Think twice before running "make install" for Pugs | moritz.faui2k3.org/irclog/ | pugscode.org | sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse (show (scanl (*) 1 [1..] !! 4)) | "Perl 6 Today" video from YAPC::Asia: xrl.us/v6op
Set by agentzh on 6 May 2007.
00:23 dduncan joined 00:24 dduncan left, obvio171 joined 00:28 dduncan joined 00:42 b00t joined 00:45 nipotaway is now known as nipotan 00:49 amnesiac joined 01:13 dolmans joined 01:33 japhb joined 01:50 mako132_ joined 01:52 yumiko joined 02:05 justatheory joined 04:09 SubStack joined 04:25 vel joined 04:44 [particle] joined 05:12 ofer joined 05:19 Alias_ joined 05:33 SubStack joined 05:59 [particle] joined 06:09 marmic joined 06:29 rindolf joined 07:03 jamhed joined 07:14 iblechbot joined 07:23 jisom joined 07:25 obvio171 joined 07:28 obvio171 joined 07:29 BooK_ joined 07:36 obvio171 joined 07:44 dmq joined 07:47 SubStack joined 07:57 obvio171 joined 08:01 Belaf joined 08:04 obvio171 joined 08:11 buetow joined 08:29 andara joined 08:46 dduncan left 09:02 obvio171 joined 09:03 riffraff joined
riffraff ciao 09:03
09:10 baest_ joined 09:11 baest joined 09:13 shady joined 09:17 franck__ joined 09:32 b_jonas joined 09:33 obvio171 joined 09:53 obvio171 joined, polettix joined 09:56 obvio171 joined
dolmans test(1,2,3) may differ from test (1,2,3), is it a bad design or not? 10:06
riffraff I think 10:07
dolmans language designing is a very hard homework, tossing to those talents. 10:10
riffraff ops sorry 10:12
wrong chan
10:15 obvio171 joined 10:18 obvio171 joined
dmq answer to fglock from yesterday: yes 10:22
riffraff is doesn't make sens that junctions work "sometimes" but not always, right?
ouch
how comes?
dmq $subrule=qr/.*/; /a (??{$subrule}) b/
or: /(?(DEFINE)(?<subrule>.*)) a (?&subrule) b/ 10:23
10:24 ruoso joined 10:25 obvio171_ joined 10:27 buetow joined
riffraff dmq, sorry, what is that? answer to me, dolmans or what? 10:32
10:34 jisom_ joined
dmq no thats in reply to fglock's question from last night/yesterday 10:36
riffraff oh sorry
dmq nothing to be sorry about. :-) 10:41
riffraff that is very ugly, if I put a junction in a variable and do is $j, $foo I get a result 10:42
but if I do is make_function(),$foo
I get another
is this expected behaviour?
10:44 hareldvd joined 10:52 b_jonas joined 10:54 chris2 joined
riffraff notices that junctions sometime don't have a proper .values() 10:56
11:00 buetow joined 11:17 fglock joined
svnbot6 r16224 | fglock++ | [PCR] - implemented <?{...}> and <!{...}> in backtracking mode 11:18
fglock dmq: thanks
avar fglock: that works? 11:19
fglock avar: yes - perl -e ' $subrule=qr/.*/; "axxxb" =~ /a ((??{$subrule})) b/x; print $1 ' 11:21
just to be sure it is not interpolating: perl -e ' my $subrule; my $rule=qr/a((??{$subrule}))b/; $subrule=qr/.*/; "axxxb" =~ /$rule/x; print $1 ' 11:23
svnbot6 r16225 | fglock++ | Pugs-Compiler-Rule-0.23 in CPAN 11:33
11:43 nipotan is now known as nipotaway 11:50 MarLaw joined
MarLaw Hello 11:50
11:52 zafoeta joined
zafoeta hi guys 11:52
moritz hi
MarLaw I'm with a friend, we've done some haskell at university and we like it, we're thinking of joining pugs if help is needed and i've did (a little) on Language Processors again at university
dmq no it doesnt interpolate. 11:54
thats the infamous bizarre parsing code that avar bumped into a while back.
and its worth noting that Perl5 and PCRE have different behaviour for named-recursion. 11:55
specifically in PCRE named recursion is atomic. wheras in perl5 much work was done to make it backtrackable.
of course we really need a superlinear cache for named recursion/(??{...})
but nobody seems to have the both the inclination and tuits. 11:56
you might find that translating subrules to named recursion brings a speedup... 11:57
moritz MarLaw: do you want a commit bit?
dmq iow: /(?(DEFINE)(?<subrule>$subrule)) a ((?&subrule)) b/x 11:58
MarLaw yeah that would be great moritz
zafoeta had one yesterday ...
11:59 obvio171_ joined
dmq that will interpolate, but to your benefit 11:59
11:59 obvio171_ is now known as obvio171
moritz MarLaw: just /msg me your email adress 11:59
dmq pokes fglock
zafoeta hey, i was reading yesterday the syopsis
there seems a lot to read
do we have to knows it first very good before we can help with any implementation coding ... ? 12:00
moritz MarLaw: btw on freenode you need to be registered to /msg... if you are not (and not keen on it either), just write it in the channel
MarLaw: unless you are afraid of spam, that is ;)
MarLaw moritz: i'm sending with /msg 12:01
not getting any ?
I think i'm registered
moritz MarLaw: no :(
MarLaw ok
'x'.$mynickname .'x@gmail.com' is enough antispam ? 12:03
moritz it is indeed ;) 12:04
zafoeta and what if @gmail is an array?
MarLaw what do you mean ?
moritz MarLaw: sadly you need a bit patience because commitbit.pugsocode.org times out :(
MarLaw: ignore him...
'..' is non-interpolating ;) 12:05
fglock dmq: i'm back
dmq did you see those comments i made? 12:06
fglock dmq: re subrules, I need them to be lexically redefinable, so I prefer not to inline
that is, no named recursion 12:07
moritz can anybody else invite MarLaw? I just get timeouts
dmq i dont see why it wouldnt be lexically redefinable.
moritz what the hell is going on on feather?
dmq its just a different approach to how you embed it.
the advantage of the named recursion approach being that it is optimizable by the compiler. 12:08
essentially its a question of latebinding versus early binding.
fglock dmq: isn't $subrule interpolated there?
dmq yes.
but thats ok. it shouldnt change for the duration of the match should it? 12:09
the thing that is special about (??{..}) is that the regex engine has to deal witht he possibility that the returned pattern could be different every time. 12:10
but if you are doing (??{ $some_qr }) then it doesnt change every time.
also (??{ ... }) has the potential for tricky bugs pre 5.10 12:11
fglock consider $rule = qr((?(DEFINE)(?<subrule>$subrule))a((?&subrule))b);
$rule will not be recompiled later 12:12
reinterpolated
dmq right, yeah if you are storing the rule then its different.
fglock ok - but I'll still consider the optimization when possible 12:16
dmq actually, it makes me think maybe we could do something to deal with scenario. 12:18
*maybe*
fglock hmm - I can't update v6.pm on CPAN; depends on Module::Compile and Sub::Multi fixes :( 12:19
12:22 riffraff joined
riffraff hello 12:22
is it possible to define my own subroutine traits in current pugs?
12:22 sonorous joined
riffraff I see I can do sub bar is foo {} and it always eval's correctly but I don't think it does anything 12:23
fglock riffraff: you can use it with a 'role'; this is being refactored, not sure how much works 12:24
riffraff so I just follow S12's trait description ? 12:25
moritz MarLaw: commit bit is finally sent, it will take a few minutes until you receive an email 12:27
MarLaw ok
thanks moritz
moritz no problem ,)
12:29 iblechbot joined
fglock I fixed the v6.pm Makefile.PL on CPAN, but I didn't update the code 12:33
@tell audreyt: have you registered the 'Pugs' namespace in CPAN? I need comaint in order to have the v6.pm modules indexed 12:35
lambdabot Consider it noted. 12:36
avar MarLaw: welcome:) 12:37
fglock @tell audreyt: actually, the problem is specific to Pugs::Grammar::P6Rule - you are the owner; I need comaint 12:39
lambdabot Consider it noted.
12:42 buetow joined 12:47 JarJarBinks joined 12:48 kanru joined 12:51 vel joined
moritz MarLaw: did you receive an email? 12:51
MarLaw nope 12:55
moritz damned 13:00
fglock @tell clkao: would you upload Sub::Multi to CPAN? I've made a small fix
lambdabot Consider it noted.
moritz fglock: can you try to invite MarLaw? connection times out for me :( 13:01
fglock trying...
zafoeta moritz: if i want to start looking at code of pugs, where do i better start of? i am reading the synopsis and want to see some examples to learn perl6. i just know perl5 13:04
also, i am quite new in Haskell 13:05
13:05 Khisanth joined
moritz zafoeta: first, check out pugs repository... 13:06
in docs/ there are a few interesting things...
zafoeta i have ti checked out already
moritz mainly docs/Pugs/Doc/
fglock zafoeta: take a look at the tests, under t/ 13:07
zafoeta just have to install the ghc 6.6
ok ..
moritz then examples/ is a good place to start
zafoeta any mac user here?
riffraff zafoeta, yep 13:08
need help to install ghc6.6?
zafoeta i had installed ghc 6.4
riffraff remove it
zafoeta updated to 6.6 and get an error
fglock MarLaw: sorry, it looks like the invitation page is not working at the moment 13:09
MarLaw no problems, i'll get it from zafoeta
riffraff I had some problems too, my solution was to remove ghc 6.4 completely then install the binaries from audreyt
zafoeta how do i remove it? sorry for the stupid question
MarLaw his the man
he's 13:10
zafoeta i am actually sitting with Lawrence in the room
:)
riffraff how did you install it? I used the .dmg package IIRC
for 6.4
otherwise I'm afraid you have to hunt it around the file system
zafoeta i used the .dmg, but it seems it has a problem removing the libraries 13:13
i think it has to do with the update i did 13:14
riffraff I see
don't nkow how to help then, I'm sorry
zafoeta anybody knows a good haskell editor for os x? 13:21
and perl6 editor for os x? 13:22
gnuvince zafoeta: vim or Emacs?
lumi zafoeta: I use emacs with haskell-mode
moritz zafoeta: vim and emacs both have perl 6 modes, but I think the vim one is better supported atm 13:23
zafoeta i see, i prefer vim then. but does vim have haskell mode as well? 13:24
rindolf zafoeta: it does, sure.
zafoeta cool 13:25
moritz vim has a mode for anything you can think of - nearly ;-) 13:26
ls -1 /usr/share/vim/vim70/syntax/|wc -l
482
Patterner Are there other editors than TextMate on OS X? 13:29
avar emacs 13:30
13:36 Casan joined
moritz avar: Patterner was asking for editors, not for OSes ;) 13:36
*SCNR*
Patterner "Emacs is a good OS, but it lacks a decent editor"
svnbot6 r16226 | fglock++ | [PCR] - fixed serialized form of '^^' '$$' 13:39
13:43 chris2 joined 13:44 jerrylee joined
riffraff moritz, I have something that vim doesn't have a mode for 13:44
NesC
13:44 [particle] joined
riffraff a C-dialect for embeded systems with concurrency & component-based development 13:45
(I had to write my small syntax.vim when doing my bachelor thesis :)
moritz riffraff: well, commit it to perl.org ;) 13:48
riffraff: of course there are many things that vim doesn't hilight...
riffraff not many really
:)
moritz but when I work with it, most files are colorfull ;) 13:49
[particle] it even highlights the need for aid in uganda
;)
moritz ;) 13:50
andara riffraff: there is a nesC mode ... 13:51
riffraff wow, it must have appeared in the last year 13:52
how lovely :)
andara riffraff: 2004, Joao Girao [email@hidden.address]
13:54 kunwon1 joined
riffraff mh 13:54
I don't have it on my vim7
andara riffraff: www.vim.org/scripts/script.php?script_id=899 13:55
lambdabot Title: nesC - nesC syntax highlighting : vim online
riffraff oh I see, exists but not included
moritz and www.vim.org/scripts/script.php?script_id=1847 13:57
lambdabot Title: NesC Syntax Highlighting - The Syntax Highlighting for NesC Files : vim online
[particle] nesc: c for nintendo? 13:59
wouldn't that be something dangerous.... 14:00
moritz are PIR and PASM syntax files uploaded to perl.org? 14:06
14:08 veritos joined 14:09 justatheory joined
svnbot6 r16227 | fglock++ | [PCR] - fixed closure call-as-method 14:19
[particle] uploaded to perl.org? 14:26
parrot has vim syntax files under editors/
so, i guess you can get them from svn.perl.org/parrot/trunk/editor/
moritz [particle]: s/perl\.org/vim\.org/ 14:27
[particle]: that's where I'd look first for syntax files
[particle] ah, no. not that i know of.
14:29 fglock joined, veritos joined
zafoeta sorry for bothering again about this problem 14:33
has any of you experienced this error when leaunching ghc Library not loaded: GMP.framework/Versions/A/GMP ?
precisely 14:34
$ ghc
dyld: Library not loaded: GMP.framework/Versions/A/GMP
Referenced from: /usr/local/lib/ghc-6.6/ghc-6.6
Reason: image not found
Trace/BPT trap
14:39 amnesiac joined
[particle] is there some other package you need to download? 14:40
i haven't installed ghc in a year or two, but i seem to recall needing something more than just the ghc base 14:41
however, it looks funny, since you're just calling ghc 14:42
i'd assume it should Just Work
moritz is there no packaging software (like apt) for mac os that does the dependency handling? 14:44
veritos Fink?
written in perl too. 14:45
riffraff macports is nice
and has ghc6.6 IIRC
14:48 DrSlump joined
DrSlump hi 14:48
moritz hi DrSlump ;) 14:51
DrSlump hello Moritz :) 14:52
moritz DrSlump: I'm case sensitive ;) 14:53
14:56 lisppaste3 joined 15:02 bonesss joined 15:16 crashmatrix joined 15:29 fglock joined
svnbot6 r16228 | andara++ | [runpugs] -Slight hitch: no Moose on feather. So installed it locally. 15:40
r16228 | andara++ | -New release going live RSN.
offby1 ooh ooh 15:41
15:41 penk joined
offby1 moose don't have feathers! 15:41
TimToady oh, horsefeathers... 15:45
svnbot6 r16229 | fglock++ | [v6.pm] - minor fix in hash declaration
fglock the 'unspace' rule from STD now loads in perl5 without warnings; doesn't mean it works, yet 16:03
avar fglock: It looks pretty easy-ish to turn PCR into a Parse::Yapp sort of thingy, you'd define a grammar and it would emit p5/p5-rx/XS which you could then ship with your module without any Pugs:: deps 16:10
fglock: is that a fair assessment?
fglock avar: sure 16:11
avar kewl
fglock actually, just use v6.pm ?
16:11 andara left
avar nah, it would be better if there were no deps aside something that implemented the match object maybe.. 16:12
fglock avar: see Rule2.pm in the PCR source; it's PCR own grammar
it is compiled into Rule.pmc (the Rule2.pm name avoids the circular dependency at installation)
avar: there is an old script that does this; I think it's somewhere in the misc/pX directory 16:15
avar yeah, that's exactly what I mean 16:16
fglock ah, here: compile_p6grammar-token.pl and compile_p6grammar.pl in perl5/PCR/
not sure if it still works 16:17
16:17 sonorous left
avar I really like the src->ast->some_emitter architecture:) 16:17
fglock avar: I'll move the scripts to perl5/PCR/util/ ok? 16:21
svnbot6 r16230 | fglock++ | [PCR] move the grammar-compiler scripts into util/ 16:24
r16231 | fglock++ | [PCR] - cleanup base directory 16:30
fglock avar: feel free to modify those grammar compiler scripts 16:32
now taking a look at Ook :) 16:34
16:36 kunwon1 joined 16:37 obvio171 joined
avar fglock: it's incomplete:), very 16:38
16:42 obvio171_ joined, obvio171_ is now known as obvio171 16:43 elmex joined 16:44 kanru joined
TimToady Juerd: I am amused that feather's /var/log has files that were last modified in 2012 and 2016... 16:45
I suspect a wormhole
gnuvince @seen audreyt 16:46
lambdabot audreyt is in #perl6 and #haskell. I don't know when audreyt last spoke.
TimToady maybe sabotage from future python programmers time travelers
that might explain why jifty/commitbit suddenly started taking 45 cpu seconds per transaction... 16:47
gnuvince: at the moment audreyt is preoccupied with Real Lifeā„¢ 16:48
moritz what's feathers load average? 16:49
TimToady about 9
gnuvince TimToady: thanks.
moritz TimToady: does comitbit work for you at least?
TimToady I did eventually get a commitbit sent to MarLaw
moritz great
TimToady (a correct one, with x on both ends) 16:50
but I had to wait till the loadavg dropped to 5
[particle] what are those numbers in, fahrenheit? i never understood that scale 16:51
TimToady currently backups are running
loadavg? 16:52
[particle] yeah.
gnuvince [particle]: if memory serves rights, it's the number of CPU secs a process has to wait before being processed
TimToady er, no...
gnuvince the three numbers show the average for the last 1, 5 and 15 minutes.
TimToady it's a time-decayed average of number of processes wishing to use the CPU right now 16:53
[particle] ah
moritz ... and other resources apart from CPU iirc
[particle] do processes typically get an equal timeslice?
moritz [particle]: that depends on their "nice" value 16:54
TimToady depends on the work profile
if several processes are completely CPU bound, and have the same nice, they should share equally
but IO always screws things up
moritz bad IO ;) 16:55
[particle] everybody hates io
moritz too
TimToady when I run top on feather, the load avg seems artificially high compared to the number of processes in R state, so it seems that it's counting in IO waits somehow. maybe processes that are done with IO are counted but don't show as R yet... 16:56
moritz TimToady: sometimes zombies contribute to the load avg as well 16:57
TimToady or maybe the algorithm has changed in the last 20 years or so.
zombies shouldn't be putting any load on the CPU 16:58
they just occupy table space
moritz TimToady: yes, but sometimes they do increase the load avarage parameter by one 16:59
TimToady I'd call it a bug if so
moritz can't be, linux is perfect[TM] ;) 17:00
TimToady ā„¢
17:00 veritos joined
moritz whatever ;) 17:01
TimToady I wish I could get my compose key to use vim digraphs...
17:01 bernhard joined
veritos TimToady: i wish i could bind keystrokes to Perl keywords 17:01
moritz veritos: that should not be too hard 17:02
veritos moritz: i know, i'm trying to write something that does that
TimToady vim has macros, but irssi isn't vim...
veritos #i<stdio.h> \n imain(v){printf("Hello");r0;} 17:03
TimToady and last time I tried changing the compose key table, it had no effect at all. apparently there's a mysterious compilation step known to the cognizenti 17:05
and my compose key doesn't, in fact, even do everything that the table I'm supposedly using specifies, but I think that's input method damage, probably 17:07
doesn't recognize any triple key sequences for instance
so it won't do em-dash, among other things 17:08
veritos TimToady: so you use all of those nifty << >> brackets and stuff?
TimToady Ā« and Ā» are the only latin-1 chars we use currently 17:09
and we don't use them as heavily as we used to
we got rid of Ā„
moritz is that 'zip' now?
TimToady zip() or 17:10
Z
veritos can't we still use as a type of quote mark? q<<foo>> or whatever?
TimToady sure
?eval say Ā«a b cĀ»
moritz can't you define that as a macro?
17:11 evalbot_r16219 is now known as evalbot_r16231
evalbot_r16231 OUTPUT[abcā¤] Bool::True 17:11
TimToady ?eval say qĀ«a b cĀ»
evalbot_r16231 OUTPUT[a b cā¤] Bool::True
veritos ah, i see.
moritz circumfix:<qĀ«...Ā»> or something?
TimToady the q language is productive and extensible. a macro wouldn't be, at least not without complicating it 17:12
[particle] "q language"... can you make it turing complete? 17:13
TimToady ?eval say q:c Ā«here is { 1+2 }Ā»
evalbot_r16231 OUTPUT[here is 3ā¤] Bool::True
moritz [particle]: all perl DSLs have to be turing complete ;)
dmq whats wit the hl thingee on the output 17:14
TimToady I just added 1+2 inside single quotes
with the help of :c
Tene dmq: ā¤ you mean?
moritz dmq: it's N/L and stands for new line
Tene that's "NL" for "newline"
dmq ah
thats a unicode character is it? 17:15
TimToady indeed
dmq interestink
:-)
TimToady ā¤ 2424 SYMBOL FOR NEWLINE
dmq so the question is does it match \n :-) 17:16
TimToady I surely hope not. :)
Tene ?eval 'ā¤' ~~ m/\n/
evalbot_r16231 Match.new(ā¤ ok => Bool::True, ā¤ from => 0, ā¤ to => 0, ā¤ str => "", ā¤ sub_pos => (), ā¤ sub_named => {}ā¤) 17:17
TimToady crap
it's emphatically *not* whitespace
dmq heh. 17:18
Tene TimToady: from 0 to 0
moritz and str => "", but ok => Bool::True
TimToady that's just evalbot character class braindamage
Tene also, it's Bool::False on my local pugs
TimToady mine too, whew!!!
I think something in PCR is deemed 'unsafe' 17:19
but fails "gracefully" instead of not matching
Tene We could just stop running evalbot in safe mode. What's the worst that could happen? ;)
TimToady system "rm -rf /" 17:20
moritz Tene: feather doesn, pugs repository lost
s/doesn/down/
TimToady how 'bout run in chroot dir?
moritz or even worse, somebody could implement a trojan horse in one of the per/haskell files...
and compromise all developer's maschines ;) 17:21
TimToady well, if someone does that we're basically hosed anyway
fortunately most script kiddies don't know Haskell...
moritz TimToady: yes, but without a trojan horse we can read svn diffs...
TimToady "can", but don't generally 17:22
moritz yes
TimToady I read maybe 1 in 50
usually when something tickles my "that's odd" detector 17:23
which could well be tickled by a trojan
17:23 fglock joined
TimToady 'course if fglock installed a trojan we'd never know... 17:24
moritz TimToady: same for you ;)
veritos at least code in perl 5 is somewhat readable. perl 5 core c code, now...
dmq its readable. 17:32
just your eyes glaze very very quickly... 17:33
veritos tokenizer code = cancer in rats. so says Camel. 17:34
TimToady yeah, it's readable like I read Chinese. I recognize most of the radicals, but haven't the foggiest idea what it means...
Juerd TimToady: Hmmm... :) 17:36
TimToady: Interesting. Probably wasn't ntp synched at some point
TimToady: During bootup, after its crash 17:37
TimToady: I knew its system clock was screwy, but not this screwy :)
TimToady I just wonder whether some locking scheme used by jifty/commitbit relies on a stable clock, is all... 17:39
but it's a long shot
and if it were that far off, you'd think the transactions would just fail 17:40
rather than taking 45 CPU seconds
unless it's a discretionary lock that is getting forced after some period of time 17:41
haven't got up the gumption yet to try to strace it...
heh, didn't know "gumption" came from Scots... 17:43
I guess a lot of them settled in Appalachia 17:46
17:50 [particle1 joined 17:51 electrogeek joined
TimToady 恗悃恂恗悃恂 & 17:54
17:56 Psyche^ joined 18:06 marmic joined 18:14 Psyche^ is now known as Patterner 18:15 MarLaw joined, SubStack joined 18:21 franck__ joined 18:23 MarLaw joined 18:38 Aankhen`` joined, mncharity joined
fglock mncharity: hey 18:49
18:56 oceanlab joined, justatheory joined 19:01 obvio171 joined
TimToady MarLaw: did you get your commit bit? 19:02
MarLaw i did! 19:03
TimToady previous try left out the second x...
moritz MarLaw: it is a custom that you add yourself to AUTHORS to test the commit bit ;) 19:04
MarLaw moritz: Wouldn't that be premature ?
TimToady sure, that's why it's enforced by custom :) 19:05
moritz MarLaw: no, we encourage new committer to do it ;)
19:05 jisom joined
MarLaw ok ;P 19:05
[particle1 once you've committed, you are an author :)
TimToady we're trying to encourage a culture of brazen committers, after all
19:08 rindolf joined 19:09 obvio171 joined 19:13 obvio171 joined 19:18 obvio171 joined
jisom other than the "FreeBSD/amd64 registerised running" thread on the ghc mailing list, has anyone heard anything since about ghc on FreeBSD/amd64? 19:19
19:24 obvio171 joined, chris2 joined
TimToady maybe you're the sucker...er, the pioneer... 19:27
jisom judging by what I read, I'm not going to try :) 19:28
others have tried and failed to get that far......someone finally succeeded in getting that far, but nothing publicly usable 19:29
moritz there's a good reason that pugs will not be the only and final implementation of perl6 ;)
lumi jisom: They tried and failed? 19:30
jisom lumi: some did, poor abstraction
was linked to the regex stuff
lumi I guess it's a bit obscure 19:31
jisom had a similar segfault in parrot recently, though, sizeof(int) != sizeof(void*)
19:31 obvio171 joined
TimToady maybe someone should write a good regex engine in Perl 6. :) 19:33
jisom I use PGE for the complicated stuff, and hand crafted for the rest.......one time wrote a module that parsed http uploads using hand crafted regex, just without the expressions 19:34
moritz where can I find a good tutorial "how to write a good regex engine"? *g*
19:34 weinig joined
TimToady I know lots of ways not to... 19:34
jisom perldoc -f index && perldoc -f substr 19:35
a lot of regexes do nothing that couldn't be done with one index call 19:36
19:37 zafoeta_ joined, riffraff joined
riffraff hi 19:37
moritz hi riffraff ;) 19:38
jisom: but regexes are just more convenient ;))
jisom but index/substr is so much faster and obscure :) 19:39
19:39 MarLaw joined, obvio171 joined
moritz is it compariable in speed to strstr? 19:40
index, i mean
19:40 crashmatrix joined
riffraff I have a spellchecker.p6 with both regex-based and substr-based impl, so much fun :) 19:40
19:42 SubStack joined
riffraff especially because I have to eval the regexen all the time so it takes ages to work 19:42
19:44 obvio171 joined
fglock jisom: what are you working on? 19:44
19:50 Loro_ joined
jisom nothing at the moment, too much work 20:00
20:04 dduncan joined 20:08 veritos joined 20:15 b_jonas joined 20:17 rashakil__ joined 20:21 justatheory_ joined 20:22 prism joined 20:32 obvio171 is now known as obvio171_, veritos joined 20:33 obvio171_ is now known as obvio171 20:38 dduncan left 20:58 dolmen joined 21:00 marmic joined 21:02 dolmen left 21:13 japhb joined 21:20 IllvilJ1 joined 21:21 lichtkind joined 21:24 weinig joined 21:28 polettix joined 21:29 Aankhen`` joined 21:39 IllvilJa joined 21:58 bazbax joined 22:00 larsen_ joined 22:30 Belaf joined 22:35 veritos joined 22:57 dduncan joined 23:00 zafoeta joined 23:02 polettix joined 23:11 SubStack joined 23:12 veritos joined
veritos TimToady: would you even consider having syntactically significant indentation in Perl 7? or dropped semicolons? 23:34
avar veritos: due to the design of perl6 you don't need a new language for those features 23:35
you just change the language grammar in your scope and call it Perl6::Pythonic or something
veritos avar: no, i mean Perl 7 as the dream perl.
riffraff dropped semicolons... yeeees
veritos or is Perl 7 like TeX pi, that will never come around until Larry kicks the bucket? 23:36
and then Perl 6 becomes 7 automatically?
TimToady I think (or at least hope) that Perl 6 gives the community the tools to decide what language they like best over the long haul. 23:46
if we get the grammar and versioning stuff right, all the rest is negotiable 23:47
polettix TimToady: is this what you meant with "rewrite of the community"? (IIRC) 23:48
TimToady no, I originally thought the community would do the redesign. the community tried, it was obviously not going to work. 23:56
too many things depended on too many other things, and all the proposed solutions had tunnel vision 23:57
dduncan when did the community try to do the rewrite?
TimToady in the several months after the announcement in 2000
a lot of mailing lists were set up...
dduncan afaik, there were meetings between people, and an RFC process for what people want in the next Perl, then TimToady took that and wrote Apocalypses from them
maybe that explains why there are so many Perl 6 language sublists that don't seem to be in use 23:59
TimToady indeed
all you have to do is read through the 361 RFCs to realize what a mess it was.
and how much bigger a mess it would have been if we'd adopted the various solutions piecemeal

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!