svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by avar on 16 November 2006.
00:12 phpError joined 00:34 lambdabot joined 00:40 nekokak joined 00:51 Lee joined 00:57 lyokato joined 01:27 mdiep joined 01:39 mdiep joined 01:41 neonse left 01:43 rintaro_ joined 01:52 mdiep joined 01:54 kanru joined 02:09 nipotaway is now known as nipotan 02:26 rintaro joined 02:30 mjk joined 02:31 dmq joined 02:34 rintaro__ joined 02:47 vel joined 03:30 esammer joined 03:44 weinig is now known as weinig|zZz 03:58 phpError joined 04:02 mako132_ joined 04:16 scaner joined 04:26 BillN1VUX joined 04:28 phpError joined 04:38 dduncan joined 05:09 mdiep joined 05:22 mdiep joined 05:28 dduncan left 05:45 phpError joined 05:55 justatheory joined 06:02 xinming_ joined 06:11 BillN1VUX left 06:15 BooK_ joined
meppl guten morgen 06:25
06:31 alexe joined
jrockway hmm, pugs from svn seems to be dying on my debian machine at "Setup: cannot satisfy dependency network-any" 06:45
it also seems upset that the revision number of my checkout is "3", because I only svk mirror'd HEAD
07:05 alexe left 07:20 frankg joined 07:35 phpError joined 07:37 jrockway joined 07:58 devogon joined 08:06 iblechbot joined 08:07 drrho joined 08:13 Yappo joined 08:14 Yappo joined 08:36 trym_ joined 08:39 kane-xs joined 08:49 elmex joined 09:33 mayur_ joined
mayur_ diff betw USE and REQUIRE? 09:35
09:39 ludan joined
broquaint The difference, mayur_, is that require is called at run-time and use is called at compile-time (basically). 09:40
See. perldoc -f use & perldoc -f require for more info, also, #perl. 09:41
buubot See: Type 'perldoc -f use' in your shell or go to perldoc.perl.org/functions/use.html
lambdabot Title: use - perldoc.perl.org
broquaint kicks buubot
mayur_ Thanks broquaint , lambdabot 09:43
09:44 marmic joined 10:07 mayur_ left
svnbot6 r14859 | kudra++ | This week 10:18
r14860 | kudra++ | Small change to template for yaml file 10:21
10:23 avar joined
svnbot6 r14861 | kudra++ | Also adding my quick instruction list to the repository; someone else may need it someday 10:27
10:39 ludan joined
ludan re 10:40
10:47 ruoso joined
svnbot6 r14862 | fglock++ | MP6-Parrot - added infix:<==>, infix:<!=> 10:54
dmq @tell audreyt ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/Testing/pcre-7.0-RC1.tar.bz2 11:00
lambdabot Consider it noted.
11:08 ruz joined 11:26 theorbtwo joined 11:27 buetow joined
Juerd Feather users, don't forget to re-start your custom apaches! 11:56
12:03 chris2 joined, TimToady joined 12:16 depdust joined 12:21 nothingmuch joined 12:36 trym__ joined 12:42 kane-xs_ joined 12:47 nipotan is now known as nipotaway 13:05 Limbic_Region joined, buetow joined 13:08 iblechbot joined 13:22 weinig|zZz is now known as weing, weing is now known as weinig 13:54 Psyche^ joined 13:55 Psyche^ is now known as Patterner 14:05 weinig is now known as weinig|bbl
Limbic_Region salutations all 14:07
dmq hi L~R 14:11
thanks for mentioning my little challenge.
Limbic_Region no worries dmq 14:13
14:13 mj41 joined
Limbic_Region I also mentioned it over on #perl but not on irc.perl.org 14:14
while I think pmichaud and fglock are probably both up to the challenge, not sure either has the time
14:22 bonesss joined 14:23 c6rbon joined
pmichaud challenge ? 14:31
svnbot6 r14863 | szabgab++ | Give file name on build failure
Limbic_Region yes, just a sec
eviltwin_b <Limbic_Region> use.perl.org/~demerphq/journal/31796 14:32
lambdabot Title: Journal of demerphq (2831)
Limbic_Region pmichaud - use.perl.org/~demerphq/journal/31796
lambdabot Title: Journal of demerphq (2831)
Limbic_Region eviltwin_b beat me to it
pmichaud mmmmmm
Limbic_Region I had mentioned yesterday it would be great if the winner of the metre of beer was PGE or P::C::R
eviltwin_b behold the power of scrollback :) 14:33
pmichaud well, PGE really depends on parrot at the moment, so that'd take a bit of work
but it's a terrific idea and challenge 14:34
Limbic_Region pmichaud - I don't think the dependency on parrot is an issue 14:35
or rather
I don't think demerphq would squabble over parrot being required to support the alternate engine 14:36
the fact that it may not be "easy" to do on the other hand....well that's why I said I wasn't sure you had the time
pmichaud right, I was only thinking that it would involve additional work
not that parrot would disqualify the solution :-)
Juerd Is embedding parrot hard then? 14:37
Limbic_Region PGE is currently written in C, PIR, or a mixture of the two? My recollection was original was in C and current is in PIR
pmichaud that's correct -- PGE is in PIR
(originally started out in C, but switched to PIR)
14:38 fglock joined
Limbic_Region Juerd - the only attempt at embedding parrot that I know of is mod_parrot by jhorwitz and I think he has been away from the project for some time 14:39
The last message to the list was Jan of 2006
pmichaud afk # support call
Juerd Limbic_Region: Right, but wasn't parrot designed to be embeddable, from the beginning? :) 14:40
fglock Limbic_Region: there is Inline::Parrot
Juerd fglock: Oh, cool 14:41
Heh, can a regex engine be written in Perl 5 itself? Then bridging to PGE should be doable :)
fglock Inline::Parrot needs an update to current parrot
PCR is written in Perl 5 14:42
Limbic_Region the 3 most likely candidates I had for winning the challenge were japhy, pmichaud, and fglock
though that was based off skill and familiarity and not time 14:43
Limbic_Region has no idea who might have the time to take up the challenge
dmq No, unfortunately a regex engine cant be written purely in perl5. 14:44
Limbic_Region Inline::Parrot was last updated 2 years ago
dmq However, i do believe that you could do what would be more or less the empty subclass test with XS that calls back into perl. 14:45
fglock Limbic_Region: for me it's a matter of priorities rather than plain "time"
dmq Its just that to implement a regexp engine you have to do c.
Although i imagine somebody suitably creative could write an XS module that allowed you to write arbitrary engines in perl. 14:46
that in itself would be a cool hack.
fglock dmq: that would be nice
Limbic_Region dmq - you mean to do a p5 plug-in regexp engine you have to do c?
dmq its mostly a question of how the regexp engine / perl core api is written. 14:47
currently yes l~r.
as i just said, i didnt do the full glue to allow somebody to do it in perl.
Limbic_Region ok. I didn't think you were implying any RE for any language had to be done in C, but that's the literal way of interpreting what you said 14:48
dmq thats something that can be written anytime.
wheras there are other things that if we want them have to be done before 5.10 is out.
So ive focused on those.
:-)
Limbic_Region for anyone with there ear to the ground (iow follows p5p), is 5.10 guestimated to be released Easter 07?
rgs ENOTUX
dmq the ground rumbled x-mas/new year recently. but rgs can say for sure... 14:49
rgs well, there's a small list of things to be closed
14:50 ludan joined
rgs (of difficult things) 14:50
dmq l~r it essentially comes down to this: a regexp engine has to be able to create a regexp_engine structure, and the complie callback defined therein has to be able to return a regexp structure. which means some perl/c glue interface that can be written by pretty much anybody, doesnt need me directly. 14:51
Limbic_Region heh
Limbic_Region suspects dmq's definition of anybody isn't quite the same as his 14:52
dmq 'anybody' meaning 'does not require a perl5 regex engine internals specialist to do' 14:53
Limbic_Region which is a very good thing and kudos to you
Limbic_Region wonders if we will get audreyt back this week 14:54
dmq Therefore meaning anybody that can write fairly straightforward XS. Which from what i recall is you. :-)
Limbic_Region dmq - um, no. I can write fairly straightforward Inline::C but you must have forgotten admonishing me for using C types instead of the perl API
Limbic_Region is sure with a week or two of focus the could get enough perlguts under his belt to fool someone into thinking he knew what he was doing though 14:55
dmq You just have to push yourself a little harder and have a copy of perlapi open.
eviltwin_b can write XS that makes perl leak references and dump core, does that count? :> 14:56
dmq heh
eviltwin_b (and I have. cyrus, anyone? still don't get what's going on there but havenever had the time to mindmeld with the api) 14:57
fglock there is something started in svn.pugscode.org/pugs/perl5/re-override and re-override-PCRE 15:00
lambdabot Title: Revision 14863: /perl5/re-override
fglock dmq: maybe it just works
dmq yeah, thats an attempt at the same thing based on the older broken api. 15:01
it was kinda doomed by the api unfortunately. 15:02
as the api meant that perl would try to use the current engine for every regex it encountered, regardless as to whether it was compiled by perl itself or whatever.
the association was regexp -> intrepreter -> engine instead of intrepreter -> regexp -> engine 15:03
15:04 gnuvince joined
Juerd rgs: Should I report the uselessness of encoding.pm as a bug? 15:05
It's not really a bug, but the way it works means that it's useful to very few people
Juerd doesn't have a suggestion to fix this in a backwards compatible way :( 15:06
15:06 hexmode joined
dmq do you have a suggestion to fix it in a forward compatible way tho? 15:06
Limbic_Region after reading perldelta in the 5.9.x branch, there are several backwards compatible breakages 15:07
dmq Cause if you do then i think it should be considered.
rgs Juerd: the problems with encoding are known
Juerd: that's why I won't make it lexical actually
it should be split in two instead
Limbic_Region: not that many 15:08
15:10 vel joined
dmq rgs: would it be feasable to make a wrapper that does the split? 15:10
rgs no, thats' a semantic split 15:11
Limbic_Region rgs - not sure what constitutes many, but 5.9.0 = 6, 5.9.1 = 2, 5.9.2 = 2, 5.9.3 = 6, 5.9.4 = 4 for a total of 20
dmq i just meant that maybe a wrapper could handle the lexicalization.
Limbic_Region rgs - I am not saying that's a bad thing btw 15:12
rgs Limbic_Region: many breakages
dmq L~R: what do those numbers mean and where did you get them.
Limbic_Region dmq - incompatible changes
dmq - from perldelta 5.9.0 .. 5.9.4
dmq breakage of workarounds of broken behaviour imo are fair game in a major release.
rgs the section named incompatible changes in perl*delta.pod
Limbic_Region isn't suggesting that they are a bad thing - actually quite the opposite 15:13
rgs example : A bareword argument to chdir() is now recognized as a file handle.
Earlier releases interpreted the bareword as a directory name.
Juerd rgs: Okay, thanks
Limbic_Region people who rely on broken or undocumented behavior shouldn't be allowed to cry when their code breaks 15:14
dmq encourage Juerd to post something that does work
Limbic_Region people write code targeting a perl not written in the last decade shouldn't be allowed to cry when the current perl no longer works for them
s/write/who write/
Juerd dmq: I don't know how to fix it without breaking existing things. The behaviour makes no sense, but there are people who use it and depend on it. 15:15
Limbic_Region does have compassion for the people who stay one rev behind and have to constantly update their code in order to upgrade though
Juerd dmq: encoding::split is on CPAN, it solves the immediate problem by doing ugly things :) 15:16
Limbic_Region or for shops where mgmt forces them to stay in the dark ages for long periods of time and then changes their mind and forces them to upgrade to current over night
dmq well the issue is how do you provide an alternative without preventing the older problematic code from working.
Juerd dmq: It requires a discussion of syntax. 15:17
dmq so if you can reimplement the broken behaviour out of the un-broken behaviour...
Juerd But "use encoding STRING" can't be it. In fact, the syntax of encoding.pm is so irregular, that I think we have a problem
As "encoding" is a nice name :)
Implementation is a matter of copy/pasting the parts from encoding.pm; Re-implementing the old interface is then a matter of using what you made, through a wrapper that knows the @_ order of the old thing. 15:20
And you need to know if STDERR matters.
svnbot6 r14864 | fglock++ | rm v6-MiniPerl6-MO/ (devel proceeds in v6-MiniPerl6/) 15:22
15:32 pmurias joined, cjeris joined 15:33 slanning joined
pmurias Limbic_Region: re: parrot embeding also done by chromatic 15:34
Limbic_Region pmurias - really? I thought he worked on extending parrot by embedding other languages (primarily C via NCI) in it (not the other way around) 15:36
pmurias it is called Embed::Parrot(or something similiar) and lives in the parrot repo 15:38
Limbic_Region cool 15:39
Limbic_Region should follow parrot more closely
pmurias regarding Inline::Parrot the version on CPAN uses IPC instead of the embbeding interface 15:41
15:43 xinming joined
pmurias you find in /misc/pX/Common/Inline-Parrot for a incomplete attempt at parrot i was working on 15:45
although chromatic's version is probably much more up to date 15:47
Limbic_Region yeah, I was just checking it out - it lives in ext/ 15:48
15:54 justatheory joined
pmurias s/you find in/you can look in/ 15:57
16:03 neonse joined, ashleyb joined, ruoso joined 16:23 heath joined
audreyt from #p5p: 16:25
lambdabot audreyt: You have 6 new messages. '/msg lambdabot @messages' to read them.
audreyt 17:21 <@audreyt> I need to sleep.
17:21 <@audreyt> and here is my entry:
17:21 <@audreyt> perlcabal.org/~audreyt/tmp/re-engin....01.tar.gz
17:21 <@audreyt> enjoy :)
17:21 <@dmq> i think you win. :-)
g'nite :) &
Limbic_Region wow
audreyt Limbic_Region++ # reminds me of it 16:26
it took longer than I hoped :)
16:31 alinbsp joined 16:34 chris2 joined 16:39 pmurias left 16:41 kanru joined 17:00 mdiep_ joined
xinming audreyt: ping 17:12
@tell audreyt I ever heard that you told someone that there are four languages which are worth learning... I read it on your blog. But I can't find the answer now, I searched all your blogs last year, and It seems that all comments are gone. So could you please tell me again? 17:14
lambdabot Consider it noted.
xinming @tell audreyt thanks. 17:15
lambdabot Consider it noted.
nothingmuch xinming: my take on it: ASM/C (doesn't really matter which), haskell/scheme 17:16
i'm guessing something like prolog too
but i dunno it
dunno what a 4th would be
but forth seems fitting ;-)
xinming nothingmuch: perl... :-) 17:17
In fact, Here, someone suggested me that I should learn python...
and I read the documentation... I didn't find any special in python compared with perl. It seems, that all thing that python have, perl does it in a convenient way... 17:18
:-)
What I like in python which might be, python use . for method call. :-P 17:19
nothingmuch i found python uninpiring
uninspiring
xinming why people suggested me python is all because of django...
nothingmuch perl is not really a fundamental language
but it's a nice wrap up of some of the good parts of all of what I mentioned 17:20
wolverian perl has culture, CPAN and linguistics. it's an experience too.
nothingmuch yes, of course
xinming and, when I saw the template for django, It's like another TT to me. :-P
nothingmuch wolverian++ # good points
xinming So, i gave up.
and then, my turn to persuade him to use perl. :-) 17:21
17:21 ruz joined
xinming at least, most people say python has better oo support. And I didn't find it either. 17:22
eviltwin_b it does (compared to perl5 at least) but that's nto actually saying much
think it does a slightly better job with the notion of class variables 17:23
nothingmuch xinming: i think perl 5's OO support is actually underappreciated
xinming hmm, well, to me, perl is a blessed hash. and in python, It's a blessed dictionary.
nothingmuch because perl lets you twiddle with the metamodel very freely 17:24
perl's problem is that it has an insane metamodel though ;-)
stashes and special arrays and stuff like that ^_^
xinming and, the most obvious thing is...
I said, I can use $a.$b
17:24 ashleyb joined
xinming very convenient... how can we do this in python... the one whom tried to persuade me shuts up. :-P 17:25
though, I know python can, But I like the handy grammer.
But, one thing I think python is great is. It's speed. 17:26
17:28 bonesss joined
nothingmuch really? 17:29
i always thought it was notoriously slow
jrockway shootout.alioth.debian.org/ 17:30
17:30 ofer1 joined
lambdabot Title: The Computer Language Shootout Benchmarks 17:30
jrockway perl wins in a lot of those
i mean, not compared to C or java obviouly
but among php/python/ruby/perl, perl is usually fastest
perl uses *a lot* of memory though
not bad for a "dead language" though
rgs both speed and memory usage will be improved by 5.10.0
17:30 stevan joined
jrockway yup 17:30
demperhq's regex engine work is awesome 17:31
rgs and Nicholas' shrinking of internal structures
Juerd 18:30 < jrockway> i mean, not compared to C or java obviouly
It's not obvious that Java would win from Perl. 17:32
For many real world applications, I expect Java to be far inferior to Perl, in terms of performance.
It's good for applications without user interface, that don't do any text processing. And that's quite a niche :)
xinming But people prefer money.. so, many people like java. :-P 17:33
Juerd Sure, Java is rather popular in commercial environments.
xinming I ever went to the site on language benmachmark. 17:34
jrockway java has jit
Juerd (Including universities, mind you!)
jrockway: So?
broquaint Java doesn't process text it shifts paradigms with XML!
Juerd broquaint: :\
xinming It seems. lisp is really attractive on both speed and memory usage. :-P
jrockway walking an optree is slower than directly executing code on the processor
getting data from registers, etc.
broquaint Did I get my terminology wrong, Juerd? ;) 17:35
jrockway i imagine parrot will give a good speed boost to most dynamic languages
(i've heard it has, but i haven't been worried about speed personally)
xinming I don't think so. :-) every release for software... the "author" will claim that there will be a better performace for the new version. 17:36
But I never see it happened... :-P
Juerd broquaint: No. You got everything right, and according to reality, which sucks.
jrockway: Sure, that's very nice for the raw operations, like maths. 17:37
jrockway: But it means very little when it comes to text processing or real time interaction.
jrockway true 17:42
17:50 hexmode joined 17:59 ruoso joined 18:19 PerlJam joined 18:20 lunny1 joined 18:22 lunny1 left, wilx` joined 18:23 alinbsp left 18:25 BooK joined 18:26 Psyche^ joined 18:27 buetow joined 18:34 Psyche^ is now known as Patterner 18:54 phpError_ joined 18:56 sweinig_ joined 19:01 weinig joined 19:22 mr_ank joined 19:23 larsen_ joined 19:47 KingDiamond joined, fglock left
svnbot6 r14865 | pmichaud++ | '.goto' in t/01-sanity/07-tailcall.t isn't defined in Synopses, 19:51
r14865 | pmichaud++ | moving to t/unspecced/ so that we can better use t/01-sanity/
r14865 | pmichaud++ | in perl6/Parrot.
19:51 marmic joined 19:55 justatheory joined 20:00 bernhard joined 20:03 thepler joined 20:04 foo\ joined 20:11 unfo joined
Limbic_Region perlbot nopaste 20:13
perlbot Paste your code at sial.org/pbot/perl and #perl will be able to view it
lambdabot Title: sial.org Pastebot - pasteling
jrockway wow: lambda-the-ultimate.org/node/613 20:15
lambdabot Title: Fold Must Fold! | Lambda the Ultimate
jrockway removing map or fold "improves" a language?
gnuvince jrockway: fold often confuses programmers, so it's not an efficient way to communicate what a program does. 20:17
20:17 TSa joined
gnuvince and in any case, that's an April 1st post. 20:18
jrockway ahh, ok
i am writing some elisp and was really wishing for fold, and that's what google gave me :)
i like functional programming, because i can usually ponder it for a few minutes and then type a line of code 20:19
much better than the don't-ponder-and-write-100-lines approach
maybe
it's also a good distraction from my PHP-infected day job :)
TSa HaloO 20:20
20:22 pmurias joined
pmurias jrockway: i enables you to reuse the keyword for something else ;) 20:23
20:43 stevan joined 20:48 Limbic_Region joined, hexmode joined
Limbic_Region hrmm 21:05
21:30 Aankhen`` joined 21:35 prism joined 22:01 premshree_ joined 22:21 beldor joined 22:44 Limbic_Region joined 23:05 KingDiamond joined 23:13 pmichaud joined 23:17 larsen_ joined 23:34 neonse left 23:38 hexmode joined 23:42 buetow joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!