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.
svnbot6 r15143 | putter++ | redsix now smokable, but with unexamined regressions. 02:17
zev I'm not sure who to report this to, but www.nntp.perl.org seems to be down 05:35
dduncan hey, you're right! 05:38
lambdabot dduncan: You have 1 new message. '/msg lambdabot @messages' to read it.
dduncan is that so?
masak what is the return value of &take? 10:16
VanilleBert ?eval &take 10:31
evalbot_r15143 \{Prim ([Pugs.AST.Internals.Val] -> () Pugs.AST.Internals.Val)} 10:31
masak is this expected behaviour? 11:00
?eval sub next($a) { return $a + 1 }; my $i = 0; gather { take $i = next($i) while $i < 10 }
evalbot_r15143 (\10, \10, \10, \10, \10, \10, \10, \10, \10, \10)
masak O_O
masak do recursive coroutines make any sense? 13:39
allbery_b recalls a contrived example of them in the icon book, so they apparently make some sense 13:53
masak allbery_b: because I found myself writing one in pugs-perl6 today, but I didn't get it to work properly 13:58
in the end I rewrote it in an iterative version, which did work
but what I failed to find out was whether the recursive version did not work out due to a bug in pugs or a but in me 13:59
s/but/bug/
eh, I mean s:2nd/but/bug/
xpika2 there sounds nothing wrong with recursive coroutines. 14:05
masak xpika2: what I was thinking, more exactly, was: what are the semantics of &yield when the coroutine has called itself a few times? 14:27
does the yielded value end up in the next higher instance of the coroutine itself, or does it percolate all the way up to the last instance of non-co-routine in the callchain? 14:29
I guess one could argue both ways
the former seems saner and less complicated, but the latter might also be useful (and would have been in my case)
masak another way of putting it is: does &yield treat recursively called instances of the same function as individual instances, or does it lump them together? 14:48
as of now, there seems to me to be no reason that the latter is the case, unless the former is somehow impossible, which I don't think it is 14:50
szabgab I cannot build Pugs again, the error line I get is: 16:07
Couldn't open "blib/lib/Prelude.js" for writing: No such file or directory
svnbot6 r15144 | audreyt++ | * PIL2JS runjs.pl: Create dirname before outputting it. 16:13
r15144 | audreyt++ | Reported by: szabgab++
TimToady masak: take $i = ... is returning an lvalue because = returns an lvalue 16:14
szabgab audreyt++ thanks, now it compiles well 16:18
btw I keep forgetting where is the pastebot?
allbery_b lisppaste3: url 16:19
lisppaste3 To use the lisppaste bot, visit paste.lisp.org/new/perl6 and enter your paste.
allbery_b there's also
@paste
lambdabot Haskell pastebin: hpaste.org/new
allbery_b but it's not so perl6-friendly 16:20
there may be others hanging out here too :)
szabgab thanks but there was a much shorter link url somewhere 16:21
TimToady I generally use sial.org/pbot/perl6 16:22
lambdabot Title: sial.org Pastebot - pasteling
TimToady have it as a button in my bookmark bar, in fact...
allbery_b actually has a pastes menu in his bookmarks bar, because different channels have different preferred pastebots 16:23
masak TimToady: why does = return an lvalue? in case I need to assign to the variable again? 16:31
how I do de-lvalueify it?
s/I do/do I/ # sigh... 16:32
TimToady I think "take item" works, probably
that's straight from P5, actually ($x = $y) =~ s/// wouldn't work otherwise
masak ah
?eval sub next($a) { return $a + 1 }; my $i = 0; gather { take item ($i = next($i)) while $i < 10 } 16:33
evalbot_r15144 (1, 2, 3, 4, 5, 6, 7, 8, 9, 10) 16:33
masak ok. I have no complaints
TimToady item is $() really, and list is @(). 16:34
masak that's nice
TimToady I just added : forms to go with the () forms
they have list infix precedence on the right
so you can instruct a zip how you want it contextualized
for instance, @@: 1,2 ¥ 3 4 16:35
equivalent to the "chunky" we've been discussing in p6l
and @@:split gets you a chunky split, instead of the default flat one. 16:36
wolverian what does a chunky split return? 16:37
wolverian is being sow
...er, slow :)
TimToady "source of wisdom" I was taking it. :) 16:38
masak also only almost follows
TimToady in p5, split just adds extra fields
wolverian source lacking of wisdom :)
masak wolverian: :)
TimToady split(/(delim)/) gets you field,delim,field,delim,field,delim 16:39
chunky would get you [field,delim],[field,delim], etc
masak ah 16:39
wolverian people capture the delim? the heresy..
TimToady @@: 1,2 XX 3,4 gets [1,3],[1,4],[2,3],[2,4] 16:40
by defaults it's just 1,3,1,4,2,3,2,4
s/s//
so you either write 16:41
for 1,2 XX 3,4 -> $a, $b {...}
or
for @@: 1,2 XX 3,4 -> [$a, $b] {...}
so the default @: works well with the default signature
and zip defaults to not needing [...] in the sig to unpack 16:42
but you can get a list of array if you want via @@:
and we can unify zip and each into just zip
and don't have to have two versions of split, map, take, etc 16:43
@@: automatically takes a list of
list of Capture and turns it into list of Array
also applies to m:g results 16:44
wolverian I kind of thought of captures as tuples, so you wouldn't need to turn them into arrays, but that's sort of different, I suppose 16:47
TimToady when you say "return 1,2,3" you have no idea what the caller wants to do with the listiness of it. 16:48
so we do it lazily
just like when you call "mumble 1,2,3"
it waits till it knows what it's binding to
1 might be in scalar context
2,3 might be list context
wolverian and it goes through captures? 16:49
TimToady Captures and Match objects are very similar
they try to capture the original "syntax" of ordered and unordered things 16:50
rindolf Hi all. 16:55
Hi wolverian, TimToady
TimToady morning | evening 16:56
wolverian hello rindolf
rindolf TimToady: evening. 16:58
I mean it's evening here.
TimToady np, I used an inclusive or
rindolf TimToady: what's up? 17:01
TimToady: have you seen the Pachelbel's Rant yet? 17:02
TimToady er, which rant is that? 17:03
rindolf TimToady: shlomif.livejournal.com/32301.html?mode=reply
lambdabot Title: Post Comment
pasteling "szabgab" at 192.117.127.193 pasted "testing pastebot" (75 lines, 1.4K) at sial.org/pbot/22664 17:07
rindolf TimToady: I've been having some fun with XML::RSS recently, which is based on XML::Parser. 17:08
svnbot6 r15145 | szabgab++ | pastebot client for those who keep forgetting the url
rindolf TimToady: one thing I don't understand is why the object given as callback to the handlers is different than the XML::Parser one. 17:09
Hi szabgab
wolverian rindolf, heh, that video is awesome. thanks. :) never knew about the canon. 17:10
rindolf wolverian: knew what about the Canon? 17:11
wolverian: and you're welcome.
TimToady yes, I've noticed that chord pattern is rather common :)
rindolf Of course, "Let it be" doesn't sound anything like Pachelbel's Canon. 17:12
wolverian rindolf, how common the chord progression is nowadays.
rindolf wolverian: I see.
szabgab rindolf: rehi 17:24
rindolf szabgab: what's up?
svnbot6 r15146 | szabgab++ | more tests for multiline comments 17:39
svnbot6 r15147 | szabgab++ | add "default" and "slurp" to perl6.vim 17:42
putter "This is a big undertaking. Perl 6 is the sort of language that only Perl can parse. But I'll do my best to get vim to." - perl6.vim :) 18:39
Schwern I have a new principle of "good ideas that are really bad in language design"! 18:42
Treat built in types differently!
lumi Nonsense, it works so well in Java
Schwern And Javascript which is what I'm beating my head against right now. 18:43
Pass by reference? Sure! As long as you're not a built-in!
putter It's fun when folks are unwilling to take a N% performance hit to get architectural clarity... and years later are still unwilling, despite it running M x 100% faster. Suggests "starved for computes" remains a primary concern. 18:51
cj Schwern!!17!
Schwern: ah, javascript. what you working on? 18:52
or is that classified?
Schwern cj: I'm reimplementing this www.wium.net/nw/evetool/ 18:53
lambdabot Title: EVE Character Tool
rindolf Hi Schwern
Schwern Mostly as an excuse to learn Javascript
Hi all
And have people send me ISK in EVE. 18:54
cj Schwern: eve looks a lot like homeworld... did they buy up the development team and artists?
Schwern Right now I'm struggling with the fact that in Javascript undefined + 5 == NaN
cj d'oh. 18:55
Schwern cj: Its only superficial.
cj if(foo == undefined){ foo = 0 }
Schwern And that I can't write something like increment(foo.prop, 5)
cj: But that's VERBOSE
cj or (foo == undefined ? 0 : foo) + 5
Schwern But that's BIZARRE
cj Schwern: yeah, well... not everything can be perl
Schwern Best I've come up with is... Object.prototype.increment = function(prop, add) { (this.[prop] = (this.[prop] || 0) + add } 18:57
cj hide it behind something more pretty :)
like that
Schwern Then foo.increment(prop, 5)
But still... yuck
rindolf Schwern: holy.
cj you could also have add + makeSureItsANumber(foo) 18:58
Schwern Its fun to learn new languages... and recoil in horror.
cj: But how do I know what property to look at on foo?
cj there's no type checking, is there?
yeah, your treatment looks good to me other than it looks nasty. limitations of the language, I guess 18:59
there's no operator override either, is there?
Schwern Dunno
cj googles
19:00
binarylord.com/work/js0wood.pdf
is what I meant
type checking 19:01
Schwern That's a language extension proposition
?
cj www.mozilla.org/js/language/js20-20...ading.html
lambdabot Title: JavaScript 2.0 Operator Overloading, tinyurl.com/23rlqy
cj oh? well, then nevermind :)
Schwern Yeah, nobody's using JS 2.0 yet
cj I didn't *read* it :)
Schwern Its all 1.5 if you're lucky.
The Javascript 2.0 spec was written back in 2000. Its just now solidifying. Mozilla is just getting around to implementing it. It'll be stable in most browsers around, say, 2010. And people think Perl 6 is taking a long time! 19:02
cj well, I s'pose $we could start pressing the IE and Mozilla folks to push support into their browsers 19:03
cj but when you rush miracles... 19:03
Schwern IE hasn't even implmented DOM level 2. 19:04
Oh bleh. 19:06
Ok, I'll stop complaining about Javascript here.
putter smokes an old pil2js (6.2.13). bbl 19:13
TimToady someone with more time than I have should rewrite cbstream.rb in p6... 19:19
rindolf Schwern: what's up besides JavaScripting? 19:24
Schwern: any progress with TAP::Harness?
Schwern Ovid has taken that over with TAPx::Parser. They're talking about gutting Test::Harness and using it as the engine making TH 3.0 19:28
rindolf Schwern: I see. 19:29
lumi TimToady: What's cbstream.rb? 19:30
[particle] timtoady: did the S03 reorg make major changes, or just move text blocks around? does it have any ramifications for existing parsers? 19:31
...it's a lot to reread
wolverian TimToady, is there a %%()? 19:36
TimToady wolverian: not yet. :) 19:41
TimToady [particle]: two changes. .accept => .ACCEPT and added @: syntax to be a listop-y kind of @() 19:42
cbstream.rb is the irc gateway that's feeding PerlMonks' chatterbox to #cbstream on freenode.
[particle] TimToady: thanks! 19:43
TimToady that's so that we get @@: to mean "chunky".
[particle] it doesn't look chunky, but it works 19:44
TimToady So @@: means pretend you bound this to a @@ and return what you'd've got if you'd done that... 19:45
And $: @: %: &: @@: are just $() @() %() &() @@() without needing the trailing paren. 19:46
basically reusing the meme that ":" turns a method into a listop. 19:48
except we tweak the precedence slightly so it governs operators like Y and XX as well. (sounds genetic...)
avar heh, more sugar:) 19:49
TimToady anyway, the implicit cbstream.rb challenge is at www.perlmonks.org/?node_id=596867 20:47
lambdabot Title: Re: I prefer to:
TimToady seems like a good chance to see how async {...} works
and, of course, to show how much prettier Perl 6 is... :) 20:49
putter TimToady: re cbstream, add it to project_planning/TASKS or TALK? though it is entirely unclear whether anyone actually looks at them... 21:50
putter smoke of old pil2js is up. 6.2.13 release, r14402, 2006-10-17. 22:04
It's rather less green than I had hoped/remembered. :/
eg all of t/builtins/arrays/ failed. 22:07
putter testing 6.2.12 pil2js... 22:15
If anyone knows a time with a good working pil2js, I'd love a pointer. tnx.
looks like 6.2.12 r10930 2006-06-26 doesn't build under ghc 6.6. :/ 22:19
definition: computer science, noun - a maze of twisty little passages 22:20
stevan: ping? 22:21
anyone know the state of Moose and such? 22:22
the v6.pm "create a p6 runtime on p5" effort? 22:23
ingy hi putter 23:45
putter hi ingy 23:46
ingy :) 23:47
putter: I myself would like to know the answer to that question 23:48
putter working pil2js?
oh, Moose/v6?
ingy yes 23:53
ingy goes off to write a better Javascript version of Digest-MD5 23:54
putter , battered by wandering a dark maze of twisty little passages, sits down against a wall to contemplate which way to go next. 23:56
I wonder if there is a phrase for that variety of yak shaving in which the yaks can teleport, can grow their hair back instantaneously, ... and have an evil sense of humor. 23:58
Pamplonan running of the yaks 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!