Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2005-05-11

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:06 cdpruden has joined #perl6
00:13 larsen has quit IRC ("later")
00:14 mugwump stevan, whereabouts do you live, out of interest?
00:16 autrijus class Point {
00:16 autrijus    has $.x;
00:16 autrijus    has $.y;
00:16 autrijus    method show () { say "Coordinate: ($.x, $.y)" }
00:17 autrijus    method set ($x, $y) { $.x = $x; $.y = $y }
00:17 autrijus }
00:17 autrijus my $pt = Point.new(:x(10), :y(20));
00:17 autrijus $pt.show;
00:17 autrijus $pt.set(30, 40);
00:17 autrijus $pt.show;
00:17 autrijus This now works.
00:17 autrijus committing.
00:17 mugwump autrijus++ # you da man
00:18 mugwump nothing else, though, I suppose?  Roles, isa, does, etc?
00:18 autrijus nothing else.
00:18 autrijus gotta get started somewhere :)
00:19 mugwump yes, indeed.  Well, I guess I can do away with Hack::Instances soon, then
00:19 mugwump Can I put defined Classes in method signatures?
00:19 autrijus yes.
00:19 autrijus and they will just work.
00:19 autrijus (envClasses gets modified)
00:20 mugwump FCKN@!   (this is a positive exclamation in case you were wondering)
00:22 mugwump Gah, I wish I didn't have a major release due this week... suppose I'd better start on it :)
00:29 mugwump hmm, I can't get to my freepan repos, nor ssh in
00:29 mugwump oh, duh.  it's not svn.freepan.org
00:30 meppl gute nacht
00:31 mugwump tschuss
00:32 * puetzk wonders if the "method set($.x, $.y)" works to make setters, or if that's only for constructors
00:35 madhouse has quit IRC ("yeah")
00:35 sproingie pugs now has real class i see :)
00:35 autrijus yes. have fun. :)
00:35 autrijus zzz &
00:36 * autrijus is quite happy that both rules and oo has landed on the 100th day :)
00:36 sproingie i've never seen anything so complex developed so fast
00:36 autrijus that's anarchism for you :)
00:36 Lopo has quit IRC (Read error: 60 (Operation timed out))
00:36 sproingie i always thought perl6 would be beyond mere mortals
00:36 sproingie then again ...
00:36 * sproingie peers at ajitrus ... not sure about that mere mortal thing
00:36 sproingie atrijus even
00:37 autrijus nah. I'm a mere mortal, this I'm quite sure :)
00:40 Lopo has joined #perl6
00:41 * Odin- scribbles an encircled A somewhere very visible.
00:50 meppl has quit IRC (Remote closed the connection)
00:50 cls_bsd has quit IRC (Read error: 104 (Connection reset by peer))
01:05 ninereasons has quit IRC ("Leaving")
01:07 stevan autrijus++
01:07 stevan real classes,..
01:07 * sproingie sets up his 32 bit jail so he can compile ghc so he can do pugs
01:07 stevan :)
01:08 sproingie *grumblegrumble*
01:08 stevan mugwump: hey
01:09 sproingie grrrargh.  i still have the music from phantom brave stuck in my head
01:09 mugwump hi
01:09 stevan so is Hack::Instances obsolete now?
01:09 mugwump quite possibly
01:09 sproingie i must say, debian makes chroots drop-dead simple
01:10 mugwump sproingie: got linux-vserver.org?
01:11 sproingie mugwump: nope
01:11 sproingie the chroot is working, don't want to mess with a vserver just yet
01:11 mugwump jail your chroots!
01:11 sproingie not planning on running anything hostile in them
01:12 sproingie i just need a 32 bit chroot, not a jail
01:12 sproingie besides, i want the chroot to talk to the same localhost as the 64 bit system
01:12 sproingie so i can run x apps
01:13 sproingie tho i suppose there's some kind of bridged interface support
01:13 mugwump I'm not trying to convince you, just thought you might be interested ;)
01:14 sproingie yah i'll probably run a vserver if i decide to start perl6nomic
01:14 Lopo has quit IRC (Read error: 145 (Connection timed out))
01:14 sproingie i was holding out for classes, and now they've landed
01:14 sproingie unfortunately the regular perlnomic is meeting the same fate as all other nomics.  stagnancy leading to a dead game
01:15 castaway_ has joined #perl6
01:16 * sproingie has a de facto static ip like all the other cable modem folks, could probably take the traffic
01:16 cls_bsd has joined #perl6
01:18 Lopo has joined #perl6
01:18 theorbtwo has quit IRC (Read error: 145 (Connection timed out))
01:28 castaway has quit IRC (Read error: 110 (Connection timed out))
01:37 cls_bsd has quit IRC (Read error: 104 (Connection reset by peer))
01:37 cls_bsd has joined #perl6
01:42 stevan mugwump: I got can() to work
01:43 stevan        eval 'defined(&' ~ $method ~ ')' ?? eval '&' ~ $method :: 0;
01:43 stevan its an ugly ugly hack,... but then again, all of Hack::Instances is one :)
01:46 mugwump hoice
01:46 mugwump choice even
01:46 mugwump so, where in the world are you, stevan ?
01:46 stevan Northeast US
01:46 stevan your in NZ right?
01:46 mugwump I am
01:47 mugwump must be small hours there now then
01:47 mugwump no, what am I on about
01:47 mugwump evening?
01:47 stevan 9:46 om
01:47 stevan pm
01:48 mugwump ok, so you're 16 hours behind me at the moment, or 8 - 24 hours ahead
01:48 stevan it is tuesday here
01:49 stevan I think you are ahead of me
01:49 mugwump wednesday afternoon here
01:49 stevan yeah
01:49 stevan so have you played with the class support yet?
01:50 mugwump sadly I'm being distracted from my real work by my professional tasks
01:50 stevan ah
01:50 stevan where do you $work
01:50 sproingie ARGH.  ghc won't build from cvs
01:50 sproingie what version of ghc does pugs need?
01:51 stevan sproingie: 6.4
01:51 sproingie oh hey that's actually out now
01:51 mugwump stevan: I'm the webmonkey for http://www.bnz.marketview.co.nz/
01:51 sproingie i can just download a binary
01:51 sproingie would be NICE if ghc actually built out of the box tho
01:55 mugwump or apt-get -t experimental install ghc
01:57 cls_bsd has quit IRC (Read error: 110 (Connection timed out))
02:08 cls_bsd has joined #perl6
02:11 autrijus has quit IRC (Read error: 104 (Connection reset by peer))
02:18 cls_bsd has quit IRC (Remote closed the connection)
02:22 sproingie mugwump: got a sources.list line i can use for experimntal?
02:22 sproingie i'll put it in sources.list this time, honest :)
02:23 mugwump I just change "sarge" to "experimental" from my normal sources
02:23 sproingie i use ubuntu
02:23 mugwump ah
02:23 mugwump deb http://ftp.nz.debian.org/debian experimental main contrib non-free
02:23 mugwump you'll need to s/nz/$your_country/
02:25 sproingie apt-get -t experimental install ghc doesn't give me anything newer than the other repos
02:26 sproingie funny thing, apt-get update gives me a GPG error, and tells me to run apt-get update to fix it
02:27 sproingie how do i install the gpg key for the experimental repo?
02:27 * mugwump shrugs
02:28 mugwump it worked for me :)
02:29 stevan autrijus: how can I get $self with classes?
02:29 mugwump actually, seems 6.4-3 is now in unstable, too.
02:29 sproingie blean
02:29 stevan (I know you are not on, but I figure you will see in the backlog)
02:29 sproingie that means ubuntu is slower than sid
02:29 mugwump you could add unstable uris to your sources.list too, but pin them to a really low priority so the ubuntu packages take priority...
02:29 stevan mugwump: class support has some missing spots, but I am going to start converting Perl::MetaModel anyway
02:30 sproingie mugwump: i'm at a complete loss as to how I'd do that ...
02:30 mugwump stevan: great... hopefully I'll be finished with this week's work tomorrow
02:30 stevan nice
02:30 cjn has joined #perl6
02:30 mugwump sproingie: http://jaqque.sbih.org/kplug/apt-pinning.html
02:32 sproingie cool.  guess the worst that can happen is that i hose my system ... which is just a chroot
02:32 mugwump sure.  and the best thing that could happen is that it works and you post a nice info page about how ubuntu users can easily pick up extra packages from debian when they're not in multiverse or whatever
02:33 mugwump no pressure
02:33 cls_bsd has joined #perl6
02:36 sproingie dammit.  same gpg error from unstable.  i'll have to ask #ubuntu
02:38 sproingie jesus.  everyone there just bleats about how evil and wrong it is
02:38 sproingie and if i ask #debian, they'll just scold me for using ubuntu
02:45 cdpruden has quit IRC ("Oops. This machine just fell asleep")
02:45 sproingie ah, found one helpful soul
03:06 sproingie yay, got darcs
03:06 sproingie now for pugs
03:07 sproingie argh,  the darcs repository on pugscode.org doesnt want to work
03:07 sproingie has it moved primarily to subversion/svk now?
03:08 obra it has, but it's regularly syncced to darcs
03:08 sproingie yahbut the darcs url is all broked
03:09 sproingie guess i can get svn
03:13 sproingie ok, im an idiot with svn... how do i check out/update pugs with it?
03:14 stevan sproingie: svn co http://svn.openfoundry.org/pugs
03:14 sproingie kinda like cvs except with http.  nice
03:14 sproingie svn update to update?
03:14 stevan yes
03:15 stevan svn help is actually very informative too
03:16 stevan basically you can do svn help <command>
03:16 stevan and this comes in handy too http://svnbook.red-bean.com/en/1.0/index.html
03:18 sproingie cool
03:18 mugwump sproingie: there's an appendix in the svnbook that explains the differences going from CVS -> Subversion.  Highly recommended.
03:18 sproingie i know some cvs, but mostly i like pointy-clicky stuff like tortoisecvs
03:19 mugwump so get tortoisesvn!  :)
03:19 sproingie have it on the windows side
03:19 sproingie wish like crazy there was some equivalent for konqueror
03:19 mugwump yeah, cervisia isn't great
03:21 sproingie whups.  guess i need to install parrot too
03:21 sproingie what's this about hs-plugins?
03:22 mugwump set pugs compiling while you get those installed; you don't *need* them for all features of pugs
03:22 mugwump just PGE/yaml/etc
03:22 Khisanth PGE seems rather significant though :)
03:23 sproingie yah, i want it to produce parrot
03:24 sproingie svk really excites me, because we used to use cvs, but the Big Borg Parent uses perforce
03:25 sproingie i hate cvs, and my opinion of perforce is "meh", but a unified interface for them, very nice
03:26 sproingie oif.  parrot configure fails
03:26 sproingie does parrot not work with gcc-4?
03:27 dvtoo_ gcc4 almost doesn't work with itself ;)
03:27 sproingie sadly, I get the exact same error with gcc-3.3
03:28 dvtoo_ what os/arch?
03:28 sproingie ubuntu breezy, x86
03:28 sproingie actually a 32 bit chroot on x86_64, but it's compiled everything else ok
03:29 sproingie does Configure.pl leave a log?
03:29 dvtoo_ unsure, it's been a while since I looked
03:29 sproingie aha
03:29 sproingie it's looking for a "cc" command
03:29 sproingie which doesnt exist
03:30 dvtoo_ odd.  there isn't one on solaris either, but it works fine there
03:30 dvtoo_ well, relatively
03:30 dvtoo_ it builds ;)
03:36 sproingie guess i'll add the cc alias and try again
03:36 sproingie already uninstalled gcc-4, but it's too bleeding edge anyway
03:36 sproingie hope ghc doesnt require it
03:37 boogie has joined #perl6
03:38 dvtoo_ I doubt it even works with 4.0.0 ;)
03:40 dvtoo_ you may not have noticed, but I'm not super impressed with the gcc4 or ghc build process, but that's probably due to how much I've been fighting with both of them lately ;)
03:40 sproingie ghc failed to build for me, so i gave up and installed it from unstable
03:40 sproingie which takes a little work since i run ubuntu
03:40 dvtoo_ yeah, this is on solaris ;)
03:40 sproingie oy
03:42 sproingie made a bunch of symlinks, seems to work ok now
03:43 Lopo_ has joined #perl6
03:44 Lopo has quit IRC (Read error: 145 (Connection timed out))
03:44 Lopo_ is now known as Lopo
03:47 f0rth is pugscc with backend Pugs broken?
03:55 sproingie yay, pugs works
04:02 sproingie argh.  i can't compile hs-plugins.  no inline haskell for me
04:03 sproingie hm.  looks like it might have built enough
04:04 sproingie nope.  no evalhaskell
04:16 justatheory has joined #perl6
04:17 Khisanth has there ever been anything on the ML regarding unloading modules?
04:20 sproingie hs-plugins seems totally broken
04:23 sproingie ahh, older snapshot seems to work all right
04:40 Ovid has joined #perl6
04:49 sproingie has quit IRC (Remote closed the connection)
04:57 rafl_ has quit IRC (Read error: 104 (Connection reset by peer))
04:57 rafl has joined #perl6
05:05 batkins_ has joined #perl6
05:05 batkins_ will/does parrot support haskell?
05:11 Ovid has quit IRC ()
05:12 mugwump parrot supports anything that will compile to it!  :)
05:15 Khisanth wasn't there someone working on a gcc frontend that would target parrot?
05:16 * mugwump smokes some crack
05:16 mugwump yeah, I can see it
05:18 mugwump damn it's gone, and now all I feel like is more crack
05:18 batkins_ haha
05:19 batkins_ has quit IRC (Remote closed the connection)
05:30 wilx has joined #perl6
05:31 puetzk has quit IRC (Read error: 60 (Operation timed out))
05:38 puetzk has joined #perl6
05:51 b00t has joined #perl6
05:56 SM_ax has joined #perl6
05:59 justatheory has quit IRC ()
06:02 Shachaf has quit IRC ("So long, and thanks for all the fish!")
06:02 rob_ has joined #perl6
06:02 scook0 has joined #perl6
06:11 rob__ has quit IRC (Read error: 110 (Connection timed out))
06:24 Jouke has joined #perl6
06:25 Jouke has left
06:32 elmex has joined #perl6
06:33 iblechbot has joined #perl6
06:46 BigBear has joined #perl6
06:47 elmex has quit IRC ("leaving")
06:54 scook0 has quit IRC ("Leaving")
07:13 larsen has joined #perl6
07:13 larsen_ has joined #perl6
07:13 larsen has quit IRC (Client Quit)
07:14 nothingmuch has joined #perl6
07:14 larsen_ has quit IRC (Client Quit)
07:14 larsen has joined #perl6
07:29 chady_ is now known as chady
07:43 pernod has joined #perl6
07:43 pernod MSN'e*
07:44 pernod has left
07:57 osfameron has joined #perl6
08:20 knewt anyone got any ideas what this error when building hs-plugins is due to? [[ Rebuilding dependencies ... Could not find module `Data.List': ]]
08:20 larsen has quit IRC ("brb")
08:21 larsen has joined #perl6
08:25 lightstep has joined #perl6
08:32 knewt hmm. snapshot from 20050424 works. that'll do.
08:34 cls_bsd has quit IRC (Remote closed the connection)
08:35 knewt aarrrggghhhh. hs-plugins is /not/ sane. not bloody DESTDIR support
08:35 knewt s/not bloody/no bloody/
08:40 BigBear has quit IRC (Connection timed out)
08:43 luqui has joined #perl6
08:45 cls_bsd has joined #perl6
09:05 Juerd 1=
09:05 Juerd s/1=//
09:09 arcady has joined #perl6
09:14 Juerd Forgot my svn password :(
09:15 webmind smart :)
09:18 dada has joined #perl6
09:21 kolibrie has joined #perl6
09:22 theorbtwo has joined #perl6
09:29 meppl has joined #perl6
09:29 sekimura has joined #perl6
09:32 gaal|work Juerd, there's a send-password link on openfoundry.
09:32 meppl has quit IRC (Remote closed the connection)
09:32 Juerd Oh, right. I forgot that it was linked to a site
09:32 Juerd Thanks
10:06 broquaint has joined #perl6
10:10 JaffaCake has joined #perl6
10:10 JaffaCake anyone tell me where the darcs repo of pugs is?
10:12 saorge has joined #perl6
10:13 saorge_ has quit IRC ("Leaving")
10:13 nothingmuch JaffaCake: sure, one second
10:13 nothingmuch but the darcs repo is not worth it... svk revisions are applied to the working dir, and then 'darcs record' is done, once a minute or so
10:13 nothingmuch so any patch you submit will be a conflict
10:13 wolverian whoa, OO. pugs++
10:14 JaffaCake just want a quick way to download it, I don't have SVN installed
10:14 nothingmuch ah
10:15 nothingmuch hmm, it seems to have disappeared
10:15 nothingmuch http://wagner.elixus.org/~autrijus/darcs/pugs/
10:15 nothingmuch want me to rsync a copy of the repo to my website?
10:15 JaffaCake that's what I thought, I wondered if it had moved
10:15 b6s elixus' hosting is moving.
10:16 JaffaCake not a big deal, I'm trying with 6.2.2 now (trying to reproduce a bug someone reported in GHC)
10:24 Odin- has quit IRC (Read error: 104 (Connection reset by peer))
10:24 Odin- has joined #perl6
10:28 iblechbot has quit IRC (Connection timed out)
10:43 pdcawley has joined #perl6
10:48 BigBear has joined #perl6
10:50 JaffaCake has left
10:53 ajs Well, I figured out where that Parrot error was coming from yesterday... now to see if I can figure out enough Haskell to formulate a patch ;-)
10:53 pdcawley Heh.
10:53 pdcawley Is anyone working on P6::DBI?
10:54 kungfuftr pdcawley: don't believe so
10:54 ajs Well, we don't have OO yet, so it's probably a bit early, unless you just want to do design.
10:54 pdcawley Didn't OO just get announced?
10:54 kungfuftr thought about DBI::PurePerl, but it uses a scary amount of perl5 foo
10:54 kungfuftr ajs: OO is here
10:54 pdcawley On the blog.
10:54 ajs Heh, sorry I'm whole hours behind ;-)
10:55 * pdcawley might half inch one of the Pure Ruby interfaces to MySQL. Not standard, but enough to get up and running.
10:55 pdcawley I find myself thinking 'Perl6OnRails'
10:57 Juerd pdcawley: Can't you use a trojan ponie to get that running?
10:57 pdcawley Is ponie running?
10:57 pdcawley Plus, where's the fun in that.
10:57 pdcawley Perl 6 babie.
10:57 Juerd Ponie exists and pugs can eval_parrot
10:58 Juerd I'm no expert, but this ought to be enough to use ponie, right?
10:58 * pdcawley doesn't know.
10:58 Juerd pdcawley: Perl 6 programming is currently a hell of a job, with list context being broken
10:58 Juerd There's no way without eval to simply add an array or hash reference to an array
10:58 Juerd As references currently flatten :(
10:58 ajs As for DBI... I think you want roles before formalizing P6 DBI. I imagine that one of the best things that could happen to DBI would be to slice off a bunch of "many, but not all databases consider this standard" stuff into roles that can be applied at run-time by using special constructors. That would reduce the footprint of DBI considerably
10:59 Juerd Instead of push @foo, [], you have to do my $i = @foo.elems; eval "\@foo[$i] = []"
10:59 Juerd Because a non-literal-integer expression in [] provides list context to the rhs, you can't just do @foo[@foo.elemes] = [] even
11:00 broquaint Couldn't someone "just" write a Parrot DBI then call that from Pugs?
11:00 pdcawley Juerd: Yikes!
11:00 Juerd pdcawley: That's what I thought when I discovered this yesterday.
11:00 ajs broquaint: yes, though that would still want P6 window-dressing.
11:00 Juerd pdcawley: I really hoped/thought/assumed pugs would have had this fixed by now
11:01 Juerd It can be that the current behaviour is intentional
11:01 Juerd In that case, it's time to discuss list context again
11:01 pdcawley Can you write 'sub pushref {...}' to handle that?
11:01 gaal|work has quit IRC ("Trillian (http://www.ceruleanstudios.com")
11:01 kungfuftr pdcawley: Rosetta ?
11:01 pdcawley Or does [] get flattened by the function call?
11:01 Juerd pdcawley: It'd be a pushscalar, but yes.
11:01 ajs what's the magic cookie that prevents agressive haskell optimizatioin and speeds up builds of pugs? I'm dying with these frequent updates :)
11:02 Juerd pdcawley: No, a scalar parameter provides scalar context
11:02 pdcawley unoptimized
11:02 ajs is that a flag to Makefile.PL?
11:02 pdcawley Are multimethods in?
11:02 broquaint But isn't the hard part writing the DB interface, ajs?
11:02 ajs NM, found it. Thanks for the poiunter!
11:03 ajs broquaint: hrm... depends. Writing a P6 DBI (even given a Parrot under-layer) is challenging. P6's relationship with its OO system is not yet clear in terms of practical coding
11:04 broquaint Surely by the time it's written perl6 will be finished anyhow ;)
11:04 ajs Writing Parrot DBD modules should be trivial enough, since you can adapt much of the P5 DBDs
11:05 knewt hmm # [[ Failed 27/249 test scripts, 89.16% okay. 80/4861 subtests failed, 98.35% okay. ]]
11:05 pdcawley ISTM that attempting to write real apps in 6 is one of the things that'll supply Autrijus with loads of tests.
11:05 Juerd I remember that hashrefs serialized as []
11:05 Juerd eh
11:05 Juerd arrayrefs as []
11:05 Juerd hashrefs as {}
11:05 nothingmuch has quit IRC ()
11:05 Juerd They're both \() now. That's horrible to work with.
11:05 pdcawley Juerd: When did that happen?
11:05 ajs pdcawley: certainly, and taking a stab at something that might change with evolving usage isn't all bad
11:06 ajs er, isn't *at* all bad, is what I meant
11:06 knewt so, am i supposed to be seeing that many failing tests?
11:06 ajs knewt: dunno. Haven't run a make test today
11:07 Juerd pdcawley: I have no idea -- I haven't used pugs for weeks
11:07 pdcawley Ah... it's a pugs thing, not a Larry thing?
11:07 kungfuftr broquaint: there are DB's that don't require C... for example csv based databases... etc
11:07 Juerd pdcawley: If it's not a bug, then we need a hundred extra operators to compensate for the loss in functionality.
11:08 Juerd references in list context must absolutely never flatten, because you can then no longer pass references in a list.
11:08 broquaint I would've thought a sqlite DBD would be fairly starightforward too, kungfuftr.
11:08 kungfuftr broquaint: yup... dbd::sqlite is self-contained
11:08 ajs Larry has changed his stance a few times on what refs do in list context, as I recall
11:09 Juerd Auto-dereferencing must only occur in SPECIFIC contexts, never in list context, never in "Any" context
11:09 * pdcawley notes that push's context should probably be 'multi scalar'
11:09 Juerd pdcawley: Signatures have no way to provide an arbitrary number of scalar context items
11:10 pdcawley ie: @foo = (); @bar = ('bar'); push @foo, @bar; # @foo = (['bar'])
11:10 broquaint I wonder if anyone has written a Parrot interface to sqlite ...
11:10 luqui Juerd: that's not true
11:10 Juerd pdcawley: And simple literals must be possible. I was unable to create a matrix of a b c d e / f g h i j / k l m n o / p q r s t / u v w x y with nested [], because of the broken list context. I had to hard code indices :(
11:10 luqui Larry has said it's something like *@foo is context(Scalar)
11:10 * kungfuftr has noted that it's entirely possible to map perl6 objects and relationships directly into oracle and pgsql
11:10 Juerd luqui: Oh. When was that?
11:10 luqui it's in some of the blessed documents somewhere...
11:10 kungfuftr if strictly typed that is
11:10 pdcawley Need to do push @foo, *@bar to get @foo == ('bar')
11:11 knewt kungfuftr: native class::dbi ?
11:11 Juerd luqui: Then I think scalar context slurpy lists need to be the default, and well, list context can go away entirely.
11:11 kungfuftr knewt: no, not class::dbi, more like tangram
11:11 luqui pdcawley: you really think that push should have multiple scalar context?
11:11 knewt kungfuftr: never really looked at tangram. maybe i should ;)
11:11 luqui Juerd: why?
11:12 Juerd luqui: List context that you don't encounter everywhere, especially if it flattens aggressively, is useless anyway, because you can do the same with **, explicitly.
11:12 pdcawley luqui: Only if you can't otherwise distinguish between @foo and []
11:12 Juerd I cannot imagine ANY occurrence of wanting aggressive flattening of literals
11:12 pdcawley Though, thinking about it, even if you can.
11:12 luqui Juerd: by agressive flattening, you're referring to the infinite dereferencing semantics?
11:12 Juerd luqui: But I'd rather just have a list context that flattens only arrays and hashes, not references.
11:13 Juerd luqui: No, flattening references.
11:13 Juerd luqui: currently, ([],[],[],[]) is a 0 element list.
11:13 luqui there's no context that does that...
11:13 pdcawley @foo and %foo *are* references.
11:13 Juerd pdcawley: No, they are not.
11:13 Juerd pdcawley: @foo in list context is its elements, in scalar context it's referenced
11:13 luqui Juerd: they are references that flatten themselves in flattening context
11:13 Juerd pdcawley: If you want a reference in list context, use \@foo.
11:14 Juerd pdcawley: This hasn't changed since Perl 5. @foo in scalar context just returns a ref instead of th enumber of elems
11:14 Juerd luqui: Whatever description suits you :)
11:14 luqui yeah, both descriptions work
11:14 Juerd I prefer to think of @foo and %bar as "real" arrays and hashes
11:14 pdcawley luqui: I rather think that 'flattening context' is a much better description of Perl 6 semantics as I understand them.
11:14 Juerd while \@foo, [] are references
11:14 Juerd That makes discussing things lots and lots easier
11:14 Juerd Of course, everything internally is a reference if it has a name :)
11:15 Juerd pdcawley: I've suggested changing the names of contexts (singular and plural), but it hasn't happened
11:15 Juerd pdcawley: Other suggestions are: item versus slurpy (as implemented in pugs)
11:15 Juerd pdcawley: And scalar versus flattening
11:16 Juerd I like item versus slurpy
11:16 kungfuftr Juerd: "steamroller context"!
11:16 luqui I like the other two :-)
11:16 Juerd kungfuftr: No, ** is the steamroller
11:16 pdcawley scalar, flattening and steamroller contexts...
11:16 Juerd kungfuftr: This is separate from list context
11:16 luqui yeah, like this:  sub foo (**@_)
11:16 Juerd luqui: Which other two?
11:16 luqui the other two pairs of names
11:16 Juerd luqui: Which names?
11:17 kungfuftr *blink* there's actually a steamroller context!?
11:17 Juerd luqui: Oh, nevermind. I understand now.
11:17 ajs yes
11:17 Juerd kungfuftr: Yes.
11:17 kungfuftr fuck me#
11:17 Juerd kungfuftr: And ** is the steamroller.
11:17 pdcawley **$thing --> take this abitrary data structure and present it as a simple list.
11:17 luqui kungfuftr: isn't it great learning about perl 6.  All your jokes are actually features
11:17 Juerd pdcawley: aggressively, unlazying.
11:17 ajs ** is like $$ but upside down
11:18 kungfuftr luqui: yeah
11:18 pdcawley Presumably composite classes will be able to declare how they decompose under a steamroller?
11:18 lightstep has quit IRC ("love is all you need")
11:18 Juerd luqui: Yesterday someone asked me in private: "Is yada yada yada really an operator?!"
11:18 luqui yeah.
11:18 luqui but I still say no to ??? and !!!
11:18 Juerd I dunno
11:19 Juerd I kind of like them
11:19 kungfuftr no to proposition S29!
11:19 pdcawley ??? and !!!?
11:19 Juerd !!! if <assertion>;
11:19 luqui they're like ..., but more agressive or something
11:19 * pdcawley really should be writing the summary.
11:19 * luqui is quite opposed to them
11:19 ajs ooh, I like that
11:19 broquaint Just c&p #perl6 logs, pdcawley ;)
11:19 pdcawley !!! if ???
11:19 Juerd pdcawley: ..., ??? and !!! are the same thing, but ??? warns, and !!! dies, while ... merely fails
11:19 pdcawley broquaint: I think not.
11:19 Juerd pdcawley: The behaviour of "fail" is configurable
11:20 luqui Juerd: I thought ??? was the one that failed
11:20 kungfuftr ah, nice
11:20 pdcawley That much I knew. And in my modules it will throw an exception.
11:20 chady is now known as chady_
11:20 Juerd luqui: I think not. Let's see.
11:20 Juerd The box on which I read mail is overloaded
11:20 Juerd This can take a while.
11:21 Juerd Reading /home/juerd/Mail/perl6-language... 2500 (59%)                          
11:21 * pdcawley would far rather catch exceptions then spend his life checking return values.
11:21 pdcawley At least truth is out of band now...
11:21 ajs pdcawley: I hope you mean, "in my modules it will (by default) throw an excpetion," since clearly the user of your module might want consistency across their entire universe
11:21 Juerd pdcawley: Me too. That's why I *love* the new "fail"
11:21 luqui Juerd: you're right
11:21 Juerd pdcawley: You can simply "use fatal" and all fails are dies
11:21 pdcawley Juerd: I know.
11:21 Juerd luqui: Just when I found the message! :)
11:21 Juerd pdcawley: I still like the undef default
11:22 luqui that's why it's configurable :-)
11:22 pdcawley ajs: That's why I'll use fail not die. But all my modules will begin 'use fatal;'
11:22 Juerd pdcawley: Because for throwaway scripts, which I write much more often, it saves you lots of evals
11:22 Juerd trys.
11:22 Juerd pdcawley: Hm
11:22 pdcawley Juerd? Why try, just let the exception kill the script.
11:22 Juerd pdcawley: "use fatal" turns fail into die, I thought
11:22 * luqui has that message *'d, so that he can chew out Larry tomorrow
11:22 pdcawley It does...
11:23 Juerd pdcawley: Then if something else fails, you die, and if you fail, you merely fail.
11:23 pdcawley But you just have to do CATCH ... {...} in an appropriate block, no need to try.
11:23 Juerd pdcawley: I think we then need something that turns fail into double-fail.
11:23 ajs pdcawley: hrm... doesn't that mean that your module will thrown exceptions regardless of the caller's wishes? Better to put that use fatal into your calling code....
11:23 luqui double-fail?
11:23 pdcawley ajs: No.
11:23 Juerd luqui: like double-return
11:23 pdcawley use is lexically scoped.
11:23 luqui Juerd: which is?
11:23 ajs yes, correct. It's scoped
11:23 pdcawley Users of any module will get whatever fail behaviour they want.
11:23 Juerd pdcawley: Yes, but die is globally killing the script.
11:24 Juerd luqui: return not only from the current sub, but also from the calling sub
11:24 luqui oh
11:24 Juerd pdcawley: Not if the *module* uses fatal.
11:24 luqui never had that need...
11:24 Juerd pdcawley: If the module uses fatal, and uses another module or function that fails, the module dies, regardless of fatal.pm setting in the caller.
11:24 ajs "use fatal" will set up exception behavior in your module, regardless of what any outer lexical scope desires
11:24 luqui Juerd: nope
11:24 Juerd pdcawley: That's why you'd need a double-fail instead.
11:24 luqui It's caller based
11:25 luqui oh
11:25 * Juerd phone
11:25 luqui I think I see what you're saying
11:25 pdcawley use fatal; says "In this module, anything I call should die to fail"
11:25 pdcawley If I then fail to catch that exception, it'll propagate out to my callers.
11:25 luqui yep
11:25 pdcawley But any 'fails' I throw will behave according to the pragmatic state of my caller.
11:26 luqui Juerd wasn't saying what I thought he was
11:26 ajs Example: no fatal; { use fatal; fail; }
11:26 luqui ajs: that isn't a call
11:26 ajs correct
11:26 wolverian no fatal; class Foo { use fatal; sub bar { fail } } Foo.bar; # I don't want this to die
11:26 pdcawley And it won't.
11:26 luqui okay --- to get things straight:  "fail" depends on the "use fatal" setting of the CALLER ONLY
11:27 wolverian ah, okay. how is that implemented, by the way? by walking caller()?
11:27 luqui is it implemented?
11:27 Juerd luqui: So "use fatal" doesn't just turn any fail of called subs into die, it turns them into double-fails
11:27 wolverian well, I don't know. I was just wondering if it was some weird scoping rule instead.
11:27 Juerd luqui: Which caller, that's the qusetion.
11:27 wolverian (which one could use in other places)
11:27 pdcawley That's the calling function right, not the calling package (cf 'warn'/'carp')
11:27 luqui Oh, the caller of the thing that's calling fail
11:27 Juerd luqui: caller of in fatal context, or top level caller?
11:27 ajs So in that case, you pick up a lexical "use fatal" inside bar. The only way that "Foo.bar" doesn't choke is if "no fatal" implicitly wraps in a catch
11:27 luqui the "double caller"
11:28 Juerd luqui: Right. Then it's okay.
11:28 luqui okay good
11:28 luqui wow
11:28 pdcawley Fail will always be a tail call...
11:28 Juerd luqui: That's exactly what I meant :)
11:28 wolverian ajs: right. that makes sense.
11:29 knewt so how will "use fatal" propagate into coroutines? just on the initial call, or differently for each re-entry?
11:29 Juerd knewt: Per re-entry, I presume.
11:30 ajs knewt: It depends. If we have the implicit catching that pdcawley seems to be suggesting, then coroutines calls would be wrapped however each invocation desires.
11:30 luqui knewt: uh... coroutines & continuations + exception handling = not thought about yet
11:30 * pdcawley could be wrong about the implicit catching.
11:31 luqui Yeah, I think each call into the coroutine could use a different fatal setting
11:31 ajs Anyone mind if I formulate this as a P6L message? I find the question intriguing
11:31 Juerd (localized sites with automatic, unoverrideable selection)--  # fuck you, drop dead, and more like that.
11:31 luqui go for it
11:31 pdcawley But it makes sense to me that any block can have CATCH blocks within it.
11:31 * knewt considers repulling & recompiling pugs later today and seeing if he gets fewer test failures
11:32 pdcawley It's currently refusing to link on this OS X box.
11:32 pdcawley (Tiger)
11:32 pdcawley whinging about an undefined symbol __DISCARD__
11:32 knewt my computer spent all of last night building ghc packages for me :)
11:33 knewt as far as pugs goes i'm getting 27/249 test script fails, and 80/4861 subtest fails
11:33 knewt and 1 unexpectedly succeeded subtest
11:34 Juerd Shouldn't a CATCH { } block imply use fatal?
11:34 pdcawley No.
11:34 kungfuftr Juerd: nope, lack of a CATCH block should imply fatal?
11:34 pdcawley Something could still be explicitly die/throwing.
11:34 Juerd That's clear.
11:36 kungfuftr you might also have lots of nested blocks and want to do the die/throw at a much higher level
11:38 pdcawley (That being rather the point of using exceptions in the first place...)
11:40 Juerd You're right.
11:41 Juerd Now I go to the office
11:41 Juerd To see if there's work for me.
11:44 iblechbot has joined #perl6
11:44 ajs message sent to list on fail / use fatal. Please send any corrections to my summary of the quandry.
11:45 ajs heh, last night's update changed the behavior of my Parrot/factorial bug. Now it just calculates the factorial wrong ;-)
11:55 gaal hi there.
11:56 nothingmuch has joined #perl6
11:58 luqui has quit IRC (Remote closed the connection)
11:58 broquaint has quit IRC ("Chatzilla 0.9.68a [Firefox 1.0.3/20050414]")
11:59 nothingmuch has quit IRC (Client Quit)
11:59 ajs Got my test case down to a simple reproducer: ./pugs -BParrot -e 'sub a($x){$x} say a(1)+1'
12:00 ajs Now to figure out where the off-by-one problem is in picking $P registers in the PIR output
12:04 vitaminmoo has quit IRC (Read error: 110 (Connection timed out))
12:08 mj41 has joined #perl6
12:13 pupilzeng has joined #perl6
12:17 ajs I just love writing code in a langauge I don't know in order to debug... it's like using an ice pick to pick up eggs
12:20 pdcawley Hhe.
12:21 scook0 has joined #perl6
12:38 broquaint has joined #perl6
12:44 b00t has quit IRC (Remote closed the connection)
12:45 cookys has quit IRC ("leaving")
12:46 cookys has joined #perl6
12:47 broquaint has quit IRC ("Chatzilla 0.9.68a [Firefox 1.0.3/20050414]")
12:52 Limbic_Region has joined #perl6
12:53 Limbic_Region There is more than 1 way to help!!! http://use.perl.org/comments.pl?sid=2646​6&amp;op=&amp;threshold=0&amp;commentsor​t=0&amp;mode=thread&amp;pid=40310#40323
12:55 nnunley has joined #perl6
13:02 scook0 Hey, does anyone know how to add type annotations to Haskell functions?
13:02 scook0 (Using the inferred type)
13:02 integral the :: things?
13:02 theorbtwo Well, the format for adding a type annotation is funcname :: type
13:02 * integral just puts nonsense for the type, then copies what the compiler says it should be
13:02 theorbtwo If you just say funcname :: a, and follow the error messages, you'll get there eventually.
13:03 scook0 That's how you did it?
13:03 scook0 I'll try that...
13:03 theorbtwo If you want to cheat, and the function is exported, you can do ghc --show-iface=path/to/foo.hi | less, and find the signature there.
13:09 scook0 Aha! If you go into ghci, then :load the module, you can use :t to find the type.
13:16 pmichaud has joined #perl6
13:18 mj41 WinXP, r2962 -  780/4861 subtests failed, 83.95% okay. 16 subtests skipped.
13:19 Limbic_Region http://perlmonks.org/index.pl?node_id=455979  # GET INVOLVED WITH PUGS!!!!
13:21 mj41 Anybody clever enough for adding "for @array.values -> $val is rw { $val++ }"  t/statements/for.t ?  More important than OO for me :-).
13:21 Juerd In case you have been hiding under a rock since before February, the Pugs project is giving us the ability to write Perl6 today.
13:21 Juerd </quote>
13:22 Juerd Even if you have not been hiding under a rock, the Pugs project is still giving us the ability to write Perl 6 today.
13:22 Juerd Limbic_Region: Use a colon, not a comma :)
13:22 * integral thought VIM did that :-P
13:23 Limbic_Region mj41 - you sure that doesn't work now?
13:23 * Limbic_Region is pretty sure it got added
13:23 Juerd mj41: @array.values is spelled @array
13:23 Juerd Limbic_Region: I thought the 100 days were rather long
13:23 Limbic_Region heh
13:24 Limbic_Region well the days themselves were long - but given what has been accomplished it took no time at all
13:28 mj41 ok 27 - for @list -> $num is rw { $num++ }
13:28 mj41 not ok 28 - for @list.values -> $num is rw { $num++ }
13:28 mj41 not ok 29 - for %hash.values -> $val is rw { $val++ };
13:28 mj41 not ok 30 - for %hash.kv -> $key, $val is rw { $val++ };
13:30 * Limbic_Region wonders what for @list.values does that for @list doesn't mj41
13:30 Limbic_Region you indicated it was more important to you than OO was
13:31 Limbic_Region AFAICT, they would do the same thing if both worked but since 1 does why squawk about the other?
13:31 mj41 autrijus said something about lvalue sub,  %hash.kv is important too
13:32 broquaint has joined #perl6
13:32 Limbic_Region the %hash.values thing makes sense to me
13:32 Limbic_Region not the array though
13:33 Limbic_Region but I am still interested in knowing why it is more important to you than OO?
13:33 Limbic_Region was the @array.values just an example and you really want something else that is similar?
13:34 mj41 because I wrote code when it was working --- http://wiki.kn.vutbr.cz/mj/attach/pugs/fool/tut.p6
13:34 Limbic_Region Juerd - A bug test is simple: write what do and what you expect first:
13:35 Limbic_Region that doesn't make a lot of sense to me?
13:35 mj41 yes example, I want "%hash.kv -> $key, $val is rw"
13:36 merlyn has joined #perl6
13:36 Limbic_Region hey merlyn - welcome
13:36 merlyn In response to your PM posting. :)
13:36 Juerd Limbic_Region: Why?
13:36 pmichaud hello, merlyn
13:36 Juerd Hey merlyn!
13:36 Limbic_Region *grin* - already getting desired effect
13:36 merlyn wow.  hangin' with the UberCoolCats
13:36 mj41 I used .kv is rw, but now it is broken. Sorry, my English is too bad. :-(
13:37 merlyn I'm not worthy. :)
13:37 Limbic_Region "write what do and" - that just doesn't parse right for me Juerd - but I am under-caffeinated ATM
13:37 Limbic_Region mj41 - that makes sense - sorry I can't help with that though - I don't speak Haskell
13:37 Limbic_Region yet
13:37 mj41 :-0
13:39 Juerd Limbic_Region: What wolud parse right for you?
13:39 dada has quit IRC ("1;")
13:40 ajs mj41: There is a serious problem if you're using the parrot back-end with the use of return values right now. I'm working on tracking it down, but the symptom is that $P{n} gets the return value, but later on, $P{n-1} is read.
13:40 jhorwitz has joined #perl6
13:40 Limbic_Region Juerd - I guess "write what to do and..." makes more sense to me - but you probably speak better English than me ;-)
13:40 Juerd Limbic_Region: No, it's not what to do.
13:41 Limbic_Region *shrug* - I have read the sentence several times and it doesn't sound right but I guess it's just me
13:42 Juerd Perhaps past tense makes it more clear?
13:42 Juerd What you did, and what you expected
13:42 Limbic_Region yes
13:42 Limbic_Region perfect
13:43 xern has quit IRC ("Lost terminal")
13:43 Limbic_Region but you notice what word you added that made it clear
13:43 Limbic_Region it isn't the tense
13:43 Limbic_Region it is the word "you"
13:43 Limbic_Region write what "you" do and ... - would also have worked for me
13:44 Juerd That was already there.
13:45 Juerd Oh, no, the word is missing indeed.
13:45 Aankhen`` has joined #perl6
13:45 * Limbic_Region even copy/pasted
13:45 Juerd Yeah, I don't know why I misread the paste as well.
13:49 Southen has joined #perl6
13:52 mj41 "for @list.kv -> $key, $val is rw { $val++; say "$key : $val" }" is more meaningful
13:52 integral *argh* GHC did something impossible :-(
13:53 Juerd mj41: Name it @array then
13:53 Juerd The name "list" is sort of like a promise you don't have a permanent association between indices and values
13:59 scook0 has quit IRC ("Leaving")
14:01 mj41 I replicated t/statements/for.t  ... so "my @list = (0..5);" is wrong name?
14:01 SM_ax has quit IRC ("ChatZilla 0.9.61 [Mozilla rv:1.7.7/20050420]")
14:07 stevan Limbic_Region++ # champion of advocacy
14:07 Juerd mj41: I personally think the word 'list' should never be used as an array identifier
14:08 Southen has quit IRC (Connection timed out)
14:08 Juerd mj41: Unless it is the single slurpy list of a general purpose function (*@foo)
14:08 Southen[a] has quit IRC (Client Quit)
14:08 Juerd mj41: Like in map's (&block, *@list)
14:09 Aankhen`` has quit IRC ("Induhvidual quote: "When it comes to nut-cutting time, the cream will rise to the top." [Time wasted online: 23mins 55secs]")
14:09 Juerd mj41: This all to avoid any confusion with lists, which are not arrays.
14:11 mj41 Juerd: ok, I think I remember Pugs with my @list = (0,,1); say @list.ref; prints "List" :-)
14:12 Aankhen`` has joined #perl6
14:19 mj41 so Can I fix all tests? supervised s/list/array/
14:19 mj41    my @pop = (1, 2, 3, 4);
14:19 mj41 -    is(+@pop, 4, 'we have 4 elements in the list');
14:19 mj41 +    is(+@pop, 4, 'we have 4 elements in the array');  
14:19 mj41 ok?
14:20 Juerd mj41: Yes
14:20 integral NB: you may need to make clean after r2968
14:20 Juerd mj41: Or simpler: The array has 4 elements
14:23 mj41 and plural form? lists -> arrays, is it?
14:24 Juerd Yes
14:30 pmichaud has quit IRC ("Chatzilla 0.9.68a [Firefox 1.0.3/20050414]")
14:34 mj41 "my @result = grep { ($_ % 2) }, @list; " is ok?
14:35 Shachaf has joined #perl6
14:39 mj41 my $array_ref1 = ("foo", "bar", "baz");
14:39 mj41 -  isa_ok($array_ref1, 'List');
14:39 mj41 + isa_ok($array_ref1, 'Array');   ?
14:52 stevan Limbic_Region++ # champion of advocacy
14:54 stevan iblechbot: you read my mind (i was going to move =:= to t/operators too)
15:10 jiing__ has joined #perl6
15:12 mj41 say (401, 201, 1).ref;   # List
15:12 mj41 my @a = (401, 201, 1);
15:12 mj41 say @a.ref;  # Array
15:13 stevan mj41: you really should put this code into a test
15:13 stevan t/pugsbugs/ is a good place for them (assuming they are really bugs)
15:14 mj41 Juerd is our expert :-)
15:14 stevan mj41: isa_ok($array_ref1, 'List'); is ok
15:15 stevan because Array is derived from List (IIRC that is)
15:15 mj41 I think they are good.
15:15 stevan this might be a bug
15:15 stevan say (401, 201, 1).ref;   # List
15:15 stevan although I don't know for sure
15:19 ajs stevan: what would you have expected from that expression?
15:22 ihb has joined #perl6
15:23 Corion has joined #perl6
15:23 broquaint has quit IRC ("Chatzilla 0.9.68a [Firefox 1.0.3/20050414]")
15:25 ihb has quit IRC (Client Quit)
15:25 Corion I would be careful about exchanging isa_ok($foo, 'List'); and isa_ok($foo, 'Array'); #in either direction.
15:25 Corion ... at least with Perl5 there are important differences and I assume there are differences in Perl6 too.
15:26 mj41 multi sub al ( List @a ) {    say 'l' ~ @a; }
15:26 mj41 multi sub al ( Array @a ) {    say 'a' ~ @a; }
15:26 mj41 al( (1..3) ); my @a = (1..3); al( @a );
15:26 mj41 prints   l1 2 3  \n  a1 2 3
15:26 jiing__ has quit IRC ("Computer goes to sleep!")
15:27 Corion mj41: Which looks to me as if Perl5 semantics survive in Perl6 in this case.
15:28 stevan ajs: I am not sure, but I suppose it would be "Array"
15:31 zuulvin_ has joined #perl6
15:32 * pdcawley could go with 'list' when it's actually describing: (a => (b => (c => undef)))
15:33 Corion pdcawley: That's Pairs. And it tshould be pronounthed like thith.
15:33 pdcawley It's a lisp (as in LISt Processing) style list.
15:33 pdcawley So I'll go with 'list' when I'm using it as one.
15:37 chady has joined #perl6
15:40 stevan pdcawley: I think the => in your example was confusing
15:41 stevan (a : (b : (c : undef)))
15:42 Corion Dotted pairs 4 life, dawg!
15:42 stevan :D
15:43 osfameron I thought you could just do a=>b=>c=>undef ?
15:43 stevan osfameron: those would end up being nested pairs
15:43 chady has quit IRC ("leaving")
15:43 osfameron ah, is that decided?
15:43 Corion osfameron: We are reminiscencing on our time spent with Lisp :)
15:44 osfameron I posted the question to p6l and it seemed that it would do the right, um, lispy thing
15:48 Jouke has joined #perl6
15:48 Jouke has left
15:49 Southen has joined #perl6
15:52 Corion r2975 - 4794 ok, 84 failed - http://datenzoo.de/pugs/win2k.html # where do the 4 additional fails come from? Are these the unlink()/Win32 fails?
15:52 mj41 perlbot: nopaste
15:52 perlbot Paste your code here and #<channel> will be able to view it: http://sial.org/pbot/<channel>
15:53 pasteling "mj41" at 147.229.221.107 pasted "List vs. Array" (41 lines, 1.1K) at http://sial.org/pbot/10070
16:00 * Corion restarts the smoke
16:05 Corion C:/DOKUME~1/ADMINI~1/LOKAL​E~1/Temp/ghc1880.hc:3256: warning: implicit declaration of function `u_iswalpha'
16:05 mama_pitufo has joined #perl6
16:06 xinming has joined #perl6
16:08 arcady has quit IRC (Read error: 110 (Connection timed out))
16:08 autrijus has joined #perl6
16:10 mama_pitufo has left
16:10 pupilzeng has quit IRC (Read error: 104 (Connection reset by peer))
16:10 * wilx clubs Corion for he uses Administrator account for work.
16:10 Corion wilx: Naah, it's for play :)
16:11 PerlJam autrijus++ still quick on the implement.
16:11 autrijus Juerd: your bug is fixed. sorry. :)
16:11 PerlJam I just read Larry's message and I come here et voila, it's done!
16:12 autrijus :)
16:12 autrijus Juerd: however, I'm not so sure about this
16:12 autrijus (1)[0]
16:12 autrijus vs
16:12 autrijus (1).[0]
16:12 autrijus vs
16:12 autrijus 1.[0]
16:13 autrijus should all three be "1"? all three be castfail?
16:13 autrijus or is ()[] a special syntax?
16:13 wilx Hmm, aren't we just changing Perl5's dark corners for Perl6' dark corners? :)
16:14 autrijus I sure hope we're not :)
16:14 PerlJam wilx: We're always doing that.  hopefully perl6 has fewer though.
16:14 Juerd autrijus: Thanks for this important fix :)
16:14 PerlJam autrijus: If we keep the analogy that perl is like english, we must have dark corners *somewhere*  ;)
16:14 autrijus np. I'm reverting (1)[0] to return 1, though
16:14 autrijus instead of castfail
16:14 Juerd autrijus: (1)[0] adds . because two bracketing operators hit eachother, 1.[0] is different
16:15 autrijus how is (1) different from 1  again?
16:15 kungfuftr Juerd: was it you who found Perl Medic disappointing?
16:15 wilx Hmm, (1) makes a literate list/array?
16:15 autrijus <- firmly of the "() is merely precedence fixer" camp
16:15 Juerd autrijus: it's not
16:15 wilx s/literate/literal/
16:15 Juerd autrijus: (1,).[0] versus (1).[0]
16:15 autrijus so (1) is always parsed as (1,) ?
16:16 PerlJam wilx: I seem to recall that Larry was going the way of python on that one (i.e. you need a comma)
16:16 autrijus right. so (1).[0] ought to be castfail
16:16 Juerd autrijus: Without anything special, () just groups, right?
16:16 autrijus and (1,).[0] is what one wants
16:16 autrijus is that the idea?
16:16 Juerd autrijus: Castfail, or perhaps be smart, because it can't mean anything else
16:16 autrijus but if be smart, then 1.[0] needs to be smart as well
16:16 Juerd Hm..
16:16 autrijus (I'm arguing the be smart is probably the right choice, btw.)
16:16 kungfuftr moo?
16:17 Juerd autrijus: I can't choose
16:17 autrijus rand(2)
16:17 Juerd Please make that int rand 2 :)
16:17 Juerd Otherwise there are still way too many possibilities ;)
16:18 autrijus ;)
16:18 wilx Hmm..
16:18 autrijus but anyway. flattening.t was testing that
16:18 wilx Writing 1.[0] to get 1 is odd.
16:18 autrijus my @a = (1..10); @a[0][0]
16:18 autrijus is undef
16:18 wilx It should be error.
16:18 autrijus that is patent nonsense.
16:19 Southen has quit IRC (orwell.freenode.net irc.freenode.net)
16:19 Odin- has quit IRC (orwell.freenode.net irc.freenode.net)
16:19 b6s has quit IRC (orwell.freenode.net irc.freenode.net)
16:19 autrijus it should either throw an exception, or return 1, imho.
16:19 Juerd autrijus: Isn't it so that postfix [] provides array context to its LHS, in which () becomes equivalent to []?
16:19 Juerd And 1 is simply wrong?
16:19 autrijus Juerd: ah but what about @a[0][0] ?
16:20 Southen has joined #perl6
16:20 Odin- has joined #perl6
16:20 b6s has joined #perl6
16:20 autrijus and also, note that in lhs, it's currently just (Val (VInt 1))
16:20 autrijus because that's what you get when you write (1)
16:20 Juerd autrijus: @a[0] goes well, @a[0][0] is a type error, because 1 is not an array
16:20 autrijus if you want the syntactic comma, (1,) is needed
16:20 Juerd BUT - beware of a possible &Num::postcircumfix:<[ ]>
16:20 autrijus sure. assuming it doesn't exist
16:21 autrijus ok. so I gather you are favoring type error for both 1.[0] and @a[0][0]?
16:21 Juerd I'm not sure "type error" is correct
16:21 Juerd But error, yes
16:21 Juerd I'm more thinking it should complain about !.can(pcf:<[ ]>)
16:21 autrijus yeah sure
16:21 autrijus that's what I mean. mmd failure etc.
16:22 Juerd Oh, yeah, crap
16:22 Juerd [] is multi, not method
16:22 Juerd I still think that's a mistake :)
16:22 autrijus *shrug* :)
16:22 autrijus i.e. I sympathize but I can't see where it makes a difference
16:23 Juerd In error reporting, and in syntax used to override.
16:23 Juerd class MyArray is Array { ... }
16:23 Juerd It would be nice if you could do all the overriding in the class, with methods
16:23 Juerd Rather than methods in the class, and other things with multisubs
16:23 autrijus I thought if you define methods, they get looked up first, then fallback to mmd
16:23 Corion r2981 - http://datenzoo.de/pugs/win2k.html - 164 failed ... that's a lot ...
16:23 autrijus i.e. invocant's method gets looked up first
16:23 autrijus then multisubs
16:24 autrijus Corion: yes, discussing that with juerd currently
16:24 Juerd autrijus: For method syntax, but I don't know if the same behaviour is defined for postcircumfix operators
16:24 autrijus Corion: lots of code depend on something equivalent to (1)[0]
16:24 Corion Aaaah
16:24 autrijus which has just been made illegal
16:24 autrijus and which I'm not exactly sure is a good idea
16:24 Juerd merlyn: You invented the list slice, right?
16:24 autrijus Juerd: I think they need to be consistent on this...
16:24 Juerd merlyn: What's your opinion about this?
16:25 * Juerd tries to separate neodymium magnets
16:25 autrijus fwiw, (1)->[0] in p5 means @1[0] in p6.
16:25 autrijus I'm sure we don't need to keep _that_ behaviour.
16:25 autrijus or at least, I sincerely hope
16:26 Juerd I agree.
16:26 PerlJam autrijus: I think that (1)[0] should DTRT (i.e. parse as if it were (1,)[0]
16:26 Juerd I HAD THEM SEPARATED FOR A SECOND
16:26 Juerd stupid magnets
16:26 autrijus PerlJam: so () normally groups, but if followed by [] or .[], it magically constructs a list
16:27 PerlJam autrijus: yes.  IMHO anyway
16:27 autrijus what about (1){1} ?
16:27 Juerd PerlJam: The consistent way to do that is to say every scalar does Array
16:27 PerlJam that should carp
16:27 autrijus Juerd: right. and then you get 1.[0][0][0] == 1
16:27 Juerd PerlJam: Because () are no longer special syntax when not postcircumfix
16:28 Juerd autrijus: Yes. Interestingly, if $/[0] is the entire match, then $/[0][0][0]
16:28 autrijus Juerd: I'm trying to choose between the three evils.
16:28 Juerd autrijus: Yes. Interestingly, if $/[0] is the entire match, then $/[0][0][0] ~~ $/ much like how 1.[0][0][0] ~~ 1 :))
16:28 autrijus 1. specialise ()[] to parse as (,)[]
16:28 autrijus 2. scalars are singleton lists, so ()[] naturally works
16:28 Juerd Allowing a single value to act like an array is useful
16:28 autrijus 3. make (1)[0] die horribly.
16:29 Juerd But I'm not sure wether it'd be annoying for error reporting perhaps
16:29 osfameron 2 please!
16:29 PerlJam autrijus: without a ruling from $Larry (or @Larry even) #3 is the safest  :-)
16:29 Juerd autrijus: My current preference is 2
16:29 autrijus 42 tests depend on #2.
16:29 Juerd PerlJam: I don't think we should pick safest when safest isn't very useful
16:30 PerlJam Juerd: I agree.
16:30 osfameron what would be the reason for breaking #2 which works in P5 and has some uses ?
16:30 Juerd PerlJam: Larry can overrule anyway, and even his own decisions, and waiting sucks.
16:30 Juerd autrijus: That does kind of indicate it's the expected dwommery
16:30 autrijus ok. anyone want to try summarising this to p6l to ask for confirmation?
16:30 autrijus Corion: ok... 42 more tests pass now :)
16:30 Corion autrijus: :))
16:34 PerlJam osfameron: um ... what's the test case for #2 to work in perl5?
16:34 Juerd grin
16:34 autrijus PerlJam: print ((1)[0])
16:35 PerlJam autrijus: um ...  that supports #1, but not necessarily #2
16:35 * PerlJam scrolls back up to make sure he was reading right
16:35 autrijus PerlJam: you was reading right indeed.
16:36 autrijus try this then
16:36 autrijus print (INC->[0])
16:36 autrijus print('INC'->[0]) too.
16:36 autrijus that is a twisted form of #2 :)
16:36 autrijus (which I agree should die)
16:36 autrijus now the question is, if anything should take its place.
16:37 autrijus so perl5 currently do #1 and a twisted weird form of #2. perl6 can either keep this, or unify them into one of #1/#2/#3.
16:38 PerlJam autrijus: I'll continue to vote #1 (so that when @Larry agrees, I can say "I told you so!"  ;-)
16:38 autrijus PerlJam: do you have five minutes to write this up to p6l? :)
16:39 Juerd I think #2 is the best for the language as a whole
16:39 broquaint has joined #perl6
16:39 PerlJam autrijus: sure.
16:39 autrijus PerlJam++
16:39 Juerd sub foo (@bar) { ... }
16:39 Juerd foo($config<foo>)
16:40 Juerd BEGIN { $config = { map { ...; $key => (/\s/ ?? $val :: $val.split) } ... }; }
16:40 Limbic_Region seen Jonathan
16:40 jabbot Limbic_Region: Jonathan was seen 2 days 14 hours 46 minutes 35 seconds ago
16:40 Limbic_Region seen Jonathan_
16:40 jabbot Limbic_Region: Jonathan_ was seen 7 days 23 hours 32 minutes 5 seconds ago
16:41 Limbic_Region autrijus et all - is there a known list of platforms that Pugs works on?
16:41 Limbic_Region I got a question at the Monastery and would prefer to point to a reference
16:41 Corion Win32 and Linux.
16:41 Corion (and *BSD)
16:41 Corion "Everywhere that GHC compiles to"
16:41 Limbic_Region ok
16:42 Corion (and it assumes that Perl5 is there as well, but Perl5 > GHC)
16:42 Corion GHC 6.4
16:42 autrijus http://haskell.org/ghc/docs/latest​/html/building/sec-port-info.html
16:42 Juerd Limbic_Region: Every platform that can run bochs, vmware, virtualpc, etc :P
16:43 autrijus or, rather, http://haskell.org/ghc/contributors.html
16:44 stevan works on Mac OSX too :)
16:45 Corion stevan: OSX is merely the commercial decline of *BSD :)
16:45 stevan Corion: true dat
16:45 Limbic_Region thanks!
16:46 * Limbic_Region is going to start taking bets as to how long it takes merlyn to stop lurking and start actively contributing
16:46 Juerd Limbic_Region: A few hours
16:47 autrijus merlyn is here? since when?
16:47 Limbic_Region since I wrote the "Get Involved With Pugs" article
16:47 Juerd 15:39 -!- merlyn [merlyn@blue.stonehenge.com] has joined #perl6
16:47 Juerd 15:40 < Limbic_Region> hey merlyn - welcome
16:47 Juerd 15:40 < merlyn> In response to your PM posting. :)
16:47 autrijus wow, 3 hours ago
16:47 autrijus Limbic_Region++
16:48 Limbic_Region [09:36] <merlyn> In response to your PM posting. :)
16:48 Juerd By the way - wtf? 3 hours? Time goes FAST
16:48 Corion r2982 - http://datenzoo.de/pugs/win2k.html - 4757 ok, 122 failed
16:48 Corion He just wants to hang out with all the cool people :)
16:48 stevan Corion: not just cool,.. but uber-cool
16:48 Limbic_Region Corion - that's why I started lurking too - now I am addicted
16:48 stevan Juerd: time moves faster in #perl6
16:49 * Limbic_Region still doesn't contribute that much but every little bit helps
16:49 Limbic_Region it's called time dilation
16:49 Juerd stevan: Apparently!
16:49 Corion It's time we get Junctions working 100%, so we can work in multiple dimensions and get it done in constant time :)
16:49 Limbic_Region it explains a lot ;-)
16:50 Limbic_Region Corion - see http://perlmonks.org/index.pl?node_id=455283
16:50 autrijus iblech: will fix * in a bit
16:52 * Corion forgot about the steamroller operator. Cute operator names are important ;)
16:52 Juerd autrijus, PerlJam: I just thought of ($aref)[0], which kind of rules out #1
16:54 Juerd Unless we really want that to return just $aref, which means () for grouping is no longer working with postfix [], in that (something that returns an array reference)[0] will return the arrayreference, and not its first element
16:54 PerlJam Juerd: yeah, I see that.
16:56 PerlJam I'm not firmly in the camp of wait-and-see-what-@Larry-says  :)
16:56 PerlJam s/not/now/
17:01 Juerd We're all in that camp
17:01 Juerd That doesn't mean we can't have personal preference
17:01 * revdiablo notices the flattening bugs have been magically fixed
17:01 revdiablo woo :)
17:01 * autrijus thinks that "iblechbot" is a fitting nickname
17:02 autrijus slurpy * fixed
17:02 Corion I'm in the camp of "Implement, and see if it survives" :)
17:03 Aankh|Clone has joined #perl6
17:03 BigBear has quit IRC ("Chatzilla 0.9.68a [Firefox 1.0.2/20050321]")
17:03 revdiablo hmm, who's in charge of the debian/ files?
17:03 autrijus scw
17:04 revdiablo the version number is out of date...
17:04 revdiablo I can commit the new changelog to fix it, just don't wanna step on any toes
17:04 autrijus just fix it :)
17:04 autrijus toestepping does not exist with a version control system
17:05 _metaperl has quit IRC (Read error: 110 (Connection timed out))
17:06 Corion ... which reminds me, I should install svk ...
17:06 revdiablo true, true. I'm still not used to it. :)
17:06 autrijus revdiablo++
17:07 Limbic_Region [13:04] <autrijus> toestepping does not exist with a version control system
17:07 Limbic_Region true or not - that is the anarchist in him talking
17:07 _metaperl has joined #perl6
17:07 nothingmuch has joined #perl6
17:08 autrijus :)
17:08 revdiablo autrijus: in your PM reply about the flattening, what's this corner case you speak of?
17:08 revdiablo on p6l somewhere?
17:08 nnunley_ has joined #perl6
17:09 autrijus revdiablo: PerlJam is supposed to write it up soonish
17:09 autrijus revdiablo: it's about the semantics of (1)[0]
17:09 autrijus oh, it's in http://www.nntp.perl.org/gro​up/perl.perl6.language/21044
17:09 Juerd autrijus: That's already been sent by PJ and has already received several responses
17:10 * autrijus mumbles something about accelerated temporal movements
17:10 osfameron has quit IRC ("Leaving")
17:10 * Corion sends the HypnoToad to put autrijus to sleep.
17:10 stevan class Time::Space { method warp { ... } }
17:12 * Limbic_Region hasn't watched TRHPS is some time
17:12 elmex has joined #perl6
17:14 autrijus iblech++ # dealing with the fallout
17:14 gaal has left
17:14 autrijus smoke back to a more respectable 94
17:15 Limbic_Region heh - let's turn failing smoke tests into a golf contents
17:15 ajs that's an old topic. Larry had quite a bit to say about that WAY BACK when the semantics of , and => were first introduced for P6
17:15 Limbic_Region s/contents/contest/
17:15 Limbic_Region person fixing most failing tests resulting in lowest overal score wins
17:15 autrijus ajs: ...yet it's not noted in A/S/E
17:15 autrijus or, afaics
17:16 ajs Let me go find the thread, and perhaps we can short-circut a lot of discussion. I recall it being very informative
17:16 autrijus cool
17:16 autrijus ajs++
17:17 Aankhen`` has quit IRC (No route to host)
17:18 knewt anything since 2982 likely to get this looking better? [[ Failed 33/250 test scripts, 86.80% okay. 115/4873 subtests failed, 97.64% okay. ]]
17:19 autrijus knewt: yes, iblech and I has been repairing them left and right
17:19 autrijus r2987 is markedly better
17:20 autrijus the $1->$0 change is large
17:20 knewt heh, cool. another long build run then :)   although it doesn't take as long as ghc did
17:20 nothingmuch autrijus++
17:21 nnunley has quit IRC (Read error: 110 (Connection timed out))
17:21 knewt my computer chugged away all night building ghc packages for me *g*
17:21 nnunley_ cd
17:22 revdiablo to get the PGE, I need parrot, right?
17:22 autrijus revdiablo: right.
17:22 ajs Some info in http://groups-beta.google.com/group/per​l.perl6.language/browse_frm/thread/54a1​135c012b97bf/98bcf09e00646b1b?q=comma&a​mp;rnum=2&amp;hl=en#98bcf09e00646b1b
17:23 Limbic_Region revdiablo - shame on you - as long as I went on and on about Parrot advocacy in #perl you are just NOW installing it?
17:23 revdiablo Limbic_Region: well, uh, er
17:23 Khisanth heh
17:24 nnunley_ Blah.  Any reason why haskell wouldn't be able to find src/Eval
17:24 autrijus nnunley_: you have old version of hs-plugins?
17:24 revdiablo Limbic_Region: I built parrot a few times, but I couldn't find anything fun to do with it. unfortunately, PASM isn't quite as fun as Perl6 :)
17:24 Khisanth Limbic_Region: so what is this Parrot you are refering to? new pet? ;)
17:24 nnunley_ How old is old?
17:24 autrijus 20050410 or later
17:25 knewt i couldn't get anything later than 20050424 of hs-plugins to build
17:25 DapperDan has joined #perl6
17:25 autrijus I've heard that hsplugins is broken for a few days
17:25 nnunley_ Mine is 20050424
17:25 autrijus I'm on 0417
17:26 * nnunley_ reinstalls hs-plugins, just to be sure.
17:26 * nothingmuch tries to get up
17:26 knewt Limbic_Region: i've built parrot, but currently have no idea how to do a DESTDIR equiv for it, so currently it's not installed and is just being used in building pugs
17:26 nnunley_ Bah.  I think darwinports tryed to help.
17:26 nnunley_ s/y/i/
17:30 ajs This is a better refernece (and the one that I'm recalling) where Larry says that $a=(10) gets 10 because () isn't magial. http://groups-beta.google.com/group/per​l.perl6.language/browse_frm/thread/24ef​8f421548b806/f119fc38427f9f3b?q=comma+o​ne+element&amp;rnum=2#f119fc38427f9f3b
17:30 dakkar has joined #perl6
17:31 autrijus so ok, (10) is not magical. what does 10.[0] do, then?
17:37 ajs 10.[0] should be an error, I would think, since Int.whatever[]is:[] is undefined
17:37 ajs Well, more to the point Any.whatever[]is:[] is undefined
17:38 ajs But (10)[0] behaves according to context, and thus should work as you expect (resulting in 10)
17:39 ajs ()'s (at least back then) were deemed to be very slightly magical in these cases: () - null list; (0) - dependent on context; (LIST) - simple precedence management, no other effect.
17:43 nothingmuch has left
17:44 autrijus (0) depends on context?
17:44 autrijus so (2+3) too?
17:44 ajs yeah, at least it did in 2002
17:44 autrijus as does pretty much all (term) without comma in them?
17:44 clkao autrijus: so are you in london yet?
17:44 autrijus clkao: spiritually
17:45 ajs so @a=(2+3) would be the same as @a=(5)
17:45 ajs hey, this is just me trying to channel old $Larry posts. I'm sure I'm wrong ;-)
17:46 sorje has joined #perl6
17:46 autrijus no, actually that makes some kind of strange sense
17:47 Corion Aieee - svn is unreachable again ;)
17:48 ajs It's as if millions of voices... bah, you know the punch line ;-)
17:48 autrijus yeah. and I'm afraid it's me who killed it ;)
17:48 * autrijus curses NFS
17:48 autrijus oh well. it will be back in, like, 7 hours
17:49 * ajs returns NFS.bless($svn)
17:49 Corion Ah, no worries - I'm not productive today anyway, and that gives me time to focus on other stuff. Like, sleep(). Or outside(). Or cook().
17:49 Corion Mmmm. cook().
17:49 autrijus right. or forme, work()
17:50 Corion Perl6 should get a work() builtin. So we can say work and work and work and die.
17:50 meppl has joined #perl6
17:53 autrijus .lived() und .laughed() ant .loved() end left(); Forsin(). :)
17:53 Juerd Corion: s/die/fail/. Upgrade yourself to 2005 :)
17:54 Juerd Nowadays, it's no exception to fail at work :)
17:54 Corion Juerd: work; fail; next work; fail; work; END { die }
17:54 Juerd Hm
17:54 stevan for (0 .. Inf) { .death and .taxes() }
17:54 Juerd "failure can be an exception"
17:54 clkao svn dead
17:54 clkao see, svk is still useful
17:55 Corion clkao: :)
17:55 stevan but it's written in perl :P
17:56 Limbic_Region Khisanth - old pet actually, I stopped active involvement around late Nov early Dec
17:56 clkao i share the same hate. but what else can you use it to boostrap the development of perl6?
17:56 PerlJam autrijus:, Juerd: Craig DeForest's email puts me over in Juerd's camp now.  
17:56 Limbic_Region revdiablo - pasm isn't quite as fun as Perl6 - are you nuts?
17:57 Khisanth PerlJam: you're justo over there for the food
17:57 Limbic_Region you don't take some kind of perverse pleasure in having an OO assembly language?
17:57 Juerd PerlJam: I have a camp? NEAT!
17:57 revdiablo Limbic_Region: oh, I definitely think it's cool, in theory
17:57 clkao so i can rewrite it in perl6
17:57 stevan PerlJam: watch out for the crazy guy who lives in the cabin on the lake
17:57 autrijus clkao: we have OO now :)
17:57 PerlJam Juerd: yeah, but you need to build a fire and get more tents
17:58 stevan autrijus: speaking of OO, how do i get the invocant in a method?
17:58 Juerd PerlJam: Pugs keeps us warm
17:58 stevan you are burning Pugs?
17:58 Juerd stevan: method ($self: $arg1, $arg2) { ... }
17:58 stevan or using them as tents?
17:58 PerlJam Juerd: friction from all of the patches as they are applied?
17:58 stevan Juerd: that doesnt work currently though
17:58 Juerd PerlJam: It's a warm community.
17:58 Juerd stevan: Does any OO work then?
17:59 stevan Juerd: sorta :)
17:59 Khisanth Limbic_Region: never saw much reason to be excited about Parrot
17:59 Limbic_Region autrijus - WRT what OO support currently exists in Pugs - using two modules with the same methods is no problem right?
17:59 stevan Juerd: I actually started cataloging what works and what doesnt
17:59 clkao why does make test want to link pugs every time?
17:59 Limbic_Region Khisanth - really?  Huh.  It has been producing working code for a long long time now
17:59 Juerd stevan: In the repo?
17:59 autrijus Limbic_Region: I think so
18:00 stevan Juerd: no not yet, I had to do $work, so i got distracted
18:00 Corion clkao: "make test" makes an optimized Pugs. I use "make unoptimized" together with prove.
18:00 * Limbic_Region likes the fact that coroutines in Parrot do what he thinks unlike Perl6's
18:00 clkao oh nnunley told me because it's ingy-ware
18:00 Odin- has quit IRC (Read error: 104 (Connection reset by peer))
18:00 Odin- has joined #perl6
18:00 Limbic_Region ok autrijus - I might port a few more things to p6 now
18:00 Juerd svn borked - great time to try svk
18:00 clkao i meant even if i run make before make test, why does it want to link again
18:00 Juerd Can someone tell me how to checkout?
18:01 clkao only if you have mirrored it with svk?
18:01 Juerd (i.e. the url or whatever it uses)
18:01 stevan autrijus: method invocants?
18:01 Juerd Oh, the server is dead entirely?
18:01 Corion Juerd: I think autrijus killed the NFS
18:01 PerlJam Khisanth: parrot is pretty neat.  I wrote a glob parser for PGE last night.  It went quite smoothly (and I hadn't really written pasm in probably close to a year)
18:01 clkao it shuld be easy for mirroring from anyone who has mirrored
18:02 PerlJam well, I wrote a first-draft glob parser anyway
18:02 clkao (!)
18:02 knewt ok, cool. with 2987 i'm down to 30 failed tests with 82 failed subtests
18:02 clkao it seems so much fun
18:03 autrijus Juerd: I'll give you somewhere else to checkout. a sec
18:03 Limbic_Region problem with Parrot (or at least my problem) was that in addition to learning PASM I also had to improve my non-existant C fu
18:03 autrijus I think PIR is nice enough, btw.
18:03 * Juerd is going to try and get his matrix.p6 working
18:03 * autrijus proceeds to commit some more .pir code into src/pge/
18:04 Juerd Or at least have it build the rule. (64 kB regex in perl 5 - thus much larger even as a Perl 6 rule!)
18:04 PerlJam Limbic_Region: Really?  I didn't think that at all (but then I probably already have the requisite C-fu)
18:05 Limbic_Region PerlJam - I never bothered to develop in C so I didn't really get used to it
18:05 Khisanth Limbic_Region: that fact it can run stuff isn't a reason to be excited about it :)
18:05 PerlJam Khisanth: what would be a reason to get excited about it?
18:06 Limbic_Region different strokes for different folks I guess
18:06 Limbic_Region I guess I joined because I was a bit like merlyn - wanting to hang out with the kool kids
18:06 Limbic_Region and just like Pugs - after seeing what was being done and how fast - I got addicted/excited
18:06 Juerd I guess I joined because. Hm, I don't know. Things started with autrijus asking lots and lots of questions ;)
18:06 Juerd That feels like years ago now.
18:07 Juerd While it has been only months.
18:07 Limbic_Region of course - with the disapearance of Dan and the squabbling on the list about how to proceed WRT Python support - I didn't have a hard time leaving
18:07 PerlJam Juerd: welcome to our time dilation! :)
18:07 Juerd PerlJam: But just a moment ago, time went fast. Now it has returned to slowness again? Oh dear.
18:07 PerlJam Limbic_Region: squabbling on the list?  Wasn't it just sam and leo?
18:07 Khisanth PerlJam: for Parrot in particular? hard to say didn't have much interest from the start and it never really moved
18:07 stevan I started because wanted Perl6, and Ruby|Python|Java|C# was starting to look attractive
18:08 Khisanth PerlJam: the fact that it was able to run everything except Perl for quite a while didn't really help the excitement :)
18:08 stevan Juerd: maybe the svn server is the source of the time-distortion
18:08 Juerd stevan: Are you trying to avoid Ruby? It's a nice language.
18:08 Limbic_Region PerlJam - pretty much those two but it was really anyone and Leo
18:08 stevan Juerd: I like it too
18:08 Limbic_Region no offense to Leo - just a lot of people seemed to disagree with him
18:09 stevan Juerd: but $work has plenty invested in perl and I really didnt want to switch if I didnt have to
18:09 Juerd stevan: I think that during a day, things go fast. But everything more than 24 hours ago feels like three times as long ago as it really was.
18:09 PerlJam Khisanth: parrot was interesting in an academic sort of way to me until I saw how quickly people wrote compilers and how easy it was.  now that we have pugs+pge I find it even more interesting because I now really believe it can accomplish its goal of being the platform for perl6 :)
18:09 Limbic_Region PerlJam - and I am not sure how much of that was a communications problem anyway
18:09 stevan Juerd: doppler effect :)
18:11 stevan Juerd: basically if I have to re-write our codebase, I would rather do it in perl6
18:11 PerlJam Limbic_Region: some communications problems, but also leo's goals seem to be different from other people's goals.
18:11 PerlJam Limbic_Region: well, maybe not goals as much as priorities.
18:11 Limbic_Region exactly
18:11 Limbic_Region in any case - I think chip has things under control
18:12 Limbic_Region seen chip
18:12 jabbot Limbic_Region: chip was seen 32 days 23 hours 5 minutes 56 seconds ago
18:12 Limbic_Region that doesn't sound right
18:12 PerlJam Juerd: you should try staying away from pugs for more than a day and then coming back.  It's like a whole new universe was created while you were away.
18:12 Juerd PerlJam: I have, and it is.
18:12 Juerd PerlJam: Imagine the feeling I had when trying pugs again after weeks.
18:12 Juerd Even though I had been following developments.
18:14 * Limbic_Region still laughs when he thinks about hardburn's excuse for not getting involved with Pugs
18:14 autrijus hm?
18:14 Limbic_Region Him:  "last time I checked, it didn't have feature X"
18:14 Limbic_Region Me:  "When was the last time you looked"
18:14 Limbic_Region Him:  "A couple of weeks ago I guess"
18:15 Limbic_Region IIRC, that feature was OO
18:15 Limbic_Region so now I will have to /msg him to stop making excuses
18:15 Juerd I think it's a good idea to not think of OO as a single feature
18:15 Juerd But who am I
18:15 ajs Juerd: I firmly agree. OO is about 20 features, 3-5 of which are large
18:16 Limbic_Region I shouldn't have used quotes there as it was a paraphrase
18:16 Limbic_Region it is more likely he said "OO support" and not "feature X"
18:16 Aankh|Clone is now known as Aankhen``
18:16 knewt yay, i have pugs installed and running :)
18:17 Juerd knewt: Congrats
18:17 elmex has quit IRC ("leaving")
18:17 elmex has joined #perl6
18:18 knewt Juerd: and down to only 82 failing subtests now :)
18:18 * Limbic_Region wanders off for $work
18:21 nothingmuch has joined #perl6
18:23 knewt heh, running sendmoremoney outputting to the terminal takes 4s for the 2nd calc, whereas redirecting to a file only takes 1s. shows how much the terminal output slows it down
18:25 Juerd For a better test, redirect to /dev/null
18:25 Juerd (NUL in dos)
18:26 knewt ooh, i like the mandelbrot
18:27 autrijus you're going to like it more running with -BParrot :)
18:27 knewt pugs: cannot cast from VError "fnord" (NonTerm (MkPos "sendmoremoney.p6" 1 1 1 1)) to [Char]
18:28 knewt that's trying to run sendmoremoney.p6 with -BParrot
18:28 knewt heh. oh yes. mandelbrot with -BParrot is just a tad faster :)
18:28 autrijus it's but a shadow of the speed of new parrot codegen :)
18:29 autrijus but I'll probably have no time to work on it until the two weeks in Leo's house next month.
18:29 knewt you're going to be at the Austrian Perl Workshop, right?
18:29 autrijus yeah
18:29 autrijus my hope is to get the codegen monad completely unified with evaluator in that two weeks
18:29 autrijus so that whenever we support a new feature in interpreter, it gets reflected in compiler instantly
18:30 knewt i'm (unless something happens) going to be there as well. need to find crash space still though
18:30 autrijus cool
18:30 knewt at worst i could possible crash at my relatives out there
18:30 * Juerd so hates that he has three things he all three wants to attend, around june 10th
18:30 Juerd Which is also my birthday
18:30 Juerd Argh.
18:31 Juerd There's one thing I'm required to go to, making the choice easy.
18:34 nnunley_ has quit IRC (Read error: 111 (Connection refused))
18:34 Juerd Hm - if the thing juni 10th starts in the evening, which I think is likely, I can perhaps be crazy and drive to austria the 8th and back the 10th
18:35 Juerd I wonder how long a drive it is
18:35 Corion Juerd: Consider flying ;)
18:35 Juerd 10 hours - no, that's madness
18:36 Juerd Corion: Isn't that very expensive?
18:36 * knewt is going by train. 12 hours worth of travel each way. ouch.
18:36 Corion Juerd: Err, yes. Well, actually, it might be on par with the train. But still very expensive, yes.
18:37 Corion Unless you find something like RyanAir, or EasyJet or whatever.
18:37 Aankhen`` Is the SVN server at openfoundry.org down?
18:37 autrijus yes. I'm switching main repo for the next 7 hrs
18:37 autrijus a sec.
18:37 Aankhen`` Alright.
18:38 Aankhen`` What's the URI for the new/alternate repository?
18:38 autrijus it's the "mirror" listed in pugscode.org main page
18:39 Aankhen`` OK.
18:43 Juerd Corion: None of the cheap ones go to austria
18:43 Juerd hm, 600 euros. No, sorry, not this year :)
18:49 Juerd Only 300 if I leave one day earlier
18:49 Juerd Weird
18:49 knewt ouch. i think the 190 euros to get from hamburg -> wien and back by train is bad enough
18:49 Juerd I haven't looked at train options yet
18:50 autrijus changes topic to: Fallback repos: http://svn.perl.org/perl6/pugs/trunk/ | pugscode.org <Overview Journal Logs> | pugs.kwiki.org
18:51 autrijus can someone test if it builds clean there?
18:51 autrijus esp with PGE
18:51 autrijus I've deleted PGE-Hs.pbc and did lots of fixes
18:51 autrijus ("there" means the fallback repo)
18:52 * knewt wonders if he can svk mirror --relocate to the fallback
18:52 clkao if they were mirrored from each other, then yes
18:53 knewt bah. "uuid is different at ..."
18:54 clkao hate
18:54 autrijus just re checkout :-/
18:55 clkao autrijus: how was the mirror created?
18:55 clkao mirroring or smerging?
18:56 autrijus clkao: smerging
18:56 clkao then you cna't mirror --relocate. bt but you can just mirror that too
18:57 autrijus that's what I did.
18:57 * clkao <- talks to knewt
18:57 dakkar has quit IRC ()
18:57 knewt hmm. i guess don't pull the history of the fallback then
18:57 autrijus yeah. just the head
18:57 knewt it'll all get merged back into the main repo when that's back up, right?
18:58 autrijus sure.
18:58 autrijus sm -I --verbatim
18:58 knewt cool
18:58 autrijus "What is the sound of SVK? Isn't it a wall that people stopped banging their head to?"
18:59 * PerlJam hasn't yet used svk and doesn't recall banging his head upon a wall
18:59 autrijus so I need to sleep. patches go to p6c as usual.
18:59 PerlJam (but I guess that might be the braindamage ;-)
19:00 clkao i used svn and cvs today and did bang my head
19:00 PerlJam What were you using old obsolete technology like cvs for?  ;-)
19:01 PerlJam autrijus: wait, don't sleep.
19:02 knewt ok, building
19:02 autrijus PerlJam: eh?
19:02 PerlJam autrijus: my compile just finished and I get a fingerprint mismatch when trying to do pugs -e 'my $a = "foo"; $a ~~ /o/'
19:02 PerlJam (yes, it's parrot 0.2.0 this time :-)
19:03 PerlJam arrot VM: Can't unpack packfile run_pge.pbc.
19:03 autrijus PerlJam: you have a src/pge/PGE-Hs.pbc?
19:03 autrijus or a run_pge.pbc ?
19:03 autrijus try svn up again (from the fallback repo)
19:03 autrijus they should be gone
19:03 autrijus like, excised
19:04 PerlJam autrijus: still run_pge.pir and no src/pge/PGE-Hs.pbc
19:04 Limbic_Region autrijus - http://perlmonks.org/index.pl?node_id=456112 # request for pugs/perl6/parrot articles for TPR
19:04 DapperDan has quit IRC ()
19:05 lumi has quit IRC (Read error: 104 (Connection reset by peer))
19:05 autrijus PerlJam: .pir is fine
19:05 autrijus PerlJam: I wonder why it ran run_pge.pbc
19:05 autrijus PerlJam: there should be no mention of run_pge.pbc anymore
19:05 autrijus PerlJam: does it still break?
19:05 PerlJam Now it says Parrot VM: Can't locate PGE-Hs.pbc, code 2.
19:06 autrijus ok, much better.
19:06 autrijus fixing
19:06 nothingmuch poor dogs
19:06 stevan Limbic_Region: hmmm, a regular pugs column
19:06 autrijus hey, fix run_pge.pir by hand and test it for me?
19:06 nothingmuch why can't they invent silent fireworks, eh?
19:06 autrijus PerlJam: that line should read
19:06 autrijus    load_bytecode "PGE/Hs.pir"
19:06 autrijus that should be all. no recompile needed
19:07 autrijus see if it magically works
19:07 PerlJam nope. Parrot VM: Can't locate PGE/Hs.pbc, code 2.
19:07 PerlJam oh!
19:07 PerlJam sorry, hold on.
19:07 PerlJam ha, much better
19:07 PerlJam er, ah
19:07 DapperDan has joined #perl6
19:08 * PerlJam needs to have his eyes checked or there need to be fewer p-things relating to perl6/pugs  ;-)
19:08 PerlJam Anyway, thanks autrijus!  go get some sleep.
19:08 autrijus so it works?
19:08 PerlJam aye
19:08 autrijus woot.
19:08 autrijus ok, committing back
19:10 PerlJam It sure would be neat if there were some CPAN-a-like for svn repositories such that if one were down, you'd automatically get sent to the next one
19:11 integral like freepan?
19:11 PerlJam freepan does svn?
19:12 PerlJam and mirroring with the smarts to take you to a mirror that's up?
19:13 nothingmuch has quit IRC ()
19:13 autrijus zzz. :) & *wave*
19:14 PerlJam integral: ah, *exactly* like freepan probably. (looking at the FAQ now)
19:14 PerlJam I'd been ignoring freepan as I'd other things nearer the surface to deal with.
19:14 * integral thinks the smarts should be client side
19:15 Juerd Damn. Why is YAPC::Europe all the way in Portugal?
19:15 integral the server shouldn't need to know that you prefer JANET and Demon over china
19:15 Juerd I want to meet other Perl people, but this way this isn't going to happen any time soon
19:16 Corion Juerd: Consider flying there ;) I'm "planning" (heh) a two-week vacation down there ...
19:17 Corion Juerd: Also, consider the Dutch Perl Ws, UK Perl Ws and maybe German Perl Ws :)
19:17 Corion (or French Perl Ws)
19:17 Juerd Corion: I can never stay anywhere too long
19:17 Juerd Corion: I went to both dutch perl workshops, but I want more :)
19:17 Corion Juerd: Hmmm - that means you'll have to live with either quick travel or cheap travel, and lots of it.
19:18 Juerd Corion: Things like Vianna and Braga are too expensive to travel to
19:19 Corion Juerd: Yeah, they're both expensive to travel to
19:19 knewt even from north germany travelling to vienna is expensive
19:19 Corion Juerd: Consider browsing through http://megaflieger.de/ (or something similar)
19:19 Juerd Corion: I have visited that kind of sites
19:19 webmind Juerd, organise a perl thingie at wth? :)
19:20 Juerd Corion: From .nl, you can get everywhere, as long as it's the united kingdom.
19:20 Juerd Or Prague, or Kopenhagen
19:20 Juerd That's about it.
19:20 integral .nl -> .uk -> .fr -> .pt
19:20 Juerd Everything else is very expensive.
19:20 webmind buses are cheap.. just horrid to travel in
19:21 Juerd webmind: And they never go the days you want to be somewhere
19:21 webmind darn
19:21 knewt for going to the uk on holiday my dad usually drives, going via .nl
19:21 webmind and hitchiking is even more unstable :)
19:21 Juerd webmind: I'd be happy to organize a Perl sub-event
19:21 Corion Juerd: Or ask on the YAPC mailing list if anyone from (or near) .nl is driving down ...
19:21 Juerd webmind: If it's the first day of WTH
19:22 webmind Juerd, well if you can find enough  people you can open a village ?
19:22 Juerd I don't grok the idea of the villages.
19:22 webmind Juerd, what was it btw why you couldn't go the other days ?
19:22 webmind grok ?
19:23 Juerd webmind: I'm going to france the night between the 2nd and 3rd days
19:23 webmind ah k
19:23 webmind Juerd, the idea with villages is to create ubs-event
19:23 webmind sub-events
19:23 webmind but kinda specialised
19:23 webmind such as a wireless village
19:23 Juerd webmind: Yes, but 4 day sub events. It groups people in ways that I think is not good for an event like wth.
19:24 Juerd BOF sessions are more useful.
19:24 webmind well the idea is localisation
19:24 webmind BOF sessions with 3000 people ?
19:25 webmind that'll get big
19:25 Juerd The problem with villages is that you can't easily be in multiple
19:25 webmind the advantage of villages is it takes load of the organisation.. which wants to do as little as possible
19:25 webmind hmm
19:25 Juerd Yea, that much is obvious.
19:26 webmind well I think you can.. but yes.. you have to place your tent somewhere
19:26 webmind but you can visit events at multiple villages
19:27 webmind although I think 4 days is a bit short
19:28 Juerd It is, for something of this magnitude, where everyone has almost equal wishes
19:28 webmind actually.. lots of people don't
19:28 webmind some people are just interested in security and maybe some code hacking
19:28 Juerd I think a simple multi-track presentation scheme works better than this village thing
19:28 webmind others more in creative issues
19:29 webmind and some in legal issues regarding privacy and human rights
19:29 Juerd webmind: I suspect most of the people who come to WTH care about all these things.
19:29 webmind well there will be a multi track prtesentation scheme aswell
19:29 Juerd Or at least most of these things.
19:29 webmind well not everyone is interested in everything.. as for what the hack atleast
19:30 webmind I'm not carring much about C code or buffer overflows
19:30 webmind other don;t care about wireless things or voip
19:30 webmind and some don't care about legal or rights issues
19:30 webmind I wish they did thoguh
19:31 Juerd We have a very different view of the people who visit wth
19:32 webmind well maybe more diverse for me
19:32 ninereasons has joined #perl6
19:32 webmind because I know there will be loads of cool people
19:32 knewt heh, pugs really manages well to push most of firefox out into swap during build-time :/
19:32 webmind just not everyone
19:32 webmind atleast.. that was for hal2001
19:32 Limbic_Region stevan - you interested in writing one?
19:33 stevan Limbic_Region: maybe
19:33 webmind Juerd, but we'll see how the village things turns out
19:34 Limbic_Region TPR is 1 x month right?
19:34 stevan i think
19:35 Limbic_Region actually http://www.theperlreview.com/
19:35 Limbic_Region it looks quarterly
19:35 stevan oh even bettet
19:35 stevan s/bettet/better/
19:35 Limbic_Region given that pugs has only been in existance for a little over 1 quarter - doesn't seem like an overwhelming task
19:35 stevan yeah
19:35 stevan although it may not make sense
19:35 stevan article 1 - we started
19:35 stevan article 2 - we are done
19:36 stevan to much time in between :)
19:36 Limbic_Region well - we are going to seriously run into the "we caught up to you @larry - waiting on further instructions"
19:36 Limbic_Region looking at the Perl6::Bible - there are still a lot of chapters left to be spec'd out
19:37 stevan Limbic_Region: that happens on a daily basis sometimes
19:37 * Limbic_Region was sorely disapointed with the result of his coroutines question
19:37 Juerd Limbic_Region: We already are
19:37 hlen has joined #perl6
19:38 Limbic_Region http://perlmonks.org/index.pl?node_id=332117 # Is a great reference guide
19:39 Ontolog has joined #Perl6
19:40 PerlJam pugs only existing for such a short time is as it should be.
19:40 PerlJam Design is the hard part and should take the bulk of the time.
19:40 Corion has left
19:40 PerlJam Implementation is just a small bit of time.
19:40 PerlJam And then Maintenance lasts forever :)
19:40 * knewt giggles at FreepanContentRequirements: [[ Only Python modules under a proprietary licence are allowed. GPL and BSD licences are particularly forbidden. ]]
19:41 Limbic_Region PerlJam - I agree though there is more to the story
19:41 Limbic_Region separating the VM was a design decision that resulted in a very long implementation process
19:42 Juerd The vm had to be rewritten anyway
19:42 Limbic_Region it is just that the piece that many people focus on is the language
19:42 Juerd Separated or not
19:42 Juerd Sure, it's more work this way
19:42 Juerd But don't think it would have been easy if it targeted *only* Perl 6
19:42 integral hmm, considering the GHC's VM is much, much older than parrot,  pugs has had a pretty long implementation too :-)
19:42 Juerd Easier, not easy.
19:43 Limbic_Region what I mean Juerd is that there is a great deal of people not aware that implementatation has been going on for a long time
19:43 Juerd You have to admit that parrot development isn't quite as rapid as Pugs
19:43 Limbic_Region no - that's not what I am saying at all
19:44 Limbic_Region I am replying to PerlJam's comment that implementation is just a small bit of time
19:44 Juerd Implementation of the language is
19:44 Limbic_Region in many respects I agree, but there is more to the story
19:44 Juerd Parrot is not Perl 6
19:44 Limbic_Region err - my last comment is to my last statement not to yours
19:45 Limbic_Region and I would say "but p6 wouldn't exist without Parrot" - but I would be dead wrong wouldn't I
19:45 Limbic_Region ;-)
19:45 Juerd Yes, you would
19:45 * Limbic_Region hugs his pugs
19:45 Juerd Perl 6 is a language, not a vm, not a program, not an library
19:46 Limbic_Region I don't believe anyone here is disputing that
19:46 Juerd So the people ignoring parrot are right
19:46 Juerd Parrot is important for perl6, eventually
19:46 Juerd But it's a sepearate project for many reasons
19:47 PerlJam Juerd: eventually being now.
19:47 Juerd And implementation of Perl 6 really has began only recently.
19:47 Juerd PerlJam: Yes, because of PGE.
19:47 PerlJam yep
19:47 Limbic_Region hmmm - it is obvious that text only communication has a way of losing unspoken meaning
19:47 Juerd Limbic_Region: You'll get used to it.
19:48 PerlJam Perl6 implementation has only really been going on since Dec 2004 or so
19:48 Limbic_Region Juerd - I don't think you got the conversation is my point
19:48 Limbic_Region stevan made a mostly joke
19:48 Limbic_Region PerlJam commented on it
19:48 Juerd afk
19:48 Limbic_Region I commented a little bit more seriously
19:49 Limbic_Region and then we started down the path of disagreement
19:49 Ontolog has left "Leaving"
19:51 knewt autrijus: ok, i don't know if the number of tests/subtests has gone up or not, but i don't have any more/less failures from the fallback 2968 than i did from main 2987. and i don't have an "unexpected success" message any more
19:52 stevan knewt: we always seem to be hovering around 70-100 failing tests
19:52 stevan lately that is
19:52 stevan it fluctuates alot
19:52 knewt Failed 30/250 test scripts, 88.00% okay. 82/4873 subtests failed, 98.32% okay.
19:52 knewt why does it relink during the make install?
19:53 stevan knewt: no idea, I hardly ever do a full make test anymore
19:53 stevan takes too long :)
19:54 knewt building is my main bit. not as bad as building ghc was though :)
19:54 knewt my computer chugged away all last night building ghc packages for me
19:54 stevan pre-built binaries ++
19:56 knewt i didn't want to put sid binaries onto hoary, so did a binary install of 6.2, and then pulled in the source package for 6.4 from sid and built that, then upgraded to the newly built packages
19:56 Southen has quit IRC (No route to host)
19:57 integral did you check out with svk, knewt?
19:58 integral If so, there may be something which regenerates the svn revision number that runs that makes the relink happen
19:58 knewt yep. well, svk.
19:59 integral that's probably why
19:59 knewt it's annoying as the link stage takes aaaaaaaaages
19:59 integral I suppose Makefile.PL shoudl detect what kind of repository the code is from
19:59 * integral is scared of Module::Install
20:00 knewt the build of pugs kindly pushes firefox out into swap for me *g*
20:00 DapperDan has quit IRC ()
20:00 integral only firefox? ;-)
20:01 knewt most everything else i run is quite small in memory usage, so although it /might/ get swapped out it isn't that noticeable
20:01 * integral normally sees GHC with 200MB vmem and 100MB resident
20:01 nnunley has joined #perl6
20:02 knewt firefox on the other hand, once the pugs build has been going for a few minutes, often takes 10+ minutes for enough to be swapped back in whilst the build is still going
20:04 integral are you on less than 256M of ram?   I can manage to do other stuff at that if I nice the build a bit
20:04 knewt 512M ram, 512M swap
20:05 integral oh, you're doing an optimized compile, right?   /me forgot about that
20:05 knewt firefox is rather low at the moment with only having a 217virt/102res footprint
20:06 Juerd If anyone wants to update my XML grammar generator and then try it out with our new rules support, please go ahead
20:06 Juerd Don't wait for me to do it, I won't have the time.
20:06 knewt but then i only have 21 firefox windows open right now, and i only restarted it 21 hours ago
20:06 ajs I've tracked down a bug to src/Pugs/Compile/Parrot.hs, I don't know Haskell, so can't figure out exactly where. tempPMC is being called one too many times on storing the result of a subroutine call, but that's as far as I can get.
20:08 integral possibly it's the compileWith on line 231?
20:08 Juerd http://perlmonks.org/index.pl?node_id=179755
20:08 Juerd (xml grammar)
20:08 Limbic_Region knewt - 21 windows open?  How many tabs in each window?
20:08 Juerd (needs to be updated to current spec, and tested)
20:09 Juerd It's written in 2002's version of perl 6 rules
20:09 knewt Limbic_Region: 1
20:10 knewt Limbic_Region: i have tab support completely disabled
20:10 Limbic_Region huh what?
20:10 hlen has quit IRC ("use vim")
20:10 Limbic_Region I don't understand knewt
20:11 * Limbic_Region assumes some good explanation
20:11 knewt i use a tabbing window manager, so i've completely disabled firefox's internal tabs
20:11 knewt means i don't have to use different keys to move between web pages than between any other window on my system
20:11 Limbic_Region knewt - I bet your tabbing window manager doesn't have an "undo close window" feature
20:12 Limbic_Region or a "disable JS in this window only" feature
20:12 Limbic_Region *shrug* - whatever works for you I guess
20:12 autrijus ajs: fixed. committing.
20:12 knewt no, but i can cope with that. the uniformity of key control is worth it
20:12 autrijus also fixing utf8ness in matches.
20:12 stevan Juerd: if you write that grammer,.. I will write the XML::TreeBuidler part :)
20:12 autrijus ajs++ # tempPMC is indeed called once too many
20:12 stevan autrijus: coding in your sleep again :)
20:12 knewt autrijus++ # can't possibly be a single person, much too productive
20:12 Limbic_Region autrijus - Larry has spoken on the single item list thread
20:13 Juerd Yeah, with a really weird reasoning
20:14 Juerd And ignoring entirely that #1 rules out (something returning arrayref)[0]
20:14 Juerd But -- I'm refraining from responding
20:14 Limbic_Region Juerd - why?
20:15 autrijus Juerd: I think larry is mistaken somehow
20:15 Juerd I have my reasons, that I wish to not discuss at this moment
20:15 Limbic_Region okie dokie
20:15 autrijus I don't see how #2 leads to (@foo)[0] being different when @foo is (1) or (1,2)
20:15 autrijus either way it will be @foo[0]
20:15 PerlJam Juerd: see? I told you @Larry would side with me!  ;-)
20:15 * PerlJam reads p6l
20:16 knewt autrijus: i think he's seeing it as (@foo)[0] producing a list of length 1 containing @foo, and then returning that item
20:16 Juerd autrijus: Yes.
20:16 knewt but i could be wrong
20:16 autrijus knewt: if so, he meant "can't be #1";)
20:17 Juerd autrijus: I think Larry lives in a weird Perl 5-ish world, where () indicates a list, and thus list context, and thus has flattening.
20:17 Juerd autrijus: But otoh he can't possibly be living there.
20:17 PerlJam Juerd: can't be
20:17 jhorwitz has quit IRC ("Chatzilla 0.9.67 [Firefox 1.0.3/20050414]")
20:18 Juerd PerlJam: ?
20:18 PerlJam Juerd: re Larry living in a weidr perl5-ish world.
20:18 autrijus (I replied)
20:18 PerlJam Juerd: well, it's time for you guys to step up and argue with Larry  :-)
20:18 Juerd PerlJam: Not me, not now.
20:18 PerlJam As apparently autrijus just did
20:19 wolverian shouldn't (@foo)[0] return @foo?
20:20 Juerd No
20:20 theorbtwo No.
20:20 wolverian why not? :) I mean. that's just the logical thing to me.
20:20 theorbtwo [@foo][0] would.
20:20 Juerd @foo[0] and (@foo)[0] must at all times be the same, because () *merely groups*
20:20 Juerd theorbtwo: () groups, it doesn't DO ANYTHING, other than mess with precedence.
20:20 wolverian right. is this what larry's post fights against?
20:21 theorbtwo Juerd: Right.
20:21 wolverian (making ()[0] mean (,)[0]?)
20:21 Juerd theorbtwo: But -- #1 implies a change to this simplicity, which I strongly oppose to.
20:21 wolverian this is python all over, except I guess one should just use [][]
20:21 Juerd wolverian: No, Larry is actually blessing that "solution".
20:22 Juerd It adds a HUGE special case
20:22 wolverian Juerd: that's what I meant. 'fights against' () being just grouping there.
20:22 Juerd Making some things terribly hard to do.
20:22 wolverian s,grouping,precedence,
20:22 Juerd wolverian: He'd be fighting against one of his own strong points of what Perl 6 should be like...
20:22 Juerd I still don't comprehend Larry's response.
20:22 Juerd My current thinking is that his "#1" is a typo, and he means "#2".
20:23 wolverian or #3.
20:23 PerlJam He definitely does not mean #3
20:23 wolverian why not?
20:23 Juerd No, nobody means #3
20:24 PerlJam because it's wrong
20:24 Juerd #3 is incredibly stupid.
20:24 wolverian okay.
20:24 wolverian it's logical to me, if there is no List there anywhere.
20:24 Juerd #1 is less stupid, but only #2 really makes sense
20:24 knewt Juerd++ # [[ ...so it isn't really useful. Let's turn it into something even less useful, I thought. ]]
20:24 wolverian then (1)[0] would be the same as 1[0]
20:24 PerlJam wolverian: I was of the #1 camp,  but I've been convinced that #2 would be most useful.
20:24 Juerd wolverian: Yes!.
20:24 wolverian Juerd: and that would be an error!
20:24 knewt i'm in the #2 camp i think
20:24 theorbtwo #3 is sounding better and better to me.
20:25 stevan I like #4 ;)
20:25 wolverian #3 sounds like the only logical choice, given that () merely sets precedenec
20:25 wolverian precedence..
20:25 theorbtwo Can't take an array element of a scalar!
20:25 Aankhen`` has quit IRC ("Induhvidual quote: "It's best not to open that can of wax." [Time wasted online: 3hrs 22mins 16secs]")
20:25 rafl has quit IRC (Operation timed out)
20:25 wolverian this is however a special case of (1)[0], and (@foo)[0] would certainly work.
20:25 wolverian (that is, the dying thing. :)
20:26 * Limbic_Region calls it a night
20:26 Limbic_Region has quit IRC (""a night"")
20:26 Juerd wolverian: And with () setting precedence, why would ($aref)[0] be illegal?!
20:27 Juerd circumfix (non-postcircumfix) () shouldn't have any special case in my very un-humble opinion.
20:28 theorbtwo Juerd, I only meant to imply a special case in the error message.
20:28 Juerd Oh
20:28 Juerd But it shouldn't die
20:28 theorbtwo "Method postcircimfix:«[ ]» not defined on Int" might be confusing for newbies who said (1)[0].
20:29 Juerd It's stupid to disallow precedence grouping
20:29 Juerd You must realise that [] is the same thing as .[]
20:29 PerlJam theorbtwo: they can RTFM to better understand that error  :-)
20:29 Juerd And thot you want to be able to call (somesub arg, arg, arg).[0]
20:29 Juerd Which equals (somesub arg, arg, arg)[0]
20:29 theorbtwo Of course, I do rather wonder why a newbie would write (1)[0].
20:30 theorbtwo Yes, Juerd, I do.
20:30 PerlJam theorbtwo: But the point is that it shouldn't error, but rather return the single scalar value
20:30 Juerd Really only #2 allows this to work well
20:30 theorbtwo If people wanted that question answered, they should have asked it, and not (1)[0].
20:30 wolverian Juerd: it wouldn't. who said it would?
20:30 G2 has joined #perl6
20:30 Juerd wolverian: What wouldn't do what?
20:31 wolverian Juerd: ($aref)[0] wouldn't be illegal.
20:31 wolverian (sorry, girlfriend took a bit of my attention :)
20:31 theorbtwo $aref knows how to postcircimfix:«[ ]».
20:31 Juerd wolverian: If ($foo)[0] is made illegal, then ($aref)[0] is illegal too. This isn't something you can defer to runtime
20:32 Juerd Well, you could
20:32 theorbtwo That's exactly what I suggested, Juerd.
20:32 wolverian Juerd: no! no! the #3 to me says that (1)[0] dies *because it means 1[0]* which is totally undefined
20:32 Juerd But that'd suck greatly.
20:32 PerlJam theorbtwo: given @a = (1,2,3), what should @a[0][0] do ?
20:32 Juerd wolverian: (1)[0] or ($foo = 1)[0] -- there must be no difference
20:32 broquaint has quit IRC (Read error: 60 (Operation timed out))
20:32 Juerd wolverian: (1)[0] or ({ 1 }.())[0] -- again, same thing
20:32 theorbtwo Die, PJ, when trying to take the first element of 1.
20:32 Juerd wolverian: You can't know until runtime what the LHS will be
20:33 Juerd wolverian: And having this error out during runtime kind of sucks if it can also do something rather useful.
20:33 knewt Juerd: i'm not a grammer person; do you know what this comment means? [[ The comments in the text of the spec are normative, not the grammar ]]
20:33 Juerd knewt: I have no idea
20:34 integral it means the words are more important than the code samples.   Believe the words, the code might be wrong
20:34 PerlJam knewt: The comments are what you should pat attention to.
20:34 PerlJam s/pat/pay/
20:35 knewt oh /that's/ a useful grammer then
20:35 Juerd knewt: No, not really. I said so in my message to p6l.
20:37 wolverian Juerd: yes, I agree.
20:37 wolverian Juerd: that is; this must be a runtime decision.
20:37 wolverian I guess #3 argues against that. I read it wrong. sorry.
20:37 Odin- has quit IRC (Read error: 104 (Connection reset by peer))
20:38 Juerd wolverian: Well, then you disagree, because I said it's better a compile time thing.
20:38 Juerd afk
20:38 Odin- has joined #perl6
20:38 wolverian oh. then I misread again!
20:38 wolverian damn, I'm tired I guess.
20:38 * PerlJam wonders if the dbi-in-parrot thing is still in the works.
20:42 theorbtwo Is there any concise, current definition of how context propigates in p6?
20:42 stevan Juerd: FYI
20:42 stevan What works in OO right now: classes can be created; attributes can be created (public and private); smart match against class (ex: $foo ~~ Foo); .isa()
20:42 stevan What does not work on OO right now: inheritance; setting the default value of attributes; setting traits for attributes (ex: has $.attr is rw); accessors are not autogenerated for attributes; .can()
20:42 theorbtwo I feel like I'm missing something in this discussion.
20:42 stevan but that was my list from this morning, so it may have changed (rebuilding right now)
20:43 theorbtwo Why would (foo arg, arg)[2] not work under any of the proposed plans?
20:43 theorbtwo (Esp plan 3.)
20:44 Shillo has joined #perl6
20:45 * Shillo waves. SVN dead?
20:45 theorbtwo What's wrong with saying that the invocant of postcircimfix:«[ ]» is always in List context?
20:45 stevan Shillo: see the topic
20:46 knewt Shillo: see topic
20:46 stevan switched to perl.org for the next ~7 hours
20:46 Shillo stevan: Just did. Problem is... how to switch to mirror?
20:46 stevan with svn right?
20:46 * Shillo nods.
20:46 autrijus theorbtwo: what about {} ?
20:46 stevan svn switch --relocate http://svn.openfoundry.org/pugs http://svn.perl.org/perl6/pugs/trunk/
20:47 autrijus theorbtwo: and in any case, how do you mandate slurpy context on an invocant again?
20:47 stevan and then just reverse it when it is back up
20:47 Shillo stevan: Woah. That one is cool. :)
20:47 autrijus theorbtwo: fwiw, you're talking about line 496 in Eval.hs
20:47 autrijus        varVal  <- enterLValue $ enterEvalContext (cxtItem "Array") listExp
20:47 Shillo svn: The repository at 'http://svn.perl.org/perl6/pugs/trunk' has uuid '3eb29b82-81e5-0310-bad4-b280dcad658b', but the WC has 'c213334d-75ef-0310-aa23-eaa082d1ae64'
20:47 Shillo Hmm.
20:47 autrijus if you change that (cxtItem "Array) into cxtSlurpyAny
20:47 autrijus Shillo: you need to checkout again
20:47 stevan Shillo: sorry
20:48 Shillo Oif.
20:48 autrijus theorbtwo: then it will work as you described.
20:48 stevan Shillo: what autrijus said
20:48 autrijus theorbtwo: the question is, why are we doing this?
20:48 autrijus isn't the invocant a single lvalue?
20:49 stevan autrijus: did clkao give you an endorsement deal :)
20:49 stevan The svk team is pround to introduce svk 1.00, a distributed version control system that powers Autrijus' productivity. While you might not be as crazy as Autrijus, you can at least use the same tool.
20:49 integral sub abc(*@foo: *@bar) just wrong?
20:49 * Shillo will wait till tomorrow, then. :) Hmm, another thing, got a funny parting bug to report.
20:49 autrijus stevan: he did ask me whether he can write that :)
20:49 PerlJam heh
20:50 autrijus theorbtwo: it's worth asking on p6l, though. I'm not entirely sure.
20:50 stevan autrijus: I hope you at least got a free copy of svk out of it ;)
20:50 autrijus theorbtwo: as it stands, (1,(2,3),4)[3] returns 4.
20:50 autrijus stevan: lol
20:50 Shillo autrijus: That's correct, though
20:50 autrijus er I mean
20:50 autrijus (1,(2,3),4)[2] returns 4.
20:50 autrijus bad offbyone.
20:51 Shillo Oh.
20:51 autrijus because it's taken to mean [1,[2,3],4][2]
20:51 knewt ah
20:51 * Shillo found another one, lemme reproduce...
20:51 knewt == 4 sounds right
20:51 autrijus because .[] imposes CxtItem on lhs
20:51 autrijus I'll ask on p6l.
20:51 Shillo (1, 2, 3) >>+<< (1, 2, 3); # prints 3
20:51 Shillo Er, prints (1, 2, 3), which is correct
20:51 Shillo But...
20:52 Shillo my @a = (1, 2, 3);
20:52 Shillo @a >>+<< @a;
20:52 Shillo *** Error: Hyper OP only works on lists: Val VUndef at <interactive> line 1, column 1-12
20:52 putter has joined #perl6
20:53 * Shillo chatted a bit with pugs REPL yesterday. :)
20:53 putter autrijus: ping
20:53 autrijus putter: pong
20:54 autrijus Shillo: test case
20:54 * Shillo would commit a test but can't (for both possible reasons, no bit and no repo :) )
20:54 autrijus not an excuse. send to p6c :)
20:55 putter kudos should probably go to ghc's simonmar for 1 day turnaround on the ghc gc bug which was crashing pugs.  Caveat - I haven't tested it yet (cvs-ghc isnt building on amd64 just now).  Expectation is it will be in tomorrow's cvs snapshots.  We'll see.
20:55 theorbtwo Also, what's your email address.
20:55 theorbtwo Shillo +|= $committer
20:56 autrijus putter: whoa.
20:56 autrijus putter++ # nice.
20:56 * autrijus waits eagerly for ghc 6.4.1
20:56 * putter too.
20:56 autrijus Shillo: "Miroslav Silovic"?
20:56 Shillo Miroslav Silovic <miro@puremagic.com>
20:57 Shillo Note, I can't simultaneously read mail to that address and commit (one is from work, commits from home. 'Complicated' ;) )
20:57 autrijus that's fine.
20:57 autrijus the email is just to send invitation (and lost password) to
20:57 autrijus please still send test case to p6c
20:57 autrijus will you? :)
20:58 Shillo Sure.
20:58 Shillo (will from Google)
20:58 autrijus google groups?
20:59 autrijus posting to google groups is an excellent way to ensure nobody can read it :)
20:59 autrijus probably need to write to perl6-compiler@perl.org, or nopaste it here and somebody else will commit.
20:59 autrijus but anyway... here you go
20:59 autrijus * Hyper operators now works on arrays too. Reported by: Miroslav "Shillo" Silovic
20:59 autrijus r2972
20:59 Shillo No, will mail from Google email address. :)
21:00 autrijus oh. good :)
21:00 Shillo I know about one-way forwarding.
21:00 Shillo Thanks! :)
21:00 autrijus :)
21:00 Shillo Oh. Regression test case, then...
21:01 autrijus yeah.
21:03 DapperDan has joined #perl6
21:03 autrijus ahh. luqui is smart and I am Wahnsinnig
21:03 Shillo Oh, @a >>+<< 1 works, too?
21:04 autrijus should be, but I don't know, write a test :)
21:05 putter i saw the off-by-one $/[1] vs $1 offset is gone.  yay!
21:05 stevan autrijus: pugs -e 'class Foo::Bar {};'
21:05 stevan *** Error: No compatible subroutine found: "&class" at -e line 1, column 1-18
21:05 stevan I will write a test for it
21:06 stevan autrijus: should I put it in t/pugsbugs or t/oo ?
21:07 autrijus stevan: either way
21:07 stevan ok
21:08 putter is it just me, or does array aliases flattening capturing parens and capturing non-capturing parens seem unfortunate?
21:09 autrijus putter: you lost me. code?
21:09 putter (i appreciate it makes array aliases on subrules nicely consistent)
21:09 kolibrie has quit IRC ("leaving")
21:11 putter "xy" ~~ ... @<a>:=(.(.)) vs @<a>=[.(.)]  yields @<a>[0] eq 'y'  vs @<a>[0] eq 'xy'.
21:11 putter err, no.
21:12 putter er, yes.  sigh.
21:12 merlyn what's the svn command to get PUGS ?
21:12 autrijus merlyn: for the next 4 hours:
21:12 autrijus svn co http://svn.perl.org/perl6/pugs/trunk/
21:12 stevan autrijus: patch of t/oo/ test sent to p6v
21:12 autrijus after that:
21:12 autrijus svn co http://svn.openfoundry.org/pugs/
21:13 merlyn what happens if I do the latter one now?
21:13 autrijus merlyn: likely timeout
21:13 stevan nothing :)
21:13 merlyn OK
21:13 theorbtwo Then it'll time out.
21:13 merlyn I'll just wait until tomorrow then. :)
21:13 Shillo Mail sent. Hope the test itself isn't buggy. :)
21:13 autrijus merlyn: nah, try it now :)
21:13 autrijus svn.perl.org should be faster for you anyway
21:13 autrijus it's physically closer
21:13 merlyn I have enough to do tonight.  I have to write a column for UnixReview
21:13 merlyn "physically closer" to what?
21:14 autrijus you can let it "svn co" in the background... :)
21:14 putter the key quotes are "an array alias on a subpattern flattens and collects all nested subpattern captures within the aliased subpattern" and
21:14 autrijus svn.openfoundry.org is in .tw
21:14 putter "If an array alias is applied to a quantified pair of non-capturing brackets, it captures the substrings matched by each repetition of the brackets into separate elements of the corresponding array.".
21:14 merlyn ahh.  Yes, .tw is more distant than anything in the US.
21:14 putter both from Damian's "capturing" post.
21:14 Shillo autrijus: Is it possible to commit to the mirror?
21:15 theorbtwo What was the switch command-line?
21:15 stevan theorbtwo: it wont work, they are diff repos
21:15 * Shillo isn't even sure it's doable with svn, mind you. :)
21:15 autrijus Shillo: it is possible, sure
21:16 * putter goes back to reversing the sense of all the array captures in Perl6::Rules t/*t...
21:16 Shillo autrijus: Hmm, so it's fine to just switch to the mirror permanently... Hmm, lemme think about it.
21:16 * Shillo runs make test, just so...
21:17 iblechbot has quit IRC (Connection timed out)
21:18 knewt gah. hateful software.
21:18 knewt not pugs btw
21:18 stevan autrijus: a quick OO question if you have a moment
21:18 autrijus stevan: sure. btw, qualified class names is fine now
21:19 autrijus and larry killed #1. yay.
21:19 stevan autrijus++ # so damn fast, it /must/ be the svk
21:19 autrijus now it's either #2 or #3...
21:19 Shillo Yep, hyper.t regularily croaks on my build. As it should.
21:19 stevan autrijus: how can I get the invocant with a method? (currently in pugs of course)
21:20 autrijus $?SELF
21:20 stevan ah
21:20 autrijus totally hacky. I guess $_ should also work
21:20 rafl has joined #perl6
21:20 autrijus but it currently doesn't... let me fix
21:20 stevan autrijus: not as hacky as Hack::Instances, so I am fine with that
21:20 autrijus I think I like "o" :)
21:21 PerlJam stevan: not only must it be the svk, but it must be that autrijus works while asleep, because just a couple of hours ago, that's where he said he was going.
21:21 Shillo One thing I kept wondering... can delegate get a reference to the instance that delegated? A12 never mentions.
21:21 autrijus it is very easily explained by the hypothesis that autrijus is a consortium of hackers.
21:21 stevan PerlJam: he works while sleeping, eating *and* showering
21:21 Shillo (that's p6l stuff, not p6c, of course)
21:22 PerlJam autrijus: indeed. That's my working hypothesis at the moment
21:22 theorbtwo Sounds reasonable to me.
21:22 knewt autrijus: all of course working for IBM </sco-conspiracy-theory>
21:22 autrijus lol
21:22 castaway_ .oO( Are they all attending the hackathon ? ) O o ...
21:23 PerlJam autrijus: either that or you've figured a way to clone yourself and have arranged schedules accordingly.
21:23 stevan personally I think autrijus has some kind of time travel device,.. its a much more plausible explaination IMO
21:23 Shillo Tests done. Failed 28/249 test scripts, 88.76% okay. 80/4854 subtests failed, 98.35% okay. (that's yesterday's build)
21:23 PerlJam stevan: yes, that's another good theory except that it's very flawed.  If autrijus had a time-travel device, why wouldn't he just pop into the future and pick up a fully functional perl6?
21:23 PerlJam s/perl6/pugs/ as to your taste
21:23 theorbtwo Paradoxes?
21:23 stevan PerlJam: and give away his secret? hell no
21:24 theorbtwo It's not as much fun?
21:24 Shillo stevan: It has other drawbacks. If you use time-travel device to create more time for yourself, you accelerate aging.
21:24 integral PerlJam: that's what he's doing, but only svk merging a few changes a day
21:24 PerlJam stevan: Ah, so he probably *has* a fully functional pugs and it just releasing it bit by bit.
21:24 stevan PerlJam: exactly!
21:25 castaway_ clever!
21:25 Shillo PerlJam: Hmm. Makes sense. ;)
21:25 theorbtwo Very clever.
21:25 PerlJam Shillo: if you travel fast enough, time slows down.
21:25 Shillo PerlJam: But what you want is to speed up. For yourself, that is.
21:25 stevan Shillo: autrijus is really 8 years old
21:25 Shillo *grin*
21:25 theorbtwo I'm not sure about accelerating aging, but it does remove free will.
21:26 Shillo theorbtwo: Nah. You just avoid looking too closely into your own life. :)
21:26 PerlJam Shillo: no, no, accelerated aging is counter-acted by local time dilation
21:27 PerlJam In short, autrijus must live in a time bubble
21:27 Shillo PerlJam: Umm... lemme start munging Lorentz transformation a bit, I'll get back to you on that one. :)
21:28 theorbtwo Well, it was pretty obvious that he sometimes lives in his own little world...
21:28 PerlJam theorbtwo: see?
21:28 Shillo theorbtwo: OTOH, if PerlJam is right, and his bubble collapses, it may collapse into a black hole that will eat the Earth...
21:28 theorbtwo Hm, I should make a test that checks if AUTHORS is in the correct format.
21:28 castaway_ heh
21:28 Shillo theorbtwo: Hangman doubles as that. :)
21:29 theorbtwo Doesn't check the sort-order, though.
21:29 theorbtwo Or that there aren't any unlisted committers.
21:29 PerlJam He just interfaces his multi-dimensional time bubble to various parts of the universe through the carbaceous life form we call autrijus.  In reality he's a transdimensional being who exists on all dimensions and in all times.
21:29 Shillo Q.
21:29 theorbtwo So he already knows how @Larry will answer his questions?
21:29 Shillo Pugs is a practical joke, in other words. ;)
21:30 PerlJam Shillo: that's one conclusion.
21:31 Shillo And if we (the humanity) pass, he'll next sic Borg on us. ;)
21:31 PerlJam Shillo: I don't think his black hole would engulf the earth.  It's not massive enough.  It would evaporate before it got the chance to eat us.
21:32 Shillo PerlJam: 'Not massive enough'. They said the same thing about the iceberg.
21:32 PerlJam Shillo: no, you need a black hole of several solar masses otherwise it doesn't hang around long enough to do damage.
21:33 PerlJam Of course, I may be neglecting the transdimensional mass.  In our space-time he might be 80 kilos, but in the others, he could store gobs or mass (or energy)
21:33 Shillo No, several solar masses is what you need to make one naturally. Once you have a blackhole, its stability grows with mass, but all you need is that is sucks more than it radiates.
21:34 Shillo And that depends on the surrounding matter density. Inside Earth... frankly, I dunno the number.
21:34 PerlJam Shillo: blackholes below a certain size evaporate *really* quickly.
21:34 Shillo 80 killos might just be enough.
21:34 Shillo Yeah, but that refers to quantum ones.
21:34 Shillo 80 killos is macroscopic already.
21:34 _why has joined #perl6
21:35 theorbtwo Quantum and normal black holes differ only in construction.
21:35 theorbtwo (I think.)
21:35 osfameron has joined #perl6
21:35 * mugwump will reserve judgement until he actually observes one
21:35 Shillo Well, yeah. Quantum ones are bottom-up, normals are top-down. :) Problem is, collapsed time-travel devices are in neither category.
21:36 theorbtwo My point is that once they're there, you can't tell the difference.
21:36 autrijus stevan: this now works:
21:36 autrijus    method set ($self: $x, $y) { $.x = $x; $.y = $y }
21:37 G2 has quit IRC ("oops")
21:38 Shillo theorbtwo: Well, I meant that in reference to size. 80 killos is a macroscopic blackhole. It doesn't live very long, but drop it in a bucket of matter, and it might be over-equillibrium suckage.
21:38 theorbtwo Nod.
21:38 theorbtwo And like any other 80kg object, it'll tend to fall downwards.
21:39 autrijus working on $_ topicalising now
21:41 autrijus done:
21:41 autrijus    method set ($x, $y) { $_<x> = $x; $_<y> = $y }
21:41 autrijus you can mix the $self: form with $_.
21:46 mj41 s/@list/@array/ patch attempt ... http://wiki.kn.vutbr.cz/mj/attac​h/pugs/patches/pugs-l-to-a.patch  ... g'night
21:46 fitzzz has joined #perl6
21:55 Juerd stevan: Thanks for the summaries
21:56 autrijus mj41: thanks, applied
21:57 Juerd I'm shocked to see that list slices go away.
21:57 nnunley has quit IRC (Read error: 148 (No route to host))
21:58 theorbtwo BTW, an 80kg black hole would have to eat 2.3e12 kg/s, by my quick approx.
21:58 Juerd Arrays have always had much greater overhead than lists. How is this in Perl 6?
21:58 Shillo theorbtwo: Hmm, I'm really curious about the method. Problem is, dunno how to estimate the evaporation speed.
21:59 PerlJam theorbtwo: earth is what? about 6e24 kg? or something
21:59 mj41 has quit IRC (Read error: 104 (Connection reset by peer))
21:59 Juerd PerlJam: I don't think a big enough scale exists :P
21:59 mugwump e30?
22:00 * Shillo has to go, anyway. Later, folks! :)
22:00 PerlJam oops, I've got to go pick up my kids.
22:00 PerlJam later all
22:00 * PerlJam &
22:00 theorbtwo http://en.wikipedia.org/wiki/Hawking_radiation gives a formula for the evap time of a black hole by initial mass.
22:00 Shillo has quit IRC ("Leaving")
22:00 mugwump 5.9736x10^24 according to wikipedia
22:00 theorbtwo (Assuming no input of matterenergy.)
22:02 theorbtwo Ah, there's a way I can get a feel for times that short...
22:02 theorbtwo An 80kg black hole would evaporate in 1 clock cycle of a 28GHz computer.
22:03 autrijus accessors are there. committing.
22:03 * knewt checks to make sure autrijus is human
22:04 theorbtwo .oO(Or can pass the turing test.)
22:05 knewt at this rate pugs will be able to pass the turing test by the end of the year
22:05 Juerd ouch.
22:07 autrijus mmm sentient language
22:08 theorbtwo Hm, autrijus could be a time-traveling android, implemented in perl 6.2831853.
22:09 autrijus nah. I'm not John Titor :)
22:11 * Juerd thinks autrijus is on some drug that employers all over the world would love to get their hands on
22:11 Odin- theorbtwo: Why would it have to be implemented in perl? By the time it's gone that far, I think Perl will spontaneously create an AI. ;)
22:12 broquaint has joined #perl6
22:13 castaway_ bro!
22:14 wolverian hmm. it takes hours for my mails to p6l to appear there.
22:14 castaway_ then send them earlier
22:14 Juerd wolverian: Are you subscribed?
22:14 Juerd castaway++  # heh
22:15 wolverian Juerd: yes.
22:15 integral are you sending from the subscribed address?
22:15 wolverian ..good point. I'm not sure.
22:15 Juerd wolverian: What integral said, and add to that: is your envelope from what you think it is?
22:15 castaway_ you mean he forgot the stamp?
22:16 wolverian Juerd: envelope?
22:16 wolverian yes, the address is the same.
22:16 Juerd wolverian: SMTP uses envelopes
22:16 Juerd wolverian: The address used in MAIL FROM may be different from your From: header
22:16 wolverian hmm. okay.
22:16 wolverian would that be bad? :)
22:16 Juerd The header should be meaningless to anything that is automated, including mailing lists.
22:17 Juerd wolverian: Messages from unknown addresses are moderated
22:17 Juerd (Though I have absolutely no idea who does this. It must be a really sucking job.)
22:17 wolverian hmmmmm. maybe it's wolverian@localhost or something silly like that.
22:17 Juerd wolverian: Can you receive bounces?
22:17 Juerd If so, it's not anything@localhost :)
22:18 wolverian okay.
22:18 wolverian how would I see the MAIL FROM address?
22:18 theorbtwo Juerd: Not that bad, I think.
22:19 Juerd wolverian: Do you send mail using SMTP, or do you use something like sendmail?
22:19 Juerd theorbtwo: Re what?
22:19 wolverian Juerd: postfix.
22:19 theorbtwo They get machine spam-filtered, /then/ human-filtered; addresses that send mail that passes both get whitelisted.
22:19 theorbtwo Do an ethereal trace.
22:19 Juerd wolverian: Then it is probably username@hostname
22:20 wolverian hmm. I can't find any postfix logs. :) I guess they are off.
22:20 Juerd wolverian: Send a message to a non-existing address on an existing domain, and see where the bounce is sent.
22:20 Juerd Try, for example, [email@hidden.address]
22:21 wolverian done. thanks for helping out, by the way. :)
22:21 wolverian I'm a bit of a noob at mail.
22:21 Juerd That was already obvious :)
22:22 Juerd SMTP is a really simple protocol, though
22:22 Juerd Knowing it helps understanding e-mail.
22:22 wolverian hmm. /etc/mailname is 'sci.fi'
22:22 wolverian I guess it could be [email@hidden.address] then.
22:22 wolverian (which would be wrong.)
22:22 Juerd wolverian: Try the bounce thing. Bounces are sent to the envelope from
22:23 theorbtwo s/are/should be/
22:23 wolverian yeah, I sent it. nothing back yet
22:23 Juerd theorbtwo: If they're not sent to the envelope from, they're not bounces. :)
22:23 Juerd theorbtwo: Then they are automated replies
22:25 theorbtwo Hm, you have a certian point.
22:25 broquaint_ has joined #perl6
22:26 Juerd I try to always have a point.
22:27 mugwump what's the point of that?
22:27 Juerd It itself is the point.
22:27 wolverian still nothing back :(
22:27 Juerd Where did you send it?
22:27 knewt hmm, while we're mentioning email, qmail--
22:27 wolverian the address you gave. :)
22:28 Juerd wolverian: The bounce goes to [email@hidden.address] and sci.fi says the user is unknown
22:28 autrijus has $.x is rw
22:28 autrijus is supported.
22:28 Juerd wolverian: So [email@hidden.address] is your envelope from, and you can never get correct error reports.
22:28 wolverian Juerd: augh!
22:28 autrijus that's pretty good for a day... see you :) *wave*
22:28 wolverian how do I fix that? hmm. /etc/alias?
22:28 theorbtwo G'night, autrijus.
22:28 Juerd wolverian: Depends on your MTA
22:28 mugwump /etc/email-addresses with exim
22:28 Juerd wolverian: In general, your hostname shouldn't be someone else's.
22:28 castaway_ tell postfix what address to use, usually
22:29 broquaint has quit IRC (Read error: 60 (Operation timed out))
22:29 Juerd wolverian: Why do you have it set to sci.fi in the first place?
22:29 broquaint_ is now known as broquaint
22:29 Juerd Make sure you use a name that resolves to your actual IP.
22:29 castaway_ or set a real FQDN
22:29 wolverian Juerd: I use fetchmail. that's my ISP's server.
22:29 castaway_ not to fetch, to send
22:29 Juerd wolverian: Why the hell is it used for *outgoing* mail.
22:29 castaway_ (fetchmail is not an MTA)
22:30 Juerd add a 7 above the ., so it is a ?.
22:30 Juerd wolverian: echo $HOSTNAME
22:30 wolverian noesis. :)
22:30 theorbtwo fetchmail thinks it's an FTA.
22:30 theorbtwo Er, MTA.
22:31 Juerd theorbtwo: In a weird way, it is.
22:31 castaway_ no it forwards to your MTA
22:31 _why has quit IRC ("ircII EPIC4-2.0 -- Accept no limitations")
22:31 wolverian Juerd: do you mean my MAIL FROM should be this computer's address?
22:31 Juerd wolverian: If that computer can receive mail for you.
22:32 castaway_ more like viceversa..
22:32 theorbtwo "Normally, fetchmail behaves like any other MTA would..."
22:32 Juerd (Has port 25 accessible from the internet)
22:32 castaway_ (maybe I'm confusing MTA with MDA.. regardless, fetchmail is hardly the problme ,)
22:32 knewt wolverian: you said sci.fi is your ISP? so you have an email account on their server?
22:32 Juerd fetchmail is an MDA
22:32 theorbtwo Your MAIL FROM [email@hidden.address] address should always be deliverable to.
22:32 wolverian knewt: yes. and I'd rather have everything incoming go there.
22:32 Juerd But it can play a limited MTA
22:32 knewt wolverian: so what is the account?
22:33 wolverian knewt: [email@hidden.address]
22:33 Juerd wolverian: Then change your username, or find a way to fool your MTA to use a forged envelope from.
22:33 knewt ok, you should be using a MAIL FROM of that then
22:33 wolverian right. I'll read postfix's docs then.
22:33 wolverian thanks a lot, all.
22:33 Juerd wolverian: Until you fix this, send no mail. You're violating the RFC.
22:34 wolverian Juerd: right. thanks.
22:34 wolverian (and sorry.)
22:34 Juerd Heh
22:34 * castaway_ suspects an awful lot of people are..
22:34 Juerd it's not our problem - it's yours, mostly.
22:34 Juerd castaway_: Yes.
22:34 theorbtwo It's some of both.
22:34 Juerd castaway_: But that's mostly because an awful lot of people use hotmail.
22:35 castaway_ If it were that terrible, then the MTA setups/config whatever should complain about it, but they dont
22:35 * theorbtwo wonders if wolverian's local machine username matches the username of some sci.fi account, who has been getting wolverian's bounces.
22:36 Juerd theorbtwo: No.
22:36 Juerd castaway_: They have no way to know the setting is wrong.
22:37 castaway_ I'm sure it could if it really tried.
22:37 Juerd castaway_: You can't tell without trying to send a normal message, that [email@hidden.address] is a valid address
22:37 Juerd castaway_: And there ends the entire possibility of checking whether the envelope from makes sense.
22:37 castaway_ You could use VRFY
22:37 castaway_ (in theory)
22:37 Juerd VRFY is not always implemented.
22:37 Juerd It doesn't have to be.
22:37 castaway_ thats why its only a theory
22:37 theorbtwo VRFY is normally not implemented.
22:38 Juerd VRFY requires the SMTP daemon to know about users, which in practice often means it must run as root.
22:38 castaway_ anyway, it would be sending a mail to itself, effectively, so I dont see why it cant test
22:38 theorbtwo (For the same reason that most hosts don't allow finger.)
22:38 Juerd Nice for sendmail, which is insecure because of this, but not for most other smtpds.
22:38 castaway_ security thru obscurity
22:38 theorbtwo Juerd: It pretty well has to anyway, no?
22:38 Juerd castaway_: No, it doesn't have to be itself. The box could be a simple forwarder.
22:38 Juerd theorbtwo: Run as root? Hell no.
22:38 theorbtwo Security through obscurity is bad for certian reasons, which do not always apply.
22:39 theorbtwo No, Juerd, know about users.
22:39 Juerd theorbtwo: That neither.
22:39 castaway_ it could be, but that could/would require an extra config or some such, since most problem installs probably arent
22:39 Juerd theorbtwo: It has to know about users to be able to fail at the RCPT TO point
22:39 wolverian I'm only seeing local delivery aliases in postfix's docs.
22:39 Juerd theorbtwo: But if it doesn't know about users, it can leave the error reporting up to the MDA, which can initiate a bounce.
22:39 theorbtwo Hm, I suppose.
22:39 Juerd theorbtwo: See how qmail is built.
22:40 theorbtwo Waiting until that late to initate a bounce means more traffic, and more importantly, it means that the bounce can go to the wrong place.
22:40 castaway_ obscuring is useful for an extra layer, as long as its not the only one.
22:40 Odin- has quit IRC (Read error: 104 (Connection reset by peer))
22:40 theorbtwo Correct.
22:40 Juerd theorbtwo: The smtpd knows absolutely nothing about the system. This way, it cannot, even if it would want to, hand out that information.
22:40 stevan autrijus++ # i take the dog for a walk and you implement method invocants, $_ as invocant, has $.attr is rw; and autoaccessors
22:40 Odin- has joined #perl6
22:40 stevan autrijus++ # for making my life much much easier
22:41 castaway_ (still, it's a shame that finger disappeared.. )
22:41 wolverian should I be looking at postfix's docs for this?
22:41 Juerd theorbtwo: Yes, it means more traffic. No, it goes the right place. The place the sender indicated it wants to use for bounces. Sure, if the sender uses the wrong envelope from, things go wrong. But things go wrong equally in all other error cases.
22:41 theorbtwo Also, it's a bad trade-off when it means that third parties can't review and fix your security.
22:41 castaway_ probably, wolverian
22:41 stevan mugwump: I am will finish the Hack::Instances removal tonight (sometime in the next 4 hours or so)
22:42 stevan mugwump: hopefully by then openfoundry will be back too
22:42 wolverian hrm. then maybe 'alias' isn't what I'm looking for.
22:42 Juerd theorbtwo: qmail is secure because it was built to be secure. It has no security bugs. DJB is a weird guy with stupid licenses, but he really does know how to write good software.
22:43 castaway_ software without bugs does not exist ,)
22:43 Juerd castaway_: It has bugs, but no security bugs.
22:43 Juerd castaway_: And if you think it does, go find one and claim the $ 500 reward.
22:43 castaway_ anyone that claims such is asking for trouble IMO
22:43 wolverian sender_canonical_maps. yay!
22:44 Juerd http://cr.yp.to/qmail/guarantee.html
22:44 * theorbtwo wonders if he can stretch the defintions enough to claim that if it delivers phishing mails, it has a security bug.
22:44 mugwump software isn't good if no-one else can understand how to use it
22:44 castaway_ theres a huge difference between assuming bugs and actually caring to find them
22:44 knewt Juerd: the fact that qmail doesn't know about users at smtp time is nasty and bad, since it means all you can do for invalid email addresses is drop the email
22:44 Juerd mugwump: qmail was written very clearly and has consistent and thorough documentation. Its usage is very evidently simple.
22:45 Juerd knewt: What do you mean?
22:45 Juerd knewt: It isn't nasty, it isn't bad. It can be very inconvenient, yes.
22:45 Juerd knewt: Try qpsmtpd if this bugs you.
22:45 mugwump if it works for you, great.  However, I really can't excuse his Dickhead special license
22:45 Juerd knewt: Qmail has a modular design. It doesn't care if you use a radically different smtpd.
22:45 mugwump that's with a capital D.  Take note.
22:45 knewt Juerd: nah, i just use exim
22:46 Juerd mugwump: I'm not going to discuss the license -- I agree it sucks
22:46 Juerd mugwump: And I ask you to not mention it further.
22:46 mugwump fine
22:46 mugwump horses for courses
22:46 Juerd DJB's licensing is a source of endless debate in which all parties agree.
22:46 Juerd I've seen it destroy IRC channels before
22:47 elmex has quit IRC ("leaving")
22:47 mugwump I just really hate it when a "free" software author doesn't really open up development to all
22:47 wolverian hmm. now I should test this.
22:47 castaway_ it's his right to do so.. shrug
22:47 Juerd Development is open to all, as long as you distribute it as patches.
22:47 castaway_ testing++
22:47 Juerd It's not free software by usual definitions.
22:48 wolverian Juerd: was the server I sent earlier to yours?
22:48 Juerd wolverian: Yes.
22:48 fitzzz (ERROR) Cannot join channel: ##php (You are banned)
22:48 fitzzz :(
22:48 wolverian can I reuse the address to test this?
22:48 Juerd wolverian: It runs qmail (with qpsmtpd), fyi
22:49 Juerd wolverian: Yes
22:49 wolverian sent.
22:49 Juerd (I hate the license, but don't mind using this software. The simple fact is that qmail's design pleases me.)
22:50 Juerd perl.org also uses qmail with qpsmtp by the way :P
22:50 * castaway_ hands Juerd the drum.
22:50 Juerd Drum?
22:50 wolverian hm. bad filename. doesn't work.
22:50 castaway_ for banging on.
22:50 Juerd castaway_: Explain?
22:51 castaway_ banging the drum is what you are doing - about qmail. (repeatedly attempting to "sell" it)
22:51 castaway_ I think we've got it now.
22:52 Juerd I love qmail, and hate the stupid arguments I continue to hear about it. There's a lot of FUD. Almost everything about the license is true, almost everything about the technical side is false.
22:52 Juerd Every sysadmin should decide for himself which MTA(s) he wants to use
22:52 mugwump great, now we've sorted that out we can MOVE THE FUCK ON TIA
22:52 * castaway_ noticed.
22:52 mugwump :)
22:52 Juerd And qmail is very probably ruled out by company policy, if there is any.
22:53 wolverian hmm. mail.info log indicates it worked this time, if it indeed shows the envelope address.
22:54 larsen has quit IRC ("later")
22:54 Juerd wolverian: Yep.
22:54 wolverian yay! thanks for helping.
22:55 Juerd You're welcome
22:55 wolverian bounce came in now. hooray.
22:59 Juerd That must be the first bounce you get in a long time.
22:59 Juerd I wonder how often people never received your message without anyone knowing :)
23:04 wolverian I don't think that has happened. I only mail to mailing lists, I think. :)
23:04 wolverian (so I see when my mail isn't delivered.)
23:08 meppl gute nacht
23:10 sahadev has quit IRC (Read error: 110 (Connection timed out))
23:11 sahadev has joined #perl6
23:18 fitzzz has quit IRC ()
23:20 meppl has quit IRC (Remote closed the connection)
23:31 Juerd Wow.
23:31 Juerd Someone dares to post with Lotus Notes
23:35 wolverian http://hop.perl.plover.com/~alias/list.c​gi?2:mss:120:200505:dghinmhgjlmjjicjnidf
23:35 wolverian wow. that is bizarre.
23:36 wolverian ah. it's fixed in perl5.9.
23:38 Juerd Hm
23:38 Juerd Can you explain what happens?
23:38 vitaminmoo has joined #perl6
23:38 mugwump the named sub declaration doesn't close over its scope
23:39 mugwump so $prefix is out of scope and undef
23:39 Juerd What is "close over scope"?
23:39 wolverian http://hop.perl.plover.com/~alias/list.c​gi?2:mss:136:200505:gnggkfjgjfjlcicgaoop "solution" (a link to the patch that fixes it)
23:39 wolverian Juerd: as in a closure.
23:40 mugwump ie, capture all the lexicals in its context
23:40 Juerd The question is: why
23:40 Juerd Obviously, $prefix is *declared* in its scope, because strict does not complain
23:41 wolverian yes. it's a perl internals bug.
23:41 wolverian as you can see in the patch. :)
23:41 Juerd pugs does it right :)
23:41 wolverian that is a bizarre thing though. I just used that kind of code as an example of closures to someone a few days ago.
23:41 wolverian I don't think I ever tested if it actually worked. :)
23:41 Juerd I think I use it all the time.
23:41 wolverian but it doesn't work!
23:41 osfameron has quit IRC ("Leaving")
23:41 Juerd You'd think I would have encountered it many times already then.
23:42 Juerd I wonder if there are preconditions for the bug
23:42 Juerd s/use it/used it/
23:42 mugwump normally you don't name closures like that, though
23:43 Juerd I do...
23:44 wolverian it's not naming a closure, it's naming a closure creator.
23:44 wolverian (except in the sense that all subs are closures)
23:45 Juerd Oh!
23:45 Juerd Yea, I'd declare $prefix inside the sub.
23:45 wolverian on another note. I wondered why '*foo = sub { 1 }; print foo' wasn't printing anything :)
23:46 Juerd use strict.
23:47 wolverian yes.
23:47 wolverian actually, I was doing: '*closure = sub { sub { 1 } }; print closure->()'
23:47 wolverian which almost worked. heh.
23:48 wolverian it even passes strict.
23:59 putter is there a time estimate on when svn checkin will be possible?

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

Perl 6 | Reference Documentation | Rakudo