'-Ofun: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6'
Set by rafl_ on 5 November 2005.
theorbtwo G'night, laz. 00:00
liz.
liz6 g'night theorbthree ;-)
leo and four me g.night 2 00:01
svnbot6 r7912 | stevan++ | Perl6::ObjectSpace - 00:16
r7912 | stevan++ | * all methods in Bootstrap now have a checked return type :)
01:17 Lopo_ is now known as Lopo
gaal dduncan: I won't have time to hack .hs myself in the coming week or so, so by all means go for it if you want. 05:16
rafl: does what work for me? 05:17
rafl gaal: pge support. You said it doesn't work on feather.
gaal SamB: er, of course. Did I imply that you can somewhere? :) The sourcy macos package "uses a binary bootstrap compiler" 05:18
rafl: sec.
rafl Does someone know under what license autrijus talks are? I want to give a Perl 6/Pugs talk on friday and would like to steal excessive from him.. :-) 05:19
pasteling "gaal" at 192.115.25.249 pasted "pge not working" (3 lines, 143B) at sial.org/pbot/14193
rafl Try to use /usr/bin/pugs? 05:20
gaal long as the flight from .ch to .hk may be, i'm pretty sure he'll be landing by then :) 05:21
pasteling "gaal" at 192.115.25.249 pasted "no go with system pugs either" (4 lines, 264B) at sial.org/pbot/14194 05:22
rafl Hm. The system pugs error confueses me.
Doesn't it work on feather only? 05:23
Read as: Is there someone who has this running?
gaal rafl: let me test on my colinux 05:24
but wait, are you doing parrot now too? wow
rafl Noone else wants to, as it seems. 05:25
There has been a request for a Debian package for months and no one felt responsible, so I did.
gaal i think there's at least one other way to put that, but rafl++ nonetheless :) 05:26
rafl What's the other way? :-)
gaal so i should just dpkg -i everything from ~rafl/public_html/debian/ ?
rafl gaal: Not libghc6-* -- it's broken from SVN currently. 05:27
gaal s/responsible/capable/, for one. or knowledgeful of debian. or temporally able. 05:28
ok
rafl Well. It's not that much work currently. One release a month isn't that much. 05:29
gaal isn't it kind of an asymptotic thing? huge inital overhead, gets easier the more frequent releases are? 05:30
rafl If you packaged it once you'll only need to bump up versions in some places and everything works again. Nothing that you need really deep Debian knowledge for.
The initial overhead wasn't that huge. I thought that first so I hestiated to do a parrot package for some weeks. I also started to do it in a branch so I could not break anything because I thought there were some major changes to the build system needed. 05:31
But in fact it was done in only some hours and I merged the branch directly back some hours after creating it. :-)
gaal i meant the whole debian gig. 05:32
rafl I don't get you then.
gaal part of the overhead is learning debian packaging, getting to know/be known by the other debian people. 05:33
if i had to package something once a year in debian, that once a year would be very hard
if i did it every week, it'd be crazy at first but then sorta ease out. no? 05:34
rafl I guess that's right.
gaal fwiw: on my test machine, PGE fails in the same way it did with eric's autobuild 05:35
not in the way /usr/bin/pugs failed on feather.
rafl It usually takes me less around 5 minutes to make a Debian package of an average cpan module now. When I started it took hours and others still had to complain about it..
gaal: I've no idea then.
gaal in any case i doubt the problem is in the package itself.
rafl The problem with /usr/bin/pugs on feather maybe is. I'm not sure though. 05:36
buu ?eval { 1 } 10:19
evalbot_7912 1
buu ?eval { 1 }->()
evalbot_7912 Error: unexpected ">" or "-" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
buu ?eval sub { 1 }
evalbot_7912 sub {...}
buu ?eval x => {1} 10:21
evalbot_7912 ("x" => sub {...})
azuroth ?eval { 1 }.() 10:25
evalbot_7912 1
azuroth ?eval -> $a { $a }.(5)
evalbot_7912 \5
azuroth slash five, eh?
theorbtwo azuroth: evalbot, for reasons that I've never quite understood, often gives one level of enreferencing beyond what you'd expect to see. 10:28
azuroth ahh 10:29
I don't think I've noticed before
buu eval: 1 10:31
buubot buu: Return: 1
love_ANGEL selam k0131zlar benimle tan0131015fmak isteyen varm0131yd0131 11:02
?
alo kimse yokmu dur 11:03
be
sizi bekliyorum ama ben
?
azuroth saluttt les pdddd, love angel! 11:04
azuroth wonders whether "for @foo -> $x, $y, $x" is a mistake in the quickref
11:07 nnunley_ is now known as nnunley
azuroth I'm going to assume it is, and change the last x to z. take that! 11:10
theorbtwo Looks like a typo to me.
svnbot6 r7913 | azuroth++ | Changed quickref's control so that 11:17
r7913 | azuroth++ | for @foo -> $x, $y, $x
r7913 | azuroth++ | is now
r7913 | azuroth++ | for @foo -> $x, $y, $z
azuroth "the list ctoring comma operator supplies this list context: sub *infix:<*> (*@things) { ... }" - should that be *infix:<,> ? 12:25
theorbtwo Quite likely. prefix:<*> also makes some sense, but not much. 12:26
azuroth I also noticed it didn't mention * providing list context (but did mention ~, +, and ?) and thought that was a bit peculiar 12:27
rafl Well, $x, $y, $x could also be useful. Even if there could be a better way like @foo -> undef, $y, $x or @foo -> _, $y, $x 12:29
azuroth I think it actually assigned x to the first value it saw. but yeah, it _could_ be useful 12:30
it's a peculiar and confusing case for being in the quickref though, in my opinion
rafl Indeed. 12:31
theorbtwo I don't think we want obfu cases in our quickref.
rafl Don't we? Ooo :-( 12:32
theorbtwo Putting "_ = foo() # Explicitly ignore the return of foo." in might be useful.
rafl Is _ really perl 6? Where is it described? 12:34
theorbtwo Donno; just seen it alluded to on the chan.
rafl But I thing _ = foo() is ugly. foo() in void context does the same. 12:35
Well, I think the _ is stolen from haskell.
theorbtwo But it's a simple example.
(_, $x, $y) = foo() would be an actually useful one.
rafl I think undef is Perl 6's _ 12:36
Just like in Perl 5.
azuroth good night, everyone 12:43
liz6 night, azuroth! 12:44
gaal liz6: hi 12:45
liz6 gaal: hi!
gaal thanks for the notes. :) 12:46
what did you mean by best practices in regard to the function call examples?
theorbtwo Hi, liz. 12:47
And gaal.
gaal hey :)
liz6 foo( $a ) as opposed to foo ($a) ?
gaal is that in PBP? I don't like it. 12:48
liz6 I think it is...
someone pls correct me if I'm wrong... ;-)
theorbtwo I wouldn't put a space between the name and the beginning of the arg list, because you shouldn't try to call perl 6 subs that way. 12:49
xinming liz6: foo( $a ) = foo $a; but != foo ( $a );
foo ($a) is become a bit like, foo list( $a ); 12:50
liz6 that's why I put a question mark in my comments...
so, please treat it as a slippo of my fingers..
wolverian rafl, tentatively, ? (I think) and ! mean optional and required params that are ignored, in signature lists. 13:23
rafl, the same could apply to any binding.
gaal sorry, was whisked away by $work 13:48
I think I've settled on `sub foo (args)` in signatures and `foo(args)` in calls. `( args )` looks weird to my eyes. 13:49
theorbtwo To mine too, gaal. 13:50
gaal though in Perl 5 I often said things like for (@{ $listref }) { ... }
theorbtwo OTOH, I don't like that calling a sub like you write the signature will silently do the wrong thing in many cases.
gaal significant whitespace is teh weird, yeah. can't say it isn't useful at times, though! 13:54
rafl autrijus: ping 13:58
Limbic_Region salutations all 14:04
kolibrie good morning, Limbic_Region 14:05
liz6 rafl: I think autri is en route from helsinki to taipei right now... 14:06
rafl liz6: Ah, OK. ETA?
liz6 not sure: I wouldn't expect anything from autri in the next 24 hours... 14:07
Limbic_Region wouldn't bet on that
liz6 ;-)
seen autrijus
jabbot liz6: autrijus was seen 23 hours 50 minutes 40 seconds ago
liz6 L~R: I guess that would be a record, no? 48 hours? ;-) 14:08
Limbic_Region liz6 - there have been times where he wasn't around in IRC but was still checking in to svn 14:09
meant to ask you about your coroutine specs
if I create a coroutine and have a yield in the body of the loop, does the coroutine work like a full-body continuation?
and not like a p5 fancy goto? 14:10
liz6 it's my understanding that execution will continue after the yield() at the next invocation
Limbic_Region IOW - can I resume inside a loop requiring initialization where in p5 (simulating coroutines with fancy goto tricks) you could not
liz6 regardless of where that is...
Limbic_Region ok - cool 14:11
liz6 off to do some shopping and then back to $work
robkinyon What's the difference between Ruby's coros and what is planned for P6? 14:59
Limbic_Region read S17
robkinyon thx
Limbic_Region rather - read S17 that's in the Pugs repository
keeping in mind it was specced by liz and autrijus (with some influence by wendy) 15:00
IOW - it is a draft that @larry needs to stamp
robkinyon wendy?
liz6 my spouse... 15:01
r0nny yo 15:02
Limbic_Region in any case, Parrot has got the goods - the decision on the p6 side is how much/little to use 15:03
and how to handle things like parameters
which can go several different ways with strong supporting arguments for each
robkinyon liz6: Oh - I didn't know you were married. :-) 15:04
limbic_region: Hmmm ... I'm liking Ruby's coros a lot
Limbic_Region robkinyon - I am Ruby ignorant, mind giving me an example? 15:05
svnbot6 r7914 | iblech++ | docs/quickref/data: Fixed somewhat critical typo, spotted by azuroth++.
robkinyon one sec
Limbic_Region and have you read Dan's "What the heck is a coroutine" blog?
robkinyon LR: Yeah, I've read it
i need to reread it after having learned Ruby, though 15:06
Limbic_Region www.sidhe.org/~dan/blog/archives/000178.html
robkinyon my eyes glazed a bit cause i didn't have any experience. :-)
Limbic_Region yeah - fortunately for me, I was in #parrot at the time and got some real time explanations
understanding how continuations and coroutines differed really solidified things for me 15:07
but I would be happy to see how coroutines work in Ruby
unfortunately - all implementations of coroutines that I know about are still just emulations of Knuth's original brain child
lisppaste3 robkinyon pasted "Ruby coros" at paste.lisp.org/display/13367 15:08
robkinyon That's from my port of Tree.pm to Ruby - it gives traverse() both implements and uses coros
Limbic_Region the case statement isn't required though right 15:09
robkinyon I use the case statement to determine which traversal method I'm supposed to do
Limbic_Region the yields could just as easily have been outside?
robkinyon yeah
Limbic_Region I don't see anything there that would lead me to believe there will be undesireable differences 15:10
robkinyon well, you mentioned a number of different WTDI 15:11
Limbic_Region 1 thing you have there is just what I was confirming with liz - that coroutines can yield and resume inside a construct requring initialization
robkinyon "construct requiring initialization"?
Limbic_Region in p5 speak, gotos can't jump inside a loop or a subroutine
robkinyon ah
gotos can't enter a scope? 15:12
Limbic_Region they can as long as initialization isn't required
liz6 from Dan's blog:
"One thing to remember is that you can yield out of a coroutine no matter how deeply you're into it. Our examples are really simple, but you could be nested ten or twenty levels of scope deep in a coroutine and still yield out--when you re-invoke the coroutine you'll be dropped back where you were, ten or twenty levels deep, with all your lexicals put back in place."
robkinyon That's how Ruby's coros work
plus, you can pass in a Proc object that has a completely different scope and yield to it 15:13
Limbic_Region in any even robkinyon, that's not the issue with different WTDI
robkinyon (which is important)
Limbic_Region the issue is with parameters
robkinyon ok ... ?
Limbic_Region let me nopaste
robkinyon k
Limbic_Region perlbot nopaste
perlbot Paste your code here and #<channel> will be able to view it: sial.org/pbot/<channel>
pasteling "Limbic_Region" at 129.33.119.12 pasted "coroutines and parameter binding" (24 lines, 356B) at sial.org/pbot/14202 15:16
Limbic_Region sorry - was filling in commentary
robkinyon huh 15:17
Limbic_Region what huh?
robkinyon lemme see what ruby does 15:18
ok - ruby has a strict type signature
so, "def foo ( a )" requires that there's a parameter every time 15:19
Also, yield will only yield to the associated block
So, you cannot yield to your caller - you only yield to the coro that you've been handed to
Limbic_Region hmmm - not sure I understand that last statement, but before you explain 15:20
does changing the parameter affect how that parameter is bound 15:21
for instance - does calling foo(3) where the yield is returning the parameter change
if you subsequently invoke with foo(5)
?
robkinyon yield doesn't yield to the caller 15:22
it yields to the associated Proc object
Limbic_Region *shrug* - I guess I need to understand Ruby to understand that distinction
robkinyon So, if you had "def foo ( a ) yield a", you would call it as "foo(3) { |x| print a }"
You associate a block with the method that has a yield in it 15:23
Limbic_Region oh, I get it
hmmm
robkinyon the caller isn't involved, save to potentially provide scope to the associated block
Limbic_Region would definately have to defer you to liz, autrijus, etc on that one
robkinyon heh
I like ruby's coros
they're very neat and clean
Limbic_Region while I can conceptually understand it - I have no idea about Parrot's implementation and therefore Perl6's ability to bend that implementation to your will ;-) 15:24
robkinyon hmm
coros, as I think you're understanding them, allow the function to yield to the caller, then resume functioning where the yield occurred the next time it's called? 15:25
Limbic_Region yes
where as a continuation can be resumed anywhere
robkinyon ok - in ruby, coros are primarily used to implement iterators
SamB Limbic_Region: Parrot is a VM -- what difference does the implementation make?
robkinyon for example, foo.each { |x| ... }
Limbic_Region SamB - a matter of emulation (or how much work it takes to get the job done) 15:26
SamB - in p5, you can simulate coroutines - but only to a certain extent
robkinyon LR: You could implement them fully, but you would have to maintain all the state yourself within the coro
Limbic_Region for instance - goto's can't jump inside constructs requiring inialization
robkinyon You'd practically have to implement a VM within each coro
SamB Limbic_Region: maybe you meant Parrot's specification? 15:27
robkinyon you'd have to unroll everything. :-)
Limbic_Region SamB - right
thanks for clarification
liz6 isn't that what continuattions are about?
Limbic_Region yes liz6
but to make robkinyon's Ruby syntax work in perl6 likely requires some manipulation 15:28
and I don't know how much or where to start
liz6 ah, ok, in that sense...
Limbic_Region which is why I deferred
SamB now if we were talking about, say, Python's VM, it would be "implementation", because there is no spec for that.
Limbic_Region and you would definately know
so WRT my earlier statement, I should have said spec and subsequent implementation of that spec (as there are no guarantees the implemenation matches the documentation) 15:29
robkinyon well, it actually doesn't take much
Limbic_Region wanders off for a sec
robkinyon In essence, Ruby's coros are more syntactical sugar for P5's & prototype that happens to be at the end instead of the beginning 15:30
And, you get a keyword for calling that subref called "yield [ARGS]" instead of doing $subref->([ARGS]) yourself. 15:31
Look at www.perlmonks.org/?node_id=506753 for a few different implementations of Ruby's Array.partition() in Perl
In Ruby, the block is at the end where it's in the beginning in Perl5 15:32
You could easily do this in P6 with a method signature
Limbic_Region liz6 - has S17 been handed off to @larry for their perusal?
robkinyon The only thing in Ruby is that you don't have to know about the block - it's just there for you unless you need to pass it along, as I did in my nopaste 15:33
Limbic_Region if so, perhaps robkinyon wants to post a question to p6.l as food for thought when they are considering it
liz6 L~R: nope, still working on it, while $work is piling up...
;-( 15:34
Limbic_Region oh, well then I will shut up and let you get back to it
liz6 suggestions / patches welcome!
robkinyon I'll wait until it's posted to ask my question
kolibrie just saved several minutes by using pugs to do some number crunching on a YAML file 15:38
(pugs)++ (Perl 6)++
r0nny (black perlvodoo)++ ;P
dada <<pugs perl6>>^++ 15:40
?eval <<pugs perl6>> 15:41
evalbot_7914 ("pugs", "perl6")
dada ?eval my $x = 1; my $y = 2; ($x, $y)^++
evalbot_7914 Error: unexpected end of input expecting term
dada oops
my perl6 is rusty already 15:42
robkinyon ?eval <<pugs perl6>>^++
evalbot_7914 Error: unexpected end of input expecting term
robkinyon ?eval <<pugs perl6>>^++;
evalbot_7914 Error: unexpected ";" expecting term
robkinyon ?eval <<pugs perl6>>^+ 1 15:43
evalbot_7914 1
robkinyon ?eval <<pugs perl6>> ^+ 1
evalbot_7914 1
dada ?eval <<1 2>> ^+ 1
evalbot_7914 1
dada ?eval my @a = (1,2); @a^++; @a;
evalbot_7914 Error: unexpected ";" expecting term
dada ?eval my @a = (1,2); @a^+=1; @a; 15:44
evalbot_7914 Error: unexpected "=" expecting "0", digit or fraction
dada ?eval my @a = (1,2); @a>>+=1; @a;
evalbot_7914 Error: unexpected ">" expecting word character, "::", term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
dada don't remember
?eval (1,2)>>+<<(2,3) 15:45
evalbot_7914 (3, 5)
dada ?eval (1,2)>>+1
evalbot_7914 Error: unexpected ">" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
dada ?eval (1,2)>>+ 1
evalbot_7914 Error: unexpected ">" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
dada isn't this a bug? 15:46
?eval (1,2)>>+<<1
evalbot_7914 (2, 3)
dada ?eval (1,2)>>++<<
evalbot_7914 Error: unexpected "<" or ">" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
dada ?eval <<pugs perl6>> >>+<< <<perl6 pugs>> 15:48
evalbot_7914 (0.0, 0.0)
dada ?eval <<pugs perl6>>>>+<<<<perl6 pugs>>
evalbot_7914 Error: unexpected "<" or ">" expecting term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
dada heh
?eval <<1 2>> >>+<< <<2 3>> 15:52
evalbot_7914 (3.0, 5.0)
Limbic_Region ?eval (1, 2) >>+<< (2, 3) 15:56
evalbot_7914 (3, 5)
Limbic_Region ?eval(1, 2, 3) >>+<< (4, 2) 15:59
evalbot_7914 (5, 4, 3)
dada ?eval <<j p 6>> >>~<< <<a h>> 16:03
evalbot_7914 ("ja", "ph", "6")
dada ?eval [~] <<j p 6>> >>~<< <<a h>>
evalbot_7914 "japh6"
dada yippee!! :-)
robkinyon ?eval <<pugs perl6>> >>++
evalbot_7914 Error: Can't modify constant item: VStr "pugs"
theorbtwo Hm, [~] <<j p 6>>»~«<<a h>>
robkinyon ?eval <"pugs" "perl6"> >>++
evalbot_7914 Error: Can't modify constant item: VStr "\"pugs\""
theorbtwo ?eval [~] <<j p 6>>»~«<<a h>> 16:04
evalbot_7914 Error: unexpected "\\" expecting term postfix, operator, ":", ",", postfix conditional, postfix loop, postfix iteration, ";" or end of input
robkinyon ?eval <$x="pugs" $y="perl6"> >>++
evalbot_7914 Error: Can't modify constant item: VStr "$x=\"pugs\""
robkinyon ?eval $x='pugs';$y='perl6';<$x $y> >>++
evalbot_7914 Error: Undeclared variable: "$x"
robkinyon ?eval my$x='pugs';my$y='perl6';<$x $y> >>++
evalbot_7914 Error: Can't modify constant item: VStr "$x"
robkinyon Gack!
?eval my$x='pugs';my$y='perl6';($x $y) >>++
evalbot_7914 Error: unexpected "$" expecting operator or ")"
robkinyon ?eval my$x='pugs';my$y='perl6';($x, $y) >>++
evalbot_7914 Error: Can't modify constant item: VStr "pugs"
robkinyon ?eval my$x="pugs";my$y="perl6";($x, $y) >>++ 16:05
evalbot_7914 Error: Can't modify constant item: VStr "pugs"
dada it doesn't do ++ on strings yet, maybe
robkinyon feh
dada ?eval my $x=1; my $y = 2; ($x, $y) >>++
evalbot_7914 Error: Can't modify constant item: VInt 1
dada no, it's something deeper
?eval my $x=1; my $y = 2; ($x, $y) >>+<< 1
evalbot_7914 (2, 3)
dada ?eval my $x=1; my $y = 2; ($x++, $y++) 16:06
evalbot_7914 (1, 2)
dada that's unorthogonal
robkinyon ?eval my $x=1; my $y=2;(++$x, ++$y) 16:07
evalbot_7914 (2, 3)
robkinyon ?eval (2,3) >>++
evalbot_7914 Error: Can't modify constant item: VInt 2
dada ?eval 2++
evalbot_7914 Error: Can't modify constant item: VInt 2
robkinyon yeah
dada that's orthogonal
robkinyon LOL
it's orthogonal to itself, but doesn't look very spec-ish 16:09
dada no no
it's unorthogonal
theorbtwo my $x=1; my $y = 2; ($x, $y) >>++ not working doesn't make sense to me. The rest do. 16:10
?eval my $x=1; my $y = 2; ($x, $y)>>++
evalbot_7914 Error: Can't modify constant item: VInt 1
theorbtwo Just making sure.
dada you can't do 2++ and you can't do (2,3) >>++
but you can do ($x++, $y++) and you can't do ($x, $y) >>++
that's unorthogonal
robkinyon or, rather ++$x,++$Y 16:11
dada well, the result doesn't matter
the fact is that ($x, $y) >>++ should be perfectly equivalent to ($x++, $y++)
PerlJam dada: depends on if () created a tuple (ala python) or not.
if ($x,$y) is a tuple, then it's immutable.
no lvalue there, only rvalue 16:12
dada nope
?eval my($x, $y) = (2,3); ($x, $y)
evalbot_7914 (\2, \3)
dada ouch
?eval my $x = 2; $x 16:13
evalbot_7914 \2
dada you see? it's orthogonal
no tuple here
theorbtwo dada: I think you are using orthogonal in non-useful ways. 16:14
The problem is that the ($a,$b) that >>++ is trying to apply to needs to work somewhat differently from a random list-forming ($a, $b).
It needs to act as the argument list to >>++, and get bound with :is rw, instead of :is copy or :is ro. 16:15
dada mmm
?eval my @a = (1,2); @a >>++; @a
evalbot_7914 Error: Can't modify constant item: VInt 1
theorbtwo OK.
Now I don't know WTF is going on.
dada it's the same, I guess
?eval my @a = (1,2); @a++; @a 16:16
evalbot_7914 [3,]
dada !!!
theorbtwo @a is taken in scalar context, get 2, ++ that, assign it back to @a.
dada yes, right 16:17
theorbtwo ?eval my @a = (1,1); @a++; @a
evalbot_7914 [3,]
dada ?eval my @a = (1,2,3,4); @a++; @a
evalbot_7914 [5,]
dada ok
theorbtwo I'm not real sure that's what it should do, but I can't think of anything saner for it to do. 16:18
broquaint It should die. Loudly.
dada 1; 16:19
theorbtwo "Die loudly" seems like a fairly good option, but how does it do that without making +@a or @a=3 not DTRT?
dada mmm 16:20
@a=3 ?
broquaint Surely it would just be: method postfix:<++>() { die "Don't do that" }, or am I missing something?
dada shouldn't that be interpreted as @a=(3) ?
?eval @a = 3; @a 16:21
evalbot_7914 Error: Undeclared variable: "@a"
dada ?eval my @a = 3; @a
evalbot_7914 [3,]
dada ?eval my @a = (3); @a
evalbot_7914 [3,]
dada it's a different thing
err, no, wait, it's not that different
broquaint Calling a scalar mutator on an array shouldn't work (unless it's overloaded, of course). 16:22
dada ?eval my @a = (42, 43, 44); @a += 1; @a 16:23
evalbot_7914 [4.0,]
broquaint Odd that it converted to a float. 16:24
dada ?eval +"1"
evalbot_7914 1.0
broquaint Wacky. 16:25
dada numification works with floats, it seems
anyway, the >>++ problem croaking about constant items is a bug, IMHO
broquaint Yeah, I thought I'd marked that up in a test before. Maybe I never got around to it. 16:26
Or figured it'd be fixed RSN.
eric256 tries feathers web chat...this is cool
?eval (+"1").ref 16:27
evalbot_7914 ::Num
dada ?eval (1).ref
evalbot_7914 ::Int
dada ?eval (+"1.2").ref
evalbot_7914 ::Num
theorbtwo (+1.ref)
broquaint ?eval 0.1.ref
evalbot_7914 ::Rat
theorbtwo ?eval (+1.ref)
evalbot_7914 Error: cannot cast from VType (mkType "Int") to Double (VNum)
theorbtwo ?eval (+1).ref 16:28
evalbot_7914 ::Int
dada ?eval (1.2).ref
evalbot_7914 ::Rat
dada mmm?
?eval +(1.2).ref
evalbot_7914 Error: cannot cast from VType (mkType "Rat") to Double (VNum)
theorbtwo Whew. If the + operator changed the type on things that were already numeric, then I'd be worried.
eric256 pugs has a couple fo number types ;)
theorbtwo Int for integers, Num for floats, Rat for rationals. 16:29
dada ?eval (+(1.2)).ref
evalbot_7914 ::Rat
dada ?eval my $x = 1.2; $x.ref
evalbot_7914 ::Rat
theorbtwo ?eval (1/3.14).ref
evalbot_7914 ::Rat
Limbic_Region would still love to have a B&D way of declaring non-mutable datatypes that would blow up with prejudice at compile time if anything tries to cast them otherwise
theorbtwo Odd; I expected that to be a Num.
broquaint use SML; 16:30
Limbic_Region ;-)
broquaint Or perhaps some fancy perl6 way: use lang :SML; # Mmm, non-backward-compatible. 16:31
theorbtwo use SML sounds right to me... not that I know what SML is...
Limbic_Region 99% of the time, I am not going to care that I have lost a few optimization possibilities to have flexibility in my programs
actually - it is probably closer to 99.95% 16:32
broquaint Standard ML, the language MJD bangs on about in his Strong Typing article.
dada perl6 -pedantic -wall
svnbot6 r7915 | kane++ | r123@coke: kane | 2005-11-09 17:39:55 +0100 16:41
r7915 | kane++ | * update S22 with kane's work in progress.. starting small with metadata -> package creation
r7915 | kane++ | * describe what we have and where we should go
r7915 | kane++ | * note the p4 prototype2 dir where live code can be found
r7915 | kane++ | * move all previous notes to the header:
r7915 | kane++ | =head1 Past Rantings/Notes/Etc
r0nny yo 16:45
xinming hmm, who wrote the cperl-mode.el for Emacs please?
Limbic_Region xinming - if it is in the Pugs repository - use blame 16:46
theorbtwo Limbic, cperl-mode is for perl 5 code. 16:47
I seem to recall it being updated by somebody on p5p lately, though. 16:48
autrijus it's since adopted p6-fu 16:49
Limbic_Region ok - but it appears that renormalist is responsible for putting it in the Pugs repository
autrijus using a (defun cperl-looks-like-perl6)
Limbic_Region wins the bet
hey autrijus
autrijus greetings. just back to .tw and the net
Limbic_Region right - liz6 suggested another 24 hours, I said don't bet on it 16:50
autrijus finished the book jhi gave me during the plane trip
Limbic_Region xinming - Steffen "renormalist" Schwigon (SCHWIGON) is the one that apparently put the file in the Pugs repository
autrijus "Unicode Demystified" - a mere 900 pages
my eyes hurt ;)
theorbtwo 900!
theorbtwo didn't know Unicode was that Mysterous in the first place. 16:51
PerlJam theorbtwo: depends on which end you're looking at it from.
autrijus I learned all about Hangul layout, Arabic shaping, BiDi and vertical, Han character composition forms, diacritical stacks, etc.
xinming Limbic_Region: hmm, thanks, But It seems, the file has problem working with cvs version of Emacs.
Limbic_Region: there will be a message while loading the file. :-/ 16:52
Limbic_Region *shrug*, I use a real editor - vi 16:53
Limbic_Region ducks
eric256 pictures purl six written vertiacly. hmmm
Limbic_Region autrijus - I speak a bit of Hangul
autrijus wonders who to charge with maintaining S15
(Unicode)
Limbic_Region it is one of the most scientific written languages ever created
xinming by the way, anyone here knows why .p6 is chosen instead of .pl6 as extension.? it seems that .pl6 be more suitable. 16:54
autrijus Limbic_Region: ooh, cool. yeah, it's very computational
PerlJam autrijus: clearly you are the best choice! ;-)
Limbic_Region which happened only about 500 years ago
PerlJam xinming: no way!
autrijus PerlJam: clearly I'm not ;) dan kogai maybe
xinming PerlJam: Why?
Limbic_Region the spoken language (much older) was previously written in chinese and only the very affluent could read/write
PerlJam autrijus: oh, Dan would be an excellent choice.
eric256 xinming: .pl, .pm, .p6 (all two letters. don't break the pattern!)
autrijus xinming: I think extensions are somewhat bogus, and you can certainly call perl6 programs .pl 16:55
theorbtwo Yeah, so far I've simply gotten away with ignoring non-latin-based languages, and ignoring the actually putting characters on the screen bit. (The former being hard since I don't know any, and the later being somebody else's problem.)
Limbic_Region then came the 3rd king (his older two brothers abdicated so he could take the thrown) got all the countries bad @$$ linguists/scientists together to create a written language accessible to everyone
xinming eric256: well, this makes sense. :-)
PerlJam xinming: .pl6 is aesthetically displeasing for a variety of reasons
xinming: besides .pl has a troubled history 16:56
theorbtwo So: Unicode isn't mysterious if you ignore the hard parts.
eric256 although if we use pl6 we could rename the languag plsix pronounced pel-ix ;) 16:57
PerlJam eric256: plsix == please icks 16:58
Limbic_Region plans on compiling all of his p6 programs to byte code so the .pbc extension makes sense ;-) 16:59
autrijus Limbic_Region: as soon as lexpad is there we can have -CPIR.PBC :) 17:00
xinming autrijus: hmm, I know, on Os which determines the file is executable is by an attribute. But On Os which isn't, the extension is used. that's why I think .pl and .pl6 is better, as some program will generate .p1 .p2 .p3 for packages(though there isn't one), hmm, IIRC, winrar use this scheme if you compress a large file and divide them into small part.
autrijus or should it be -CParrot.PBC
xinming: I don't think winrar uses .p1 etc... or I might be wrong
LimbicRegion grrr - I hate when that happens 17:01
theorbtwo No, it uses r?? -- rar and r00, r01...
xinming autrijus: winrar don't use .p1 or .p2. It use r01 r02 r03...
autrijus right
LimbicRegion autrijus - jhorwitz showed up in #parrot today asking about the state of Pugs/Parrot interopability (specifically with objects) 17:02
I mentioned the "awaiting lex implementation"
and pointed him to your "day before parrot release" journal entry
xinming hmm, why choose p6, In my opinion, It might be for the compiled byte-code.
autrijus the bytecode would be .pbc for the parrot runtime
LimbicRegion: oh nice 17:03
PerlJam xinming: Are .pl and .pl6 files for Perl Libraries and Perl 6 Libraries respectively?
Limbic_Region was only joking about compiling everything to bytecode though
PerlJam xinming: and what about .plx files? 17:04
xinming PerlJam: .pl is stand for the "current" version of perl(maybe perl 5, maybe perl 6), and .pl6 for specified version of perl, for perl 6
PerlJam the trouble with having a good memory is that you don't forget stuff ;)
Limbic_Region PerlJam - the problem with having a good memory with bad recall is that you feel even more stupid that you hadn't forgot something but needed someone else point it out to you 17:05
geoffb [OT] Anyone happen to know a decent free tool that can easily diagram relationships in a SQL DB? $client is allergic to ASCII art, I'm allergic to MS Access, and the sooner this gets dealt with, the sooner I can be a Perl 6 Playa again . . . :-) 17:06
xinming PerlJam: It doesn't conflict with .pl? scheme. :-)
Limbic_Region wonders if there is a test that can actually measure memory without the limitations of recall
geoffb - you lost me at "free" 17:07
geoffb sigh
Time to relearn PIC . . . .
theorbtwo You could try dia, I hear good things about it.
Of course, that'd be nonautomated.
autrijus geoffb: SQL::Translator::Producer::GraphViz 17:08
Limbic_Region geoffb - www.databaseanswers.com/modelling_tools.htm
autrijus (iirc.)
Limbic_Region google is your friend
geoffb theorbtwo, no problem, it's only 10 tables, half of which are just audit versions of the other half, with absolutely trivial relationships . . . but $client just hates doing things the easy way . . . 17:09
Limbic_Region, point taken
autrijus, hmmm, thanks for the idea
Limbic_Region didn't realize there were so many free options out there
Limbic_Region wonders if he should have fought the purchase of Erwin a little harder now 17:10
Limbic_Region wanders off for lunch
stevan hullo autrijus :) 17:12
autrijus greetings stevan 17:14
robkinyon autrijus: I have a question re: single-pass interpreters (like Ruby/Perl5), if you don't mind 17:21
autrijus sure
robkinyon I was looking at Ruby's source to see if I could add in user-defined lexically-scoped syntax 17:22
The way it's written, the keyword and ops lists are both static and the lexer is hard-coded
PerlJam perl5 is single pass? I've always thought of it more like a 2-steps-forward-one-step-back kind of compiler :)
robkinyon Shouldn't it be possible to have a table of funcps that respond to the longest-match lexeme, with context thrown in there for good measure, calling the top funcp on the stack (to provide user-defined capabilities) ? 17:23
eric256 perljam...i think i heard its a two pass, but the second pass progrresses right behind the first pass. that ameka any sense? ;)
robkinyon And, that, in theory, shouldn't be too much slower than the current switch table
autrijus robkinyon: I think you just described perl6 macros. 17:24
robkinyon Oh
:-)
In theory, that's not too hard to implement (assuming a strong graps of C which I had and lost)
PerlJam gives robkinyon a strong graps so that he can apply it to C 17:25
robkinyon The only part I'm finding problematic in my mental model is how to determine which lexemes are legal where. For example, within the condition of an if-statement, certain things are legal that aren't legal outside it, and vice-versa 17:26
So, one would associate a lexeme within a given context to a given funcp ... ?
Or, am I off base? 17:27
autrijus robkinyon: yes. the context is called grammatical category
robkinyon ahhh
autrijus the lexeme is parsed using "is parsed" specification
robkinyon wasn't lucky enough to take compiler theory in college
autrijus the funcp is the macro
robkinyon "is parsed" specification? 17:28
autrijus and it's not comp theory, it's perl6 :)
robkinyon heh
hmm
autrijus for some bit about macros, see S06 (and A06 for fun) 17:29
for grammatical categories, see S12
robkinyon I was thinking about hacking this into the Ruby executable
seeing as ruby's codebase is significantly easier to work with than Perl's
autrijus I'm not too sure about that :) 17:30
you may have better luck with yc/yarv... maybe
robkinyon yc/yarv?
autrijus the nextgen ruby runtime
yarv is ruby's parrot
see this www.atdot.net/yc/compile 17:31
also www.atdot.net/yarv/
stevan YARV++
robkinyon I want to avoid working with Parrot until I see Parrot actually implement a full range of stuff
Plus, I like how everything in Ruby is an object, including the runtime
stevan robkinyon: not parrot,.. its an unrelated project
robkinyon Oh - so ruby won't be targeting parrot? 17:32
stevan robkinyon: who knows,. but that is not what YARV is :)
robkinyon: see the inst table www.atdot.net/yarv/insnstbl.html
the first 50 are all you really need,..everything after that is an optimization 17:33
#26 newrange
autrijus: speaking of bad-ass language runtimes :)
autrijus: any more thoughts/work on the sendmsg stuff? 17:34
autrijus stevan: you saw that I'm converting the entire PIL special forms into sendmsgs
stevan autrijus: no I didnt,.. where? 17:35
autrijus next step is introduce Arglist and Siglist objs
to replace even more special forms
stevan nice
did you see I added a closure::signature last night
autrijus stevan: see use.perl.org/~autrijus/journal/27490 on Apply, Bind and Assign
yeah I saw that
stevan so that is supports return type checking :)
theorbtwo Oh, speaking of, did you see... 17:36
autrijus very nice
theorbtwo ?eval my ($a, $b) = (1,2); ($a, $b)>>++;
evalbot_7915 Error: Can't modify constant item: VInt 1
rafl autrijus: I want to give a talk about perl6/pugs on friday. I'd like to steak something from your pugs talks. How are they licensed? Am I allowed to steal?
theorbtwo AFAICS, that's getting treated as j-random-list, and not like an arglist. 17:37
autrijus rafl: I hereby release it in the Microsoft Permissive License ;)
rafl: or SQLite "license", or CC-by-2.0, or BSD3, or whatever 17:38
rafl autrijus: I interpret that as a "Yes" :-)
autrijus that interpretation would be correct.
sili charge royalties
rafl autrijus: I'm also happy to hear your Pugs and Haskell talks in march :-) 17:40
robkinyon autrijus: Thanks 17:41
autrijus rafl: :D 17:42
rafl I'm still in doubt about the visual basic talk, though. 17:45
autrijus: Hey! Stop making new talks or updating them.. I get a mail every time! ;-) 17:47
autrijus couldn't quite bring myself to journal... tomorrow then 18:00
good night folks :) &
geoffb good night, autrijus
Simonff Hi - can anyone help me with a Pugs compilation error? 18:09
geoffb Simonff, nopaste it
Simonff geoffb: www.rafb.net/paste/results/IY7AkL50.html 18:10
geoffb Hmmm, someone got this a day or two ago . . . I don't recall the answer, but it should be in the IRC logs (see pugscode.org -> IRC logs in the left menu) 18:11
Simonff That was me, on Sunday. I was told to try the SVN version, but the problem persists. 18:14
geoffb oh.
sorry about that
Simonff np :)
svnbot6 r7916 | asavige++ | added golfex.t misc golf examples 18:30
r7917 | asavige++ | fixed comments in golfex.t 18:48
robkinyon In #ruby-lang, someone just asked "How do I write a function that accepts a single array containing two numbers?" 19:10
I'm curious - what would the P6 signature for that look like?
iblech robkinyon: my sub foo ([$first_elem, $second_elem]) {...} IIRC 19:12
(Which will, of course, work with both foo @array and foo $arrayref)
robkinyon my sub foo ([Int $width, Int $height]) { ... } ? 19:13
iblech Yep.
see dev.perl.org/perl6/doc/design/syn/S06.html, "Unpacking array parameters"
robkinyon What about: my sub foo ([Int $width, Int $height], $other_param) { ... } ? 19:14
iblech Works too
robkinyon when is this going to be released again?
iblech You can even say things like ([Int $width, Int $height = $default_height, Int ?$optional_depth])
theorbtwo When it's ready.
iblech :)
robkinyon theorbtwo: *sighs* 19:15
nothingmuch if it's an array taking two params it might as well be a tuple 19:28
cognominal_ this is a form pattern matching (in the haskell sens). I wonder why perl6 does not go the whole way and does not explicitely support algebraic types 19:59
s/form/form of/
robkinyon Question: Will there be an analog to Ruby's :foo? 20:04
That creates a singleton symbol called :foo that can be used everywhere
21:05 spinclad_ is now known as spinclad
mugwump ?eval "RT"++ 21:53
evalbot_7917 Error: Can't modify constant item: VStr "RT"
mugwump :)
?eval ($x="RT")++
evalbot_7917 Error: Undeclared variable: "$x"
mugwump ?eval (mr $x="RT")++
evalbot_7917 Error: No compatible subroutine found: "&mr"
mugwump ?eval (my $x="RT")++
evalbot_7917 Error: No compatible subroutine found: "&my"
mugwump ?eval my $x ="RT"; $x++
evalbot_7917 "RT"
mugwump There you go, obra - RT++ is still RT 21:54
obra :) 21:55
coral interesting. it's ok to modify an instance of a constant but not the constant itself? 22:05
cognominal_ ?eval say (0..)[0] 22:36
evalbot_7917 Error: unexpected "(" expecting block construct, ":", term, term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
obra seen autrijus
jabbot obra: autrijus was seen 4 hours 36 minutes 10 seconds ago
cognominal_ ?eval say (qw(a) )[0]
evalbot_7917 Error: unexpected "(" expecting block construct, ":", operator, ")", term postfix, postfix conditional, postfix loop, postfix iteration, ";" or end of input
cognominal_ say qw(a)
?eval say qw(a) 22:37
evalbot_7917 a bool::true
wolverian ?eval say (0...)[0]
evalbot_7917 0 bool::true
wolverian cognominal_, you missed one .
cognominal_ thx
sub take(Int $n, Code &f) { (1..$n).map:{ f() } } 22:38
my $compteur = coro { for 1... -> $i { yield $i; } };
my @a;
say ( @a = take 10, $compteur ) while @a[0] <20;
apparently 1.. is not lazy because it loops without returning anything 22:39
if I give an finitie integer as upper bound everything is fine 22:40
eric256 opps...don't hit F5 in the web based IRC chat ;) 23:06
cognominal_ encore une de ces conneries de la nasa pour eviter une collision d'un asteroid avec la terre sur /. 23:34
ils doivent vraiment avoir du mal a justifier leur budget. 23:35
buu ?eval say +(0..)[2] 23:49
evalbot_7917 Error: unexpected ")" expecting term
buu Eh?
?eval say (0..) 23:50
evalbot_7917 Error: unexpected "(" expecting block construct, ":", term, term postfix, operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
buu ?eval my @x = (0..); say @x
evalbot_7917 Error: unexpected ")" expecting term
buu Wtf?
?eval my @x = 0..; say @x
evalbot_7917 Error: unexpected ";" expecting term
wolverian ...
not ..
buu ?eval my @x = 0...; say @x
evalbot_7917 pugs: out of memory (requested 1048576 bytes)
buu heh
So much for 'lazy'
?eval say (0...)[0] 23:51
evalbot_7917 0 bool::true
buu ?eval say (0...)[3]
evalbot_7917 3 bool::true
buu ?eval say (0...)[180]
evalbot_7917 180 bool::true
buu Interesting.
sili eh? what is that? 23:57
infinity?
mugwump ?eval say (0...)[-1] 23:58
evalbot_7917 pugs: out of memory (requested 1048576 bytes)
sili i think that means yes.
mugwump should be inf! :)

Powered by Cro and the Raku® Programming Language.

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

Thank you!