly in t/syntax or maybe t/subroutine
the former, actually.
vladtz yes IÄŗl add it to goto.t?
gaal no, since it's probably isolatable. 20:14
vladtz subroutine (always confused by former and latter).
gaal i meant syntax :) but whichever makes more sense to you. remember to use fail("FIXME parsefail") if it still hard-parsefails. 20:15
vladtz it fits nicely into t/subroutines/sub_return_values.t 20:18
gaal cool 20:19
vladtz zzzzzzz 20:23
jabbot pugs - 1256 - 20:37
pugs - 1255 -
theorbtwo 's going to be AFK for at least an hour, possibly much longer -- I'm going to ci my current testgraph stuff -- it's certianly not bugless, but no more buggy then the current. 20:38
kungfuftr evening all 20:42
Corion Hmm. Sorting the command line seems to "work", at least in my testbed ... 20:46
jabbot pugs - 1257 - testgraph: CSS and HTML-escaping updates 20:47
cm hello 20:54
Corion lr!
Yay! My -c / -e option sorter seems to pass its TODO tests now, at the price of a Haskell warning :) 21:06
gaal congrats, Corion :) 21:07
pjcj u
Corion gaal: Thanks - I haven't tested it yet against the rest of the suite :))
Limbic_Region hey Corion 21:08
Corion Hi li!
err - lr, I meant :)
gaal &
Limbic_Region tis ok 21:09
Corion It's some ugly code I've created there, but it should be better than what we have now :)
kungfuftr pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m | FBSD_4.11:r1257(2/3212) win2k:r1197(105/3116) Linux:r1197(105/3116) MacOSX:r1197(105/3116) **[793 tests TODO]** 21:26
pugscode.org <<Overview Journal>> | pugs.kwiki.org | logged: xrl.us/e98m | FBSD_4.11:r1257(2/3212) win2k:r1197(105/3116) Linux:r1197(105/3116) MacOSX:r1197(105/3116) **[793 tests TODO]** 21:27
kungfuftr bah 21:27
cm ..been kung-fu fighting..
kungfuftr HUH! 21:28
fritzz install pubs atm
on windows ...
:|
kungfuftr fritzz: that'll be a pint of bitter for me then 21:29
fritzz hehe
Corion Shouldn't this be valid Perl6 code? ("-h", "--help").map():{ $_, "-w $_", "$_ -w", "-w $_ -w" } 21:30
(that is, (list).map(){block} ? )
fritzz i just installed GHC 21:33
added it to my path
but when i perl Makefile.pl: *** Cannot find a runnable 'ghc' from path 21:34
when i type ghc, its there ...
cm reboot :D
fritzz lol
cm close the console
open it again 21:35
fritzz i did
cm for that the environment variables get updated.. :|
hmm
fritzz if i type ghc, i can reach the bin
cm bizarreness
fritzz ghc: no input files
yup yup
Limbic_Region is there a space in the path to where ghc is? 21:36
IIRC, there is/was a bug with that 21:37
Corion writes a parsefail todo test to test for C<< ("foo","bar").map() >>, because C<< ("foo","bar").sort() >> works :)
What is the protocol for adding parse fail tests like mine above? 21:38
fritzz ill just reboot
darn windowzz !
Limbic_Region Corion - I will search the channel log
Corion Limbic_Region: Thanks! 21:39
Limbic_Region I remember there being something said about not commenting it out as it will never be uncommented
kungfuftr Corion: todo_eval ? 21:40
Corion kungfuftr: Don't ask me :) I'll use whatever you suggest :)
Limbic_Region thinks kungfuftr is correct Corion - still searching though 21:42
2087 [21:14] <gaal> test writers: please mark your failing tests (those that must be commented out because even eval doesn't protect from them) with "fail('FIXME: parsefail');" 21:43
that was on 2005-03-16
and then later on
fritzz jeez
same thing after a reboot
cm might it be the spaes? :o 21:44
Limbic_Region on the 17th
it looks like nothingmuch added the todo_eval stuff Corion
fritzz i added it to the user path, let me try with the system path
Limbic_Region fritzz - for the 3rd time - are there spaces in the path to ghc?
IIRC there is/was a bug with that
fritzz yes 21:45
%20 ?
Corion Limbic_Region: Heh. So it's gaal vs. nothingmuch ;-)
Limbic_Region no Corion
gaal was present in the later conversation on the 17th
Corion Limbic_Region: Ah :)
Limbic_Region fritzz - not url spaces
cm fritzz: use the short name
fritzz ah !
cm fritzz: like Progra~1 instead of Program Files, or such
fritzz sorry for being so lame folks
cm i believe that could work :| 21:46
Limbic_Region or use subst to give pugs its own drive letter
Limbic_Region has a tendency to bend things to his will versus just fixing them
fritzz work now :)
Limbic_Region now if you had only listened to us 10 minutes ago
:P
fritzz compiling in progress 21:47
can i submit report ?
Limbic_Region Corion - colabti.de/irclogger//irclogger_log...xt=checked
check the second to the last entry on the 16th
and then read the 17th
you be the judge of what is correct
fritzz pugs installed successfully :) 21:51
cm on win32? :) 21:57
jabbot pugs - 1259 - Parse bug: Map applied to constant list
pugs - 1258 - Partially reworked command line parsing
cm fritzz: nice. what kind of requirements does it have besides GHC?
fritzz: Perl 5?
fritzz sound like
look like he installed itself in my perl5 dir
now i can finally mess with perl6 named arguments :) 21:58
Corion I went with the todo_eval_is route :)
r1259(0/3221) 22:01
(on win2k)
... and I should maybe change 5 of my unexpectedly succeeding tests from todo_ to real tests :) 22:02
Are the current pugs build number stored in any internal variables ? Like somewhere in (the Perl6 equivalent of) Config.pm ? Or $], whatever its name now is ? 22:06
integral pugs_config.h and pugs_version.h contain #defines, as used in Help.hs
The Help module also exports some version information variables, which then don't get used in Config.hs 22:08
Corion integral: :)) So it's not readily available from within Perl6 code :)
integral Not yet :-) I'm about to sleep, but if I wasn't, I'd hack Config.hs to use Help.hs or the two header files 22:09
Corion integral+
integral++
integral hmm, actually, ... 22:10
Corion Enoguh hacking from me - tomorrow, I might add the implicit conversion of C<-l> to C<-e "BEGIN{...}"> - now I just need to find out what to put there instead of the dots :)
Gah. Pugs doesn't even like "BEGIN { ... } " yet :) 22:13
fritzz my precious BEGIN
usefull to parse config file onload :) 22:14
Corion fritzz: Also very much needed to implement C< -l > :) 22:15
(the command line option, not the -X function)
fritzz what is C< -l > ? 22:16
22:16 khisanth_ is now known as Khisanth
Corion fritzz: The "-l" command line option (I use Pod escape syntax) 22:18
fritzz: See perldoc perlrun
nothingmuch stevan: ping 22:27
do we have any vim ninjas present? 22:30
rgs nothingmuch: ask 22:32
nothingmuch perl6.vim should not gobble up =kwid, =pod, =etc
they should be treated as comments
at least for starters 22:33
it is giving me a hell of a time trying to parse the HTML output of Text::VimColors, as I need to flatten some <spans>, due to the way it currently parses L<S02/"Section" /skip regex/> stuf
rgs show me a file that demonstrates the bad behaviour 22:34
nothingmuch t/operators/hyper.t 22:35
Corion Err - how do I specify an "\n" string literal in Haskell ? is chr(13) the only way ? 22:36
pjcj What's the problem with using \n? That should be fine, I think. 22:42
Corion pjcj: At least with GHC, I get a literal "\n" in the string then :) 22:43
rgs nothingmuch: do I put only "=cut" as an end marker ? 22:44
nothingmuch uhuh
afaik
either way, it should do for now
pjcj Corion: go into ghci, type 'putStrLn "a\nb"' 22:45
rgs nothingmuch: Lexer.hs only lists =pod and =cut as markers.
Corion pjcj: Weird... I'll try again ... Maybe putStrLn interprets stuff?
nothingmuch i think it skips =anything
rgs ( in comments )
nothingmuch since =kwid doesn't seem to bother it 22:46
rgs ok
Corion Yeah - simply typing "a\nb" returns "a\nb" and not "a<newline>b"
rgs ok, Parser.hs has the definition of pod blocks 22:47
nothingmuch: commited 22:48
nothingmuch rgs++ 22:49
do you have a colorscheme that could work well?
the pink one is not that nice
it shares too much colors 22:50
rgs I don't know
nothingmuch and i don't think it's fair to impose mine
it's a bit too nasty
rgs I'm colour blind. Everyone finds my custom colorscheme ugly
Juerd I'm colour deaf
I can't hear the difference between green and red very well 22:51
rgs: What colours can you see?
nothingmuch hmmm
rgs Juerd: this question doesn't make much sense 22:52
nothingmuch syn region p6POD start="^=[a-z]" end="^=cut" <-- makes perfect sense, but doesn't work for me
perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
Juerd rgs: Why not? Do you see no colour at all then? 22:53
rgs Juerd: I see colors. Less than other people.
Besides this, I have difficulties remembering color names in English.
Juerd Do you know as which you see them?
rgs I can't see violet, for example 22:54
pasteling "nothingmuch" at 212.143.91.217 pasted "how perl6.vim colors pod:" (5 lines, 163B) at sial.org/pbot/8675
Juerd rgs: Never go to Belgium by the way - the traffic lights there do not make sense if you can't see the difference between colours 22:55
They don't consistently stick to the red/yellow/green pattern 22:56
nothingmuch sideways and stuff?
theorbtwo returns.
nothingmuch hurrah! theorbtwo is here!@
Juerd nothingmuch: Well, there are traffic lights with dual yellow, or 2-column thingies where next to the main green light is a red arrow 22:57
theorbtwo Have I been missed?
jabbot pugs - 1260 - Treat pod blocks as comments w.r.t. vim
pjcj Corion: \n is a single character. '\n' is legal and length "a\nb" is 3.
nothingmuch oh my
Juerd nothingmuch: And green/red/yellow/green happens too
rgs Juerd: I've been to Brussels twice
Juerd rgs: Ah, glad you survived :)
Juerd thinks most of these weird lights are in Antwerp 22:58
The city you drive into at 120 km/h, after which the motorway suddenly becomes a normal road with traffic lights and everything
nothingmuch rgs: any clue why that still happens?
rgs nothingmuch: well, in my tests the whole pod block is highlighted in the p6POD class 23:00
nothingmuch that's odd
i just noticed it's only doing that for HTML output
and that :TOhtml from that window works properly 23:01
rgs FWIW I have this in my &stl option, to debug highlighting : (%{synIDattr(synID(line('.'),col('.'),0),'name')})
nothingmuch ah, i think i see: 23:04
:let perl_include_pod=1
Text::VimColors is passing that,
rgs this variable is used by syntax/perl.vim 23:05
time to sleep. 23:10
stevan nothingmuch: hello 23:13
nothingmuch hola
stevan you pinged?
nothingmuch yeah, but i forgot why
stevan ok
nothingmuch =) 23:14
stevan pjcj: was that your email about code coverage on p6c? 23:15
Corion pjcj: Weird - I must have done something wrong before, because now it works exactly as you say (and what I expected when I started :) )
pjcj: Thanks!
pjcj stevan: yeah - I mentioned it briefly here and autrijus asked for a mail to p6c 23:16
Corion: good ;-)
stevan Devel::Cover for perl6 would be really cool. I hope that works out
Corion promotes a todo_is test to is() :)
Win2k:r1261(0/3229) 23:18
good night all! My work is done by tainting Main.hs with lots of ugly code ... 23:20
nothingmuch is frustrated 23:29
theorbtwo Eh, nm? 23:32
With the syntax hilighting for the tests?
I'm rather fustrated by this apparent firefox bug. 23:34
nothingmuch yes 23:35
vim highlights ok, but then 'source /usr/local/lib/perl5/site_perl/5.9.1/Text/VimColor/mark.vim' yields:
>Comment>=pod<Comment<
>>Hyper operators L&lS<<>Number>03<Number<>>/<<>Constant>"Hyper operators"<Constant<>>&g<<
stupid
jabbot pugs - 1261 - Multiple -e switches work now 23:37
pugs - 1262 - fixed AUTHOR in pugscc; worked on PodStr 23:57

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!