6.2.11 released! | pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: sial.org/pbot/perl6 | < audreyt> my fingers juxpatosed
Set by nothingmuch on 24 February 2006.
00:12 iank joined, rgs joined, iank left 00:26 macli left 00:48 robkinyon joined 01:05 f0rth_ joined 01:08 khisanth_ joined 01:09 khisanth_ is now known as Khisanth, lao_ joined 01:16 Bit-Man joined 01:33 joepurl joined 02:13 hexmode joined 02:14 lumi joined 02:32 xern joined 02:49 justatheory joined
svnbot6 r9220 | fglock++ | * misc/pX/Common/iterator_engine_p6grammar.pl 02:59
r9220 | fglock++ | - working Perl 6 compiler - about 100 lines of code are now written in perl6
r9220 | fglock++ | - started a Prelude
r9220 | fglock++ | say 'compiling Prelude';
r9220 | fglock++ | sub infix:<+> { eval(' $_[0] + $_[1] ', :lang<perl5>); }
r9220 | fglock++ | - optimized rules - use Text::Balanced for implementing <literal> and <code>
r9220 | fglock++ | - 124 tests for p6rule.pl
r9220 | fglock++ | - note: spec change '$<>' to '$()' not applied yet 03:00
03:17 mako132_ joined 03:35 Shachaf joined 04:00 aufrank joined
aufrank hello all! 04:00
azuroth hi frank
aufrank if anyone has some reading time and is feeling docs-ish, I'd appreciate it if you'd commit fixes to the 'Variable scoping' section of Perl6/Overview.pod that I wrote today 04:02
04:10 vborja joined
aufrank has to go to bed 04:11
gnight!
azuroth actually, what's with perl6/overview line 159... "say @mixed.elems; #last element, prints 1.23" ? 04:18
04:22 ajs_home joined 04:37 typonaise joined 04:43 beppu joined 05:00 kanru joined
ayrnieu ?eval my @a = <a b c>; @a.elems 05:25
evalbot_9220 3
ayrnieu ?eval my @a = <a b c>; say @a.elems
evalbot_9220 OUTPUT[3 ] bool::true
05:27 K joined 06:09 Arathorn joined 06:17 integral joined 06:20 joepurl joined
GeJ Morning all 06:31
ayrnieu 'morning.
PerlJam doesn't this channel ever sleep? ;-) 06:34
azuroth only when I'm not here 06:40
Kattana Its always morning somewhere.
..Its everytime anytime. 06:41
szabgab On the wiki of OSDC::Israel someone started to transcribe the opening Talk of Larry Wall 06:46
with some help it could be finished in a few hours wiki.osdc.org.il/index.php/Main_Page
exact link: wiki.osdc.org.il/index.php/Larry_Wa...re_Perfect 06:47
06:49 pdcawley joined 06:59 fandango joined
ayrnieu ?eval [<1 2>] 07:02
evalbot_9220 ["1", "2"]
ayrnieu ?eval my @a = <'a b' c>; my @b = <<'a b' c>>; [[@a], [@b]]
evalbot_9220 [["\'a", "b\'", "c"], ["\'a", "b\'", "c"]]
ayrnieu as neither Juerd nor SDF are likely to get back to me soon, I think I'll go drink a bit. *wave* 07:04
ingy audreyt++ 07:08
^^ wiki style development pioneer
obra ingy: I like to call her the first "anarchitect" 07:11
miyagawa anarchitect++
ingy :)
ingy heads to $cafe on %scooter 07:12
obra ingy: are you back in .tw? 07:14
ingy obra: yeah 07:15
obra: do you give free commit bits on jifty?
obra ingy: not yet. 07:18
we'd like to see a patch or two first.
"I'm not as cool as audrey"
ingy or me 07:19
:p
obra Fair enough
obra is uncool
Just a boring business guy
ingy just let go
obra Need to hire more people. Then I can let go more
<- hiring perl hackers and UI designers. 07:20
ingy :)
&
07:26 Aankhen`` joined 07:27 jp-autar1 is now known as jp-autark, TimToady joined, r0nny_ is now known as r0nny
bd_ hm, does svn.openfoundry no longer run https? 07:28
07:36 pdcawley joined 07:49 fglock joined 07:51 ghenry joined 07:52 iblechbot joined
fglock Hi! anyone around to discuss bootstrapping p6 grammar? 07:54
I've got this line: rule sub_decl { sub <ws> $fix := (infix|prefix|postfix) \: \< $id := (.*?) \> <ws>? <block> } 07:56
TimToady fglock: I think we're most of us rather trashed from late dinner last night. The auction went on a long time.
fglock np - just in case someone reads the backlog... 07:57
and then I want to declare a sub - but the grammar was not compiled yet, it was only parsed 07:58
TimToady In the long run foofix:<?> has to be something recognizable as a separate term, but I understand you're bootstrapping.
07:58 marmic joined
fglock so it needs to parse, compile, and continue parsing, right? 07:59
TimToady Perl 5 arranges that any sub is callable as soon as it's defined.
fglock ok
TimToady Indeed, a BEGIN is just a special sub to Perl 5. 08:00
fglock so as soon as a 'rule' is parsed it should be compiled. I'm simulating this by splitting the program in chunks and parsing/compiling each one separately 08:01
svn.openfoundry.org/pugs/misc/pX/Co...grammar.pl - just in case. I'm aware there are errors in p6 syntax, but it mostly works 08:03
08:03 shmar joined
shmar lo all 08:03
webmind good morning 08:06
fglock TimToady: how about: rule sub_decl { sub ... { $().compile; return ... } } 08:08
08:11 cookoo joined
TimToady fglock: sorry, bit distracted getting ready to go out and sightsee in Jerusalem, so may have to bug out at any moment, but go on... 08:12
08:12 drrho joined
TimToady Hmm. 08:13
That's telling a string to compile itself. Seems a bit odd.
It's also a bit strange to refer to the complete match via $() when you haven't finished matching... 08:14
fglock not sure about the syntax, sorry - I mean, once there is a match, compile the ast
TimToady One could do a temporary compile to stupid implementation, then recompile smarter at CHECK time. 08:15
fglock like if it were just a sub prototype 08:16
TimToady Yeah, ... kinda works that way already. 08:17
Hopefully, one just re-uses the same compiler, but with more information.
you don't really need two separate compilers, which would be bad policy anyway.
The fuzz that can precompile needed bodies can presumably also be fuzzed into the 08:18
runtime direction to do on-the-fly optimization. 08:19
All a matter of knowing what's needed right now, and what you can get away with based on the current status.
fglock mm - lazy compilation 08:20
08:20 cookoo left
TimToady CHECK time is just the most usual place to think about that. 08:20
Obviously anything that changes the current grammar has to at least have its interface installed immediately. 08:21
08:22 cookoo joined
fglock TimToady: thanks! 08:22
TimToady welcome 08:23
08:29 Ara4n joined, fglock left 08:31 putter joined
TimToady off to do the tourist thing. tata 08:31
putter bah. missed fglock. 08:32
have fun. 08:33
08:39 putter left
r0nny yo 08:40
08:52 Migo joined 08:54 shmar joined 09:11 Wassercrats joined
Wassercrats I just know one of you. 09:12
09:14 Wassercrats left 09:18 rantanplan_ joined 09:22 kane_ joined 09:45 shmar joined 09:50 lypanov joined 09:58 lypanov joined 10:07 chris2 joined
lypanov morning chris2 10:07
chris2: today my world power adaptor kit will arrive!
shame about the errm.. lack of computer to use it for :P
chris2 hi lypanov ;) 10:08
10:32 Aragone is now known as Arathorn 10:33 nnunley joined 10:38 pdcawley joined 10:42 nnunley joined 10:53 kanru joined 10:56 feng joined 11:36 KingDiamond joined 11:48 KingDiamond joined 12:01 kanru2 joined 12:10 kanru2 is now known as kanru 12:15 iblechbot joined 12:31 pdcawley_ joined 12:32 juerd joined, juerd is now known as Juerd
rafl Hello Juerd 12:35
GeJ greetings Juerd. 12:36
hum, dunno if someone told you, but I think lilo was looking for an admin on feather yesterday.
Juerd Hi 12:38
GeJ: I see. Could be. I have lilo on ignore...
lypanov lol 12:42
Juerd What's so funny? :)
Who doesn't ignore lilo... :)
lypanov Juerd: the fact that the idea is sane ;) 12:43
mauke oh, should I remove my "lol" filter? 12:46
I seem to be missing a line 12:47
Juerd 12:55
12:56 KingDiamond joined
Juerd mauke: Any else missing? 12:57
mauke no, thanks 12:58
lypanov mauke: lol 12:59
13:09 fglock joined
fglock putter: pong 13:10
Juerd: cgi-irc in feather has stopped 13:13
Juerd: it may be a problem with https - wget gives 'SSL23_GET_SERVER_HELLO:unknown protocol' 13:16
svnbot6 r9221 | audreyt++ | * scratchpad with nothingmuch on designing piln desugaring for Code
r9221 | audreyt++ | objects in order to run on p5 ObjectSpace
Juerd Can't fix it now anyway, so you'll have to cope without 13:18
fglock Juerd: np - I've just learned irssi :) 13:20
13:22 kolibrie joined
audreyt rehi from tel aviv airport 13:23
going to fly to vienna, then duesseldorf, arriving at midnight, _then_ deliverying two talks starting from tomorrow morning 13:24
rgs german perl workshop ?
audreyt yeah
fglock++ # woot, compiler!
rgs please hug Nicholas for his pumpkinging then 13:25
audreyt will do!
boarding in 6mins... bbiab. :)
fglock audreyt: hi! I'd appreciate a code review, specially in the part written in p6
audreyt will definitely do it during flight 13:26
fglock thanks! p6grammar.pl is the main file
audreyt I look forward to turn the much normalized p6 parsing as committed by recent Synopses to this 13:27
but I need to run now. will sync back tonight
audreyt waves &
fglock that's S05? link?
audreyt it's all in Perl6::Bible 13:28
see p6l for my commit logs
and larry's
fglock ok! have a nice trip!
audreyt mostly, "my $f = sub x {...}" (without trailning newline) is valid
also much clearer macro rules
and import rules
so "use" and "BEGIN" can be written as macros relatively easily
will journal about htem
13:28 alexe joined 13:34 KingDiamond joined 13:46 gaal joined
gaal audreyt: will have had a nice flight! 13:46
lypanov hehe 13:47
Juerd rafl: Try again 13:48
rafl Juerd: Same IP address?
Juerd I think so. 13:49
14:05
That happens every time I come out of suspend... 14:06
I don't know why.
rafl Stop suspending then.
Juerd Give me your charger, then.
rafl Erm, no.
Get an 8-cell battery.
Juerd Yeah, yeah
I know
lypanov refrains from using his normal response "buy a mac" 14:13
14:16 Qiang joined 14:17 mauke_ joined 14:31 fglock left 14:37 juerd joined, juerd is now known as Juerd 14:38 GeJ joined 14:45 GeJ joined 14:47 hexmode joined 14:49 nothingmuch joined
nothingmuch morning 15:05
15:06 alexe left
anatoly morning 15:08
actually, why morning? :) 15:09
webmind it's afternoon here
anatoly yep, here too
webmind well, good afternoon then :)
anatoly right, good afternoon :)
turning evenegish though, and then it'll be a good evening, I guess? 15:10
15:12 justatheory joined
nothingmuch yeah, you could say it's evening 15:14
gaal: ping
ingy hi nothingmuch
nothingmuch ingy: please release Spork::Formatter::Kwid 15:15
Spork tests fail on account of it not being a dep
and adding it as a dep won't work, since it doesn't exist
ingy mmmm Spork 15:16
hmm that's near the top of my inifinite todo list 15:17
lypanov came back from his first trip to the states with a spork
lypanov was most happy
nothingmuch at the OSDC dinner we ate with sporks 15:18
lypanov grins 15:20
15:22 Sandrina joined
gaal nothingmuch: pong 15:23
I got your stuff
nothingmuch goodness
i also dropped my phone somewhere on the way to the train station
gaal eep
nothingmuch so i'll be dropping by for a visit say saturday
gaal cool, i think i'll make pizza and also invite bsb and lumi over 15:24
we can have an afterparty :)
what's svk for "svn st", from my local wc to the upstream?
nothingmuch huraah 15:25
15:25 fglock joined
lypanov gaal: i just use svn di /trunk . 15:27
nothingmuch hola fglock
gaal lypanov: "path /trunk is not a checkout path." 15:29
lypanov gaal: sorry thats your umm.. forgot the name. your um... thingy
fglock nothingmuch: hi! 15:31
15:31 mauke_ is now known as mauke 15:32 vel joined
fglock looks for what to do next 15:34
svnbot6 r9222 | gaal++ | r9243@sike: roo | 2006-03-01 16:46:12 +0200 15:38
r9222 | gaal++ | * Some last cleanups to Peek talk. Online at
r9222 | gaal++ | perlcabal.org/~gaal/peek/start.hml
anatoly moose, gaal 15:46
GeJ gaal: I get a 404 on the URL 15:50
15:51 robkinyon joined 15:54 kanru joined
nothingmuch seen luqui? 15:57
jabbot nothingmuch: luqui was seen 31 days 8 hours 10 minutes 27 seconds ago
nothingmuch ugh
his mail bounces, too
gaal: btw, i am renaming OSDC to ØSDC
please help me hack gabor's server 15:58
Arathorn zero source? :D 16:01
nothingmuch 2haha 16:07
• [syeeda:~] nothingmuch % cd /tmp/
• [syeeda:/tmp] nothingmuch % mkdir foo
mkdir: foo: File exists
• [syeeda:/tmp] nothingmuch % mkdir moose
mkdir: moose: File exists
svnbot6 r9223 | fglock++ | iterator_engine - implemented '??' quantifier 16:12
gaal rehi 16:21
s/hml/html/ in that URL
hey toly! 16:22
16:27 r0nny joined
anatoly hey gaal 16:32
svnbot6 r9224 | fglock++ | iterator_engine - implemented <?non_capturing_subrule> 16:39
nothingmuch is avoiding his sister 16:43
she doesn't know we bought her an ipod for her birthday
and now she wants me to burn her CDs
;-)
Juerd In a talk here ("What's wrong with Parrot and Perl 6"), one major argument is that there are 0 lines of code yet for Perl 6. 16:45
This ignores pugs, this ignores the 35000+ lines in unit tests... 16:46
I don't even know how to begin to react to that...
And then, it discusses pugs...
Ah well, Audrey will convince the audience tomorrow :)
fglock about rules: <!abc> is a negation of '<abc>' or 'abc' ? 16:47
kolibrie nothingmuch: burn one or two, just so she doesn't get suspicious
lypanov Juerd: lol 16:48
nothingmuch kolibrie: the surprise party is saturday
i told her my itunes is broken and that we'll do it saturday or something instead
lypanov hehehe
kolibrie nothingmuch: :)
Juerd fglock: Probably :)
nothingmuch Juerd: hehe
Juerd: whose giving the talk? 16:49
Juerd nothingmuch: Marc Lehmann
Assuming "Who's"
nothingmuch i was speaking inglesit
Juerd Argument: "Weak typing != Strong typing", "Developers are biased to strong typing, as you can see in the many examples."
nothingmuch who giving the talk belongs to? 16:50
Juerd Ignoring that weak typing, and anything in between strong and weak typing, is still possible.
nothingmuch uh
is he against weak typing?
or for it?
Juerd For it.
Or so is my interpretation
Current bashing: "-> becomes ., like the rest of the world uses". 16:51
nothingmuch okay
Juerd Memo to all: no pun in documents...
nothingmuch err
perl 6 *has* strict typing
is he giving a perl 6 sucks talk without having read the synopses etc?
fglock types on a weak keyboard 16:52
Juerd Argument wrt the dot: java versus prel 16:53
perl
nothingmuch hahahaha
Juerd "Why does it have to look like java?"
nothingmuch finds it borderline-pathetic that people are so hung up on syntax
Juerd "Look at this lwall quote: C uses i++, while java uses system().foo().bar().i().value().please().increment();"
This is FUD.
There is no ground in this whatsoever.
Moving from -> to . will *not* make $i++ go away.
nothingmuch in fact 16:54
did he notice that . is 50% of the length of ->?
16:54 rantanplan_ joined, TimToady joined
nothingmuch that's like haviing i+ autoincrement ;-) 16:54
TimToady: is there any draft on AG stuff i can read?
Juerd nothingmuch: Yes, but that was not important.
nothingmuch if not, i started writing a draft based on roles 16:55
but i'd like a lexical form of 'but' first
since that'll make it much nicer
Juerd Argument: "The bless fundamentally incompatible thread didn't end in concensus, decision or implementation"
nothingmuch maybe 'temp', actually
err
Juerd As if only measurable progress is progress...
nothingmuch we still have bless
it's just going to remain ugly and slow, is all 16:56
are you arguing back?
Juerd No.
TimToady nothingmuch: not that I know of.
Juerd What does "lueckenhaft" mean?
Juerd wants a dictionary that is text browser compatible
mauke dict.leo.org/?search=lueckenhaft
oh
fragmentary, incomplete
nothingmuch TimToady: is there a genral direction in which the tree processing stuff was going to grow? or is it already clear that it's going to be attribute grammarish 16:57
mauke Juerd: works in w3m
Juerd Ah 16:58
TimToady I think AG is actually somewhat limited, but I'm not sure where to go from there.
Juerd Title of slide "Who wants or needs Perl 6"?
I immediately reacted "I do"
"Why?"
"Neat new features"
"Like?"
"@several"
The reply to which was, more or less, but you don't *need* that.
And: but with Parrot it won't be possible (one of the features I named was threading in a nice way) 16:59
nothingmuch TimToady: i'll read the SYB article first
it's supposed to be interesting
IMHO ags are not very flexible
but capture a specific form of processing very very well
Juerd I'm considering giving my Perl 6 myths lightening talk tomorrow or the day after
nothingmuch Juerd: *sigh* 17:00
i hope it helps (audrey, your lightning talk)
i had a pretty lame talk at OSDC about compiling Perl 6
and why I think it will work out in the end
Juerd Most of this talk, though, is technical and against Parrot. I can't judge or argue that, but I think that in that he's right. 17:01
nothingmuch things that Perl 6 will be able to do but that Perl 5 can't (too hard to hack guts, too few strong/static semantics)
are his slides online?
fglock TimToady: is <!abc> a negation of '<abc>' (or 'abc') ? is <!'abc'> right?
Juerd I don't know
"Perl 6 is blocking [Perl 5]" 17:02
nothingmuch isn't the CPAN still growing polynomially?
Juerd Most of his slides are citations, though
The real remarks are in speech
nothingmuch hmm 17:03
TimToady <!'abc'> is short for <!before 'abc'>, though in this case the quotes don't do anything.
17:03 Aankhen`` joined
TimToady <!abc> just means treat fail if <abc> would succeed at this point. 17:03
s/treat// 17:04
Juerd "autoboxing, another neat word we took from Java" 17:05
nothingmuch err
java didn't have that for a long while
and it reads "performance"
fglock TimToady: ok
nothingmuch it's completely non user visisble
(except in java)
Juerd These arguments are so irrelevant to the actual language that I can't argue with them.
I'd need to research stuff.
I hate this.
nothingmuch ask him to put an IRC screen up 17:06
i'll be happy to argue with him
=)
Juerd Nah, would take too long, I think.
And the audience is German
nothingmuch we had an IRC screen at OSDC
during the auction
loads of fun
Juerd (All the English citations are (correctly) translated -- English in the slide, translated in speech)
"?? !!" is questioned WITHOUT ANY DISCUSSION of why it is there. 17:07
And without any discussion; next slide. 17:08
Unicode characters would make the code unreadable. 17:09
The point in the slide being "Code (un)readable?"
"Who proves that the new operator works? Hm, well, pugs, maybe, but..." 17:10
s/new operator/operator move around/
Or however you'd translate that
nothingmuch *sigh* 17:11
Migo: what do you think of Perl 6's fudgery after a few days with it?
Juerd Hm, finally a real argument
Twigils are hard, and there's "twigilitis". This one I agree with.
nothingmuch all twigils have a long form 17:12
Migo nothingmuch: mixed filling. My first reaction was against "?? !!", against unicode, against haskel, whatever :-)
Juerd "Buzzword compliant Microsoft vocabulary?"
nothingmuch $*IN --> $GLOBAL::IN
if he prefers that
Juerd nothingmuch: See perlvar versus English.
For reading code, you need to know them.
nothingmuch i guess i can see that, but twigils have a sort of consistent feel to them
and the variable *names* are typically good enough 17:13
Juerd Agreed, but they do make things harder.
nothingmuch Migo: what do you feel about this now that you have gotten to know more? svn.openfoundry.org/pugs/ext/Getopt...Process.pm
Juerd I think for good, he thinks for bad.
That's just gut feeling.
Now being bashed: STM and non-locking
nothingmuch anyway, it sounds like he hasn't tried writing any 17:14
Juerd But I don't understand it all.
nothingmuch HHAAAAAAAAAAAAAAAAAAAAAAA
he is arguing *for* threads?
err
locks
Juerd I don't know
nothingmuch okay
Juerd "This is a neat feature. Do I need it? No."
nothingmuch let him use forks
Juerd I can only translate some of what he says.
He doubts Parrot'll handle it, I think.
nothingmuch frankly, i doubt parrot will handle it too, in it's current state 17:15
it's seems prematurely optimized
Migo nothingmuch: I am not really against "?? !!", unicode should be completely optional, haskel is ok in small amounts. :)
Juerd Right, I agree with most of the parrot bashing.
nothingmuch but that's not to say that it can't be fixed
parrot is not a production target for anything 17:16
they have lots of freedom
Juerd "Why sponsor/grant only key developers?"
nothingmuch and it *is* fast and featureful, so far
yeah, i would love to get $2 a week for hacking perl 6 ;-)
Juerd Well, parrot is not fast.
He rightly argued against some much cited benchmarks
That they are unfair. He then proposed a fair benchmark, and perl5 or gcc4 would win from parrot. 17:17
nothingmuch well, i won't argue with that, i really haven't looked into it
but i dunno
Perl 6 could still run on Mono, CLR, JVM, Perl 5 optrees
Juerd Right
nothingmuch in fact, we are looking into compiling from pugs to Perl 5's optree right now 17:18
that will not be any slower ;-)
rgs said it'll be hard
but probably not as hard as getting a production parrot
TimToady depends on how much emulation you have to do, nothingmuch
nothingmuch anyway
TimToady: you mean for fast or for working?
fast will not be easy, i buet
bet
Juerd "Every idea is welcome, like this [stupid] thing: if ... { } uncond { } cond { }"
TimToady for fast.
rgs hmm PIL to p5 opcodes ? 17:19
nothingmuch TimToady: ofcourse
Juerd As if that was even close to concensus.
nothingmuch i'm mostly interested in stable and interoperable
that's what i'm optimistic about
and I am guessing that it'll be much faster than the current haskell runtime
rgs: aye
anyway, dinnertime
then syb time
then bsg time
rgs nothingmuch: which perl 5 are you targetting by the way ?
syb ? bsg ?
you guys speak in english 6 ? :p 17:20
nothingmuch rgs: i have no idea, i've just started poking around the perlguts in the hopes of finding optree stuff
;-)
scrap-your-boilerplate - an article that's supposed to be a nice contrast to AGs
(attribute grammars)
Juerd I speak English 0.1
Beta.
(c) Google
nothingmuch bsg: battlestar galactica
we need a purl here
syb?
bsg?
that would have worked more easily =)
anyway, ttyl 17:21
rgs see you
nothingmuch Juerd: please mock him publicly saying he looks like a fake møøse, and send my warmest regards to balas and audrey
Juerd Ah, nice. Discussion. Leo versus Marc
rgs TimToady: do you have an opinion on that PIL to P5 ops thing ?
Juerd afk
rgs need really to "make peephole optimizer optional" I think, for that to be easier to hack 17:22
TimToady It would be easy to make it bug compatible with P5. :-)
Juerd Unfortunately I don't understand much of the kind of German that Leo speaks :(
rgs bug compatibility, you're speaking like a maint pumpking ! :)
Juerd To me, it sounds like a combination of Scandinavian and German :) 17:23
mauke where is he from?
rgs with some parrot assembly thrown inside for good measure
Juerd .at
mauke haha, good luck
svnbot6 r9225 | fglock++ | iterator_engine - implemented <!negated_subrule>
rafl Juerd: Shall I translate for you? 17:24
Juerd I just hear "aonhu sehuantoheu snthaoeunthaeous nthaosnetu haoeu benchmarks aoeunsthaenuth aonethsnthoe just in time compiler aoesunthntehonethusnth compiler oaenuhnshnh compile time snthnthoeunt"
rafl: Feel free
rgs Juerd: at least you got the idea :) 17:25
Juerd Why does he say "und so weiter, und so (vort?)" 17:27
(both)
mauke fort
usw. usf.
Juerd Is that an idiom, to have both?
mauke yes
Juerd Isee 17:28
mauke "und so fort" doesn't even exist by itself 17:29
Juerd I see
I knew usw
mauke leo translates it as yada-yada-yada :-)
Juerd We have enz ("En zo voort") in Dutch
mauke nice 17:30
fglock TimToady: does '<word> & <short>' means the match must be a short word?
Juerd Nicholas enters the stage. 17:34
Ahhh, english. Good.
rgs applauses
TimToady fglock: it means the two subrules must match with the same length. 17:35
spinclad in english we sometimes have 'and so forth, and so on', which is likewise redundant
TimToady (over the same string)
17:36 mj41_ joined
spinclad Juerd: I spoke with lilo about feather, referring him/her to you (ha, what did I know?) for the real story. He/she was seeing a lot of access from feather, asked about CGI, I didn't know of any (wrong again). He/she may have blocked cgi-irc from feather now? (Otoh that doesn't seem to match fglock's symptom earlier.) 17:37
Juerd spinclad: Could be.
spinclad: I'll talk to him as soon as I can.
spinclad: It'll probably be easily resolved if he knows that only authenticated users can access. 17:38
spinclad Juerd: I did mention that.
Juerd I see
spinclad (but not about CGI, of course) 17:39
Juerd What was his stance?
spinclad he didn't answer back after I braindumped what I knew of #irc access on feather (typically ssh -> screen -> irssi) 17:40
Juerd Okay.
Thank you for letting me know in detail. I appreciate it.
spinclad np, glad to help, glad to not make things worse i hope
Juerd We'll see. 17:41
I think that aside from his global message spamming inclination, lilo's a reasonable guy.
Arathorn nothing wrong with global message spamming :D 17:42
spinclad too much is wrong but his level doesn't happen to bother me
Juerd spinclad: It used to be tenfold as much.
spinclad that would be much then. 17:43
nothingmuch lambda-the-ultimate.org/node/1332 17:51
rgs lambda-the-ultimate.org/node/1313 was interesting too 17:55
17:55 salty-horse joined
nothingmuch pushes to todo list 17:56
;-)
back to syb
17:56 dduncan joined 18:15 putter joined 18:16 Shachaf joined 18:17 salty-horse left
putter surfaces from perlguts, attempting to use the regexp hooks bsb++ turned up to swap in an alternate engine. that part (sort of) works. but setting up the regexp struct so $1,etc get set, even if you ignore leaking, blech. But ingy++ for Inline::C. 18:23
Steve_p That's just crazy 18:25
gaal rehi 18:29
fglock putter: hi! 18:30
putter hi fglock
re crazy, it should actually be straightforward. two hooks, regcomp and regexec. comp creates a regexp struct which wont get optimized away (// works for my old 5.8), and overwrite a non-essential field in it with the pointer you wish to pass from comp to exec. in exec, do whatever, and set up the the capture offsets. only the captures arent showing up. :( 18:33
18:33 chihchun left
putter and xs is having some funky issues calling a sub defined in one module but accessed via CORE. 18:33
fglock (I'm intermittent - meeting) 18:35
putter :)
fglock putter: regex is compiled to bytecode, right? does generating bytecode let you access more basic/useful routines, perhaps? 18:36
putter: did you see the latest i_e_p6grammar.pl? 18:43
18:43 Arathorn is now known as Aragone, KingDiamond joined 18:48 shmar joined
shmar lo all 18:48
gaal hey shmar 18:52
fglock putter: perl.plover.com/Rx/paper/ 18:54
19:15 levengli joined 19:16 levengli left 19:24 gaal joined 19:33 draconit joined 19:42 cookoo left 19:45 DavidCraft joined 19:46 DavidCraft left 19:50 samuel joined 20:01 putter joined 20:04 samhain1138 joined
putter re bytecode, it's basically replacing the C function which does the regexp search, and (with the unclear but possible assistance of other parts of perl:( ), sets up the match values. So it should simply be a matter of copying out that functionality. But this is perlguts, which makes the simple incomprehensible. ;) 20:08
re Rx, neat! thank you. added to the (mostly doc about what isnt working) file I'm about to commit (and then perhaps abandon). 20:09
re latest grammar, probably not the latest. looking... 20:10
fglock: neat :) 20:17
oh, fribble. 20:33
svnbot6 r9226 | putter++ | misc/pX/Common/Regexp-ReplaceEngine.pm created. Perl5 has hooks to permit interception of regexp compilation and execution. This is an attempt to use them. No-one really has yet. The objective is perl -we 'qr/a normal regexp/; {use Regexp::Perl6; qr/<ws>?<word>/}'. Extensive comments included. Callbacks to &regexp_hook_exec,etc subs work. Next big step is getting $1,etc to work. Volunteers encouraged. It is written in Inline:
putter And a big bsb++ for turning up the hooks. 20:34
20:34 amnesiac joined
putter Hmm. There is a line length limit in svnbot. Have to remember that. The rest of the log reads: It is written in Inline::C, and would deploy as a cpan XS module. 20:35
ingy++ # Inline::C 20:36
20:43 SamB joined
svnbot6 r9227 | putter++ | Regexp-ReplaceEngine.pm: Added link to bsb's pure-perl $1,etc setter, and a "Risk" that a source filter using it might be an easier way to go. 20:48
20:50 cdpruden joined, larsen joined 20:55 vborja joined 21:17 Arathorn joined 21:35 DesreveR joined 21:43 frederico_ joined 22:45 DesreveR is now known as r0n|mess 23:17 pdcawley joined
mauke < Yaakov> I think perl6 should be spec'd in French, the Lingua Franca of programming language specification. 23:30
svnbot6 r9228 | rootmj++ | * probably fixed 23:37
mauke wow, that was fast 23:38
mj41_ examples/tutorial_gen/ probably fixed :-) ... good night 23:39
23:42 elmex joined
svnbot6 r9229 | putter++ | Move Regexp-ReplaceEngine.pm from Common to Aside as insuffiently interesting. 23:43
r9229 | putter++ | Variable interpolation happens before the regexp compiler hook gets the pattern.So while this hook approach is fine for creating say a reentrant p5 engine, even one with rules, it can't handle p6 regexps with variables correctly.
23:45 jalex joined 23:49 xinming 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!