pugs.blogs.com/dist/Perl6-Pugs-6.2.13.tar.gz released! | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by audreyt on 17 October 2006.
Alias_ Well, it would be nice from a PR point of view to have it formalised and wrapped up in a candy-shell installer (windows) or apt-get'able (etc) 00:00
And might actually finally mean I don't have to listen to the Duke Nukem jokes any more :)
avar ascii art of the duke in the installer?:) 00:01
Alias_ That would be hilarious :)
[particle] i think duke nukem is under misc/ somewhere in the repo... ;)
Alias_ A camel with a shotgun and strippers
avar heh
TimToady I think when people start playing with 6.2.13 they'll realize it's really getting rather solid already. 00:02
Alias_ nods
I was just thinking in more symbolic terms
TimToady and I really don't mind the anonymonks proving themselves to be idiots in front of everyone. well, okay, I mind a little... 00:03
00:04 buetow joined
TreyHarris TimToady: which anonymonks are you referring to? not names, just what comprehension results in the subset of anonymonks to which you refer? :) 00:07
TimToady it seems a little odd to use the word "comprehension" on that particular subset 00:12
00:12 dakrone joined 00:13 hexmode joined 00:16 weinig joined
TimToady Parents' Back To School Night & 00:17
00:18 nekokak joined 00:46 Pomin joined, mako132_ joined 00:59 lollan joined 01:01 buubot joined 01:11 stef_ joined
svnbot6 r14411 | mgrimes++ | Fixed a number of old /regex/ style smartlinks and broken smartlinks. 01:14
r14412 | fglock++ | v6/emitter(2) 01:41
r14412 | fglock++ | - added README
r14412 | fglock++ | - fixed Value->Scalar conversion
r14412 | fglock++ | - updated HACKING, TODO
r14412 | fglock++ | - some code cleanup
01:43 mdiep joined 01:55 cognominal joined 02:03 cjeris left 02:04 Alias__ joined 02:15 dduncan joined 02:16 frederico joined
dduncan fyi, I just smoked the final release of Pugs 6.2.13; m19s28.vlinux.de/iblech/stuff/pugs-...802cd.html 02:16
lambdabot Title: TAP Matrix - Wed Oct 18 01:16:18 2006 GMT, tinyurl.com/yec5dj
dduncan the results are the same as for pre1, except that it was faster this time
specifically, it took 90 minutes today, vs 103 minutes for pre1 the other day, and around 135 minutes a couple months ago when I last did smokes of Pugs 02:17
I attribute the 13 minute difference to that I was doing some other resource-intensive activities on the computer (watching videos, etc), when smoking pre1, but basically nothing was happening today
02:18 eviltwin_b joined, shachaf_ joined
dduncan my system in both cases is Mac OS X 10.4.8, Perl 5.8.7, GHC 6.6.0, PowerPC G4 dual-1Ghz CPU, 1.25GB of RAM 02:19
and the smoke was setup to use 4 threads
so about 2 threads per CPU
that is, concurrency=4 02:20
that is all
so 135 mins down to 90 mins is certainly a significant speedup
what changed since the 135 and the 90 is mainly that GHC was updated from 6.4.1 to 6.6.0, and Pugs itself had large updates ... no other system changes of great significance to smokes 02:21
02:21 HCoyote joined
dduncan that is all 02:21
all settings re the pugs make system were the default 02:22
02:22 Alias_ joined
dduncan s/since/between/ 02:23
02:31 justatheory joined
svnbot6 r14413 | Darren_Duncan++ | ChangeLog : added template for upcoming release 6.28.0, based on the one for 6.2.13 ... tweaks may be needed 02:35
02:37 HCoyote left
TreyHarris [particle] is showing off smartlinks at the seattle perl user group meeting as we speak :) i use them every day and i forgot how very very cool they are. 02:56
audreyt future is here just not evenly distributed 03:03
diotalevi_ smartlinks? Boo. I wish I was at SPUG. 03:07
03:11 weinig is now known as weinig|sleep
cognominal better distributed on pugs than parrot? :) 03:23
audreyt maybe different timelines..
cognominal I am working on the node^Wpugscapture pmc btw 03:24
when done I will update PGE to use it 03:25
I need to support shift, unshift, iterators, freeze. Still a lof of things missing 03:26
03:27 christopher joined
audreyt *nod* cool 03:27
cognominal worked on the french Perl 6 wikipedia entry too
03:32 AzureStone joined 03:34 lisppaste3 joined 03:35 root4o joined
root4o hi, can u point me to gather() , take() explanation..or tell in several words their purpose 03:36
audreyt root4o: it's for the pattern of
my @result; while ... { ... for ... { ... push @result if ... } } return @result
now you write instead:
gather { while ... { ... for ... { take $x if ... } } } 03:37
see also search.cpan.org/dist/Perl6-Gather/Gather.pm
lambdabot Title: Perl6::Gather - Implements the Perl 6 'gather/take' control structure in Perl 5 ..., tinyurl.com/vxftl
root4o 10x
Alias_ It might just be me, but that seems like it has fairly limited gain 03:38
Although I can see how it would simplify certain expressions 03:39
audreyt Alias_: you can't refactor with push
since it's depending on a lexical 03:40
of course you can
Alias_ I get that
audreyt our @result; sub gatherer { local @result; ... taker() } sub taker { push @result }
Alias_ It's just a smart grep
audreyt but it's considered badstyle
anyway it's handy :)
Alias_ frets about keyword explosions stealing all the useful terms 03:41
say is evil
It has nothing whatsoever to do with saying anything
But then hey, I'm sure it's cleanly worked around in P6 03:42
audreyt it actually hooks into a text to speech engine
but we're not revealing that yet
;)
Alias_ heh
audreyt (nb: "print" has nothing to do with printing for a very long time now; we're not using teletypes)
cognominal I suggested to rgs to add cowsay to 5.10 but he declined 03:43
Alias_ It prints pixels on a canvas
But yes, I get your point
It's more like ->send
root4o many times in Perl5 i was tring/doing something like this (cant remember if it works now) 03:44
action1, action2 if condition;
could in some way do this shortcut in perl6
i mean two actions
and if at the end
Alias_ if at the end for complex cases of if is bad style anyway 03:45
audreyt root4o: that is valid p5
root4o yeah, but it was for simple stuff
Alias_ Multiple actions?
not simple
root4o no two simple actions
eviltwin_b it'll work in perl5 as long as you'r very careful about lists
audreyt Alias_: comma as stmt separator
eviltwin_b: yup 03:46
root4o print 'hi', $x = 12 if $cond;
eviltwin_b that'd be a prime example of "as long as you're very careful" :)
(rather of the opposite...)
audreyt in p6 there's something like
{say 1; say 2}() if 3
which is rather unambiguous to my eyes
eviltwin_b hm, can't you use a block that way (minus the parens) in perl5? 03:47
hm, no, onlyas "do {...}"
root4o one other thing that bugged me , not a big deal ... is that I had to do 03:48
my $x;
$x = 55 if $cond;
instead of just :
audreyt ?eval my $x = 55 if $cond;
03:48 evalbot_r14401 is now known as evalbot_r14413
evalbot_r14413 Error: Undeclared variable: "$cond" 03:48
audreyt ?eval my $x = 55 if 1; $x
evalbot_r14413 \55
audreyt there you go.
root4o my $x = 55 if $cond;
audreyt yeah. fixed
root4o :) 03:49
audreyt (my $x) can appear anytime $x can
which is wonderful :)
my $x[10] = 2;
et
etc
declarators are finally real expressions
root4o I use gentoo.... of cource when I install pugs it doesn't include tests and examples...I was wondering, I can write test with installed pugs...i.e. 03:54
it doesnt need additional stuff..
and I do not need to correct some PATH's? 03:55
audreyt you can simply write tests and run them with installed pugs
but "svn co svn.openfoundry.org/pugs/" will get you the test corpus in t/
root4o ok...10x, will try
lambdabot Title: Revision 14413: /
audreyt you had a commit bit yet? 03:56
root4o no
audreyt ok. what's your email addr?
root4o iraptor_at_gmail
audreyt welcome aboard. try editing AUTHORS to add your name and "svn ci" it back to test the commit 03:57
root4o where can find info in which part tests are missing
audreyt http//spec.pugscode.org/
click into specs and find noncovered parts
_so_ happy that question can be answered straightforwardly
smartlinks++
root4o: check your mail for login 03:59
I need to get lunch and pack; bbiab. have fun :)
Alias_ pack?
audreyt pack my stuff for plane tonight to portland
root4o cant load http//spec.pugscode.org/ .... is it something with my connection or it is down 04:02
Alias_ You missed the colon? 04:03
04:04 b00t joined
root4o no havent...just when wrote in the irc 04:04
sorry
hmm it seems my connection ... 04:06
04:06 Eidolos joined
root4o test 04:06
christopher root4o: any SYN link on smoke.pugscode.org also will give you the synopses 04:07
root4o 10x 04:08
lambdabot Title: Pugs Smoke Reports 04:09
root4o yes I'm getting this 04:10
how can I get supported methods for a class and/or attributes i.e. 04:20
Int.get_all_methods()
or 55.get_all_methods()
is there such beast
audreyt root4o: $obj.^methods is canonical 04:33
it's not yet implemented, sorry
04:35 Yappo2 joined
root4o bad ;) 04:35
04:36 shachaf_ joined
audreyt working on that though :) see search.cpan.org/dist/Moose/ for perl5 land prototype 04:38
lambdabot Title: Stevan Little / Moose - search.cpan.org
04:46 Alias_ joined
root4o yeah, i've tried it .. very cool 04:47
04:59 Yappo2 is now known as Yappo 05:03 root4o left 05:17 justatheory joined, b00t joined 05:21 Alias__ joined 05:26 BooK joined 05:34 AzureStone joined 05:50 buetow joined 06:07 hoowa joined 06:10 marmic joined 06:17 dakrone joined, flw joined
flw hi audrey 06:18
audreyt flw: greetings 06:23
flw Can you read Chinese on your workstation? 06:25
audreyt probably yes
try typing some chinese?
(or should we go over to #perlchina?) 06:26
flw ok, see you in #perlchina
to audreyt: ęˆ‘č®°å¾—ę‚Øę›¾ē»č®²čæ‡ļ¼Œęœ‰äø€äøŖ Module åÆ仄ꊊ Perl ęŗē؋åŗå˜ęˆ TAB 和ē©ŗę ¼ē»„ęˆēš„äø€ē‰‡ē©ŗē™½ļ¼Œå®ƒēš„名字叫什么ļ¼Ÿ 06:29
Alias__ (funny when you can't render it) 06:31
audreyt it's more mysterious that way
Alias__ (square) (square) (square) (square) (square) (square) (square) (square) Module (square) (square) (square) (square) Perl...
:) 06:32
audreyt to translate: which module was that you mentioned that turns Perl programs to tabs and spaces?
see, totally not mysterious then
Alias__ Well of course, but it has a nice parallel to listening to people talk in foreign languages 06:33
flw Alias__, I'm sorry, my english is very poor.
Alias__ Except for the words taken from langauges that stick out, they might as well be saying (square) (square) (square) (square) Module (square) (square) (square) (square) etc
flw: Don't apologise
My English is very poor, and I speak it as a first language :) 06:34
06:45 flw_ joined 06:49 iblechbot joined
Eidolos flw: To answer your question, Acme::Bleach. :) 07:03
flw_ Eidolos: Thank you! 07:04
07:05 flw_ is now known as flw 07:06 kanru joined 07:16 bsb left 07:22 zakharyas joined 07:49 chris2 joined 07:50 cognominal joined 08:11 elmex joined 08:20 b00t joined 08:24 mjk joined 08:25 andara joined
dduncan fyi, just smoked r14413 ... same results as with the last release ... m19s28.vlinux.de/iblech/stuff/pugs-...3add0.html 08:32
lambdabot Title: TAP Matrix - Wed Oct 18 08:22:43 2006 GMT, tinyurl.com/wpdz8
08:34 fayland joined
dduncan I will probably not do any more smokes until some large change has been checked in, such as to enable the new metamodel, or to remove the design cludges that GHC 4.4.1 requires but 6.6.0 doesn't 08:34
audreyt right. that seems prudent 08:35
dduncan incidentally, if it were me, I would make removing 4.4.1 support the first major change 08:36
you were probably itching to remove those cludges, weren't you?
wolverian 6.4.1, you mean :)
dduncan yes, that
audreyt well yes. in fact I am
but rafl is still solving 6.6 bindist on debian
and freebsd ports havn't been updated yet 08:37
those two represent a fairly major portion of pugs devbase
so I'd hold for another week or two
dduncan sounds like a plan
audreyt 1..1
wolverian ubuntu 6.10 is slated to have 6.4.2, too
clkao audreyt: hey you know all this us flight gel thing right? i assume you have loads. i throw away some shampoo in boston. so just fyi and fly safe. zzz time. 08:39
dduncan audreyt, does your busy world tour by chance include the "Open Source Meets Business" conference in Germany this January; www.heise.de/veranstaltungen/2007/ho_osb/en/ ? 08:41
lambdabot tinyurl.com/yd2a7v
dduncan if so, you may want to check out the plat_forms study going on there
fyi, www.plat-forms.org/
lambdabot Title: Home
dduncan I won't be there, but it might be interesting for you, perhaps
audreyt dduncan: if someone flys me there, maybe, but prolly not 08:42
dduncan what would it cost to fly there for you?
audreyt usd$1k or a bit more 08:44
dduncan I see
Alias__ I've changed my mind about gather/take 08:45
Mostly because I wish I had it right now
dduncan on a separate matter, I notice that the smartlinked Synopsis16 going with my last smoke has POD errors in it; see the end of: m19s28.vlinux.de/iblech/stuff/pugs-...s/S16.html 08:46
lambdabot Title: S16, tinyurl.com/y4bpg7
dduncan not sure if this is the original S16 or the auto-edited one 08:47
dduncan back in the future 08:48
audreyt Alias__: good :) Perl6::Gather on CPAN is the answer :) 08:49
Alias__ I don't like it that much 08:50
:)
I may as well be using Acme:: modules in production
Alias__ &
audreyt Perl6:: is the new Acme
actually it's the old Acme
Patterner it's Acme's New Clothes
08:51 kanru joined
theorbtwo Acme: The next generation. These are the voyages of the crazy channel #perl6, on it's continuation mission, to seek out new features and new circular dependencies. 09:30
TreyHarris we only have one circular dependency: on circular dependencies 09:31
09:54 dduncan left 10:18 andara left, andara joined, andara left 10:33 penk joined 10:52 andara joined 11:14 xinming_ joined 11:25 nipotan is now known as nipotaway 11:26 Psyche^ joined 11:34 Psyche^ is now known as Patterner 11:36 theorbtwo joined 11:41 mako132_ joined 11:42 HiTech69 joined 11:43 mdiep_ joined 12:09 fglock joined 12:18 iblechbot joined 12:23 Limbic_Region joined
Limbic_Region salutations all 12:31
fglock Limbic_Region: hi 12:32
audreyt greetings 12:36
Limbic_Region thought you were on a plane audreyt
audreyt the one good thing about prioritypass lounges is that you don't have to hunt for electricity... 12:37
no, that was checkin; the actual take off is 2hr in the future
audreyt meanwhie repairs pugscc
Limbic_Region do you have an itinerary.txt of where you will be and what you will be doing for the next month? 12:43
audreyt Oct18 arrive at Portland; Oct20 speak at Intel; Oct23 speak at OOPSLA; Oct 24~26 Jifty Hackathon; Oct27 speak at Amazon; Oct29 arrive at Sao Paulo 12:44
Oct30~Nov2 Pugs/v6 hackathon 12:45
Nov3 speak at YAPC::SA
Nov6~10 at Rio; Nov13 back to .tw 12:46
fglock audreyt: our sessions on YAPC::SA are contiguous - this gives us 3h to talk, how about planning something together
audreyt end of itinerary
fglock: sure, we can figure something out when we meet :)
fglock with TimToady too - it would be nice 12:47
audreyt maybe we'll just put a text-to-speech device to #perl6
and then we type ahead into our laptops
maybe too crazy 12:48
audreyt hacks her slides some more -- now at "Dual Core" section with Pugs.Eval at one side and Pugs::Runtime at another 12:49
fglock we could have somebody typing what we are talking, and put #perl6 on the projector :)
audreyt yeah :) 12:52
12:53 Lorn joined, Psyche^ joined
svnbot6 r14414 | audreyt++ | * Pugs.Compile.Pugs: Repair -CPugs, and by extension pugscc 12:57
audreyt ponders, now that libpugs contains libperl, whether to also link in the PAR loader 12:58
it's tiny (<100k)
and gains us multiversioning automagically 12:59
need to think about it some more. 13:00
fglock audreyt: module versioning for ext/ ? 13:02
svnbot6 r14415 | audreyt++ | * pugscc: s/autrijus/audreyt/
nothingmuch @seen cmarcelo 13:03
lambdabot I saw cmarcelo leaving #perl6 1d 7h 24m 59s ago, and .
audreyt fglock: the ability to install multiple versions of, say ,CGI.pm 13:04
and use them in the same program distinguished by long names
and use precompiled binary modules 13:05
etc, etc. see tsee's directory on CPAN, including PAR::Repository, PAR::Dist, etc
nothingmuch starts the shittiest 13:06
MO refactoring of all
making methods have a ->compile method
13:06 cmarcelo joined
nothingmuch cmarcelo: ah, i was just hunting for you 13:06
cmarcelo nothingmuch: pong! :) didn't code much, but finished reading paper about software integration that gaal++ recommended to get some (more) ideas on how translate things.. 13:09
nothingmuch cool
nothingmuch is going to do the last yucky refactoring to MO
13:09 Psyche^ is now known as Patterner
cmarcelo the plugin refac? 13:11
nothingmuch not yet
that's easier from an engineering perspective
giving methods a ->compile method
and clearing up the attachment api
ingy: ping 13:17
fglock nothingmuch: will you create an intermediate language? 13:18
nothingmuch no
MO is orthogonal to the language
audreyt modularity++ 13:19
nothingmuch i will create MO::Emit::Perl5 soon
which should compile an "idealized" MO class into regular perl 5 packages
using hacks like overriding can and wrapping to emulate private methods, etc
audreyt which in Pugs's dual core vision
nothingmuch precisely ;-) 13:20
audreyt will then allow a single declaration to generate embedding native to both runtimes
so we don't have to play AUTOLOAD games anymore
nothingmuch huzzah
audreyt huzzah indeed!
nothingmuch we're so awesome
group hug everyone
fglock :)
audreyt hugs nothingmuch
nothingmuch hugs audreyt 13:21
cmarcelo: ?
fglock: ?
beware of clkao though, i found out some facts about him
fglock hugs all
cmarcelo hugs everyone!
nothingmuch chuck norris type facts, that is
audreyt hugs each(@perl6) 13:22
nothingmuch @everyoneĀ».hug();
lambdabot Unknown command, try @list
[particle] is always hugging himself
fglock nothingmuch: I'm abstracting the base perl5 types in Pugs::Emitter::P6::P5(2)
stevan wonders what he has stumbled into
nothingmuch base types == unboxed?
audreyt hugs stevan whilst he's still wondering 13:23
stevan :)
nothingmuch: robkinyon agreed to give us Sub::Compose 13:24
nothingmuch MUAHAHAHAA
stevan one more thing to slow down Moose's compile time :) 13:25
fglock nothingmuch: compile-time boxing as-needed, FETCH/STORE with typing, and a few other things
nothingmuch fglock: sounds useful
stevan fglock: yes, very useful
fglock so the emitter don't have to worry about basic optimizations
[particle] presented moose at seattle.pm last night
stevan [particle]: cool, how did it go? 13:26
[particle] pretty well.
i used my smartlink impl as an example
cmarcelo [particle]: slides? :) 13:27
[particle] agentzh: ping
cmarcelo: no slides, just code
fglock nothingmuch: these things are inlined, so you don't have the subroutine call overhead, or lexical scope problems 13:28
nothingmuch *nod*
[particle] svn.perl.org/parrot/trunk/lib/SmartLink.pm
lambdabot tinyurl.com/y8uokd
[particle] and svn.perl.org/parrot/trunk/t/tools/smartlinks.t
lambdabot tinyurl.com/y7jw9s
[particle] and of course, the smoke output from pugs 13:29
audreyt at some point those technologies could go back to CPA...
*CPAN
fglock stevan: I think you could use it in Moose, through Module::Compile 13:30
audreyt I mean, as standalone dists, such as search.cpan.org/dist/Test-Chimps/ and search.cpan.org/dist/Task-Smoke/
lambdabot Title: Zev Benjamin / Test-Chimps - search.cpan.org
stevan fglock: yes, this is what I am thinking 13:31
Moose::Autobox, while fun, has a *huge* performance penalty, even more so than regular Moose :)
[particle] moose are slow creatures 13:32
stevan actually once they get moving,.. they are pretty fast,.. long legs and all 13:33
[particle] yeah, then try stopping them...
stevan www.fugly.com/media/IMAGES/Crazy/moose1.jpg 13:34
audreyt Moose.pm: SIGINT ignored
stevan yup
fglock stevan: very simple example of compile-time preprocessing: svn.openfoundry.org/pugs/perl5/Pugs.../01-bool.t
lambdabot tinyurl.com/ym8uzh
tinyurl.com/yapk44
audreyt Moose.pm: SIGSTOP ignored (etc)
stevan audreyt: ;)
Reaction is currently the largest Moose code base,.. and mst says startuptime is still managable 13:35
13:38 idiotmax joined
nothingmuch i so love darcs amend record 13:38
nasty commits are a thing of the past
fglock I've been thinking about the emitter model PCR uses - maybe it could be used for emitting optimized v6.pm code - try several emitters, and use the faster one that succeeds
nothingmuch i haven't had a commit message titled "oops" in a long while
audreyt imagines a GA-based emitter
nothingmuch fglock: talk to stevan about the Sub::Comose future plans
GA?
audreyt genetic algorithm aka spamassassin 13:39
nothingmuch oi oi oi oi oi oi
audreyt "train your emitter"
nothingmuch that's a nice dream
fglock audreyt: yes, like that
nothingmuch but debugging these things.... oh my
audreyt nothingmuch: not anymore crazy than your runtime profiling feedback idea
Khisanth hmm easier to train than spam :)
audreyt that is, both are terminally crazy
nothingmuch heh
audreyt which is the norm around here...
stevan what happens when you emitter evolves to become smarter than you?
13:40 ruoso joined
nothingmuch stevan: skynet 13:40
audreyt: touche
;-)
audreyt then you port yourself to it...
fglock Sub::Comose?
nothingmuch Compose
Sub::Compost;
=)
stevan is suddenly replaced by v6 code
audreyt Stevan The Bootstrapped 13:41
stevan I wonder what that would smell like ... eaww
audreyt interesting. Sub::Compose is called "combine" in pugs codebase... 13:42
fglock looking up Sub::Compose in CPAN
nothingmuch audreyt: yesterday we discussed refactoring it
i want it to be monad inspired
stevan fglock: the one on CPAN is just a toy/experiment
it is too ambitious to be useful
nothingmuch that is, using conjoin is a port of the Maybe monad, etc
so that you can write highlevel compositions
and then inline them
== very good for type checking
audreyt nothingmuch: scala's monad comprehension is a good model for that
stevan and restrict the types of subs we accept as well 13:43
13:43 diotalevi_ joined
nothingmuch where the type subs are very small and hence the overhead is big (calling 10 subs for a param check is =() 13:43
audreyt 1>stevan
audreyt what's this infix 1> operator?
stevan scala has monads?
nothingmuch it's a shell redirect
stevan audreyt: I think he is pipiing your STDOUT to me :)
audreyt stevan: yes, with sugar for comprehension as well
nothingmuch i was wondering if i should redirect stderrr too 13:44
stevan audreyt: cool
audreyt it's the thing Haskell 98 took away from 1.3 and I'm trying to restore it
stevan nothingmuch: nah,.. I have enough issues of my own :)
nothingmuch anymoose
back to my refactor... nearly done!
(attr::compiled::Attached is gone)
13:44 dakkar joined
nothingmuch (compiled == attached per definition now) 13:44
stevan wow,.. audreyt can sleep and code at the same time,.. and nothingmuch can babble and code at the same time
nothingmuch 's idle loop is babbling 13:45
fglock stevan: Set::Infinite, DateTime::Set::*, DateTime::Event::* all use composition
audreyt 's idle loop is coding...
stevan audreyt idle? no? I dont believe it :P 13:46
fglock stevan: so you have a sub body with perl5 code, and known input & output types?
audreyt stevan: in any case, scala's monad comprehension is just their list comprehension 13:47
see scala.epfl.ch/examples/files/simple...reter.html for a demo
lambdabot Title: sources/examples/files/simpleInterpreter.scala, tinyurl.com/zfdh8
audreyt the doc for that is scala.epfl.ch/intro/comprehensions.html
lambdabot Title: The Scala Programming Language: Sequence Comprehensions, tinyurl.com/zcv6o
audreyt "Every datatype that supports the operations filter, map, and flatMap (with the proper types) can be used in sequence comprehensions."
cmarcelo fglock: re PCP: just installing PCP and PCR from Pugs tree and the other deps from CPAN is enough to start hacking? or there are more modules that *must* be from pugs tree? (this answer is a good candidate to be "shifted" in PCP/HACKING file..) 13:48
fglock stevan: this looks a bit like Pugs::Emitter::Perl6::Perl5-2
stevan audreyt: nice :)
fglock: yes, I think it is probably along the same lines
fglock cmarcelo: installing from CPAN is a good start, then update from Pugs tree 13:49
stevan fglock: but keep in mind that we are gutting the current CPAN release and re-writing it to be more "production" useful
13:49 dakrone joined
stevan audreyt: hmm, this reminds me a little of the LISP loop macro 13:50
the compresions that is
fglock stevan: PEPP2 has several implementations for each "sub", depending on want()
stevan fglock: Sub::COmpose will be much simpler than that 13:51
fglock stevan: so it can choose the best implementation
cmarcelo audreyt: why comprehension syntax sugar was took away from Hs? it just works for lists then? I thought that was just more sugar over do-notation...
audreyt cmarcelo: it was. then they had no idea how to generate good enough warnings for the simple case (list comprehension) and then took it away
stevan fglock: the plan is to just deparse subs and then compose them back together with eval 13:52
and avoid the sub call overhead
audreyt I tried to lure edwardk to hack it back in the GHC hackathon but it didn't happen... maybe do it myself some time (or maybe you can do it :))
stevan fglock: and to support various ways of recombining as well (conjoin, disjoin, chain, etc) 13:53
cmarcelo audreyt: I bet on you since i'm still reading the ghc commentary :o)
and, what you mean by "good enough warnings for the simple case"? 13:54
audreyt s/warnings/error messages/
cmarcelo ah, it was hard to identify when list monad was being used and emit good enough msgs.. 13:55
audreyt right
fglock stevan: cool - you can also support more high level things, like atomic 13:56
audreyt but if you use -fmonad-comprehensions then you asked for it
13:56 cjeris joined
audreyt Pugs can certainly use more of that 13:56
fglock stevan: Set::Infinite also supports undo, which is used for backtracking :)
cmarcelo fglock: (i'll try cpan v6 and then upgrade from pugs tree to see if it works smoothly) 13:57
fglock cmarcelo: the v6 in svn currently passes less tests than the CPAN version 13:59
audreyt actually Test.pm's :depends<> handling seems to have broken it wrt v6.pm 14:00
fglock audreyt: I'll take a look, thanks 14:01
cmarcelo i see. but in hackaton context things will happen in svn repo, right? 14:02
stevan fglock: it's possible, my first priority is a good conjoin as that is immediately useful for speeding up Moose type constraints, but the goal is to create a nice framework so that you can eaisly make new combination types
[particle] are there any examples of refactoring from duck-typing to role-based composition? 14:03
fglock cmarcelo: yes, I think we will need to move emitter-2 to misc/Common, and keep the CPAN version in perl5/ for a while 14:07
stevan: btw, I wrote this while trying out binding and typing implementations in plain p5: svn.openfoundry.org/pugs/perl5/Pugs...example.pl 14:09
lambdabot tinyurl.com/tdx9d
fglock x= is overloaded to typed assignment, and = does binding 14:10
nothingmuch 's idle loop is babbling
wtf 14:11
oh, wrong window
up + enter
urffff
fglock stevan: the nice thing is that most operations only have the indirection overhead 14:12
stevan x= ?? 14:16
nothingmuch $foo x= 3; # "$foo$foo$foo"
methinks
fglock stevan: it implements typed STORE 14:17
stevan huh
fglock the type is an object stored in the $var (try running the code)
the object also holds the value, which can be accessed directly (no OO overhead!) 14:18
so STORE is expensive, but FETCH is not 14:19
stevan oh wow,.. 14:20
:)
this is kind fun :)
fglock I'm also using it for boxing Array and Hash, so they can be OO and unboxed at the same time 14:21
stevan yeah
wow,.. this is really cool
I never knew it existed :)
fglock oops - did I invent something? 14:22
audreyt stevan: you should hang around #perl6 more :)
[particle] :)
stevan audreyt: so true :) 14:23
fglock: what does overloading '=' do?
fglock it does binding - both variables share the same cell
stevan oh i see
fglock actually, it shallow-copy the Type object :) 14:24
nothingmuch woot
i can now easily write Method::Accessor, Method::Constructor etc
and thus remove all the hackish closures from MO itself
stevan fglock: cool, so we can actually do autoboxing without having to use autobox :)
14:25 hexmode joined
fglock yes - that's what emitter-2 is about (plus some other cool optimizations) 14:25
audreyt Moose::OutOfTheBox
fglock like, it knows how to deal properly with Pair at p5 level
gaal transgressive antlers! 14:26
fglock and lazy lists
gaal David Bowie should be so proud
stevan fglock: very nice!!!!!!! 14:27
[particle] perl 5 is a large box with many dark corners
fglock stevan: thanks!
stevan ok,.. I hate to go, but I have a deadline,.. but fglock lets talk about this again, very soon 14:28
nothingmuch stevan: the problem converting this to perlland, of course, is that making people type x= is difficult ;-)
fglock++ # nice reuse of an otherwise useless operator 14:29
[particle] yes, and accidental binding will be common
stevan nothingmuch: change is hard :P
nothingmuch yep =/
audreyt I hate to go too but I have a plane :) *wave*
nothingmuch have a nice flight
stevan waves
[particle] waves
gaal moose moose!
fglock you can use v6 :)
Patterner A moose once bit my sister. 14:30
[particle] did she get superpowers?
stevan fglock: I look forward to the day direct usage of Moose is no longer needed and I can just use v6 straight :)
fglock gotta go too - lunch &
stevan: just rewrite Moose to v6?
stevan circularity!
nothingmuch stevan: i need a minion
stevan nothingmuch: me too ;) 14:31
Patterner It wasn't a radioactive moose :(
fglock it will power Pugs & v6.pm :)
14:31 fglock left
gaal moose recognition: www.simplebits.com/notebook/2004/11...ntest.html 14:32
lambdabot Title: SimpleBits | Mystery Photo Contest Thing, tinyurl.com/ylvap8
stevan grumbles some and return to $work work 14:33
14:46 diotalevi_ joined 14:57 osfameron joined
osfameron afternoon 14:57
I'd really like to try feather.perl6.nl:8080/cgi-bin/runpugs?
lambdabot Title: Run Perl 6 Now -- in your browser!
osfameron but silly work firewall dosn't like :8080
is there any chance of having it listen on :80 as well?
Juerd osfameron: No, none. 15:03
osfameron: :80 is strictly static content, for security and maintaintenance reasons. 15:04
This excludes finger.pl, which is practically still static. It doesn't interact with the user.
Juerd wants to make that fully static too one day
15:05 weinig joined
Juerd weinig: Da's niet veel 15:05
15:05 bernhard joined
osfameron oh well 15:06
ta anyway
Juerd osfameron: Can you access port 443 (https)? 15:17
osfameron: Because then you could request a feather account, and then either:
1. connect with ssh, and try pugs locally on the server 15:18
2. set up an ssh tunnel to access port 8080
The https port is used for both https and ssh.
dakkar using the semi-magical proxy by BooK?
Juerd Yep :)
theorbtwo Of course, you could also get the code and throw up a server yourself that does run on :80. 15:19
Juerd I had the same idea long time ago, but never came around to implementing it. I saw his talk about Net::Proxy at YAPC::Europe, and then saw that this was a far superior thing than I would ever have created myself :)
osfameron I could also just run pugs locally :-)
Juerd theorbtwo: But then, that means that you can run pugs locally, which means there's no reason for a web terminal anymore :) 15:20
osfameron but feather seems handy as I'd just get out of sync
Juerd osfameron: You can request an account by sending me an email.
osfameron Juerd: yeah, 443 is good for me (it's how I irc)
theorbtwo True, but running a copy on port 80 means that the next guy wouldn't have that problem.
Juerd osfameron: Include your real name, and your requested username
osfameron Juerd: cool, can you privmsg me your preferred email?
Juerd theorbtwo: If someone has a spare box, sure
[particle] juerd: are you backlogged with account requests?
Juerd theorbtwo: But I prefer not to provide such things
[particle]: Yes, by one account. I read the mail confirming the requested login today. 15:21
15:21 vel joined
[particle] ok, great 15:21
Juerd [particle]: Sending a message with all necessary information in one time can save a few days :)
[particle]: I'm currently more-than overloaded re work
We moved our entire office
[particle] juerd: ouch
Juerd osfameron: No. Is a public message okay with you? [email@hidden.address] 15:22
[particle] juerd: i'll let coke know you're working on it
Juerd [particle]: The ouch factor is that this wasn't planned.
osfameron Juerd: thanks :-)
[particle] Juerd++ 15:23
unplanned move of entire office? yikes!
andara juerd: what do you think of the load caused by runpugs? 15:26
15:27 vel joined
Juerd andara: So far, it has been acceptable. Do you renice the processes? 15:28
Who's using ssh over port 443 right now by the way? 15:29
I notice that sslh uses quite a lot of CPU power :(
read(5, "", 4096) = 0
write(2, "Use of uninitialized value in co"..., 101) = 101
Oh :(
andara juerd: no, how would i do that? 15:30
Juerd andara: man nice
BooK: ping 15:31
BooK: paste for you:
read(5, "", 4096) = 0
write(2, "Use of uninitialized value in concatenation (.) or string at /usr/share/perl5/Ne"..., 101) = 101
select(24, [4 5 6 7 8 9], [5], NULL, NULL) = 1 (in [5])
write(2, "Use of uninitialized value in length at /usr/share/perl5/Net/Proxy.pm line 245.\n", 80) = 80
Juerd restarted sslh
andara Juerd: :-) I can nice them on launch of course. Where whould you like them? 15:32
Juerd andara: I don't understand the question
Just start the process with nice 15:33
nice pugs foo.p6
BooK Juerd: pong
andara Juerd: I meant, any prefered -n value? But if just 'nice' is fine, no bother. 15:34
Juerd BooK: Do you happen to know this bug? I lack tuits to debug
andara: 10 is fine
BooK Juerd: can you push a bug report on rt?
Juerd even if you use 3 of 4, it already allows others to still use the box should runpugs turn into an infinite cpu hog, and all the limitations already on it fail :) 15:35
BooK Net::Proxy sometimes dies without reason (or message) on me
Juerd BooK: I haven't seen it dying yet. But then, I don't use it myself
BooK: Sure
BooK which is why I run it in a while true ; do ... ; done loop :)
thanks
Juerd Done 15:36
I only start it on boot 15:37
15:37 cmarcelo joined
Juerd Oh, no, I do have it in a loop. 15:37
I don't remember setting that up :)
afk
15:43 fglock joined
ingy nothingmuch: pong 15:47
15:53 buetow joined 15:57 Pomin joined 15:59 hexmode joined
rafl audreyt: The debian extralibs thingy is waiting for igloo (the debian ghc maintainer) uploading some extra packages. 16:19
16:31 weinig_ joined 16:39 andara left 16:45 araujo joined 16:50 kanru joined 16:53 thepler joined 16:56 weinig_ is now known as weinig 17:01 justatheory joined
TimToady @tell audreyt MOH is just a rotated HOW. :) 17:11
lambdabot Consider it noted.
gaal also WHO 17:13
[particle] also OHM
gaal also Ī£IO 17:14
TimToady those are more like anagrams than rotations... 17:16
svnbot6 r14416 | fglock++ | v6/emitter - moving the unstable version to pX/Common
gaal WHO is two different kinds of rotation 17:18
Ī£IO is HOTEL-script
TimToady doesn't work well in a sans-serif font... 17:19
17:22 weinig is now known as weinig|bbl 17:30 ofer1 joined 17:31 hexmode joined 17:41 kensanata joined 17:51 beppu joined, Limbic_Region joined 17:52 theorbtwo joined
svnbot6 r14417 | fglock++ | v6 - merged with CPAN version 17:53
r14417 | fglock++ | - devel version of the emitter is now in pX/Common
18:00 weinig|bbl is now known as weinig 18:04 dmq joined 18:29 pmichaud_ is now known as _pmichaud_ 18:30 _pmichaud_ is now known as pmichaud
TimToady hi patrick, how's things? 18:30
gaal any openfoundry-related folks around? 18:31
TimToady seems unlikely, given the time of night 18:32
gaal gotta try my luck, maybe someone's pulling an audrey 18:33
svnbot6 r14418 | fglock++ | v6 - Test.pm works again 18:34
18:37 prism joined
TimToady audrey's not even pulling an audrey... 18:38
it's amazing how quite this channel gets when audrey is in an airplane.
*quiet
gaal The Elk Audreyborous 18:39
fglock hi 18:41
18:46 cdpruden joined 19:10 takanori joined
Khisanth TimToady: only until wireless connections on planes become common :) 19:17
TimToady troonuff 19:19
19:32 mauke joined 19:34 iblechbot joined 20:07 Eimi joined, Daveman joined 20:10 ruoso joined
cj TreyHarris: btw, I was just giving you a hard time about spending all your working hours on this channel 20:15
ashleyb is Perl6::Rules depreciated in favor of Pugs::Compiler::Rule? 20:19
PerlJam heh. 20:20
ashleyb: did you read the docs for Perl6::Rules?
ashleyb I quickly skimmed them
But I didn't find anything that compared them to Pugs::C::R 20:21
so I didn't find any expert confirming my hunch, so that's why I asked.
PerlJam ashleyb: Let me draw your attention to all of the Notes in the docs for Perl6::Rules
ashleyb: they all follow a similar form to this: "due to limitations of the Perl 5 regex engine, the Perl6::Rules module does not currently support this feature" 20:22
ad nauseum
TimToady P6::R was abandoned before PCR was started...
fglock ashleyb: Pugs::Compiler::Rule trades speed for features, except that it doesn't have many features yet 20:25
PerlJam fglock: but you're also not hampered by the direct reliance on the perl5 RE engine 20:26
fglock yes, most limitations of Perl6::Rules don't apply 20:27
ingy hi cj
cj hey there ingy. long time no see
ingy did you go to spug last night?
ashleyb I assume it has the "feature" of implementing all of S05? 20:28
ingy oh btw I got a motorcycle recently
ashleyb (in p5 syntax that is)
ingy do you still have yours?
cj yeppers. Sounds like lots of fun things are happening in the world of perl6. I got commit rights, too, so I'll see if I can't make some mischief. Of course I've got /work/ work to do for a while...
ingy: I gave it to my brother. My wife doesn't want to become a single mom :)
ingy :P 20:29
cj I bet I could borrow it for a jaunt or two, though. Fall is my favorite time for riding
SamB cj: who doesn't have commit access to the pugs repo?
cj SamB: tdbptdpbh
SamB: I feel like I'm special, therefore I'm special.
ingy why does everyone assume that motorcycle == death
PerlJam SamB: There are about 6 billion people who don't have access to the pugs repo (yet)
cj ingy: ask the folks in emergency rooms
ashleyb SamB: no one is going to answer that question. 20:30
SamB: because if they answer it here, they get lassoed with a bit.
20:30 ruoso joined
ashleyb or they already have one. 20:30
ingy I've been riding nearly 20 years with no major incidents
fglock ashleyb: take a look at the tests in t/ - this gives a broad view of what works
ashleyb k
ingy I'm lucky I know. but I'm pretty lucky in general
PerlJam ingy: The logic goes something like "people are stupid, ergo you're going to screw up or someone around you is going to screw up and you're dead" :-) 20:31
ingy: do you wear a helmet?
cj ingy: my incidents haven't been "major" in the sense of needing to spend time with a doctor, but I've fallen off/down quite a few times in ways that could very easily have hurt me badly
ingy I do in now :)
I do now
cj PerlJam: he's wearing one *right now*!!!
PerlJam heh
not *that* kind of helmet!
;-P
cj now I'm curious what type of helmet you think he's wearing... 20:32
ingy I once rode to the Illinois state capitol with 1000's of hogs to protest the helmet law
Illinois is still a free state
PerlJam ingy: There was an accident on a major highway here the other day. An 18-wheeler switched lanes into a car which spun out and hit another car which was knocked into a lady on a motorcycle. Guess who was the only one to get seriously hurt? 20:33
ingy But now I rather like helmets ad leather
fglock ashleyb: take a look also in lib/Pugs/Grammar/Rule2.pm - but to you that syntax you need to install the v6.pm module too
ingy PerlJam: your mom?
;)
cj fglock: you mean "youse"? 20:34
get back to work, cj
fglock you use
ingy I know shit happens. I just want to be on a motorcyle when I die.
fglock for you to use
ingy and I want to die old so I try hard to be safe
ashleyb fglock: I"m still confused 20:35
fglock: what's in Rule2.pm that I need?
fglock ashleyb: I mean, take a look at the source code, to see what kind of stuff is supported
ashleyb ok 20:36
fglock it is compiled with PCR itself, with v6.pm help
SamB prefers to think that everyone has commit access, some of them just don't know it yet
20:36 lanny joined
lanny Hi all. I'm seeing conflicts in perl5/Pugs-Emitter... but haven't modified anything in those directories. Any suggestions? 20:38
PerlJam What's the perl6 version of perl 5's \b (word boundary) ? 20:39
fglock lanny: try rm and svn up
lanny: the directories were moved today
ashleyb is there a vim syntax highlighter for perl6?
lanny And lose my 2 line change on a single file!? Are you crazy??
Ok. Thanks. :)
[particle] pj: <?wb> 20:40
20:41 justatheory joined
fglock lanny: if you commit only the file you changed, it will not check elsewhere 20:41
ci filename
PerlJam [particle]: and are there synonyms for fore and aft boundaries? (ala vi's \> and \< )
kolibrie ashleyb: util/perl6.vim
ashleyb thx
fglock svn ci filename
lanny fglock: Good reminder. I'm used to popping to top of tree, svk update, svk st, then committing.
fglock lanny: yes, that's safer 20:42
lanny Yes. The result of a misguided adulthood as a software developer. :(
[particle] pj: i don't think so 20:43
20:43 larsen joined
[particle] pj: perhaps <after wb> and <before wb> ? 20:43
fglock PerlJam: word>> and <<word
[particle] whoa, is that new? 20:44
i need to test those
PerlJam [particle]++ :-)
fglock the unicode version should work in PCR, the ascii version had problems with M::C
svnbot6 r14419 | lanny++ | * docs/feather/index.html - expand pugscode.org to include irc and smoke 20:45
20:45 vel joined, thepler joined
[particle] hrm, missing = in =item header in S05 20:45
lanny fglock++ Thanks again.
[particle] just before Backslash reform
fglock home & 20:47
20:48 fglock left 20:53 justatheory_ joined
ingy seen audreyt 20:56
jabbot ingy: audreyt was seen 6 hours 26 minutes 36 seconds ago
TimToady note however, that the definition of word boundary there was wrong up until about a minute ago... 20:58
20:58 explorer joined 21:02 seano joined
seano gaal: ping? 21:03
21:06 Limbic_Region joined, SubStack joined
[particle] timtoady: did you see the improper formatting? 21:09
-item * 21:11
+=item *
S05:1014
21:15 weinig is now known as weinig|bbl
seano any good reason for having ":e foo" always run in an empty environment? 21:25
seems to me you can always just use ":r" if you want that... 21:26
(alternatively, is there some way to print the haskell-form of a value besides :e?)
theorbtwo Hm, I thought that was fixed some time ago... 21:28
seano not here (== CVS, er svn, as of this morning) See src/Pugs.hs:336 21:31
and I don't think "fixed" is the word, as someone went to some trouble to add a flag controlling this.
21:32 dduncan joined 21:48 Franky joined 21:59 Franky left 22:00 HiTech69 joined 22:05 weinig|bbl is now known as weinig 22:20 SubStack joined 22:28 mako132_ joined 22:29 nperez joined 22:44 nothingmuch_ joined
gaal seano: pong 22:49
seano: what does svn annotate say about it? 22:50
seano: I'm just about to write the openfoundry folks, unless you have already :)
seano hang on a sec, let me check..
gaal ingy: audreyt's conti-commuting 22:51
seano what would you hope/expect it to say?
22:52 cmarcelo joined
seano mostly just says "audrey was here", and doesn't list the lines I've changed. 22:53
Okay, I'll mail the openfoundry folks.
gaal seano: would help you find the commit revision, so you can track down the ci log and see if it comes with a rationale.
seano oh, other question.
ok.
gaal seano: ok, I'm about to turn in for tonight. If you get a response from them (might happen in a few hours) ask here to get reinvited 22:55
there are plenty of people who can send you a commit bit :)
seano okay, will do. thanks.
gaal good night :) zzZ& 22:56
ingy gaal: thanks 23:07
where is she headed?
oops. ENOGAAL
23:08 justatheory joined
Limbic_Region seano - I can get you a commit bit 23:09
just need an email address
seano sure -- my name at cs dot ucsd dot edu 23:12
(er name == nick) 23:13
Limbic_Region give me a minute 23:14
commit bit on its way 23:16
typically first commit is adding yourself to the AUTHORS file 23:17
don't be afraid to commit something to the wrong place if there is no one around to give you advice
commit it and it will get taken care of
seano what about features you find personally useful? 23:18
Limbic_Region seano - pugs features, perl 6 features, or modules that you think would make pugs more fun or what?
seano e.g. I don't want the pugs shell to print results for evaluating things that end with a semicolon (think matlab).
things to make development life more livable for the moment. 23:19
Limbic_Region *shrug* 23:20
I dunno as I never use pugs shell (I write a file as I do with Perl)
I would guess you would want to discuss that with the channel though
seano freaking masochists...
seano mutters that, that is...
Limbic_Region is definately a masochist 23:21
seano k -- hey channel: anyone object if the shell doesn't output the evaluation result on things ending in a semicolon?
[particle] freak dances
seano e.g. you can do "@x = something_really_big();" 23:22
instead of "@x = srb(); 1"
Limbic_Region seano - I don't think it currently prints anything out without print or say
ooh, it does
hrm
seano yup...
[particle] pugs> my @a= (1..10); 23:23
my @a= (1..10);
(1, 2, 3, 4, 5, 6, 7, 8, 9, 10)
pugs>
Limbic_Region seano - I would suggest making it configurable
[particle] ah, too late :)
how about a flag, like use/no 'echo';
Limbic_Region right
seano yeah. $?STFU 23:24
er, $*STFU that would be.
Limbic_Region: no dice on the invite -- I already have an account, so it doesn't seem to do anything. 23:25
Limbic_Region: I sent them an email, so I can just wait.
Limbic_Region ahh 23:26
well, it could have been eaten as spam
I don't hand out commit bits very often but almost inevitably I end up using a second email address for the requester because the 1st never shows up
[particle] seano: the best way forward is probably probably to write a pragma... that's what you'd do in perl 5, no?
seano [particle]: no, this is only relevant in the shell. so a commandline option and/or rc file 23:27
23:28 Alias_ joined
seano [particle]: well, I guess you *could* use a pragma, but I don't think pugs has those yet. 23:28
Limbic_Region seano - what is you current OpenFoundry acct name?
seano seano :)
Limbic_Region ok - going to add you that way 23:29
seano - done 23:30
you should just be able to use your current acct/pw
seano Limbic_Region: sweet -- works.
Limbic_Region takes a bow 23:31
23:33 mako132_ joined
Limbic_Region personally regarding the pugs shell 23:33
I was thinking a :flag
like :h = help
:q = quiet would have been nice, but :q is taken for quit 23:34
[particle] :s for shhhh
Limbic_Region :s = silent mode perhaps
svnbot6 r14420 | seano++ | added self
Limbic_Region high fives [particle]
seano yeah, that makes sense. maybe traipse down the slippery slope and have :s ARG..., short for "stty" ;) 23:35
svnbot6 r14421 | seano++ | * Turn of "strict 'vars'" for -e scripts via the non-standard variable 23:37
r14421 | seano++ | $?STRICT. This will go away with lexical pragmas. (Note: you can
r14421 | seano++ | also de-strict your pugs interactive experience with "$?STRICT=0".)
23:39 dmq joined
Limbic_Region salutations demerphq 23:41
ok, well welcome aboard seano 23:42
Limbic_Region is calling it a night
TTFN all
23:43 mdiep joined 23:53 mako132_ joined

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!