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.
wolverian summer time ends in one hour here. 00:00
don't generalise. :)
theorbtwo Oh, damn, I'm an idiot.
No, wolverian: it ends the same moment all over europe.
That time is in one hour.
wolverian ah, you mean european summer time.
yes, correct.
eric256 the parens are for pairs. its the .perl way of outputing it. 00:34
wolverian pairs are higher precedence than , I think 00:35
s,pairs,=>,
eric256 didn't say they were needed. ;)
a little lighter green now
on the tests pages. ;) 00:36
jrussell hiya .. i'm using the combine() function from Math::Combinatorics. i'm not getting the unique combination i'm supposed to get 00:45
eric256 jrussell...you know this is a room for perl6 i.e. pugs development right? you might want #perl 00:46
jrussell oh sorry i didn't 00:47
thank you
svnbot6 r7770 | eric256++ | Test-Catalog: Added colums to the file listing, added yellow shading for todo items. 00:59
theorbtwo Woo, it's 1 o'clock again (in England). 01:00
eric256 ?eval localtime();
evalbot_7770 \Time::Local.new(("day" => 30), ("hour" => 2), ("is_dst" => bool::false), ("min" => 0), ("month" => 10), ("picosec" => 811255000000), ("sec" => 25), ("tz" => 3600), ("tzname" => "CET"), ("wday" => 1), ("yday" => 302), ("year" => 2005));
eric256 is that 3600 seconds?
eric256 is amazed by the usefullness of running top in one terminal while working in another...and promises not to leave it running when not needed ; 01:22
svnbot6 r7771 | eric256++ | Converted all http links to use the L<" format 02:18
eric256 466 test files with 344 links...not too shabby. 02:27
anyone know when S29 will be commited? 02:57
svnbot6 r7772 | eric256++ | Created SAE index for tests.pugscode.org, some code speed up and clean up as well. 03:17
autrijus eric256: you mean to svn.perl.org? 03:22
I think it makes sense to ask @Larry on p6l about this 03:23
and if yes I can commit it into svn.perl.org
wolverian OT - MS's new Consolas font is the best coding font I've seen in a while. (at least if you like purty shapes and antialiasing. :) 03:28
autrijus ooh url? 03:29
oh, wait, that'd be asking for warez ;) 03:31
wolverian I d
on't actually know its license
autrijus can't easily find it online though 03:32
wolverian right 03:33
hmm. perl5 feels like walking on ice nowadays. slippery fun, but frustrating. 03:34
autrijus I thought it's been like that for ages :) 03:35
medieval castle and all that
Consolas screenshots look nice
wolverian but it's only now that I've realised it.
autrijus not sure I'm giving up Andale Mono (another MS font) though
wolverian Andale Mono is very nice too.
Consolas, Andale Mono and Bitstream Vera Sans.
autrijus and Anonymous for lower res
that's about it for me
I still remember the first day I got the then-cutting-edge antialias andale mono + mlterm set up 03:36
I feel my life has changed :)
even better, I imported Apple LiGothic using pfaedit (now known as fontforge) so it works on freetype 03:37
wolverian eh, I'm using 17% of my gmail space and I haven't even checked it in weeks
autrijus spams?
wolverian no, p5p apparently. :)
autrijus aha :)
autrijus backlogs p6l and wonders if some more sleep is warranted (4:43am) 03:44
wolverian well, that was the first time I've ever used Storable::dclone 04:08
eric256 autrijus i meant svn.perl.org 05:04
svnbot6 r7773 | eric256++ | Minor tests documentation clean up. 05:48
eric256 night all 05:49
gaal heya 05:56
eric256: just missed you - but nice work on the tests referencing!
Khisanth oO 06:14
Bundle::Pugs::SmokeKit requires 5.8.7? 06:18
gaal errr, it shouldn't 06:26
oughta be 5.006001. where'd you get that? 06:27
Khisanth CPAN :)
0.03
gaal i wonder what might want 5.8.7 there, maybe one of the deps... 06:28
Khisanth the use 5.008007; wants 5.8.7 06:29
gaal where's a use 5.008007?
Khisanth the Makefile.PL 06:30
gaal "use 5.006001"
Khisanth search.cpan.org/src/GAAL/Bundle-Pug...akefile.PL does not agree with you! :) 06:31
gaal !
fixing
must have fixed and forgotten to upload :/
uploaded. 06:34
commute& 06:35
Khisanth wonders why ghc requires over 600MB to compile Pugs.Run 06:51
xerox Howdy! 07:20
wolverian morning 07:21
xerox What's this Sunday's pretty-Unicode-sigil? :-)
pmurias morning
gaal Khisanth: because that file includes the precompiled prelude. 07:22
but 600mb is a little much.
are you precompiling other modules besides? Test.pm etc?\
Khisanth: $CPAN/authors/id/G/GA/GAAL/Bundle-Pugs-SmokeKit-0.04.tar.gz 07:23
and thanks for the report :)
wolverian xerox, Ī©, the camel's butt
xerox pokes wolverian
wolverian am I really the only one who sees it? :)
xerox What does it do ?
I see it :D
wolverian you can figure out what it does. I just made it up. 07:24
xerox Hmpf.
xinming seen autrijus 08:24
jabbot xinming: autrijus was seen 4 hours 6 minutes 17 seconds ago
xinming wonders why pugs developement progress is slown down... 08:26
liz6 no worry, it will speed up again this week... ;-) 08:27
xinming Why the commits aren't so much as before...
liz6: hmm, what's happening? 08:28
I finally got a rest after I come back to WenZhou.
dduncan I'd say its partly all those conferences that people are going to
autrijus seems to be going to be chain-attending several of them 08:29
liz6 exactly, and autrijus should be able to find some "quality time" this week...
xinming Yeap, I think so, iblench and other people aren't here as usual these days 08:30
svnbot6 r7774 | yiyihu++ | Correct a translation problem. 08:31
xinming oops. I should use mistake instead of problem. :/
dduncan good night folks
xinming where to get the irc log for this channel? I can't access the link given in www.pugscode.org 08:32
rashakil I'm seeing it at colabti.de/irclogger/irclogger_log/...2005-10-30 just fine 08:34
Though, that's where the pugscode.org link redirects to. 08:35
xinming rashakil: thanks, I really don't know why I can't open directly through pugscode.org 08:45
buu So you got the log? 09:11
nothingmuch office mannars suck 09:12
i had a friend who needed a ride a few days ago
i picked him up and he spend a few minutes at work
i was on the second screen, coding
he was on the first screen, watching a movie 09:13
today someone remarked that "it's not really accepted to share company code"
fsck! the dude doesn't even know to code, and i'm more responsible than that
and what I was writing is also intended to be opensourced eventually, so it's not even sensitive IP or anything
xinming buu: Yes 09:15
xerox poor nothingmuch :( 09:16
xinming nothingmuch: what does fsck mean?
nothingmuch tr/s/u/
xinming phuck and suck?
ooops.
nothingmuch ex
cept more geeky ;-) 09:17
there was a sprocketdata (colo company) banner a while ago
J. Random Hacker calls up his colo company's support and says:
"Hi, I need you to fsck my server" or something like that
xinming I hear that some companies even don't have a good programmar, most people are hourly programmar I ever heard. 09:18
nothingmuch so the support guy says "Uh, sir, please don't use language like that"
so he says "I really need a fsck", ... eventually support guy hangs up
and then epilogue was "We at SprocketData know what fsck really means" or somesuch
gugod a awesome, educationally meaningful story 09:20
autrijus xinming: irc.pugscode.org 09:36
liz6 morning autrijus! 09:37
autrijus greetings liz!
the only thing between me and mad pugs haxx0ring is your WEP password :) 09:38
xinming autrijus: I can't even access to these site. :-)
liz6 autrijus: come and get it then...
autrijus xinming: it worksforme
xinming autrijus: I have problem get on the subdomains on pugscode.org
autrijus xinming: check your dns? it worksforme for sure 09:39
scook0 I have noticed some problems in the past with irc.pugscode.org
xinming autrijus: I think this might be filtered by Great Firewall
autrijus: I changed DNS, not works.
autrijus xinming: or use this - colabti.de/irclogger/irclogger_log/perl6
yeah, the Great Firewall of China is indeed a wonder of the world
xinming autrijus: yes, I use that now, marked favourate.
autrijus cool 09:40
re lack of commit, mostly because of I don't have net
secondly because PickAxe is such a wonderful read.
xinming where are you now? 09:42
autrijus xinming: liz6's place
ecto, netherlands
liz6 echt, netherlands to be more precise... ;-)
autrijus er.
xerox autrijus: did you meet dons lately? ;)
autrijus blames this ecto blog reader thing
but yes. echt 09:43
xerox: no, but I'm going to stay at his place on december 20(ish)
xinming wants to travel around the world...
r0nny_ rafl_: ping? 11:02
rafl_ r0nny_: pong 11:03
11:09 scook0_ is now known as scook0
Khisanth gaal: only if pugs compiles the other modules by default 11:10
r0nny_ re 11:25
rafl_: is the svk mirror up ?
rafl_ r0nny_: No. 11:28
r0nny_ :/ 11:34
rafl_ autrijus: ping 11:35
autrijus pong 11:36
rafl_ autrijus: Does parrot embedding work for libparrot.a as well as for libparrot.so?
autrijus rafl_: only .a I think, not sure ghc supports .so on lnx
rafl_ Well, how to test that? I'm currently preparing the Debian packages for parrot and would like to build the shared lib only. 11:37
autrijus I don't know... just build libparrot .so and -lparrot 11:39
and see if it finds that 11:40
gaal Khisanth: by default only Prelude.pm is compiled. But also, by default the maximum heap size for GHC is much lower: it should not (and need not) go to 600mb with anything we're doing (yet). just to make sure, can you take a look at your config.yml?
also: what GHC/platform are you using? 11:41
autrijus praises the Consolas font. 12:12
wolverian autrijus, isn't it nice :) it's fun!
autrijus the comma is magical
wolverian yes. the <> are perky, as are $ and Ā¢ (the vertical lines are slanted) 12:13
autrijus as it should be
so all in all fairly nice
wolverian ' and ' are the same though. that's weird. 12:14
r0nny_ rafl_: might it be possible to dcc me a tarball of the actual version ?
yo 12:39
autrijus journal up :)
autrijus switches to journal-before-hacking 12:40
r0nny_ i got a problem with yaml evaluation
autrijus yes?
r0nny_ there is a yaml file, i want to compute to a deep hash
svnbot6 r7775 | iblech++ | * Usual svn props.
r7775 | iblech++ | * t/pugsbugs/class_name_and_attribute_name_confliction.t: Fixed the test (the
r7775 | iblech++ | return value of a submethod BUILD is not the return value of a .new()); the
r7775 | iblech++ | test passes 3/3 now.
r0nny_ but it only maks the name of the first top-level key to a empty value 12:41
autrijus post a sample?
or better, write a test :)
(and commit it)
r0nny_ r0nnyger.mine.nu/~r0nny/config.yaml 12:42
autrijus take a moment to convert it to test format? 12:43
r0nny_ btw - the reuslting hash is {('CC' => '')}
autrijus nod, will fix
r0nny_ how to convert it to test format ?
maybe im just using the wrong indentication 12:44
autrijus r0nny_: look at t/unspecced/eval_yaml.t
add to it
the identation is correct afaics
(you are a committer right?) 12:45
r0nny_ no commiter
autrijus ok, your email?
r0nny_ [email@hidden.address]
plz dont missuse ;P
autrijus welcome aboard :)
sure
r0nny_ hmm
weit
autrijus an invitation should be in your inbox in a couple mins
svnbot6 r7776 | iblech++ | * Fixed typos in L<> links in four tests. 12:46
autrijus I'll brb -- fetch drink
svnbot6 r7776 | iblech++ | * util/catalog_tmpl/directory.tmpl: Colorized the ok/todo/failed nums
r7776 | iblech++ | (with the same colors as the smokeserv uses) and added title="..." attributes
r7776 | iblech++ | for maximum usability.
r0nny_ i think someone put me on some list about 4 or 5 days ago
hmm
*trying to remember*
autrijus yeah, that's AUTHORS file
not committer list
r0nny_ hehe ;P
autrijus so you'd still need to register to set a password for svn commit :)
gaal shouldn't there be a yaml doc separator before the CC: line?
autrijus gaal: optional
gaal "---\n"
autrijus only the very old YAML.pm insists on it
it's long gone rom spec
s/rom/from/ 12:47
perl5 really needs a first-class syck binding
it's sad to see RoR people praising YAML
gaal i'm too used to using implementations as ad-hoc specifications :/
autrijus and binding is there for py, php, ocaml, etc
r0nny_ gaal: i removed it after it dod not work ;P
autrijus except perl ;)
(well to be fair there is a very primitive YAML-Parser-Syck)
but no dumper part, and the parser is old too
any takers? :) 12:48
if not I may try to trick liz into doing it :)
gaal :)
autrijus drink &
r0nny_ !eval eval("CC:\n FOO: bar\n",:lang<yaml>) 12:51
autrijus it's ?eval
r0nny_ ?eval eval("CC:\n FOO: bar\n",:lang<yaml>)
evalbot_7776 {("CC" => {("FOO" => "bar"),}),}
r0nny_ ?eval eval("CC:\n FOO: '%bar do bar'\n",:lang<yaml>) 12:52
evalbot_7776 {("CC" => {("FOO" => "%bar do bar"),}),}
r0nny_ ?eval eval("CC:\n FOO: '%bar do bar'\n bar: '%FOO'\n",:lang<yaml>) 12:53
evalbot_7776 {("CC" => {("FOO" => "%bar do bar"), ("bar" => "%FOO")}),}
r0nny_ why does it work with evalbot, but not on my files -_- 12:54
omg 12:55
i think i found the problem
=<fil> doenst slurp right :/
fil=file 12:56
wolverian =$fh
theorbtwo =<filename> should to, iirc...
r0nny_ since i do eval(=<config.yaml>,:lang<yaml>)
theorbtwo (But perhaps I recall incorrectly, or perhaps the spec changed while I wasn't looking, or perhaps it is unimplmented.) 12:57
wolverian ?eval slurp "foo" 12:59
evalbot_7776 Error: No compatible subroutine found: "&slurp"
wolverian arr!
r0nny_ eval and =<file> wont work together well 13:00
rafl_ autrijus: Did linking against a shared lib work? 13:01
xinming who edit the test in t/pugsbugs/class_name_and_attribute_name_confliction.t ? 13:05
hmm, the updated version doesn't show the bug, 13:07
rafl_ xinming: use svk/svn blame
Or svk log.
autrijus rafl_: I've never used that with GHC. 13:09
it's sad that "svk annotate" and "svk praise" are both longer than "blame"
wolverian: slurp is IO no? 13:10
r0nny_ i got a syntax error: line 2, column 4
at tmp-Prelude.pm line 119, column 30-64
autrijus therefure unsafe
xinming hmm, autrijus, can attribute has the same name as class name?
autrijus xinming: yes
xinming hmm, Ok, revert the file in t/pugsbugs/class_name_and_attribute_name_confliction.t please. 13:11
autrijus: hmm, that's a bug, the test is changed into fit pugs... not the bug report. :-( 13:12
autrijus xinming: iblech is responsible 13:13
the log says
return value of a submethod BUILD is not the return value of a .new()
r0nny_: can you commit the test that triggers the syn error? :)
autrijus can't fix something that is not a test...
r0nny_ syn error ???? 13:15
autrijus r0nny_: the "tmp-Prelude.pm" syntax error
xinming ?eval class a {has $.a; method update { $.a; } }; class b { has $.a; submethod BUILD { a.new( a => $.a ).update; } };class c { has $.b; submethod BUILD { b.new( a => $.b ); } };c.new( b => 30 ); 13:16
r0nny_ autrijus: im not sure, what made it
evalbot_7776 (no output)
autrijus r0nny_: nopaste the program?
xinming this will cause pugs run infinitely.
r0nny_ atm the error doesnt happen 13:17
autrijus k
r0nny_ the =<> op wont work with variables :/ 13:18
autrijus =$fh ?
oh you mean the string form
r0nny_ i mean filenames
autrijus =open($filename) perhaps? 13:19
xinming ?eval class a { has $.a; method update { $var -= $.a; } };class b { has $.a; submethod BUILD { a.new( a => $.a ).update; };};b.new( a => 20 );
evalbot_7776 Error: No compatible subroutine found: "&update"
xinming autrijus: these are 2 bugs. :/
r0nny_ prelute error again
xinming But iblench might not understand what I want to test really. he changed the class name. 13:20
autrijus r0nny_: nopaste the code? 13:21
xinming: hm, add some comments to it, then
and revert that part of change
xinming don't know how to do that exactly. :-( that's why I want you to revert it. 13:22
r0nny_ rafb.net/paste/results/nZRCGy89.html
autrijus iblechbot: ideas?
r0nny_ i think line 9 causes it
autrijus k, fixing 13:24
I think the =open form is broken for some reason 13:25
r0nny_ hmm
rafl_ Why are haskell libs linked statically by default?
autrijus rafl_: speed mostly 13:27
and versioning
because of separate compilation
ghc user manual has something to say about that
r0nny_ autrijus: it breaks also with @foo= =$fh
hmm 13:28
hw to slurp a file ?
how
autrijus @foo = slurp($fh); 13:29
@foo = $fh.slurp;
rafl_ autrijus: Thank you.
autrijus rafl_: np
r0nny_ damn - it even breaks if i use slurp 13:30
autrijus weird 13:31
r0nny_ what kind of variable returns slurp ? 13:33
if fails on eval("file".fslurp,:lang<yaml>
rafb.net/paste/results/MY5Lfi72.html 13:35
is there a way to tell eval to use a file instead of a string ? 13:36
anyone ? 13:38
btw - ith there a way, to keep the tmp-Prelude.pm for checking ? 13:39
autrijus r0nny_: eval to use a file?
it's called require :) 13:40
r0nny_ autrijus: has require a :lang<foo> param ?
and does it have a result like a hash ?
autrijus no and no I'm afraid 13:41
although it would make sense to do so
let me look at the slurp problem some more 13:42
I think it's 3 bugs ;)
stevan 13:46
!
gaal wasn't tmp-Prelude renamed to Prelude.pm a while back for better error messages? 13:47
anyway you should have it in the source tree root, r0nny.
is "fslurp" a typo? 13:48
autrijus r0nny_: right. are you using svn or 6.2.10? 13:49
r0nny_ oh - i just see its 6.2.9 13:56
autrijus er. grr.
svn co svn.openfoundry.org/pugs # please do this :) 13:57
stevan_ hola senor autrijus
autrijus hola senor stevan_
svnbot6 r7777 | yiyihu++ | The tests are right, reverted and add some comments about the test.
13:57 stevan_ is now known as stevan
autrijus how's your schedule today? 13:57
r0nny_ i think there was a way, to do evn ebuilds
stevan autrijus: I have about 1.5-2 hours right now
r0nny_ will do ine for pugs
autrijus r0nny_: you on gentoo?
stevan: ok, are you on macosx or on non- ? 13:58
stevan yes, OS X
r0nny_ autrijus: yeah
autrijus stevan: cool, give me ~10min
stevan ok
autrijus I'll fetch ibook and SEE with you
do you have a public IP?
stevan ok
autrijus or should I set up one in feather?
stevan I dont
autrijus that's fine, we'll use feather
stevan feather is good
autrijus fetch ibook & 13:59
r0nny_ wich version of parrot does the svn version need ? 14:05
autrijus r0nny_: also trunk 14:06
r0nny_ so it needs the svn version of parrot, too
hmm
autrijus yeah
but not neccessary
because you can build without parrot linking
that's fine
you can get rules support for 0.3.0 as well 14:07
r0nny_ ok - im going to use parrot-0.2.3 till i got a parrot svn ebuild
autrijus sure... then rules may break
but other things will all work
stevan: skype + seE? 14:08
svnbot6 r7778 | iblech++ | t/pugsbugs/class_name_and_attribute_name_confliction.t: Explained the current
r7778 | iblech++ | problems of the test in detail (with proposed fixes).
r7779 | stevan++ | Perl6::ObjectSpace -
r7779 | stevan++ | * first sketch for the object space, this has a long
r7779 | stevan++ | long long long long way to go.
r7779 | stevan++ | NOTE: based heavily on fglocks++ and putters++
r7779 | stevan++ | Perl6::Value/PIL::Run work, however all bad ideas
r7779 | stevan++ | are completely mine ;)
stevan autrijus: lemme download skype, but I have SEE ready 14:09
theorbtwo Is SEE better then gobby?
autrijus theorbtwo: yes.
not groundbreakingly better, but better
mostly it's the scrollbar and cursor chasing
stevan: connect to feather
theorbtwo Nod. 14:10
stevan with SEE?
autrijus yeah
see://
stevan k
autrijus set up scary tcpxd and ssh-R magic with Kwiki::Edit::SubEtha
stevan is installing skype now 14:11
autrijus btw, other people with SEE are welcome to join also. 14:13
xinming ?eval class a { has $.a; method update { $.a; } };class b { has $.a; submethod BUILD { a.new( a => $.a ).update; }; }; b.new( a => 20 ); 14:14
evalbot_7779 Error: No compatible subroutine found: "&update"
xinming hmm, is this right behaviour?
autrijus prolly not 14:15
that's a good test
works if you rename?
r0nny_ *afk* 14:16
xinming autrijus: hmm, could you please check out the file and tell me where should I change? :/ 14:17
autrijus a sec
which test is this again?
xinming t/pugsbugs/class_name_and_attribute_name_confliction.t 14:19
?eval class a {has $.a; method update { $.a; } }; class b { has $.a; submethod BUILD { a.new( a => $.a ).update; } };class c { has $.b; submethod BUILD { b.new( a => $.b ); } };c.new( b => 30 );
evalbot_7779 (no output) 14:20
xinming is a bit confused about indirect creating the instances.
autrijus BUILD doesn't return an instance
it's not like new
it modifies stuff
stevan xinming: BUILD is passed the recently created, but not initalized instance 14:21
xerox How do you read '$'? :)
autrijus dollar
xerox What's its sense?
autrijus "singular" 14:22
xinming hmm, for my second example, c.new( b => 30 );
c.new, call BUILD in c class, right?
:-S 14:23
But I didn't return an instance, I just use BUILD to call another class, 14:25
just use BUILD to call another constructor. 14:30
autrijus yup
so it is a bug alright... 14:31
if you rename the "a" class to "foo" does the bug go away?
xinming autrijus: yes, If I change the class name into upper case, the bug will disappear 14:33
?eval class A { has $.a; method update { $.a; } };class B { has $.a; submethod BUILD { A.new( a => $.a ).update; }; }; B.new( a => 20 ); 14:34
evalbot_7779 B.new(("a" => 20),);
xinming ?eval class a { has $.a; method update { $.a; } };class b { has $.a; submethod BUILD { a.new( a => $.a ).update; }; }; b.new( a => 20 );
evalbot_7779 Error: No compatible subroutine found: "&update"
xinming ?eval class A {has $.a; method update { $.a; } }; class B { has $.a; method BUILD { A.new( a => $.a ).update; } };class C { has $.b; submethod BUILD { B.new( a => $.b ); } };C.new( b => 30 ); 14:36
evalbot_7779 C.new(("b" => 30),);
xinming I don't think we need to use method or submethod here, as the Synopsis said we can BUILD as method,
we don't need to care about method or submethod 14:37
?eval my $x = 100; class A {has $.a; method update { $x -= $.a; } }; class B { has $.a; method BUILD { A.new( a => $.a ).update; } };class C { has $.b; submethod BUILD { B.new( a => $.b ); } };C.new( b => 30 ); $x.say 14:46
evalbot_7779 70 bool::true
autrijus nod 14:51
ok, I got it
will fix
(conf calling with stevan)
xinming autrijus: hmm, where do I have to change in my test? 14:53
autrijus mm 15:09
ods15 ?eval my $a = 3; $a = $a++; 15:12
evalbot_7779 \3
ods15 ?eval my $a = 3; $a = ++$a;
evalbot_7779 \4
ods15 ?eval my $a = 3; ++$a = $a;
evalbot_7779 Error: Can't modify constant item: VInt 4
ods15 ?eval my @a = qw(1 2 3); for (@a) { say pop @a } 15:13
evalbot_7779 3 2 1 undef
ods15 ?eval my @a = qw(1 2 3); for (@a) { say pop @a }; 1
evalbot_7779 3 2 1 1
ods15 ?eval my @a = qw(1 2 3); for (@a) { say "$_ -> ".pop @a }; 1 15:14
evalbot_7779 Error: unexpected "f" expecting ";", statements or end of input reserved word
ods15 ?eval my @a = qw(1 2 3); for (@a) { say "$_ -> " ~ pop @a }; 1
evalbot_7779 1 -> 3 2 -> 2 3 -> 1 1
ods15 ?eval my @a = qw(1 2 3); for (@a) { say ",$_ -> " ~ pop @a }; 1
evalbot_7779 ,1 -> 3 ,2 -> 2 ,3 -> 1 1
ods15 heh odd
xinming autrijus: ping 16:01
autrijus xinming: pong 16:13
stevan++ # very successful design session 16:15
subethaedit++ skype++ # technology
notes committed. questions and feedbacks and sanity checks etc welcome
svnbot6 r7780 | autrijus++ | * Object Space design notes, with Stevan 16:17
autrijus liz6: you got Consolas font? they are amazing 16:25
can transform one's psyche into the outer space
liz6 interesting... ;-) 16:26
autrijus (ok, perhaps not that, but still impressive)
stevan autrijus++ # always enjoyable hacking sessions :) 16:31
autrijus fyi, url: svn.openfoundry.org/pugs/docs/notes..._space.txt 16:32
stevan I will continue on the core runtime stuff, and read more about YARV/PyPy 16:34
obra Consolas?
stevan has also dug out his Pickaxe book too
obra: it is the new MS fixed with font 16:35
for Visual Studio on Longhorn I think
obra stevan: nicer than andale mono?
. o O { URL? }
stevan I prefer monaco myself .... but digging up URL from history
obra: blog.bluecog.co.nz/archives/2005/08...olas-font/ 16:36
obra Want. ;) 16:37
stevan this is kind of funny too : weblogs.asp.net/cazzu/archive/2005/...Porno.aspx 16:38
"Consolas in spanish is as close as it gets to the "masturbate" verb inflection"
stevan is off to do errands in meatspace, adios & 16:39
gaal autrijus: does Consolas have Unicode goodies? 16:41
autrijus gaal: let me look 16:44
it's latin cyrrilic greek 16:45
not comprehensive, not bad for monotype
gaal I will never leave Courier New it seems :-(
theorbtwo So it has everything core wants, but not everything Set can use. 16:46
gaal: Fontsets are the answer.
autrijus gaal: you don't have fallbacks? :)
theorbtwo BTW, is it possible to do usable Hebrew monospace?
gaal It'll always be ugly, since some letters really are tiny and narrow. (yod, vav, nun) 16:47
yod and vav in fact look almost identical except for size. 16:48
err, yod and reish that is.
autrijus dduncan: hey :) we are finally getting to get your modules working For Real 16:49
gaal autrijus + theorbtwo: windows is my desktop environment these days, so not really :( 16:51
xinming Is parenthesis needed in method call within a class? 16:52
gaal not if it has no parameters.
how's ./ spelled this week?
xinming ?eval class A { method foo( *@a ){ @a.perl }; method bar( *@a ) { .foo( @a ); }}; my @x=("f","b");A.new.bar( @x ); 16:53
evalbot_7780 "[\"f\", \"b\"]"
xinming ?eval class A { method foo( *@a ){ @a.perl }; method bar( *@a ) { .foo @a ; }}; my @x=("f","b");A.new.bar( @x );
evalbot_7780 Error: unexpected "@" expecting operator, postfix conditional, postfix loop, postfix iteration, ";" or "}"
gaal btw, nice paper about namespaces and haskell: research.microsoft.com/Users/simonp...g_conf.pdf 16:54
xinming wishes that parenthesis can be omited if a method is called within class.
autrijus ooh, more spj paper goodies. 16:59
xinming: what would this mean
.foo 1 + .bar 2
if you don't require parens?
gaal i don't see how this in particular can be used in pugs, but it's great for when you write eg gtk code. 17:00
xinming .foo( 1 + .bar 2)
liz6 or .foo( 1 ) + .bar( 2 ) ? 17:01
xinming first method will take all things as parameters since meet the ';'
rafl_ What's Perl6-ObjectSpace for? 17:02
autrijus I think it's principle of most surprise, either way
rafl_: unified runcore of Perl6-* for p5
with the metamodel bootstrapping bit reusable among js hs pir etc backends
so all backends can finally get same obj semantics
rafl_ autrijus: It superseeds perl5/Perl6-*?
autrijus which is what 6.28.0 is all about
yes.
xinming This is a bit like `sub foo ($a){...}; sub bar($b) {...}; foo 1 + bar 2 `; 17:03
rafl_ autrijus: We'll need a Makefile.PL for it so it gets built with the other perl5 stuff.
autrijus rafl_: we will, but we'll do that upon release, not now when we are not even sure of the layouts
xinming what does `foo 1 + bar 2` mean is same as .foo 1 + .bar 2
rafl_ autrijus: OK. 17:04
autrijus xinming: right. that is ergonomics domain; I just note that it looks unfamiliar to me
xinming: but if you perceive it as ergonomic, please post to p6l :)
gaal how does a p5 runcore unify things for js hs pir etc backends? 17:05
autrijus gaal: by separating the metamodel bootstrapping sequence from the underlying Chaos (primitives) 17:06
gaal: so basically, think of this YARV thing
we implement a YARV like VM with p5
then spec our metamodel as a set of YARV instruction sets
the VM is so minimal that it doesn't handle anything other than prim obj dispatch and opaque storage 17:07
gaal how does PIL figure in this?
autrijus think of it as "the subset of parrot we need to port to p5 to run p6"
PIL is core language 17:08
gaal PIL, not pir
autrijus the obj space is core runtime, so to speak.
core as in "we can add sugar and optimization and shortcut all we want"
but if you implement this, we can run p6 on your runtime, although very very slowly 17:09
gaal so: p6 -> PIL -> ..? p5 with objectspace?
autrijus p6 -> PIL -> p5codegen -> p5objspace
p6 -> PIL -> jscodegen -> jsobjspace
and p5objspace share with jsobjspace the same metamodel
I mean the same code
gaal oh; okay, so
hm, how's that possible?
they're coded in p5 now
no? this is a p5 file: rt.openfoundry.org/Foundry/Project/...m?rev=7779 17:10
gaal needs deconfusing. 17:11
integral this is one of those cases where's it's just all layers of indirectionn...
gaal what is the base runcore written in then? 17:12
p6 too? (ideal, i guess)
autrijus gaal: that is an empty file ;)
gaal but then what is the recent commit? just thunks (in the API sense, not the closure sense)? 17:13
autrijus thunks yes
gaal: the MM::* will conform to an API not unlike the YARV instruction table
xinming autrijus: hmm, how man commiters from mainland by the way? Why I ping you just now is because of this question... But forgotten after you responsed. :-/
autrijus or the PyPy Bytecode cable
xinming s/man/many/
autrijus xinming: I don't know, check AUTHORS
probably only you
s/cable/table/; # freudian 17:14
gaal: ah, I've read that spj paper. makesense
gaal so www.atdot.net/yarv/insnstbl.html is our inspiration for a pugsy runcore/vm?
xinming autrijus: I've checked that, But hard to find as some Characters are the same. 17:15
autrijus gaal: only the OO parts not flow
gaal: so really just 20~30 opcodes
and this is API not really bytecode
so they nest, unlike (say) PIR ;)
gaal pritty pictures would be useful at this point, i think. 17:16
autrijus it's basically a API over Chaos.pm
yes. I'll see to that.
gaal autrijus++ # omni, er, science.
autrijus ;)
xinming pugs now can do many things... I just now rewritten a small perl 5 script of mine into perl 6. :-P 17:17
gaal this is instead of doing the OO runcore coupled into the current haskell runcore, yes?
xinming much smaller in perl 6
gaal (testing my understanding) 17:18
autrijus gaal: the hs runcore will implement the same api
which allows me and stevan work in parallel without dup work
gaal hmm, but it's not like OO itself (the MM) is written in P6 and then we get runcores for free in all languages? how come? 17:19
autrijus so, the runcore can be written in procedural p6 17:20
and the runtime just has to support that part and things will Just Work
that is still the plan
gaal uh huh, that's how i was imagining it
autrijus but if the runtime decides it's too slow
as it would be
gaal :)
autrijus then the runtime will naturally ask
can I reimplement the MM? 17:21
and they all do
but then we get mismatch
because the outside API of MM is huge
and varies from lang to lang
so each bootstraps differently
which is good, but would be better if we can say
"implement this set of prims fast enough, and you get MM in your language natively"
gaal gotcha now. 17:22
autrijus so it's like, refactor the JS MM, the P5 MM, the Hs MM, and try to find the midlayer of limited vocab
where they all have in common
and settle the complexity there
this is practical because, well, the upper layer MM keeps changing every week 17:23
;)
gaal :-)
autrijus so we get same motivation as PIL, only at the OO space.
(which is natural seeing we are moving to 6.28.x)
so think PIL as the way to constrain the complexity space for the imperative runcore in 6.2.x
and ObjSpace doing the same for 6.28.x semantics. 17:24
I'm sure we can find more abstractions and indirections for grammars, importing from Patrick
gaal Perl6::Core::Hash demands string keys - is this an oversight or an implementation leakage? 17:26
Khisanth ack!
Setup: waitForProcess: interrupted (Interrupted system call) 17:27
gaal Khisanth, your GHC is loopy? Version? 17:28
autrijus gaal: er what?
the Str key is for Opaque
look again (and maybe fix the wording)
Opaque obj instances are essentially attrs of name=>container slots 17:29
and name here is fixed Str
gaal Core::Hash / hash is just a slot dict for objspace objects? 17:32
autrijus oh you are talking about code 17:33
sorry.
thought you were talking about design notes 17:34
gaal oh, i'd better read those :)
autrijus yeah :) the code is a mess.
and subject to lotsa moving.
gaal url for notes? 17:36
autrijus svn.openfoundry.org/pugs/docs/notes..._space.txt
gaal ah, found it thanks
autrijus feel free to edit them and sprinkle with questions and reformat and maim and torture the text 17:37
gaal better than torturing hackers 17:39
however! i need to debug this c beast for $work and get home first :) 17:40
autrijus have... appropriate amount of fun 17:42
gaal appropriate amount of coffee you mean! 17:43
thanks :) &
r0nny_ re 18:02
autrijus yo 18:04
r0nny_ finally found a ssh cappable pc :)
autrijus excellent :) 18:07
Juerd Blegh, I hate it when people hide behind Anonymous Monk. 18:08
(And change their HTML writing style accordingly, so you can't even use simple analysis to find out who's behind it... :)) 18:09
autrijus resists the temptation to suggest Consolas Monk
Juerd The font? :) 18:11
autrijus yeah. I was using Anonymous
xinming r0nny_: why not use putty?
Juerd Ah :)
I like Consolas
I hate Microsoft for having pretty fonts.
autrijus eh, isn't that somewhat destructive :) 18:12
Juerd It's the only way I could think of to turn this positive thinking about MS into something negative ;)
luqui bonjour nothingmuch 18:15
eric256 morning
Juerd Ah, hi luqui
I was hoping to find you online some time :)
luqui: Would it be possible to change evalbot so that it does a /nick, rather than reconnect, on svp updates? 18:16
luqui er...
Juerd luqui: I don't know about the possibilities of IPC with Pugs
autrijus sockets?
Juerd autrijus: Signals or something alike would be easier
luqui yeah, evalbot would have to talk to a perl 5 IRC interface
r0nny_ xinming: putty makes a pc ssh-cappable 18:17
Juerd luqui: Anyway, if you think it's doable and not too much work, could you please?
eric256 how does it update now? i was looking at the code the other day an it doesn't seem built to do that. ;)
autrijus Juerd: we can have them... once Liz gets some cycles, we'll do a design run
Juerd If it's either not possible or too much work, forget it for now
autrijus Juerd: it's mostly S17.
Juerd autrijus: I see
autrijus or the lack of.
luqui autrijus, do we have IPCish stuff in pugs?
r0nny_ rafl_: ping?
autrijus luqui: only sockets and kill
Juerd luqui: -e $file? :)
(MS-DOS age IPC) 18:18
luqui Hmm... I just think it would be nice to run the evalbot IRC interface with the latest version of pugs
autrijus it would
luqui but I guess that isn't quite necessary 18:19
autrijus luqui: how's AG coming in wither p5, p6, or pir land?
eric256 it doesn't effect its eval capabilites (or affect...i'll be darned if i can remember which of those to use)
luqui L::AG 0.07 in perl5 is quite nice nowadays
and other than that I've been focusing on school
Juerd luqui: Does it really matter, as long as the ?eval's are executed by the latest?
luqui but perl6 is my next target
Juerd, right, probably not 18:20
any particular reason why, other than annoyance?
Juerd luqui: I'm afraid of k-lines 18:21
luqui ?
Juerd k-lines are server or ircnet-wide IP-bans
autrijus luqui: cool, hopefully OO will be robust enough for you to serve as the first victim^Wtest
luqui ahh
Juerd I don't know how tolerant freenode is 18:22
luqui okay, that's definitely a valid concern
considering that it would be banning feather
Juerd Yes. And that would hurt more than just evalbot, you see.
autrijus can talk with ops about that no?
Juerd I do like evalbot, though
autrijus I mean kline policies
Juerd autrijus: I've tried, to no avail 18:23
And freenode lacks well defined rules
autrijus evalbot used to eval the last version without /nick ing
iirc
luqui yes
eric256 autrijus: still does
luqui :-p
Juerd The revision in the nick is useful. 18:24
nothingmuch evening.
Juerd Especially when pasting the output, or reading a log, later in time
nothingmuch autrijus + Liz: please publish quick drafts for review
I am very opinionated on threads but very lacking on time
luqui so autrijus, what makes io so important?
nothingmuch and would like to bring up ideas, issues, etc
autrijus nothingmuch: will do; you on apple and hence SEE reachable? 18:25
luqui: zero keywords
luqui: clean semantics without cruft
eric256 is there any reason S29 is in the regular repo? i mean why is it on a seperate site? couldn't the author working on it just get a commit bit?
nothingmuch yes/
autrijus so you can think of it as Scheme in the OO age
nothingmuch i'll leave my laptop unsleeping with see open
luqui okay, that makes sense 18:26
nothingmuch or the wiki is also good
autrijus and scheme = important if you are into dirty^Wimpure functional langs
luqui :-p
nothingmuch actually even better, since I won't be here tonight and can't contribute
and wiki is documented for everyone
autrijus nothingmuch: liz is still fighting with mysql I believe
so won't be tonight either way
nothingmuch okay
nothingmuch never heard of fighting mysql
Juerd zero keywords 18:27
autrijus mumbles something not very positive about mysql3.
Juerd we can do that!
unicode has way more than enough characters ;)
autrijus Juerd: lol
nothingmuch APL is *last centuries* line noise... Perl 6 is the future!
Juerd Sometimes I think that PHP was invented more or less the same way.
nothingmuch today someone on #perl said 'lolz' every tiime someone made a counterpoint to his pugs trashing 18:28
Juerd "We can do that too!" without sufficiently profound knowledge and wisdom regarding "that".
luqui ugh
nothingmuch in the end we got to explaining it all with the phallic aspects of unicorns
which brings me to something very troubling: www.boingboing.net/2005/10/29/nasas...penis.html
appearantly this dancing penis has a fetish for feet, and doesn't think that much of hands 18:29
Juerd nothingmuch: ...
nothingmuch anyway, I have to go sand some shelves
eric256 ...should test files be able to link to other test files? might make sense to be able to lead a person to the final home for a test. i think right now we have quite a few duplicate tests. just a though 18:33
t 18:34
svnbot6 r7781 | eric256++ | Fixing existing test-cross references. 18:39
autrijus eric256: I don't see why not. 18:40
eric256 once i finish my seemingly unending look for existing but broken links.. ;) thats getting pretty close to done though i think. then i'll work on reorganizing a little bit ;) 18:43
luqui bbiab
autrijus eric256++ # Organizer
eric256 not normaly. just want to stay active while i wait for OO and sense i normaly have a hard time finding tests and documentation it seemed like a natural job for me ;) 18:45
there is also a mix of commenting styles driving me mad. lol. pod, kwid and # ;) /me wonders if anyone is working on a perl6tidy yet. that would be good to apply to most tests too 18:47
autrijus theoretically that is just pugs -CPerl6 18:49
eric256 but i want a --style=HTML option so a simple css file colorizes the output. ;) that would be sweet 18:50
-C doesn't have a Perl6 option yet?
Khisanth gaal: 6.4 18:53
autrijus eric256: hence "theoretically"
eric256: actually you want vimhtml 18:54
with perl6.vim
that'd do
and Works Today (tm)
eric256 hmm. i'll have to look at that. i was thinking it would be nice with the tests.pugscode.org html...i wonder if i'm the only one using that? 18:55
Khisanth gaal: sial.org/pbot/14013 the config.yml
autrijus I do look at that.
it is nice.
eric256 getting there. and whoever does the smoke server linked to it so you can click failed tests there and see them! and hopefully a link to the SA or E that explains why the test should pass. ;) 18:56
i wonder if documenting broken/todo tests would be the next step, they are the ones that would probably need links to documents the most
autrijus iblech is the smoke server guy... though code is in utils/smokeserv/
right
and it's probably a matter of modifying the smoke html matrix 18:57
eric256 yea he already did. ;)
autrijus sigh I'm so out of touch ;)
luqui can anyone recommend a good cross-platform network library for c++? 19:13
theorbtwo eric256: Make a hitlist of erroring tests without links to documentation nearby in random order? 19:14
eric256 theorbtwo..yea that might work. or just list the broken tests with links to them, most don't have any documentation to speek of ;) 19:19
theorbtwo That's another idea. 19:23
I'm off to the pub.
autrijus luqui: NSPR? 19:24
svnbot6 r7782 | eric256++ | Still fixing cross reference links. :)
eric256 later all. /me needs to not stare at tests for a while ;)
autrijus ciao :)
autrijus dinnereth &
gaal Khisanth: try 6.4.1 19:59
Khisanth hrm what does ghc do with opengl? 20:06
integral it includes an OpenGL binding... 20:07
geoffb <wearily> Hi all . . . 20:16
wilx ilogger? 21:40
:) 21:45
eric256 iblech++ .. just noticed he was helping with L<> links ;) 22:16
rafl_ r0nny_: What? 22:20
r0nny_ re 22:25
rafl_: do u have p4 access atm - i need to know the co cmd line 22:26
rafl_ r0nny_: I've never used perforce directly. svk has a p4 backend.
r0nny_ hmm 22:27
rafl_ r0nny_: kane uses p4 himself. You should ask him how to use it or get some docs.
r0nny_ he told, there was no access
rafl_ There is public access. 22:30
At least for reading.
r0nny_ im foing to try svk access 22:33
-f+g
svnbot6 r7783 | eric256++ | perl.t - working on de loopifiying. ;) 22:37
eric256 anyone know how to force svk to update my local copy with the newest remote copies? 22:40
rafl_ eric256: pull? 22:41
Or do you have local changes that should be overwritten? 22:42
eric256 not working. i mean it updated some files. but one file in particular will just not update
svnbot6 r7784 | eric256++ | Checking in a new template. Test-Catalog
eric256 local changes that need over written
actualy deleted the file localy and need to re-get it
rafl_ svk revert -R .; svk pull
eric256 tries that...thanks
it worked!!! what joy 22:43
thanks
rafl_ np. 22:47
svnbot6 r7785 | iblech++ | * Usual svn props.
r7785 | iblech++ | * util/catalog_tmpl/directory.tmpl: Oops, swapped the colors of todo/failed
r7785 | iblech++ | test nums, fixed.
autrijus yo. 23:28
svnbot6 r7786 | eric256++ | Finaly finished looking at all tests. Anything that looks like a link is now parsed correctly and cataloged. Now to make sure they all actualy point somewhere.
autrijus liz++ # S17 beginning to take shape 23:29
amos hmm. lib/pugs/hack.pod says "grammer" instead of grammar, and docs/src/Junc.* say list instead of pair at some point 23:36
autrijus amos: you have a committer bit? 23:41
if not, email please
(email address, that is) 23:42
23:48 amos is now known as azuroth
azuroth no, I don't 23:48
[email@hidden.address] 23:49
autrijus welcome aboard :) 23:52
please fix away
azuroth cool ^_^
autrijus don't forget adding your self to AUTHORS :)
azuroth I'm just learning haskell now, but hopefully I can get used to it soon 23:53
autrijus woot 23:54
dduncan a darwin smoke of 7782 is now up 23:56
clkao *yawn* 23:57
dduncan lots of objectspace and test updates before that ... and after 23:58