The Return of the Journal : pugs.blogs.com/ | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by GammaRay on 31 December 2005.
00:02 khaladan joined 00:03 borkened joined 00:48 daja77_ joined
svnbot6 r8637 | Darren_Duncan++ | r1897@Darren-Duncans-Computer: darrenduncan | 2006-01-10 16:59:46 -0800 01:04
r8637 | Darren_Duncan++ | /ext/Rosetta : move/renamed 4 files from /docs to /lib/Rosetta, as they are or will be POD files; removed 'srt' from 3 .pm files; deleted the README
01:11 stennie joined
dduncan regretfully, it seems that svk doesn't propagate "moves" or "copies" as such when a client is setup in the mirror+local duality, and changes are made against the local ... while the right thing happens locally, svk push just seems to turn them into ordinary adds and/or deletes against the main repository, so those lose history information being attached to the moved/copied files 01:11
svnbot6 r8638 | Darren_Duncan++ | r1901@Darren-Duncans-Computer: darrenduncan | 2006-01-10 17:09:11 -0800
r8638 | Darren_Duncan++ | /ext/Rosetta : oops, didn't actually delete README last time
01:15 phredmoyer joined, phredmoyer left 01:17 Amnesiac joined 01:23 stennie joined 01:26 Cryptic_K joined 01:39 luqui joined 01:42 Shachaf joined 01:50 lisppaste3 joined 02:19 alexe_home joined 02:41 verve_ joined 02:42 xah_ joined 02:43 venk joined 02:46 hcart1 joined
svnbot6 r8639 | Darren_Duncan++ | r1905@Darren-Duncans-Computer: darrenduncan | 2006-01-10 18:48:04 -0800 02:51
r8639 | Darren_Duncan++ | /ext/Rosetta : updated all 6 pod-only files for consistent formatting and some updated content
03:03 Amnesiac joined
Debolaz Hey Amnesiac. 03:08
Amnesiac hey Debolaz, how's everything? 03:12
03:13 mjl69 joined
Debolaz Amnesiac: Good. Just started writing my own software license. 03:19
dduncan something everyone needs to do at some point 03:43
for what type of software?
Debolaz dduncan: Just a game I'm writing. It's not really something serious, just a learning project. 03:46
dduncan I hear U 03:47
03:54 drbean joined 04:11 stevan_ joined 04:15 Cryptic_K joined 04:24 xah joined 04:25 K__ joined 04:42 hcart1 left 05:35 r0nny joined 06:17 flw joined
svnbot6 r8640 | gaal++ | * improve error message for $x := $undeclared 06:28
06:31 Cryptic_K joined
GeJ greetings all 06:33
gaal hey GeJ 06:36
flw hi GeJ 06:37
GeJ hum, sweet, YAML, YAML::Syck and JSON::Syck in the ports. 06:42
Damn, it's nice to have a couple of commiters hanging around over here :) 06:43
06:49 greister joined
gaal I must go off to $work. I made a checkin of my further attempts with some comments in the hope that someone may pick it up and figure out what went wrong. 06:52
svnbot6 r8641 | gaal++ | mulitsub export: document a few unsuccessful attempts
gaal have a nice day :) & 06:53
06:59 nothingmuch joined
GeJ morning nm 07:00
nothingmuch morning 07:01
what's cooking?
07:03 binary42 joined
GeJ not much. no headache this morning. 07:07
07:11 binary42 joined
nothingmuch sorry, someone bugged me 07:22
headaches are bad =(
\are you sick? or hung over, etc/
07:27 ycheng joined 07:30 lisppaste3 joined 07:48 DesreveR joined
buu ?eval 1 07:48
07:49 joepurl_ joined 08:04 iblechbot joined
dduncan fyi, I have some module v-num changes to commit tomorrow, to get in before 6.2.11 goes out, but otherwise I'll hold off any further changes until after the 6.2.11 release 08:23
which is the weekend, I've heard
08:26 K__ joined 08:47 nnunley joined 08:55 dolmans joined 09:02 ilogger2 joined 09:06 kane_ joined, amv joined, chip joined, qmole joined, cookys joined, PerlJam joined, Juerd joined, pjcj joined 10:06 bsb joined 10:43 dmo joined 10:46 PJF joined 10:50 G2 joined 10:57 elmex joined 10:58 Odin- joined 11:12 stennie joined 11:17 Odin- joined 11:22 adrian_ joined, adrian_ left 11:26 dada joined 11:31 SamB joined 11:34 pdcawley joined 12:06 chris2 joined 12:50 abram joined
bra is now known as quuxman 12:51
13:00 nnunley joined, iblechbot joined 13:03 stennie left 13:15 kolibrie joined 13:25 lisppaste3 joined 13:36 nothingmuch joined 13:38 bsb joined
esreve is now known as r0nny 14:00
wolverian audreyt++ # the D&R is hilarious :) 14:04
14:06 stevan_ joined 14:19 Rolan joined 14:26 feng123 joined
tevan is now known as stevan 14:26
14:35 Qiang joined 14:39 vel__ joined 14:52 xern_ joined 15:02 eric256_ joined 15:05 eric256_ left, nothingmuch joined 15:24 Eimi joined 15:30 rantanplan_ joined, lisppaste3 joined 15:31 bernhard joined
gaal wolverian: D&R? 15:40
wolverian gaal, search.cpan.org/src/AUTRIJUS/YAML-S...21/COPYING 15:41
15:42 amnesiac joined
gaal considered "This software might be used" as a license. 15:43
mnesia is now known as Amnesiac 15:44
15:53 nomeata joined 15:58 hexmode joined 16:08 nothingmuch joined
ingy seen audreyt 16:20
jabbot ingy: audreyt was seen 1 days 3 hours 56 minutes 20 seconds ago
16:30 nomeata left 16:32 Nei joined 16:36 Pc03 joined 16:40 Pc03 left 16:44 chris2 joined
clkao win 20 16:45
Nei xD 16:47
keybindings not your thing?
webmind sometimes finds it easier to type /win 34, then to do alt-b 16:48
but I'm not a 10finger typer eiher
Nei interesting
after all you need 5 or 6 more keys for the command 16:49
any idea why it's easier?
webmind I think because I find it harder to locate the key with the meta+<key> 16:54
it requires an extra mental step to convert 34 to b 16:55
I have it mainlt on the higher end the 30+ channels since I visit those less often 16:56
and I see a highlight on 37 and wonder now how do I get there again
17:01 marmic joined
Nei yea I have the same problem 17:03
that's why I have the act show keys instead of numbers
webmind how do you do that ? 17:05
Nei replacement perl script 17:06
17:10 feng joined 17:14 justatheory joined
webmind ah 17:16
Nei its hacked together in perl5 though 17:17
webmind where can I find it ?
Nei personally I'm using adv_windowlist script, you can find it on ai.onetrix.net 17:18
webmind thanks
Nei it is quite flexible but the drawback ist that it always reserves an extra statusbar just for the act 17:19
webmind hmmm
Nei if you don't like that I think there is also chanact script available on the irssi website
webmind k
I'll take a look, thanks :)
Nei but I don't know whether that can display keybindings >_>
I couldn't get it to
17:41 Amnesiac_ joined 17:43 nothingmuch joined 17:44 G2 joined
gaal yo 17:48
So, it turns out that the YAML rule deserialization problem isn't in YAML at all! look: 17:58
?eval "a" ~~ /a/; say "alive"
ugh.
well, even with a pugs that works, that doesn't.
so PGE embedding appears to be broken with new parrots 17:59
and the YAML stuff very likely works as-is.
18:04 sapper joined 18:10 sapper joined 18:28 Lartuv joined 18:32 Limbic_Region joined 18:36 ajs joined
nothingmuch audreyt: i have a nice perl 5 hack for you 18:36
audreyt nothingmuch: oh? (I'm about to sleep) 18:38
nothingmuch audreyt: catalyst.perl.org/random/nothingmuc..._x_stashed 19:05
19:13 elmex joined
Limbic_Region my guess is she is asleep by now 19:16
19:23 penk joined 19:29 gantrixx joined
gantrixx I was wondering if anyone had a reaction to nazarijo's comments on slashdot today 19:29
obra What were tehy? 19:32
Nei basically /. says 'python is god, perl sucks' 19:33
rep slashdot is never wrong
Nei hahaha
gantrixx actually I think his comments were more along the line that python is replacing perl 19:35
and it is easier to do some things in python than in perl
and I don't think nazarijo speaks for all of slashdot
it's just one person's opinion
but it is a point that I've been speaking about for some time now 19:36
in industry many of my clients are loosing interest in perl
obra Well, These days, many are jumping to Ruby, not Python
theorb Yes, well, this is one major reason for the perl 6 project. 19:37
gantrixx my other point was that if Perl6 takes too long to become available to the masses, there may be no interest in it since it has been replaced by what is considered newer technology 19:38
theorb I doubt it.
Whenever perl 6 comes out, it'll be considered new technology.
gantrixx Perl++
theorb ...hm, or, on second though, not.
I quite agree.
gantrixx I did have one client who was interested in using Perl6, but I had to tell him to wait a year or so 19:39
theorb But me thinking perl++ doesn't help me get a job at it.
gantrixx He is content to continue with Perl5, but he is augmenting his tool set with Python scripts now
If it were me, there would be no Python or PHP, just Perl and mod_perl 19:40
I like Perl, I've spent a lot of professional time with it
theorb The question is, what is easier in Python and PHP, and how can we fix it?
gantrixx but I think, like the article says, Python is gobbling up much of the Perl marketshare
Well the only saving grace is that Perl6 incorporate the best things from Python, Ruby, and PHP 19:41
xinming gantrixx: Things will change after perl 6 is out. And the world will rock with perl 6. :-) 19:42
theorb Yes, I expect things will change after perl 6 is out, what with the comming apoclypse and all.
gantrixx and Perl6 must be made available to the masses sooner rather than later
theorb Exactly.
gantrixx Perl6 is not ready for prime time until ActiveState has a distro for win32 and at least one major Linix distro includes it 19:43
xinming gantrixx: I don't agree. :-) where is parrot, and then where is perl 6. :-)
gantrixx xinming, I don't understand your comment 19:44
xinming gantrixx: And compile parrot on win 32 isn't that hard than perl 6 IMHO.
gantrixx please explain
xinming gantrixx: hmm, I mean, after the perl 6 can be self-hosting, perl 6 is desined to run on top of parrot.
rafl Actually compiling parrot on win32 is a big mess. 19:45
gantrixx OK, well when I refer to Perl6 I'm talking about Perl6 + Parrot
rafl But I think that'll be changed as soon as someone works on parrot milestone 4..
xinming hmm, I still wondering, If perl 6 will can be compiled to binary. hmm, Or, Parrot can compile it's byte code into binary. hmm, I don't mean embed a parrot in the binary. 19:46
gantrixx Well I'm assuming that at first Parrot will be bundled with Perl6 19:48
rafl xinming: Thought it already can do that (even if it's prove of concept, atm)
gantrixx I don't see languages like Python promoting how they will run with Parrot
Parrot is a Perl6 creation
and everyone else has adopted a wait and see attitude toward it 19:49
theorb I don't think anybody outside of perl6 hackers much care about parrot... and not even all of them. 19:50
rantanplan_ xinming, you can create native binaries out of parrot bytecode. But you have to link it agains libparrot.so. 19:51
rafl theorb: Maybe because parrot isn't mature enough. But I think it will be attractive for a lot of other languages as it lets you get rid of lots of hard tasks during writing a compiler.
obra seen alias
jabbot obra: I havn't seen alias, obra
gantrixx I agree that Parrot is a good idea and should eventually be considered it's own seperate entity, but it has to prove itself with Perl first 19:53
xinming ever heard some people would port gcc to parrot... I can't imagine what the world will be if gcc is ported to parrot. Maybe the kernel only need to run the Parrot?
theorb ...and I think that parrot is an interesting idea, but is moving too slowly and holding back perl6. 19:54
...but perhaps I'm just not seeing the current picture -- I've gotten rather behind.
xinming: Making gcc target parrot in the general case is a /lot/ easier said then done. 19:55
gantrixx Yes, I think that Perl6/Parrot is starting to sound like Mozilla to many people. Great promises but very very slow results
theorb Doing it with language interop is probably simply impossible.
gantrixx My only hope is that Perl6 will be as successful as Mozilla/Firefox
theorb
.oO(Help me, Audry-wan Konobi, you're my only hope...)
19:56
rantanplan_ But its interesting how many new languages are created for Perl6 ;) PASM, PIR, Jako .. 19:57
theorb It certianly is, but I wouldn't want to write anything serious in any of them. 19:58
xinming gantrixx: I won't agree, It's a bit like Java, It isn't that slow.
rafl rantanplan_: What's Jako?
theorb To me that's not a sign of how easy it is to write a language that targets parrot, but how difficult it is to write a good one, or we wouldn't keep getting more of them.
Jako is a language that was created around the time parrot first got usable PMCs. 19:59
rantanplan_ rafl, its higher level than PIR, but can be directly translated to PASM. So its like a scripting language for Parrot, pretty near to Assembly but looks and feels like a scripting language.
gantrixx hey guys, I'll chat with you later, I'm on the company clock, I need to do some more Perl5 coding
theorb To me, it looks like pugs was flying along until it started trying to target parrot, and since then has gotten bogged down in a twisty maze of mini-languages, all starting with P. 20:00
rafl rantanplan_: Sounds cool. Can I use it with pure parrot or is it something in languages/?
rantanplan_ rafl, its in the languages/jako directory of the Parrot tree. Makes use of Perl5 and Parrot. 20:01
rafl rantanplan_: Why Perl5?
rantanplan_ Dont know, the parser and lexer stuff seems do be in Perl5.
To generate the AST. 20:02
theorb Because it was easier to write it in perl5 then much of anything else.
rantanplan_ This is one think i still dont really understand. All parser of all Parrot Languages seem to be in another language like Perl5, Python, C etc.. (Or flex, bison) isnt there a Parser engine in the Parrot Core? 20:03
rafl Maybe it can be done with PGE and TGE soon.. 20:04
theorb Only recently, and they still aren't finished.
rantanplan_ I would be interested in porting nawk to Parrot, but dont wanna rewrite the parsing stuff if parrot supports it natively ;) 20:06
20:10 xah joined
rantanplan_ Hehe, Parrot AWK == pawk. 20:12
20:12 dduncan joined 20:20 leo__ joined, leo__ left 20:28 scrottie joined 20:32 justatheory joined 20:34 Limbic_Region joined
webmind hmm, feather has a k-line on efnet ? 20:45
Juerd webmind: Good chance. 20:50
21:45 frederico joined, sapper joined 21:47 nnunley joined 21:51 PJF joined 21:55 penk joined 22:00 Odin- joined 22:06 penk joined 22:21 marmic joined 22:28 nnunley_ joined 22:29 justatheory joined 22:39 elmex joined
nunley is now known as nnunley 22:40
22:55 Southen joined 23:06 nnunley_ joined 23:07 putter joined
putter re python and perl<[56]>... In a corporate setting, it seems clear that python is generally preferable to perl5. 23:13
But the key thing re python and perl6, is that _python should be a perfectly fine way to write p6 code_. As long as p6 doesn't keep losing hook-iness. (Leaving goto to a hairy module, and the talk of disappearing globals, has me worried...) 23:15
As long as p6 contains a superset of the language/runtime features other languages need (at least on backends which support continuations), doing transliteration-based implementations of those languages is remarkably easy. 23:16
Emphasize remarkably. One of the key advantages of perl5 over python when you are playing "interesting" games is the much greater "mess of hooks". With python, one can find one self in a "you cant get there from here". 23:18
One will usually spend _most_ of ones time dealing with the impedance mismatches. As long as p6 remains hooky and a superset, there are no impedance mismatches. Implementing a language is merely the grind of dumping its spec to code. 23:22
For instance, consider goto. Now gone from p6(?). The JavaScript spec is written with intra-method control flow described as goto step N. One can regexp massage the spec into p6 code with goto's. But with p6 not having goto's, you now have to massage, by hand, most(?) of the methods. 23:26
Which ends up being a major part of the total development time, a even larger likely source of bugs, and results in distinctly less maintainable code which no longer cleanly mirrors the spec. 23:27
23:28 scook0 joined
putter So re p6 and python, it should be straightforward to have use Language::Python::OnPerl6::ByTransliteration; ...python code..., 23:33
23:33 lisppaste3 joined
putter and do piln->RPython to use the pypy codegen, etc, etc. 23:33
23:36 bsb left
putter We just need the tuple of pilng, fleshed out oo, working regexps and type system, and we'll be set to go. 23:37
But without all four, its hard to do anything more than toys 23:38
<end missive> 23:39
putter looks forward to a Cambrian Explosion of p6 code ;)
23:40 trym joined
putter Oh, and part of the beauty of piln and multiple backends, is that p6 need not be dependent on parrot. There are lots of path to native codegen once one has already dealt with and boiled away the hairy semantics of the language (p6). 23:44
s/path/paths/
23:48 hcarty joined
putter stevan, audreyt: it might be an interesting thought experiment to create a "wishlist" - tasks which are on the pugs critical path, which someone else can do (well defined, dont require you're global perspective), which you will have to do if someone else doesnt, near term (a few days or weeks), and... 23:54
I think that's it. Basically trying to accelerate the critical path.
cheers & 23:57