svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com
Set by avar on 16 November 2006.
00:01 Psyche^ is now known as Patterner
Debolaz Speaking of parrot, is there anywhere I can find the discussion/rationale behind a register based approach? 00:04
luqui Debolaz, I don't remember where the doc is; but the surface reason is that we have tons of register based research to draw on; the deeper reason is that stack-based VMs do not like ILP very much 00:05
and there is probably an even deeper reason that I don't know about 00:06
?eval class List is also { method shuffle() { $.pick($.elems)) } }
evalbot_r14926 Error: ␤Unexpected ")"␤expecting ":", term postfix, operator or "}"
luqui ?eval class List is also { method shuffle() { $.pick($.elems) } } [1..5].shuffle
evalbot_r14926 [3, 4, 1, 2, 5]
luqui ?eval class List is also { method shuffle() { $.pick($.elems) } } <w 0 0 t>.shuffle 00:07
evalbot_r14926 ["0", "t", "w", "0"]
wolverian whoa. luqui++
jrockway Debolaz: the reasoning was that optimizations for register-machines are well-known
since every *real* machine is register-based, and there are pretty good optimizing compilers for them
almost every, i mean
(further reasoning is documented somewhere on the parrot website, btw) 00:08
Ovid There's a List class separate from Array? Which inherits from the other? 00:10
(I'm feeling very slow right now)
luqui I left perl about the time the debate was going on
my angle was: List is the value type for Array
wolverian luqui, should that be @.pick?
luqui wolverian, possibly, yes
jrockway ?macro circumfix:</* */> is parsed /.*?/ { '' } /* foo bar */ 1;
lambdabot Unknown command, try @list
jrockway ?eval macro circumfix:</* */> is parsed /.*?/ { '' } /* foo bar */ 1;
evalbot_r14926 Error: ␤Unexpected "/.*?/"␤expecting trait or block
luqui luqui, an Array is a reference to a variable containing a List
wolverian reference? what's that? ;) 00:11
luqui but on these issues I seldom get my way :-), so it's probably something else
Debolaz jrockway: But those real machines are register based because that's much more practical to implement in hardware aren't they? In many software cases, especially when allowing unlimited registers, it seems somewhat inefficient (Sort of like an obfuscated stack) to use registers. That being said, I'm not expert on the sbuject.
jrockway Debolaz: right, but that's where the research efforts have been made
there'a *a lot* of literature on register-based machines
even if they're "not as good"
luqui Debolaz, also, stack machines don't like ILP; they cause mispredicts too much
er, cache misses
py1hon doesn't code for stack machines tend to be bigger too as it's littered with pushes and pops? 00:12
luqui yeah, but whatever.
py1hon and i'd think stack based machines would be harder to do jit compiling on, not that i know
Debolaz I just feel that perhaps chosing a register based machine because more optimizations are known for it is sort of like chosing fortran over perl based on it being many possible optimizations in the context of fortran.. ie, it doesn't really counterargue using a stack machine in itself.
jnthn If you are interpreting the code and have no JIT available and you have a stack, you gotta maintain the stack pointer. = slow
luqui jnthn, yeah, that's the ilp thing; all your instructions are dependent on the stack pointer. 00:13
jnthn And yeah, JIT on a stack is stack => register mapping, whereas JIT with register is essentially just register minimisation.
Which if you have lots of registers and a Parrot sub doesn't use many, may be a no-op.
jrockway although my machine doesn't have a register that can store pmcs :) 00:14
luqui jrockway, really? they're just pointers...
jnthn A Parrot register doesn't really, just a pointer to one. :-)
Your hardware stores pointers just fine. 00:15
Ovid For shuffle, can't you just extend Array with the sig (*@list) and use it with both lists and arrays? (Or am I being totally naive again?)
py1hon arrays can't do o(1) insertions and deletions in the middle... 00:16
Ovid Wait, I meant (*@list is copy). That's assuming that slurpy arrays don't default to 'is copy' 00:17
jrockway ah :) 00:18
luqui Ovid, I'm not sure
I implemented it as a method
so...
Ovid Well, Larry suggested it might be that way, but who knows?
luqui ?eval class List is also { method shuffle() { @.pick($.elems) } } <w 0 0 t>.shuffle 00:19
evalbot_r14926 ["0", "t", "w", "0"]
luqui hmm.
?eval <w 0 0 t>
evalbot_r14926 ("w", "0", "0", "t")
luqui why is it returning a ref
Ovid ?eval <w 0 0 t>
luqui ?eval <f o o b a r>.pick(6)
evalbot_r14926 ("w", "0", "0", "t")
("b", "r", "o", "a", "o", "f")
jrockway ref?
luqui the square brackets
Ovid ?eval <w 0 0 t>.pick(3) 00:20
evalbot_r14926 ("0", "0", "t")
Ovid ?eval <w 0 0 t>.pick(5)
evalbot_r14926 ("0", "w", "0", "t")
luqui ?eval class List is also { method shuffle() { @.pick($.elems) } } @{<w 0 0 t>.shuffle}
evalbot_r14926 Error: ␤Unexpected "{<"␤expecting "@" or "::"
luqui ?eval class List is also { method shuffle() { @.pick($.elems) } } list <w 0 0 t>.shuffle
evalbot_r14926 ["0", "t", "0", "w"]
luqui ?eval class List is also { method shuffle() { @.pick($.elems) } } *<w 0 0 t>.shuffle 00:21
evalbot_r14926 Error: Odd number of elements found where hash expected: VNum Infinity
luqui ?eval class List is also { method shuffle() { @.pick($.elems) } } *(<w 0 0 t>.shuffle)
evalbot_r14926 Error: Cannot cast from VNum Infinity to Pugs.AST.Internals.VCode (VCode)
luqui oh
* doesn't do that anymore
?eval class List is also { method shuffle() { @.pick($.elems) } } =<w 0 0 t>.shuffle 00:22
Ovid Why should "<w o o t>.pick(5)" only return 4 elements? That's silently discarding the information that I'm trying to pick too many elements.
evalbot_r14926 Error: Can't modify constant item: VUndef
luqui Ovid, you want it to die?
Ovid I always want things to die when they're a serious error doing something they shouldn't do. 00:23
Debolaz Also, what is the adventage with a junction over a list that would be difficult to achieve using lists?
audreyt luqui: the obvious solution is the correct one: @()
luqui ?eval class List is also { method shuffle() { @.pick($.elems) } } @(<w 0 0 t>.shuffle)
evalbot_r14926 ["0", "w", "t", "0"]
luqui hmm..
apparently not
audreyt why not?
Ovid Of course, that might just be the "strong typing" pseudo-bigot coming out! 00:24
audreyt arglist flattening is |<< btw
?eval class List is also { method shuffle() { @.pick($.elems) } } |<< <w 0 0 t>.shuffle
evalbot_r14926 (["0", "t", "0", "w"],)
audreyt but that doesn't quite work here
have to be inside a function call
luqui what is |<< ? 00:25
audreyt ?eval class List is also { method shuffle() { @.pick($.elems) } } <w 0 0 t>.shuffle
evalbot_r14926 ["w", "t", "0", "0"]
audreyt ?eval class List is also { method shuffle() { @.pick($.elems) } } | <w 0 0 t>.shuffle
evalbot_r14926 ["0", "t", "w", "0"]
audreyt ?eval class List is also { method shuffle() { @.pick($.elems) } } sort | <w 0 0 t>.shuffle
evalbot_r14926 ("0", "0", "t", "w")
Ovid In other words, if there's a serious debate about what a non-sensical method call should return, I want it to die a quick and painful death.
audreyt luqui: prefix | is the new *
Ovid Of course, that might be the rum talking.
audreyt S03/Argument Llist Interpolation
luqui Ovid, I'm never sure about TimToady, but he might disagree with you
Ovid, you're not too good with haskell, are you? 00:26
(ha, speaking like I am! :-p)
audreyt you are quite good :)
audreyt goes making Really Important phone call. bbiab &
Ovid I went through a Haskell tutorial a long time ago and I remember precisely squat (aside from the fact that i can recognize guards when I see them) 00:27
luqui Ovid, it might be a good exercise to make it die when you try to pick too many elements
audreyt Ovid: web.archive.org/web/20060327034654/...review.pdf
lambdabot tinyurl.com/y94rfc
audreyt ponders the ethics of linking into now-404'ed preview pdf of a book
luqui Ovid, src/Pugs/Prim/List.hs line 70 if you're interested 00:28
Ovid Audrey: thanks!
audreyt but that's the best material I've found so far
and I will buy the book once it comes out, so I guess that's ok :)
Ovid Audrey: no worries. I won't tell anyone and this will just be between us :) 00:30
luqui and me, now
Ovid Shh!
I just noticed that you linked to something that is from the University of the city I just moved to (and am now moving away from). Cool! 00:31
TimToady is never quite sure about TimToady either... 00:32
luqui knew that, he just didn't say
Ovid luqui: there's absolutely no point in making it try to pick from more elements than the list has, so anyone trying to do that is arguably trying to do something wrong. 00:33
audreyt this certainty thing is certainly overrated...
00:33 Ovid left
luqui oh, the old hit-and-run argument approach :-) 00:33
jrockway heh, i printed a paper copy of that book a few weeks ago ( i think )
00:34 Ovid joined
Ovid Oops. 00:34
Just lost my chat log, too :(
audreyt there's always irc.pugscode.org
Debolaz checks out the book. 00:35
TimToady ?eval <a b c>.pick(*)
evalbot_r14926 Error: pick not defined: VList [VStr "a",VStr "b",VStr "c"]
TimToady looks like the implementation is not yet complete...
luqui ah 00:36
I wonder what the internals of * look like
Ovid Thanks, Audrey.
TimToady matched by Whatever in the sig
Debolaz The biggest obstacle to learning haskell for me so far has been the tutorials going out of their way to just saying things the way they are but instead try to make up every imaginable analogy for it.
TimToady other than that, doesn't have much internals...
luqui looks like it is represented by VNum Infinity 00:37
jrockway i just read through one of audreyt's talks on haskell
TimToady that's a hack
jrockway very accessible and entertaining
TimToady arbitrarily large is very different from infinity
Debolaz jrockway: Url? 00:38
luqui goes to perpetuate the hack in his implementation of pick
TimToady and * is really only "arbitrarily dwim" with a slight bias toward largeness.
or smallness, in the case of *..0 00:39
Ovid jrockway: which intro? I read one and was terribly confused, so it must be a different one (this being from a guy who went from OO to Prolog but doesn't quite grok the lambda calculus)
jrockway let me look around for it
feather.perl6.nl/~audreyt/osdc/haskell.xul
Ovid Tjat 00:40
That's the the one I read. I guess I'd have to hear the accompanying talk.
jrockway just type them into ghci and follow along 00:41
and then read the book
then you'll be the next audreyt, perhaps :)
Ovid I'll give that a shot. The fact is, I confess I'm just not as quick as the rest of you. I'm just persistent :) 00:42
jrockway hehe
some of the examples need let in front of them, i think
TimToady I am often persistent briefly. 00:43
jrockway let double = (*) 2 instead of "double = (*) 2" as per the slides
Ovid TimToady: don't say that. I'll stop believing your honesty!
So anyone willing to make a prediction about a timeline for Perl 6 alpha built in PGE? 00:47
jrockway people are willing to predict, but i bet you're tired of the answer ;)
Ovid (Warning: a O'Reilly blog post might follow that if there aren't enough caveats) 00:48
luqui is certain that people don't want to hear my answer...
Ovid Well, I've predicated "2007", but likely in the first half.
jrockway christmas 2007? 00:49
svnbot6 r14927 | luqui++ | Implemented @list.pick(*).
jrockway ?eval (1,2,3).pick(*)
evalbot_r14926 Error: pick not defined: VList [VInt 1,VInt 2,VInt 3]
luqui jrockway, be patient! :-)
Ovid :P
jrockway *whir of evalbot self-compiling itself*
luqui ?eval [1..31].[5..*] 00:51
evalbot_r14926 ()
luqui hmm
?eval (1..31).[5..*] 00:52
evalbot_r14926 (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31)
luqui that seems, wrong
the former, that is
jrockway ?eval [1..31]
evalbot_r14926 [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31]
jrockway ?eval [1..31].WHAT
evalbot_r14926 ::Array
Ovid Agreed. References should resolve themselves.
TimToady there are no references. :) 00:53
Ovid We'll call them "spoons".
jrockway what's the difference between [...] and (...) ?
sporks
TimToady Arrays are mutable
luqui I'm afraid of touching that bug though.
TimToady but that shouldn't matter for this
luqui I've been there before
TimToady ?eval [1..31][-1] 00:54
evalbot_r14926 \31
luqui was it the dot?
?eval [1..31][5..*]
evalbot_r14926 ()
TimToady ?eval [1..31].[-1]
py1hon array references wrap around?
evalbot_r14926 \31
TimToady subscripts wrap around
jrockway py1hon: as per perl5, right?
TimToady yes
luqui py1hon, but not as in modular arithmetic
just negative numbers 00:55
?eval [1..31][42]
evalbot_r14927 \undef
jrockway ok :)
py1hon jrockway, i'm a c monkey =/
jrockway ?eval (1,2,3).pick(*)
evalbot_r14927 (1, 2, 3)
TimToady however, once we get range objects, they may behave more toward
jrockway py1hon: learn perl :)
Ovid ?eval item [1 .. 31][-1]
py1hon i did a long time ago, then forgot it all ;p
evalbot_r14927 31 00:56
jrockway i only program in C for the mental exercise
using tools like valgrind has a certain satisfaction
py1hon i do tons of sockets programming for work...
valgrind is so awesome
jrockway well, even better is a language that doesn't let you leak memory :)
not as fun though
luqui I program in C++ for games (and I'm ashamed), haskell for mathematical inquiries, and perl for everything else
Ovid ?eval (1..23).pick(*)
evalbot_r14927 (11, 20, 7, 9, 1, 4, 22, 16, 18, 17, 8, 23, 14, 15, 6, 10, 12, 3, 2, 5, 19, 13, 21)
sili memory leaks are good thing. it keeps certain people employed
jrockway sadly kde's C++ base seems to lead to unstable apps 00:57
Ovid Yeah, but they can never forget the pain!
jrockway whereas gnome's python base (now) is quite solid
sili what doesn't kill you...
luqui is amused at the fact that .pick(*) is the same number of characters as .shuffle\
sili makes you study liberal arts
py1hon lol
or pure math?
Ovid ... leaves you in a half-dead, unreaped state.
luqui py1hon, hey!
py1hon mmm graph theory :) 00:58
luqui I was almost killed by my math final today
that was the hardest test I have ever taken (but I did well ;-)
py1hon i didn't finish my take home math final because i was finishing my project that i put off till the last minute =/
Ovid Hey, I don't want to hear about Python. lambda is fundamentally borked and is going to stay that way :(
py1hon graduate course, too
luqui Ovid, how come 00:59
TimToady ?eval ('a'..'c').pick:*
luqui my python-fu is weak
evalbot_r14927 Error: ␤Unexpected ":*"␤expecting term postfix or operator
TimToady ?eval ('a'..'c').pick: *
evalbot_r14927 ("c", "b", "a")
TimToady heh. hoist by my own petard
luqui :-)
sili python does function-level scoping like javascript/ruby, right? that's really ick.
Ovid basically, lambda only allows a single statement (unless it's changed recently) 01:00
jrockway python never excited me really
luqui dislikes function-level scoping too
jrockway no better than perl, but really really ugly syntax
luqui jrockway, you say python's syntax is ugly?
sili luqui: it's just... dumb. there's no argument for it aside from the fact you can omit my()-like constructs
Ovid www.artima.com/weblogs/viewpost.jsp...ead=147358
lambdabot Title: Language Design Is Not Just Solving Puzzles 01:01
jrockway __init__.[__whoa__].[__underscores_arent_that_cool__] '''''''comment'''''''
:)
luqui jrockway, wow, you're the first person to call python ugly when comparing to perl
that I've heard
jrockway perl's not beautiful, but it's concise
easy to type, easy to edit, easy to read 01:02
not so easy to parse
luqui I think python looks cleaner, but then again I think italian sounds poetic.
it doesn't really matter, because I don't know either language
Ovid Though now that I google for it, it appears that some folks seriously disagree with me: ivory.idyll.org/blog/nov-06/i-just-dont-get-it
lambdabot Title: Daily Life in an Ivory Basement : /nov-06/i-just-dont-get-it
jrockway if you want really ugly, look at PHP
no postfix control statements, everything requires parens, etc. 01:03
TimToady someone really needs to write a refactoring PHP-to-Perl6 translator...
jrockway that would be potenitally easy
php is just a simple bison grammar 01:04
T_FUNCTIONNAME etc.
it's slightly insane
i found a bug the other day, but i couldn't bring myself to send them a patch :/
luqui I think the "refactoring" makes it a little tougher :-)
jrockway yeah
how to make dumb php smart
TimToady "here's the namespaces you should have used, dolt!"
jrockway who needs namespaces when you have 8 different ways to split a string? 01:05
sili pretty soon someone's going to write a language that supports anything you could ever want in a language, then it's just a matter of giving it a syntax
cough cough.
jrockway php is a great language to learn for one thing
luqui Ovid, hmm, the double colon seems slightly analogous to :: in haskell, but not very much so
jrockway it's every combination of poor design, all combined into one language
sili php isn't great for anything. take that back right now!
jrockway fine, i take that back 01:06
it's interesting to see why php is popular
mostly, people feel smart when they do things the compiler should do for them
like differentiate between when to use split and preg_split 01:07
Ovid Because it's so easy to get up and running with PHP. That's not something which should be ignored.
jrockway a smart language would say, "hey, this regex is just a string" and DWIM
TimToady er, unlike P5, which botches that...
jrockway p5 is included with every sane OS in the default install 01:08
what could be easier?
catalyst is hard to install (i'm told), but so are the PHP frameworks
and they're not worth the effort IMHO
TimToady I think "sane OS" indicates the null set.
jrockway non-insane then?
luqui wow.. it's amazing how much I disagree with guido. 01:09
thanks for the article Ovq
er, Ovid
sili but he's our benevloent dictator for life!
how dare you!
Ovid :)
TimToady I hope I'm just benevolent dictator till I can Get It Right... 01:13
luqui you mean for life?
01:13 lyokato joined
luqui hmm.. actually longer than that 01:13
TimToady Life pluse 182 years...
*plus
jrockway btw TimToady, have you read this? steve.yegge.googlepages.com/ancient...uages-perl
lambdabot Title: Stevey&#39;s Home Page - Ancient Languages: Perl􏿽xC2
jrockway pretty hilarious 01:14
TimToady sure
it's a nice rant
polettix I don't fully get this in van Rossum's article: "This is also the reason why Python will never have continuations, and even why I'm uninterested in optimizing tail recursion.". Why this opposition to tail recursion *optimisation*? It should be benevolent with the caring programmer and not-malevolent with the uncaring one 01:18
jrockway i learned the other day that elisp of all languages doesn't support tail recursion 01:19
pretty insane IMHO 01:20
luqui polettix, tail recursion optimization has some trade-offs, and it is good for programming styles that make good use of it. python's "one true style" does not.
jrockway what are some of the tradeoffs?
polettix luqui: does this mean that enhancing tail recursion would bash those not using it? 01:21
Ovid Because tail recursion is something that many programmers don't quite 'get'. Guido's point seems to be that he wants a language that most programmers can understand. However, most programmers 20 years ago wouldn't have understood Python, so I don't follow his logic.
jrockway PHP is a language that everyone can understand
polettix ah, so it's for a readability issue
Ovid Of course, I could easily be misunderstading Guid
jrockway i personally would prefer a langauge that *I* can understand :)
Ovid ... misunderstading Guido, so that's not an objection. 01:22
luqui polettix, that is my impression. I'm not heavily enough in implementation-land to defend it though.
Ovid Ah! I misspelled it again!
"misunderestimating!
luqui "misunderstandimating"
polettix it seems some time that I don't update perl6 - when did it jump on new GHC 6.6?
jrockway before the latest release release, i think 01:23
Ovid I note that TimToady i largely silent on this topic, but given the politics of the situation, I guess that's understandable :) 01:24
TimToady Guido is not willing to give up indentation, and I'm not willing to give up list flattening. :) 01:25
allbery_b discovered perl and python at the same time. he stuck with perl. :)
(after playing with both for a while)
jrockway i looked at python when i wanted a "real" object system, but python didn't satisfy 01:26
i would have been happy with perl6 = perl5 + java classes (+MI)
TimToady people keep talking about Python like it's the true successor to Lisp, but I don't buy it.
py1hon if it was, it'd have macros :)
Tene I think lisp variant allowing indentation to be used in place of ()s would be interesting. 01:27
py1hon and oo based on multimethods
Tene entertaining, at least.
jrockway that would be hard
py1hon Tene, in practice that's what happens, when you write lisp you let your editor indent and mostly ignore the parentheses
jrockway how would you handle cond, for example 01:28
01:28 lyokato joined
Ovid Tene: I thought about that based on billhails.net/Book/, but I need tuits! 01:29
lambdabot Title: Exploring Programming Language Architecture in Perl (cover art by the author)
jrockway something like: (let ((a 10)) (cond ((< a 10) t)((> a 10) t)((= a 10) nil))) 01:30
would it be let \n \t a 10 \n \t \t cond \n \t \t \t ... ?
Ovid I would have called that language "Cor" (massive bonus points to anyone who knows why) 01:31
jrockway google from "cor python lisp" returns some possibilities 01:32
Ovid Nope.
jrockway oh. :( 01:33
Ovid The original Ovid wrote love poems to "Corinna". No one knows if she existed or if she was a literary device.
TimToady well, hey, the two are not mutually exclusive. :) 01:34
Ovid There's some speculation that she might be the woman he was forced into marrying for his first marriage (an arranged marriage).
01:34 seano joined
luqui TimToady, I don't understand the paragraph beginning with "It is allowed to specify" in your recent change 01:35
Ovid So it might be the case that Ovid wrote many of his love poems to a woman he was forced to marry, but had the bad grace to actually fall in love with.
01:35 Aankhen`` joined
TimToady luqui: the following example doesn't clarify it? 01:37
luqui not really
so you're talking about a sub taking a capture, and then continuing to bind?
but doesn't the capture eat up everything?
TimToady no, it's a lookahead. :)
luqui so there is backtracking? 01:38
TimToady no, just snapshotting
luqui sub foo (|$cap, $n) { say $n }; foo(3,4,5,6); # ?
is it 6?
01:38 mako132_ joined
TimToady no it'd be 3 01:39
except there are too many args
luqui sub foo (|$cap) {...} foo(3,4,5,6) # legal?
01:39 Aankhen`` joined
TimToady certainly 01:39
$cap is \(3,4,5,6) 01:40
luqui so the comma isn't so much a comma as an en-passant capturer?
01:40 Aankhen`` joined
TimToady which comma? 01:40
luqui after |$cap
(|$cap) and (|$cap, $n) are very different
TimToady I originally didn't have one. Someone else put it in.
luqui looks it over again 01:41
TimToady But it really doesn't matter, if |$cap is "zero width".
luqui but in my first example, foo(3,4,5,6) is illegal?
TimToady no compatible routine found... 01:42
luqui one last test:
sub foo (|$cap, $a, $b, $c) {...} foo(1,2,3)
$a is 1, $b is 2, $c is 3, $cap is \(1,2,3) ?
TimToady that's the intent.
luqui okay 01:43
yeah, the comma is what is bugging me
but I kind of grok the reasoning
it's more like a processing queue than a specification list
TimToady I originally specced it (|$cap $a, $b, $c)
luqui or something
why the change?
TimToady someone thought it was a typo
luqui oh, you were serious about someone else changing it 01:44
TimToady and I thought maybe it didn't matter.
but it could influence mmd
luqui wants to hash out the consequences of a processing-queue model 01:45
TimToady either that, or you have to write the most generic form as (|$cap, *) or some such
luqui wants also to go home now
back in an hourish probably
TimToady and pragmatically, I figured if people thought it was a typo once, they'd do it over and over.
luqui uh huh 01:46
neither is really obvious
one for syntax, one for semantics
luqui &
Ovid OK, It's almost 2 in the morning here. I have to run. Bye folks. (And read the Peter Green's translations of Ovid's works. He's brilliant) 01:47
TimToady decommuting &
01:49 Aankhen`` joined 01:51 Aankhen`` joined 02:00 mr_ank joined 02:01 Azure-BOT joined
Debolaz What's the proper way to call a subroutine referenced in $foo? 02:17
$foo($bar)?
Debolaz recalls eval
eval my $foo = -> $x { say $x }; $foo(42); 02:18
?eval my $foo = -> $x { say $x }; $foo(42);
evalbot_r14927 OUTPUT[42␤] Bool::True
Debolaz :o
02:22 amf joined 02:28 Psyche^ joined, Psyche^ is now known as Patterner 02:37 awwaiid joined 02:42 nipra joined 02:47 luqui joined
TreyHarris ?eval my $foo = -> $x { say $x }; $foo.(42) 02:56
evalbot_r14927 OUTPUT[42␤] Bool::True
TreyHarris that syntax works too.
03:54 bsb joined, ssig33 joined
ingy hola 04:16
luqui hola ingy 04:18
04:33 dduncan joined 05:14 mjk joined 05:23 Eidolos_ is now known as Eidolos 05:33 nipra joined 06:15 BooK_ joined 06:28 jdv79 is now known as capo 06:36 seano joined 06:43 capo is now known as jdv79 06:55 xinming joined 07:09 devogon joined 07:28 justatheory joined 07:39 ofer1 joined 08:00 iblechbot joined 08:19 rahikkala joined
svnbot6 r14928 | lwall++ | change call to callsame/callwith/nextsame/nextwith 08:33
09:45 falseep joined 10:02 luqui joined 10:11 elmex joined 10:23 dduncan left 10:54 ruoso joined 10:56 buetow joined
Nei am I still here? test test 11:08
jrockway ?eval Nei.exists
evalbot_r14928 Error: No compatible subroutine found: "&Nei"
Nei dang! 11:09
thought so
jrockway declare yourself, human :)
luqui ?eval 'Nei'.exists 11:10
evalbot_r14928 Error: No compatible subroutine found: "&exists"
luqui aww
Nei ^^
luqui ?eval 1 < 5< 7 11:11
evalbot_r14928 Bool::True
luqui ?eval sub postfix:<!> ($x) { [*] 1..$x } 9!
evalbot_r14928 362880 11:12
luqui ?eval class List is also { method stupid() { ([+] @.elems) % 2 == 0 } } [1,4,5,7,3] 11:13
evalbot_r14928 [1, 4, 5, 7, 3]
luqui ?eval class List is also { method stupid() { ([+] @.elems) % 2 == 0 } } [1,4,5,7,3].stupid
evalbot_r14928 Bool::False
luqui ?eval class List is also { method stupid() { ([+] @.elems) % 2 == 0 } } [1,4,5,7,2].stupid
evalbot_r14928 Bool::False
Nei perl6 will make it easy to provide a mirc script like irc scripting glue 11:15
avar why does make test always compile something even though I've just done "make" and haven't svn up-ed 11:55
12:00 chris2 joined 12:02 elmex joined 12:10 avar joined 12:13 Psyche^ joined 12:21 Psyche^ is now known as Patterner 12:40 mako132_ joined 12:50 xinming_ joined 12:59 xinming joined 13:22 buetow joined 13:26 weinig|bbl is now known as weinig_, weinig_ is now known as weinig 13:48 kanru joined 13:52 weinig is now known as weinig|bbiab 13:53 gnuvince joined 13:54 gnuvince joined 14:11 weinig|bbiab is now known as weinig 14:36 iblechbot joined 14:42 weinig is now known as weinig|bbl 14:46 rahikkala left, chris2 joined 15:03 c6rbon joined 15:04 bonesss joined 15:05 mdiep_ joined 15:07 azazello joined 15:13 awwaiid joined 15:28 mdiep joined 15:33 buetow joined 15:52 hexmode joined 16:00 marmic joined 16:32 penk joined 16:33 ruoso joined 16:36 penk joined 17:08 lisppaste3 joined 17:10 kanru joined
svnbot6 r14929 | wolverian++ | r32@pupu: wolverian | 2006-12-19 19:24:00 +0200 17:24
r14929 | wolverian++ | update the hq9+ example
wolverian augh, can I make --verbatim the default somehow? 17:25
17:42 weinig|bbl is now known as weinig 18:03 coumbes joined 18:07 TSa joined 18:15 buetow joined 18:31 ozo joined
nothingmuch Coleoid: i know the problem 18:34
just back from a 4 day hike
will take care of it soon 18:35
18:46 drrho joined 18:50 justatheory joined 18:51 weinig is now known as weinig|bbl 18:52 justatheory joined 18:55 justatheory joined 19:05 prism joined 19:13 prism joined 19:18 autark joined 19:19 prism joined 19:21 larsen_ joined 20:17 cookys joined 20:39 Aankhen`` joined 21:04 chris2 joined 21:07 weinig|bbl is now known as weinig 21:09 ozo joined 21:12 Ovid joined
Ovid ?eval my @result; <a b c c d> ==> map { $_ xx 2 } ==> @result; @result.perl.say; 21:13
evalbot_r14929 Error: ␤Unexpected " map"
Ovid I gather that feed operators are not yet implemented? 21:14
When they are, will the following be legal syntax? <a b c c d> ==> map { $_ xx 2 } ==> my @result; 21:15
Or wait, looks like I would do: my @result = < a b c c d > ==> map { $_ xx 2 }; 21:16
jrockway Ovid: i think the synopsis says ==> @result 21:19
@result <== map { .. } <== @input
Ovid Looking at it now:
@oddsquares =
(@nums ==> grep { $_ % 2 } ==> sort ==> map { $_**2 });
jrockway so feed operators return the end of the pipeline
makes sense
Ovid Needs parens, but it's still not implemented. Looks like the t/operators/feed.t is all :todo tests.
jrockway am i allowed to refer to list and scalar context in perl6, or is that gone? 21:20
sili still there
jrockway so what do the feeds return in scalar context (without parens) ? 21:21
?eval (1,2,3) ==> my @output
evalbot_r14929 Error: ␤Unexpected " my"
jrockway ?eval @in1 ==> map {} <== @in2 # :) 21:22
evalbot_r14929 Error: ␤Unexpected " map"
jrockway that should be fun :) 21:23
21:24 prism joined
Ovid Hmm, I wrote a program to split the 99 problems into separate Perl 6 test files. I wonder if those would be OK to commit? 21:25
I imagine they might slow down the test suite even more. 21:27
jrockway are they really tests? 21:31
maybe put them in examples?
Ovid Ah, that makes sense. I am writing them as .t files, but examples/ would be much better.
jrockway however, having them as tests seems like a beginning of a good idea to me 21:34
really you should give each example a test suite 21:35
to make sure the example works, and the perl6 implementation can run it and get correct results
i think that would be pretty awesome
99 working programs with good test coverage
Ovid Each problem has been split into it's on .t file (and it turns out that there are only 84 problems, go figure) and I'm slowly working through them. I just don't know if t/ is really a good place to put them since they *are* examples. 21:36
Still, they are also tests. 21:37
jrockway what is the output of your .t files?
allbery_b new section?
jrockway i only briefly looked at the problems, but they're things like "get the last element of a list"
allbery_b perhaps rigging the test harness so you can toggle its inclusion?
Ovid They are standard Perl 6 tests using the 'Test' module. 21:38
jrockway (easy in perl, harder in lisp :)
Ovid Oh, some of the problems get *much* harder. There are other beginner ones like 'packing elements in a list into sublists and stuff'.
Of course, I don't think I'll be tackling the sudoku solver right away :) (problem 97) 21:39
And I'm not even remotely able to write an Ada identifier syntax checker! 21:40
Oh, wait. It actually turns out to be easy. Never mind :) 21:41
Junctions are broken in Pugs? 21:44
Wow. They're horribly broken!
?eval use Test; plan 1; is 'b', any('a' .. 'h'), 'junctions should work'; 21:46
evalbot_r14929 Error: *** Unsafe function 'use' called under safe mode␤ at -e line 16, column 7-138
Ovid :(
Debolaz :~( 21:47
Ovid ?eval sub a (Str $g, Str $e) { my $t := $g eq $e; $t ~~ s:P5:g/#/\\#/;say('whee!',$t) } a('b', any('a'..'h')); 21:58
evalbot_r14929 OUTPUT[whee!1␤whee!␤whee!␤whee!␤whee!␤whee!␤whee!␤whee!␤] (Bool::True)
Ovid Yeah, that looks like a bug to me. 21:59
Ah, it's calling a() once every every item in the any() junction. 22:00
22:08 amf left
Ovid OK, gotta run. Night all. 22:13
22:28 silug joined 22:29 gnuvince joined
jrockway ?eval 'b' == any('a', 'b', 'c') 22:36
evalbot_r14929 (Bool::True)
jrockway ?eval 'b' == any('a'..'z')
evalbot_r14929 (Bool::True)
jrockway that's horribly broken?
?eval sub a(Str b){ b == 'a' } a(any('a'..'c')) 22:37
evalbot_r14929 Error: ␤Unexpected "Str"␤expecting "\\", ":", "*", parameter name or ")"
22:37 gnuvince joined
jrockway ?eval sub a(Str b){ b == 'a' }; a(any('a'..'c')) 22:37
evalbot_r14929 Error: ␤Unexpected "Str"␤expecting "\\", ":", "*", parameter name or ")"
22:38 sili joined
jrockway ?eval sub a (Str b){ b == 'a' }; a(any('a'..'c')); 22:38
evalbot_r14929 Error: ␤Unexpected "Str"␤expecting "\\", ":", "*", parameter name or ")"
jrockway jrockway-- # :(
22:39 sili left, dduncan joined 22:58 dmq joined
dmq seen audreyt? 22:59
allbery_b ?seen audreyt 23:06
lambdabot audreyt is in #ghc, #haskell and #perl6. I last heard audreyt speak 22h 31m 53s ago.
23:09 Psyche^ joined, weinig joined 23:14 Aankhen`` joined 23:25 Psyche^ is now known as Patterner 23:27 Aankhen`` joined 23:32 Aankhen`` joined 23:35 Aankhen`` joined 23:45 sonorous joined 23:49 Aankhen`` joined 23:52 Aankhen`` joined 23:53 luqui joined 23:54 xpika 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!