-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.
rep why? 00:01
Juerd_ It spasm
spams too
This is why I hate the oplessness of freenode channels. 00:02
It doesn't encourage spreading op-rights to cover multiple time zones
Khisanth well it usually isn't truely opless, just appears that way
Juerd_ And when something happens, nobody knows who is able to handle it.
Khisanth: I know that. 00:03
But there are very practical reasons to have ops carry a visible mark.
I know that these advantages do come with disadvantages
But having been an op in several channels for almost a decade, I have only experienced such denial of service attacks at most five times. 00:04
So I don't really see how this single pro can outweigh the huge cons
Khisanth that pro is irrelevant with cloaks anyway 00:06
Juerd_ That's true too
Then I see no remaining reasons to hide opness
afk # z 00:13
00:59 knewt_ is now known as knewt
buu Juerd_: Er, my understand behind the op hiding is that is discourages users from attempting to provoke ops, treating them differently, and so on 01:29
Khisanth buu: you forgot cower in fear 02:03
buu Khisanth: Or that, I suppose. 02:07
04:17 SamB is now known as Melchior, Melchior is now known as SamB
geoffb Good $time, nothingmuch 07:05
nothingmuch hola 07:07
que pasa? 07:08
geoffb hacking on gjbforth
and you?
geoffb smacks his virtual forehead: split and splice are *not* the same, duh ... 07:12
nothingmuch err, sorry, didn't notice activity 07:15
lots of catalyst work lately
rewrote session and authentication stuff 07:16
making a movie
geoffb So I've seen . . . congrats
geoffb mumbles something about "cursed DO LOOPs" and wanders off to bed 08:08
08:19 Aankh|Clone is now known as Aankhen``
Juerd_ buu: Exactly treating ops differently is a good thing, not a bad thing. 08:47
buu: And provoking ops is a problem that solves itself, because the op gets angy and rids the channel of the provoker. 08:48
integral suggests that the lack of ops is not a result of the hiding policy, but rather 08:49
that people who only have op powers can't give more people op powers
Juerd_ It's a combination. 08:50
If someone wants to complain about the channel, they should talk to ops. An often heard complaint is that there are no ops in a certain time span. This is then fixed by spreading +o bits. 08:51
However, both the complaining and the raising of power are impossible here. 08:52
Complaining because the average visitor has no idea who to turn to.
integral hrm, I don't really understand complaints
Juerd_ And in fact, even analysis of the problem is impossible, because one doesn't *know* if there are any ops around in a certain time.
integral actually it's totally false that you can't get a list of the ops. 08:53
Juerd_ Enlighten me
integral just ask chanserv for the access list
Khisanth autrijus, hcchien and obra :)
integral who all have the power to make more ops :) So looks like it's their fault... 08:54
Juerd_ So
If this is possible, then why the fuck is it hidden in IRC-land?
integral because this isn't IRC-land! That's why it's all run with services, rather than traditional IRC mechanisms 08:55
Juerd_ I hate freenode.
Khisanth :)
integral this isn't a traditional IRC network, blah blah, it's your fault for being too traditional, blah blah, lilo is God, blah blah :) </summary>
Juerd_ fscking morons.
GeJ what a weird name for a mount point... 08:56
Juerd_ Lilo should be keelhauled the very next time he sends a global message that wasn't necessary.
buu hahaha 09:09
I agree, every time lilo sends an unecessary global message, at least 3 electrons lose their lives. 09:10
integral: (just a note, it's impossible to get a list of ops in, say, #perl, for example)
integral I know :) I was just hoping everyone else didn't ;) 09:11
buu Heh
gaal is $=POD specced (or however it's spelled)? 09:17
Khisanth seems to be the only one like that 09:22
r0nny moin 09:24
yo 09:41
i got a problem with the metod dispatcher
for some reason it puts itdelf into a endless loop, where it should call a method of a class attribute, wich got a method with the same name 09:42
anyone got an idea how to deal with this ? 09:47
amyome ? 10:22
does pug have a way, to ask for the full path of the actual script 11:16
scook0_ r0nny: not that I'm aware ... I assume $?FILE is not good enough for your needs 11:29
(you could theoretically append $?FILE to pugs' initial working-directory, as a hacky solution) 11:31
11:31 scook0_ is now known as scook0
scook0 r0nny: btw, are you still having dispatch troubles? 11:32
r0nny re 11:33
scook0: yeah
atm i tryed to solve them with method renaming
scook0 have you checked the parse tree, to make sure it's parsing correctly?
(I ran into parsing-related method-call problems a while back...might still be some bugs) 11:34
also, AFAIK pugs just uses the relative file path internally (if one is given), so you would need to reconstruct the full path on the inside anyway 11:37
r0nny hm 11:38
btw - is it possible, to make parrot run haskell code ?
scook0 r0nny: well, I imagine it should be possible to make Haskell functions callable from C using FFI, then use whatever FFI Parrot has to call these functions 11:46
but I'm not an expert on any of these topics :|
r0nny it might be usefull to have pugs itself run in parrot, and then slowly migrate from the haskell implementation to the perl6 implementation 11:47
btw - does parrot support c++ bindings ? 11:49
scook0 I don't know ... you could always try asking #parrot 11:51
SamB r0nny: you could do what Haskell users do 11:54
and write C-callable wrappers
scook0 I think other people have had similar ideas about migration -- what with the Prelude being written in P6 and all
r0nny SamB: writing c-callable wrappers for komplex stuff is HUGE paine
SamB r0nny: which is why you autogenerate them 11:55
r0nny there are generators ? 11:56
scook0 although interfacing with c++ directly can be painful too
r0nny scook0: jep - there is written huge paine in the inside of name manglers, and uncatched exceptions 11:57
scook0 because those things aren't as stable/portable as plain C calling conventions, right? 11:58
r0nny they are not as portable
12:03 bradb_ is now known as bradb
pmurias there has been some talk about swig and parrot on perl6-internals in the past 12:34
fglock writes a p6 parser in p5 12:36
just for fun
pmurias fglock:What tools are you using? 12:38
fglock pmurias: it's a barebones thing - just plain perl 12:40
I'm looking for examples of the difficult parts - will look at t/ 12:42
r0nny re 12:43
pmurias re == regular expresions ? 12:44
r0nny pmurias: re just means im back 12:51
GeJ pmurias: regular expressions do not exist. They never existed. Rules is what you're looking for.</jedimindtrick> 13:06
nnunley r0nny: Haven't gotten the rest of the story, but I believe there's a PMC in parrot called NCI (native call interface) 13:08
pmurias GeJ: is there a two letter [acronim?] for rules, like rx is for 'regular' expressions 13:14
scook0 pmurias: well, I believe the syntax for a rule literal is `rx/.../` 13:24
(see the relevant Synopsis -- the expansion is 'rule expression' IIRC) 13:25
fglock I have a working parser - I wonder if it belongs to /misc (there is a message there about moving things to freepan) 13:49
autrijus fglock: that message is gone 14:38
it does belong in misc -- extra points for a new ext/ README
er I mean a misc/README 14:39
(just committed a ext/README)
autrijus finally finished the three $job assignments... plane to .au is tomorrow
svnbot6 r8026 | autrijus++ | * Throw away throw_away/ now parrot/t/compilers/pge/p5regexp is there. 14:41
r8027 | autrijus++ | * Remove the no-longer-relevant misc/PORTING_HOWTO.
r8028 | autrijus++ | * Moves misc/README to ext/ and reword it significantly;
r8028 | autrijus++ | misc/ will need its own README too...
fglock autrijus: thanks
autrijus g'nite for now... tomorrow would be the massive backlogging day :)
fglock: np :)
rafl Juerd: ping 14:45
ingy_: ping 14:52
svnbot6 r8029 | kane++ | * add open issues 15:10
r8030 | kane++ | * damn you rangerrick 15:16
Juerd rafl: pong 15:23
rafl Juerd: I will do a perl 6 hacking workshop on GPW. Can we use feather as the hack box? 15:24
Juerd rafl: Of course
rafl: Just realise it won't handle more than three compiles at the same time :)
Normally, timezones fix this ;)
rafl Juerd: May I add accounts for the participants during the workshop? I'll remove them afterwards.
Juerd: We want to use pugs, not compile it.
We want to do perl 6 in practice. Port modules of the participants to p6, etc. 15:25
Juerd rafl: Under the usual terms: I need realname and mail address
rafl: And they can of course keep them after the GPW if you want.
rafl: If everything goes according to plan, I'll be at th egpw myself, and perhaps I can do a short talk about feather before your workshop 15:26
kane_ waves at vasi
Juerd Then they know what they're using, and are hopefully already motivated to request an account
vasi hi again kane :-)
kane_ w00t, my first stalker :)
rafl Juerd: OK.
Hello kane_, vasi.
vasi heh
Juerd rafl: Are you a GPW organizer?
rafl: Or should I mail someone? 15:27
kane_ vasi: more p6 people here, and by proxy sixpan folks (like rafl)
rafl Juerd: I organize it.
Juerd rafl: Then hereby I submit this talk idea ;)
rafl Juerd: Hm..
Juerd: Please do that at the beginning of my workshop.
Juerd Whatever fits in the schedule. I'm very flexible :) 15:28
rafl Juerd: Well, it's a bit late for new talk submissions, btw.
kane_ vasi: adding all your rantings to the S22 draft now :)
Juerd I think at the beginning of your workshop is a good time, because then they can *immediately* put it to use :)
vasi kane, so yeah anyhow i'm not saying the naive approach is a bad start
just that you have to be very very careful that you don't get stuck with it
rafl Juerd: And they need it for the workshop. We don't have a local server that can be used as hack box.
kane_ vasi: our current dependendy engine (cough) already takes roughtly the arguments it should for the smart version too 15:29
rafl I doubt that binary modules will be used very much, btw.
vasi kane, i have a module as part of Fink that does a bit of the basic "here's a system state, what's wrong with it?" work
rafl It's a good thing, though.
kane_ vasi: point me at the code? 15:30
vasi yup, doing that
Juerd I need a power outlet.
Juerd wants power-over-wifi :)
vasi it's not well tested, and doesn't do much resolution, just detects problems
and it's dpkg-specific 15:31
kane_ vasi: it's the thought that counts :)
rafl Juerd: I order 10 of them if you've developed it.
vasi kane, well i built it just to play around with different dep engine ideas
so i could have an algorithm do some operations, see what broke
Juerd rafl: You're on the list now ;)
rafl Juerd: On the top of it, I hope! :-) 15:32
svnbot6 r8031 | kane++ | * add vasi's suggestions about the dep engine
kane_ vasI; a sane start surely...
vasi the only actual resolution is in 'resolve_install', to work around a dpkg bug
kane_ oh god, you are watching me!
Juerd rafl: I use *IRO
Any way in, random out.
vasi it's complicated compared to the naive approach, but simplistic compared to a real dep engine :-) 15:33
kane, i also highly recommend you take a look at the code for smart.pm
whatever :-)
kane_ the python one, right?
Juerd rafl: "I've added you to the set" doesn't sound wel :)
vasi it's in python, and it's badly documented
rafl Juerd: What did Sven say, btw?
vasi but it's still worth looking at
kane_ vasi: tbh, i'm happy with a minimal approach that Works (tm) and have teh people with clue fixing up the parts Real Good
vasi i think it was transaction.py or policy.py that had the basic stuff
Juerd rafl: "yep... you're already on my list!"
vasi kane, yeah just something to be aware of as you move forwards 15:34
rafl Juerd: Great. Any news from your client?
kane_ vasi: *nods* adding all this stuff to the pod, as you'll seen in a few mins :)
Juerd rafl: And he mentioned traveling by train, but I'm not too comfortablewith that, because I don't like trains
rafl Juerd: You can hack while driving. That's hard by car. 15:35
vasi yeah, so all i'd do at this point is have a _dep.pl with an opaque API, to make sure you don't introduce ugly side effects
kane_ vasi: anything else in the prototype that made youw onder? :)
Juerd rafl: No - I sent the first reminder yesterday. Next one tuesday, phonecall friday.
vasi kane, looking
Juerd rafl: I know it's practical to move by train -- except for baggage (I should not carry heavy stuff) 15:36
rafl Juerd: OK. Let's hope you get your money until chrismas.
Juerd It'd be nice to have it before christmas. Otherwise, no presents for family.
rafl Juerd: Well, a laptop and coffein is enough for *C3.
kane_ damn, time's flying
4:30 already
Juerd rafl: air mattress
rafl Juerd: A sleeping bag is enough. But with enough coffein you won't need that. 15:37
svnbot6 r8032 | kane++ | * add vasi's comments about a 'smart' dependency engine
vasi where's the alternatives code at?
kane_ hardcoded in install.pl/uninstall.pl 15:38
this is 'perl-style', not debstyle btw
Juerd rafl: In any case, there are still other reasons for avoiding the train - some of which I won't mention, because they're overly silly.
rafl Come on.. entertain us with silly reasons..
Juerd Heh 15:40
vasi it doesn't really deal with changing alternatives or anything yet
kane_ vasi: it does, 'perl-style'
see =head1 Alternatives in the pod file
vasi uh, unless i'm missing something, install.pl is just unconditionally linking them? 15:41
kane_ see install.pl line 95 and 105 15:42
and 117 -- which will cancel linking if needed
vasi yeah, my brain must be damaged today, yeesh 15:43
rafl kane_: I'll start with porting the current prototype to something more modular and pluggable now, OK?
kane_ ralf: sure.. i'm not fast enough today i think :)
i'll code on a meta.inf -> perl-object thing onw.. that's fast and easy 15:44
rafl s/ralf/rafl/ :-)
kane_ flrarghn!
rafl "meta.inf -> perl-object" -v
kane_ oh, no ugly hashes that can't be verified, but a decent object that can be extendend and queried
rafl OK 15:45
vasi hmm...so if a package isn't 'preferred' at install time, it just doesn't register alternatives at all?
kane_ exactly 15:46
Juerd Wha - traveling to Berlin by train (and back) costs EUR 167.80
rafl Wow..
vasi kane, that doesn't sound right, unless i'm confused as to what alts are supposed to be about 15:47
kane_ vasi: there should probably be an 'update alternatives' script that can do this Later One
vasi seems very assymetrical
kane_ vasi: alts are different from debian alts.. did you check the pod?
svnbot6 r8033 | fglock++ | * misc/Parser-Mini - experimental 'thin' p6 parser
r8034 | kane++ | * add dir structure
r8034 | kane++ | * touch Meta.pm file
vasi kane, i see that the rules are different 15:48
ie: the flow chart
but i'm not sure WHY
kane_ vasi: we allow multiple versions of a package to be installed
they both will probably come with teh same script, at least the same manpage
vasi yeah, that's ok
kane_ but only one can be the 'official' version to the OS
vasi but shouldn't they all at least be registered as providers of the alt? 15:49
kane_ so we link that one to /usr/bin and /usr/man
vasi and then the policy engine chooses from all the providers
kane_ ### XXX this should be a policy test!
i think that's what hte comment says :)
vasi oh i see why we're confused, your idea of 'registering' an alternative is very different from mine 15:50
for this system, alts are ONLY for different versions of the same package
kane_ who was it again taht said (overloaded terminology)++ ? :)
vasi ok, i thought that was IN ADDITION
kane_ which means we probably need another name
rafl kane_: What will your META.pm do? I currently want to write JIB::Config which also deals with META info a bit.
kane_ rafl: translates the contents of META.info to a queryable object 15:51
rather than a stupid hash
rafl kane_: OK. You will need the JIB::Config for that, won't you?
kane_ rafl: don't think so.. but maybe i misunderstand your purpose 15:52
vasi kane, alright looks pretty good otherwise, at least to my cursory inspection :-)
kane_ vasi: thanks for your stamp of approval :)
vasi: and if you want a comitter bit for the dep engine... ;)
rafl kane_: Do you deal only with META.info contents or also with opening the files, etc? 15:53
vasi heh, got enough work on my plate at the moment :-/
kane_ rafl: opening the file too.. given file location of META.info, feed it to yaml, parse, evaluate, sanity check, etc
vasi kane, is the next version (modularized, what you're working on now) intended for use? 15:54
kane_ vasi: yes
rafl kane_: OK, so you won't need the META.info filename. No need for JIB::Config first.
vasi kane, great
kane_ rafl: i think so too.. basic parts first -- we need a config too obviously :)
rafl So I'll do JIB::Config now.
vasi i eagerly await the chance to break it :-)
kane_ vasi: eh.. thanks.. i guess ;) 15:55
rafl: i'm adding basic files like JIB.pm for $version and a makefile.pl too 15:57
rafl kane_: Hm. The current config does a lot wrt dep handling/package name parsing, etc. I think that should be JIB::Package.
kane_: OK.
kane_: My config will only export some fixed values. The rest will be in another package. 15:58
kane_ sec, phone 16:00
rafl kane_: What's files.list for? 16:05
lollan hello 16:10
rafl Hello lollan
lollan :)
svnbot6 r8035 | rafl++ | r22573@ata: rafl | 2005-12-01 17:10:06 +0100 16:11
r8035 | rafl++ | * Added a first JIB::Config version.
lollan so the new version of perl is out !
i did'nt know since know
svnbot6 r8036 | rafl++ | r22574@ata: rafl | 2005-12-01 17:10:47 +0100
r8036 | rafl++ | * Fixed a syntax error
lollan !tutorial
rafl lollan: Actually perl 6 isn't out yet. 16:13
PerlJam lollan: The current implementation of perl6 is called pugs and it's mostly a proof-of-concepts implementation in haskell. 16:14
lollan ok
PerlJam lollan: pugs is not *the* perl6 compiler
lollan ok
rafl But it's a nice toy, though. 16:15
lollan thanks everibody
PerlJam but if you want to play with perl6 stuff now, pugs is it.
lollan ok
what is a haskell ?
rafl A pure functional programming language.
lollan ok
Limbic_Region perlbot haskell 16:16
perlbot www.haskell.org/ - a purely Functional, Strongy Typed Programming Language that is used for Pugs - the Perl 6 Compiler
lollan thanks 16:17
kane_ rafl: sorry, long phone call -- should be documented in _inc.pl i think it's the list of installed files 16:29
$Fileslist = 'files.list'; # .packlist equiv
svnbot6 r8037 | rafl++ | r22577@ata: rafl | 2005-12-01 17:45:28 +0100 16:46
r8037 | rafl++ | * Added a first JIB::Package prototype.
r8038 | kane++ | * add first JIB::Meta 16:50
r8038 | kane++ | * add dummy meta file
r8038 | kane++ | * add basic tests
kane_ rafl: mind if i simplify package.pm a little bit? 16:51
rafl kane_: Go ahead. I actually haven't done much perl coding in the last months. I hope to get used to it again soon. 16:53
kane_ rafl: np -- it's good we get code done... if we can see ways to make it easier, that's secondary ;)
btw, i'm adding some vim/emacs directives at the bottom that make your <tab> expand to 4 spaces.. that would make my editor a lot happier ;) 16:54
svnbot6 r8039 | rafl++ | r22579@ata: rafl | 2005-12-01 17:52:00 +0100 16:55
r8039 | rafl++ | * Added ability to export all the config vars in JIB::Config.
r8040 | kane++ | * add editor directives 17:01
r8041 | kane++ | * make JIB::Package use object::accessor 17:13
r8041 | kane++ | * steal regex implementation from CPANPLUS::Module
r8041 | kane++ | * add basic tests
kane_ rafl: what are you coding on now? 17:14
rafl kane_: I coding my evening meal.
kane_ rafl: heh ;)
ralf: how about we make the config OO too? saves variable exportation 17:15
rafl Yes. A singleton?
kane_ i think so 17:16
rafl OK, I'll do that.
kane_ k
and speaking of evening meal, i should do that too 17:17
bbl &
svnbot6 r8042 | kane++ | * metafile is called 'meta.info' not 'info.info' ;) 17:19
pmurias perlbot help 17:25
perlbot Syntax: (fact) :: tell (who) about (what) :: (who) > (what) :: learn (what) as (info) :: relearn (fact) as (info) :: phone (phone number) :: shorten (url) :: shorten it :: search (module) :: docs (module) :: perldoc -f (function) :: jargon (term) :: math :: fortune :: flip :: host (type) (record) ::rot13 :: roll (die) :: tempconv (temp) :: scramble (text) :: 8ball :: slap :: diss :: what time is it :: top/bottom (number) karma :: geoip (ip)
dada geoip 17:35
perlbot: geoip
perlbot Record for Italy
dada wow, it works
svnbot6 r8043 | fglock++ | * misc/Parser-Mini - can parse '/examples/algorithms/attrgrammar.p6'; 17:38
r8043 | fglock++ | - emits a tree using Data::Dumper
r8044 | kane++ | * make JIB::Package use params::check and Object::Accessor 17:44
r8044 | kane++ | * update tests
ingy_ rafl: pong 17:45
rafl ingy_: I'm the new Debian maintainer for YAML.pm. Please take a look at this: bugs.debian.org/cgi-bin/bugreport.cgi?bug=305526 17:46
ingy_ rafl: awesome... 17:55
I'm pretty far into a new release
svnbot6 r8045 | kane++ | * update JIB::Package docs 18:02
r8045 | kane++ | * update tests for ->package allow handler
r8046 | rafl++ | r22589@ata: rafl | 2005-12-01 19:02:02 +0100
r8046 | rafl++ | JIB::Config is a singleton now.
rafl leaves till this evening. Bye!
fglock research idea for misc/Parser-Mini - plug it into the perl5 runtime and see if it works 18:35
stevan fglock++ # nice stuff 18:45
fglock thanks stevan! 18:51
I'm reading about rules and macros - let's see to what is possible to implement 18:53
s/to// 18:54
stevan nice
fglock macros are basically subs that execute at compile time (that's easier than I thought) 18:58
stevan fglock: yeah,.. but they operate on the AST 19:00
macros take an AST as an arg, and return an AST 19:01
at least LISP style macros do
fglock no problem, my AST is perl data - not unlike what PIL-RUN executes 19:02
stevan cool
stevan curses $work and returns to writing HTML templates :( & 19:05
fglock now trying to find out how to plug Rules into this - maybe using Inline::Parrot and PGE
bye stevan
Aankhen`` goes to sleep. 19:29
fglock needs a lightweight metamodel to make tests 19:32
actually, I need a p6 metamodel written in p6 19:37
19:42 _SamB_ is now known as SamB
svnbot6 r8047 | fglock++ | * misc/Parser-Mini/parser.p6 - p6 version (doesn't compile yet); 20:02
r8047 | fglock++ | - first attemp to make a parser that parses itself
fglock oops - no Data::Dumper in p6 yet? 20:15
wolverian fglock, that isn't very perl6-ish, btw. 20:16