-Ofun: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by apple-gunkies on 11 November 2005.
rafl clkao: ping 00:09
clkao rafl: hi 00:17
rafl: 'sup? 00:20
rafl clkao: An svk problem: svk sync -a works fine if one of the servers is down. svk pull -a dies after the first server that's not reachable. I think pull should behave the same way sync does.
clkao uhm. n noted 00:28
rafl clkao: Thank you. 00:30
clkao rafl: you hack on js stuff as well? 00:31
rafl clkao: Not much. I only fix things that are necessary to distribute it well. 00:33
clkao ah ok
audreyt stevan_: what's the diff between cls.isa and cls.is_a ? 01:05
stevan_ cls.is_a checks the class object, cls.isa checks against the name 01:06
audreyt k
clkao heard of the language "io" ?
buu Help, I can feel my sanity slipping away!
audreyt add it as comment in .pil?
clkao: yes, minilang is under its influence
stevan_ audreyt: will do 01:07
01:07 stevan_ is now known as stevan
audreyt or rather, I designed the semantics when drugged by io 01:07
clkao oh. someone is forking it. trying to chosoe between monotone, svk, and vesta
stevan forking Io? why?
clkao ""Moose" blurgle.blogspot.com/2005/12/moose.html
stevan the language has no keywords at all.. it is highly flexible
audreyt stevan: to add fork() capability to it? ;) 01:08
stevan hmm
audreyt blurgle.blogspot.com/2005/12/thats-...ve-up.html
clkao lol 01:09
audreyt: right, so you made jsperl5 backend pass 99% of the tests, did you notice?
audreyt sounds like something we can target
stevan :)
audreyt clkao: no I did not, wow that's impressive, what is the 1%?
clkao the 1% is 'use' 01:10
you did, because all testse using "Test" are now skipped
audreyt heh...
;)
the namespace bug is still unresolved? why?
I did grep and didn't find the bug
clkao no. i thought you said you gave up last night?
er?
./pugs -CPugs -Iext/Test/lib /tmp/test.pl |grep Test::foo ? 01:11
audreyt yeah 01:12
I gave up jsperl5
not that
that is fixed
communication error 01:13
I gave up "use perl5:..." in js
the grep shows nothing here
stevan audreyt: any thoughts on next METHOD? 01:14
svnbot6 r8338 | stevan++ | PIL/Native/Bootstrap.pil - adding comment about difference between is_a() and isa()
r8337 | stevan++ | PIL/Native/Bootstrap.pil
r8337 | stevan++ | - added ::Class.isa() because it is different than ::Object.isa()
r8337 | stevan++ | t/pil/*
r8337 | stevan++ | - renamed metamodel.t to basic_object.t
r8337 | stevan++ | - added inheritence.t to test basic inheritance
clkao audreyt: oh ok. 01:15
audreyt stevan: it'd require exceptionhandling in minilang
it's essentially a throw followed by a dispatcher loop level catch 01:16
stevan audreyt: why not a lexical sub?
audreyt you are allowed to call your next method and then still return?
clkao i hate exception handling, and the people designing api for that
audreyt i.e. postprocess it
clkao it's never done right! 01:17
audreyt clkao: I'm trying to avoid that
stevan audreyt: well, that is what call() was supposed to do
clkao bugzilla.mozilla.org/show_bug.cgi?id=321021
for your enjoyment
stevan but I have been thinking that there is no need for call() and next METHOD
audreyt self.call ?
svnbot6 r8339 | rafl++ | r23721@ata: rafl | 2005-12-19 19:04:25 +0100
r8339 | rafl++ | * JIB::Repository:
r8339 | rafl++ | + Don't use Math::Conbinatorics anymore
r8339 | rafl++ | + Fixed doc typo
audreyt oh?
stevan audreyt: I would prefer self.next_method or something similar 01:18
audreyt oh so just nomenclature
stevan actually I dont like the self. part
audreyt &?NEXT_METHOD
svnbot6 r8340 | rafl++ | r23722@ata: rafl | 2005-12-19 19:12:17 +0100
r8340 | rafl++ | JIB:
r8340 | rafl++ | * Removed usage of File::Spec in favour of Path::Class.
theorbtwo Shouldn't it be a method on the current arglist?
stevan &?NEXT_METHOD is good 01:19
svnbot6 r8341 | rafl++ | r23723@ata: rafl | 2005-12-19 19:12:40 +0100 01:21
r8341 | rafl++ | * Some more comments on JIB::Repository.
r8342 | rafl++ | r23724@ata: rafl | 2005-12-19 19:29:02 +0100
r8342 | rafl++ | * Fixed a warning in the JIB::Package test suite
clkao what does jib mean? 01:22
rafl JIB Is Brilliant.
clkao brilliant.
rafl It's the package manager for Perl 6.
clkao can you promoise mean it won't suck?
rafl Or it will be or I hope it will be or whatever..
audreyt ...because .p6p would be too mundane :)
rafl clkao: It's in early stages of development. Help us to make it suck less. :-) 01:23
clkao nice try :)
i will take a look when i am bored
rafl :-(
But well.. I'd like to see my svk problem fixed first, right.. 01:24
clkao heh
svnbot6 r8343 | rafl++ | r23725@ata: rafl | 2005-12-19 19:29:47 +0100
r8343 | rafl++ | * Use Test::NoWarnings in t/020-JIB-Package.t
rafl Hm. I always forget to use --lump when pushing lots of changes :-( 01:25
audreyt that's fine, more karma for you
rafl Is it counted somehow?
audreyt if I thought --lump should be the default, I'd have made it the default ;)
clkao grins 01:26
audreyt it's counted in an intangible fashion :)
rafl At least my machines load is high enough so the push is pretty slow. So it doesn't kill svnbot6 like the last time..
clkao it's also viral - people will know what cool version control system you use with this svn repository
stevan audreyt: take a look at line 467 in svn.openfoundry.org/pugs/perl5/Perl...timized.pm
rafl audreyt: How much karma gets me eternal fame?
audreyt rafl: 1
clkao when you go there you will change your name and perhaps gender 01:27
rafl How's the karma count initialized?
clkao s/go/get
are we nearly there yet?
audreyt rafl: it's initialized at 1 01:28
clkao: what's your lexical context? "we"? "there"? ;)
stevan: got it, implementing
stevan audreyt: excellent :)
clkao that's for the compiler to figure out
stevan :D
clkao or prebranching
audreyt the compiler complains that they are unbound
clkao that's so 1970 01:29
audreyt stevan: so you can call NEXT_METHOD with different args, right?
stevan no 01:30
audreyt stevan: and tail call optimization should apply there as well?
no?
stevan I dont think you should be able to
that was what call() was for
audreyt why not?
stevan hmm
well i guess you should be allowed to
audreyt after all you can do that for &?SUB
seems weird to disallow it for &?NEXT_METHOD 01:31
stevan but it should default to the original @_
audreyt (also I wonder if we should just rename it to &?NEXT)
svnbot6 r8344 | Darren_Duncan++ | r1465@Darren-Duncans-Computer: darrenduncan | 2005-12-20 17:29:56 -0800
r8344 | Darren_Duncan++ | /ext/Rosetta-Incubator : reorganized the DATA TYPES OVERVIEW section of Language.pod to cut on redundancy
r8343 | rafl++ | r23725@ata: rafl | 2005-12-19 19:29:47 +0100
r8343 | rafl++ | * Use Test::NoWarnings in t/020-JIB-Package.t
stevan &?NEXT works too
audreyt stevan: there is no @_ in minilang
stevan actually I like that better
dduncan that was weird
audreyt and explicit rebound is clearer
stevan @_ in the abstract sense :)
rafl Oh no.. merge conflicts. I hate that..
audreyt k :) 01:32
rafl I always screw something up when merging conflicts. :-(
clkao gr!
sorry. using -l is better
trunk svk has the ability to do sm -IC
audreyt stevan: I have this weird feeling that &?NEXT can't be written in minilang 01:33
stevan it cant
it needs to be in haskell
audreyt but it's weird
stevan yeah i know
audreyt because &?SUB is minilang
stevan ?
audreyt so is $?CLASS and $?SELF
they are perfectly valid first class values
dduncan sometimes I think that svk skips or merges all sorts of weird things
stevan can be written in,.. or used in?
audreyt &?NEXT is something else
audreyt thinks about how to define it 01:34
dduncan eg, on my commit it says ...
... t/pil/metamodel.t - skipped
but there is no such file in my dir
audreyt stevan: the problem comes when you want to return &?NEXT outside your scope.
stevan dduncan: I renamed that
retun &?NEXT? as in return the sub?
clkao audreyt: oh, the other thing i wanted to ask you
audreyt stevan: yeah
clkao since i am starting to implement stuff for jsbackend in perl6
stevan that shouldnt be possible I think 01:35
clkao such as hyperoperators. those seems to belong to a common shared prelude
audreyt stevan: but &?SUB you can carry it out no problem
stevan yeah
clkao and individual backend shuld be able to override the default implementation if they have better choise
stevan maybe this is another special type
audreyt clkao: hyperops are handled implicitly in the default runcore
stevan: maybe it's not. let me think
clkao i mean the perl5 runcore could well share what i've done 01:36
stevan audreyt: can't &?NEXT just "freeze" the scope it is defined in?
avar <random thoughts> something like multi sub ncircumfis:<4 a b c d> {} .. for a 5 b 4 c 3 d would be cool
an arbitary number of -fixes ;)
clkao btw, namespace fix confirmed. hurrah! zzz 01:37
audreyt stevan: yes
stevan: and takes the parmlist for the next candidate as its parmlist
clkao audreyt: it seems default export list is broken
audreyt clkao: yup. feel free to post it to src/perl6/Prelude/Hyper.pm or something like it
stevan next canidate? 01:38
audreyt stevan: next in the mro
stevan yes
audreyt and if it's empty you don't get a &?NEXT
stevan can &?NEXT throw an error in that context?
clkao alright.. i will try to catch you sometime for that, linking the runcore build system etc
stevan &?NEXT should alwasys be defined I think
clkao audreyt: modify the test.pl to put plan 1 after 'use Test' 01:39
audreyt stevan: not sure about that... we don't have $?CLASS outside a class
clkao: k
dduncan fyi, my current goal is to have the first runnable version of Rosetta, which at least can create/drop an empty database, by Christmas ... in perl 5 and 6 01:40
audreyt dduncan: ooh, excellent
stevan I mean always defined in the context of a method body
clkao normal runcore works, can you see if pil output is sane? so i can see if it's pil2js specified problem
dduncan today's focus is on design docs ... Language.pod is somewhat analagous to the synopsis
clkao why isn't -CPIL prettyprint 01:41
dduncan and SQL::Routine is analagous to PIL
audreyt clkao: I think I know why... fixing
clkao k cheers 01:42
quick, let me kick the smoke start and go sleep
dduncan to put that in perspective, audreyt, that milestone is like your pugs 6.0.0 that just ran "hello world" 01:43
anyway ...
audreyt dduncan: that is the most important milestone of all :)
and 6.0.0 runs "hello world" but not say("hello world")
dduncan as they say, you don't get anywhere without the first step
clkao :D 01:44
svnbot6 r8345 | rafl++ | r23726@ata: rafl | 2005-12-19 19:32:31 +0100
r8345 | rafl++ | * More consistent naming of JIB tests.
clkao are we nearly there yet?
audreyt clkao: sure we are! </here-is-a-cookie>
clkao thank you miss tang! merry christmas! 01:45
svnbot6 r8346 | rafl++ | r23727@ata: rafl | 2005-12-19 19:33:34 +0100
r8346 | rafl++ | * Remove the error log after testing the error.
audreyt *smile*
clkao: so, gaal's lexical export work broke PIL export completely, yay
clkao does this translate to it won't get fixed within 18 minutes? 01:46
svnbot6 r8347 | rafl++ | r23728@ata: rafl | 2005-12-19 19:48:40 +0100
r8347 | rafl++ | * Some POD for JIB.pm.
audreyt it does. sorry about that
will research a fix only after gaal is up and about
but!
you can apply this single line patch to make it work ;) 01:47
clkao XD
clkao reminds me the voodoo fix i did with richardc today
svnbot6 r8348 | rafl++ | r23729@ata: rafl | 2005-12-19 23:26:47 +0100
r8348 | rafl++ | JIB::Repository:
r8348 | rafl++ | + Added force option to add_package that allows to put a package into a
r8348 | rafl++ | repository even if it already exists.
r8348 | rafl++ | + Fixed index file generation
clkao wgr, i am not reminding myself. stupid finger
oh, do you still update commit/committer graph? :) 01:48
audreyt not really ;)
perlcabal.org/~autrijus/tmp/bandaid-clkao.diff
cheers
that should get your smoke running 01:49
now to go back to &?NEXT...
svnbot6 r8349 | rafl++ | r23730@ata: rafl | 2005-12-19 23:38:53 +0100
r8349 | rafl++ | JIB::Repository: Improved forced add_{package,file}
r8350 | rafl++ | r23731@ata: rafl | 2005-12-19 23:39:22 +0100 01:51
r8350 | rafl++ | * Added test suite for JIB::Repository
r8351 | rafl++ | r23732@ata: rafl | 2005-12-20 01:41:40 +0100 01:52
r8351 | rafl++ | JIB: Read the abstract from JIB.pm instead of hard coding it into Makefile.PL
r8352 | rafl++ | r23733@ata: rafl | 2005-12-20 01:43:59 +0100 01:55
r8352 | rafl++ | * Increased the number of some JIB tests to be able to put some tests before them.
stevan audreyt: I should have a &?NEXT test for you in a moment 01:57
svnbot6 r8353 | rafl++ | r23844@ata: rafl | 2005-12-21 02:39:54 +0100
r8353 | rafl++ | JIB::Meta: use JIB::Constants.
r8352 | rafl++ | r23733@ata: rafl | 2005-12-20 01:43:59 +0100
r8352 | rafl++ | * Increased the number of some JIB tests to be able to put some tests before them.
audreyt stevan: good, by then I should have an implementation
clkao ok, smoke running 01:58
svnbot6 r8354 | rafl++ | r23845@ata: rafl | 2005-12-21 02:40:25 +0100
r8354 | rafl++ | JIB::Package::Source: Fixed a pod error.
clkao night all.
btw
"are we nearly there yet" where we = Pugs, there = 2 * PI 01:59
audreyt g'nite
dduncan nite
audreyt and yes we are, for some value of near and some value of yet :)
clkao oh, t/subroutines/currying.t last test is broken. was working yesterday 02:01
svnbot6 r8355 | rafl++ | r23846@ata: rafl | 2005-12-21 02:40:54 +0100
r8355 | rafl++ | JIB: Added pod and (stub) pod coverage tests
clkao take a look if you have time
*** cannot cast from VList [] to Pugs.AST.Internals.VCode (VCode)
amv 03:58:49 < clkao> ok, smoke running
hmm.. just forget it. 02:02
clkao nvm, it fails on a todo one.
but fatal 02:03
amv the mouse is quite a har peripheral to use..
+d
rafl Can someone please run the tests in misc/sixpan/JIB? I'm not sure if my merge was successful.
audreyt rafl: doing so
stevan: implemented
stevan audreyt: you bet me,.. I am almost done 02:05
svnbot6 r8356 | autrijus++ | * &?NEXT implemented -- untested, unproven, etc.
dduncan regarding $!, was it ever explicitly decided that this will be an ENV var rather than a global? 02:06
audreyt rafl: still chasing deps 02:10
dduncan: no.
it needs to be written into synopses
poke luqui?
(or lwall)
dduncan they're not on this channel now 02:11
audreyt rafl: list deps: Package::Constants and Module::Load::Condiditonal 02:12
dduncan: yeah, so p6l...
rafl: all test pass when I installed P::C and M::L::C
rafl audreyt: Can you please add the deps you found to Makefile.PL? 02:18
audreyt done 02:19
rafl Thanks! 02:20
clkao gr!
Preprocessing library Pugs-6.2.10...
Setup: got error code while preprocessing: Pugs.Embed.Parrot
svnbot6 r8357 | autrijus++ | 03:18 < rafl> audreyt: Can you please add the deps you found to Makefile.PL?
rafl is about to rewrite smokeserv 02:27
I'd like to send yaml data instead of html files, but the using YAML.pm for deserializing a >3MB YAML dump is fucking slow. 02:28
Is there a faster implementation that's usable from perl somehow?
audreyt are you versed in SWIG or XS? 02:29
svnbot6 r8358 | stevan++ | t/pil/S12_example.t - &?NEXT -- tested, proven, etc. :)
audreyt I had for a very long time wanted YAML::Syck
there is YAML::Parser::Syck
which may be what you want
stevan audreyt: &?NEXT seems to work :) 02:30
I will start testing more cases
audreyt good, because it only worked for one level :)
I fixed that --- r8359 02:31
rafl audreyt: I know XS pretty well.
stevan only one level?
audreyt rafl: ok, in that case, grab latest libsyck and put a SWIG/XS wrapper around it should get you a yaml implementation that works with other languages
rafl: you can start with what YAML::Parser::Syck has and gofrom there
rafl audreyt: YAML::Parser::Syck looks good. Thanks. 02:32
audreyt rafl: it's very long overdue for perl to get a fast yaml implementatnion -- every other language does a syck binding
rafl: it maybe sriously outdated.
ingyware :)
rafl ingy: PING :-)
dduncan okay, I sent a poke about $! to p6l
svnbot6 r8359 | autrijus++ | * improve &?NEXT support for chained NEXT calls --
r8359 | autrijus++ | t/pil/S12_example.t all pass.
dduncan after confirming what s2 currently says
audreyt ok
dduncan++
stevan: it's multilevel now 02:33
theoretically... until proven otherwise
stevan ok,.. I will put some tests together
audreyt wunderbar
anything else you need before roles?
if not I can move to containers and autoboxing 02:34
also, &?NEXT is now completely okay to pass outside methods
as it's a self contained closure
stevan I think that is good
nice
audreyt representable in minilang
so I'm happy :)
stevan cool
audreyt you have to choice of binding &?NEXT to nil or otherwise 02:35
stevan should we load the Bootstrap.pil by default?
audreyt s/to/the/
when there is no &?NEXT, that is
I think binding to nil is easier to test than missing symbol
stevan yes 02:36
you dont like it just failing?
audreyt it has to be completely constructed for each level of &?NEXT 02:37
and there needs to be a way to know whether we are at the end of MRO chain or not, I think
so binding to nil trumps binding to something that throws warning
as otherwise we would not know if that's actually the next candidate
that happens to be throwing exception itself 02:38
i.e. there'd be no way to distinguish the terminal case with the normal case
stevan ok
audreyt implemented 02:39
azuroth good afternoon
audreyt greetings azuroth
stevan: and yes, once you are happy with B.pil we will load it by default 02:40
svnbot6 r8360 | autrijus++ | * &?NEXT is now bound to nil on the end of MRO chain.
r8360 | autrijus++ | (all S12_examples.t tests still pass.)
audreyt I'd like to see roles before working on container types :)
(so that leaves autoboxing for me currently)
stevan audreyt: ok, once we have roles we can load it by default 02:41
audreyt: are you going to add any methods other than unboxed to the boxed types? 02:42
audreyt you mean default methods?
I don't know, do you think there should be any?
stevan nope
audreyt good then.
stevan I would like those to be in roles 02:43
audreyt we do need .unboxed though
because we did away with attr
$!val that is
but nothing else
stevan well, we need a way to dispatch to the native methods
audreyt yes, and I'm more comfortable with .unboxed than $!val
stevan me too
audreyt it's strange that we are in almost constant agreement, modulo five minutes here and there :) 02:44
stevan yes, we tend to convince one another when in disagreement 02:45
audreyt :)
I'll go find some food... bbl
meppl english speaking people speak "autrijus" like "audrey"? 02:50
sili_ a tree juice.
mmm mayple.
audreyt meppl: no, similar, but probably not identical
meppl k
i know a american man who calls "fritz kleiner" [fr?ts] [kla?n?r], but the people call him "fred cleaner" [fr?d] [kli?n?r] with english 'r' of course 03:03
svnbot6 r8361 | stevan++ | t/pil/next.t - check &?NEXT deeper calls & calls with params 03:58
ingy rafl, audreyt: I have been cooking on the YAML Syck stuff. 03:59
rafl ingy: How's the whole rewrite of YAML.pm going? 04:00
ingy there will be 3 YAML implementations soon, all sharing the same API classes: YAML.pm (YAML::Classic) YAML::Lite (Simpler, stabler pure perl) and YAML::Syck
rafl it is going well 04:01
audreyt where is the code?
and more importantly, where can rafl and I jump in?
rafl And when is 'soon'?
:-)
audreyt ingy: and when are you going to taipei so we can chat? :)
ingy I have basically refactored the current YAML.pm into a new setup and it works. 04:02
once I port the tests I'll release it and we can cobble together YAML::Syck
the code is in my svn repos 04:03
and audreyt, maybe this weekend
audreyt url to your repo again? 04:04
ingy rafl: soon means who knows
rafl ingy: How's the performance compared with good old YAML.pm? 04:05
audreyt easily 10x ;)
ingy: cool -- saturday afternoon I've got a conseling session, but otherwise I'm free
ingy svn.kwiki.org/ingy/YAML-Dev/ 04:06
the current performance can be expected to be the same or slightly slower :) 04:07
audreyt rafl: interested in takign YAML::Syck and run with it ? :)
ingy since it is just a refo
audreyt (I was referring to Syck's perf vs YAML.pm)
ingy I have a plan as to how YAML::Syck should be cobbled together, so I don't mind if rafl runs, but it should be on rails ;) 04:09
audreyt lol 04:10
stevan audreyt: ./pil -e '{ "one" => [ 1, 2, 3 ] }' does not work
ingy fyi, syck == simple yaml cobble-yourself-together-a-parser kit
so cobble is the appropriate term 04:11
audreyt stevan: fixing
you can code asif it's fixed :) 04:12
stevan always do :)
ingy hi stevan 04:13
stevan hey ingy
ingy: are you still working on that database idea you had in Toronto?
ingy stevan: I am working on a database idea called Goodboy :) 04:16
not sure what I talked about in Toronto 04:17
I'm also working on Perldoc and Kwiki-Kwiki
arrrrrg
ingy needs to prioritize
stevan knows the feeling
ingy I just woke up, was planning to work on Kwiki today 04:18
stevan lately pugs always bubbles to the top, and $work sinks to the bottom :)
ingy now you guys have forced me into working on YAML.pm release
well and I still have to do 5-8 hours of $work each day 04:19
stevan don't blame me,.. it's audreyt fault :)
wait,.. you *just* work up?
ingy but hey, I'm in Taiwan!
stevan s/work/woke/
ah
that makes more sense
ingy the great island of tuits!
stevan :) 04:20
stevan should try to convince his boss to send him to YAPC::tw :P
I will tell him I can pick up extra tuits 04:21
obra next year, it's yapc::asia in tokyo
stevan oh
stevan has always wanted to stay in a capsule hotel
ingy actually hcchien is planning a ::tw for after ::jp 04:22
at least that's what I heard in the winds.
stevan: I actually talked to miyagawa about staying in a capsule but he says I should just sleep in a comic book cafe 04:23
stevan :)
ingy I think they rent private rooms by the minute afaict
it's all pretty freaky over here :p
stevan short naps would be economical 04:24
ingy in .tw a dollar is 3 cents
svnbot6 r8362 | autrijus++ | * PIL.Native.Parser: allow arbitrary expressions in NativeMap keys
r8362 | autrijus++ | and values. It's desugared to {}.push() if any of key/value
r8362 | autrijus++ | expressions are compound (non-literals), otherwise it's still a
r8362 | autrijus++ | literal (thus preserving nested seqs and maps).
ingy I call .tw dollars, pennies
which makes it easy 04:25
stevan :D
ingy a $500 bill is really a $5 bill
which is *really* a $15 bill
the $5 coin looks like a nickle 04:26
and to $10 looks like a quarter
and there is no $25
audreyt 05:24 < ingy> in .tw a dollar is 3 cents
stevan recalls going to East Berlin back in the mid-80s ... now /that/ was a good exchange rate
audreyt watch for value recursion!
(i.e. "a dollar is 3 cents" implies a cent is 1/3 dollar) 04:27
ingy stevan: so anyway a comic book cafe is $1/minute
stevan I assume the yen exchange rate is not as good (for americans that is)
ingy which is roughly $1.80 USD
audreyt stevan: I think the {...} fix works; the key position is a bit too laz 04:28
s/laz/lax/
which makes {1.add(3) => 4} work
ingy so $20 would get you a place to crash
stevan audreyt: it is working for my purposes for now
audreyt not sure if it's a good thing or not, but anyway it works
k
stevan I am converting the MRO tests
audreyt TDD++ # tests far more portable than docs 04:29
ingy stevan: yeah I don't know anything about tokyo
TDD?
stevan Test Driver Development
avar Driven?
stevan yes, that :)
ingy DDD? 04:30
DDT
stevan Drug Driven Development?
audreyt Dungeons, Dragons and Debugging?
ingy Drug Driven Testing
audreyt then we just need Development Driven Drugs
to complete the cycle
obra invest in biotech
ingy or meth labs 04:31
stevan sort of a carrot and the stick idea,.. but with a joint and some kind of lego robot connected to source control
ingy audreyt: anything for me to do in pugs these days?
avar Will TDD decrese my TCO and improve by FAT?
(Funny Acronym Ratio) 04:32
;)
stevan audreyt: I think once I finish the MRO tests I will be ready to start roles
Skud avar: shirley you mean FAR 04:33
avar raTio ?;)
rafl ingy: Besides increasing YAML.pm performance so I can write a good new smokeserv? :-)
audreyt ingy: YAML::Syck definitely 04:34
avar will YAML be a standard perl6 module?
audreyt but if rafl can handle that... then I'd like a p5 impl of kwid
avar I hope so, then I can exchange ruby & perl stuff out of the box
audreyt avar: either that, or builtin
it's pugs builtin 04:35
avar audreyt: fucking awsome;)
;)
audreyt and is highly likely to be stdlib among p6 implementations if JIB uses it
avar JIB?
audreyt still not sure about sqlite
JIB is rafl+kane's way of doing a lightweight package manager for perl6 -- because perl6 supports installation of multiple versions and authorities of the same module 04:36
see docs/AES/S22draft.pod
and misc/sixpan/
stevan I am really starting to like this mini-language,.. who needs Perl 6 :P
avar yeah, I've read that
rafl We're not yet sure about what format to use to store the JIB meta information. But YAML is at least very likely if we have a good implementation of it.
audreyt reputedly it stands for "JIB Is Blamed"
because everybody blames package managers 04:37
stevan: good, because that's going to be pug's assembly (aka PIR) for the forseeable future
rafl But the docs say "JIB Is Brilliant"!
audreyt yeah, I was just glibbing :) 04:38
rafl Since today it's written down somewhere and I can tell people to RTFM instead of asking me ;-) 04:39
audreyt :)
avar I like blamed 04:41
rafl I wouldn't like it if I were JIB.pm..
audreyt: How's the Haskell implementation that's mentioned in YAML.pm's POD going? 04:42
stevan audreyt: did you change the way things are printed?
svnbot6 r8363 | xgl++ | fixed t/pugsbugs/parsing_do_my.t
stevan s/printed/pretty printed/ 04:43
audreyt stevan: not sure... not intentionally anyway
rafl: I went with a syck binding, so that para can go away
stevan Seqs were printing on a single line, but it seems if the string is very long,.. it wraps each element onto it's own line 04:45
rafl audreyt: para?
stevan maybe this is something from Text.PrettyPrint
audreyt raparagraph 04:46
stevan: it is... if it causs you inconvenience I can disable it
stevan yes please :)
it makes it hard to test
rafl audreyt: Oh, OK. Yet another reason for a new YAML.pm release ;-) 04:47
audreyt stevan: fixed 04:50
stevan audreyt: thanks
this will complete the MRO test
so I will start on Roles tomorrow morning
audreyt yay! 04:51
audreyt celebrates another milestone in the the long march
s/the the/the the the the the/; # really long
svnbot6 r8364 | autrijus++ | * PIL.Native.Pretty: always prettyprint to one line, to
r8364 | autrijus++ | make stevan++'s testing easier.
r8365 | xgl++ | fixed t/pugsbugs/parsing_do_my.t; 04:57
r8366 | stevan++ | t/pil/* - more &?NEXT tests and a bunch of MRO tests; tomorrow .... we do roles
stevan signs off ... back in 9 hours & 05:04
ingy heads off to work on YAML.pm :p 05:05
sili_ mm 05:07
fix the "or die;" because it confuses me. 05:08
avar or die "Error " ~ int rand 0xdeadbeef; 05:12
gaal g'morning 05:15
xgl midnight here
gaal g'midnight_there 05:16
oops? pil broke? sorry about that (still backlogging) 05:20
xgl what happened? a bunch of svn updates? 05:23
rafl ingy++ 05:24
gaal so, how come it broke? 05:27
audreyt gaal: pil didn't handle lexical export, that's all 05:31
not really broken, more like... bitrotted
in addition to that, to avoid namespace leaking I put (in Pugs.hs) {} around automagic inlining of "use Test" so it becomes "{use Test}" and hence lex symbols won't be exposed -- need some way for inlining _and_ main lexical export to work 05:32
gaal we use Test in Pugs.hs 05:34
? 05:35
audreyt no, in clkao's test.pl
home.clkao.org/~clkao/tmp/test.pl
but those two are separate bugs 05:36
avar what do you call something that gets more expensive the more data you process (in a non-linear fashion)
I've completely forgotten the word 05:37
xgl resource hog?
audreyt exponential?
xgl unscalable?
azuroth exponential..?
gaal I really need to get up to speed with PIL: what it implements+where, what needs to be maintained there after "main" runcore changes
audreyt gaal: I too, and I really need to start blogging about it.
or p6c it. 05:38
gaal also, did you see my wonderings about exports of variables?
audreyt yes I did
avar audreyt: ah, exponential
audreyt if it's global vars in the module site, how is it different from sub exports? 05:39
gaal sub decls have a MkCode in the parser 05:40
var decls always MkObject in Eval
(unless I'm mistaken.) 05:41
also, no word yet from @Larry about it having indeed to be a global... 05:42
ruleVarDeclaration always returns a Pad; reducePad SMy does newObject (err, not MkObject) 05:43
STemp I don't even know what we're supposed to do with 05:44
and (say) SOur... I don't know where the var is allocated :)
audreyt it's allocated upon runtime entry, not at compiler time 05:45
I think it should
be allocated at compile time, that is
gaal what func? 05:46
oh i see it 05:47
reduceSym _
audreyt yeah
gaal btw please tell me you can't export memebers :-) 05:48
("has is export") 05:49
# makes no sense
How does moving variable creation to compile time affect deparate compilation? 05:50
audreyt class is a closure... exporting per-closure-entry stuff generally makes no sense 05:51
gaal s/dep/sep/
audreyt it shouldn't affect it. each var should be serializable
also, I think SGlobal vars are already created upon parsing if I'm not mistaken
gaal I don't see code for that by vgrep 05:52
audreyt actually I think they all are... into Pad
in unsafeEvalLexDiff line 693
it's just Pad gets refreshed by Eval if it's SMy
gaal I don't understand how that doesn't get run many times in the chained case. This function is weird. 05:56
...(sym, expMaybe) <- option... recurses
so the innermost element would be created several times, no? 05:57
audreyt it might... the chained case is I think very broken...
gaal ok, so ignoring it for a moment. 05:59
audreyt yup.
gaal lexDiff contains a bunch of transformers that already reference created memory?
well, created vars
I'm not sure it does, because I put in a trace and it showed me MkPad [] or soemthing similar 06:00
ah, i still have the print
MkPad (padToList [])
this was for a simple "my $var is export" 06:01
oops, "our"
audreyt "our" creates no lexDiff 06:02
it just installs the symbol
gaal hmmm right 06:03
audreyt (which, btw, is broken as well -- it should install a lex alias)
gaal really? wow
why? performance?
audreyt no, just so this can work
package Foo; our $blah; package Bar; say $blah;
which is part of p5 and p6 spec, weird as it may seem.
so maybe the correct fix is to make our install the lex alias 06:04
and then export that lex alias in lexdiff.
maybe. I'm not exactly sure
gaal I know how it's part of p5
audreyt (and did I mention I was about to get food? turned out I didn't and it's been for hours :))
gaal but in p6, doesn't package X mean package X { } ?
ok in the "my" case this looks good:
MkPad (padToList [("$v",[(<ref>,<ref>)])])
audreyt it means {} for closure purposes
but still it shares a pad with outer
it's very weird and very special case. 06:05
I'm not sure if it's worth keeping anyway... but that's for p6l to decide
I think the Syn still preservs the behaviour
gaal okay, you get food, i go earn food
audreyt okay :) sorry I didn't help much
gaal you helped clarify a lot! :) 06:06
but this stuff's a little harder than it looks.
audreyt indeed
food & 06:07
gaal bon mousse :)
audreyt :)
GeJ greetings all 06:34
xgl greetings
dduncan question ... what is the Perl 6 way of indicating a weak reference? 07:55
eg, if $foo and $bar each contain a reference to the other, one is weak so they can autodestruct 07:56
in perl 5 you use Scalar::Util::weaken() to do what I want
Khisanth my $foo isa wimp; :P 08:01
jp-autark isn't a real GC going to fix that?
Khisanth hmm yes 08:02
since ther would be not ref to count :) 08:03
d'oh, there and no
gaal is in a maze of nested namespaces, all differently scoped 09:36
clkao Preprocessing library Pugs-6.2.10... 10:01
Setup: got error code while preprocessing: Pugs.Embed.Parrot
Build failed: 256 at util/build_pugs.pl line 86.
gr?
masak hm 10:03
t/builtins/math/NaN.t says "plan 2" and then contains 5 tests
is that really correct?
should i change it?
gaal masak: please do 10:11
masak changes it
svnbot6 r8367 | masak++ | number of tests planned should be 5, not 2, in t/builtins/math/NaN.t 10:14
10:38 kgtfr|konobi is now known as konobi
masak can I rewrite this line 11:01
$arg = defined($ctl[$CTL_DEFAULT]) ?? $ctl[$CTL_DEFAULT] !! 0;
as
$arg = $ctl[$CTL_DEFAULT] // 0;
in perl6?
rafl It would be $ctrl[$CTRL_DEFAULT] || 0 in p5 as well as in p6. 11:02
Whops. I didn't see the defined.
Yes, // is ought to work in perl6. 11:03
masak rafl: good. thx
gaal masak: // works in pugs. 11:09
11:26 lhooq is now known as dada
meppl guten morgen 12:07
wolverian hm, why does svk ci also push? 13:53
audreyt because you operate it on a mirror path instead of a local branch 13:56
wolverian oh. can it not do that for a mirror, or do I just branch?
audreyt just branch. 13:58
svk cp -m 'branch' //mirror/pugs //pugs
mirrors are by definition mirrors
audreyt goes sleeping... & 14:08
wolverian audreyt, thanks! 14:09
svnbot6 r8368 | iblech++ | Usual svn props for all this new files :) (t/pil, src/PIL, src/Data) 15:19
Qiang is there a roadmap on where Pugs is going ? where is pugs now ? 16:00
kane-xs over at pugscode.org 16:05
svnbot6 r8369 | kane++ | r580@coke: kane | 2005-12-21 16:58:45 +0100 16:06
r8369 | kane++ | * Add JIB::Dependency + tests
r8369 | kane++ | * Make JIB::Meta hold a JIB::Dependency
r8369 | kane++ | * rename some test files
r8369 | kane++ | * add more test packages/meta.infos under t/src
r8370 | kane++ | r625@coke: kane | 2005-12-21 17:27:59 +0100 16:30
r8370 | kane++ | * oops, forgot to add Dependency.pm
r8370 | kane++ | * move config test to 003
webmind errors are rewarded? ;) 16:34
theorbtwo is having container/value problems with the sentance "The Pugs project was started on February 1st, 2005 by Audrey Tang."
wolverian 'sentence' :) 16:35
webmind theorbtwo, hmmm
understandable
rewriting history 16:36
gaal doesn't see the problem? 16:54
integral no problem if history is under version control
stevan anyone know if there is a limit with the runInteractiveCommand function in pugs? 16:55
I am seeing this *** /bin/sh: runInteractiveCommand: resource exhausted (Resource temporarily unavailable) 16:56
not sure if this is pugs ,.. or me
oh well,.. I will look closer after lunch 16:59
stevan food &
gaal I still don't see the problem. That is a simple proposition, and it happens to be true as well. 17:03
theorbtwo But on feb 1st, it was Autrijus Tang who started the Pugs project. 17:04
gaal "I started writing this sentence at 19:04, IST" is also a simple proposition that happens to be true. 17:05
Ahhhh, OK.
Well, this has been addressed by theories of reference.
I forget who it was that made the point about referential failures. 17:06
theorbtwo Thus my original statement -- that I am having container/value problems.
gaal Suppose I say to you "see that guy drinking gin over at the bar?"
(and he happens to in fact be drinking ouzo) 17:07
theorbtwo Hmm...?
gaal we may have been referring successfully to the same person.
Smari gaal: Couldn't you clobber that with a fuzzy predicate?
theorbtwo The question is more if "Last week, the guy over by the bar drinking ouzo killed a man in Reno." 17:08
gaal Smari: what do you mean clobber? Could I have gone and said, "see that guy drinking a clear liquid at the bar?"
theorbtwo Now, assuming we aren't in Reno, could that sentance possibly be correct? 17:09
Smari gaal: Assume a truth value due to high probability.
gaal Sure I could have, but people communicate effectively w/o hedges like that all the time.
theorbtwo I would you have to say "...the guy who is now over by the bar..."
Smari Alas, I found out 10 minutes ago that I couldn't take a course in symbolic logic next term, because it clashes with a manditory course in metric space theory. Blehrg.
gaal Smari: I don't think probabilities need enter the picture.
Smari gaal: Aye, that's true. 17:10
theorbtwo s/I would you/Or would you/
Smari theorbtwo: Wouldn't current time be asumed? 17:11
gaal theorbtwo: most metaphysics worth their salt can deal with name changes. If not, consider it a good reason to change metaphysics. :)
"George Orwell wrote _1984_ in 1948" 17:12
It isn't his real name. And even if he took the alias *after* 1948, it would be reasonable to call that proposition true. 17:13
theorbtwo I was unaware that George Orwell was a pen name. 17:22
Smari theorbtwo: His real name was Eric Arthur Blair IIRC 17:25
stevan gaal: ping 17:42
gaal: cancel ping :) 18:10
svnbot6 r8371 | stevan++ | PIL/Native/Bootstrap.pil 18:18
r8371 | stevan++ | - skeleton for Roles and Role system bootstrapping (no implementation)
r8371 | stevan++ | t/pil/MRO.t & t/pil/bootstrap.t
r8371 | stevan++ | - added some basic tests for ::Role, moved other tests around
r8371 | stevan++ | Test::PIL::Bootstrap
r8371 | stevan++ | - added explicit :r flag to Pipe::open call, didn't help though :/
clkao why oh why oh why the smoke run is not working, but direct invocation works for jsperl5 18:26
rafl nothingmuch: ping 19:08
nothingmuch rafl: pong 19:36
rafl nothingmuch: See #catalyst 19:37
gaal stevan: pong but false 21:23
stevan gaal: :) 21:27
dduncan audreyt, the openfoundry is off again 22:03
audreyt dduncan: yes, expect it to be back within 1~2hrs 23:53
audreyt backlogs
mm, antiping packets
clkao audreyt!! 23:54
21 hours to my flight
audreyt theorbtwo: valid reference frames assumes dynamic scoping, as gaal demonstrated 23:55
theorbtwo: so it's not lexical but depends on the env frame carried by the caller (you)
which is true as no document canlexically determine the extent of all its references
clkao: congrats!
clkao 38 hours to be sit in hotsprings 23:56
audreyt woot

Powered by Cro and the Raku® Programming Language.

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

Thank you!