Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2007-05-02

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 [particle] it is SLOOOOOWWWW
00:00 [particle] in fact it just timed out
00:00 TimToady same here
00:00 TimToady so maybe something on feather end
00:01 [particle] likely
00:01 TimToady there are now 4 copies of commitbit running on feather
00:01 [particle] what are the anglicized japanese words for plan and ok?
00:01 TimToady 3 in R state
00:02 [particle] eew
00:04 TimToady well, "puran" and "ookei" are both understood, but of course they have lots of other words
00:06 geoffb How does a word containing r and l together get converted?  For example, how would a native Japanese speaker pronounce "url" or for that matter "perl"?
00:08 pasteling "[particle]" at 144.81.84.216 pasted "japanese imports ;-)" (19 lines, 494B) at http://sial.org/pbot/24640
00:09 geoffb [particle]: You switched them?  Evil.
00:09 [particle] parrot++
00:09 [particle] i'm writing an example of subclassing the Exporter PMC i recently created
00:09 TimToady depends on whether the r is vocalic or not.  English vocalic r just results in lengthening so that perl comes out "paaru"
00:10 TimToady I don't know how they pronounce "url"
00:11 prism joined #perl6
00:12 veritos joined #perl6
00:14 TimToady 'course, you should really spell them プラン and OK to match typical usage...
00:15 [particle] my irc client no likey unicode
00:15 veritos [particle]: what client
00:15 [particle] trillian
00:16 veritos ah, i see. mirc maybe a better bet?
00:16 DarkWolf84 it's fine in my client
00:16 [particle] is that win-friendly?
00:16 veritos yeah
00:16 DarkWolf84 I use Xchat
00:16 [particle] hrmm, does it do aim/yahoo/msn as well?
00:17 veritos [particle]: no :( meebo.com for those tho
00:17 DarkWolf84 no but it do irc well
00:21 DarkWolf84 I don't know katakana but I think プラン is pronounce paru not paaru
00:23 TimToady that's "plan"
00:23 veritos left #perl6
00:23 DarkWolf84 oh fine my mistake
00:24 TimToady paaru would be パール
00:25 DarkWolf84 i saw the little circle on the first chana
00:25 TimToady but almost certainly spelled "Perl" since it's a proper name
00:25 DarkWolf84 and make mistake
00:26 bonesss joined #perl6
00:27 DarkWolf84 and made mistake*
00:27 DarkWolf84 my english again
00:28 DarkWolf84 May be is's time lo learn katakana
00:29 bonesss`out changed the nick to bonesss
00:35 TimToady basically, there are three makes running on feather right now...
00:35 TimToady so the load avg is bouncing between 10 and 17
00:36 TimToady and it appears that jifty can't wake up commitbit fast enough (or vice versa) to respond in what firefox thinks is a timely fashion
00:39 TimToady agentzh++ for running his make niced
00:57 TimToady well, I reniced karhu's background make, and the load avg dropped to 9, but after 50 cpu seconds commitbit still can't make my firefox happy, so there must be some spinlocking or some such contention going on
00:57 TimToady I note that there seems to be two base copies of commitbit running, so maybe one is blocking the other
01:00 TimToady I'm giving up for now.  Need a commitbit expert...
01:00 [particle] obra: ping
01:00 [particle] hrmm, actually i think it was audrey who did that coding
01:06 |OUF|HiTech69 joined #perl6
01:14 rissy joined #perl6
01:16 drupek12 joined #perl6
01:25 SubStack joined #perl6
01:29 audreyt looking into it (commitbit).
01:31 * davidfetter waves to audreyt
01:31 audreyt hi.
01:33 * obra really needs to find the cycles for another round of commitbit hacking
01:33 obra I removed 30% of the code last week. about  3 more hours and I'll have something that hrts less than what we have today
01:39 rhr re pipes: I'm thinking $io = Pipe.to/from: "cmd" for auto-fork+exec; ($r, $w) = Pipe.pair for no-fork; ($proc, $r, $w) = Pipe.fork for auto-fork without exec.  Is that a good interface?
01:42 avar is there any debugger for perl6 yet?
01:43 audreyt avar: no, currently there's only pugs -d which is more of a tracer than a debugger
01:43 avar ;(
01:44 audreyt a debugger for Pugs.Eval can be done rather painlessly by introducing a new argument to envEval
01:44 audreyt i.e. swap out the current tracer with something that resembles the interactive shell
01:48 laye joined #perl6
01:56 avar are there any modulen in p6 on cpan? how would I go about adding some?:)
01:56 avar language: in META.yml would be neato:){
01:57 audreyt :D yes.
02:00 geoffb audreyt: back just to fix commitbit, or back for a while?
02:05 Yaakov audreyt: Hello!
02:06 Yaakov audreyt: Will you be at YAPC::NA this year?
02:08 audreyt Yaakov: unfortunately no. will be at oscon though.
02:08 Yaakov Ah. Well, that's a disappointment.  I was hoping to see you.  Maybe next year.
02:08 gnuvince_ audreyt: Quick note, the YAPC::Asia presentation was amazing
02:08 audreyt geoffb: yes, back just to fix commitbit... maybe back for a while. but need to get some sleep first (3hr.)
02:09 audreyt gnuvince_: thanks. I was actually having a nervous breakdown of sorts toward the end
02:09 audreyt but the presentation slides themselves is I'd say not bad :)
02:09 geoffb Yes, please, do sleep!
02:09 gnuvince_ Really?  I must admit I did not notice.  I guess the material was too absorbing :)
02:09 Yaakov audreyt: I hope you are well.  I watched your very compact YAPC::Asia video.  I was enjoyable.
02:09 audreyt danke.
02:10 Yaakov Heh, "it", of course.
02:10 rhr indeed, that was a great talk
02:10 Yaakov Though I hope I am also enjoyable occasionally.
02:10 gnuvince_ I made a joke on your blog about the number of slides you have in your presentation, but I thought that it actually helped.
02:10 gnuvince_ It really "flowed" well
02:10 audreyt oh yes. it's all about the subtle subconsciousness hints and brainwashing
02:10 gnuvince_ :)
02:11 audreyt glad you enjoyed it :) really glad :)
02:11 * audreyt restarts apache and hope it fixes commitbit
02:11 Yaakov audreyt: Sleep well.
02:21 rhr I was also thinking it would be cool if there were some way to specify arbitrary io redirection for things like system, Pipe.to/from, etc. without having to do close/dup/fork/whatever by hand, or getting into weird sh syntax
02:21 nipotan changed the nick to nipotaway
02:22 rhr something like: system "cmd", :ioredirect( :stdin($some_io), :stderr('/dev/null'), :fd42($secret_channel, :r) )  Is that too unwieldy?
02:41 geoffb joined #perl6
02:54 mako132_ joined #perl6
03:03 justatheory joined #perl6
03:17 REPLeffect joined #perl6
03:32 awwaiid joined #perl6
03:39 stevan__ joined #perl6
03:46 kunwon1 joined #perl6
04:07 quantboy joined #perl6
04:09 agentzh joined #perl6
04:10 agentzh Hmm, i see 4 broken smartlinks appear over the night. anyone here willing to offer a quick fix? ;)
04:14 agentzh auto-smoking on feather tends to take many hours recently :(
04:15 agentzh Juerd: 1:00 ~ 4:00 AM (feather localtime) is a very busy time period for the feather world?
04:29 rhr another IO question: is there a way to do io to/from a variable (both scalars and lists)?
04:31 rhr example: $decrypted_data = Pipe.from "gpg --passphrase-fd 3", :ioredirect( :fd3( IO.fromstr($passphrase) ) )
04:53 goban joined #perl6
04:54 Alias_ joined #perl6
04:55 SubStack joined #perl6
04:58 agentzh left #perl6
05:00 rhr yet another: IO.tee(IO) to link output streams, e.g. $*ERR.tee($logfile)
05:10 dduncan joined #perl6
05:34 Southen_ joined #perl6
05:43 devogon joined #perl6
05:49 nelhage joined #perl6
05:59 rfordinal joined #perl6
06:04 SubStack_ joined #perl6
06:07 iblechbot joined #perl6
06:10 SubStack changed the nick to SubStack_
06:22 marmic joined #perl6
06:26 CindyLinz changed the nick to CindyLin1
06:27 TimToady rhr: we will probably end up doing most such IO redirections using feed operators.
06:27 TimToady (commitbit finally sent, by the way)
06:31 rhr got it, $thanks xx 1e6
06:31 rhr how would that work with a feed?
06:38 ozo joined #perl6
06:39 TimToady io("filespec") ==> $variable kinda things
06:40 Alias_ TimeToady: For a second there I thought you'd gotten involved with CommitBit :)
06:41 TimToady nah, just wondering why feather wasn't keeping up, mainly.  as soon as the going got tough, I quit...
06:42 rhr ah, that's a better way to do that.  what do you think of the whole :ioredirect thing?
06:43 TimToady seems a bit heavy
06:44 TimToady but doubtless something like that is going on underneath, even if we have the notation convenience of feed ops
06:45 moritz TimToady: yesterday I saw a pice of rubbish on eweek.com (iirc) titled "the 100 most influential people in it", you were no. 100 ;)
06:45 TimToady basically, the feeds can give us directionality, but can't establish the type of the io handle
06:46 TimToady yeah, I'm just famous for almost being not famous
06:46 buetow joined #perl6
06:48 moritz TimToady: which means you don't have to spend too much time in interviews, so you can do more perl stuff ;)
06:50 rhr actually, I don't quite see how your feed example would work in my gpg example.  How would you declare an IO object that reads from a scalar, but could be passed as an fd to a child process?
06:53 elmex joined #perl6
06:54 rhr or should you have to do the piping explicitly?
06:54 TimToady my guess is that feeds can handle the ordinary connection of one thing's stdout to the next thing's stdin, but you'd have to use adverbs to modify stderr and extra fds
06:55 TimToady io(...) ==> run("cmd") ==> $result for instance
06:55 franck__ joined #perl6
06:55 rhr yeah, that would get rid of the clunky syntax in most cases
06:55 TimToady io(...) ==> run("cmd", :err</tmp/foo>) ==> $result for instance
06:56 TimToady something like that
06:56 TimToady alas, my eyes are going crossed
06:57 TimToady so I should probably turn in for the night
06:57 * Alias_ accidentally encountered the $foo = { $^a }; notation today in a Perl6:: module, is that still in the Perl 6 language?
06:57 TimToady yes
06:57 Alias_ where it makes the { } block into sub ($a) { $a }
06:57 Alias_ It feels ... wrong...
06:57 TimToady well, except not a real sub
06:57 Tene Alias_: howso?
06:58 TimToady more like -> $a { $a }
06:58 Alias_ or inelegant, or something
06:58 Alias_ Tene: I'm not entirely sure
06:58 Alias_ I think possibly that I could see them getting too seperated form the subject of their impact to remain not-confusing
06:59 Alias_ $foo = { $^a ... 25 lines of code ... $^b }
06:59 TimToady the subject of their impact?
06:59 SubStack joined #perl6
06:59 TimToady "doctor, it hurts when I do this..."
06:59 Alias_ Their impact is to transform the block to a sub and modify the argument list
06:59 TimToady placeholders are for little blocks, not big blocks
06:59 TimToady they're for when the order of params is obvious
06:59 Alias_ When was the last time people didn't take a good idea way too far :)
06:59 TimToady if it hurts to use them, use -> instead
07:00 Alias_ Lets just say I could see MYSELF being tempted into abusing them thusly :)
07:00 TimToady when was the last time I prevented them from it?
07:00 TimToady :)
07:00 Alias_ I think also that maybe it was that there's two effects, not one
07:00 TimToady I think the culture will enforce a reasonable balance here
07:00 Alias_ the sub'ification and the argument addition
07:01 Alias_ Honestly, I find it hard to put a precise argument in place
07:01 TimToady no, it's already a block
07:01 Alias_ ah
07:01 Alias_ maybe it was the docs then
07:01 TimToady $a = { ... } is just the same, only defaults to $_
07:01 Alias_ Seemed to imply a double effect
07:01 Alias_ And that without a $^a the { } was not a sub
07:01 TimToady it's not really a sub with either...
07:01 Alias_ (in perl5 parlance)
07:01 TimToady if you return from it, it doesn't see those curlies
07:02 Alias_ One of these days I need to sit down and read the Perl 6 syntax stuff end to end
07:02 Tene Alias_: you can do horribly twisted things with macros, too.
07:02 TimToady shh!
07:02 Alias_ macros just seem like syntactic sugar for currying to make C coders more at home
07:02 TimToady Lisp, actually... :)
07:02 Alias_ right
07:03 * Alias_ is also learning a little Oz atm
07:03 Tene Alias_: if you want the language to prevent you from doing awkward things when you explicitly want to, perl isn't the language for you. ;)
07:03 Alias_ I couldn't help thinking that Oz needed TimToady's linguistic touch to it
07:03 Alias_ It's nice, but it's terminology feels hideous to me
07:03 TimToady the main reason P6 has macros is that P5 has source filters.
07:03 Alias_ ya
07:04 Alias_ And don't think I'm not supremely greatful we're killing source filters
07:04 Alias_ grateful
07:04 TimToady and a common p6 theme is reducing the scope of damage in every dimension possible, including grammatical
07:04 Alias_ ya
07:04 Tene Hmm... could we implement source filters in p6? ;)
07:04 TimToady if something doesn't scope right, it's probably indicative of a missing scoping concept
07:05 TimToady shh!
07:05 TimToady .oO(don't anyone tell Tene that you could implement "use Python" in p6...)
07:08 rissy joined #perl6
07:09 TimToady well, now my mouth is flopping open, so
07:09 TimToady zzz &
07:19 ozo joined #perl6
07:23 jisom joined #perl6
07:25 rhr to answer my own question, I guess it would look like my IO $pphr = io <== $passphrase; $decrypted_data = Pipe.from: "gpg --passphrase-fd $pphr.fileno()";
07:26 rhr so maybe there's no need for more sugar
07:31 BooK joined #perl6
07:32 franck__ joined #perl6
07:39 bsb left #perl6
07:42 polettix joined #perl6
07:49 goban guys i want to learn perl, but should i start with perl 5 or wait til 6 comes out?
07:49 Alias_ Perl 5
07:50 goban also i know there is no official date but is there an estimate how long?
07:50 wolverian yup. it won't be a wasted effort.
07:50 goban k thanks
07:50 moritz goban: nobody knows when perl6 will be final, so don't wait until then
07:50 Alias_ goban: Next yearish
07:50 Alias_ not this christmas, but some christmas
07:50 moritz cause every day will be christmas once p6 is out ;)
07:51 goban fed up with java after starting a game server in it :(
07:51 moritz well, servers are not always the best start ;)
07:51 Alias_ hmm
07:52 moritz but learning perl is never a bad idea ;)
07:52 Alias_ I think I might go home early today and work on my repository issues
07:52 goban yeah was a tough first project but it runs many games and has a good GUI :)
07:52 Alias_ goban: Perl is basically a "get stuff done" language
07:52 qmole joined #perl6
07:52 Alias_ no nonsense, no pretentions, fast development, not always as low level as you'll need for some tasks
07:52 moritz goban: a good place to start with many links is this one: http://perl-begin.org/
07:52 lambdabot Title: Perl Beginners' Site
07:53 Alias_ But Perl isn't afraid to let you use other languages where appropriate
07:53 goban yeah watching the perl 6 today video and excited about classes, fp, being able to use haskell/other stuff ...
07:53 Alias_ Speaking of which I need to get Oz.pm working
07:54 Alias_ or released or something
07:54 Tene goban: well, quite a bit of Perl 6 is available now.
07:56 dduncan Alias_, what does that do?
07:58 Alias_ dduncan: Wrapper for the Mozart Oz compiler
07:58 Alias_ Which is a fairly neat multi-paradigmic language that seems like a cleaner and more usable language for logic/constraint programming than Prolog :)
07:59 Alias_ And also apparently has built-in network-transparent parrelism support
07:59 dduncan okay
07:59 moritz sounds quite cool ;)
07:59 Alias_ I have a constraint problem in a project I'm working on, and Perl isn't really going to cut it
08:00 moritz yaltl - yet another language to learn
08:00 Alias_ So I'm looking for a good language to write logic programs in
08:00 Alias_ moritz, I try to only learn languages when I have a concrete problem I can't solve effectively in any languages I already know
08:01 Alias_ I started "coding" with HTML, then learnt JavaScript when I hit the limits of HTML, then learned Perl when I hit the limits of JavaScript
08:01 Alias_ Ignoring some basic of several variations in childhood
08:01 Juerd No, PM
08:01 Alias_ ?
08:02 Juerd Eh
08:02 Juerd Is for agentzh, but he isn't here so the nick completion didn't work :)
08:02 Alias_ ah
08:02 * Alias_ googles for "programming language PM" and got www.wall.org/~larry/pm.html
08:03 moritz Alias_: I learned a wide variety of languages, partly because I needed them, or other sucked for the job I wanted to do, or because they were used in school/university
08:03 Alias_ I try to avoid learning new languages as such
08:03 rfordinal_ joined #perl6
08:03 moritz Alias_: and I quite like learning new languages - as long as I have enough time ;)
08:03 Alias_ I just try to get enough of a feel to slot them into my mental database of "what things are good for"
08:06 Alias_ hrm
08:06 * Alias_ fixes YAML::Tiny bugs
08:06 Alias_ Empty Comments: check
08:06 Alias_ Hash Keys With Spaces: check
08:06 Alias_ Now I just have yet another module that works on my laptop but isn't released due to repository-fu
08:07 franck__ joined #perl6
08:11 buetow joined #perl6
08:25 rfordinal changed the nick to rfordinal_
08:35 pbuetow joined #perl6
08:41 andara joined #perl6
08:59 demq joined #perl6
09:06 pbuetow joined #perl6
09:22 chris2 joined #perl6
09:41 dduncan left #perl6
09:43 buetow__ joined #perl6
09:43 zamolxes joined #perl6
09:49 penk joined #perl6
09:57 pbuetow joined #perl6
10:04 rissy joined #perl6
10:06 iblechbot joined #perl6
10:27 rissy joined #perl6
10:29 rissy joined #perl6
10:31 rissy left #perl6
10:58 ofer joined #perl6
11:04 agentzh joined #perl6
11:05 flyby5 joined #perl6
11:12 rindolf joined #perl6
11:29 rfordinal joined #perl6
11:35 nipotaway changed the nick to nipotan
11:40 VanilleBert joined #perl6
11:46 VanilleBert left #perl6
11:54 ruoso joined #perl6
11:56 buetow joined #perl6
11:56 flyby5 The news is that our beloved Head of Freenode Staff _christel_ works together with the anti Rob Levin network OFTC now to fight against trolls.
11:56 flyby5 And all your donations are used to pay a lawyer, so that Freenode can be a home of script kiddies like trelane, while people with justified complaints are removed.
12:01 Limbic_Region joined #perl6
12:03 Limbic_Region salutations all
12:04 fglock joined #perl6
12:11 Limbic_Region ?seen audreyt
12:11 lambdabot audreyt is in #haskell and #perl6. I last heard audreyt speak 9h 59m 54s ago.
12:11 Limbic_Region ok, good
12:13 moritz oh, lambdabot recognizes commands starting with '?'?
12:13 moritz that's rather new, isn't it?
12:19 svnbot6 r16158 | fglock++ | Perl-6.0.0-STD.pm - fixed a few mismatched brackets
12:22 svnbot6 r16159 | fglock++ | [PCR]
12:22 svnbot6 r16159 | fglock++ | - regex grammar: added <rule: string...> and <rule( params, ... ) syntax
12:22 svnbot6 r16159 | fglock++ | - new ast node 'call'
12:22 svnbot6 r16159 | fglock++ | - rules can have positional parameters
12:22 svnbot6 r16159 | fglock++ | - "Whatever" {*}
12:22 svnbot6 r16159 | fglock++ | - fixed ratchet quantifier - rollback .pos on fail
12:22 svnbot6 r16159 | fglock++ | - added an initial double-quoted parser
12:22 svnbot6 r16159 | fglock++ | - allow aliasing to a literal
12:22 svnbot6 r16160 | fglock++ | [v6.pm]
12:22 svnbot6 r16160 | fglock++ | - grammar - added prefix:<|>
12:22 svnbot6 r16160 | fglock++ | - added: proto token rx { ... }
12:22 svnbot6 r16160 | fglock++ | - fixed signature parsing
12:22 svnbot6 r16160 | fglock++ | - added: Module::CompileV6
12:22 svnbot6 r16160 | fglock++ | - fixed unicode << ... >> parsing
12:22 svnbot6 r16160 | fglock++ | - allow [ ... ] as regex delimiter
12:23 drupek12 joined #perl6
12:24 moritz fglock: are you working on STD.pm in order to use it in kp6?
12:25 stef_ joined #perl6
12:29 fglock moritz: there are several possibilities,
12:29 fglock STD -> pugs is one likely target,
12:29 iblechbot joined #perl6
12:31 fglock but STD -> v6.pm -> perl5 could work
12:32 fglock ( STD -> kp6 -> mp6 -> Any ) would be nice, but the kp6 project is not  good enough yet
12:32 lisppaste3 joined #perl6
12:33 moritz from what I've read the regex integration of pugs is not so very well developed atm
12:35 fglock the plan is to desugar regexes into perl6 code,
12:35 fglock like mp6 does
12:36 fglock "token" is a macro in mp6
12:36 moritz so you want to use kp6 for desugaring, so that pugs can parse its output?
12:40 Zaba joined #perl6
12:40 Zaba No. In the foreseeable future all Perl 6 programs will require the resources of the Parrot engine. (in official FAQ)
12:40 Zaba is it true?
12:41 fglock will likely use PCR for parsing, and write a PCR-Pugs backend
12:41 Zaba hm ok
12:41 fglock Zaba: don't know
12:42 moritz Zaba: since there can be an arbitrary number of perl6 implementations that statement is just a guess
12:43 Zaba hm
12:43 Zaba is the test suite included in official pugs distribution?
12:43 fglock v6.pm currently chokes at two methods in STD; the regexes parse ok
12:44 moritz Zaba: it's not ready yet, and it lacks generallity...
12:44 Zaba isn't v6.pm for perl5?
12:44 moritz Zaba: but it's in the pugs repository under t/
12:44 Debolaz joined #perl6
12:44 Zaba okok
12:44 fglock Zaba: yes, it is a perl6-in-perl5
12:44 Zaba and i'm about pugs
12:46 fglock re STD: the problems are in 'method heredoc' and 'method EXPR'
12:48 fglock Zaba: 'pugs' is perl6-in-haskell, 'Pugs' is the project which includes many other goodies
12:48 Zaba yeah
12:51 fglock moritz: re STD - but running 'regex' macros in pugs requires some libraries, that are not written yet
12:51 moritz fglock: ok, thanks
12:52 rindolf Hi all.
12:53 rindolf Hi fglock, moritz, Zaba
12:53 Zaba moin
12:56 rindolf Zaba: moin moin
13:00 stevan__ joined #perl6
13:00 DarkWolf84 joined #perl6
13:05 moritz hi rindolf ;)
13:06 rindolf moritz: what's up?
13:06 rindolf I'm going out to bike soon.
13:07 fglock ?eval macro token ( $a ) { ' "macro returned " ~ ' ~ $a } token '123'
13:08 evalbot_r16157 changed the nick to evalbot_r16152
13:08 evalbot_r16157 "macro returned 123"
13:08 diakopter joined #perl6
13:09 moritz rindolf: not much, having vacations at my girlfriend's...
13:11 * Zaba slaps GHC around with a industrial-strength trout with nuclear fishhead, super ultra dragon plus(tm)
13:11 moritz Zaba: did ghc survive?
13:11 moritz ;-)
13:11 Zaba moritz, i hope
13:11 Zaba NO
13:28 Zaba sux
13:36 miyagawa joined #perl6
13:37 Zaba why does ghc lag so much?
13:39 moritz what do you mean by "lag"?
13:39 moritz that it takes long to compile something?
13:39 moritz how much memory does your computer have?
13:43 Limbic_Region moritz - perhaps you missed Zaba's departure
13:50 fglock left #perl6
13:52 moritz Limbic_Region: I did... I /ignore joins and leaves, that's too much line noise IMHO ;)
13:54 Zaba joined #perl6
13:56 franck___ joined #perl6
14:02 mj41_ Hi, so I definitely lost my commit passwd . User name rootmj ( rootmj_konf <a_t> seznam.cz ). Audrey once tried to changed it, but it didn't work. Any sugestions?
14:02 vel joined #perl6
14:03 moritz do you want a new username?
14:03 Arathorn joined #perl6
14:05 veritos joined #perl6
14:06 mj41_ yes, i don't care. I prefer mj41
14:06 mj41_ you can user email rootmj <a_t> seznam.cz , thanks
14:07 veritos we giving you a commit bit?
14:07 moritz why not?
14:07 veritos i'm just asking if that's what's happening.
14:08 Zaba hmm
14:08 Zaba ghc sucks
14:08 veritos Zaba: why now
14:08 moritz I'm trying, but I can't reach commitbit.pugscode.org
14:08 moritz track seems to be down as well :(
14:08 Zaba veritos, lags the whole system
14:08 moritz Zaba: not enough memory?
14:09 Zaba moritz, no, just does
14:09 Zaba even on few-liners
14:10 [particle] moritz: timtoady and i haven't been able to reach commitbit since yesterday
14:10 kolibrie I wonder if that is part of the stuff Juerd disabled - though I think he was only trying to disable trac
14:10 [particle] even via http://svn.pugscode.org:9999 it's timing out
14:10 buetow joined #perl6
14:10 moritz mj41_: then I can't help you atm, sorry :)
14:12 kolibrie I am almost getting something for http://svn.pugscode.org:9999, but feather's load is at 14, so it's not finishing
14:13 moritz now I reached the web page, and submitted the data
14:13 moritz mj41_: please tell me if an email reached you
14:14 veritos anyways, welcome to our little project.
14:20 fglock joined #perl6
14:22 Zaba pugs: user error (*** *** Can't locate strict.pm in @*INC (@*INC contains: /usr/lib64/perl6/x86_64-linux /usr/lib64/perl6 /usr/lib64/perl6/site_perl/x86_64-linux /usr/lib64/perl6/site_perl /usr/lib64/perl6/auto/pugs/perl6/lib /usr/lib64/perl6/site_perl/auto/pugs/perl6/lib .).
14:22 goban joined #perl6
14:23 Zaba mh
14:23 vel6608 joined #perl6
14:23 Zaba is it wrong to use strict;?
14:23 ayrnieu seems like a reasonable Perl6 error.  What's wrong?
14:23 ayrnieu zaba - 'use v6;'
14:23 Zaba is it wrong to use strict; in perl6?
14:23 Zaba ayrnieu, i'm using pugs, do i really need?
14:23 kolibrie Zaba: yes, strict is on by default
14:24 kolibrie 'use strict' will fail
14:24 Zaba okay
14:24 Zaba no need to use v6; in pugs?
14:24 Zaba it's aready v6 -,-
14:25 kolibrie it is nice to specify 'use v6' or 'use v6-alpha' so that you know it is a perl 6 program
14:25 kolibrie required, of course, if you are using v6.pm
14:25 ayrnieu Which also allows perl5s to fail nicely, after perl6 starts to go by binaries named 'perl' and people shebang accordingly.
14:26 ayrnieu he set it on fire as a sacrifice to the gods of sticking to a set budget.
14:27 ayrnieu mischan.
14:27 Zaba hm
14:27 Zaba ayrnieu, does it need pugs in shebang atm?
14:28 ayrnieu It needs whatever binary you installed, which is probablt 'pugs'.
14:28 amnesiac joined #perl6
14:28 kolibrie unless you call it as 'pugs script.pl'
14:28 Zaba hm ok
14:29 kolibrie since I use pugs and v6.pm, I rarely put in the shebang line
14:34 franck__ joined #perl6
14:38 riffraff joined #perl6
14:38 riffraff hi
14:38 riffraff someone has an idea on why if I do "use Test" from interactive pugs I get *** user error (***
14:38 riffraff        Unexpected "(@"
14:38 riffraff ?
14:39 moritz riffraff: set PERL6LIB correctly
14:39 Zaba riffraff, i got undef
14:39 mj41 joined #perl6
14:40 moritz that is, to blib/lib/ of the pugs repository
14:40 Zaba pugs 6.2.13 though
14:40 riffraff moritz, was this needed two weeks ago? I don't remeber ever doing that
14:42 riffraff and I gues is blib6/lib, right?
14:42 Zaba is 6.2.13 old?
14:43 moritz riffraff: no, it was not needed
14:43 moritz and blib6/lib/, yes
14:43 riffraff Zaba, I am on the development version, I think you are on the last release
14:43 Zaba riffraff, o'rly?
14:44 riffraff moritz, I see thank you
14:44 amnesiac joined #perl6
14:44 moritz np
14:47 veritos joined #perl6
14:51 zamolxes joined #perl6
14:51 riffraff would it make sense if I implement Str.subst as a dummy "make local copy, use s//" and commit it?
14:52 riffraff it is annoying that it is missing but I suck at haskell hacking more than I do at perl's, and this is telling a lot
14:54 TimToady eventually we'll want everything implemented in Prelude.pm, even if circular
14:54 TimToady then various implementations can optimize more or less of that into lower-level code
14:54 TimToady and break the circularity
14:55 TimToady nothingmuch's "circularity saw"
14:55 TimToady so that would be fine
14:56 riffraff yay
14:57 Zaba joined #perl6
15:01 Zaba shall i change my isp?
15:03 Zaba eh..
15:03 Zaba hmm
15:03 Zaba is there socket library for perl6?
15:03 Zaba yet*
15:03 Zaba or..
15:04 riffraff by the way, I remeber a Pattern role, is there also a Replacement thingy, covering the things that can be used in substitution (or maybe substitution-like things that I can't imagine now) ?
15:06 Zaba hmm
15:06 Zaba how does s/// work in perl6?
15:06 TimToady Zaba: you can get at most p5 libraries by saying "use perl5:Socket" currently
15:06 Zaba ah ok
15:06 Zaba TimToady, it's IO::Socket btw
15:06 Zaba thanks
15:07 TimToady s/// works the same as long as you don't use brackets
15:07 TimToady the bracketed form wants a pseudo assign for the 2nd half
15:07 TimToady s[foo] = "bar"
15:07 Zaba ok
15:08 TimToady ?eval $_ = "1 2 3"; s:g[\d+] += 42; $_
15:08 evalbot_r16160 changed the nick to evalbot_r16157
15:08 evalbot_r16160 \"1 2 3"
15:08 TimToady hmm, yeah, that was broken for some reason...
15:09 TimToady ?eval $_ = "1 2 3"; s:g[2] += 42; $_
15:09 vel joined #perl6
15:09 evalbot_r16160 \"1 44 3"
15:10 TimToady something wrong with \d+ in evalbot, probably thinks generating Unicode aware \d is unsafe
15:11 TimToady ?eval $_ = "1 2 3"; s:g[<[0..9]>+] += 42; $_
15:12 evalbot_r16160 \"1 2 3"
15:12 TimToady ?eval $_ = "1 2 3"; s:g[ [ 0 | 1 | 2 | 3 | 4 ]+ ] += 42; $_
15:13 evalbot_r16160 \"43 44 45"
15:14 TimToady hmm, now that I recall, it was more on the order of not returning the match pos correctly
15:14 TimToady ah well
15:15 svnbot6 r16161 | karhu++ | Greatly fixed perlintro.
15:18 svnbot6 r16162 | mj41++ | [util/PaP6-testing]
15:18 svnbot6 r16162 | mj41++ | * fix some bugs, some cleanup
15:18 svnbot6 r16162 | mj41++ | * die -> croak
15:18 svnbot6 r16162 | mj41++ | * document verbose level
15:31 goban joined #perl6
15:32 xern joined #perl6
15:34 moritz karhu++ #good job
15:59 riffraff is it possible to make a regex into a string?
15:59 riffraff interpolation just gives "<Regex>"
16:00 moritz perhaps .perl?
16:00 moritz doesn't work atm in pugs
16:01 moritz ?eval my $a=regex {\d+}; $a.perl
16:01 evalbot_r16160 "\\\{regex}"
16:02 riffraff because I don't know how to use a regex object in s, so I was trying s/$rgx/rplcement/ but it cannot interpolate.. thought of converting to str
16:02 riffraff now I understand the problem is much deeper
16:02 avar ?eval my $a = token { \d+ } ; ("123" ~~ $a).perl
16:02 moritz avar: that's just the Match object
16:03 avar ya
16:03 evalbot_r16160 *** Cannot parse regex: :ratchet(1)[ \d+ ]␤*** Error: ␤"Match.new(\n  ok => Bool::False, \n  from => 0, \n  to => 0, \n  str => \"\", \n  sub_pos => (), \n  sub_named => \{}\n)"
16:04 moritz ?eval my $s='\d+'; m/<$s>/ ~~ "123"
16:05 riffraff I think a regexen object doesn't hold a copy of the string it was created from
16:05 riffraff but it should, imo
16:05 evalbot_r16160 *** Cannot parse regex: <$s>␤*** Error: ␤Match.new(␤  ok => Bool::False, ␤  from => 0, ␤  to => 0, ␤  str => "", ␤  sub_pos => (), ␤  sub_named => {}␤)
16:05 moritz riffraff: at least its AST
16:06 riffraff I'd prefer the string, so it is eval'able again
16:06 riffraff like numbers or strings
16:06 riffraff but maybe .perl would be enough
16:07 riffraff and it is just my ruby background showing :)
16:07 TimToady in p6 regexen are not string, they're programs
16:08 TimToady if you want to use another regex in yours, use /<$rgx>/
16:08 TimToady bare $rgx is always a literal match
16:08 moritz TimToady: so .perl shouldn't reproduce everthing?
16:08 TimToady so it wouldn't do what you want anyway
16:09 TimToady it should, but there's little point in decompiling and recompiling when you can just use a ref
16:09 riffraff makes sense
16:09 TimToady arguably, treating regex as strings was the biggest design mistake in Perl 1..5
16:09 moritz so does m/<$regex>/ only works when $regex is a regex, or should it work with a string as well?
16:10 TimToady it should work if it's a string
16:10 TimToady don't know if it's implemented tho
16:10 moritz ok
16:11 TimToady ?eval my $pat = "abc"; $pat ~~ /<$pat>/
16:11 avar TimToady: I guess you'll be happy to know that I've recently been fixing your screwups in 5.10 :)
16:11 moritz it seems <$foo> is not implemented at all in pugs
16:11 avar regexes can be objects now
16:11 avar (or any scalar)
16:11 TimToady but not the assumption that $foo interpolates before regex compilation
16:12 TimToady I carefully avoided the multiple interpolation bugs of shell scripting, only to fall into the same trap with regexen
16:12 evalbot_r16160 *** Cannot parse regex: <$pat>␤*** Error: ␤Match.new(␤  ok => Bool::False, ␤  from => 0, ␤  to => 0, ␤  str => "", ␤  sub_pos => (), ␤  sub_named => {}␤)
16:14 avar TimToady: yup, but you can do "str" =~ sub { $re => "(ook|eek)" } I guess
16:14 avar but ya, I haven't changed the parser/tokenizer
16:16 bernhard joined #perl6
16:17 TimToady biab &
16:23 veritos joined #perl6
16:25 cognominal joined #perl6
16:27 justatheory joined #perl6
16:31 riffraff looking at src/Pugs/Eval.hs: I can see a reduceSyn "subst" doesn't this mean that this should be actually parsed ?
16:31 cognominal joined #perl6
16:40 cognominal joined #perl6
16:44 cognominal joined #perl6
16:47 concept11 changed the nick to ofer
16:47 ofer changed the nick to concept11
16:54 lichtkind joined #perl6
16:56 franck__ joined #perl6
17:06 fglock joined #perl6
17:07 kanru joined #perl6
17:08 cognominal joined #perl6
17:08 Debolaz joined #perl6
17:11 sapir joined #perl6
17:13 fridim is there an argument of the choice of haskell over ocaml (more perf?), or did you choose it just because you love it more ? :)
17:13 moritz fridim: audreyt implemented pugs as a haskell exercise first ;)
17:13 riffraff is there an obvious way of calling pugs' internal functions fro perl6? I can use Pugs::Internals::sprintf, for example, but how do I call mkRegex defined in RRegex.hs
17:14 fridim moritz, ok :)
17:18 Zaba hmm
17:19 Zaba snowing.pl from examples doesn't work here
17:19 Zaba pugs 6.2.13
17:19 Zaba ah
17:19 Zaba does
17:19 Zaba just Slooow
17:20 justatheory joined #perl6
17:22 gaal riffraff: you... don't, but if you need it for exploratory purposes just add something to Prim.hs
17:23 riffraff yeah I was thinking of that
17:23 Zaba pugs is so slow =(
17:23 Zaba it idles when i start some script for about 1/3 of minute eating CPU a lot
17:24 gaal riffraff: re: reduceSyn, that's a hacky thing. Something (Parser.hs) had constructed a Syn node and reduceSyn is the evaluator moosing it
17:24 Zaba then begins to slowly execute it
17:24 Zaba any ideas why?
17:24 gaal Zaba: do you have a blib6/lib/Prelude.pm.yml file?
17:24 gaal also, did you build with "make soon"?
17:24 gaal or just "make"?
17:24 sapir what's make soon?
17:24 Zaba gaal, i installed it from portage
17:24 riffraff point is: if I want to implement X in perl6 I need Y (available in haskell) at the perl level. Otherwise I can implement X at the haskell level but then we have less portability no?
17:24 gaal make soon happens faster but gives a much slower pugs
17:25 * Zaba greps logs
17:25 sapir make fast?
17:25 gaal sapir: renamed to make soon, because it built a _slow_ pugs :)
17:25 gaal riffraff: well, yes, that's the tradeoff.
17:25 sapir ah, I think in 6.2.13 it's still make fast
17:25 gaal probably, yes.
17:26 Zaba is 6.2.13 very old?
17:26 sapir because make soon doesn't work here :)
17:26 gaal kinda.
17:26 Zaba well
17:26 sapir also that would explain why it's slow for me too :)
17:26 Zaba build was VERY time and mem/cpu consuming
17:26 Zaba i'm not sure if it was fast soon whatever
17:26 gaal Zaba: how much free RAM does your machine have?
17:27 Zaba gaal, 768mb
17:27 gaal pugs can easily eat half a gig at build time
17:27 Zaba and the build was _veeeery_ mem/cpu consuming
17:27 gaal oh, another tip: use a snapshot GHC 6.7
17:27 Zaba so even GNU screen _lagged_
17:27 gaal well, builds that aren't CPU consuming aren't doing their jobs...
17:27 sapir and I have 256megs . . .
17:28 Zaba ghc seems to be an excessive hog here
17:28 Zaba even on fewliners
17:28 gaal sapir: you *might* get away with an optimized build, but you'll have to tweak the heap size in config.yml
17:28 Zaba 03:33:43 < gaal> Zaba: do you have a blib6/lib/Prelude.pm.yml file? <-- yeah it loads Prelude
17:28 gaal and don't run anything else at the time :)
17:28 riffraff gaal, about reduceSyn: we are er.. abusing the parser to parse a method call? Anyway, does it actually work? :)
17:29 gaal riffraff: dunno, what does smoke.pugscode.org say? :)
17:29 riffraff ?eval subst("hello",/h/,'f')
17:29 Zaba it's even slow on saying it can't find something in @*INC
17:29 gaal Zaba: okay, but note it won't find it if you cd elsewhere
17:29 sapir do I have to make clean to fix whatever make fast did, or is just making enough?
17:29 evalbot_r16162 changed the nick to evalbot_r16160
17:29 evalbot_r16162 Error: No such subroutine: &subst
17:29 riffraff not working :)
17:29 gaal sapir: better make clean.
17:30 Zaba time pugs -MNonexisting
17:30 Zaba real    0m2.244s
17:30 Zaba user    0m2.120s
17:30 Zaba sys     0m0.060s
17:30 sapir ah well
17:30 gaal that isn't a meothd call. :)
17:30 Zaba it's use Nonexisting;
17:30 Zaba and it takes THAT long
17:30 sapir so ghc_heap_size is the size of the heap that ghc uses when it compiles?
17:31 Zaba time perl -MNonexsisting
17:31 Zaba real    0m0.006s
17:31 Zaba user    0m0.004s
17:31 Zaba sys     0m0.000s
17:31 gaal sapir: yes.
17:31 Zaba that's a REAL difference!
17:31 riffraff gaal, well, it does'nt work the other way either IIRC
17:31 riffraff ?eval "hello".subst(/h/,'f')
17:31 sapir so if I set it lower it uses less memory? won't it run out?
17:31 evalbot_r16162 Error: No such method in class Str: &subst
17:31 gaal Zaba: there have in fact been improvements in pugs in that area. on my machine time pugs -e1 is a little under a second now. sure, could be much better. and should, eventually.
17:31 moritz riffraff: did you commit your changes?
17:32 riffraff I don't have changes :)
17:32 Zaba gaal, yes it should be
17:32 moritz riffraff: ok :(
17:32 Zaba and it's not now
17:32 gaal sapir: it'll do its best to GC more aggressively
17:32 Zaba perl5 lacks a lot of perl6 features i need
17:32 sapir ok
17:32 Zaba though perl6 is unusuable
17:32 gaal sapir: thus avoiding swap, which if you hit, you might as well leave the computer and go on vacation
17:32 riffraff I need to either: understand how to use a pattern object with s, or pull VSubst (I think) to the perl level
17:33 gaal riffraff: looking
17:33 Zaba question is, what should i use before christmas?
17:33 sapir gaal: practically what I did when I compiled it. closed basically everything (including X) and left it going for the day
17:33 gaal Zaba: that's up to you, your goals and the environment you're running in.
17:34 riffraff Zaba, perl with some perl6 modules ?
17:34 riffraff s/perl$/perl5
17:34 gaal sapir: pugs encourages longer vacations :)
17:34 sapir gaal: ah. it's a feature :)
17:34 Zaba no
17:35 gaal sapir: seriouslier now, if you're planning on a lot of Hs dev work, you pretty much want lots of ram.
17:35 SubStack joined #perl6
17:36 gaal otherwise - use make soon, or be content with pugses others build...
17:36 gaal and, try a ghc 6.7 -- creates faster pugsed faster
17:36 gaal but don't build your own ghc ;-)
17:37 gaal Zaba, sapir: bottom of this page: http://www.haskell.org/ghc/dist/current/dist/
17:37 lambdabot Title: Index of /ghc/dist/current/dist
17:37 sapir ok next question: I run make, it gets up to PIL2JS which tells me the 'name' method doesn't exist and to rerun Makefile.PL, which I'm pretty sure I've already done
17:37 gaal unpack in ~ , cd to the dir, follow the README
17:38 gaal uh, I think you can ignore PIL2JS
17:38 gaal for now.
17:38 sapir I mean the make tells me
17:38 gaal does the make die?
17:39 sapir yup. I'm trying make -k now
17:39 gaal if it happens again nopaste the log please
17:40 sapir if so, it'll probably be after I clear out everything, get ghc from that link, and get pugs from svn
17:41 gaal ok. while you're at it maybe "make realclean" or even wipe out the working copy and re-co
17:43 gaal riffraff: I'm getting the feeling that Syn "substr" is a fossil
17:43 gaal making sure...
17:44 Psyche^ joined #perl6
17:44 gaal ah, no. substLiteral
17:45 gaal P.P.Literal:626
17:45 riffraff 626?
17:46 gaal line
17:46 gaal src/Pugs/Parser/Literal.hs line 626
17:46 riffraff yes yes I was trying to understand what it means:)
17:46 gaal that's where the parser sees s/elk/moose/
17:46 gaal (or tr///) and puts the Syn "substr" thing
17:47 Zaba hmm
17:47 riffraff ok
17:47 gaal anyway: the reason your "substr" wasn't working is because you were talking about the other substr
17:47 gaal the function that has the same name in p5
17:47 Zaba ?eval my $s = "Hello wotld!\n"; $s =~ s/tl/rl/; print $s
17:47 evalbot_r16162 OUTPUT[<Subst>] Bool::True
17:48 Zaba doh i hate it
17:48 riffraff gaal, subst, no 'r'
17:48 gaal "~~"
17:48 gaal riffraff: well. no such prim...
17:48 riffraff exactly :)
17:48 gaal let me see if the prelude knows it
17:48 Zaba where did the old s/// behavior go?
17:48 riffraff no, it does not AFAICT, I was just trying to implement it on t of s///
17:49 Zaba ?eval my $s = "Hello wotld!\n"; print $s ~ s/tl/rl/; print $s;
17:49 evalbot_r16162 OUTPUT[Hello wotld!␤<Subst>Hello wotld!␤] Bool::True
17:49 gaal Zaba: "=~" is spelled "~~" now
17:49 riffraff but since I suck, I could not get how to make the magic happen and go down to pugs internal behaviour, the one used for s///
17:49 Zaba ?eval my $s = "Hello wotld!\n"; print $s ~~ s/tl/rl/; print $s;
17:49 evalbot_r16162 OUTPUT[1Hello world!␤] Bool::True
17:49 Zaba gaal, porting is pain.
17:50 Zaba ?eval my $s = "Hello wotld!\n"; $s ~~ s/tl/rl/; print $s;
17:50 gaal a lot of it is mechanical.
17:50 evalbot_r16162 OUTPUT[Hello world!␤] Bool::True
17:50 Zaba it seems different still to me
17:50 gaal this particular case will be taught to raise a warning...
17:50 gaal Zaba: what does?
17:51 Zaba gaal, nothing, and it's why 30% of people will stick with perl5 which 'is enough for their purposes'
17:51 Zaba s/nothing/everything/
17:51 gaal Zaba: that's okay
17:51 gaal took a while to transition from Perl 4, too.
17:52 Zaba which kind of interpreter does perl 5 use? and perl 6?
17:52 gaal what do you mean?
17:52 Zaba on kind i meant, how does it work
17:53 Zaba like, does it just go by character and see what's it, or uses some kind of flex/bison/antlr-like tools?
17:53 gaal Zaba: Perl 5 has a heavily hacked yacc grammar
17:53 Zaba oh
17:53 Zaba and pugs?
17:54 Zaba gaal, then why is it so fast?.. ;-)
17:54 gaal Zaba: Pugs has a parser built with Parsec, a Haskell library. But Perl 6 will eventually parse itself with Rules
17:54 Zaba compared to pugs even, perl5 is flying
17:54 Zaba to current pugs that is
17:55 gaal Zaba: speed isn't the top priority now.
17:55 Zaba gaal, then it's another reason to stick with perl5 ;-)
17:55 gnuvince -Ofun
17:55 gaal Perl 5 is based on code that's been tweaked for very, very long.
17:55 Zaba i see
17:56 Zaba gaal, i see, and rewriting everything from scratch in better way is better than tweaking on
17:56 Zaba though the question is, doesthe new attempt really work better, or at least same fast/wahever criteria as the previous?
17:56 gaal we have kind of a large feature set. some of it was very hard to do in c
17:56 Patterner changed the nick to Psyche^
17:57 Zaba gaal, C isn't really good for fast and advanced parsing/lexing
17:57 Zaba at least it might be fast, then code turns unreadable
17:58 gaal riffraff: okay, the thing is s/// works at the AST level and Prim looks like it doesn't -- but it can :)
17:58 gaal riffraff: what you do is create your Syn "substr" node
17:58 gaal then evaluate it, in Prim
17:59 gaal with evalExp
18:00 gaal riffraff: see for example Prim:315, try
18:00 gaal basically it guards the code with a exception moose, but the point is it runs it
18:00 gaal evalExp $ App (Val $ VCode sub) Nothing []
18:00 gaal that means,
18:01 gaal "evalExp (this function call)"
18:02 gaal you'll need something like evalExp $ Syn "substr" str subst adverbs
18:02 gaal look at the code in Literal.hs to see how to create default adverbs
18:03 riffraff I see but I don't think I need to do that, the parsing should be correctly handled by the current parser, I just need to fill in the code for the method, more like "say"
18:03 riffraff then try
18:03 riffraff or am I wrong?
18:03 * Zaba just wants to implement an irc bot in perl, with real man's OO, not perl5's
18:03 gaal Zaba: well, nobody's claimed pugs was the righteous heir to the throne. so far no other implementation has some close to pugs' coverage.
18:04 gaal Zaba: but that said, look into Moose.pm (perl 5)
18:04 rindolf Hi gaal
18:04 Zaba gaal, i did
18:04 Zaba and i decided to take a look at perl6 before diving into dirt
18:04 gaal riffraff: you are wrong :) what you want is to fire the evaluator code at runtime that deals with substitutions
18:05 gaal hello riffraff
18:05 gaal hello rindolf
18:05 gaal <- another tabcompletion casualty
18:05 rindolf gaal: hi hi.
18:05 rindolf gaal: are you at home now?
18:05 gaal riffraff: you're right in that you need no special parser support
18:05 gaal rindolf: I am.
18:06 rindolf gaal: nice.
18:06 Zaba cya
18:06 rindolf gaal: say hi to Foo for me.
18:06 gaal thanks
18:06 gaal rindolf: the point is to reuse the actual substitution implementation
18:06 gaal grrr
18:07 gaal riffraff: see above.
18:07 riffraff yep
18:07 sapir gaal: the advantage of mp6 etc. is? they bootstrap from perl5?
18:07 riffraff but would'n that cause misterious effects related to when the regexen are interpolated, when the code is executed and so on, that don't exactly match the semantics of the method?
18:09 gaal by the time you hit Prim, your rx is already parsed
18:09 riffraff ok
18:10 sapir gaal: or do they?
18:11 gaal sapir: you'll have to talk with fglock, but I think their idea is that they are very quickly Perl 6
18:11 sapir easier to complete p6 in a p6-like than in haskell?
18:12 gaal that is, there's the minimum possible p5 code in them and they are mostly implemented in successively more comlpex subsets of p6
18:13 gaal sapir: completing p6 take a lot of moose; possibly the parser will be easiest in p6. but the hs runtime does a lot of thing well, and gives it to pugs for free
18:13 gaal examples: GC, concurrency, io...
18:13 Casan joined #perl6
18:14 sapir so Parrot is intended to replace Haskell for that purpose?
18:14 gaal btw, our runtime is a hybrid one; we *also* have a perl5 interpreter available to do what it's good at
18:14 gaal for example, unicode stuff
18:15 gaal Zaba would have been happy; it takes 0 wallclock secs to load :-)
18:15 gaal well, parrot is one backed
18:16 gaal pugs can compile to several things
18:16 jisom joined #perl6
18:16 laye joined #perl6
18:17 integral hmm, what's the current status of pugs's compilation stuff?  Has it fully moved to the new internals yet, and is pir output broken or working?  Or, how do I find out these answers?
18:18 gaal integral: re PIR I don't know. re the other stuff, backlog ~two days ago
18:18 gaal a discussion between cmarcelo and myself
18:18 theorbtwo joined #perl6
18:20 integral thanks
18:20 gaal sure
18:23 buetow joined #perl6
18:25 bonesss joined #perl6
18:25 Arathorn left #perl6
18:34 rindolf gaal: are you still developing on Windows?
18:50 [particle] joined #perl6
18:51 atp joined #perl6
18:53 geoffb joined #perl6
19:00 Decker joined #perl6
19:19 Yappo joined #perl6
19:19 Limbic_Region joined #perl6
19:20 svnbot6 r16163 | rhr++ | First commit, added myself to AUTHORS
19:21 jisom joined #perl6
19:22 theorbtwo joined #perl6
19:29 Limbic_Region hey theorbtwo
19:41 svnbot6 r16164 | rhr++ | Rough draft spec of Pipe.to and Pipe.from
19:43 rhr TimToady: comments?
19:45 rhr what about my ideas for Pipe.pair and Pipe.fork?  should the latter return a (IO $r, IO $w) but True/False for parent/child?
19:49 TimToady it's a good start.  Pipe.fork doesn't feel right.  the objects in question are processes, and pipes just happen to connect them some of the time
19:50 TimToady (I also think solutions involving "but True" are a code smell showing the wrong class is being asked to do something.)
19:51 rhr agreed, I was hoping you'd come up with something better :)
19:52 rhr how about Pipe.pair?
19:53 benny joined #perl6
19:54 TimToady sure, that's just a wrapper around the pipe() syscall.
19:55 rhr ok, I'll add it.  does the interface (returns List of IO) sound good?
19:56 penk joined #perl6
19:56 TimToady good enough for now, but maybe we should have some official way of combining multiple one-way pipes into something more like a two-way socket pair.
19:56 rindolf Hi TimToady
19:56 TimToady rindolf: howdy
19:56 rindolf TimToady: what's up?
19:57 TimToady oh, the usual
19:57 rindolf TimToady: BTW, did you phrase the Perl Artistic licence yourself?
19:57 TimToady the original one?  yes.  it was written sloppily on purpose.
19:57 rindolf TimToady: hmmm...
19:57 penk joined #perl6
19:57 rindolf TimToady: why?
19:57 rindolf TimToady: the FSF considers it non-free.
19:58 TimToady I wanted corporate lawyers to be able to see 50 ways to wriggle out of it so they'd let their programmers use Perl
19:58 rindolf TimToady: heh
20:00 Aankhen`` joined #perl6
20:00 SubStack joined #perl6
20:01 wolverian has anyone gone to fsf about artistic 2, by the way? I think they still haven't updated it
20:01 svnbot6 r16165 | rhr++ | Add Pipe.pair
20:02 wolverian that is, they still have this: "The Artistic License 2.0 is currently in a draft phase with The Perl Foundation. ..."
20:02 rhr so we want some EntangledIO role, with .closer and .closew or some such?
20:05 rhr or do we just add options to IO.close?
20:05 TimToady phone&
20:06 rhr but if we use normal IO objects for that, .fileno doesn't know what to return...
20:10 veritos joined #perl6
20:23 veritos heya
20:25 riffraff in pugs internals is there a generic way to lift a val from a generic type Val to something specific like VStr ?
20:27 SamB riffraff: why don't you see what typeclasses VStr is an instance of?
20:30 riffraff bvecause I want to pass a Val that I know is a VStr to something that just wants that
20:31 riffraff but probable I'm approaching the thing from a wrong perspective :)
20:32 SamB what I mean is that any generic mechanism must be in a typeclass that the various value types, like VStr, are instances of
20:32 svnbot6 r16166 | rhr++ | Spec Socket.pair, analogous to Pipe.pair
20:33 Xach joined #perl6
20:33 riffraff dunno what to look for :)
20:35 veritos riffraff: search for a line "type VStr"
20:37 fglock joined #perl6
20:38 riffraff i was looking for instance  vstr
20:38 riffraff but nevermind, I sidestepped the problem :)
20:39 veritos left #perl6
20:41 pack|pizza joined #perl6
20:44 fglock hmm - I'm back to 'what to do next?'
20:44 Xach left #perl6
20:46 marmic joined #perl6
20:48 TimToady fglock: either pick the easiest thing, or the hardest thing, or something in the middle.  :)
20:49 fglock kp6-in-C with embedded perl5 seems to be -Ofun, for some weird reason
20:50 fglock I guess I like "fast", rather than "features"
20:50 fglock but "eval" in C is difficult
20:51 fglock kp6-in-parrot would be duplicated effort
20:51 fglock kp6-in-pugs is very slow
20:53 fglock STD-in-PCR-in-Pugs would need some serious backtracking-library implementation & profiling
20:53 riffraff feature I'd like today: variable interpolation in regexen :)
20:54 fglock riffraff: I don't think I can fix that; it depends heavily on the haskell/perl5 bridge
20:55 [particle] kp6-in-parrot should be faster, no?
20:57 [particle] fglock: think you could team up with pmichaud and implement something together?
20:57 fglock [particle]: re faster: not sure, kp6 implements OO at high level
20:57 [particle] it seems one path you haven't really pursued is multiple hackers on the same project :)
20:57 [particle] well, parrot implements oo at low level
20:58 [particle] you should be able to modify kp6 to take advantage of that
20:58 fglock re team: sure, but I'd like to keep an eye on 6-in-6
20:58 fglock rather than implement perl6 *with* parrot
20:59 [particle] okay, then the parrot emitter can be smarter about oo
20:59 fglock the PMC would need some extra goodies,
21:00 fglock like "modified" flags that kp6 needs at BEGIN time
21:00 [particle] parrot classes have an 'instantiated' flag
21:01 [particle] hrmm, i guess you want that for objects, though
21:01 fglock ok
21:01 fglock for containers
21:01 [particle] you could certainly build a Ref pmc that has a 'modified' flag
21:01 fglock but that's just a detail
21:01 [particle] s/Ref/Container/ # whatever
21:02 beppu http://lolgeeks.com/wp-content/​uploads/2007/05/lolgeeks016.jpg
21:04 fglock otoh, p6-in-parrot already has a plan and a history, I don't think it would reuse much of my work
21:05 gaal riffraff: does fromVal do what you want?
21:06 gaal it's monadic, so you need pureStr <- fromVal x
21:06 gaal used all over Prim
21:06 [particle] is running Test.pm enough to say you've implemented enough of p6 to build p6-on-p6?
21:06 fglock [particle] any ideas on how we can help each other?
21:06 fglock [particle]: no
21:07 gaal rindolf: did you mean pugs? rarely
21:07 [particle] i guess the problem with defining a subset of p6 is *which* subset
21:08 riffraff gaal, I'll check it pout, thank you
21:09 [particle] fglock: pm is planning a more kp6-like approach to parsing STD, iirc
21:09 [particle] you should probably email him, as he hasn't been reliably on irc lately
21:10 [particle] i'd be happy to be part of the conversation, if it'll help
21:10 gaal np
21:10 fglock btw, kp6 can be desugared to the current p6parrot, using lots of workarounds
21:11 [particle] i wonder if that's any faster than pugs
21:11 fglock which means it could run 6-on-6
21:11 gaal z&
21:12 fglock I'm not reliably on mail :) irc is easier
21:13 [particle] perhaps tomorrow, then. he's not available now
21:18 miyagawa joined #perl6
21:20 fglock [particle]: I think my problem with p6parrot is that it depends heavily on parrot,
21:20 fglock that is, you can't work around bugs
21:21 fglock in kp6 you have the emulation layer until everything is fine
21:22 fglock of course, this doesn't help much for parrot's development
21:23 fglock got to go home &
21:23 [particle] ok, we'll talk more later
21:26 SubStack joined #perl6
21:36 dduncan joined #perl6
21:42 autark joined #perl6
21:42 riffraff pouch, fromVal was so much easy to spot, I wonder why I missed it
21:48 riffraff what type declaration is x :: !Foo ?  
21:51 gaal riffraff: same as x :: Foo, except the field is strict
21:54 riffraff ok, I think I understand :)
21:55 gaal you can ignore it for most purposes
21:55 gaal basically it's an optimization thing
21:55 avar ook
21:55 gaal apart from some cases where it influences semantics
21:56 svnbot6 r16167 | avar++ | testing commit
21:56 gaal (for some algorithms it radically influences space behavior)
22:01 riffraff ok
22:01 drupek12 joined #perl6
22:04 jjore-w joined #perl6
22:07 ruoso joined #perl6
22:08 mr_ank joined #perl6
22:19 lichtkind joined #perl6
22:20 * avar fails at writing a module in p6
22:20 avar stupid build systems...
22:23 polettix joined #perl6
22:27 Limbic_Region joined #perl6
22:31 * Limbic_Region wonders if the bold green (TimToady) and bold red (audreyt) in the irclog coloring has anything to do with perl 6 coming out on Xmas
22:37 MindUs joined #perl6
22:37 MindUs Free phone calls all around the world ----->  http://callfree.point-serv.com/en/
22:37 lambdabot Title: Free phone calls all around the world | Free Calls
22:37 MindUs left #perl6
22:41 MindUs joined #perl6
22:41 MindUs Free phone calls all around the world ----->  http://callfree.point-serv.com/en/
22:41 lambdabot Title: Free phone calls all around the world | Free Calls
22:41 MindUs left #perl6
22:41 offby1 *sigh*
22:42 offby1 that guy was spamming #emacs a day or so ago too
22:42 justatheory joined #perl6
22:43 dduncan joined #perl6
22:52 gnuvince_ joined #perl6
22:52 cmarcelo joined #perl6
22:53 gnuvince_ Hey, weird thing, I'm trying to install pugs on my new Feisty Fawn installation, and I get this message: Can't locate v6.pm in @INC
22:54 gnuvince_ Any hint?
22:55 [particle] maybe install is broken?
22:55 qmole cpan install v6?
22:55 gnuvince_ qmole: cpan won't run because of the same error :)
22:55 [particle] you can install v6 from pugs, too
22:55 qmole not sure where you'd get that message, i thought v6 and pugs were separate
22:56 [particle] cd v6/v6-alpha or something
22:56 Tene gnuvince_: 'make install' is broken right now
22:56 Tene you've got perl6 stuff in your perl5 lib/site_perl/auto
22:57 Tene you'll need to remove it manually.
23:01 gnuvince_ Tene: so what do I do?
23:01 gnuvince_ reinstall perl5?
23:01 Tene gnuvince_: I just told you what you needed to do.
23:01 Tene reinstalling perl5 won't help
23:01 Tene try this:
23:02 Tene find /usr/lib/perl5 -ctime -1
23:02 Tene that will show you all the files in /usr/lib/perl5 that were created in the past day
23:03 gnuvince_ Thank you
23:11 drupek12 joined #perl6
23:18 justatheory joined #perl6
23:38 catch23 joined #perl6
23:47 goban joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo