»ö« | perl6.org/ | nopaste: paste.lisp.org/new/perl6 | evalbot usage: 'perl6: say 3;' or rakudo: / pugs: / std: | irclog: irc.pugscode.org/ | UTF-8 is our friend!
Set by Juerd on 28 August 2009.
00:00 elton_mcz joined, tak_ joined, tak11 left 00:08 elton_mcz left 00:10 Whiteknight joined 00:13 ruoso joined, ruoso_ joined 00:15 ruoso_ left 00:20 frew_ joined, meppl left 00:38 ihrd joined, ihrd left 00:40 alexe_ joined 00:41 alexe_ is now known as alexe 00:43 sevvie joined, hercynium_ joined
mikehh rakudo (5960161) builds on parrot r411796 - make test / make spectest (up to r28216) PASS - Ubuntu 9.04 amd64 (g++) 00:47
rakudo - t/spec/S03-operators/arith.rakudo - TODO passed: 120, 131-132
rakudo - t/spec/S02-names_and_variables.rakudo - Non-zero wait status: 11 (Segfault after passing tests)
colomon On Complex.sin -- seems like more or less the same thing is happening to it as happened to Num.sin. Only difference is it looks like it really is a Complex being passed in. 00:56
Is it possible Complex.sin is calling a Parrot Complex.sin function rather than the Complex.sin defined in Complex.pm? I added a "say" in Complex.pm's sin, and it is never called, even on the Complex.sin calls that work. 00:57
00:57 hercynium left, hercynium_ is now known as hercynium, hercynium_ joined 00:58 hercynium left, hercynium_ is now known as hercynium 01:15 alexe_ joined, frew___ joined 01:16 frew_ left 01:17 alexe left, alexe_ is now known as alexe
colomon rakudo: say 0.0.cosh; 01:27
p6eval rakudo 446d49: OUTPUT«1␤»
colomon rakudo: say 0.0.cosh('degrees');
p6eval rakudo 446d49: OUTPUT«57.2957795130823␤»
colomon Ummm... that can't possible be right, can it? 01:28
01:29 frew___ left
TimToady did you mean arccos? 01:29
colomon No, I meant cosh, hyperbolic cos. 01:30
Rakudo is treating it as if the output of the function was an angle measure, and so adjusting the output by base.
TimToady hmm, that might work if the slope of the curve were .5 :) 01:31
er, 1 01:32
but somehow cosh never struck me as an identity function
wayland76 gets it -- cosh never struck me :) 01:33
TimToady we're civilized in America; we don't cosh people, we club 'em 01:34
01:34 dukeleto left 01:36 Whiteknight left
wayland76 Well, I learned about coshes from American comic strips :) 01:37
01:37 jaldhar left
colomon I can't recall ever hearing of "cosh" for a blackjack-type weapon before. 01:38
wayland76 Well, you obviously don't read the right kind of comics :)
I think it was a Phantom comic, possibly one older than I am 01:39
colomon I have been reading comics older than I am (by a few months, anyway) lately, but they were circa-1970 Iron Man issues. 01:40
Anyway, can someone out there confirm I am not insane? The hyperbolic functions take an angle as an argument, not return an angle. If that is correct, then both the implementation and tests for cosh (and presumably the others) are flat out wrong. 01:42
Take a "hyperbolic angle" according to wikipedia.
TimToady doesn't look like you can measure a hyperbolic angle in "degrees"... 01:47
interestingly, you can think of ordinary angles as areas in a sector of a circle 01:48
01:48 sri_kraih_ joined
TimToady it just happens to be linear in ϑ 01:49
anyway, I'd guess someone just cargo-culted the code from circular functions 01:52
colomon TimToady: where are you getting the degrees thing?
TimToady which degrees thing? when talking about ϑ I mean circular functions
colomon and if we use that logic, do we chuck out the different bases on Complex sin / cos as well?
"doesn't look like you can measure a hyperbolic angle in "degrees"" 01:53
TimToady it's a measure of area
01:53 agentzh joined
TimToady according to wikipedia 01:53
colomon "Just as a point on a circular segment sweeps out an area proportional to the circular angle, so too it can be shown that with the above definition a point on a hyperbolic sector sweeps out an area proportional to the hyperbolic angle" 01:54
Also according to wikipedia. ;()
TimToady sure, I just doubt that anyone measures it in degrees for hyperbolics
but I'm certainly not claiming to be an expert... 01:55
01:56 Caelum left
TimToady haven't thought about hyperbolic functions for 35 years... 01:56
01:56 Caelum joined
TimToady that was back when I was intelligent but stupid 01:56
colomon They don't seem to have a lot of day-to-day uses, for sure. 01:57
araujo guesses now TimToady is just .... intelligent?? 01:59
TimToady now I'm an idiot, but a very smart one
araujo hahaha 02:00
wayland76 I engage in hyperbole but don't think much about its function :)
TimToady I prefer to be elliptical
angles are conic sections too, but they don't get much respect 02:02
02:04 payload left, arthur-- left, rfordinal left, Confield left, cblaptop left, krakan left, betterwosld left, jql left, BooK left, chipdude left, hatseflats left, psychoschlumpf left, cj left, pioto left, |Jedai| left, tarbo2 left 02:05 BooK joined, hatseflats joined, cj joined, Jedai joined, sri_kraih left, sri_kraih_ is now known as sri_kraih 02:06 meppl joined 02:07 __ash__ joined
colomon phenny: tell pmichaud calls to Complex.sin appear to be rewired to calls to Parrot's Complex.sin function rather than Complex.pm's. Not sure if this is worth fixing right away, as am having existential crisis about whether Complex trig should support anything other than radians. 02:08
phenny colomon: I'll pass that on when pmichaud is around.
02:10 psychoschlumpf joined
__ash__ colomon did you try mapping Perl6 Complex to parrot's complex like how he did with Num and Float 02:11
02:16 hercynium left 02:22 payload joined, JimmyZ_ joined 02:40 payload left 02:45 Confield joined 02:52 stephenl1 is now known as stephenlb 03:05 jaldhar joined 03:09 cdarroch left 03:17 jaldhar left 03:23 sri_kraih left, jaldhar joined 03:24 frew_ joined 03:27 sri_kraih joined 03:28 jaldhar left 03:29 jaldhar joined 03:30 dukeleto joined 03:39 mepplock joined 03:40 japhb left 03:46 rez2009 joined 03:47 rez2009 left 03:55 meppl left 04:08 mepplock left 04:10 __ash__ left 04:14 tak_ left, synth left 04:23 dukeleto_ joined 04:28 jql joined 04:30 zloyrusskiy joined 04:32 dukeleto left, dukeleto_ is now known as dukeleto 04:40 meppl joined 04:51 kent\n left 04:53 kent\n joined 05:01 iblechbot joined 05:09 Avada joined, Avada left 05:18 sundar_ joined
pmichaud phenny: tell colomon that I'm thinking that perhaps we should ignore Parrot's built-in Complex type altogether and just write our own from scratch in Perl 6, similar to how we did Rat 05:19
phenny pmichaud: I'll pass that on when colomon is around.
pmichaud: 02:08Z <colomon> tell pmichaud calls to Complex.sin appear to be rewired to calls to Parrot's Complex.sin function rather than Complex.pm's. Not sure if this is worth fixing right away, as am having existential crisis about whether Complex trig should support anything other than radians.
Tene I'm working on one of the facebook puzzles, and some of this stuff would be much easier in Perl 6. :) 05:23
05:25 agentzh left 05:26 Confield left 05:34 justatheory left 05:36 agentzh joined
sundar_ Hi.. I'm having some issues compiling parrot under Windows. Is there any guide or anything specifically for Windows? 05:42
05:48 dukeleto left
wayland76 sundar_: I recommend the parrot channel on irc.parrot.org -- all the Parrot people hang out there 05:48
( #parrot channel, I should've said) 05:49
05:53 frew_ left 05:54 tilt joined
tilt \o/ 05:54
can perl6 connect to mysql? :P 05:56
does it have a CGI module? :)
dalek kudo: a0494f8 | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION.
 06:05
06:11 japhb joined
Tene tilt: Perl 6 can connect to SQLite. There's been some work on MSql support, but it's not quite there yet. 06:18
agentzh Tene: really awesome... 06:19
Tene tilt: For CGI, the answer depends on exactly what you want to do. The Perl 5 CGI module is very large, and includes a lot of stuff. Much of it is available in different individual libraries in Perl 6.
tilt: You should probably start looking at the web.pm repo, although remember that it's a work in progress, just like everything else around here. :)
Tene considers erecting one of those 1995-era "Construction In Progress" animated GIFs. 06:20
(give the anti-camelia crowd something legitimate to complain about...) 06:21
06:28 Bzek joined 06:30 rfordinal joined 06:34 iblechbot left 06:40 orafu left, orafu joined 06:49 rfordinal left 06:59 Hammit joined 07:02 stephenlb left 07:07 rbaumer left 07:09 pdcawley left 07:11 carlin joined 07:17 sundar_ left 07:19 mj41_ left 07:26 mikehh left 07:28 mj41 joined 07:37 rfordinal joined
Matt-W Morning 07:49
agentzh good localtime :) 07:50
moritz_ oh hai
araujo iah ho 07:52
;) 07:53
agentzh is always waiting to use perl 6 extensively in $production :D 07:55
though looking at coco lua atm :P 07:56
07:57 icwiener joined
araujo coco lua? 07:58
agentzh araujo: coco.luajit.org/ 08:00
agentzh is currently pondering merging it into nginx's event model.
08:03 KyleHa joined 08:14 hanekomu joined
JimmyZ_ rakudo: my int $y = undef; 08:21
p6eval rakudo 446d49: OUTPUT«Malformed declaration at line 2, near "int $y = u"␤in Main (src/gen_setting.pm:3469)␤»
JimmyZ_ rakudo: my int $y = Nil;
p6eval rakudo 446d49: OUTPUT«Malformed declaration at line 2, near "int $y = N"␤in Main (src/gen_setting.pm:3469)␤» 08:22
JimmyZ_ rakudo: my int $y = 'a';
p6eval rakudo 446d49: OUTPUT«Malformed declaration at line 2, near "int $y = '"␤in Main (src/gen_setting.pm:3469)␤»
JimmyZ_ rakudo: my int $y = 1;
p6eval rakudo 446d49: OUTPUT«Malformed declaration at line 2, near "int $y = 1"␤in Main (src/gen_setting.pm:3469)␤»
JimmyZ_ rakudo: my int $abc = 1;
p6eval rakudo 446d49: OUTPUT«Malformed declaration at line 2, near "int $abc ="␤in Main (src/gen_setting.pm:3469)␤»
moritz_ JimmyZ_: "malformed declaration" has nothing to do what you put on the right side
JimmyZ_: it means that rakudo can't cope with left side 08:23
JimmyZ_ rakudo: my $abc = 1;
p6eval rakudo 446d49: ( no output )
JimmyZ_ moritz_: rakudo doesnot support int? 08:24
moritz_ JimmyZ_: no
no native types at all
JimmyZ_ rakudo: my Int $abc = 1; 08:25
p6eval rakudo 446d49: ( no output )
JimmyZ_ oh, I thought It had.
08:26 patmat_ is now known as patmat, hanekomu left 08:31 szabgab left 08:38 dukeleto joined 08:55 ihrd joined, ihrd left 08:56 KyleHa left 08:59 rfordinal left 09:12 cmv joined 09:17 payload joined 09:29 Jedai left, Hammit left 09:32 M_o_C joined 09:37 masak joined
masak good late morning, cheese speleologists. 09:37
phenny masak: 10 Sep 17:13Z <pmichaud> tell masak I expect us to have a decent REPL by December. Probably sooner than that.
masak sounds good.
phenny: tell pmichaud that when I run `parrot examples/pge/demo.pir`, enter 'regex a*' and then 'pir', I get a Null PMC access in get_string(). has something bitrotted? 09:38
phenny masak: yeah, sure, whatever
masak o_O 09:39
JimmyZ_ Is phenny a bot/
masak I used to think so... :)
M_o_C It should be.
JimmyZ_ phenny: Are you a bot?
rakudo: constant Num π = atan2(4,4) * 4; π.say; 09:41
p6eval rakudo 446d49: OUTPUT«3.14159265358979␤»
JimmyZ_ Can perl6 ouput more precision? 09:42
masak doesn't know 09:47
JimmyZ_ :( 09:48
09:48 alexe left
carlin from inamidst.com/phenny/modules/tell.py: elif rand > 0.999: response = "yeah, sure, whatever" cute :) 09:49
masak one in a thousand. :)
09:58 Whiteknight joined 10:08 Whiteknight left
Matt-W Reminds me of a power I got in Champions Online last night. Experimental Blaster: 1% chance of turning target into a teddy bear 10:10
masak that would explain all the hugging on this channel. 10:12
Matt-W :) 10:14
10:14 sharada joined
carlin hugme should have a 1 in 1000 chance to refuse to hug someone ;) 10:19
Matt-W no that would be wrong 10:21
hugme: hug carlin 10:22
hugme hugs carlin
10:32 JimmyZ_ left 10:33 pmurias joined
pmurias if i have clone a repo from github is it possible to have git reuse data when cloning my fork? 10:38
* cloned
masak doubts it 10:39
though I'm pretty sure github does that, internally. they would be foolish not to.
pmurias i remember reading that they do it 10:40
10:44 ruoso left
pmurias fatal: github.com/pmurias/perl/tree/git@gi.../info/refs not found: did you run git update-server-info on the server? 10:45
what the heck?
my fault 10:51
moritz_ pmurias: if you just 'remote add' the second clone it should reuse packs 10:52
10:52 araujo left
moritz_ if you make a completely independent clone, not (obviously) 10:52
10:54 cognominal left 10:58 payload left 11:02 cognominal joined 11:38 spachy joined 11:40 szabgab joined, spachy left 11:42 szabgab left, szabgab joined 11:44 synth joined 11:45 mj41 left, icwiener_ joined 11:47 mj41 joined 11:51 icwiener_ left 11:55 jaldhar left 11:57 payload joined 12:01 icwiener left
masak suddenly I'm not so sure Bielefeld exists. 12:02
moritz_ masak: it doesn't.
masak see? :)
moritz_ at least I don't believe what THEY tell me
masak probably wise. 12:03
moritz_ @seen Su-Shee 12:06
phenny: seen Su-Shee 12:07
where are the bots when you need them?
wayland76 see? lambdabot was good for something after all :)
masak hugme: hug Su-Shee
hugme hugs Su-Shee
masak the important bot is here :)
wayland76: oh lambdabot is quite OK, albeit somewhat chatty. it's purl that's the great nemesis. 12:08
12:11 takadonet joined
takadonet morning all 12:11
moritz_ @foo 12:12
masak @morning, @takadonet. @@@ 12:13
M_o_C moritz_: At least phenny is in a bad mood today. 12:14
moritz_ he, talking about random features for hugme... 12:20
randomly refusing hugging might be a bad idea
masak nod.
moritz_ but it could look like "*hugme hugs masak and flushes" or so 12:21
masak randomly giving extra hugs might be a good idea :)
moritz_ yes 12:22
uhm, "to flush" is very ambiguous 12:23
it can mean become red in the face
or you can flush a toilet
masak "blush" might be better.
moritz_ indeed 12:24
masak in fact, the more variations on hugs, the better, I think. 12:25
"* hugme gives moritz_ a long, warm hug"
"* hugme jumps forward and hugs moritz_ forcibly" 12:26
"* hugme takes moritz_ in its arms and applies an excellent hug"
moritz_ in rare cases, hugme could also handle out positive karma ;-) 12:27
masak aye. 12:28
12:32 szabgab left, szabgab joined
masak spectesting today took 88 minutes. :( 12:33
moritz_ did it produce many exit status 11?
masak that, too.
a frightening amount, actually. 12:34
some of them are 10's.
and 4's.
moritz_ masak: then could you please comment on trac.parrot.org/parrot/ticket/998 to tell them that it happens on your platform too?
masak will do.
moritz_++
is the slowness related to this, you think? 12:35
moritz_ no
slowness and non-zero exit status appeared at different parrot revisions 12:36
sadly I was having a really hard time bisecting them
colomon + 12:37
phenny colomon: 05:19Z <pmichaud> tell colomon that I'm thinking that perhaps we should ignore Parrot's built-in Complex type altogether and just write our own from scratch in Perl 6, similar to how we did Rat
moritz_ that would make it easier to have different slot types 12:38
ie a Complex of Rats or so
moritz_ likes the idea
colomon Huh. I don't know if it was just getting a good night's sleep or what, but all my doubts about how to handle these issues seem to have dissolved this morning. :) 12:40
masak great! 12:43
pugs_svn r28217 | colomon++ | [t/spec] Unfudge tests for Num.sin / Num.cos with explicit base, fixed by pmichaud++'s commit d0355a5dde0973e01834d58de099e53b97fe70a6. 12:44
colomon May just mean I am confidently wrong now, but what the heck! [:) 12:45
12:45 jaldhar joined
masak I've gotten a lot of productive things done by being confidently wrong. 12:45
as long as you're still impressionable, being confident in any direction is usually a boon.
takadonet loves using multiple cores to run the spectest 12:52
moritz_ me too
it makes testing more convenient, but it doesn't solve the underlying problem
masak something has definitely happened. testing used to take under 30 minutes here. now almost 90. 12:53
colomon Glad that slowdown wasn't just my imagination! 12:54
moritz_ would be glad if the slowdown was just imaginary 12:59
13:00 nihiliad joined 13:02 tilt left
moritz_ anyway, until these errors-at-exit are fixed I can't meaningfully work on the test suite without breaking rakudo 13:02
the test summary is just useless if you have to scroll several pages, reading carefully if there are any new/relevant ones
colomon I wonder why I'm not getting these errors at exit? 13:03
13:03 ruoso joined
colomon (On the other hand, it seems like I keep on getting bus errors after running standalone p6 scripts...) 13:03
moritz_ colomon: which platform (CPU, OS) are you on?
colomon OS X Intel. 13:04
moritz_ masak can reproduce them - bacek can't
13:06 zloyrusskiy left
moritz_ maybe related to compiler versions? 13:10
13:10 carlin left 13:18 kidd joined 13:20 agentzh1 joined, ab5tract left 13:28 agentzh left 13:30 ssm joined 13:35 payload left
masak urgh, and now I get an ugly 'failed assertion' when running the squerl test file. :( 13:37
13:40 Gothmog_ left 13:44 carlin joined, JimmyZ joined 13:45 pmurias left, pmurias joined 13:47 Gothmog_ joined 13:57 alester joined 13:59 Confield joined 14:02 krunen joined 14:07 patspam joined 14:08 JimmyZ left
masak rakudo: ((Temporal::DateTime.new(date => Temporal::Date.new(:year(2010), :month(4), :day(1)), time => Temporal::Time.new(:hour(0), :minute(0), :second(0))).epoch - time) / (60 * 60 * 24)).ceiling.fmt("%d days left until April!").say 14:09
p6eval rakudo 446d49: OUTPUT«Use of uninitialized value␤Use of uninitialized value␤Use of uninitialized value␤Use of uninitialized value␤Use of uninitialized value␤202 days left until April!␤»
14:16 Bzek_ joined
colomon I'm a little disturbed at the moment by how many of the trig.t tests are actually wrong.... 14:16
(And correspondingly, by how many of the trig implementations in Rakudo are wrong, because they are passing the tests.) 14:17
14:19 perletc joined 14:22 patspam left, patspam1 joined 14:24 justatheory joined, Bzek_ left
krunen std: class A {...}; class A {} 14:25
p6eval std 28217: OUTPUT«===SORRY!===␤Illegal redeclaration of package symbol A at /tmp/yzK5PhGbLr line 1:␤------> class A {...}; class A⏏ {}␤FAILED 00:02 37m␤»
14:26 Bzek left
krunen hmm.. no predeclaration yet 14:26
frew__ what's the recommended way to get rakudo running on win32? 14:30
colomon frew__: If you download that "Padre Standalone Plus Six" thing, it has rakudo in it. 14:36
Probably about a month old now, but that's the easiest way I know of to get it. 14:37
rakudo: atan(:x(1))
p6eval rakudo 446d49: OUTPUT«No applicable candidates found to dispatch to for 'atan'␤in Main (/tmp/8Jp1xqLOyp:2)␤»
szabgab colomon, frew__ next release of Almost six is due after the next release of Rakudo 14:38
frew__ ah, cool 14:39
colomon szabgab++ 14:41
14:43 payload joined 14:48 Psyche^ joined
colomon As long as I'm overhauling trig stuff -- does the "named args" section of S32-trig/trig.t make sense to anyone? 14:51
It's all skipped in Rakudo, and as near as I can tell, it all depends on the name of the first argument to the trig functions. 14:52
Because of the way they are currently implemented in Rakudo, these tests will never pass, and they seem pretty daft to me. 14:53
14:54 kidd left 14:55 szabgab left 15:01 __ash__ joined
masak as a parallelism layman, I enjoyed research.sun.com/projects/plrg/Publ...Steele.pdf -- perhaps others interested in parallelism will, too. 15:01
actually, it's enjoyable just to leaf through a talk littered with Lisp, Haskell and Fortress code. :) 15:02
frew__ colomon: if I remember correctly that's how the named args are for all the functions 15:03
colomon: I think the named args for those are silly, but I added them into the tests because they are in the spec 15:04
15:04 agentzh1 left, SmokeMachine joined 15:05 Patterner left, Psyche^ is now known as Patterner, justatheory left 15:07 nihiliad left
colomon In the spec where? 15:08
15:08 patspam1 is now known as patspam
frew__ lemme look 15:09
15:09 jferrero joined
frew__ svn.pugscode.org/pugs/docs/Perl6/Sp...umeric.pod 15:11
method atan2 ( Num $y: Num $x = 1 ) # I figured that since it's documented, we should test it 15:12
carlin urgh, I'm getting a bunch of random seg faults 15:14
masak carlin: during spectesting? 15:22
carlin masak: Nope, while running scripts. Usually on exit after the script has run, but I had one half-way through before 15:23
masak I've had many on exit today as well. 15:24
less than a week 'til next release, eh? :)
colomon frew__: Next time TimToady shows up I'll check with him, but I'd be surprised if that were his intent for the rest of the trig functions. atan2 is a very exceptional function. :) 15:25
pmichaud good morning, #perl6
phenny pmichaud: 09:38Z <masak> tell pmichaud that when I run `parrot examples/pge/demo.pir`, enter 'regex a*' and then 'pir', I get a Null PMC access in get_string(). has something bitrotted?
masak pmichaud: morning. what phenny said.
pmichaud masak: I suspect that demo.pir has indeed bitrotted 15:26
masak ok.
pmichaud better is probably to use Perl6Grammar
pmichaud checks
colomon Good morning pmichaud!
15:27 payload left
pmichaud nopasting... 15:27
gist.github.com/185365 # new way to do demo.pir stuff 15:29
15:29 payload joined
pmichaud (Perl6Grammar has a REPL :) 15:29
masak ah. 15:30
works fine here too. thanks. 15:32
pmichaud although something is wrong with the example I just gave -- since it's a token we shouldn't be building a stack for backtracking
masak indeed.
15:33 brunov joined
brunov Hi there! 15:33
I'm reading s02
And either I don't quite understand something, or there seems to be an incosistency regarding embedded comments and the definition of bracket characters 15:34
masak uh huh.
what inconsistency?
brunov In s02, "lexical conventions", it says "Characters with no corresponding closing character do not qualify as opening brackets."
masak snowmen, for example. 15:35
scnr. :)
brunov further down, when defining embedded comments, it says "comments require the "#" to be followed by a backtick ("`") plus one or more opening bracketing characters."
but then the third example, the embedded comment uses the character "X" as a bracket
and since "X" doesn't have a closing characer, it shouldn't count as a bracket, right? 15:36
pmichaud third example isn't an "X" 15:37
brunov oh
TimToady probably missing char in your font
brunov oooooooh
*_*
dumb terminal
sorry! 15:38
masak I see a 「 here.
brunov ok, makes sense now. Thanks everyone!
masak brunov: no problem! see if you can come upon a UTF-8-aware console.
15:38 zloyrusskiy joined
masak it's worth it. :) 15:38
brunov ok, I'm using Ubuntu's default, which I think it is gnome-terminal. I'll probably have to switch if I want to read these correctly. Thanks again! 15:39
avuserow just successfully demoed a small CompSci project written in Perl 6
TimToady I use gnome-terminal with utf-8
probably just some config somewhere 15:40
LANG or something
brunov TimToady: cool, I'll definitely look into it
masak might be fonts, too.
avuserow: that's great! do you have the source somewhere, so we can ooh and aah?
pmichaud avuserow: or even... a blog posting? ;-) 15:41
avuserow hm. I could blog about it. I'll just need to clean up a bit and get rid of the deadline cruft. :) 15:42
masak avuserow: just don't overdo it! release early! :)
carlin hmm, I also see an X in perldoc from gnome-terminal but 「in vim 15:43
masak vim is correct.
TimToady brunov: for the moment you might want to read the html versions under perlcabal.org/syn 15:44
15:46 zloyrusskiy left
brunov TimToady: yes, thanks! Some POD markup will be nowhere near as harmful as not having the proper characters rendered in UTF-heavy docs like these 15:47
TimToady but I understand the desire to read the pods :)
especially if you want to edit them too... 15:48
15:48 nihiliad joined
TimToady speaking of which, do you have a commit bit to the pugs repo, where the pod is stored? 15:48
brunov I'd be happy to!
15:48 jferrero left
TimToady then /msg me your email and preferred nick, and I can send a commit bit 15:49
(svn nick)
brunov TimToady: I don't think so, no. I've been lurking the channel/blogs and such, but just today I decided to get started reading the synpopsys
15:49 __ash__ left
TimToady well, it's just nice to get instant gratification when the itch to edit comes upon you, is all :) 15:49
but anyhoo, any time you want a commit bit, show up here, and someone can help, likely 15:50
masak is re-reading the Apocalypses 15:51
colomon TimToady: Speaking of messing with the specs (and whatnot), I have just locally rewritten all the hyperbolic trig functions to allow you to specify them in degrees, gradians, etc.
I decided it was more consistent, even if most sane people will probably avoid using them in anything but radians. 15:52
pmichaud reads backscroll
TimToady colomon: more bothersome to me is all the string args where the spec uses an enum 15:53
colomon I will work on cleaning that up once the functions work right. :)
sometime next week, probably.
15:54 masak left
colomon Once my changes have more completely infected the trig.t file, it will be easy to do the switchover there. 15:55
TimToady what was the Q about atan2? 15:57
15:58 agentzh joined
colomon TimToady: The spec specifies a variable name for every argument to every trig function. frew__ wrote up a bunch of tests to make sure you could use named arguments for them all. 15:58
My feeling is that the only such test that really makes sense is the atan2 test, because the $x and $y arguments do not come in the natural order, and really mean something. 15:59
s/test/tests/
For the other functions, the $x argument is just a placeholder.
TimToady yeah, it's sort of redundant testing of a general property of the language 16:00
colomon Well, it's a test rakudo completely fails at the moment, as none of the trig functions are implemented in terms of $x. 16:01
16:02 meppl left
pmichaud also Rakudo doesn't know how to accepted named arguments into positional parameters yet 16:02
*accept
TimToady yes, it doesn't do the general property in question
colomon pmichaud: Do we intend to re-write the trig functions into that format once it does? 16:03
TimToady on the plus side, you'll get a lot more passing tests when it is implemented :)
pmichaud ...into what format?
TimToady you're confusing named args with enums
colomon pmichaud: write now it is our Num multi method acosh($base = 'radians') is export
pmichaud the 'radians' part needs to change, yes. (more) 16:04
colomon According to the spec it should be Num multi method acosh ( Num $x: TrigBase $base = $?TRIGBASE ) is export
pmichaud the spec is correct :-)
colomon Not the radians, the Num $x: bit
TimToady pretty much all the numeric functions try to stick with $x and $y
pmichaud yes, we'll want to do that 16:05
colomon (Even putting a space between "func" and the open parenthesis?)
pmichaud otherwise acosh(:x(...)) won't work
16:05 justatheory joined
pmichaud the space isn't critical :) 16:05
colomon Then I vote we leave the tests.
Though I suspect a bunch of them will need to be rewritten.
TimToady which space are you referring to, def, or call? 16:06
colomon A huge quantity of the trig tests were wrong. I've got a big patch to push through as soon as I've run a proper spectest. (So in a couple of hours, at current speeds.)
TimToady: "func (" is what the spec says.
I would have thought it would be "func(".
pmichaud I might also suggest eliminating the "Num" constraint on $x, but I'm not as picky about that. 16:07
TimToady the space doesn't matter on the definition side
colomon TimToady: Interesting and confusing. :)
pmichaud: Oh, absolutely! We've already implemented a few of them for Any and Complex as well, after all....
TimToady I suspect common practice will eventually be to omit the space, but historically signatures have usually had the space in most languages
colomon must go eat lunch. Back soon. TimToady++, pmichaud++ 16:08
TimToady interestinly though, on the call side, with a space sin (...) turns into a unary named prefix, not a listop, according to S03 16:10
colomon How will take interact with the function definitions? 16:11
TimToady which probably means the current multi method is wrong, and should be prefix:<func> instead, unless we change our mind about supporting traditional math notation 16:12
func() is always a function call
func ... is named unary, same as in Perl 5, for trig ops
but most of the other named unaries have gone away
don't know how confusing it would be if trig functions turned into listop precedence 16:13
it's also a little confusing if they *don't*
colomon Does that imply that the multi method should not have $base as an optional parameter?
(afk again) 16:14
TimToady operators can have named args
pmichaud how about optional positionals?
TimToady but you'd have to call it either as a function, or with a :base adverb
pmichaud okay.
TimToady but if there's a good chance of calling it with multiple args with some frequency, that's a good argument for making it listop, not unary 16:15
16:15 rfordinal joined 16:16 ruoso left
TimToady if the mathematicians would be happy, I wouldn't mind blowing away named unary precedence, but that's a hard "if" to predict 16:16
16:16 agentzh1 joined, agentzh left
TimToady it would break things like: if sin $x == 2 {...} 16:17
well, not that that would ever be true :)
Juerd Aw, but overriding sin can be so fun! 16:18
TimToady but the precedence would be surprising if sin were a listop
16:21 cdarroch joined
avuserow masak, pmichaud: auzon.blogspot.com/2009/09/applying...jects.html 16:22
16:23 cmv left, cmv joined
PerlJam avuserow: That's your blog? 16:24
avuserow Yep 16:25
PerlJam avuserow++
and again avuserow++
avuserow @karma auzon
hm. forgot how the bots here work
16:26 carlin left 16:30 cotto left
pmichaud avuserow++ # nice post! 16:33
PerlJam avuserow: did you also code a perl5 solution for your prof? 16:35
16:35 frederico joined 16:37 sevvie left
TimToady does MAIN actually work in rakudo now? 16:37
avuserow PerlJam: no. I did shell out to Perl5 in one part to handle binary data
TimToady: yes. MAIN seems to work quite well.
PerlJam avuserow: Is he going to actually run your program?
TimToady heh, dint knowit 16:38
avuserow PerlJam: no, it was strictly "do you have code?" and "did you demonstrate it?"
pmichaud ...multi MAIN works in rakudo?!? 16:39
hmmmm.
TimToady why didn't you use gather/take instead of push?
16:39 nihiliad left
avuserow didn't remember about gather/take 16:39
TimToady oh wait, then it wouldn't be a scripting language
it would be FP :) 16:40
ah well, tmtowtdi
16:41 icwiener joined, nihiliad joined
avuserow the assignments in this class are very loosely speced, so I could've done functional without them having a problem 16:41
TimToady well, I hope your prof is impressed :) 16:42
16:43 agentzh1 left
pmichaud gist.github.com/185407 # multi MAIN does work in Rakudo! jnthn++ 16:43
actually, jnthn++ probably didn't even realize that the work he did caused this to start working :) 16:44
it just falls out of the other semantics
16:44 patspam left
TimToady jnthn++ and everyone-else++ 16:44
pmichaud we don't do option processing yet, though
avuserow jnthn++ # for multi-dispatch, especially for MAIN 16:45
16:45 Chillance joined
colomon frew__++ # For adding the positional trig tests 16:49
16:51 patspam joined 16:54 pyrimidine joined
takadonet hey pyrimidine 16:56
pyrimidine o/
16:58 cmv left
Tene pmichaud: github is going to be migrating providers soon, and it looks like the transition might not be without problems. 17:03
pmichaud Tene: url?
PerlJam It's a good thing git makes complete copies of the repo on clone :)
Tene www.engineyard.com/blog/2009/engine...ransition/ 17:04
17:06 abra joined
PerlJam "In fact, the transition from Subversion to Git has been one of the most aggressive technology transitions that I’ve ever seen!" That's so true, but really an understatement IMHO 17:06
(regarding rails)
pmichaud well, I don't know what sort of problems we're likely to encounter 17:07
anyway, the rakudo account is a paid account, so hopefully it'll get a measure of improved service over a free one
Tene pmichaud: Sure. Ideally, there won't be any problems. 17:10
Just wanted you to not be surprised if they have problems running on the new provider during th emiddle of the release, or something.
pmichaud is there a transition date published somewhere?
Tene I can't find it mentioned anywhere on github's website. 17:11
pmichaud Me neither.
Tene Although, that post is about an hour old. 17:13
pmichaud well, we'll keep an eye on it. :)
I suspect neither github nor engineyard want any black eyes here :)
Tene I'm a little bit irked by "... we didn’t want to subsidize that quota for non-Ruby developers." 17:15
pmichaud really? why?
frew__ I agree 17:16
well, they seem to have billed themselves as free for all Open Source is all 17:17
pmichaud github did that, not engineyard
frew__ I think I misunderstood soemthing 17:18
EY doesn't own GH?
pmichaud EY *hosts* GH
frew__ ok
my bad
pmichaud I don't think they own them.
frew__ I take it back
they can subsidize whoever they want then
:-)
pmichaud afaict, engine yard has been providing free(ish) server hosting for github
frew__ got it 17:19
Tene pmichaud: I don't actually object, and they can do whatever they want, it just sounded a little weird... 17:20
pmichaud EngineYard is explicitly a ruby site for ruby-hosting.
frew__ I think that for those of us not intimate with EY, GH, and rails the article is hard to understand 17:21
right
pmichaud I think it would be very hard to write an article that is easily understood. :)
frew__ I'm not criticizing
pmichaud it will be interesting to see what github has to say about it :)
frew__ I'm saying we are on the outside so it makes it hard
pmichaud sure
anyway, as I said, Rakudo has a paid account, so hopefully we'll get some special care 17:22
frew__ do we need special care with anything?
pmichaud I'd like to make sure we have as few service interruptions as possible :)
frew__ ah, right
Tene github and Engine Yard often disagree with each other, from what I've seen. GitHub posts announcements saying "We're down because of engineyard fails" and engineyard responds with "No, you're down because your app can't scale and will fail entirely when doing more than two repository exports at a time" 17:23
pmichaud or, put another way, github isn't likely to be around long if they annoy their paying customers :)
frew__ you know you could do a mirror to gitorious or something, and because git works the way it does it would be easy to just mirror back and forth
I think that's what Moose does (athough it's shadowcat instead of gitorious) 17:24
17:26 sevvie joined 17:32 takadonet left
pugs_svn r28218 | colomon++ | [t/spec] Try to create proper test cases for the "forward" hyperbolic trig functions, as well as a tan and cotan. 17:36
colomon Okay, I've just pushed the morning's big trig patch. :)
If my $work continues to involve long compiles, there may be an afternoon patch as well.
dalek kudo: 84296e4 | (Solomon Foster)++ | src/setting/Num.pm:
Fix implementations of the "forward" trig functions.

Note that many if not most of the inverse trig functions remain broken.		 17:37
17:52 jonathanturner joined, PZt left 17:53 iblechbot joined 17:57 nihiliad left, nihiliad joined 18:01 frederico left
TimToady phenny: tell masak, I particularly like the slide with "(this step requires ingenuity)" which is obviously the ??? before Profit! 18:04
phenny TimToady: I'll pass that on when masak is around.
18:05 patspam left
colomon Is there a way to make git automatically do a git diff --check before committing? Apparently I cannot remember that on my own... 18:05
TimToady phenny: tell masak also, we're missing the list split primitive; I wonder if two more or slurpy arrays in a binding should auto-split a list 18:10
phenny TimToady: I'll pass that on when masak is around.
Tene colomon: chmod +x .git/hooks/pre-commit 18:12
18:14 xenoterracide left
colomon Oooooooooo 18:15
18:16 __ash__ joined
colomon Tene++ # I'd have known it myself if I'd just read the next chapter of Pro Git. :) 18:17
__ash__ Is the right way to multi-thread in Parrot a Task?
colomon Wow, easy to set up client side hooks might be enough to get me to switch to git for work.
Tene __ash__: no, not as I recall. 18:18
My vague memory is that tasks are run in the same thread they were created from, usually, in parrot right now.
__ash__ ah, okay, i was reading docs.parrot.org/parrot/latest/html/...y.pod.html and it gave me that impression 18:19
Tene you want to use a ParrotThread.
__ash__: *please* feel free to try it and let me know what you find out.
__ash__ I was planning doing some tests with it, making Tasks and all in parrot 18:20
Tene I was rathe runable to sort out the various bugs with Threads and HLLs in Parrot the other day. :(
__ash__ I'll let you know, i have been reading over that and the tests and the Task core stuff, to see what I can learn, just thought i'd ask in case someone knew already
18:21 patspam joined 18:23 patspam left, patspam joined 18:24 patspam left 18:25 patspam joined
pugs_svn r28219 | lwall++ | [S06] supply missing list splitting primitive as a form of binding, masak++ for 18:29
r28219 | cf research.sun.com/projects/plrg/Publ...Steele.pdf
18:29 zloyrusskiy joined
__ash__ from PIR can you get a block as a function pointer? like a pointy block? 18:35
TimToady phenny: tell ruoso please read the Guy Steele slides mentioned above and think about hyper iterators in terms of balanced (more or less) list splitting 18:36
phenny TimToady: I'll pass that on when ruoso is around.
18:40 Jedai joined
__ash__ Tene: yeah, Tasks's aren't parallelized by default, they go to the built in schedule of the current instance of parrot and lets parrot do whatever it wants, so you can make a schedule that uses green threads or pthreads or forks or whatever you want, kind of nice i guess, but if you want a real posix thread then you have to use parrotthread 18:55
Tene __ash__: did you figure out the block thing? 18:59
19:02 szabgab joined 19:03 justatheory left 19:05 justatheory joined 19:10 justatheory left 19:13 zloyrusskiy left 19:19 __ash__ left 19:30 justatheory joined 19:34 M_o_C left 19:36 frew__ left 19:39 abra left 20:00 lambdabot left 20:03 lambdabot joined 20:12 payload left
dalek kudo: 3b63817 | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 436 files, 14267 (69.2% of 20603) pass, 5 fail

S32-str/samecase.rakudo aborted 5 test(s)		 20:13
20:15 icwiener left 20:22 jaldhar left 20:23 patspam1 joined, patspam left 20:28 thisguy joined 20:30 cmv joined, cmv left 20:31 __ash__ joined, cmv joined
__ash__ Tene, no i never figured out the block thing 20:33
20:33 |Jedai| joined
Tene __ash__: are you in #parrot on irc.perl.org ? 20:34
__ash__ no, irc.perl.org likes to get mad at me, sometimes wheni connect from my school campus it just hangs, i'll try
lisppaste3 tene pasted "parrotthread example for __ash__" at paste.lisp.org/display/86957
20:36 |Jedai| left 20:37 thisguy left
__ash__ Tene: do you know how to get a block's invoke function in perl6? like, specifically what I am doing is a simple async function that takes a block and then gives it to your example to execute 20:40
20:41 payload joined
lisppaste3 tene pasted "Perl 6 async implementation for __ash__" at paste.lisp.org/display/86958 20:41
Tene __ash__: like that?
20:42 |Jedai| joined 20:46 perletc left, Jedai left
__ash__ yeah, like that, i didn't know you could do it that way, intersting 20:48
Tene __ash__: you don't even need the inline PIR... I just used that so I coul dmake a parrotthread. 20:49
20:50 pyrimidine left, PZt joined
__ash__ hm... thats a lovely debug message 20:50
Tene __ash__: oh, by the way, that doesn't work. 20:51
Because of Parrot's current problems with threads.
__ash__ thats the wrong way to do it?
Tene forgot to mention that...
__ash__: it's the right way to do it, but Parrot is broken.
__ash__ okay, thats what I was curious about
anyone have any ideas about whats going on? 20:52
is it that ParrotThread run_clones from within a HLL don't retain the HLL?
20:53 simcop2387 left 20:54 simcop2387 joined
Tene __ash__: the problem is that Parrot just does the wrong thing when it's trying to clone the interpreter. 20:57
__ash__ but didn't the pure PIR version work?
Tene __ash__: the pir version I posted earlier does a lot less. 20:58
It doesn't have PIR classes, it doesn't have any dynamic libraries loaded...
and those are the two things that I've run into it failing to copy correctly.
When it enters the code path that's supposed to copy those, it does something broken instead, or runs into some data it' snot expecting, o rsomething. Those code paths are completely untested, afaict. 20:59
__ash__ Hmmm, sounds like some simple tests would be helpful then, so, its specicially dynlibs and PIR classes? 21:00
Tene __ash__: I'm not entirely sure of the details. For the former, look at ticket 757 in parrot's trac.
for the latter, add this at the beginning of the .pir:
.loadlib 'libsqlite3' 21:01
or any other .so on your system
__ash__ hm... mine didn't crash, i ran it 3 times, i added the .loadlib outside of the main sub though, should it be inside the main sub? 21:02
Tene no, outside.
21:03 frew_ joined
Tene try adding this as the first line inside the main sub: 21:04
.include 'test_more.pir'
__ash__ What system are you on btw? I am on OS X
Tene linux x86_64
__ash__ do i need to touch a test_more.pir?
Tene no, it will be loaded from Parrot's include directory. 21:05
__ash__ alright
Tene do you actually have libsqlite3.so ?
__ash__ that one failed for me
yeah, i have libsqlite3.so
Tene Hmm.
The issue revealed by including test_more.pir has a possible fix in a pair of patches on tt757 21:06
I don't understand what's going on well enough to be confident applying them.
__ash__ i have /opt/local/lib/libsqlite3.dylib and /usr/lib/libsqlite3.dylib (OS X uses .dylib not .so)
Tene nods.
21:16 sharada left 21:20 |Jedai| left 21:28 alester left 21:29 Whiteknight joined 21:34 justatheory left 21:35 sharada joined, zamolxes joined
pugs_svn r28220 | fglock++ | mp6 - simplified 'make' inside tokens 21:37
21:39 patspam1 left 21:52 __ash__ left 22:00 rdice joined 22:05 pmurias left 22:11 frew_ left 22:12 justatheory joined 22:16 jdv79 joined
jdv79 just wondering if stream and exception mgmt is better in p6? 22:16
pmichaud Yes. 22:17
TimToady but define "stream"
jdv79 handles like pipes and file handles
TimToady and then define "and" 22:18
jdv79 they are only related in my mind cause i've hit annoyances in p5 with them
they are not related. just grouped in the same question.
TimToady whew! :)
any particular annoyances you want to know about? 22:20
esp with streams, since we think we have the exceptions very much under control 22:21
22:21 cmv left
jdv79 just capturing, redirecting, and tying them. esp. the STD handles. 22:21
TimToady most of that now falls naturally out of "context variables" 22:22
jdv79 it seems like the abstraction isn't high enough so different ways of seemingly doing the same thing cause odd errors 22:23
TimToady well, there will always be issues syncronizing the abstractions with the various OSes ideas of concrete
*synchronizing 22:24
jdv79 i did see a blurb about context vars lately. sounds very nice.
TimToady certainly redireciton we can do on a subcall by subcall basis
and filehandles are just objects, so it's pretty easy to substitute anything else that implements the correct roles
filehandles aren't "special" symbols like they are in P5 22:25
jdv79 ok, thanks. that's what i was looking for.
TimToady and file operations tend to map nicely into lazy lists, if you want them to 22:26
however, p6 will be a little pickier about expicitly closing files for the same reason
*explicitly
otherwise it gets really easy to get race conditions in a lazy language 22:27
and you can't rely on timely destruction to close your files when you might be running on a random GC on a random VM 22:28
tradeoffs...
jdv79 makes sense. 22:29
TimToady basically, anything transactional should be orthogonal to control flow, as much as possible 22:30
22:31 mikehh joined
TimToady hopefully we can give support for transactions that looks like control flow, however 22:31
see STM and such
mikehh rakudo (3b63817) builds on parrot r41220 - make test PASS / make spectest (up to 28220) FAIL - Ubuntu 9,04 i386 (g++) 22:32
rakudo - 1 direct fail - t/spec/S12-introspection/walk.rakudo - Non-zero exit status: 1 - Parse errors: Bad plan. You planned 10 tests but ran 8.
rakudo - ./perl6 t/spec/S12-introspection/walk.rakudo - fails with (after ok 8):
rakudo - set_pmc() not implemented in class 'Failure'
rakudo - in regex _block424 (t/spec/S12-introspection/walk.rakudo:92)
rakudo - called from method Perl6Object::WALK (src/gen_setting.pm:1871)
rakudo - called from Main (t/spec/S12-introspection/walk.rakudo:94)
rakudo - t/spec/S03-operators/arith.rakudo - TODO passed: 131
rakudo - I got 112 tests PASSing but exiting with Non-zero wait status: 11
jdv79 i thought STM was dropped for now cause it was too much
TimToady talking evenutalities here
*eventualities 22:33
mikehh on amd64 I got - t/spec/S03-operators/arith.rakudo - TODO passed: 120, 131-132
TimToady we want to get to a model where transactions can be written on top of an underlying event model
mikehh so 120 and 132 don't pass on i386 22:34
TimToady much like a TCP connection doesn't have to worry about the underlying back-and-forth
but it's a long-term vision, I suspect 22:35
we will have to build it up post-6.0.0, I think
22:36 sri_kraih left, sri_kraih joined
TimToady basically, we'd like to take a monadic view of transactions as a computation of a series of events, but hopefully hiding the mental anguish of your typical monadic languages 22:39
sharada TimToady: by using delimited continuations? 22:40
22:42 s1n left
TimToady perhaps, this is all subject to negotiation as we go forward, and our understanding of the problem matures from the viewpoint of the naive programmer 22:42
22:42 meppl joined, sevvie left 22:43 nihiliad left 22:44 lambdabot left 22:46 perletc joined 22:52 M_o_C joined
jdv79 wait, so how will scope guards be done then? 22:54
22:58 SmokeMachine left
TimToady depends on what you mean by "scope guards", I suppose 23:12
23:14 lambdabot joined 23:17 Caliginous_ joined
Caliginous_ Hi! 23:17
is there perl mod to do same shit as flashgot or flashget?
jdv79 do x when this block is "done" 23:18
Caliginous_ done?
jdv79 upon exiting a scope run x
Caliginous_ where do I fine done?
jdv79 there is no done 23:19
23:19 donri joined
donri can you do something similar to inheritance and/or mixins with grammars? 23:20
TimToady yes, grammars are just classes that happen to be based on cursors 23:21
Caliginous_: in Perl 6 there will be a LEAVE block. were you asking about Perl 5 or Perl 6? 23:22
Caliginous_ no!
is there perl mod to do same shit as flashgot or flashget?
or can I run flashgot from Mozilla::Mechanize? 23:23
TimToady are you asking about Perl 5 or Perl 6?
diakopter Caliginous_: I wrote a Perl 5 script to do multithreaded downloading (using WWW::Curl::Multi) but it works on win32 only afaik.
lambdabot diakopter: You have 2 new messages. '/msg lambdabot @messages' to read them.
donri TimToady: cursors?
TimToady donri: abstraction of a match state in a particular language 23:24
Caliginous_ I ried that
it said no messages
23:24 perletc left
donri roger 23:24
TimToady but all token/rule/regex are really just methods, given that base 23:25
23:25 jaldhar joined
TimToady they're just always called on a cursor, and always return a cursor 23:26
but there are certainly ways of attaching other information to such a cursor, which is how we get match trees and ASTs
donri TimToady: can i treat a grammar both as a class and a role or just either? 23:28
23:32 cdarroch left 23:42 japhb left
donri also means you can have methods and attributes in a grammar huh? 23:45
23:48 s1n 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!