6.2.10 released! xrl.us/hxnb | geoffb's column (/. ed): xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6 | www.geeksunite.net
Set by integral on 25 October 2005.
teddyns eh :) 00:00
r0nny and even the spec isnt complete
Odin-LAP Nor will it ever be, I suspect. 00:01
r0nny anyone here knows the svn url of parrot ?
Odin-LAP: i think around 2007 the spec is done ;P 00:02
geoffb r0nny, svn.perl.org/parrot/trunk
r0nny thx
teddyns why so long ? :-/
geoffb np
Odin-LAP r0nny: Until someone comes up with a really cool idea, perhaps. ;)
r0nny hmm
geoffb teddyns, there is MUCH that the designers needed to learn to make a language so all-powerful.
(That sounds like a joke. It's not.) 00:03
webmind ey Odin-LAP :)
r0nny well
they will never reach perl7
Odin-LAP geoffb: Yeah. Substantial parts of which were the implementation of CommonLisp-with-syntax. :)
teddyns omg :/
geoffb Odin-LAP, heh 00:04
webmind uhm
Odin-LAP webmind: Greetinks. You're here too, eh?
teddyns ,thats too bad
webmind perl6 -will- finish imho
r0nny since murphys law tells its not possible
webmind just takes the time it'll take
Odin-LAP, ofcourse.. big perl6 fan :)
teddyns we are able to run 00:05
perl 5.8 scripts with perl6 right ?
r0nny yeah
Odin-LAP geoffb: Which, I note, is not a trivial accomplishment. It's quite amazing, how well it seems to have succeeded.
r0nny bttw - will pugs die if my libperl is with ithreads, and i use threads ?
geoffb Odin-LAP, nodnod 00:06
Odin-LAP webmind: Well, I wonder at the definition of 'finish'. Sure, you'll have it in a running, production-ready form eventually.
webmind Odin-LAP, well it's extendable 00:07
Odin-LAP, but I see finish as a production ready product
with a stable syntax 00:08
perl6 is certainly heading that way
with goos speed
good
azurot_tafe is there anywhere I can find all the synopses? or are there just numbers missing? 00:16
r0nny install Perl6::Bible from cpan
azurot_tafe ahh, ok 00:17
geoffb azurot_tafe, there are still some missing, but that's as complete as it gets outside of @larry's collective mind
azurot_tafe cool. thanks :D 00:18
geoffb np 00:19
azurot_tafe ah, so it looks like all of the available ones are up on eric256's feather anyway 00:22
oh no, never mind
Juerd 01:27 < azurot_tafe> ah, so it looks like all of the available ones are up on eric256's feather anyway 00:43
eric256's feather? :)
Khisanth yes, he rooted the box yesterday! :P 00:44
Juerd Heh
People needing root access can just ask for it :)
If I trust them, and they really need it (or it would help me much if they had it), they can get it 00:45
sili i need root 00:49
00:50 rafl_ is now known as rafl
azurot_tafe ;-p I mean, his feather.perl6.something/~eric256 ^_^ 00:58
byesss
:q
Juerd sili: Why? 01:03
sili bragging rights
and i really need them.
Juerd ok 01:04
sili i'd actually like to figure out how parrot works
Juerd You don't need root access for that :)
rafl autrijus: ping 01:32
svnbot6 r7814 | stevan++ | Perl6::ObjectSpace - 02:27
r7814 | stevan++ | * added optional parameters to closure
r7814 | stevan++ | * cleaned up some more of the bootstrap code
r7814 | stevan++ | - added some tests for the Class (simple ones for now)
06:38 scook0_ is now known as scook0
coral does VInt have an upto method (from ruby) or equivalent? 09:42
VInt.upto(n) to get VInt..n in perl
nwc10 autrijus: in the S17 draft there's no mention about sharing-or-not of variables. 10:11
oh wait. I think I see. 'shared, transactional variables by default' 10:12
blackman selam kanal 10:32
hi?
nwc10 Hi
blackman where are you?
nwc10 London 10:33
blackman ok
turkey
dou you know??
nwc10 Ah right. I was trying to work out what language "selam" was in so I could find something to translate it
blackman :))) 10:34
you are f or m?
nwc10 M
blackman me too
clkao nwc10: you know there's london pm meeting tomorrow and first timer can get free beer 10:37
nwc10 I didn't know the second part. Or at least notice
clkao pretends nwc10 is newcomer, at least on #perl6 :P
nwc10 I've been here before. 10:38
clkao scarily you are the guy behind my computer
azuroth ?eval 10 | 2 11:18
evalbot_7814 2
azuroth ?eval 10 +| 2
evalbot_7814 10
azuroth ?eval 10 +| 3
evalbot_7814 11
azuroth ?eval 10 | 3
evalbot_7814 3
rafl_ Shouldn't that return a junction? 11:21
azuroth ohh right
rafl_ ?eval (10|3) 11:22
evalbot_7814 3
rafl_ ?eval (10|3|10|10|10).pick
evalbot_7814 3
rafl_ ?eval (10|3|10|10|10).pick
evalbot_7814 3
wolverian evalbot doesn't output junctions right 11:23
azuroth hmm
wolverian ?eval any "a".."d" 11:24
evalbot_7814 "a"
wolverian works in pugs itself.
azuroth ahh 11:25
junctions are so cool 11:26
webmind hmm /me keeps miing up names 11:27
what are >>*<< contrustructions calles again ?
wolverian hyperops
webmind right
wolverian see S03, "Hyper operators" 11:28
(we need an @index into the synopses.. any takers? :)
(a la lambdabot)
azuroth night 11:29
wolverian good night 11:30
azuroth ?eval 5 < 10 < 15 11:49
evalbot_7814 bool::true
azuroth ?eval 5 < 2 < 15
evalbot_7814 bool::false
azuroth ?eval 5 < 20 < 15
evalbot_7814 bool::false
12:23 Lopo_ is now known as Lopo
svnbot6 r7815 | iblech++ | * Pugs.Parser: Parse \ in subroutine signatures (sub foo (\$arglist)), 13:16
r7815 | iblech++ | so we don't have to add eval()s around arglist tests.
r7815 | iblech++ | * t/data_types/lists.t: \(1,2,3) does no longer create an array of refs,
r7815 | iblech++ | per latest S03 (r6417).
r7815 | iblech++ | * New t/data_types/arglist.t: Added tests for arglists/tuples (\(1,2,3) creates
r7815 | iblech++ | an arglist now, as does sub foo (\$arglist); *$arglist flattens arglists).
verbalkent im new to perl, are there any online begginer sites,a or communities? 15:00
webmind perl or perl6 ?
verbalkent regular perl i suppose 15:01
Juerd verbalkent: perlmonks.org 15:02
verbalkent thnk u sir
PerlJam heh "regular perl" 15:06
That may just become a common phrase in the not too distant future
integral vs. "diet perl" and "new perl"?
liz6 cherry perl? 15:07
PerlJam all of the above
Khisanth PerlJam: and that would be perl6's fault!
PerlJam Khisanth: no, it would be the fault of those people who resist the perl6!
verbalkent :( did i come off too n00bish? 15:21
integral verbalkent: IRC quickly teaches people to spell :)
verbalkent spelling is overated 15:22
theorbtwo Spelling is one of the primary ways to size up your character in a text-based environment. 15:23
eric256 feather.perl6.nl:8027/cgi-bin/wiki.pl slowest incomplete implementation of a wiki ever ;)
webmind perl6 wiki?
theorbtwo Of course it lies just as much as personal grooming and clothing taste, the pimary way to size up your character in a real life environment.
eric256 porting www.perlmonks.org/?node_id=503926 to p6
verbalkent hmm, cant argue witht hat 15:24
spelling means alot, but grammer and punctuation must be fodder in your case :p 15:25
integral "can't"
"a lot"
theorbtwo Spelling, grammar, and puncuation are all important -- grammar quite likely the most important. 15:26
verbalkent integral has the grasp
lol, sort of like a pocket secretary
PerlJam "secretary"? 15:28
verbalkent exactly
the secret that scared the w out of wary
steve197888888 12Hi all.. 15:45
liz6 yo!
steve197888888 12I have a small Problem: $data->{Body}->{Call}->{return}->{item} is an array
12but I don't know how to run through it.
12foreach doesn't work. 15:46
eric256 @{$data->{Body}->{Call}->{return}->{item}}
btw stever this is a channel about perl6 now perl5...you probably want #perl
steve197888888 12I wanted to go to the newbee channel.. 15:47
12they told me to get here :)
liz6 who's they?
steve197888888 12the guys from the perl channel
eric256 they must have misunderstood and thought you meant new perl 15:48
steve197888888 12OK I C.
12Thanks Wi will go there then eric. Thanks for your help! 15:49
kolibrie eric256++ # very diplomatic
eric256 kolibrie...its early, i promise to get grouchy later. ;) 15:50
whats the syntax for perl5 regexes ? 15:51
theorbtwo rx:perl5/.../, IIRC.
eric256 thanks 15:52
rafl Or rx:p5//
nothingm1ch obra: ping 16:22
eric256 lol if you use a given block in the main body of a program...it seems to be ending the program when it matches a block...seems being key here. 16:23
obra hey
eric256 ?eval given 1 { when 1 { say "test" } } say "test2";
evalbot_7815 test test2 bool::true
obra nothingm1ch: what's up? 16:24
eric256 hmmm... or not.
nothingm1ch is rt.cpan.org still being upgraded?
obra yes. I spent much of yesterday gnashing my teeth after doing a full import and having the new server cry 16:25
nothingm1ch ouch
just wanted to make sure because I suspected cache issues or something
obra hm? what cache issues? 16:26
still not seeing a new server? ;)
nothingm1ch i'm seeing "Unsatisfied dependency chain in Joins Users_2 at ..."
obra ?! 16:27
ok. hang on a sec
try now?
nothingm1ch obra: thanks 16:36
sorry for not noticing
obra no worries. glad it's better 16:37
nothingm1ch now i forgot what I actually wanted to do in rt.cpan.org ;-)
obra :/
nothingm1ch i think i just wanted to link someone 16:38
Juerd eric256: What makes it slow? You didn't use a sound software design, I hope. 16:47
eric256 Juerd. loading HTML::Template and Textile are whats slow. but i plan on factoring those out once its working 16:49
just using them as a crutch for the moment 16:50
Juerd Use regexes instead of templating
I don't know what Textile does
eric256 it handles all the Wiki formating.
Juerd Go back in time if you want to use pugs in production, especially when it's CGI
eric256 like i said, they will go once i get the rest converted
Juerd Consider mod_pugs 16:51
eric256 link? 16:52
Juerd undef 16:53
eric256 nearly has new,edit,and view completly ported. ;) then just build my own template engine. ;) 17:10
obra isn't micromason already ported?
eric256 ?eval my %wiki; my $topic = "hello"; %wiki<$topic> = "hello"; %wiki.perl.say; 17:11
evalbot_7815 {("$topic" => "hello"),} bool::true
eric256 what am i doing wrong there?
theorbtwo Looks fine to me. 17:12
(Except for the redundant say.)
eric256 look at the output
theorbtwo Oh, wait, nevermind.
eric256 lol
theorbtwo You want curlies, not angles.
eric256 thanks. figured it was something like that
theorbtwo ?eval my %wiki; my $topic = "hello"; %wiki{$topic} = "hello"; %wiki.perl.say; 17:13
= "hello"; %wiki.perl.say;
?eval my %wiki; my $topic = "hello"; %wiki{$topic} = "hello"; %wiki.perl.say;
evalbot_7815 {("hello" => "hello"),} bool::true
xerox Why ".perl.say" ?
eric256 becausei felt like it ;)
xerox ?eval my %wiki; my $topic = "hello"; %wiki{$topic} = "hello"; %wiki.say; 17:14
evalbot_7815 hellohello bool::true
xerox Ah, like a prettyprinter.
theorbtwo .perl prettyprints; .say outputs.
xerox ?eval my %wiki; my $topic = "hello"; %wiki{$topic} = "hello"; %wiki.perl;
evalbot_7815 "\{(\"hello\" => \"hello\"),}"
xerox I see :D
eric256 ?eval my %wiki; my $topic = "hello"; %wiki{$topic} = "hello"; %wiki; 17:15
evalbot_7815 {("hello" => "hello"),}
eric256 eval bot does some magic for us ;)
theorbtwo evalbot++
eric256 can someone help me convert a p5 regex to a rule? \b([A-Z]\w+[A-Z]\w+) 17:17
its for WikiWords
?eval "this is a TesT" ~~ /\b([A-Z]\w+[A-Z]\w+/ 17:18
evalbot_7815 Error: cannot cast from VUndef to Pugs.AST.Internals.VCode (VCode)
eric256 ?eval "this is a TesT" ~~ /\b([A-Z]\w+[A-Z]\w+)/ 17:19
evalbot_7815 Error: cannot cast from VUndef to Pugs.AST.Internals.VCode (VCode)
PerlJam eric256: \b(<[A..Z]>\w+<[A..Z]>\w+)
althought I'm not sure PGE knows about \b yet.
(heck, I don't even know if it groks \w)
eric256 ?eval my $x = "hello"; $x ~~ /\w/; 17:20
evalbot_7815 Error: cannot cast from VUndef to Pugs.AST.Internals.VCode (VCode)
eric256 thats not the same error i get on the command line...soo odd 17:21
no error on command line, error about PGE.pbc in interactive and that completely different error in eval bot 17:22
theorbtwo Whee, eric.
Limbic_Region ?eval say [+] 1,2,3 17:23
evalbot_7815 6 bool::true
theorbtwo Does no error imply that it works as expected?
eric256 theorbtwo...not a clue what it implies.
theorbtwo ?eval say [.{}] {a=>{b=>{c=>42}}}, qw(a b c)
evalbot_7815 42 bool::true 17:24
Limbic_Region ?eval say [+] @_
evalbot_7815 Error: Undeclared variable: "@_"
eric256 doesn;'t matter. pugs doesn't implement subst yet.....so how do i do regex replacements in pugs right now? 17:25
Limbic_Region used Pugs to win a golf contest 17:28
perlmonks.org/index.pl?node_id=504986
eric256 ?eval my $entry = "this is a TesT"; $entry ~~ s:perl5/\b([A-Z]\w+[A-Z]\w+)/"<a href=''>$1</a>"/; $entry; 17:29
evalbot_7815 \"this is a TesT"
eric256 sure... running that in my prog gives me the "*** Can't modify constant item: VUndef "error 17:30
geoffb Limbic_Region++ 17:33
eric256 ?eval my $entry = "this is a TesT"; $entry ~~ s:perl5/\s([A-Z]\w+[A-Z]\w+)/[$0]/; $entry;
evalbot_7815 \"this is a TesT"
eric256 anyone got a clue on this one? /me is frustrated. lol 17:34
geoffb ?eval my $entry = "this is a TesT" ~ ""; $entry ~~ s:perl5/\s([A-Z]\w+[A-Z]\w+)/[$0]/; $entry; 17:35
evalbot_7815 \"this is a TesT"
geoffb ?eval my $entry = "this is a TesT" ~ "a"; $entry ~~ s:perl5/\s([A-Z]\w+[A-Z]\w+)/[$0]/; $entry;
evalbot_7815 \"this is a[TesTa]"
geoffb You weren't matching the second \w+ 17:36
eric256 hehehe. duh
geoffb :-)
eric256 thanks. me talks a walk around the room breathing deeply. thats what i get for using someone elses regex! 17:37
geoffb :-)
.oO( The first \w+ should probably be \w+? and the second \w* )
 17:38
eric256 yea thats what i changed to. 17:39
feather.perl6.nl:8027/cgi-bin/wiki.pl?topic=Test thanks geoffb! 17:40
working wiki ~~ ish 17:44
obra eric256: have you considered porting cwest's minimalist wiki? 17:52
eric256 i dont know what that is soo....no. lol 17:58
porting the recent one on perlmonks...well the port is mostly done, now working on removing the need for HTML::Template since it isn't very fast when loading perl5:HTML::Templat;
stevan autrijus: ping 18:02
liz6 I think autrijus is in $work mode at the moment... 18:03
stevan liz6: ahh, ok,.. if he pops back into reality and feels like talking objectspace,.. I have some questions/thoughts to discuss with him 18:04
liz6 will pass it on... 18:05
stevan liz6: muchas gracias
nothingm1ch wishes he was in a condition to discuss threads
Juerd rehi
nothingm1ch i was busy with work
now I'm sick
18:06 nothingm1ch is now known as nothingmuch
Juerd Being in a condition usually means being compared 18:06
nothingmuch compared?
Juerd Yes
nothingmuch to what?
Juerd Most conditions are comparisons
liz6 nothingmuch: maybe you need some async { get_better() } ?
Juerd liz6: Then just the other thread gets better ;)
nothingmuch liz6: well, my skin around the eyes are swollen
liz6 $health : shared ? 18:07
nothingmuch so I'll need a blocking sem-put on that resource
Juerd Shared health. Hmmm... Interesting idea :)
nothingmuch till I can see well enough to read
liz6 nothingmuch: best wishes!
nothingmuch grazie
Juerd nothingmuch: Be careful to avoid deadlocks
nothingmuch goes to have a salt bath 18:08
Juerd nothingmuch: Have fun, and get well soon
nothingmuch thank you
stevan thinks justatheory is having a bad "net" day 18:10
Juerd justatheory: Please, fix whatever is wrong and stay. 18:11
eric256 is not sure that replaceing HTML::template bought him much speed.... 18:12
Juerd autrijus: You mention in your journal that you need more pause time for jokes and to digress less, but I think the quick jokes and digressions added to the overall (positive) feel of the presentation 18:14
justatheory Juerd: I would, but what's wrong is that a clipboard fell and cracked my PowerBook monitor. :-(
Juerd autrijus: I don't know if others would agree with me on this, though
justatheory: You say that as if it's unfixable
eric256: Have you profiled the thing yet?
justatheory Juerd: I say it as if it's a tragedy. And it is.
But I've been trying to get the laptop to use an external monitor, and kept going to sleep and waking up. Sorry 'bout that. 18:15
Khisanth hmm
eric256 i've never profiled any p6 code ;) 18:17
liz6 fetches autrijus for some dinner and will be back later
Juerd Whooo! 18:18
FINALLY
after all these years
a sticky drink in my keyboard
I thought it'd never happen
eric256 feather.perl6.nl:8027/cgi-bin/wiki.pl?topic=TesT 18:19
working pugs wiki....could be prettier. but it works! ;)
Juerd cdltbn56.epf r9ntc8gh b,4;ei jxr9lcntwh bmwcghtbwm jk23',a o;qj jk
y5f6piu bmnthbm w mx
sorry about that
It uses wiki syntax? That's unfortunate 18:20
PerlJam Juerd: unplug it and hose it down with de-ionized water and let it dry completely before using it again. This will minimize future stuck keys from latent sugar. 18:21
Juerd PerlJam: I think I'll just buy a new one. I should be using more ergonomical keyboards than this one anyway
I have no reason to believe any sticky substance has reached the membrane too 18:22
geoffb Goldtouch++, BTW
(If you're about to buy one)
Juerd And I have 40+ casing-and-keys-compatible keyboards in stock, so I can replace parts :) 18:23
goldtouch? /me googleth
geoffb www.goldtouch.com
eric256 Juerd. no. it doesn't use Wiki anything. lol. just linking between preformated pages. andyour the bastard you requested a wiki way back so be nice
r0nny yo
Juerd geoffb: Ah, I see. No, not what I had in mind.
eric256: Oh, I am nice! 18:24
geoffb: I like the kinesis-ergo "countered" keyboards.
geoffb I thought it was a bit odd at first, but man are my wrists happy to not be stuck with whatever stupid angle some KB designer thought was "average" 18:25
r0nny portages svn feature just rocks :) now i allways have a propper install from svn :)
eric256 what was taht kwid formating thing before?
geoffb (MS Natural KB)-- # Assuming all people are the same
Juerd I can't load goldtouch.com
Jooon eric256: more intutive pod format, and parsed differently
Juerd geoffb: It moves the strain to your shoulders if you're skinny/narrow 18:26
geoffb Juerd, hmmm, odd, it's printed right on the KB in friendly letters. :-) Lemme check
eric256 Jooon...i would need enoug info to implement it! ;)
Jooon eric256: wasn't my spec ok? :) ingy knows
r0nny darn - still erros on Prelude.pm
geoffb Juerd, you mean the MSNKB?
Juerd geoffb: With a kinesis keyboard, my wrists can be straight, and my shoulders natural (upper arms parallel with body)
eric256 google gives that address. but it never loads
Juerd geoffb: o 18:27
Hey
My key is't workig aymore
geoffb heh
r0nny does soneone know, where it installs prelude ?
Juerd Ah, I do't eed it ayway
r0nny i need to comment the pir line out for a test
Juerd You still kow what I'm sayig, right?
geoffb If you meant the goldtouch, someone gave me a great piece of advice once -- put the KB in your lap, almost shoved against your belt.
Juerd It's ot as if the is used really much, or aythig.
geoffb: I've used alike keyboards 18:28
geoffb (MS Natural KB--) # Again, this time for the "hand rest" that gets in the way of the lap trick
eric256 www.fentek-ind.com/ergo.htm lots of funky keyboards
geoffb Hmmm, goldtouch.com is timing out for me too. I wonder if their having DC problems
"they're"
theorbtwo Timeout for me as well.
Juerd I should be able to discover ways to type a lot without this specific letter... 18:29
eric256 Juerd just don't usa any words with it in them. ;)
geoffb What was that novel that someone wrote without ever using "e"?
r0nny anyone ??
eric256 thats gotta be horrible to read.
Juerd Without Es? The whole book? 18:30
geoffb And decently springy keys are nice, too
Juerd, yep. It was to prove it could be done. Which in English is REALLY hard
Juerd geoffb: That's right :)
eric256 r0nny. your questions are confusing. "where it installs prelude?" do you mean where prelude is pre compiled into pugs? there is an ENV flag you can set for that i beleive. but what does that have to do with pir? 18:31
Juerd geoffb: It's hard to avoid this key too
geoffb When I took typing classes in high school, we had some manual typewriters, some IBM Selectrics and some Apple IIs. Everyone wanted the Selectrics for speed exams
Juerd geoffb: It takes much more time to type a lie
eh
li.e
haha
r0nny eric256: i just want to comment out the eval option for pir, wich casues problems in prelude.pm for some reason 18:32
Juerd Ah, well, it also takes more time to type a lie...
r0nny eric256: it allways tells, there is a syntax error at this line
geoffb heh
Juerd I were to say line, though (hurrah for copy/paste)
geoffb :-) 18:33
eric256 so comment it out r0nny then make
Juerd altermatively, I could just start typimg M's imstead
geoffb Thankfully, English is an incredibly redundant language. And yet, somehow shorter than all the Romance languages, which says something.
Juerd geoffb: heh
eric256 www.spinelessbooks.com/gadsby/ 18:34
r0nny eric256: is there a way, to use another file as prelude on runtime ?
Juerd Okay, with IRC, the lack of this key is easy to work aroud, but CLIs are less flexible...
r0nny since i use portage to install it from svn 18:35
Juerd Oh, hit. 18:36
Moe key ae ow failig
eric256 LOL
r0nny...why do you make your life so difficult? ;)
and what did you do to prelude.pm to get that error in the first place?
theorbtwo Juerd, if a-f are OK, and you use gnome, hold down crtl-shift, and type the hex uncode char code of the missing key.
r0nny eric256: my %foo=evalfile("foo.yaml",:lang<yaml>); 18:37
Juerd adbcdefghijklmropqrtuvwxy
r0nny eric256: keeps to fail all the time 18:38
Juerd whee do2er that exta d c2ome f2om?;2~
ard the 2r?;2~
uh oh
it'r quickly gettirg much wore
r0nny lol
theorbtwo Juerd: I think it's time to unplug and swap out.
Juerd theorbtwo: what'r the fur ir that?;2~
thir i geat ertetairmairt 18:39
iff you car rtill follow what I'm rayirg
typirg
geoffb easy
eric256 r0nny, why would that command cause an error on a different line in prelude.pm? that makes no sense. you are chasing the wrong problem
Juerd G2o2od, ther I have r oear2o to rr~~~~~~~~~~~~~~~~~~~rrrrrrrrrrrrrrrrrrrrrrrrrrrrrr
eric256 perhaps your yaml is badly formated? or that option just plain doesn't work. either way a new prelude.pm won't help you
Juerd hr~it~
r d7ba~u5krpac~e~~
Khisanth Juerd: it just looks like you are drunk :)
Juerd hcchien: rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr 18:40
furc~
dramitr
r0nny eric256: it allways failed in the pir line
Juerd ~~~~~~~~~~~~~~~a~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
r0nny eric256: so i want to try disabling
Juerd h~ehe~, ~thri~ ~~~~~~~~~~~~~~~~~~~~i~ ~~t r~gd~rrrrrr.
rmedyr~ ~he~p m~e~ ~he~e!
geoffb *Now* It's gone beyond my ability to keep up
oh, but I got that last line!
Juerd geoffb: Ca~' tr~~ay~ r~~it~r'r ea~y r~f ~me r~.t ,r~wri.t~h2o.t~ ~ar~y t~a0~rpa~~ ~ ~ay~ ~.t ~rp ~.h ~.igd~r
lllllllllllllllllllll 18:41
geoffb hmmm, just below comprehension, that
Juerd rrlyla
~y ~lhi
eric256 slaps Juerd and dumps a bucket of water on his head 18:42
r0nny. on a script with no evalfile it still fails on that line? 18:43
Juerd reached the end of this thing's life
geoffb He's back!
Juerd copy~paste sucks 18:44
eric256 www.themodernword.com/scriptorium/perec.html <-- author of a book with no e's ... apparently there have been 2. must not be that hard 18:45
Limbic_Region besides the potential infinite loop, did Abigail have anything to say with the talks autrijus presented? 18:49
r0nny eric256: same with slurp 18:54
eric256 what error do you get when slurping? 18:56
r0nny syntax error: line 2, column 2 18:57
at Prelude.pm line 66, column 30-64
wich is the text Pugs::Internals::eval_parrot($code) 18:59
eric256 i believe thats a syntax error IN the file you are evaling
hince the first line/colm...the second is begging to trace up the call tree
r0nny well - if i dont eval, there is no error 19:00
eric256 or call list or whatever we call that. lunch later.
right. because you arn't evaling the code, the code you are evaling has the error, or it thinks it does
the yaml implementation is limited
errors reported in Prelude.pm are off by a couple of lines i beleive, something about the compile process
later 19:01
r0nny darn 19:05
allways it dies on eval slurp "foo.yaml",:lang<yaml> 19:06
oh -and pugscc tells it cant find the module Pugs.Internals if i do pugscc 19:07
cool 19:10
foind the sweetness of the do command from perl5
Juerd re 19:27
Got another keyboard from my office :)
It's newer, and it actually feels different too. But it's the same keyboard. 19:28
Same model, bought at the same time, just not used until now.
Limbic_Region: Not that I can remember
Limbic_Region: Why this rather specific question?
apple-gunkies excuse me folks, is there a synopsis or other document that describes the current thinking on perl6 macros? 19:43
S6 is pretty short on details :) 19:44
r0nny apple-gunkies: u mean the howto for html comments ? ;P 19:50
19:52 _SamB_ is now known as SamB
kolibrie apple-gunkies: I don't know of anything other than S06, and perhaps p6l 19:52
r0nny anyone got an odea why Prelude.pm might break at the pir eval line ? 20:01
apple-gunkies thanks guys, guess I'll start digging through p6l. 20:02
kolibrie apple-gunkies: report your findings, if you find anything 20:03
obra gugod! hi! 20:12
pmurias hi 20:16
what sdl bindings are availble for pugs? 20:17
i'am currently aware of chromatic Parrot::SDL
u 20:19
wanted to /u :) 20:20
eric256 wonders if there will be a way to make a code block atomic. 20:23
atomic { my $fh = open ($file); $fh.print($data); $fh.close; } ... could be very nice ;) 20:24
theorbtwo ?eval my $x=1; atomic { $x++; die }; $x; 20:25
evalbot_7815 Error: No compatible subroutine found: "&atomic"
theorbtwo Oh, could have sworn it was already there.
PerlJam theorbtwo: me too.
eric256 maybe its not safe? 20:26
PerlJam ?eval my $x=1; atomically { $x++; die }; $x;
evalbot_7815 Error: No compatible subroutine found: "&atomically"
PerlJam I think we just have the wrong name.
(or maybe it's just not there) 20:27
But it ought to be. We've talked about it enough.
eric256 can;t find it in the documents. not that that means alot
eric256 peaks at the KWID module. and wonders how hard it would be to port 20:33
pmurias mention of atomic can be found in one of the talks 20:36
geoffb pmurias, 'use perl5:SDL;' works as well.
oops . . . that's what I get for typing around a child
eric256 lol
Kwid::Event::Parser exists...hmmm
stevan eric256: that is missing a LOT of functionality 20:48
and event based parsing it probably not the right approach for Kwid 20:49
eric256 i have no idea. i don't even know the Kwid syntax... i thought i'd heard rumbling that tests would be documented with it 20:52
feel free to play on feather.perl6.nl:8027/cgi-bin/wiki.pl its pretty low tech for the moment. looking for a quick Wiki-ish formating system for it 20:53
stevan eric256: in the early days of Pugs,..there was talk of kwid replacing POD
but I think that aspect has died in favor of the PerlDoc stuff
eric256 PerlDoc? hehe
whats that? 20:54
stevan search.cpan.org for it,.. It was a DOM-like object structure for parsing POD, Kwid, and any other misc format
eric256: very nice (wiki.pl) ... and it's almost as fast as kwiki :P 20:55
eric256 thanks. it doesn't do much and it stores everyting with $fh.print($data.perl); but its a start. 20:56
i'm thinking about haveing it write the HTML files, and just re-write the HTML on edit. then viewing is fast and editing is the same speed 20:57
pmurias anybody interested can see the beginnings of my perl6 tutorial at www.yourcargo.org/draft.txt 21:04
svnbot6 r7816 | autrijus++ | * More S17 work: a vastly more improved alarm(), among other 21:05
r7816 | autrijus++ | random changes and topics.
autrijus that doesn't work -- www.yourcargo.org:8080/draft.txt does
pmurias as you can it's in it's very early stages yet (94 lines)
eric256 pmurias ... i'm sure you could put something like that in the repo under docs
stevan hey autrijus :)
autrijus hey stevan.
sorry, was fried last night
stevan autrijus: no problem 21:06
autrijus and hence didn't do much but finished the pressure centre that is $job today
and got alarm() among other thing specced
eric256: DBI::PurePerl is only 400 lines in its main logic -- interested? :)
(and also DBD::mysql::PurePerl) 21:07
eric256 interested in....porting? no thanks. ;) not yet at least. 21:08
stevan DBD::MySQLPP is very nicely written code IMHO
eric256 is there a way i can checkout a single directory? i want to push this wiki into the examples section. is that doable? and then check the wiki back out into my cgi-bin directory? or is that a n no? ;) 21:10
autrijus with svk? 21:12
sure, do this
svk import --to-checkout //mirror/pugs/examples/yourwiki /path/to/yourwiki
pmurias stevan: Also wonderfully modularized...
autrijus dduncan: have you looked at DBI::PurePerl and DBD::mysql::PurePerl? 21:15
dduncan: I'm wondering if it makes sense to port them over -- liz6 noticed that in all p6 examples, we start using
use perl5:DBI;
which should be a pretty good indication that it's a worthy module to have under 6.28.x
xerox autrijus! 21:16
autrijus hey xerox
eric256 does it matter where i run that command autrijus?
dduncan autrijus, do you mean the perl 5 ones, or is there a perl 6 version now?
autrijus eric256: nope
xerox You did show (an early-staged) Nymphaea in the Haskell.xul slides! Woot! you++ !
autrijus dduncan: no, I mean porting the 400 or so lines of p5 into p6 21:17
xerox: eh what? what?
xerox autrijus: there is that slide 'cairo + Gtk2hs' and then *poof*, the L-System renderer. I wanted to thank you :-)
autrijus oh right :) 21:18
dduncan if you want me to do this, I can consider it ... but the one I do may be more of a look-alike that doesn't follow all the little details
autrijus dduncan: sure, lookalikes are fine
I'm just looking for things that can exercise OO stuff that also have max practical usability 21:19
xerox (Though you could have chosen a best shot ;-))
autrijus though a thought would have perl6 DBI working with perl5 DBDs
mmm scary thought
liz6 autrijus: yep, 2 scary..
dduncan this said, I feel I should put forth an invitation to the people that are designing DBI-2, should they want to participate 21:20
PerlJam dduncan: people? I thought it was just Tim these days.
svnbot6 r7817 | eric256++ | Directory for svk import.
r7818 | eric256++ | Adding Wiki port of www.perlmonks.org/?node_id=503926
autrijus dduncan: yeah. the last time it came up, parrot was nowhere, our obj system doesn't really exist, so there was not much work done 21:21
Tim said he'd take a look but didn't report back
dduncan true, the current effort is parrot focused, for a multi-language solution, but they may be interested in this
stevan Tim recently asked for a p5 roles impl for DBI2
robkinyon was building it
dduncan on the other hand, I could just make a clone of DBI1 in perl 6, with the assumption it will be rewritten after 21:22
I'll do something, anyway
PerlJam dduncan: you can't. What will happen is that you're start on the clone of DBI1 and then realize that you could make it better here and there, and you'll end up with DBI1.5 :) 21:23
dduncan but it may take liberties in improvement, such as the perl 6 Test.pm isn't exactly the same as the perl 5 Test::More
autrijus peripheral stuff are fine :)
s/stuff/changes/
dduncan sure, we'll make it DBI 1.5
autrijus cool! dduncan++
PerlJam dduncan++ indeed 21:24
pmurias it would be cool to be able to use DBI transparenty from browsers
Khisanth PerlJam: meh there already modules with same name as their p5 equivalent except they have a different interface!
dduncan I'm going to have to install MySQL on my system again, though, in order to test it with
Khisanth PerlJam: don't you mean he will end up with DBI6 ;)
PerlJam heh
autrijus you can then port DBD::NullP and DBD::Sponge :)
(<100 lines) 21:25
and then stevan can port his wonderfully sick but practical module
dduncan which is ...
stevan autrijus: which one?
autrijus (which is considerably more lines)
DBD::Mock
stevan :)
PerlJam a parrot DBI would be nice though. Let all languages reap the benefits. 21:26
autrijus PerlJam: or a P6 DBI with a reasonable P6->Parrot codegen
PerlJam or that.
autrijus and optimize the heck out of that (codegen)
in any case I suspect DBI::PurePerl to be easier to port than the full DBI XS to Parrot 21:27
dduncan damn straight!
PerlJam maybe.
parrot makes many things simpler.
autrijus it's not about parrot :)
(it's about grokking the XS part)
DBI::PP to PIR may be interesting as well. 21:28
timtowtdi
dduncan all we really need though is a new code base that follows the same behaviour as the old version ... doesn't have to be a line by line or sub by sub port
autrijus dduncan: right
PerlJam Get the interfaces right and who cares what the code looks like :-)
dduncan and that makes it easier, since the DBI API is dead simple
autrijus it's basically just a way to query stuff from DBDs
and very few abstractions between them 21:29
PerlJam btw, if I'm not mistaken, DBI<->DBD does the same trick of having intermediate "languages" that meet in the middle. This seems to be a good pattern. 21:34
autrijus well, it's a set of calls 21:36
but it's reasonably minimal and well specified, so yeah 21:37
integral surely that pattern's trivial since it occurs in any extensible layered situation...
stevan DBI->DBD also works on the idea of "reasonable defaults"
writing a basic DBD is actually fairly simple,.. it's when you get into the fine grained details that it gets harder 21:38
autrijus integral: though to be nicely extensible, the pattern can be subtle to find 21:39
stevan autrijus: so any object space revelations you want to share?
autrijus stevan: thanks to you I was able to comprehend diagrams on PickAxe about metamodel and eigenclasses without squinting 21:40
stevan autrijus: :) glad I could help
Pickaxe++ # i picked it up again recently
autrijus stevan: er no, except they have this weird ability to inspect all managed objects in existence
and I started thinking of doing that in hs
and then it occured to me that 21:41
stevan we need to spec the GC?
autrijus all the scalar array and hash contains would be counted as well
as they are objects in p6 but not in ruby
(I journaled about that)
stevan hmm, thats a good point
I am starting to think maybe we need to seperate our core objects from our user obejcts 21:42
autrijus so have to layers of runtime objects?
s/to/two/
stevan maybe
autrijus this just means there are things that don't get into MM/ObjSpace
right?
so you can still call methods on them but not extending them as such?
stevan what do you mean by "things"
autrijus things, as in objects in the runtime 21:43
stevan well I think we can make them accessible for extension and such, but we just don't count them as user objects
autrijus oh ok
so just a flag
stevan I mean,.. the GC should never need to reap Class, Object, Module, Package, etc
they should never go out of scope anyway 21:44
same with Scalar, Hash, Array, etc
they are key to the language itself
autrijus but specific Class instances are reaped
as is Hash containers
stevan yes, they should be
but the original ones,.. no,.. they should always stay around
autrijus it's just the ::Hash object itself that is not a "user obj"
but consider 21:45
::Hash := ::MyHash; # whee!
suddenly the old hash stuff is outta scope
dduncan so I'll checkin a shell of this DBI in an hour or so ... but a name ...
stevan why is it out of scope?
dduncan I thought of naming it "DBI1"
autrijus it's not reachable
unless you have hash containers that's alive
dduncan and any drivers, "DBD1::*"
stevan ah,.. I see
theorbtwo autrijus: Don't the extant hashes... exactly. 21:46
stevan well maybe the whole "classes are but an illusion" idea of luquis works here
autrijus elaborate?
dduncan the 1 is intentional so that any uses of this module that become legacy don't interfere with the more solid "DBI" version 2 being worked on for the future
stevan well, if ::Hash is just a Role in the user space, and it is created in the outermost scope, then it should live long enough 21:47
autrijus dduncan: I think in light of multiversioning it's a bit defensive, but I'm fine if you do that :)
stevan the underlying impl will be in accessible ... luqui's "concrete type"
dduncan alternately, what "version" should I use?
autrijus stevan: ok, I understand and agree.
dduncan 1.5.0 perhaps?
theorbtwo OTOH, it'd be nifty if ::Hash = ::MyHash magically made {} and my %foo create MyHashes.
autrijus dduncan: sure
theorbtwo: that is the idea. 21:48
stevan wonders if the crack in England is that much better than the crack in Germany, and why theorbtwo didnt bring any to the hackathon
autrijus doesn't see anything wrong with that idea, at least not yet
theorbtwo stevan: Huh? 21:49
stevan it is an interesting idea,.. but it's like overloading CORE::* stuff,.. you never know who is doing weird stuff, and what that will break
theorbtwo Yes, well, I wouldn't suggest actually doing so.
stevan theorbtwo: just joking,.. it is semi-insane
theorbtwo Esp assigning to it without temporalizing.
dduncan on the other hand, I can just call it "DBI" and "0.0.1", and this can be evolved into the actual version by Tim etc later 21:50
autrijus dduncan: yes.
PerlJam dduncan: DDBI (Duncan's DBI :)
stevan PerlJam: that would be DDDBI (Darren Duncan's DBI)
dduncan I'm not doing that ... that's what Rosetta is
liz6 isn't Perl 6 able to load modules by author?
stevan DBI-0.0.1-cpan:DDUNCAN should be fine IMO 21:51
autrijus liz6: yes.
liz6 stevan: right
PerlJam dduncan: what stevan said
liz6 so no need to use something other than DBI as namespace, right?
autrijus use DBI--cpan:DDUNCAN;
liz6 cool
autrijus liz6: right. that's what I meant by "in light of multiversioning" 21:52
stevan praises the 3 part namespace,.. all hail the @Larry
autrijus prays we get that in p5
dduncan should I include the cpan:DDUNCAN right from the start?
autrijus dduncan: sure why not
liz6 autrijus: I was interpreting mult-versioning more strictly: different versions of the same module / author to co-exist
stevan dduncan: you could also use email:dduncan@something.net or url:www.dduncan.net 21:53
autrijus liz6: nod
dduncan okay ... should help test the system, since I don't think anyone else is doing that
autrijus s/email:/mailto:/
and http:// is fine without url: :)
(at least afict)
dduncan if I have to associate it with myself, then my CPAN id is fine
theorbtwo ++ to whoever realized that it's cpan:, and not cpan://. (Perlmonks got that wrong.)
stevan yeah,.. url:http: is a little redundant
autrijus more importantly, it's spceed to take a URL :) 21:54
s/URL/URI/
and a URI scheme named url would be strange :)
theorbtwo There is a URI scheme named urn, BTW. 21:55
...and yes, it is strange.
stevan autrijus: have you had any time to look at the Bootstrap file I have been hacking on? 21:57
autrijus MM-Bootstrap?
stevan MetaModel::Boostrap actually 21:58
autrijus the procedure to bootstrap from any capable chaos?
stevan yes that is the idea (although not reality yet)
obra 'evening
stevan hey obra
autrijus to get there you need the native p5 closures to go away 21:59
decomposing them into yet more prim calls
stevan autrijus: what native p5 closures? the sub {} stuff?
the closure->new() objects are implemented in the runtime,.. they just need some type of executable chuck to work with
s/chuck/chunk/ 22:00
autrijus I mean the stuff like 22:01
my @nonemptyseqs = (map { (@{$_} ? $_ : ()) } @seqs);
they are trivial 22:02
stevan oh yeah,.. _merge needs help,..
eric256 can i ask...why aren't you doing all this work in p6/haskell? i think i've asked before..but i'm still curious. i mean i figure its faster to prototype in p5. but arn't you now in the proccess of integrating it into pugs?
autrijus but that's also portability problems may rise
since it's not all messages
but the idea is sound and... you know, this is much much more readable than any of the previous boots
stevan++ 22:03
stevan autrijus: I am thinking some kind of data driven approach might make sense
where we spec the metamodel out as XML or something ;) and each backend builds a minimal translator of some kind
autrijus maybe. but what you are doing can already be subject to static analysis 22:04
stevan true
autrijus because the call graph (treating assignments as bindings and inlineable) represents the XML structure
so yes, I think we will do that (eventually)
with the canonical sequence in p6
stevan I might take the PyPy approach too of using list iterators rather than foreach loops
autrijus and do PIL->BootCode->P5 22:05
that's trivial to write
stevan that makes sense
autrijus iterators makes much more sense, yes.
(I was looking at the foreachs and wondering what to do)
stevan PyPy++ # :) 22:06
autrijus indeed
stevan ideally I can break it all down into message sends, and what little can't be done that way will need to be implemented by the runtime
autrijus though I suspect the "what little" there can also become kernel msgs 22:07
stevan mumbles something about doing it all in pure lambda calculus
autrijus and the runtime just supplies the prim
stevan autrijus: quite true
autrijus stevan: that's what PIL is, you know :)
at least, readily translatable to
I like the idea of kernel msgs in this setting
stevan yeah me too 22:08
I think _merge might be a Kernel message actually
it would get ugly as message sends I think
autrijus you can decompose it into a few kernmsgs 22:09
theorbtwo thinks this project is starting to have lots and lots of different languages and turtles-all-the-way-down syndrome.
autrijus one handling the emptyseq
and one or two in each loop
theorbtwo ...but I also think that I might think this because I don't understand what's going on.
autrijus theorbtwo: that's mostly because insufficient documentation :) 22:10
(and visio diagrams)
I'm looking to fix that as well.
theorbtwo I saw that there was lots of progress on a PA02 on your journal, but have not yet actually read it.
autrijus theorbtwo: what stevan is doing here is to write the metamodel bootstrapping sequence in a language agnostic way
theorbtwo Yes.
stevan s/stevan is doing/stevan is trying to do/ 22:11
theorbtwo That seems like overkill; don't we only have one language to bootstrap?
autrijus ok, stevan is trying to attempt to do
stevan loves to throw out code,.. so you never know ;)
autrijus theorbtwo: er, in a runtime-language-agnostic way
of course they all boot the same p6 model
theorbtwo I'd really be much happier, all in all, to be able to run perl6 on one backend then almost be able to run it on 7. 22:12
stevan theorbtwo: which one?
autrijus which one, is the problem
I'd like p5
if I can only have one
stevan me too
easier to transition
theorbtwo I suspect rolling our own might be the most reasonable. 22:13
autrijus but then p5 is not really rigorous, and unless we do some abstractions like this, the lessons learned there can't be portable to hs, pir, js, or whatnot.
theorbtwo No impedence mismatch.
autrijus mumbles something about parrot and p5 is also supposed to be "our own"
eric256 p1-5 didn't have these issues. or did they? or are we planning for the next 100 years?
autrijus eric256: p4 to p5 went thru the same pain 22:14
theorbtwo p1-5 was our own, I think.
autrijus eric256: also, because in p5 there is only a runtime, not a language
stevan invokes the -Ofun switch :)
autrijus eric256: so when the runtime becomes hard to change, the language suffers
(vice versa)
theorbtwo The problem is that the interpreter kept getting more and more hairy.
autrijus stevan: you collect 1000 zorkmids :)
theorbtwo (And I'm saying the same thing as autrijus, but from a different point of view.) 22:15
Is the subset of perl6 that we can run now different then when we introduced the idea of multiple backends?
stevan theorbtwo: interpreters are hairy beasts,.. this idea (language-agnostic bootstrap) should actually help (I hope)
theorbtwo So do I.
eric256 okay...so this is the way to keep perl6 from running into a wall ilke perl5 where changing the language becomes difficult of even impossible.
autrijus theorbtwo: more things has been supported, some only found on js 22:16
eric256: yes, basically
stevan eric256: p6 will never stop changing while @Larry is still around :P
eric256 scratchs his head agian and begins to anxiously continue to hurry up and wait for real OO ;)
theorbtwo But I'm worried that it'll take 6 months to write this generator, vs 6 days to write the code that it would generate.
autrijus the PIL and MM and other things we invent are partially defensive measures against an ever changing spec. 22:17
extensibility, modularity, and other good things may result from it, but fun (and survival) are the initial motivations -- at least for me :D
eric256 theorbtwo++...it does feel like you could just implement something in pugs and perfect it there. instead of all this back and forth..but i understand thats how it looks from my seat, when reality might be entirly different. ;) 22:18
stevan theorbtwo: the generators should be fairly easy actually, and they guard against runtime specific bugs
theorbtwo I saw a great quote in somebody's /topic the other day -- why spend 5 days of your life writing something that you can spend five years of your life writing a thing to automate writing it.
r0nny autrijus: ping?
eric256 lol /me notices that r0nny wasn't paying much attention. lol
theorbtwo Or something like that, but more elegant, because I am quoting from memory.
autrijus theorbtwo: as do I... but then, I have brainwashed myself enough that I do believe that p6-running-on-p5 will be the ticket into "hack hack hack" phase 22:19
r0nny: sure
r0nny eric256: afaik i was afk eating
autrijus theorbtwo: I may be wrong, but that's the assumption I'm operating under :)
r0nny is there a new svn version where it wont tell a syntax error, on the call to the pugs internal pir evaluator ? 22:20
stevan theorbtwo: read about PyPy,.. it is a very interesting approach
autrijus (happily, my assumptions are lexical and only affects my block)
r0nny: eval('print 1', :lang<pir>);
like this?
theorbtwo Donno. One thing that worries me is that if you can mix p5 and p6 too easily, there's no motivitation to port your code to p6. Instead you write in a mix of p5 and p6, and when that's too slow, compile your p6 down to p5 and start hacking on the generated p5.
autrijus r0nny: it worksforme 22:21
stevan theorbtwo: it wont be compiled down to pure p5,.. it will still require the runtime
r0nny autrijus: it fails on stuff like eval slurp "file.yaml";:lang<yaml> - tells the error is the call to the pir evaluator
autrijus theorbtwo: I guess the motivation would be that p6 is easier to write than p5.
r0nny: eval(slurp("file.yaml"), :lang<yaml>);
theorbtwo But if p5 is the primary backend of p6, then p6 will never be faster then p5.
autrijus r0nny: I think your semicolon are in the wrong place 22:22
r0nny oh
i just typed wong here in irc
well - it fails, if i use evalfile
stevan has to run,.. I will backlog on this later, adios &
theorbtwo r0nny: You should copy-and-paste more.
autrijus theorbtwo: however, I don't see hs and js faster than p5 really 22:23
theorbtwo stevan: Right -- it won't be compiled down to pure p5, it'll be compiled down to p5-with-baggage-on-top.
autrijus theorbtwo: that is not entirely true -- sky__ pointed out that we can compile to OPs and introduce new OPs 22:24
r0nny ok - on eval slurp "config.yaml" , :lang<yaml>;
stevan mumbles something about "we should all just use Java" and runs out the door
autrijus first as loadable XS, but maybe even into p5blead, no less, just to support pugs
r0nny it tells syntax error: line 2, column 2
at Prelude.pm line 66, column 30-64
autrijus theorbtwo: and that would be as fast as we'd like.
r0nny: are you using svn head? 22:25
r0nny and line 66 col 30-64 is the call to the damn pir evaluator
theorbtwo Wasn't part of the point to get rid of the icky perl 5 runtime?
r0nny yeah
Version: 6.2.10 (r7815)
autrijus theorbtwo: er yes, and people are working on that as well :) 22:26
it's just takes some time to unscrew a previously screwed project... but I think we're doing well :) 22:27
theorbtwo Hm.
eric256 perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
eric256 r0nny can you paste your yaml file there
autrijus theorbtwo: and also, if parrot dissipates and YARV becomes real nice, we don't even have to roll our own.
theorbtwo YARV?
autrijus theorbtwo: Ruby2 VM
theorbtwo Ah.
autrijus or rather, Ruby 1.8's new VM as well
already runs miniruby
theorbtwo It seems unlikely that parrot will unscrew; trying to be everything to everybody is /hard/. 22:28
autrijus theorbtwo: try www.atdot.net/yc/compile
r0nny r0nnyger.mine.nu/~r0nny/config.yaml
its the output of using perl5 data and a yaml output 22:29
theorbtwo Hm. I'm off, I think. 22:30
autrijus theorbtwo: yeah, but the gist is that I'm not really a VM tuning person, so I wouldn't want to write and maintain a fast, efficient, general-purpose runtime that's competitive to other shared runtime out there
r0nny eric256: any idea from the yaml file ?
autrijus theorbtwo: but I want to use p6, so my strategy is to target it to whatever VM that's best out there. :)
other people's motivations may, of course, differ
eric256 yes. the bult in YAML processor is very limited, as i stated before 22:31
autrijus r0nny: that worksforme
./pugs -e 'say eval(slurp("config.yaml"), :lang<yaml>)'
or even
./pugs -e 'say eval(slurp("config.yaml"), :lang<yaml>).perl'
r7598
er I mean r7816 22:32
eric256 but that file processes fine
r0nny autrijus: same errkr
autrijus r0nny: very strange -- did you change Prelude.pm somehow?
r0nny no
eric256 same here clear back on r7711
autrijus r0nny: your platform? 22:33
r0nny: I wonder if a clean rebuild would help
r0nny gentoo linux - x86
clean ?
autrijus make clean
perl Makefile.PL ; make unoptimised
if that fails, I'm willing to take a shell account and go look, but that's after I wake up
eric256 you might revert your SVN checkout too
autrijus right, or use "svn diff" to check 22:34
r0nny might it be a problem to have a binary version of ghc ? 22:36
autrijus no, that shouldn't matter at all.
r0nny i wish i knew, why it dies all the time 22:37
buu Are the l33t new perl rules implemented in pugs?
autrijus buu: yes 22:38
buu Hurrah!
Woah. They look exactly like BNF
autrijus r0nny: so if you run this
./pugs -e 'say eval(slurp("config.yaml"), :lang<yaml>).perl'
in the command line
it dies with the same error, right?
r0nny yeah
autrijus what if you change "yaml" to "perl6" there? 22:39
r0nny undef as ret val
no error
autrijus what if this? 22:40
./pugs -e 'say Pugs::Internals::eval_yaml(slurp("config.yaml")).perl'
r0nny syntax error: line 2, column 2
at -e line 1, column 5-58
it doesnt like pugs::internals
hmm
autrijus that's very wrong 22:41
mm
r0nny darn - pugs::parser kales a mess while being compiled 22:42
eh makes
autrijus kales?
mm? meaning it takes a lot of ram?
r0nny it makes the x11 framerate about 1 fps 22:43
autrijus oh wow. maybe renice it
your kernel isn't very preempt, I see :)
r0nny wich basically means it takes a hell of shitload of work for the vm
allready niced down to 19
autrijus I'm sorry. :/ did you really "make unoptimised"?
(straight "make" for GHC 6.4.0 will leak) 22:44
r0nny only compiling monotone had whorse results on my pc
i think im going to make it need ghc 4.1-r1 22:48
the -A200 opt "should" have worked
dduncan I think I'll defer the DBI thing for a few days or weeks 22:55
fyi, in order to test this with MySQL etc, we need an IO::Socket replacement that is functional
liz6 yes, that would seem necessary... ;-) 22:56
dduncan does Haskell have the libraries etc to hook up to that implement sockets?
hmmm ... I see there's a Net::IRC module ... but does it work? 22:57
Khisanth aren't the svnbot6 and evalbot_7818 both written in Perl 6?
r0nny omfgdh 22:58
Khisanth surely that requires working sockets ...
dduncan are they?
looking ...
r0nny watching the configure process of ghc makes me think a insane person did this freaking autofoo setup
eric256 Khisanth the code for them is in the repository under exampes/net 23:00
but yes i beleive they both are, well eval bot has a perl5 helper script
autrijus dduncan: we have sockets 23:02
dduncan: examples/network/
liz6 is calling it a day 23:03
a day
dduncan okay, I see the examples ... this will take me awhile to do, though, and I'll probably continue my own modules concurrently 23:14
autrijus that's fine, no rush :) 23:21
rafl autrijus: FYI, our Module::ScanDeps, PAR::Dist and PAR modules will be in Debian soon.
autrijus cool 23:22
r0nny rafl: can the depenca scanner be used for other stuff, then modules ? 23:23
eh depency 23:24
rafl I'm also considering to extend dh_perl (a script that finds dependencies for perl packages during the bild) and dh-make-perl (a script that makes debian packages from perl module from scratch) using Module::ScanDep?
r0nny: what stuff for example?
autrijus r0nny: yes
rafl: that'd be nice.
(though I need to sleep in a couple mins) 23:25
r0nny source files > object files > library file(s)
im still on the build-system thing 23:26
maybe it can be integrated into sixpan for creating binary modules in a nice way
rafl I doubt that. Different distributions will have other names for different libraries. 23:27
Or maybe none at all (win32).
sixpan would need to provide "The Source Of The World" to make that work.
r0nny not really 23:28
autrijus waves &
r0nny cu autrijus
rafl: basically all i need to change betwen platforms is binary format and prefixes/suffixes 23:29
should be configured in a yaml file
btw - hadnt parrot a arch-independ object format ? 23:30
rafl How can I specify and check that I need libgmp3 in at lease version 4.1.4?
r0nny rafl: such stuff will be part of the configuration framework, not the building/depency part 23:31
but basically everything should be made by some yaml config files
it can easyly be the first file od a tarbz2 23:32
rafl I really doubt that this can be portable.
r0nny why ?
where do u see problems
i really need to know such stuff - maybe i forgot simple stuff wich makes some things bad 23:33
rafl Different platforms have different packaging systems. Some don't have such a thing at all.
If a dependency is available with the packaging system it should install it using this system
r0nny could do that with extensions 23:34
rafl If not it needs to be downloaded and built (including all deps) and installed where also other software can use it.
23:34 khisanth_ is now known as Khisanth
rafl How to write an extension that can deal with every package management system that exists out there? Impossible, imho. 23:35
It's almost impossible to really find the name of a perl package in Debian from the module name on cpan.
r0nny hmm 23:36
then dont support till they got sane names
maybe help them keep perl6 module names clean from the beginning
rafl Of course. Only support gentoo as it's the only sane alternative to <insert whatever here> 23:37
r0nny hmm
well under gentoo Foo::Bar is just mapped to dev-perl/Foo-Bar
rafl Perl is cool because it runs on almost everything and is highly portable. 23:38
r0nny well - there is allways the way of creating own deb files, and installing them with apt
eh dpkg
rafl Not supporting something because we want to have some magic for installing the whole world from cpan isn't good.
r0nny well - just make the debian folks keep the perl6 module names clean 23:39
rafl The goal of sixpan is to provide something to install perl packages. If there are other libs require it can tell you that, but it can't install it in the normal case.
I'm going to, of course. But there's more out there than debian and gentoo.
r0nny i got an dea for this 23:40
well - iach package with external depencyes could have a yaml file maps os/dist to package names need to be installed
also the package-maintainers of the distro could do some of this work 23:41
guess debian and gentoo wil have it the fastest, *bsd and other unix-aliked folowing 23:42
rafl I, as a potential author of perl modules, have exactly one platform I write my stuff one. Others may have two or ten or whatever. No one will have every platform you can think of. So it won't work for every platform.
r0nny but win32 might be problematic
rafl We can't rely on others. It simply needs to work as it is.
r0nny then it needs to relly on sane pkg managers 23:43
and i dont know any conplete sane out there
rafl OK. You convince developers of every operating system flavour out there to obey our rules first. Then we can implement your ideas. 23:44
r0nny well - there should be a concept allows package-maintainers to add the needed informations easy, so the package-developers dont have to worry about all platforms
make it nice and sane, so they start using it :) 23:45
at leas i hope so :)
rafl For sure..
r0nny well - i know it will be sane under gentoo (u know the :: to - mapping) 23:46
debian folks coudl do, too
rafl Stop thinking only about Gentoo and Debian.
r0nny just make it perl6-mod-Foo-Bar or something
bsd might be more problematic
since i dont know hopw ports work
but i know, who to ask 23:47
and win32 need something else
basically unix needs some simple stuff, an then could do
but win32 might be more complicated
rafl BSD isn't a real problem either. Other systems won't have a package management at all or they don't package the library a module needs and maybe not even the build-depends and probably the build-depends won't even run there, ... 23:48
r0nny then the modules just should tell them all depencyes at once (i hate autofoo configures show only one a time - makes it bad) 23:49
rafl Fine. Unfortunately that doesn't solve anything. 23:50
r0nny no package manager means user needs to install
think of LFS users and such
they have no pkg-manager at all
so they need to know what they need
ping? 23:52
rafl Yes. That the least common denominator what is useful to implement. Tell the user what non-cpan dependencies are needed and maybe give him some hints about them
r0nny yeah
btw wvk mirror ? 23:54
eh svk
rafl Nothing yet. Too busy with Debian stuff to take a look at perforce and how to mirror it in a sane way. 23:55
r0nny hmm
well - i stay away from debian - it broke multiple times on me - in a non-recoverable way 23:56
rafl Yes. That happens often and it's always Debians fault. 23:57
r0nny who knows
i know ppl never had any problems with it
rafl I usually reinstall it at least once a week. But that doesn't hurt because I don't need to compile everything from scratch. *scnr*
r0nny well - usually my system only breaks on hw failure 23:58
afaik only free-bsd's stuff s more powerfull on this 23:59
but free-bsd cant do xfs, and im a xfs fanbay ;P
eh fanboy

Powered by Cro and the Raku® Programming Language.

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

Thank you!