Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2009-09-15

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 patspam1 joined #perl6
00:08 Confield joined #perl6
00:12 colomon Is it possible to export variables from a .pm file?
00:16 frew_ joined #perl6
00:17 Limbic_Region joined #perl6
00:24 rhr_ joined #perl6
00:35 kcwu joined #perl6
00:38 colomon Ack, where are perl 6 heredocs when I need them?
00:47 __ash__ joined #perl6
00:49 __ash___ joined #perl6
00:58 maja_ joined #perl6
00:59 colomon How do you do a multiline quote with interpolation?
01:01 avuserow colomon: double quotes worked for me as of the last Rakudo release
01:02 rhr joined #perl6
01:02 colomon Really?  Errr... but that would require escaping all the double quotes in the quote, which would be pretty awkward in this case.
01:03 sm` joined #perl6
01:05 colomon Oh, qq// is it.
01:06 colomon but it breaks horribly on my code because of included {}s which are meant to be text rather than code.  ARgh.
01:07 colomon avuserow: thanks!
01:20 rhr_ joined #perl6
01:34 ihrd joined #perl6
01:55 sevvie joined #perl6
02:04 agentzh joined #perl6
02:05 colomon Huh.  Does \N do weird things to performance?
02:10 colomon geez, I've forgotten basic perl 5 programming.
02:21 s1n pmichaud: ping
02:23 SmokeMachine joined #perl6
02:27 colomon rakudo: my $a = qq[{pi}     is pi]; say $a;
02:27 p6eval rakudo a9ff30: OUTPUT«3.14159265358979     is pi␤»
02:28 colomon Weird.  I'm losing the whitespace after the {} in my code at the moment...
02:57 pugs_svn r28240 | colomon++ | [t/spec] Check in very early draft version of the code to auto-generate a full
02:57 pugs_svn r28240 | suite of trig tests.
02:59 colomon Right now that code is at least an order-of-magnitude uglier than I would like it to be.  I was hoping the wiser heads out there could take a look and advise.
03:00 colomon I'm also getting a segmentation fault now, no idea why, will worry about it in the morning.
03:02 he_ joined #perl6
03:02 BinGOs joined #perl6
03:02 TimToady joined #perl6
03:02 baest joined #perl6
03:02 Patterner joined #perl6
03:02 szbalint joined #perl6
03:02 c1sung joined #perl6
03:02 Tene joined #perl6
03:02 [particle] joined #perl6
03:02 japhb joined #perl6
03:02 szabgab joined #perl6
03:02 huf joined #perl6
03:02 literal joined #perl6
03:02 IRSeekBot joined #perl6
03:02 Ryan52 joined #perl6
03:02 astinus joined #perl6
03:02 tomaw` joined #perl6
03:02 wolverian joined #perl6
03:02 cookys joined #perl6
03:02 sunnavy joined #perl6
03:02 bloonix_ joined #perl6
03:02 bigpresh_ joined #perl6
03:02 edenc joined #perl6
03:02 cosimo joined #perl6
03:03 pugs_svn r28241 | colomon++ | [t/spec] Fix typo in generate-tests.pl commit.
03:03 eiro joined #perl6
03:03 kcwu joined #perl6
03:03 pjcj joined #perl6
03:06 jantore joined #perl6
03:11 ihrd left #perl6
03:12 frew_ joined #perl6
03:12 carlin rakudO: class Alpha { has $!beta; method gamma { my $!beta = 'omega'; while $!beta { say 'Delta'; }; }; }; my $a = Alpha.new; $a.gamma;
03:12 carlin rakudo: class Alpha { has $!beta; method gamma { my $!beta = 'omega'; while $!beta { say 'Delta'; }; }; }; my $a = Alpha.new; $a.gamma;
03:12 p6eval rakudo a9ff30:  ( no output )
03:12 carlin rakudo: class Alpha { has $!beta; method gamma { $!beta = 'omega'; while $!beta { say 'Delta'; }; }; }; my $a = Alpha.new; $a.gamma;
03:12 p6eval rakudo a9ff30:
03:12 p6eval ..OUTPUT«Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤Delta␤…
03:17 carlin Ideally, that first one should have either failed or complained
03:18 s1n phenny: tell pmichaud we need to send out this months' hackathon notice, ideas for locations? saxby's in allen or mckinney is free and usually pretty quiet (my recommendation is saxby's in allen). see http://www.wififreespot.com/tex.html
03:18 phenny s1n: I'll pass that on when pmichaud is around.
03:22 carlin rakudo: class Alpha { has $!beta; method gamma { my $!beta = 'omega'; say $!beta; }; }; my $alpha = Alpha.new(beta => 'delta'); $alpha.gamma;
03:22 p6eval rakudo a9ff30: OUTPUT«delta␤»
03:24 carlin rakudo: class Alpha { has $!beta; method gamma { $!beta = 'omega'; say $!beta; }; }; my $alpha = Alpha.new(beta => 'delta'); $alpha.gamma;
03:24 p6eval rakudo a9ff30: OUTPUT«omega␤»
03:27 carlin rakudo: class Foo { method bar { my $!baz = 'lorem'; say $!baz; }; }; Foo.bar;
03:27 p6eval rakudo a9ff30: OUTPUT«No such attribute '$!baz'␤in method Foo::bar (/tmp/y3FBPZsapq:2)␤called from Main (/tmp/y3FBPZsapq:2)␤»
03:29 * carlin thinks trying to declare a $!foo variable with my should fail or warn but isn't sure ...
03:39 russell__ joined #perl6
03:41 russell__ left #perl6
04:06 maja joined #perl6
04:08 felipe joined #perl6
04:13 ingy diakopter: hi
04:16 molaf joined #perl6
04:24 elmex joined #perl6
04:27 elmex joined #perl6
04:35 elmex joined #perl6
04:35 elmex joined #perl6
04:36 dukeleto joined #perl6
05:59 dukeleto joined #perl6
06:21 sri_kraih joined #perl6
06:26 iblechbot joined #perl6
06:31 alexn_org joined #perl6
06:35 |Jedai| joined #perl6
06:38 elmex joined #perl6
06:41 orafu joined #perl6
06:46 barney joined #perl6
06:47 abra joined #perl6
07:26 frettled joined #perl6
07:26 frettled good morning
07:26 konenkovs joined #perl6
07:33 frettled joined #perl6
07:41 dukeleto frettled: howdy
07:44 kidd joined #perl6
08:06 moritz_ good day everybody
08:08 cmv joined #perl6
08:08 moritz_ I just updated http://rakudo.de/ with new charts (number of commits to rakudo per month; number of projects known to proto)
08:12 Matt-W Good morning
08:15 frettled moritz_++ for the good morning news :)
08:18 * moritz_ bows
08:19 bionoid Hi guys. Did those "Solved in Perl 6" slides make it online?
08:19 moritz_ yes
08:20 moritz_ jnthn.net -> slides
08:20 bionoid great, thanks moritz_
08:25 cmv joined #perl6
08:26 Grrrr joined #perl6
08:30 masak joined #perl6
08:31 masak oh hai, amigos.
08:31 Matt-W I see jnthn made extensive use of unicode support for those slides :)
08:32 moritz_ Matt-W: I think he had a translator :-)
08:32 moritz_ oh hai masak
08:32 moritz_ masak: see bottom of rakudo.de for new charts :-)
08:33 Matt-W I thought jnthn spoke Japanese
08:33 Matt-W I may have imagined that though
08:34 masak oi! jnthn's slides are up? \o/
08:34 moritz_ he also casually kicks Chuck Norris' in the guts when he seems him :-)
08:34 moritz_ masak: aye, on jnthn.net
08:35 masak moritz_++ # modules graph
08:35 Matt-W Those are good slides
08:36 snarkyboojum joined #perl6
08:36 masak curious question: why do the module count start at 20?
08:36 moritz_ because that's where projects.list started
08:36 moritz_ before it had a different name
08:37 masak ah, right.
08:37 masak now I remember.
08:37 masak modules.list or something.
08:37 moritz_ and even before that it was inline as a Perl 6 data structure, or something
08:38 moritz_ of course you're welcome to contribute a more accurate history :-)
08:41 masak I might. :)
08:42 moritz_ I think generating the data once and putting it into projects.hist.0 and then simply concatenating that and the new file would be rather easy
08:43 masak nod.
08:46 masak er. slide 15 of "Solved in Perl 6" is a bit not-recommended...
08:46 masak or rather, line two of that program is redundant. :)
08:47 Matt-W the shuffle a list into random order one?
08:47 masak that one.
08:48 Matt-W I know nothing about that bit of the list api
08:48 moritz_ what's wrong with it?
08:48 masak >>. doesn't enforce any ordering.
08:48 Matt-W Oh yes I thought that
08:48 moritz_ aye
08:48 Matt-W But I'm not sure it's enough to provide actual randomness
08:48 Matt-W Because chances are if you do it twice it may well come out the same
08:48 masak people who assume it does should be punished by the compiler. :)
08:49 Matt-W At least in a non-threading implementation
08:49 masak the idea with that construct is to be able to introduce threading.
08:49 Matt-W Yes, but at the moment it's linear
08:49 Matt-W So would make a very poor example
08:49 moritz_ and even in threading implementations there are chances that on long lists early items are processed before later items
08:49 Matt-W 'ordering not enforced' is not the same as 'random'
08:50 meppl joined #perl6
08:50 masak Matt-W: I think my point is that the program introduces two sources of non-ordering: one intentional, and one discouraged.
08:51 moritz_ aye, it would have beeen better to just .say for @order
08:51 Matt-W yes, in this case you want to print out in order guarenteed
08:51 Matt-W So you should use a for
08:53 masak same in slide 26. :)
08:56 Matt-W slide 26, ordering doesn't matter...
08:56 Matt-W however if that autothreads it could behave strangely
08:56 Matt-W unless there's line-based buffer locking on STDOUT, which is entirely possible
08:56 Matt-W in which case it's okay
08:57 Matt-W I think we're going to find that there are quite a few cases where we might reach for the hyperoperators but where the lack of ordering and potential threading means we have to reach for loops instead
08:57 Matt-W Almost like we need an ordered linear version...
08:57 moritz_ I think that say() and print() should be atomic operations
08:57 Blixtor joined #perl6
08:58 Matt-W moritz_: there's a good argument for that, yes
08:58 moritz_ Matt-W: well, the order version is called "for" :-)
08:58 Matt-W on a per-filehandle basis
08:58 moritz_ *ordered*
08:58 Matt-W moritz_: yes but .name.say for @objects doesn't have quite the same ring to it...
08:59 Matt-W still, it might encourage people to write purer things which work better with hypers
08:59 Blixtor Hi, I've a problem running spectests t/spec/S03-operators/arith with current head, is this expected?
08:59 moritz_ Blixtor: whta's the problem? A few passing TODO tests are expected
09:00 Blixtor perl6 binary consumes more and more memory
09:00 Blixtor and test does not seem to come to an end during 'make spectest'
09:01 masak I'm not sure it'd be a good idea to make print and say atomic.
09:01 masak I'm not expert, but my feeling is that sometimes you'd actually want different threads to print at the same time.
09:02 masak min you, I see the allure of writing @products.sort(*.get_price)>>.name>>.say and similar things. it's very convenient.
09:03 masak on the other hand, saying '.say for ...' isn't very hard either.
09:04 masak ooh! slide 32 is *very* nice.
09:04 masak (the 'multi win(::T, T)' part)
09:04 moritz_ Blixtor: did you update to the latest rakudo?
09:04 moritz_ masak: aye, jnthn++ fixed that a few months ago
09:05 Blixtor moritz_: yes, I did 'git pull'; perl Configure.pl --gen_parrot; make; make spectest
09:05 masak moritz_: I know. I just haven't seen that example before.
09:05 Blixtor moritz_: or is this the wrong way of doing things?
09:06 moritz_ Blixtor: no, that's the right way - I'm trying it here now
09:06 Blixtor moritz_: great, thanks for checking this out!
09:07 am0c joined #perl6
09:07 moritz_ Blixtor: are you on a 32 or 64 bit platform?
09:08 Blixtor moritz_: 64bit, open solaris
09:08 moritz_ arith.t takes 14 seconds here, and doesn't consume huge amounts of memory...
09:08 moritz_ Blixtor: if you run 'make t/spec/S03-operators/overflow.t', does that pass for you?
09:08 Blixtor I killed the test at ~1 gig ...
09:08 moritz_ quite understandable :-)
09:09 Blixtor moritz_: yes, that runs fine
09:09 moritz_ so it's not an issue with big numbers, and I'm at loss here
09:10 jnthn masak: The slide that uses >>. to output sorted results is Wrong.
09:10 jnthn That's what I get for writing slides while jetlagged...I maybe shoulda corrected it before putting them online, but forgot.
09:10 Blixtor @moritz_: anything I can do to help debug this?
09:10 lambdabot Unknown command, try @list
09:10 jnthn (It came up in the talk.)
09:10 masak jnthn: good. :)
09:10 masak jnthn: very nice slides, by the way.
09:10 masak I'm reading 'The Way to Rakudo *' right now.
09:11 moritz_ Blixtor: you could try to rm -rf parrot_install/* and then reconfigure again, if there's an old version of parrot interfering somehow
09:11 Blixtor moritz_: I have this problem since an update yesterday ... a couple of weeks earlier the spectest was already running through fine for me
09:11 jnthn masak: Thanks.
09:11 Blixtor moritz_: I'll try that and let you know then, thanks
09:11 jnthn It's *so* tempting to use >>. for that. :-/
09:12 jnthn I think I may randomize the dispatch somehow.
09:12 mikehh rakudo (a9ff309) builds on parrot r41266 - make test / make spectest (up to 28241) PASS - Ubuntu 9.04 amd64
09:12 mikehh rakudo - t/spec/S03-operators/arith.rakudo - TODO passed:   120, 131-132
09:13 lisppaste3 moritz_ pasted "segfault-at-exit workaround" at http://paste.lisp.org/display/87086
09:13 masak ooh! a mountain metaphor!
09:13 mikehh Note on i386 (r41265) the test only TODO passes 131 and two segfaults after (exit 11) - t/spec/S12-attributes/class.rakudo and t/spec/S14-roles/basic.rakudo
09:14 moritz_ mikehh: when you're on i386 again, could you please spectest with the patch in http://paste.lisp.org/display/87086? (no hurry really)
09:14 moritz_ those two are know (though they used to have a different exit status, iirc)
09:15 mikehh moritz_: I will probably be able to do it after #ps
09:16 mikehh moritz_: I have to on this platform for $work at the moment :-{
09:16 moritz_ mikehh: sure, no hurry
09:16 mikehh stay
09:17 moritz_ it's enough if I get the results before the Rakudo release on Thursday :-)
09:17 jnthn masak: Was a nice excuse to show more pretty pictures. ;-)
09:20 masak (pretty pictures)++
09:20 riffraff joined #perl6
09:20 masak nice to see that over the weekend, the less-than-awesome memory errors on Rakudo exit have been solved. someone++
09:21 moritz_ somehow I think it's NotFound++ or WhiteKnight++
09:29 Blixtor moritz_: reconfigure did not help me, still the same problem with memory usage growing by around 15-20 MB/5sec
09:30 moritz_ Blixtor: last idea: try with a newer parrot. echo 41266 > build/PARROT_REVISION; perl Configure --gen-parrot && make spectest
09:30 moritz_ Blixtor: if that still doesn't help, please open a ticket by writing a mail to rakudobug@perl.org
09:33 Blixtor moritz_: I'll try that
09:52 dakkar joined #perl6
09:55 colomon Blixtor: that sounds kind of like a Rat bug I fixed a couple of weeks ago.  Have you updated your spectests lately?
09:55 zamolxes joined #perl6
10:03 Blixtor colomon: I was (incorrectly?) under the impression that 'make spectest' does the update automatically
10:03 colomon it may?
10:04 carlin It does
10:04 colomon svn info t/spec
10:06 carlin It runs `cd t/spec && svn up`
10:07 colomon okay, so much for that thought
10:07 colomon have you identified the failing line of arith.t?
10:22 Blixtor colomon: no, I've not been digging into this yet, I'm still trying the last suggestion of moritz_
10:22 Blixtor colomon: and doing other things on the side ;)
10:23 colomon I understand.  :)
10:30 Blixtor colomon: ok, the problem is still there, so I'll start digging and see if I can find the problematic testcase
10:33 payload joined #perl6
10:35 hanekomu joined #perl6
10:52 Gothmog_ joined #perl6
11:14 lisppaste3 Blixtor pasted "perl6 binary eats up all memory" at http://paste.lisp.org/display/87090
11:15 Blixtor moritz_: d'oh, I found the problem
11:15 moritz_ Blixtor: do tell
11:15 Blixtor moritz_: I had some problems in the past with this spectest, so I played around with the test a bit
11:16 Blixtor make spectest does the svn update
11:16 Blixtor now there was some conflict while doing the update, which I didn't see
11:16 Blixtor and the make does not complain about it
11:17 Blixtor the conflict markers in turn lead (in this specific case) to the described behaviour of the perl6 binary
11:17 Blixtor have a look at  http://paste.lisp.org/display/87090
11:17 Blixtor this is a minimal test case that leads for me to the described behavior
11:18 colomon wacky!
11:18 Blixtor of course I'd forgotten about playing around with the spectest in the first place ;)
11:19 Blixtor this seems to be some really weird parsing issue, as if you change the 'say' line under the conflict markers slightly (remove the # in the string), it leads correctly to a syntax error
11:20 moritz_ Blixtor: could you please submit that as a bug report?
11:20 Blixtor moritz_, colomon: can you reproduce the behavior with my pasted code?
11:20 colomon give me a moment
11:21 jauaor joined #perl6
11:21 moritz_ Blixtor: I have to rebuild rakudo first :-)
11:21 colomon (just ran it through std.pm)
11:21 Blixtor moritz_: I'll submit the bug report if one of you can confirm - not that there is something else fishy with my checkout here
11:21 frederico joined #perl6
11:21 mberends joined #perl6
11:22 Blixtor moritz_: I'm no longer trusting myself ;)
11:22 moritz_ yes, confirmed
11:22 colomon watching it in activity monitor now
11:22 moritz_ running for 2 minutes now
11:22 jauaor good localtime()
11:22 Blixtor moritz_: watch out, it's getting big rather quickly!
11:23 moritz_ yes, at least 10M per second
11:23 moritz_ jauaor: OH HAI
11:23 colomon yeah, up to one gig vm here...
11:23 jauaor moritz_: HAI FOR YOU TOO
11:23 jauaor :P
11:23 Blixtor ok, I'll file the bug report then, thanks for your help guys!
11:24 sm` joined #perl6
11:24 colomon Blixtor++
11:24 masak Blixtor++
11:25 colomon moritz_: when you have some spare time, can you look at t/spec/S32-trig/generate-tests.pl ?
11:26 moritz_ colomon: I'll probably queue that for tonight
11:26 colomon it's unforgivingly ugly write now, but I don't have any good ideas for making it better.
11:26 colomon thanks!
11:27 fredrecsky joined #perl6
11:31 Blixtor what information about the revision of Rakudo should I include in the bug report? And how do I obtain that?
11:31 colomon git log
11:31 Blixtor I'm trying to follow the instructions at http://www.rakudo.org/developers-guide
11:31 colomon if you used git to get it
11:32 colomon the top commit string from that
11:32 moritz_ git rev-parse HEAD
11:33 Blixtor ok, thanks
11:33 colomon (which is the same as I said, only smarter)
11:34 Blixtor jup, I noticed ;)
11:34 Blixtor bug report is filed!
11:35 moritz_ Blixtor++ indeed
11:36 Blixtor maybe this information about how to get the version/revison could be included on the page I referenced ...
11:37 masak Blixtor: feel free to submit a patch.
11:37 moritz_ bah, I can't remember my rakudo.org password or user name
11:38 Blixtor masak: is the webpage source somewhere in the repository?
11:38 moritz_ no, it's not
11:38 moritz_ I can fix it as soon as I have recovered my password
11:39 iblechbot joined #perl6
11:40 masak Blixtor: oh, sorry, you actually meant "page". I thought you were reading from a file in the Rakudo repository...
11:40 Blixtor masak: no, I meant the page at rakudo.org (that I googled up when looking for information how to submit a bug report)
11:40 Blixtor moritz_: great, thanks!
11:42 mikehh_ joined #perl6
11:43 moritz_ updated.
11:49 Blixtor moritz_++
11:52 Blixtor I'm running the complete (and now unchanged (; ) 'make spectest' now. Is anybody interested in the results? Can I submit these somewhere to help?
11:52 moritz_ just give a short summary here
11:52 moritz_ we don't have a smoke aggregator yet
11:53 Blixtor moritz_: ok, will do
11:58 payload joined #perl6
11:59 KyleHa joined #perl6
12:01 colomon moritz_: You pasted "work around segfaults at exit" earlier.  What was that about?
12:02 moritz_ colomon: about working around segfaults at exit
12:02 moritz_ last week man rakudo spectests segfaulted (or die with bus error) on the newest parrot
12:03 moritz_ s/man/many/
12:03 moritz_ but it seems to be fixed now anyway
12:05 colomon huh.  I never saw that, but that generate-tests.pl script I wrote started segfaulting, apparently at the end of the program, around the time I tried to add the hyperbolic trig tests to mix.
12:05 colomon was wondering if there might be a connection, because darned if I can see what I could be doing that would cause it.
12:05 moritz_ colomon: please update to today's parrot and see if it still segfaults
12:06 colomon Just svn up parrot and rebuild?
12:06 moritz_ svn up parrot, rebuild parrot, reconfigure and recompile rakudo
12:07 colomon Wow, that is a LOT of changes in parrot!
12:07 colomon mostly involving .gitignore, on further inspection.
12:15 takadonet joined #perl6
12:15 takadonet morning all
12:16 patmat left #perl6
12:22 payload joined #perl6
12:22 masak #moose is a friendly channel, almost up there with #perl6.
12:22 araujo joined #perl6
12:26 moritz_ that's good to hear
12:29 moritz_ masak++ # the April list
12:30 * Matt-W waves
12:31 * wayland76 also waves, sending a a fine spume of salt spray into the evening breeze
12:31 wayland76 (actually just going to bed -- night all)
12:31 Matt-W night
12:33 synth joined #perl6
12:34 payload1 joined #perl6
12:35 masak moritz_: oh, you liked that one - nice. I've been toying with the idea of making another list with features instead of bugs.
12:36 masak and not so much from RT as from RAKUDO comments in the various projects out there.
12:36 moritz_ I don't know if it's worth the effort, pmichaud and jnthn will have lots and lots to do to just cover the features listed in the ROADMAP :/
12:37 masak true.
12:37 masak and only 198 days left. o_O
12:38 moritz_ t/spec/S32-str/flip.rakudo ..................................... Failed 10/13 subtests
12:38 masak hm, maybe I should make an "April countdown" web app using Web,pm... :)
12:40 Matt-W :)
12:40 Matt-W yes you should
12:40 Matt-W do it now
12:40 Matt-W Have it render the days remaining using SVG.pm
12:40 masak yummy -- but slightly non-platform-indep
12:41 moritz_ the basics of SVG work in pretty much every browser except in IE and konqueror
12:41 masak hence 'slightly non-platform-indep' :)
12:44 Matt-W pfft
12:44 Matt-W trivialities
12:44 Matt-W :P
12:45 Matt-W do a simple version for the poor souls who have to use IE
12:45 masak I would probably do the simple version regardless.
12:46 masak but throwing something up on feather and redirecting daysleftuntilrakudostar.com to it (or something) would be kinda fun.
12:46 moritz_ or perl6.com :-)
12:46 masak and then fixing S32/Temporal, so that one can show the code without being ashamed...
12:47 moritz_ oh wait, perl6.info would fit better
12:47 masak I had a thought the other day... maybe separating Time, Date and DateTime is already a proliferation of classes. I saw Ruby use only Time for all of those, and I kinda liked that.
12:52 moritz_ anyway, if somebody's looking for a nice Perl 6 task, implementing :by on Ranges would be kinda cool
12:52 moritz_ it's not parsed yet
12:53 moritz_ but you can construct it with Range.new(:from(1), :to(5), :by(0.5))
12:53 masak I heard it'd be difficult before PGE changes.
12:53 moritz_ yes, but the range logic can be done before
12:53 masak oh, definitely.
12:54 moritz_ that's something that requires care (:by only works for numeric ranges; reversion of ranges must be thought of etc.), but it's not too hard all in all
12:54 masak there are numerous lurkers in here. maybe one or two of them will rise to the task. :)
12:54 moritz_ I'd be happy to guide and answer questions, of course
12:54 masak :by only works for numeric ranges? since when?
12:54 alexn_org joined #perl6
12:55 moritz_ always
12:55 masak 'a'..'c' :by(.5) used to be allowed.
12:55 moritz_ and what did it produce?
12:55 colomon ?????
12:55 masak I know we've talked about it, at least.
12:55 masak <a a b b c>
12:55 moritz_ that's kinda scary
12:55 masak yes, but scary good, right? :P
12:56 moritz_ anyway, reading S03 would be required to do the job :-)
12:56 moritz_ (and simplifying it if needed, or only partially implementing it)
12:57 colomon moritz_++: New parrot does indeed seem to have cured my seg fault.  Yay!
12:57 masak hm... did we ever use any of the ideas from http://paste.lisp.org/display/83194 ?
12:57 moritz_ not my doing at all.
12:58 masak moritz_: at least you reported the bug.
13:02 ruoso joined #perl6
13:03 masak wohoo! spectests pass!
13:03 jnthn Sheesh, 100 days and we drop to single digits...
13:04 jnthn At this rate I'm going to feel like I'd better like, write some code or something. :)
13:04 * moritz_ thinks that one can see jnthn's vacation on the commits-to-rakudo graph :-)
13:04 jnthn The commits to blizkost graph on the other hand... ;-)
13:04 * jnthn did some hacks while at the conference
13:05 PZt joined #perl6
13:06 Matt-W jnthn++
13:06 Matt-W I was thinking
13:06 Matt-W How does one go about doing Parrot-level bindings to an external library?
13:07 Matt-W And making it so you can say use Library; in perl 6
13:08 masak Matt-W: you're very welcome to find out and write up a tutorial. I sure'd like to know as well.
13:09 Matt-W heh
13:09 Matt-W I walked into that one didn't I :)
13:10 Matt-W jnthn++ might find out though while he works on making Blizkost work with use Module :lang<perl5> or however it works...
13:10 takadonet OMG 15500 test passed!
13:10 Matt-W yay!!
13:10 jaldhar joined #perl6
13:13 * masak looks at Camelia and smiles warmly
13:13 moritz_ ok, parrot 1.6.0 release is tagged
13:15 jnthn Matt-W: My current patch that I'm hacking on will make use CGI:lang<perl5> work.
13:16 moritz_ wow.
13:16 jnthn Matt-W: Actually the "install something under CGI in the namespace" bitg already works.
13:16 jnthn (locally)
13:16 jnthn Trouble is I've f00ked up calls on packages somehow.
13:16 jnthn It's like I'm not refcounting something properly, but I don't see where yet.
13:17 Matt-W Cool
13:17 Matt-W I'm really impressed with that
13:17 jnthn Once I fix calls on packages, then I've probably done enough "basic" parameter and return value marshalling to make various things usable.
13:17 Matt-W I was kind of thinking of looking at something to process GObject Introspection data and turn it into Parrot language bindings
13:18 Matt-W Have to look at Parrot's NCI though
13:18 moritz_ jnthn: can arrays and hashes be passed around yet?
13:18 jnthn After that it's importing, which I think I know how to do, followed by being able to round-trip any Perl 6 object (well, actually any Parrot object) into Perl 5 land.
13:18 Matt-W That's going to be scary
13:18 jnthn the second of which I don't know how to do, and looks hard.
13:18 jnthn moritz_: Well, not really...
13:19 jnthn moritz_: References to them perhaps can.
13:19 Matt-W that's object roundtripping isn't it
13:19 jnthn moritz_: There's *loads* of bits to fill out.
13:19 jnthn Matt-W: I can round-trip Perl 5 scalars already, afaict.
13:19 jnthn Basically, I want to get the "big pieces" working first.
13:19 jnthn And then it's easy for others to flesh out details, if they wish to.
13:20 jnthn Anyway, my aim for my trip hacking is that by the end of the month, Blizkost is far enough along that you can use Perl 5 DBI will enough to actually do something useful.
13:20 jnthn *well
13:20 moritz_ $jnthn.clone for ^10;
13:20 Matt-W That would be super cool
13:21 Matt-W Yes we need more jnthns
13:21 Matt-W We need a quad-core jnthn
13:22 jnthn IMO, one of me is quite enough for me to put up with. :-)
13:22 dalek rakudo: c22d81b | moritz++ | build/PARROT_REVISION:
13:22 dalek rakudo: bump PARROT_REVISION to the 1.6.0 release
13:22 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c22d81b7e1c2314835d16c0c9895152aa4c2b05a
13:23 Matt-W jnthn: nonsense, you're no trouble at all
13:23 moritz_ well, at least not to us :-)
13:24 JimmyZ joined #perl6
13:25 jnthn Yes, I'm well behaved here. :-)
13:25 jnthn $admiree probably doesn't need another 9 of me admiring her though. ;-)
13:26 Matt-W That's okay
13:26 Matt-W One of you can admire her full time
13:26 Matt-W The other nine of you can write code
13:26 Chillance joined #perl6
13:26 jnthn \o/ finally...I found a version of PuTTY patched up to handle Korean chars properly...
13:27 jnthn plz PuTTY devs incorporate those patches... :-)
13:28 Matt-W I think you might have to tell them somewhere they actually look
13:28 * moritz_ tries to put some pictures into his next blog posts - and fails half of the time :)
13:28 jnthn Matt-W: Aye, probably. :-)
13:28 jnthn Matt-W: I rather suspect though that they're already aware of them.
13:28 Matt-W Probably
13:28 Matt-W I assume the author of the patches has told them
13:29 jnthn And either will incorporate them or have sound reasons for not doing so.
13:29 jnthn PuTTY is a nice bit of software.
13:29 Matt-W it is
13:31 jnthn ooh! My patched PuTTY also makes hyperlinks clickable.
13:31 jnthn \o/
13:32 carlin Could solaris be added to test 4 in S02-magicals so that it doesn't fail (or am I missing the point of the test?)?
13:32 moritz_ carlin: patches welcome
13:34 carlin It's a one word change :) I was just thinking there might be a reason that its not there
13:34 moritz_ carlin: do you have a commit bit for pugs already?
13:35 carlin Yip
13:35 moritz_ no, probably just that nobody thought of it before
13:35 Blixtor carlin: I just noticed as well ;)
13:36 Blixtor my results from 'make spectest': Failed 5/422 test scripts, 98.82% okay. 3/15713 subtests failed, 99.98% okay.
13:37 Blixtor t/spec/S12-attributes/class.rakudo and t/spec/S14-roles/basic.rakudo lead to Parrot stacktraces
13:37 moritz_ these two are known
13:37 moritz_ (and not easily fixable)
13:38 Blixtor the one mentioned by carlin fails for me as well (also 'solaris' missing)
13:39 Blixtor and I have test 171 of t/spec/S03-operators/arith.t failing as well as test 11 of t/spec/S32-num/power.t
13:39 __ash__ joined #perl6
13:41 Blixtor looking more closely, they both seem to test 1**Inf and lead to NaN on my machine ...
13:42 Blixtor so overall no big problems to report, good!
13:44 masak shouldn't 1**Inf lead to NaN?
13:44 moritz_ rakudo: say 1**Inf
13:44 p6eval rakudo a9ff30: OUTPUT«1␤»
13:47 moritz_ masak: I think it makes sense to return 1 here
13:48 pugs_svn r28242 | carlin++ | [S02-magicals/config.t] Add solaris as a known OS
13:51 masak moritz_: hm, I think you're right.
13:51 masak rakudo: 0 ** 0
13:52 p6eval rakudo a9ff30:  ( no output )
13:52 masak rakudo: say 0 ** 0
13:52 p6eval rakudo a9ff30: OUTPUT«1␤»
13:52 moritz_ rakudo: say 1.001 ** Inf
13:52 p6eval rakudo a9ff30: OUTPUT«Inf␤»
13:52 masak 0**0 should yield NaN, though.
13:52 moritz_ I long resisted against 0 ** 0 being defined. I had no success.
13:53 masak there are several good arguments for it being NaN.
13:53 moritz_ I know.
13:53 masak what did people bring as arguments against?
13:54 moritz_ that the IETF floating point standards defines 0 ** 0 to be 1, or so
13:56 masak then they are wrong, too. :/
13:57 Whiteknight joined #perl6
13:58 sevvie joined #perl6
13:59 arnsholt This pages argues (from mathematics) in favour of 0 ** 0 == 1: http://mathforum.org/dr.math/faq/faq.0.to.0.power.html
13:59 arnsholt Their arguments aren't too shabby
13:59 moritz_ http://irclog.perlgeek.de/perl6/2007-02-06#i_-54180 for previous discussion
14:01 moritz_ I wonder why they don't even mention that 0^0 could be 0
14:01 moritz_ (unless I missed it on fast skimming)
14:01 arnsholt They mention it
14:01 moritz_ that would make the function 0^x continuous
14:02 arnsholt Yep. But it seems they prefer x^0
14:03 __ash__ are there any callbacks you can register right now?
14:03 masak __ash__: what do you mean?
14:03 moritz_ on the meta class? don't think so
14:03 masak quietfanatic and I were discussing it on Twitter.
14:03 pnu joined #perl6
14:04 __ash__ like maybe you could register an 'inherited' callback? so when a class inherits from another class it calls that callback, or a trace function?
14:06 masak __ash__: interesting; that was the exact use case quietfanatic had as well.
14:07 masak __ash__: I checked with the Moose people; they don't have it yet, but it shouldn't be so hard to add, they said. http://gist.github.com/187247
14:08 __ash__ kk, i'll look at that, Moose, isn't that perl5's object system that sorta mimics perl6's object system?
14:09 moritz_ sorta, yes. There has been a lot of cross-inspiration
14:09 masak I think Rakudo can borrow quite a lot of the MOP infrastructure from Moose.
14:10 masak when tuits allow, I'll have a look at the Moose tests and see which ones can be translated to the Perl 6 spectest suite.
14:12 molaf joined #perl6
14:15 __ash__ jnthn was saying he was going to try to implement most of the SMOP OO API on perl foundation
14:18 Blixtor left #perl6
14:21 jnthn If you want to hook inheritance independent of meta-class, best is to .wrap the trait handler.
14:22 jnthn Or at least, that's what comes to mind first.
14:22 masak jnthn: do you have a working code example?
14:23 jnthn masak: It's kinda a PITA to do it right now.
14:23 jnthn because
14:23 Matt-W trait handlers are still a bit washy?
14:23 jnthn 1) The nice syntax for selecting a multi-variant isn't implemented.
14:24 jnthn (meaning getting hold of the multi you wish to wrap is tricky)
14:24 jnthn and 2) even though you can do it, you're currently going via a non-spec interface that is likely to change/break in the next month or two.
14:25 Matt-W Is there any spec at all on selecting a particular multi variant yet?
14:25 jnthn Or put another way, you can probably do this today if you really want to, but if you're going to then know you're using stuff that is going to move under you.
14:25 jnthn Matt-W: Yeah, IIRC it's like &foo:(Int, Int) or similar.
14:26 jnthn std: multi foo { }; &foo:(Int, Int)
14:26 p6eval std 28242: OUTPUT«ok 00:04 58m␤»
14:26 jnthn aye, I think that.
14:26 jnthn nyi though
14:26 jnthn (though I guess not that hard)
14:26 Matt-W Cool
14:26 Matt-W like the syntax
14:26 Matt-W It makes good sense
14:26 jnthn The way you can do it now is by introspecting the signature.
14:27 jnthn But what Rakudo currently provides is a stop-gap.
14:27 Matt-W mmm
14:27 Matt-W ouch
14:27 jnthn Signatures are going to get revised drastically in October (provided I survive Asia ;-))
14:28 jnthn One thing I want to do under my grant (plz comment ;-)) is get the signature introspection API into the synopses.
14:28 * masak sincerely hopes jnthn survives Asia
14:28 masak grant URL?
14:28 jnthn masak: So far I'm doing a lot better than surviving. ;-)
14:28 jnthn masak: http://news.perlfoundation.org/2009/09/hague_grant_application_rakudo.html
14:29 jnthn Matt-W: I'm not sure off hand how the syntax is parsed.
14:29 jnthn Matt-W: I think it's probably straightforward to implement though.
14:31 hanekomu jnthn: are you still in Tokyo?
14:32 literal rakudo: sub foo ($a, $b = $a + 1, $c) { say $b }; foo(2);
14:32 p6eval rakudo a9ff30: OUTPUT«too few arguments passed (1) - 4 params expected␤in sub foo (/tmp/UxJlGr42W7:1)␤called from Main (/tmp/UxJlGr42W7:2)␤»
14:32 literal rakudo: sub foo ($a, $b? = $a + 1, $c?) { say $b }; foo(2);
14:32 p6eval rakudo a9ff30: OUTPUT«3␤»
14:32 jnthn hanekomu: no
14:33 jnthn hanekomu: I'm in its anagram. ;-)
14:33 hanekomu ah, nice
14:33 hanekomu (wrongly romanized anagram, that is)
14:33 jnthn Yes, it is. :-)
14:33 jnthn Heh.
14:33 hanekomu Toukyou != anagram(Kyouto)
14:33 jnthn Romanization is an inprecise art. :-)
14:33 hanekomu yes
14:33 hanekomu Kyoto station itself is impressive
14:33 Matt-W Friend of mine lives in Nagasaki
14:34 Matt-W Are you going there at all?
14:34 hanekomu also, if you have time, visit the Fushimi Inari shrine - thousands of Torii, very nice
14:34 jnthn Matt-W: Planning to day-trip Nagasaki from Hakata.
14:34 Matt-W Cool
14:34 Matt-W he says it's nice there
14:34 jnthn hanekomu: That's the one on the outskirts of Kyoto with loads of red pillars?
14:34 hanekomu yes
14:34 Matt-W he's not a Perl guy though, so won't be trailing you with a parasol and some cool drinks
14:34 hanekomu a bit to the south of the station
14:35 jnthn hanekomu: Aye, I'd got that on my "really want to see" list.
14:35 jnthn Kyoto has an insane amount of stuff to see.
14:35 hanekomu yes
14:36 hanekomu Nijo castle is also fantastic. And Nishiki market. And... :)
14:36 nihiliad joined #perl6
14:36 jnthn I'd planned to day-trip to Osaka from here, but...well...I'll probably make it an afternoon/evening trip so I can have more time in Kyoto. :-)
14:36 * hanekomu misses it
14:36 jnthn hanekomu: You lived here once, or just visited?
14:37 hanekomu no, just visited for a week this May (with my mother)
14:37 hanekomu but I plan to return to Kyoto, and Tokyo and other places in Japan
14:37 jnthn Yeah
14:37 jnthn It's my first time in Japan.
14:37 jnthn but I'm rather sure it won't be my last. :-)
14:37 hanekomu And all the nice matcha-based sweets they have there.
14:39 jnthn Mm. And tempura. :-)
14:39 jnthn I don't eat fish and seafood, but there's still a lot of nice food to be had in Japan even so.
14:40 jnthn the quality is consistently between good and excellent.
14:40 hanekomu yes.
14:40 jnthn And actually does often go with price curve.
14:41 jnthn That is, I've never paid a lot for something that wasn't reallly good.
14:41 hanekomu Hm, Oyakodon
14:42 hanekomu Ah, I remember - there's a restaurant called Katsukura on the 11th floor of Isetan (The Cube, Kyoto station) - lots of meat, and pickles, and other side dishes.
14:42 hanekomu http://www.bento.com/kansai/rev/7027.html
14:43 jnthn tonkatsu!! :-D
14:44 hanekomu better description, and better pics here: http://paulstravelpics.blogspot.com/2008/09/kyoto-restaurant-reviews.html - scroll down a bit
14:46 jnthn hanekomu: Ah, this is one of a chain?
14:46 hanekomu you get to grind sesame with a mortar, then mix with tasty sauces.
14:46 hanekomu yes
14:49 jnthn hanekomu: Yeah, the Pauleys took me to one of the chain in Tokyo.
14:49 hanekomu ah
14:49 jnthn It *is* very, very good.
14:50 Psyche^ joined #perl6
14:50 SmokeMachine joined #perl6
14:50 donaldh joined #perl6
14:53 justatheory joined #perl6
14:57 pmurias joined #perl6
15:02 pmurias ruoso: hi
15:02 ruoso hi pmurias
15:02 phenny ruoso: 11 Sep 18:36Z <TimToady> tell ruoso please read the Guy Steele slides mentioned above and think about hyper iterators in terms of balanced (more or less) list splitting
15:02 pmurias ruoso: what should i submit as the gsoc code sample?
15:03 ruoso pmurias, I think the multi tests are very much illustrative
15:03 masak ruoso: that's http://research.sun.com/projects/plrg/Publications/ICFPAugust2009Steele.pdf if you're curious.
15:04 ruoso yeah... I was looking at it the other day...
15:04 ruoso still trying to parse much of it
15:06 carlin darnit, parrot failed to build on Haiku
15:07 __ash__ jnthn: in src/builtin/guts.pir is that where you would need to pun a role if your doing an explicit role method call? (the $var.Role::method(); issue)
15:08 Matt-W carlin: awwwww :(
15:08 __ash__ rakudo: role A { method foo { 'a' } };  A.new.A::foo;
15:08 p6eval rakudo a9ff30: OUTPUT«Method '!select' not found for invocant of class ''␤»
15:09 carlin Probably because of how out-of-date the gcc is
15:09 * carlin tries to upgrade it
15:10 ruoso I'm mostly trying to understand what does it mean to have two slurpies in a signature...
15:14 sharada joined #perl6
15:14 pmurias joined #perl6
15:14 jnthn __ash__: . new has already punned
15:14 jnthn Ah, I see the isue, hmm...
15:15 __ash__ rakudo: class A { method foo { 'a' } };  A.new.A::foo;
15:15 jnthn ruoso: I thought you could only have one slurpy array and one slurpy hash (and slurpy scalars are just unpackers in disguise)
15:15 p6eval rakudo a9ff30:  ( no output )
15:15 Eevee joined #perl6
15:16 jnthn __ash__: It probably needs some fix in the punner.
15:16 jnthn IIRC.
15:16 __ash__ k, i'll look at that then
15:16 jnthn I did know what needed to be done to fix it at one point, but I forgetted it.
15:17 __ash__ no worries, if i figure it out i'll let you know
15:20 cono joined #perl6
15:20 barney joined #perl6
15:21 Exodist joined #perl6
15:25 abra joined #perl6
15:27 pmichaud good morning, #perl6
15:27 phenny pmichaud: 03:18Z <s1n> tell pmichaud we need to send out this months' hackathon notice, ideas for locations? saxby's in allen or mckinney is free and usually pretty quiet (my recommendation is saxby's in allen). see http://www.wififreespot.com/tex.html
15:27 pmichaud phenny:  tell s1n saxby's in allen sounds good to me
15:27 phenny pmichaud: I'll pass that on when s1n is around.
15:27 cdarroch joined #perl6
15:28 KyleHa joined #perl6
15:28 jnthn pmichaud!! :-)
15:28 * jnthn waves from Kyoto
15:29 pmichaud oh, sorry, I was waving in the wrong direction I think :-P
15:29 pmichaud how's Kyoto/Japan/Far East?
15:31 jnthn It's an awesome place.
15:31 jnthn Kyoto has far too much to do. :-)
15:33 payload joined #perl6
15:35 __ash__ so, jnthn, in the future the right way to add an inheritence hook is to setup a wrap on a trait?
15:36 jnthn __ash__: If you want to do so independent of meta-class, that's probably the place...
15:36 jnthn __ash__: Otherwise, wrap add_parent on the metaclass you care about;.
15:36 masak oh hai, pmichaud!
15:38 __ash__ so there wouldn't be a generic object hook you could setup to do that? no function like object.HOW.inherited() is called when a subclass inherits from something?
15:40 jnthn I don't think there's anything like that in the default meta-model.
15:40 jnthn But it'd probably not be a hard module to write.
15:40 jnthn (Once we have other bits in place, I guess...)
15:43 __ash__ I am working on a role based persistence mechanism (that kinda mimi's DataMapper from ruby) and that would be really useful for STI, although for now i am dumping things to a file, since I can't connect to a DB at the moment, but one day, it should be able to do STI
15:44 colomon joined #perl6
15:44 __ash__ right now i am still in the planning/what can i do/whats the right way to do it phase
15:46 __ash__ s/that kinda mimi's/that kinda mimics
15:46 jnthn STD...STI...great, now we covered both acronyms for it... :-)
15:47 __ash__ single table inheritance :-P
15:47 jnthn oh, phew
15:47 * jnthn reminds himself that STD = short for Standard
15:48 stephenlb joined #perl6
15:49 masak jnthn: Perl 6 is contagious!
15:49 masak jnthn: alternatively, "Perl 6: spread the love!"
15:50 jnthn masak: all(<eww lol!>)
15:51 masak :)
15:59 * jnthn -> sleep
16:00 asiekierka joined #perl6
16:00 asiekierka left #perl6
16:03 justatheory joined #perl6
16:03 colomon joined #perl6
16:04 donaldh left #perl6
16:07 nihiliad joined #perl6
16:09 ms11 joined #perl6
16:17 ruoso jnthn, http://irclog.perlgeek.de/perl6/2009-09-11#i_1492820
16:17 pmurias diakopter: hi
16:17 diakopter pmurias: howdy
16:19 pmurias still alive ;) had to wake up at 4 hour at night to retrieve a laptop i lost during a train trip
16:19 diakopter you retrieved it, sounds like?
16:19 * ruoso probably suffer a bit more parsing that slides because he's not fluent in LISP
16:21 pmurias diakopter: yes, fortunatly
16:22 pmurias diakopter: how did your viv stuff do?
16:23 diakopter pmurias: I got vivjs to work... and the V8 context embedded in perl was able to handle/parse the quasi-JSON edition of STD.pm, but I tried pretty-printing it with json2.js' JSON.stringify() and it exploded into wanting a few more GB of memory than available (b/c of all the duplicated references)...
16:23 ruoso TimToady, I *think* the non-flattening behavior of Perl 6 is already very friendly to parallelism
16:23 diakopter but at least V8 was able to parse it and instantiate it
16:24 ruoso TimToady, in a way that multi foo(*@a, *@b, *@c) would expect a multidimensional capture where the first dimension has 3 elements
16:24 pmurias diakopter: do you plan to write the compiler in js?
16:25 diakopter interpreter, yeah
16:25 pmichaud that would be incredible
16:26 diakopter STD.pm in JSON is 11MB :)
16:26 abra joined #perl6
16:27 * diakopter thanks to the (anonymous?) person who wrote Acme::JavaScript::V8 last year...
16:27 ruoso TimToady, theoretically, finding the elements in the "first dimension" is the way we ask the list to "split itself" in a hyper iterator
16:28 colomon rakudo: say 31415926535 / 10000000000;
16:28 p6eval rakudo a9ff30: OUTPUT«say requires an argument at line 2, near " \x{2}31415926"␤in Main (src/gen_setting.pm:2565)␤»
16:28 colomon rakudo: say 31415926535 / 10000000000;
16:28 p6eval rakudo a9ff30: OUTPUT«1␤»
16:28 colomon That strikes me as wrong....
16:28 __ash__ thats not 1.... isn't that PI?
16:28 pmurias diakopter: you plan to interpret perl6 with an js interpreter rather than compiling it to js?
16:29 __ash__ or part of PI
16:29 colomon __ash__: yes, I'm trying to convert to radians in rational numbers.
16:29 pmichaud rakudo: say 31415926535 / 10000000000.0
16:29 p6eval rakudo a9ff30: OUTPUT«3.1415926535␤»
16:29 __ash__ oh, those silly integers (and integer division)
16:29 pmichaud rakudo: say 31415926535.PARROT
16:29 diakopter pmurias: well, the ast is "compiled" to a JS "expression tree" object...
16:29 p6eval rakudo a9ff30: OUTPUT«Int␤»
16:29 pmichaud rakudo: say 31415926535.WHAT
16:29 p6eval rakudo a9ff30: OUTPUT«Int()␤»
16:29 colomon Oooh, it's an overflow
16:30 pmichaud rakudo:  say 314 / 100;
16:30 p6eval rakudo a9ff30: OUTPUT«3.14␤»
16:30 colomon I counted the wrong number of digits.
16:30 colomon rakudo: say 314159265 / 100000000
16:30 p6eval rakudo a9ff30: OUTPUT«3.14159265␤»
16:30 pmichaud still shouldn't give "1", likely.
16:30 colomon pmichaud: probably the overflow hoses the GCD calculation routine.
16:30 pmichaud rakudo: say 31415926535 / 1000000
16:30 p6eval rakudo a9ff30: OUTPUT«-2147.483645␤»
16:31 diakopter pmurias: but the (stackless/continuation-returning-style) interpreter model I'm using allows the emitter to swap in "compiled" JS too... so it can all start out interpreted, and then as optimization phases are added, more can be "compiled"
16:31 pmichaud I can't figure out why it thinks that 31415926535 is an int.
16:31 colomon Are you sure it does?
16:31 pmichaud rakudo: say 31415926535.PARROT
16:31 p6eval rakudo a9ff30: OUTPUT«Int␤»
16:31 colomon rakudo: say (1415926535 / 1000000).WHAT
16:31 p6eval rakudo a9ff30: OUTPUT«Rat()␤»
16:32 colomon whoops
16:32 pmichaud rakudo: say (31415926535+1).PARROT
16:32 colomon rakudo: say (31415926535 / 1000000).WHAT
16:32 p6eval rakudo a9ff30: OUTPUT«Num␤»
16:32 p6eval rakudo a9ff30: OUTPUT«Rat()␤»
16:32 pmichaud rakudo: say (31415926535-1).PARROT
16:32 p6eval rakudo a9ff30: OUTPUT«Num␤»
16:32 pmichaud looks like Rakudo has trouble with integer constants that fall outside of the range of 32-bit ints
16:32 colomon that number is too big for a 32-bit int, right?
16:33 pmichaud yes
16:33 pmichaud in most operations Rakudo auto-promotes it to a float
16:33 pmichaud rakudo:  say 31415926535
16:33 p6eval rakudo a9ff30: OUTPUT«31415926535␤»
16:33 pmichaud that's.... weird
16:33 diakopter pmurias: since the jseval() function exposed in Perl-land bypasses V8's JS eval() and sends the JS source straight to the v8 engine, jseval() is just as slow(/fast) as any other top-level JS code in V8
16:34 ruoso diakopter, one thing you might want to evaluate is how far V8 can implement Perl 6 semantics, and what are you going to do about unsupported semantics...
16:35 __ash__ hmmm, on my computer perl6   say 31415926535 / 10000000000; printed 3.1415926535
16:35 __ash__ is that because i have mine compiled for 64 bit?
16:35 pmichaud __ash__: probably.
16:35 diakopter ruoso: :) hey, it's an interpreter! any set of semantics can be emulated! ;)  srsly though, what specifically do you have in mind
16:35 __ash__ so parrot defaults int's capacity to the size its compiled for? (32 or 64)
16:36 ruoso diakopter, the thing is that in order to emulate Perl 6 semantics, you're probably going to need to implement a runtime library on top of V8, which might prove way slower than you expect (i.e. KindaPerl6)
16:36 ruoso diakopter, otoh, you might simply choose to implement a subset of Perl 6
16:36 ruoso "whatever V8 supports"-Perl 6
16:37 pmichaud I've just added masak++'s mmorpg article and KyleHa++'s "how I got started" articles to http://rakudo.org/how-to-help
16:38 diakopter ruoso: definitely, of course.  But that kind of delay won't be noticable anyway compared to the 2-4 seconds STD.pmc takes to startup/parse input :P
16:38 ruoso that considering you need to run STD every time
16:39 ruoso but if you're able to compile to some sort of bytecode, compile-time might be irrelevant
16:40 pmurias and you can work around the STD startup time by using app-persistent
16:40 ruoso diakopter, note that a Perl 6 subset that can be run in V8 is still way cool
16:41 pmurias diakopter: you are targeting V8 only?
16:41 diakopter pmurias: yes, b/c it's linked into perl with xs
16:42 pmurias so you don't plan for the code to work in browsers?
16:42 diakopter so STD can call out to it when it hits the end of a BEGIN{} and then resume
16:42 diakopter well sure, if gimme5 was ported to a gimmejs
16:43 pmurias hmm, if we rule out eval why do we need a parser at runtime?
16:43 diakopter b/c it's mutable if STD is still running...
16:44 sharada joined #perl6
16:45 pmurias once the program parser you could freeze everything and load it from json/js?
16:46 pmurias a perl 6 on js that *doesn't* run on browsers doesn't seem very usefull
16:47 pmurias there is nothing wrong with developing using a standlone V8 at first if it's possible to browserize"
16:47 pmurias it later
16:48 diakopter yep; all it would need is a gimmejs... which would eliminate the Perl dependency
16:49 moritz_ once the compiler works it's a gimmejs ;-)
16:49 diakopter likewise with rakudo using STD.pm; all it needs is a gimmenqp or gimmepir or gimme_Perl6_rakudo_r388382whatever
16:51 diakopter moritz_: yeah :)
16:55 pmurias or cat if once rakudo matures enough
16:55 pmurias joined #perl6
16:56 moritz_ we're working on it ;-)
16:58 colomon joined #perl6
17:01 jaldhar joined #perl6
17:01 pugs_svn r28243 | colomon++ | [t/spec] Update to trig test generator -- still ugly as sin, but now correctly generates a full suite of real number tests on all the forward trig functions.
17:02 colomon It's currently generating 3718 passing tests on my branch of Rakudo.
17:03 colomon Still need to do a ton of work for Complex numbers and all the inverse trig functions.
17:06 smash_ joined #perl6
17:06 smash_ hello everyone
17:07 colomon hello
17:15 payload1 joined #perl6
17:22 garron joined #perl6
17:23 pmichaud http://pmichaud.com/perl6/rakudo-tests-2009-09-15.png   # that's a really nice looking graph.  colomon++  moritz++  ash++  (let me know if I need to ++ anyone else)
17:24 desertm4x joined #perl6
17:24 * moritz_ was barely involved </gandalf>
17:27 dalek rakudo: 1c35342 | pmichaud++ | docs/spectest-progress.csv:
17:27 dalek rakudo: spectest-progress.csv update: 436 files, 15500 (71.5% of 21679) pass, 0 fail
17:27 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1c353426c2359d0005653b2221d9d789e2eb8b50
17:30 sharada joined #perl6
17:36 synth joined #perl6
17:45 jauaor joined #perl6
17:54 dalek rakudo: 42ff49e | pmichaud++ | docs/announce/2009-09:
17:54 dalek rakudo: Initial draft of 2009-09 release announcement.
17:54 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/42ff49e7d69d0bd93002eb37bc6c85f8566e462c
17:54 dalek rakudo: 48b5d4d | pmichaud++ | docs/ChangeLog:
17:54 dalek rakudo: Some ChangeLog updates.
17:54 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/48b5d4d63e02a51df1e8d724df6a90b4edb9ef62
17:56 M_o_C joined #perl6
18:01 sevvie joined #perl6
18:02 colomon joined #perl6
18:08 diakopter std: five(say(3),say(4,sub five{}))
18:08 p6eval std 28243: OUTPUT«ok 00:02 38m␤»
18:11 smash_ rakudo: say "1"
18:11 p6eval rakudo a9ff30: OUTPUT«1␤»
18:12 smash_ rakudo: say (1 if 0).perl
18:12 p6eval rakudo a9ff30: OUTPUT«[]␤»
18:12 blaze-x_ joined #perl6
18:13 mikehh__ joined #perl6
18:13 blaze-x is there a way to debug perl6 grammars? I'd just like to see what is going on, how my teststring is parsed
18:14 moritz_ when it's parsed successfully, you you inspect $/.perl
18:14 moritz_ or $/.chunks.perl
18:17 blaze-x and when it parses unsucessfully? Anything I can inspect then?
18:17 pmichaud I still hope to put a debugging mode into the regex engine
18:18 moritz_ blaze-x: Regexes current support PIR code blocks
18:18 moritz_ in the form {{{{ PIR HERE }}}}
18:18 moritz_ you can put print statements inside those
18:19 blaze-x I see, but that might just be a bit too much, I'm just learning perl6, and don't know any PIR. I'll try to build up from a small example.
18:20 blaze-x Maybe there is a way to put print statements inside grammars for debugging purposes?
18:20 moritz_ rakudo: 'foo' ~~ m/ f {{{{ print "1" }}}} o/
18:20 p6eval rakudo a9ff30: OUTPUT«1»
18:20 moritz_ rakudo: 'foo' ~~ m/ f {{{{ print "1" }}}} oa/
18:20 p6eval rakudo a9ff30: OUTPUT«1»
18:21 moritz_ you see that it gets as least as far as the 'f'
18:21 moritz_ if you only want to print a constant, PIR is no rocket science
18:21 blaze-x I see, interesting... having a try thanks.
18:22 smash_ blaze-x: you can use $ ./perl6 --target=parse script.p6 to see the parsing tree
18:23 moritz_ smash_: that's no help for self-written regexes
18:23 smash_ hmm, i must have misread his question then
18:24 diakopter TimToady: bug in  viv --ast : children of the Comma node are not emitted
18:29 rfordinal joined #perl6
18:31 diakopter yep; 13:14 < blaze-x> and when it parses unsucessfully? Anything I can inspect then?
18:31 diakopter oops
18:31 diakopter yep; STD::Comma is not transformed
18:37 Util joined #perl6
18:46 smash_ RT#67866 can be resolved
18:48 moritz_ rakudo: say any("foo","bar").WHAT
18:49 p6eval rakudo a9ff30: OUTPUT«Junction()␤»
18:49 moritz_ when we have tests, yes
18:49 colomon rakudo: say asin(sin(2.35619449019234));
18:49 p6eval rakudo a9ff30: OUTPUT«0.785398163397453␤»
18:50 moritz_ I've updated the ticket accordingly, smash_++
18:50 colomon pugs: say asin(sin(2.35619449019234));
18:50 p6eval pugs: OUTPUT«*** No such subroutine: "&asin"␤    at /tmp/gqMC42ejDc line 1, column 5-32␤»
18:50 moritz_ rakudo: say 2.35619449019234 - pi
18:50 p6eval rakudo a9ff30: OUTPUT«-0.785398163397453␤»
18:50 colomon oh, I see.
18:51 colomon it doesn't have to be a full pi away, it can be pi/2 in the right circumstances
18:51 pmichaud say sin(0.785); say sin(2.356);
18:51 pmichaud rakudo: say sin(0.785); say sin(2.356);
18:51 p6eval rakudo a9ff30: OUTPUT«0.706825181105366␤0.707244293145898␤»
18:51 moritz_ iirc returned angles are normalized to the range from -pi to pi
18:51 colomon okay, time to rewrite my first stab at the asin test.  :)
18:52 colomon moritz_: think it may actually be -pi/2 to pi/2.
18:52 moritz_ that's not enough for every function that returns an angle
18:54 colomon it's enough for asin and acos, it's not enough for atan2.
18:57 SmokeMachine joined #perl6
19:01 smash_ moritz_: what would be the most appropriate file in your opinion to add the tests for RT#67866 ?
19:03 moritz_ smash_: t/spec/S03-junctions/misc.t
19:04 moritz_ something like any(1, 2).WHAT  === Junction or so
19:04 * smash_ nods.
19:05 moritz_ (btw when I take a ticket with the comment that it needs tests, others are very welcome to provide those; I just act as a placeholder)
19:05 zloyrusskiy joined #perl6
19:16 diakopter std: say(:3,8,9)
19:16 p6eval std 28243:  ( no output )
19:16 diakopter (crash, I think)
19:17 [particle] or rebuild
19:17 moritz_ timeout on the server, maybe
19:17 moritz_ [particle]: rebuild shouldn't affect the running copy, except for slowness
19:18 diakopter what's the timeout limit
19:18 moritz_ 8s
19:19 moritz_ sh: line 1:  2657 CPU time limit exceeded /usr/bin/perl tryfile /tmp/4tiOdoFiXk >> /tmp/L4Ki81iK0U 2>&1
19:19 diakopter hrm. well, it infinitely loops locally
19:19 frederico joined #perl6
19:19 moritz_ and infinity > 8s ;-)
19:20 pmichaud for sufficiently large values of infinity :-)
19:20 diakopter TimToady: bug in STD.pm/gimme5:  std: say(:3,8,9) # loops whilever
19:20 moritz_ so far I found infinity always large enough ;-)
19:21 moritz_ [particle]: I read in the #ps logs that you want to release Rakudo next week - was that just a thinko/typo?
19:21 diakopter memory usage doesn't increase and recursion warning doesn't appear
19:21 [particle] moritz_: thinko
19:22 [particle] i'm releasing this week
19:22 moritz_ great
19:22 [particle] i'll do a practice run tonight
19:22 moritz_ great
19:22 [particle] ~8h
19:22 pmichaud [particle]: are you still expecting to do a "live release"?
19:22 pmichaud er, "live release demo"?
19:22 [particle] looks like my perl meeting is cancelled :(
19:22 pmichaud :(
19:22 pmichaud awww, howcome?
19:23 [particle] ENOGROUPLEADERSHIP
19:23 pmichaud eh, just do it as a video podcast or something, then!
19:27 diakopter std: say(7,8,9)
19:27 p6eval std 28243: OUTPUT«ok 00:02 38m␤»
19:27 sharada joined #perl6
19:27 diakopter moritz_: have you tried putting STD.pm in mod_perl
19:28 diakopter STD.pmc I mean
19:28 moritz_ diakopter: no
19:28 diakopter I wonder if its response time via p6eval would improve
19:28 diakopter a web service of sorts.
19:29 moritz_ IMHO not really worth the trouble
19:29 fredrecsky joined #perl6
19:31 moritz_ pmichaud, [particle]: the current release announcement draft lists old deprecation notices...
19:31 moritz_ what's your preference? write them as "have been removed", or delete them altogether?
19:32 diakopter TimToady: also, bug in STD.pm/gimme5/viv -  std: say(7,8,9) # term missing args member
19:35 mikehh_ joined #perl6
19:35 iblechbot joined #perl6
19:35 frederico_br joined #perl6
19:40 diakopter TimToady: scratch that last
19:42 __ash__ joined #perl6
19:43 cmv joined #perl6
19:43 cmv joined #perl6
19:54 frederico joined #perl6
19:55 pugs_svn r28244 | fglock++ | [mp6] Javascript emitter, runtime tweaks
20:03 smash_ moritz_: just attached a simple patch in RT#67866 that adds tests, if those aren't enough just reply to the message
20:04 moritz_ smash_: you can commit them right away... have you got a commit bit for pugs yet? (that's where the tests live)
20:04 smash_ moritz_: i know, but i do not have a commit bit for pugs repo
20:05 fredrecsky joined #perl6
20:05 moritz_ I can change that if you tell (or /msg) me your email address
20:09 justatheory joined #perl6
20:13 pugs_svn r28245 | smash++ | * add tests for RT#67866
20:13 frederico_br joined #perl6
20:15 smash_ moritz_: ok, added tests to pugs repo
20:16 moritz_ smash_: great
20:19 meppl joined #perl6
20:19 moritz_ smash_: I've closed the ticket now, thank you
20:19 smash_ moritz_: great.. one down, many to go
20:24 Confield joined #perl6
20:26 frederico joined #perl6
20:26 zamolxes joined #perl6
20:32 kidd_ joined #perl6
20:34 fredrecsky joined #perl6
20:38 mikehh rakudo (48b5d5d) builds on parrot r41279 - make test / make spectest (up to 28245) PASS - Ubuntu 9.04 amd64
20:38 mikehh rakudo - t/spec/S03-operators/arith.rakudo - TODO passed:   120, 131-132
20:45 frederico_br joined #perl6
20:45 rfordinal joined #perl6
20:46 rfordinal3643 joined #perl6
20:47 pmichaud I think I want to mark those arith.t tests as SKIP instead of TODO
20:47 pmichaud they're false passes
20:47 moritz_ I think we should move them to overflow.t
20:47 pmichaud +2
20:47 pmichaud afk, grocery shopping
20:48 moritz_ I have a local patch that removes the int() function. When I'm done spectesting and I'm still sufficiently awake I'm happy to do that
20:48 rfordinal joined #perl6
20:48 sri_kraih_ joined #perl6
20:49 moritz_ rakudo: say log(4063328477) / log(2)
20:49 p6eval rakudo a9ff30: OUTPUT«31.9200148498582␤»
20:49 mberends :-) proto's installed-modules branch is starting to use ~/.perl6lib
20:49 moritz_ mberends++
20:55 pmurias joined #perl6
20:59 mberends moritz_: it's not ready for general use yet. One lesson is that manual testing is soooo tedious. There is no test suite for proto. (I wonder if that's easy to create).
21:00 moritz_ mberends: I've learned that lesson for SVG::Plot
21:00 moritz_ and it's really not easy to write meaningful tests for that, too
21:01 mberends I read your remarks earlier
21:02 moritz_ (and yes, testing stuff that works on the file system is always ugly)
21:03 mberends indeed, and because proto does bootstrapping, it's ugly inside too
21:04 * moritz_ wonder how many people are still using the int() function, despite early deprecation notice
21:04 pmurias moritz_: whats the replacement?
21:04 * mberends was, in Temporal module and tests :/
21:04 moritz_ pmurias: .Int
21:05 pmurias rakudo: say 1.79.In
21:05 pmurias rakudo: say 1.79.Int
21:05 p6eval rakudo a9ff30: OUTPUT«Method 'In' not found for invocant of class 'Num'␤»
21:05 p6eval rakudo a9ff30: OUTPUT«1␤»
21:05 Whiteknight joined #perl6
21:05 mberends rakudo: say 3.6.floor
21:05 p6eval rakudo a9ff30: OUTPUT«3␤»
21:17 Util rakudo: eval q[ class B0rk { say $.a; }; ]; eval q[ class Chef { say $.b; }; ]; say "Done";
21:17 p6eval rakudo a9ff30: OUTPUT«Done␤»
21:18 Util eval q[ class B0rk { say $.a; }; ]; eval q[ class Chef { say $.b; }; ]; say $!;
21:18 pmurias moritz_: a sub int {die 'We are terribly sorry but you should use the .Int method instead of the int function'} would cure that swiftly
21:18 Util rakudo: eval q[ class B0rk { say $.a; }; ]; eval q[ class Chef { say $.b; }; ]; say $!;
21:18 p6eval rakudo a9ff30: OUTPUT«Lexical 'self' not found␤»
21:21 mikehh BTW if I run t/spec/S03-operators/arith.t on amd64 it fails (not ok) 6 tests - 130, 133, 191, 192, 193 and 197
21:22 Util When I am logged in as "util", RT displays links to allow me to change existing tickets.
21:22 Util I need to open a new ticket, but I see no link to allow that.
21:22 Util Is some other privledge needed for my RT ID, or am I just missing something obvious?
21:23 moritz_ Util: it's only possible by mail (rakudobug@perl.org)
21:23 Util moritz_, OK, thanks
21:31 mikehh t/spec/S03-operators/arith.t passes SKIP tests 121, 123, 198
21:35 moritz_ I'm moving some of those tests to overflow.t (which is not run by default)
21:36 mikehh_ joined #perl6
21:52 pugs_svn r28246 | moritz++ | Kill off prefix:<int> (as per r25890)
21:52 pugs_svn r28246 |
21:52 pugs_svn r28246 | Also remove a huge chunk of commented-out Perl 5 code form pack.t.
21:52 pugs_svn r28246 | It confused me greatly twice already - if anybody wants it back,
21:52 pugs_svn r28246 | it can be found in the perl 5 git repo.
21:54 dalek rakudo: 27dcc8d | moritz++ | docs/announce/2009-09:
21:54 dalek rakudo: mention context vars in release announcement
21:54 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/27dcc8d2875904494529cc752b9e5fdf62eaf1fd
21:54 dalek rakudo: c3d93dc | moritz++ | docs/announce/2009-09:
21:54 dalek rakudo: mark deprecation notices as being a thing of the past
21:54 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c3d93dc6b63de8650090fc1b2b6dfe6d6c284319
21:54 dalek rakudo: 9a61441 | moritz++ | src/builtins/any-num.pir:
21:54 dalek rakudo: sub int() is gone (as announced in the 2009-08 release), give a helpful error message instead
21:54 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/9a61441ec3281f0306f88f67d47eeee5a3ae35f4
21:54 pugs_svn r28247 | moritz++ | [t/spec] move arithmetic tests with large numbers to overflow.t
21:54 * moritz_ -> bed
22:10 estrabd joined #perl6
22:11 Helios- joined #perl6
22:15 snarkyboojum joined #perl6
22:35 tak11 joined #perl6
22:37 mikehh joined #perl6
22:41 snarkyboojum joined #perl6
22:43 rbaumer joined #perl6
22:49 rbaumer joined #perl6
22:51 synth joined #perl6
22:51 hercynium joined #perl6
22:53 rbaumer joined #perl6
23:09 KatrinaTheLamia joined #perl6
23:14 blaze-x joined #perl6
23:14 lisppaste3 joined #perl6
23:23 synth^2 joined #perl6
23:24 sm^^^ joined #perl6
23:47 diakopter whee
23:49 bgs100 joined #perl6
23:51 frikipedista8928 joined #perl6
23:51 frikipedista8928 hola auxilio
23:54 frikipedista8928 hola
23:54 frikipedista8928 hola
23:54 frikipedista8928 hello
23:54 frikipedista8928 s.o.s.
23:55 frikipedista8928 hi
23:56 frikipedista8928 left #perl6
23:56 diakopter ok

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo