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.
00:31 jisom joined 00:37 grayson joined 00:40 Sandrina joined, Sandrina is now known as sandrina_l 01:02 joepurl joined 01:07 knewt joined, nothingmuch joined
Kattana "Perl6: Like Perl, but after major body-alteration surgery, just to make you happy. Perl6's body certainly looks younger and firmer, if you can overlook the frankensteinian scar tissue. Moreover, there are certain... enhancements. All of perl6's joints bend both ways; all of perl6's body parts are fully reconfigurable any way you like them. If your friends could see what you do with Perl6, they'd learn more about your taste than yo 01:12
avar Kattana: got cut off 01:13
nothingmuch that's kind of sick
someone with joints that can go both ways 01:14
nothingmuch pushes the image out of his head before sleep
Kattana pleasent dreams nothingmuch
jisom throw in the Frankenstein's monster image into the mix
nothingmuch thanks
jisom no problem
01:38 xern joined, Supaplex joined 01:42 stevan joined
stevan (scrap your boilerplate)++ # really cool stuff, and I am only on page 3 01:43
obra Getting Real?
stevan 2 step recursive traversals 01:44
type safe casts,..
all sortsa cool stuff
obra ?
stevan its hard to explain, I expect that I only barely grok it myself 01:45
its a Simon Peyton Jones paper :P
www.cs.vu.nl/boilerplate/#papers 01:47
I am just reading the first one at this point
obra Oh! it's not 37signals new book
stevan oh no :) 01:48
I didnt know they had a book, let alone a new one
obra they have two
the new one is on "getting real"
stevan ah 01:50
"Discover the smarter, faster, and easier way to build a successful web app." 01:51
give us $19 and we will tell you
:P
what! no guinsu knife set included?
obra "It's called rails" 01:52
stevan :) 01:53
01:53 Khisanth joined
stevan Ruby on Rails, it slices it dices, is even cuts juliane fries! 01:54
Khisanth bleh
02:11 handon joined 02:20 vel joined 02:22 amnesiac joined 02:39 mauke_ joined 02:58 mauke_ is now known as mauke 03:26 justatheory joined 04:00 revdiablo joined 04:29 szbalint joined 04:32 beppu_ is now known as beppu 04:47 f0rth joined 05:57 K joined 06:22 jp-autark joined
ayrnieu ?eval sub foo { 'hello'.say }; sub bar { foo; 'world'.say }; bar; &foo := { 'goodbye'.say }; bar 06:23
evalbot_9251 OUTPUT[hello world goodbye world ] bool::true
azuroth nifty 06:25
jisom the question is, does it do runtime checking for unsafe commands.... 06:26
azuroth as in, sub foo(Carrot $a) {..}; &foo := sub (Animal $a) { ...}; type checking? 06:27
jisom as in the "rm -rf /" type 06:28
azuroth ohh, right
pretty sure it does
ayrnieu wait, what are we talking about?
jisom evalbot 06:29
azuroth doing an "eval `rm -rf /`"
ayrnieu ?eval (open "/etc/motd", :r).readline 06:31
evalbot_9251 Error: No such method in class IO: "&readline"
06:40 Alias_ joined 06:46 beppu joined 06:48 beppu joined 06:51 huhlig joined 06:59 fandango joined 07:03 Draconit joined 07:04 beppu joined 07:15 jisom joined
huhlig is there a good site to learn perl6 yet? 07:20
revdiablo huhlig: Well, you can read the specs: dev.perl.org/perl6/doc/synopsis.html 07:22
huhlig: Keep in mind, it's all still a work in progress
ayrnieu huhlig - pugs also comes with examples and tests.
huhlig when are they expecting some form of solidification 07:23
revdiablo When it's ready
huhlig heh
07:33 Draconit_ joined
buu CHRISTMAS 07:39
Solidification by christmas!
I love that joke.
07:45 lao joined 07:47 gabor529 joined 07:48 Aankhen`` joined 07:58 handon joined 08:06 ghenry joined 08:08 orafu joined 08:10 marmic joined 08:26 bsb joined 08:32 typonaise joined 08:34 handon joined 08:36 _freenode joined
gaal putter: oh, go ahead and hack on Prelude.pm 08:37
the only thing is that currently, an installed pugs won't find a precompiled Prelude, probably 08:38
that's a simple matter of programming (and of some directory structure layout design) to fix
we don't inline the precompiled prelude bytecode inside the pugs executable, so there's need to tell pugs where to find it on the disk 08:39
currently there's one hardcoded location: in the source tree.
if there were an easy, nonhacky, portable way to inline stuff in the executable, the gzipped version of the bytecode (which we know how to handle) is just 42k 08:41
so all I was saying is that we should not release until we fix that 08:43
but that doesn't mean we can't also develop more P6 prims.
08:45 gaal joined
pasteling "ayrnieu" at 68.13.103.15 pasted "my rendition of $PUGS/examples/p6explain/p6explain" (19 lines, 480B) at sial.org/pbot/16122 08:47
08:50 iblechbot joined
miyagawa audreyt: ping 08:56
09:04 handon joined 09:15 clkao joined 09:16 handon joined, Aragone is now known as Arathorn
Alias_ miyagawa: She's still a bit thin on the ground 09:33
Arathorn her wave function has yet to be collapsed 09:34
miyagawa k
09:37 lumi joined 09:53 rindolf joined
rindolf Hi all! 09:53
How can I query CPAN for the list of distributions a certain CPAN id has?
ayrnieu in CPAN shell? "ls $author" 09:57
10:00 elmex joined
Alias_ rindolf: A one off for just you, or in code? 10:01
because if it's just you, y'know, search.cpan 10:02
rindolf Alias_: in code. Programatically.
10:02 Draconit_ joined
Alias_ Do you mind using CPANPLUS? 10:02
It apparently has a much better API for doing that sort of thing
Or so everyone I know that does that sort of thing says 10:06
Personally, I want a SQLite DBIx::Class interface :)
Like JSAN has
10:07 nnunley joined 10:19 rantanplan_ joined
rindolf Alias_: I already whipped up a script using CPANPLUS. Sorry for bothering you. 10:19
Let me nopaste it.
Alias_ ah, ok
pasteling "rindolf" at 88.155.200.148 pasted "CPANPLUS script to return the distributions of an author." (12 lines, 321B) at sial.org/pbot/16123 10:20
rindolf Alias_: sial.org/pbot/16123
10:21 Gothmog_ joined 10:25 szabgab joined
samhain1138 rindolf: you can expand Author, and $author->ls :) 10:25
something like: my ($author) = CPAN::Shell->expand('Author', $id); 10:26
10:26 Gothmog_ joined
szabgab when running pugs it says "Loading Prelude... " that takes nearly forever, maybe 5secs ? Can I get rid of that delay ? 10:26
rindolf szabgab: hi! 10:33
jp-autark isn't it supposed to read prelude.yaml?
PreludePC.yml even.
rindolf szabgab: I sent you a patch to add some new authors and distributions to the Misc. Content Page on www.perl.org.il 10:34
szabgab rindolf: thanks
rindolf szabgab: you're welcome. 10:35
I added Gaal and Yuval Kogman.
Juerd that takes nearly forever, maybe 5secs 10:37
lol.
samhain1138 hehe :) 10:40
10:47 Alias_ left 10:51 rindolf left
Migo szabgab: when I first ran "make unoptimized" and trivial "time ./pugs -Iext/Net-IRC/lib -MNet::IRC -e 1" it took 1 minute 40 seconds. Now it takes 33 seconds on my machine. 10:57
gaal szabgab: what's your cwd when you run pugs? 11:01
szabgab I compiled optimized - so I guess this should be faster - and I only wanted to run say "hello world"
gaal: anything I use the symlink /usr/bin/perl6 pointing to pugs in my compilation directory
gaal: it is much fatster if my cwd is the same where pugs resides... 11:02
gaal right. it's a known issue: we don't look for the precompiled prelude in a system installation place
I was discussing it with putter eariler if you want to see the details 11:03
szabgab gaal: no, I just want it to work fast :)
gaal well, we want that too :)
Migo "time ./pugs -e 'say <Hello, World!>'" is 3 seconds for me. 11:04
szabgab real 0m2.242s
the above for me
Migo so, I guess, it is loading modules that takes huge time 11:05
szabgab 2 sec is already reasonable for me 11:06
for now
gaal Migo, szabgab: these are known issues and we're still working to improve them. 11:07
szabgab gaal: thanks 11:08
audreyt rehi 11:09
I _finally_ got network!
gaal yay!
Juerd Hi, audreyt
audreyt suffers from massive withdrawal
Juerd Welcome back :)
gaal heard your stuff was mismoosed? :(
audreyt my luggage has come back last night
no net @ hotel, no net @ conf site (until now) 11:10
samhain1138 gaal: so why is it so much faster for me? real 0m0.252s?
gaal samhain1138: (a) you are in the pugs build path (b) you made optimized (c) your machine is fast. :)
Migo I would say, very fast :)
szabgab he removed the sleep 2 from the code... 11:11
Migo mine is 800MHz, 192Mb physical memory
samhain1138 (a) no, (b) just make -- is optimized the default? (c) I WISH! ;)
audreyt optimized is default
samhain1138 ok
thanks :)
gaal 192 is lower than I've ever tried, but migo may be able to get away with an optimized build too 11:12
if you set the heap size to about 150 in config.yml
run perl Makefile.PL again
Juerd Migo: If you're upgrading anything, upgrade RAM. It's cheap and helps a lot.
Migo: 512 MB costs around $50 if you look around a bit.
gaal I'm curious if it works. (~200 worksforme) 11:13
Juerd: that sounds like a good price
Juerd The hoersaal is too hot, outside it is too cold :|
Migo gaal: yes it took 2 days to compile it optimized on Hackathon. Whenever network was lost, Pugs.AST.Internals and Pugs.Parser were recompiled again and again. 11:14
gaal network!?
Migo I compiled on my home machine, laptop had no haskel 11:15
gaal anyway 2 days means your machine was thrashing. you really need to throttle GHC's ram usage
ah
Migo better split them to smaller pieces :)
szabgab Migo: at least you spend 2 days with useful work :)
gaal Migo: commits welcome :) 11:16
but in fact that's exactly what lumi, audreyt, and integral were working on
among other things
Migo gaal: I know. And it was my mistake not to use screen(1) in the first place. 11:17
integral why did you need to compile it optimized?
gaal also anatoly with the mad math skillz :)
Juerd 12:18 < gaal> anyway 2 days means your machine was thrashing. you really need to throttle GHC's ram usage 11:19
Yeah, I think it may be a good idea to disable swapping, if it's a home box or laptop, that may crash.
Migo integral: it seems to be 2-3 times faster (30 seconds vus 90 seconds) 11:20
11:20 Alias_ joined
Alias_ audreyt: ping? 11:20
audreyt Alias_: pong
Alias_ joy!
integral Migo: oh, you were doing lots of smoking?
Alias_ audreyt: OK, so I'm still working on making Bundle::CPAN sane on Win32 11:21
Migo integral: no, just "-MSome::Module -e 1"
integral oh! a perl6 developer, excellent!
Alias_ audreyt: And next in my sights is Module::Signature, which is troublesome
audreyt k
Alias_ audreyt: It seems to install even without gpg or Crypt::OpenPGP
When it shouldn't
And then on running complain that there's no gpg 11:22
And occasionally freeze up the process in the process
A bigger problem is that the bugs are somewhat nondeterministic
But the big one is that Module::Signature shouldn't install on Win32 without either gpg or the OpenPGP feature enabled 11:23
Can either 1) you fix it 2) Let me steal M:Sig
(for a while) 11:24
audreyt sure, you are now M::Sig comaint
Alias_ thanks
Is it in SVN somewhere, or do I borg it to my SVN? 11:25
audreyt borg it
Alias_ ok
audreyt hm, or place it to modinstall/ ?
Alias_ You don't have a top-level somewhere inside that same area?
ala svn.phase-n.com/svn/cpan/trunk/ 11:26
audreyt no... do whatever you'd like :) 11:27
audreyt is still backlogging
Alias_ ok
I'll borg it for now
I want to set up an orphanage some time soon anyway I think
To let the collection of clueful people co-operatively maintain stuff like that
Will Mod:Sig freak out if it isn't itself signed? 11:29
rafl kane-xs: ping 11:30
miyagawa audreyt: are you with Larry? 11:32
audreyt miyagawa: no, we parted in .il a while ago
Alias_: prolly not
miyagawa k
audreyt: Matz will be in YAPC::Asia 11:33
and hackathon too
audreyt oooooooh
Sasada? Takahashi?
miyagawa yeah Takahashi will speak at YAPC 11:34
not sure about Hackathon. I'll call him
I've never talked with Sasada before. Probably you want to mail them
11:44 mj41_ joined
audreyt okay, will do 11:47
lunch, bbiab & 11:54
11:56 salty-horse joined 12:10 handon joined, kane_ joined 12:15 salty-horse left 12:34 szabgab left 12:35 salty-horse joined
salty-horse hi. i was facinated by audreyt's style of presentation and failed to find much information about it. i started a wikipedia stub here: en.wikipedia.org/wiki/Takahashi_Style, and welcome everyone who's interested to contribute 12:36
Alias_ Takahashi Style? or Method? 12:37
wolverian en.wikipedia.org/wiki/The_Takahashi_method
salty-horse i'm not sure :) a redirect/rename is in order
....
oops. en.wikipedia.org/wiki/Takahashi gave me nothing :( 12:38
wolverian nod, I use en.wikipedia.org/wiki/Special:Search?search=%s
(as a magic bookmark)
lypanov salty-horse: www.presentationzen.com/ 12:39
salty-horse wolverian, me too :)
lypanov salty-horse: (presentationzen.blogs.com/presentat...e_ta.html)
salty-horse lypanov, i linked to taht article in my now-defunct stub 12:40
lypanov :)
its an interesting blog
salty-horse i updated en.wikipedia.org/wiki/Takahashi 12:43
stevan miyagawa: when is the hackathon scheduled for? 12:54
stevan realllllllyyyy wants to go now 12:55
Juerd (Someone's using *MS Powerpoint*!) 12:57
stevan is suprised to find that plane/hotel rates are still reasonable 12:58
stevan ponders begging his $wife and $boss to let him go 12:59
13:00 elmex joined
miyagawa stevan: it's March 31 and April 1st. 13:00
the conference is March 29-30
rafl Juerd: Who? 13:05
sahadev coffee? 13:06
sorry, wrong window
pasteling "samhain1138" at 217.132.59.2 pasted "Problem compiling 6.2.11: no such key: 'rpath_blib'" (197 lines, 14.6K) at sial.org/pbot/16126
rafl samhain1138: Embedded parrot? 13:07
Juerd rafl: Someone who just tried the beamer in the break 13:08
rafl Juerd: OK
WTF, then.
13:09 chris2 joined
samhain1138 rafl: that's more-or-less on purpose (ok, ok: it's stolen from Gentoo's 6.2.10 ebuild ;)) 13:09
13:11 fglock joined
samhain1138 rafl: i'm (naturally) basing my 6.2.11 ebuild on the 6.2.10 ebuild (which worked, honestly ;)) 13:11
rafl samhain1138: Yeah, but parrot changed. 13:14
svnbot6 r9252 | fglock++ | iterator_engine - updated README
r9252 | fglock++ | - started 'macro' implementation
samhain1138 rafl: i'm using the same parrot, if that's what you mean 13:15
(i'm just compiling 6.2.10 again and it works so far :))
Juerd Why does everyone seem to need logging? I very rarely use such stuff.
rafl samhain1138: Which version? Maybe newer pugs needs newer parrots
samhain1138 rafl: 0.4.0 13:16
rafl Juerd: For debugging only. And alrso very rarely.
Juerd For debugging, which I also don't do much, I have print statements and a debugger...
svnbot6 r9253 | fglock++ | iterator_engine - p6 tests are *.p6t; p5 tests are *.t 13:17
Juerd Ah well, I'm probably weird.
rafl samhain1138: Could be caused by that. What does that output: parrot-config --dump
samhain1138 doesn't have parrot-config for some reason... 13:21
13:22 _freenode joined, kolibrie joined
pasteling "samhain1138" at 217.132.59.2 pasted "parrot-config" (256 lines, 8.1K) at sial.org/pbot/16127 13:23
13:28 _freenode is now known as typonaise
rafl samhain1138: Yes, that's caused by an old parrot. 13:30
samhain1138: It is caused by an old parrot. 13:33
samhain1138: Unfortunately the current parrot doesn't work either.
samhain1138: I wanted to fix that with leo today.
samhain1138 rafl: heh, what should work then?
rafl samhain1138: Only p5 embedding. 13:34
samhain1138 rafl: ok, thanks :)
lypanov p/win 2
um
jp-autark cvs parrot works though 13:40
?
fglock re: macros - if a macro has an 'is parsed' trait then it doesn't need an argument list, right?
audreyt fglock: right. 13:43
fglock: "is parsed" subsumes the sigs part
fglock audreyt: thanks!
audreyt: I'm implementing macros in i_e 13:44
Alias_ swears at Windows 13:47
Bloody stale .dll locks
lypanov Alias_: reboot! 13:48
Alias_ lypanov: Programmatically, not this instance
I _really_ need to hunt down that method for deleting in-use file 13:49
hmm... 13:50
ok, so you can move in-use files
13:51 kisu joined 14:08 Qiang joined 14:19 nothingmuch joined
nothingmuch morning 14:27
14:28 pmurias joined
kolibrie nothingmuch: morning, glad to hear your sister's party was a success 14:36
anatoly it's not morning, nothingmuch :)
kolibrie anatoly: it's always morning somewhere
Alias_ except when the sun blows up 14:37
Then it's morning EVERYWHERE!
(briefly)
kolibrie and then night for a long time
anatoly what was that story by Azimov 14:40
when night finally came and they saw the stars
rgs "nightfall" 14:41
anatoly Right. Thanks. 14:42
14:47 vel joined
svnbot6 r9254 | audreyt++ | * fix prelude generation on win32, due to CRLF problems reported by Ran. 14:47
14:52 Chip_Zero joined
svnbot6 r9255 | audreyt++ | * Instead of forcing BEGIN evaluation of populating ::EXPORTS for 'is export' functions, make it evaluate at module body's runtime -- eventually the sub expression itself will be part of the runtime desugaring as well, for '$x = sub f is export {...}'. 14:52
nothingmuch kolibrie: hi 14:55
svnbot6 r9256 | audreyt++ | * workaround to get PIL2JS compiling -- PIL wasn't handling VRefs correctly.
nothingmuch anatoly: err, okay
actually when the sun blows up it's morning everywhere not so briefly 14:56
14:56 justatheory joined
nothingmuch novas usually take about 18 months to finish, IIRC 14:56
kolibrie nothingmuch: thanks to you and gaal, I almost understand parsec now, I think
it is amazing how clear haskell is after initial comprehension is gained 14:57
nothingmuch kolibrie: =)
haskell is very simple
(... from a certain perspective ;-)
Alias_ (the mirror universe) 14:59
rafl samhain1138: OK. Actually the solution seems to be harder than we thought. 15:03
broquaint ping, fglock. 15:08
unping 15:11
audreyt clkao: pil2js should be repaired now -- the fix is actually done several days back 15:14
but net here is so extremely spotty
got a lot of designing/thinking/syncing done though 15:15
nothingmuch audreyt: i had something to ask you yesterday
but now you will never know, it's too late
audreyt with Nicholas on perl5 -lpcre, perl bounty, darkpan, etc
and with Andk on sixpan, a better only.pm for p5, and the toolchain
sadly the conf will end in an hour or so, and with it the network 15:16
rgs andk on sixpan ? hope I'll see the slides
audreyt I'll try to get better net ~48hrs from now when I'm in Cologne
rgs: no, just discussions, not slides -- thogh they did have a BOF here on 6pan
fglock broquaint: unpong
rgs ok
audreyt I'll journal what I can about it
rgs thanks
svnbot6 r9257 | fglock++ | iterator_engine - macro declaration is parsed and installed in the
r9257 | fglock++ | grammar. TODO - install the macro body
clkao audreyt: oh ok. 15:17
Alias_ audreyt: I'm going to N/A Module::Signature on Win32 outright
rgs just notices "perl5 -lpcre" and goes "huh?"
Alias_ audreyt: Just FYI...
rgs that would be for the PIL on P5 thing ? 15:18
audreyt Alias_: sure
broquaint Hey fglock, I was trying to figure out iterator_engine_p6compiler.pl but gave up :)
audreyt rgs: er no, the possibility of solving all p5 re problems now and forever by switching to pcre
and rid the porters of a maintainence nightmare
Alias_ audreyt: It appears that half of IO::, and in particular IO::Socket::INET, is completely fucked on Win32. And somehow IO:: has survived in the core without any of the problems being fixed
audreyt and gain some nice features along the way (varwidth lookbehind, etc)
Alias_: :/
rgs you mean, transfer the maintainance nightmare somewhere else :)
Alias_ I'm still not sure how it managed to make it past testing 15:19
audreyt rgs: well, it's not unlike syck vs YAML.pm.
Alias_: N/Aing is fine for the time for me.
Alias_ ok
audreyt but if IO:: chain is broken, gbarr needs to know.
Alias_ Oh, I'm filing bugs :)
audreyt cool :)
Alias_ There's a ton of bugs in rt.cpan.org
fglock broquaint: it compiles itself - you'd better start with iterator_engine.pl, which is the bare rule engine 15:20
Alias_ But some seems to be dismissed with "why do you care, it's in the core"
audreyt Alias++ # pro player at this CPAN game
Alias_ ugh
fglock then goto p6rule.pl, and then to p6compiler
Alias_ audreyt: It helps to not get sucked into the languages themselves :)
audreyt indeed :)
Alias_ That's one Yak this little black duck is not going to shave 15:21
fglock broquaint: there is an exponential growth in complexity
broquaint Hrm, perhaps I don't quite understand it then, fglock. Could you sum it up in a sentence or two?
audreyt it's nice to hear that TimToady gives Nicholas goahead to commit his perl5-to-perl5 translator into a perforce branch
so all can see it soon
it's basically annotating PPI tree with B:: tree
Alias_ huh? what?
audreyt so you have close correspondence of which area of source code are responsible for which opcode
Alias_ That sounds crazily evil 15:22
audreyt actually it's the other way around, sorry
annotating a (very pessimised) B:: tree
with its document tree correspondents
rgs against which perl has TimToady patched ?
audreyt rgs: 592 iirc.
Alias_ So taking the code tree and embedding pointers into the document tree?
fglock broquaint: p6compiler compiles Perl 6 - it does this by following a grammar, which is defined in p6 itself - so it reads the grammar one sentence at a time
audreyt Alias_: yeah, and inventing opcodes like "junk"
rgs there will be conflicts.
audreyt for comments
Alias_ right
fglock broquaint: and generates the compiler dinamically 15:23
rgs OP_JUNK++
audreyt rgs: right, but I care about getting into public review first
Alias_ Must say that's a good idea though
audreyt we can survive conflicts :)
Alias_ Hadn't occured to me to do both trees in parallel and match
Mostly because I don't work under the language
broquaint The perl6 is compiled into perl5 right, fglock?
Alias_ Must be hella expensive
audreyt Alias_: indeed. I think it's a nice feature -- it even dumps yaml 15:24
Alias_ PPI trees are pretty big
fglock broquaint: right
audreyt (via syck)
Alias_: and you can control it with an env
Alias_ runs away
audreyt if that env is not set, then there's minimal perf slowdown
Alias_ Glad you found it useful at least
audreyt if it's set, p5 switches to the mad lexer
(mad = misc. annotation declaration or some such)
and attach to each opnode madprops 15:25
(misc. annotation declaration properties)
to point to document tree
Alias_ head esplodes!
audreyt to be dumped out with madskills
into a madtree, which is then fed into "nomad"
Alias_ hahaha
audreyt to extract the parallel trees out and dump to yaml
Alias_ names++
fglock broquaint: what it makes it hard to get is that the compiler is not there - p6compiler is like a factory, which makes a compiler on demand - or something like that 15:26
Kattana its so ingenious i thought of it myself which is what led me to #perl6 ^__^ 15:27
broquaint So how might I use it to compile a simple perl6 script? 15:28
Kattana (ingenious in this case is also a synonym for lazy)
15:29 Draconit_ joined
Alias_ audreyt: Do you have any cygwin ninjas floating around? 15:29
audreyt: I noticed pjf doing some thing
fglock broquaint: perl iterator_engine_p6compiler.pl <sourcefile> 15:30
broquaint: there is a sample file, if I didn't erase it by mistake - let me check
broquaint Hrm, I keep getting errors of: syntax error in program '$PROGRAM_SOURCE' at '$PROGRAM_SOURCE' 15:32
Even for simple "Hello World" programs.
fglock broquaint: the current grammar file is very simple - try say 'something' (with single quotes - I didn't install double quotes yet) 15:33
broquaint: iterator_engine_p6compiler.p6t is the sample file I use for tests 15:34
broquaint: it will warn that 'macro' is not yet working, because that's what I'm working on
broquaint Weird, iterator_engine_p6compiler.p6t works but: say "Hello World"; # doesn't :/ 15:37
15:37 Migo1 joined
fglock broquaint: double quotes don't exist yet - it needs an entry in the grammar file - that's easy to add, but I'm trying to make macros work first :( 15:38
broquaint: try single quotes
broquaint Cool :) 15:39
fglock broquaint: :) 15:40
stevan audreyt: ping
svnbot6 r9258 | fglock++ | recovered iterator_engine_p6sample.p6 15:41
fglock broquaint: take a look at the grammar source file - iterator_engine_p6prelude.p6 - everything the compiler can do is specified there
audreyt stevan: pong! 15:42
stevan: theme for vienna^2 hackathon is: objspace on pir
stevan audreyt: nice
I would be into doing that
audreyt you have timeslice?
wonderful!
stevan I think 15:43
audreyt the 'thon happens 4 days from now
for ~1wk
it'd be wonderful to have you on gobby/skype/etc
stevan two projects have been "launching" for a week now, the $clients cant make up their @minds
audreyt in .il we figured out how to serialize the code/pad/etc to objspace
stevan nice
audreyt as well as lang simplified enough so linking would be possible
stevan :)
audreyt so it's time to write them down as APIs, and promote S12.5 etc to Perl6::Doc tree 15:44
et cetera
it'd be wonderful if you can work on the P5 mapper
stevan ok,.. I feel about ready to write 12.5 now, after Class::MOP
audreyt so we can (when compiled to perl5) interop with native p5objs
excellent!
stevan cool, I will make as much room as possible in my schedule 15:45
I have a (sort of) unrelated question though, not sure if you can even answer
how releated is SYB and LINQ?
are andres and SPJ working together at all?
15:46 stennie joined
audreyt LINQ is primarily emeijer 15:46
they are @ .us
SPJ and JaffaCake (S.Marlow) are @.uk 15:47
stevan ah
audreyt LINQ is based more on monad comprehensions
(ala Haskell 1.4)
as well as emeijer's previous woek on HaskellDB and HSP
so, not directly related
but generic programming is a focus (as is relational ones) on CLR3 15:48
15:48 lumi joined
stevan the whole generic traversal thing in SYB 1 and the LINQ movie when taken together, kind of melded in my mind 15:49
but that might have just been my brain melting
SYB is the first Haskell stuff that really made sense to me,.. not sure if thats cause I am finally groking haskell now, or SYB is just so damn cool
fglock finds a cool way to implement macros 15:50
samhain1138 rafl: hmmm, what is the solution? 15:51
rafl samhain1138: No solution yet. 15:54
samhain1138: Well. Actually there is a solution. Link installable_config.o into libparrot.
samhain1138: But unfortunately there is no implementation yet.
fglock re: macro - dump the macro body AST into perl5, and { return } it when the macro rule matches 15:57
lunch & 15:58
15:58 fglock left
rgs perl 6 on perl 5 question on P5P : www.xray.mpe.mpg.de/mailing-lists/p...00081.html 16:00
16:12 fglock joined 16:15 iblechbot joined
fglock rgs: thanks! 16:16
rgs what did I do ? 16:17
rgs looks behind him
16:17 Draconit_ left
fglock rgs: sent a link :) 16:18
rgs while you were off channel. 16:19
fglock rgs: I just backlogged 16:20
rgs you're irc addicted. 16:21
nothingmuch irc addiction is bad 16:25
Odin- What isn't? 16:27
nothingmuch well
err
nothingmuch firmly believes that balance is everything
Odin- Or, to be more specific, what addiction isn't?
nothingmuch so i actually couldn't agree more
16:28 tcliou joined
nothingmuch but irc as opposed to say tea addiction, or møøse addiction is not so bad 16:28
Odin- stamps a taijitu on nothingmuch. 16:29
huhlig is perl6 functionably useable yet or is it still in a great enough flux that nothign that doesent need to be written 6 times should be tried
audreyt huhlig: rewritten 6 times, probably not.
the APIs are in flux, that's true, which is why we started the docs/Perl6/ tree to document them 16:30
if you are looking for writing new production code in perl 6
then perhaps wait another couple months 16:31
till we get ->perl5 compiling in shape along with a fairly complete API docs
Alias_ heh, production perl 6
audreyt (but you can help -- want a commit bit?)
szbalint :) 16:32
nothingmuch rgs: so, i'm playing a bit with -MO=Terse
rgs Terse sucks. use B::Concise instead
nothingmuch okay
audreyt bbl -- hotel &
nothingmuch ciao audreyt
rgs nowadays Terse is only a wrapper around Concise 16:33
nothingmuch ah
rgs see you audreyt
fglock is there a switch in Data::Dumper (p5) that makes it not show '$VAR1 =' ?
nothingmuch anyway, i was thinking: the optional optimizer is useful for learning too
1 + 2 is optimized into 3
rgs yes
nothingmuch and that's not good for poking around
so yet another reason
rgs no, that's constant folding, only part of peephole optimisation
pmurias s/::/ruleop/ is alternation([constant("..."),constant("...")]) implicitly ^ anchored ? 16:34
nothingmuch yes, ofcourse =)
rgs the optimiser turns some ops into nullops
nothingmuch doesn't it just zap them?
rgs with concise you can see what they were before (with B::Terse you can't)
(for example)
nothingmuch aah
okay, tha'ts a good thing
Migo1 fglock: Terse 16:35
fglock Migol: thanks! 16:36
Migo1 hey, it is "1", not "l" :) 16:37
nothingmuch fglock: also note that for some things you need a special flag
to make it really emit perl 5 code
fglock Migo1: fixed!
nothingmuch not just something that looks familiar
for complex structures
there's like 31985895 closed bugs on Data::Dumper's RT queue about that ;-) 16:38
(one of them is mine)
fglock nothingmuch: it's a simple tree - I generate it myself, it's non-recursive
16:38 cookys joined
nothingmuch fglock: ni that case you're in the clear, IIRC 16:38
fglock nothingmuch: iterator_engine ast tree - I need to return it from the rule that parses the macro 16:39
nothingmuch rgs: was this ever finished? gisle.aas.no/perl/illguts/op.html
rgs no 16:40
nothingmuch shame
rgs and it's probably a bit outdated w r t blead 16:41
nothingmuch okay
Migo1 Data::Dumper has very silly API, I always wrap it into something like encode_value of search.cpan.org/src/MIGO/Arch-0.5.1...edIndex.pm 16:45
will .perl accept parameters for different styles?
nothingmuch ofcourse not, there is only one true way to style Perl 6 code
Migo1 in one line? 16:47
nothingmuch $*USER.style 16:49
seriously though ;-)
i guess it will have it
like
:indent_type( tab || space );
:indent_width( ... ); 16:50
....
rgs :brace_placement( one_true_way ) 16:52
Kattana And it will automagicaly generate an IDE at compile time?
Migo1 I hope the defaults will be less brain-dead than in Data::Dumper. :) I should yet find a usage for default $Terse = 1; $Indent = 2;
I mean $Terse = 0; 16:53
svnbot6 r9259 | fglock++ | iterator_engine - macros! 17:05
17:14 bsb joined 17:28 Eimi joined 17:30 nnunley joined
nothingmuch so 17:37
we're sitting at the table for dinner
and some time near the end my 17 (Today!) year old sister says 17:38
"I can't understand this evolutionary defect... Why is it that men get sexually aroused by rectal stimuli?" (ofcourse, with more blunt wording)
so then my 7 year old sister starts giggling and repeating sub-phrases 17:39
chris2 that's an evolutionary defect?
nothingmuch 10 points for style
-100 for practicality
well, her conjecture was that it doesn't help ensure procreation
i think what she's missing is that perhaps the same system that makes both the "right" parts and the "wrong" parts work is simpler when generalized 17:40
and thus perhaps it's actually a direct product of evolution
Alias_ The general purpose answer would be that she probably hasn't looked into it in enough detail yet
integral there are some monkeys that use sex a _lot_ for community building and bonding, inc. homosexual sex
nothingmuch Alias_: aye
chris2 i'm trying to connect it with the fact that assholes are older than mouthes
rgs larger bottom -> larger route to make babies pass -> less risks of mother dying 17:41
stevan nothingmuch: so you are saying that anal sex is merely a refactoring done by ($diety|$evolution)?
nothingmuch chris2++
Alias_ rgs: I think he may mean physical stimulo
nothingmuch stevan: no, i'm arguing that the feature was added at the same time
chris2 anal sex is just a blow-job that got stuck in evolution...
stevan :D
chris2 ;)
rgs Alias_: ah yes. I read it backwards.
nothingmuch stevan: and that since this helped in one way 17:42
stevan nothingmuch: so it's in the spec then?
nothingmuch the by product also remained
no
it's one of those "accidental features"
stevan thumbs through his copy of Anal::Sex::Bible but cant seem to find the reference
chris2 still, the closeness of rectum and genitials is given in almost any creature
stevan perhaps I have the wrong version
chris2 notices he is not in #haskell-blah. wow :P 17:43
stevan nothingmuch: a more detailed question is "why do men have prostates"
chris2 ever tried to eat with an open mouth?
oops, confused two words :P
stevan tries to figure out which two 17:44
chris2 prostates != hemorrhoids
stevan not at all
one usually is associated with bad feelings, the other with good,.. they couldnt be more diff 17:45
chris2 prostate controls flow of semen or urine, no?
stevan is not a doctor.... please turn your head and cough
nothingmuch stevan: two reasons 17:46
or wait
stevan chris2: some men are stimulated sexually when their prostate (something inside their anus) is stimulated
chris2 ever had to cough while the doc holds your balls?
nothingmuch is this about their location?
or their actual purpose?
chris2 no
Alias_ stevan: So we are encouraged to get those big dumps out and not hold them in
chris2 the prostates are somewhere different to my knowledge
nothingmuch point 1: semen has to be slightly cooler than body tempreture, I don't know for what reason
point 2: it probably encourages struggling males to chop each other's gonads off 17:47
stevan Alias_: there you go,.. evolution at it's finest,.. its all about the shit
nothingmuch thus ensuring that only the best fighters are left
chris2 then, to check if you got a big prostate, the doc actually fingers you...
nothingmuch through a catalyzing protocol of ensuring that the losers can't reproduce
;-)
chris2 nothingmuch: know why we have finger nails, btw?
stevan ok,.. all this talk about anal sex is making me hungry,.. and besides,.. my pizza is ready :P
nothingmuch chris2: to keep various items stored in a convenient location? 17:48
stevan: =)
stevan heads up stairs for food &
chris2 nope. to get rid of your ear wax
so you can hear the mammuths comming
nothingmuch ouch
actually, we don't really need to remove ear wax unless:
a. we have some kind of medical condition that creates too much of it
b. we used q-tips 17:49
there are these tiny hairs that help the wax move out
chris2 it depends
nothingmuch and if you just use warm water they aren't messed with
otoh qtips flatten them
chris2 most asian people have different ear wax, or so i heard
they lack the gene for stickyness
nothingmuch hasn't heard of that
what does it do?
chris2 it makes the ear wax sticky 17:50
nothingmuch no no, i mean what is the effect of non sticky wax?
does it just squirt occasionally when you turn your head?
chris2 "Asian people tend to have very dry (flakier) wax build-up opposed to a white person's oily wax build-up."
nothingmuch oh 17:51
17:51 netanya_goy joined
chris2 maybe 17:51
nothingmuch ouch
chris2 "Thus, traditionally Asian people use wooden ear picks to clean out their ears instead of Q-tips, which aren't as thorough for this type of wax." 17:52
be careful :)
Alias_ I have to wonder if waxy ear wax is an adaptation for the cold 17:56
17:57 jeh joined, huhlig joined
nothingmuch lypanov: i just started making stevan write us some piln docs 18:06
stevan denies all invovlement in nothingmuch's plot to document PIL^N
the first rule of PIL^N club is ... 18:07
lumi There was some terrorist thingy at the Church of the Annunciation in Nazerath (this is of possible interest to szabgab, anatoly, timtoady, bsb who were there with me a week ago)
stevan nothingmuch is not allowed in :P
18:08 rantanplan_ joined
nothingmuch imagines a bunch of perl hackers looking like perl syntax in a church 18:12
PerlJam lumi: Great! Now you've implicated them to the FBI/CIA/NSA/whoever is watching the Internet ;-) 18:13
lumi It's full of internets! 18:14
But apparently not a terrorist thing, rather three crazies
salty-horse with fire crackers 18:20
huhlig muhahah 18:22
huhlig tells the FBI
18:22 bsb joined 18:26 feng joined
stevan ( bsb | audreyt | gaal ).ping() 18:28
szbalint Is that a junction? 18:34
stevan yeah, but it wouldnt actually work 18:35
ping( bsb | audreyt | gaal ) would be more correct IIRC
salty-horse something like ()>>.ping() would work as well, no? 18:38
actually, what will ping() do with a junction? 18:39
stevan whatever ping wants to do :) assuming sub ping (Junc $people) { ... } 18:40
salty-horse my irc client uses perl5 bindings only 18:41
and it doesn't even have ping() :/
huhlig hmm 19:11
anyone here use norton firewall?
stopkeylogger 19:12
19:12 bsb joined
huhlig hmm 19:12
19:15 Arathorn is now known as Aragone
ayrnieu yeah, that's great, welcome to ONE WEEK AGO. 19:19
huhlig heh 19:24
didnt think it would affect anyone here
this is kind of a unix channel
integral ayrnieu: TWO WEEKS AGO 19:25
19:29 dduncan joined 19:31 iblechbot joined 19:37 sahadev joined
ayrnieu integral - oh. Mind still in the weekend. 19:41
19:41 larsen joined 19:42 bsb joined 19:47 pmurias joined 19:48 Aankh|Clone joined
gaal stevan: pong! 19:50
audreyt: ping 19:51
xerox pang! 19:53
svnbot6 r9260 | pmurias++ | A version of iterator_engine.pl with comments in Smart::Comment format. 19:54
r9260 | pmurias++ | For consideration by fglock.
20:01 Kattana joined 20:07 bsb joined 20:12 amnesiac joined
bsb pong stevan 20:22
20:35 kulp joined, Shachaf joined
stevan gaal: bsd: I am having issues running "make pili" 20:40
nothingmuch's best guess was that it tries to use gzip despite lack of zlib 20:41
ld was getting undefined symbols _gz*
gaal reproduced. 20:47
I thought we had MissingH but I can't find it now? 20:49
or is this actually not enough to have missingh, and we need zlib for faststring? 20:50
20:52 bsb joined 20:53 duckyd joined 20:55 SamB joined 21:02 marmic_ joined 21:06 avar joined 21:07 kulp left 21:27 fglock joined
fglock pmurias: re comments - nice! is there a way to selective enable Smart::Comments on specific subroutines? (haven't looked at the docs yet) 21:29
pmurias fglock: hi 21:37
fglock pmurias: hi!
pmurias use Smart::Comments before the sub;no Smart::Comments after
fglock pmurias: thanks! 21:38
21:38 DesreveR joined
pmurias shall i merge the changes into the main version ? 21:39
fglock pmurias: I wished there was a way to enable from the command line --show-comments='sub xxx' :)
pmurias: sure - please leave it commented out in svn, as the users will not see debugging messages 21:40
pmurias regarding the switch I we could either use a custom Smart::Comments, or change it's private subroutines at compile time 21:42
or use a second source filter
if it's Smarts internal change rarly, the second option is the best 21:43
21:45 DesreveR joined
fglock pmurias: I'd rather use an ENV variable to configure what to show - but I don't want to add another project to my list :) 21:45
pmurias is there a way to determine the which subrutine a sub was called from? 21:50
PerlJam pmurias: perldoc -f caller
:-)
That morphs into CALLER:: or something in perl6 though
(I think) 21:51
21:51 Chip_Zero left
pmurias forgot to look at the secong page of the caller desctiption in the camel(third edition,polish second by the way :) 21:53
svnbot6 r9261 | fglock++ | iterator_engine - macros can return perl5 code using: 21:55
r9261 | fglock++ | return ' eval( \' print "xxx\n" \', :lang<perl5> ); ';
r9262 | fglock++ | iterator_engine - removed debug info 21:58
22:01 pdcawley joined 22:03 DesreveR is now known as r0nny
pmurias a more difficult question, how do i get the subroutine a anonymous subroutine was *defined/created* in? 22:03
in p5 22:04
:(
fglock: the problem with --show-comments is how do you name an anonymous sub :) 22:07
fglock: ignoring that it's 7 lines of code in the %ENV version 22:09
22:14 larsen_ joined 22:21 Nouk joined 22:26 elmex joined
pmurias good night, hopefully I'll dream of solution to that problem 22:29
22:31 jeh joined 22:42 ruz joined 22:43 nothingmuch joined
nothingmuch møøseness 22:43
22:50 Limbic_Region joined
fglock Limbic_Region: hi 22:50
23:01 frederico_ joined
fglock frederico_: ola - voce e de Sao Paulo? 23:04
svnbot6 r9263 | fglock++ | iterator_engine - added very simple parameter binding to macros (need more work)
frederico_ fglock, eu sou sim 23:09
fglock, mas da cidade de taibate
fglock frederico_: ah ok, vc e do grupo de taubate :) nos vemos na cascavel-pm! (preciso ir agora) 23:10
frederico_ ok
t +
fglock & 23:11
23:11 fglock left 23:15 stennie joined 23:17 macli left
rep DCC SEND "string" 0 0 0 23:23
oops, wrong channel.
23:27 netanya_goy left
nothingmuch seen stevan? 23:31
jabbot nothingmuch: stevan was seen 2 hours 50 minutes 5 seconds ago
nothingmuch =(
23:33 stennie left 23:39 mako132__ 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!