Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2005-09-27

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:01 putter hi stevan
00:01 svnbot6 r7153 | putter++ | PIL-Run - updated name mangling of types for mm2.
00:01 svnbot6 r7153 | putter++ | pugs-p5.pl: p5 expressions are now evaluated in the correct package.  And ->perl is now shown, rather than ->str.
00:01 putter let's see... why did I ping...
00:03 putter ah.  I figured out ->name.  so the only thing is the  Str->new()->isa('Class') #=> 1 bug report
00:04 putter and when DESTROY's in Array and Hash and Scalar began working, the gc errors became pervasive.  So I commented out the meat of the DESTROYs.
00:06 stevan Str->new()->isa('Class')??
00:07 putter ah,
00:10 stevan putter: hmm, that is the tricky part
00:10 putter bah.  it should be something like  $::Class->new('$.name'=>'Foo')->new()->isa('Class')   but gc errors stop me from actually running it
00:10 stevan Str->isa() is actually calling $instance_of_class->isa()
00:10 stevan gc errors?
00:10 stevan what ones?
00:11 putter oh, that's a new one,
00:11 putter (in cleanup) Method (DESTROYALL) not found for instance (Dispatchable=HASH(0x133b2d0)) at /home/net1/perl6/pugsxpl2/perl5/PIL-Run/../Perl6-MetaModel2.0/lib/chaos.pl line 450
00:11 stevan hmm
00:11 stevan the problem is that GC tries to call all the DESTROY methods
00:12 stevan which just totally wreaks havoc
00:12 frodo72 has quit IRC (Connection timed out)
00:12 stevan I will have to give this some thought
00:12 putter ok.
00:12 stevan maybe I should call, then remove all the DESTROY methods at the end of runtime
00:12 putter so if all the DESTROY methods are commented out, the gc problems should go away?
00:12 stevan no idea
00:12 stevan I never tried it
00:12 putter k
00:13 stevan let me try the removing DESTROY methods thing
00:13 stevan I think that should wokr
00:13 putter ok
00:13 stevan as for the Str->new()->isa(Class) thing
00:13 stevan thats a bug,... not sure how to fix it at the moment
00:14 stevan the problem is that there should be a class method version of isa()
00:15 putter k
00:19 stevan putter: can you make a test out of that GC issue?
00:20 stevan can you reproduce it on a small scale?
00:23 frodo72 has joined #perl6
00:27 putter hmm... let's see...
00:29 stevan putter: I think I might have gotten in
00:29 stevan s/in/it/
00:31 putter well, this   ./pugs -B perl5 -e 'say 3'  shows the old gc problem... :)  and... (oh, ok, I'll wait and try that)
00:31 svnbot6 r7154 | putter++ | pugs-p5.pl: repl reverted to p6_to_s from p6_to_perl.  No non-essential changes!  They may expose bugs, creating ambiguity vis mm2 transition.
00:34 stevan putter: commiting now ..................
00:34 stevan ok,.. its there
00:35 stevan try that,.. let me know if it complains
00:35 putter trying...
00:36 putter ./pugs -B perl5
00:36 putter p5ugs> :5 $::Class->new('$.name'=>'Foo')->new()->isa('Class')
00:36 putter Use of uninitialized value in concatenation (.) or string at /home/net1/perl6/pugsxpl2/perl5/PIL-Run/../Perl6-MetaModel2.0/lib/chaos.pl line 513.
00:36 putter        (in cleanup) Method (DESTROYALL) not found for instance (Dispatchable=HASH(0x13409d0)) at /home/net1/perl6/pugsxpl2/perl5/PIL-Run/../Perl6-MetaModel2.0/lib/chaos.pl line 450
00:36 putter ...
00:36 stevan this might interest you : http://users.ipa.net/~dwighth/smalltalk/bluebook/bluebook_imp_toc.html
00:36 svnbot6 r7155 | stevan++ | Perl6::MetaModel 2.0 -
00:36 svnbot6 r7155 | stevan++ | * chaos.pl now attempts to handle some GC issues
00:36 svnbot6 r7155 | stevan++ |    - in an END block we collect all active classes, then call DESTROY
00:36 svnbot6 r7155 | stevan++ |      on all of them, then set a flag to tell the dispatcher not to
00:36 svnbot6 r7155 | stevan++ |      accept anymore requests for DESTROY
00:36 svnbot6 r7155 | stevan++ | * added remove_method to ::Class
00:36 svnbot6 r7155 | stevan++ | * added FETCH('{}') to ::Package which gives back the namespace hash
00:37 svnbot6 r7155 | stevan++ |    - this is not ideal, but works for now
00:37 svnbot6 r7155 | stevan++ | * added some more tests re: the Str->new()->isa(Class) issue
00:37 frodo72 has quit IRC ("Ciao a tutti / bye to all")
00:38 putter neat.  read the books years ago.  added to my "readables" list.  thanks! :)
00:38 putter still complains.  replicable as described.
00:39 stevan putter: do you think you can boil it down to a test case for the MM?
00:39 stevan I am not sure what your repl is doing that is causing this
00:40 putter :5 is just             print eval("package ".&p6_main()."; ".$1),"\n";         warn $@ if $@;
00:40 stevan what is p6_main()
00:40 luqui has joined #perl6
00:40 stevan hey luqui
00:41 luqui hi stevan
00:41 luqui Language::AttributeGrammar cpanned
00:41 stevan what is that?
00:41 luqui an attribute grammar module
00:42 luqui it's a tree traversal thingydingy
00:42 stevan ah
00:42 stevan cool
00:42 * luqui just realized that it probably works on graphs pretty well too
00:42 luqui yay for sensible algorithms!
00:43 stevan putter: what version of perl are you using?
00:44 stevan so luqui what does it do? is it on search.cpan.org yet?
00:44 nothingmuch luqui++; # i will check tomorrow
00:44 luqui I just uploaded it ten minutes ago, so probably not
00:44 stevan luqui: wanna no paste it?
00:44 nothingmuch http://pause.perl.org/incoming/
00:44 nothingmuch no need
00:45 luqui there are four tests :-)
00:45 stevan nothingmuch: check this out http://users.ipa.net/~dwighth/smalltalk/bluebook/bluebook_imp_toc.html
00:45 luqui I couldn't think of a way to test it modularly
00:45 luqui so I just used some examples
00:45 nothingmuch luqui: that's what happenned to me with blondie
00:45 luqui so I know when it breaks, but it will be a pain to figure out how
00:45 nothingmuch i retrofitted a few later
00:45 nothingmuch now i will add more, and then refactor
00:46 nothingmuch s/now/tomorrow/
00:46 putter p6_main() just returns a package name.  the same error occurs with a bare eval($1).
00:47 nothingmuch btw, stevan, i was thinking of macro processing as nothing more than making the actual evaluation of code a compile time directive
00:47 putter v5.8.5
00:47 nothingmuch sort of like, to actually run a program you have to say something like BEGIN { apply(&main) }
00:47 putter I tried doing a file with just  use Perl6::MetaModel; $::Class->new('$.name'=>'Foo')->new()->isa('Class');  but that gives me a different (but in some ways similar) error.
00:47 nothingmuch that *really* confuses macros with functions, which I like
00:48 luqui heh
00:48 Amnesiac has quit IRC (Read error: 104 (Connection reset by peer))
00:48 putter Use of uninitialized value in concatenation (.) or string at perl5/Perl6-MetaModel2.0/lib/chaos.pl line 513.
00:48 putter Method (BUILDALL) not found for instance (#<=(9)>) at perl5/Perl6-MetaModel2.0/lib/chaos.pl line 450
00:48 nothingmuch code emission to a "static" backend means that there is no Eval node and no Val of type AST anywhere but &main
00:48 stevan putter: weridness
00:49 stevan putter: can you run the MM2 test suite?
00:49 putter running...
00:49 stevan nothingmuch: *cough* thick crack smoke *cough*
00:49 nothingmuch stevan: yup =)
00:49 nothingmuch forth has tought me the value of IMMEDIATE
00:50 luqui that is, everything is a macro in such a backend?
00:50 stevan luqui: I only sort of understand you Attribute grammer thing
00:50 stevan was there a p6l thread on this?
00:50 luqui yeah, tough to describe, esp. today when I am not in a documenty mood
00:50 nothingmuch stevan: there was a nice talk i linked to
00:50 nothingmuch it has an example of dumping a treeish structure as HTML
00:50 stevan nothingmuch: where?
00:50 nothingmuch with indentation
00:50 nothingmuch i don't remember
00:51 stevan bah,... what good does that do me
00:51 nothingmuch backlog ;-)
00:51 luqui but I'm looking for features to add.  I wrote this module before ever using an attribute grammar for anything
00:51 luqui so I have no idea what would be useful
00:51 nothingmuch ah, i think i found it
00:51 nothingmuch http://www.win.tue.nl/ipa/archive/falldays2003/AgTutorial.pdf
00:52 saorge has joined #perl6
00:53 stevan putter: I see your latest issue
00:53 nothingmuch good night
00:53 stevan $::Class->new('$.name'=>'Foo')->new()->isa('Class') wont work becuase Foo needs to have $::Object as it's superclass
00:53 stevan good night nothingmuch
00:54 geoffb good night, nm
00:54 nothingmuch btw.. new website up: http://www.beyahad.info/
00:54 nothingmuch point your hebrew-speaking-bi-polars there please
00:54 luqui night
00:55 luqui hey nothingmuch
00:55 nothingmuch huh?
00:55 luqui go to http://www.behayad.info
00:55 luqui :-)
00:55 nothingmuch i'm not bi-polar
00:55 nothingmuch i just built the website for a friend =)
00:55 luqui uh huh
00:55 luqui sure
00:55 stevan suuuuurrre for a friend
00:55 stevan just like luqui's friend who needed Viagra
00:56 nothingmuch nono, that was for my dog
00:56 nothingmuch uh, i mean, luqui's dog
00:56 nothingmuch friend
00:57 luqui haha
00:57 stevan putter: this works for me print eval("my \$Foo = \$::Class->new('\$.name'=>'Foo'); \$Foo->superclasses([ \$::Object ]); \$Foo->new();");
00:58 luqui why would you eval that?
00:58 stevan luqui: it is eval'ed because it is inside putter's repl loop
00:59 luqui i see
00:59 stevan he is having GC issues,.. we are trying to debug em
01:00 coet has quit IRC (Read error: 104 (Connection reset by peer))
01:00 coet has joined #perl6
01:01 luqui eew gc
01:01 luqui er, yay gc, eew gc issues
01:01 stevan eew ref counting GC,...
01:02 putter perl5/Perl6-MetaModel2.0/t/08_multi_methods.t
01:02 putter 1..12
01:02 putter ok 1 - The object isa PrettyPrinter
01:02 putter Can't call method "can" on an undefined value at /usr/lib/perl5/site_perl/5.8.5/Class/Multimethods/Pure.pm line 86.
01:02 putter # Looks like you planned 12 tests but only ran 1.
01:02 putter # Looks like your test died just after 1.
01:02 stevan odd
01:02 luqui uh oh
01:02 putter other than that, looks good.
01:03 putter trying...
01:03 vcv- has quit IRC (Read error: 104 (Connection reset by peer))
01:03 luqui that looks like my module :-p
01:04 luqui yay for bad error checking luke
01:04 vcv-- has joined #perl6
01:04 putter Use of uninitialized value in concatenation (.) or string at perl5/Perl6-MetaModel2.0/lib/chaos.pl line 513.
01:04 putter #<=(9)>
01:04 luqui oh wait --- stevan, did you note my latest change to C::MM::P
01:04 stevan putter,.. this commit will fix that one
01:04 stevan luqui: no idea
01:04 robkinyon has joined #perl6
01:04 luqui forgot to tell you that it would break your rewrite of internal_multi
01:04 luqui I refactored
01:04 robkinyon obra: ping
01:05 stevan luqui: you punk!!
01:05 luqui on the other hand, I made it so you don't have to rewrite anything
01:05 luqui just pass a custom registry object
01:06 stevan putter: what version of C::MM::Pure are you using?
01:06 luqui so that's probably where putter's problem is coming from
01:06 putter checking...
01:07 putter 0.10
01:07 stevan I have 0.06
01:07 stevan ok,.. upgrading now
01:09 obra robkinyon: heya. autrijus suggested I ping you about PDF::Template
01:09 svnbot6 r7156 | stevan++ | Perl6::MetaModel 2.0 - small fix for annoying GC stuff
01:09 obra robkinyon: I'm looking for complex examples.
01:12 stevan luqui: what do I need to do to fix this?
01:13 Amnesiac has joined #perl6
01:14 luqui so you know how you rewrote internal_multi?
01:14 luqui with basically the same code as it had before, but a small change at the end
01:14 stevan ok
01:14 stevan wanna just fix my code in chaos.pl?
01:14 stevan :)
01:14 luqui fair enough
01:14 robkinyon Uhh ...
01:14 luqui suppose I deserve it :-p
01:14 robkinyon Heh
01:14 stevan luqui++
01:14 robkinyon PDF::Template is in an interesting place right now
01:15 robkinyon What kind of complex examples are you looking for?
01:15 * stevan has to attend to $real_life, bbiab
01:15 * robkinyon & kid's screaming
01:18 robkinyon back
01:21 jiing is now known as jiing__
01:21 jiing__ is now known as jiing
01:22 robkinyon obra: I can come up with some complex examples, if you want
01:24 buu What happened to the eval bot?
01:27 luqui oh right
01:28 luqui I should start him up again
01:28 luqui pugs wasn't building, so evalbot wasn't updating
01:28 obra robkinyon: Thanks :)
01:33 svnbot6 r7157 | luqui++ | Brought MM2's Class::Multimethods::Pure usage up-to-date with my interface-breaking change.
01:33 vcv-- has quit IRC (Read error: 104 (Connection reset by peer))
01:35 joepurl has joined #perl6
01:35 luqui putter, that's your queue, try again :-)
01:36 luqui er, cue :-)
01:36 * luqui laughs at himself
01:37 buu I don't suppose anyone has even heard of a portmaster?
01:37 Supaplex i has
01:38 buu Really?
01:38 Supaplex yup
01:38 buu Have you ever, say, configured one?
01:38 Supaplex on occasion, yes
01:38 buu REALLY?!
01:38 buu Can you help me make this hideous piece of trash talk to another server via telnet?
01:38 geoffb heard of one, yes, fought with one yes, configured one, no
01:39 Supaplex you mean telnet out of it?
01:39 buu Yeah
01:39 Supaplex there is the telnet command, isn't there?
01:39 buu I'm trying to dial in to it and get connected to a telnetd running on a local box
01:39 Supaplex it's like uber simple last I recall
01:39 Supaplex ohh, you'll need to login as !root
01:40 buu Er, not that part
01:40 buu When ever I dial up portmaster at the moment I get a "press <RETURN> to begin Login"
01:40 Supaplex so telnet say to your cisco router from the portmaster
01:40 buu Just a linux box
01:41 Supaplex k, 1m - logging in to the pm4
01:41 buu I have a modem, ya? I want to dial a number and connect to the portmaster. I want the portmaster to transfer me over ethernet to another box running telnet, so I'm left going directly from my modem to the telnet box
01:45 Supaplex so the modem on your pc dials the pm, and you login to a terminal session, then you want to telnet via tcp to the linux box
01:45 ods15_ has joined #perl6
01:45 buu Yes!
01:46 buu Well, what do you mean by "login to a terminal session"?
01:46 Supaplex not ppp
01:46 Supaplex like the bbs days ...
01:46 buu Ah. Well, I was kind of assuming I would use ppp, but I'm fairly agnostic as long as it works
01:47 Supaplex depends on the account you have. if you have ppp connectivity, you can just telnet from your pc to the linux box, can't you?
01:48 Supaplex well, I guess you could telnet to the pm, then telnet again off the pm to the linux box
01:48 buu Well, it's slightly more complicated
01:48 buu I'm just using a pc as a test
01:48 buu Eventually it's going to be a dial-up modem in a small embedded system way the fuck out
01:48 ods15 has quit IRC (Read error: 110 (Connection timed out))
01:49 Supaplex I see...
01:49 Supaplex so what's the simplest part of your objective you're attempting to complete? (let's break this complicated thing down into bite size pieces)
01:50 mauke has quit IRC (Remote closed the connection)
01:50 mauke has joined #perl6
01:50 buu Er
01:50 buu Get telnet over modem on this pc?
01:50 buu Via the pm
01:51 Supaplex so you've got putty, right?
01:51 Supaplex humm does putty handle modem's directly (via serial?)
01:51 buu Actually, I was using hyperterm
01:51 Supaplex before you answer that, you want ppp, right?
01:51 buu Sure.
01:51 buu And yes, hyper term handles the modem directly. I give it the number and it dials it.
01:51 Supaplex so you can putty (use telnet) to the pm
01:52 Supaplex ok, then that's not likely to be ppp. that's straight up serial terminal
01:52 buu Er, yes, I can telnet to the pm just dandy.
01:52 Supaplex unless you talk ppp ;)
01:52 Supaplex ok
01:52 buu But when I dial the pm, I get the "press enter to begin login" dealy.
01:52 Supaplex and you can run command line utils off the pm
01:52 evalbot_7157 has joined #perl6
01:53 buu Sure?
01:53 Supaplex really
01:53 luqui ?eval [+] 1..10
01:53 evalbot_7157 55
01:53 Supaplex can you paste the transcript to a pastebin or something?
01:53 buu Supaplex: Transcript of what, exactly?
01:53 Supaplex you attempting to hyperterm, including that snippet where the delay is
01:54 buu Er, I open hyperterm, say "dial here", it displays an alert box saying "connecting..", eventually it connects, then I just get the message I mentioned above. I hit enter, it asks for a password.
01:55 buu "Press <RETURN> to begin Login Session"
01:55 buu "Password:"
01:56 Supaplex http://rafb.net/paste/results/jVw2oP82.nln.html is what I get
01:56 buu I've tried lots of passwords and such, but nothing works.
01:56 Supaplex well you have to have a login on the pm :)
01:56 buu Supaplex: Yes, that's talking via telnet
01:56 buu I've got telnet to the pm working fine, it's not the issue here.
01:56 buu I'm trying to dial *through* the pm
01:56 Supaplex as in my sample, you're having issues paste the lithium> prompt, right?
01:57 Supaplex paste/past/
01:57 buu No!
01:57 buu My issue has nothing to do with your paste!
01:57 Supaplex ok
01:57 buu That is using telnet to talk directly to the portmaster.
01:57 buu I can do that fine and dandy.
01:58 buu I'm trying to use my modem to dial up, go *through* the portmaster, and talk to another box using telnet via the portmaster
01:58 Supaplex after you're dialed into the pm, you're trying to telnet?
01:58 buu Yes!
01:58 Supaplex can you ping google or browse www sites?
01:58 buu Except, I'm not really because I don't know how to configure it.
01:58 buu Er, I don't get anything past the login prompt when I dial in.
01:59 Supaplex yea, that's nothing fancier then getty attached to $SHELL - no tcp/ip or ppp involved. you're not routing anything
02:00 buu Um. What?
02:00 buu The portmaster has a shell?
02:00 Supaplex ignore that, first see if you can ping/browse etc while connected
02:00 buu I CAN'T CONNECT
02:01 Supaplex then quit using hyperterm!
02:01 Supaplex to dial ...
02:01 buu What else should I use?
02:01 Supaplex inetwiz
02:02 buu Look. I can dial in and "connect" to the pm
02:02 buu But *all* it gives me is the "press enter for login" dealy
02:02 buu I can't get past that.
02:03 justatheory has quit IRC ()
02:04 Supaplex but if you're using hyperterminal to dial, you're only connecting to the pm via a serial console over modem. (like mgetty)
02:04 Supaplex bbiam
02:05 buu Er, I don't think so. The the box is only connected to a modem. This modem is connected (eventually) to a portmaster
02:06 buu (Going for hot dogs, I shall return!)
02:06 ods15_ has quit IRC (Read error: 110 (Connection timed out))
02:06 putter luqui: thanks! updating...
02:07 putter syntax error at perl5/Perl6-MetaModel2.0/lib/chaos.pl line 559, near "my "
02:08 coet has quit IRC (Read error: 104 (Connection reset by peer))
02:08 coet has joined #perl6
02:09 luqui that's... odd
02:09 luqui didn't happen for me
02:10 luqui unless the test suite doesn't depend on chaos.pl
02:10 putter my $registry = {   had two { {
02:10 luqui yeah, I see that.  weird
02:10 luqui did you fix it?
02:11 putter 7158.
02:11 svnbot6 r7158 | putter++ | chaos.pl: fix typo
02:12 luqui what's the name of that module that allows you to croak out of a set of modules?
02:12 Supaplex buu, you can connect to a pm with minicom, hyperterminal etc and not have a ppp connection established.  It's one of the handy ways to troubleshoots a pm box when your data provider is having issues to the pop (point of presence) location.  I've done that dozens of times to login and see no ppp sessions active, because radius will not authenticate (because the data transport is down, telco PRI's work fine). if you want docs from the pm about this, I coul
02:12 Supaplex luqui: Carp ?
02:13 evalbot_7157 has quit IRC (Remote closed the connection)
02:13 putter the test file use Perl6::MetaModel; print eval("my \$Foo = \$::Class->new('\$.name'=>'Foo'); \$Foo->superclasses([ \$::Object ]); \$Foo->new();");  no longer gives warnings, but result in  #<AnonClass=(9)>, which seems incorrect.
02:13 luqui no, the one where if module A calls module B, and I "own" both of those, if I croak from B, it will look past A to find the caller
02:13 putter s/result in/results in/
02:13 evalbot_7158 has joined #perl6
02:14 putter err, huh?
02:14 luqui putter.  hmm.
02:14 putter oh, a Carp comment. ;)
02:14 luqui I think you're back in stevan's turf
02:14 putter sounds like.  my thanks for your help. :)
02:15 luqui Carp::Clan, that's it
02:15 buu Supaplex: No, I'm connected via modem
02:16 Supaplex and your modem transports a serial connection over a telco line.
02:16 putter luqui++
02:17 buu Supaplex: Um. Ok?
02:17 Supaplex and if you were connected via ppp while using hyperterminal, you'd beable to browse etc
02:17 buu Supaplex: Wouldn't I have to get past this login prompt?
02:18 Supaplex yea, you'd have to have a login for it. not to be confused with a radius login that ppp users would use
02:18 putter stevan: the .t error has changed to > Can't use string ("PrettyPrinter::pretty") as a HASH ref while "strict refs" in use at perl5/Perl6-MetaModel2.0/lib/chaos.pl line 555.
02:18 putter > # Looks like your test died before it could output anything.
02:18 buu Supaplex: I created a user on the pm, but this thing doesn't ask me for a username, just a password
02:19 putter however, the gc error on exit is gone! :)  luqui++ stevan++
02:19 nothingmuch has quit IRC (Read error: 113 (No route to host))
02:19 Supaplex what portmaster is it?
02:19 buu Supaplex: a portmaster 3, I beleive
02:20 buu believe!
02:20 Supaplex I'm not even sure if we have any of those in service
02:20 * Supaplex checks
02:21 putter stevan: if Array's DESTROY is again enabled, the result is  (in cleanup) Bad instance (FORCE=SCALAR(0xac41a0)) at /home/net1/perl6/pugsxpl2/perl5/PIL-Run/../Perl6-MetaModel2.0/lib/chaos.pl line 52
02:23 Supaplex nope, just a armfull of pm4's and some 3com stuff
02:23 Supaplex anyway, time to commute
02:24 buu Well, thanks for the time any way.
02:25 putter fglock: @a=(3,4) no longer works.  and push(@b,3) fails with Signature does not match (as does for()).  I haven't yet looked at it closely.  hopefully tomorrow.
02:28 ods15 has joined #perl6
02:35 putter so, not bad for a day's fiddling.  stevan++ fglock++ luqui++
02:36 putter good night all &
02:36 putter has quit IRC ("Leaving")
02:44 justatheory has joined #perl6
02:44 justatheory has quit IRC (Client Quit)
03:09 luqui has quit IRC ("This computer has gone to sleep")
03:20 luqui has joined #perl6
03:24 luqui autrijus, ping
03:30 vcv has joined #perl6
03:33 coet has quit IRC (Read error: 104 (Connection reset by peer))
03:33 coet has joined #perl6
03:46 SamB has quit IRC (Read error: 110 (Connection timed out))
03:50 Amnesiac has quit IRC ("Leaving")
04:09 elmex_ has joined #perl6
04:17 elmex_ has quit IRC ("Lost terminal")
04:19 kakos has quit IRC (Read error: 110 (Connection timed out))
04:25 elmex has quit IRC (Read error: 110 (Connection timed out))
04:27 justatheory has joined #perl6
04:28 justatheory has quit IRC (Client Quit)
04:37 coet has quit IRC (Read error: 104 (Connection reset by peer))
04:37 coet has joined #perl6
04:43 joepurl has quit IRC ("Lost terminal")
04:43 joepurl has joined #perl6
05:00 SM_ax has joined #perl6
05:06 Amnesiac has joined #perl6
05:26 r0nny has joined #perl6
05:30 svnbot6 r7159 | yiyihu++ | * Fix some typo.
05:34 evalbot_7158 has quit IRC (Remote closed the connection)
05:35 evalbot_7159 has joined #perl6
05:42 Amnesiac has quit IRC ("Leaving")
05:54 GeJ has joined #perl6
06:02 autrijus luqui: pon
06:02 autrijus greetings \camels
06:04 autrijus listening to brendon's invited talk:
06:04 autrijus "Javascript in 10 years"
06:05 luqui autrijus, I was going to ask you what a chained attribute was, but i figured it out
06:05 autrijus cool. it's a bidirectional attr
06:05 luqui yeah.  except it doesn't seem to make a lot of sense
06:05 luqui lhs.num = @lhs.num + 1
06:05 luqui seems to be, um, wrong
06:06 autrijus that does not look like correct indeed. where does that come from
06:06 luqui let me find the link
06:06 luqui http://www.cs.uu.nl/~arthurb/data/AG/AGman.html#htoc27
06:07 luqui I know what it's saying, that the synthesized attribute is one more than the inherited attribute of the same name
06:07 luqui But it's a wicked shorthand (if I'm grokking it)
06:08 luqui oh, also, Language::AttributeGrammar is cpanned :-)
06:08 autrijus ooh.
06:08 autrijus and yes it's a wicked pun.
06:08 autrijus it makes sense but evil
06:08 luqui so I'm not doing it in my module.
06:08 autrijus ooh, luqui++
06:10 luqui anyway, besides autocreated attributes, are there any important features?
06:10 luqui (and no, I'm not including a typechecker :-p)
06:11 autrijus are you working with variant (tagged union) types?
06:12 luqui uh, it can, provided they implement the right interface
06:12 coet has quit IRC (Read error: 104 (Connection reset by peer))
06:12 coet has joined #perl6
06:12 luqui (a method for each child node is enough)
06:13 luqui One of my goals is to make it work nicely Parse::RecDescent without being coupled to it
06:14 autrijus ok.
06:14 autrijus makes sense, since P::RD is the closed match on CPAN
06:14 autrijus s/closed/closest/
06:16 autrijus "javascript (previously known as "mocha") suffered from the "marketing scam" where imitating java's suboptimal standard library at that time"
06:16 autrijus <brendon> and some people still think I'm Gosling. it's very confusing
06:16 autrijus s/where/by/
06:17 * luqui doesn't associate JS with Java at all anymore :-)
06:19 autrijus not "object oriented" but "object based" -- i.e. not distinguishing constructors from other regular methods
06:19 autrijus "conventions enforced by culture"
06:20 luqui :-)
06:20 autrijus "each ajax application needs its own OO system with its own reflections anyway"
06:21 coral mm, Parse::BNF in javascript
06:21 coral "click here to test the parser described in the RFC above"
06:21 coral pardon the dreaming
06:21 luqui coral, it shouldn't be that hard
06:21 luqui parsers are easy.  fast parsers are hard. :-)
06:21 QtPlatypus luqui: Famious last words.
06:22 coral it's already on CPAN..
06:22 luqui oh, I figured "pardon the dreaming" meant it didn't exist
06:22 coral no, that's better phrased as "pardon the idea without motivation to implement"
06:22 QtPlatypus QtPlatypus: BTW is there a good way to do default values for object attributes in pugs?
06:23 * coral looks at QtPlatypus
06:23 autrijus autrijus: sure
06:23 autrijus BUILD ($.x = 3) { ... }
06:23 * coral reverts to osmosis phase
06:24 luqui class { has $.x = 3;
06:24 luqui } # not supported?
06:24 QtPlatypus luqui: Thats currently a paser error.
06:25 autrijus luqui: no, not sure how to interact with BUILD for one ( I guess BUILD overrides)
06:26 autrijus luqui: but a proper mm2 integ should fix that
06:36 pdcawley Morning all.
06:36 * luqui just realized that Parse::RecDescent::FAQ has the question "Is Parse::RecDescent LL(1)?  LL(n)?  LR(1)?  LR(n)?" and never answers it
06:36 nothingmuch has joined #perl6
06:36 luqui evening pdcawley
06:36 luqui no, I suppose it is morning
06:38 iblechbot has joined #perl6
06:39 QtPlatypus ?eval "a" eqgt "a"
06:39 evalbot_7159 Error:  unexpected "g" or "e" expecting operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
06:39 QtPlatypus ?eval "a" eq "a"
06:39 evalbot_7159 bool::true
06:39 QtPlatypus ?eval "a" gt "a"
06:39 evalbot_7159 bool::false
06:39 QtPlatypus ?eval "a" egt "a"
06:39 evalbot_7159 Error:  unexpected "e" expecting operator, postfix conditional, postfix loop, postfix iteration, ";" or end of input
06:39 QtPlatypus Thought so
06:39 luqui ?eval "a" ge "a"
06:39 evalbot_7159 bool::true
06:40 QtPlatypus Thanks luqui
06:43 r0nny damn
06:44 luqui darn
06:44 luqui r0nny?
06:44 r0nny i got a socket $foo - how to read data from it ?
06:44 luqui in perl 6 land?
06:44 r0nny yeah
06:44 * luqui doesn't know
06:44 r0nny argh
06:44 luqui probably with the iterator
06:44 QtPlatypus =$foo
06:44 luqui yeah, that one
06:45 saorge has quit IRC (Read error: 110 (Connection timed out))
06:45 r0nny hmm
06:45 r0nny now it "works"
06:46 r0nny wtf
06:47 r0nny after i kill te connec ted programm it prints everything i send to it, but adds 2 \n for each \n in the input
06:48 r0nny for =$bar{say;}; <- what did i break ?
06:49 luqui yeah
06:49 luqui because say appends a \n after it prints
06:49 luqui and =$bar doesn't strip the newlines (but maybe it should now?)
06:49 luqui use print instead of say
06:50 r0nny there is still a problem
06:50 r0nny it wont print anything, till i close the connection
06:50 r0nny how munch does = want to read ?
06:52 luqui one line usually
06:52 luqui but it may be laziness creeping in...
06:52 * luqui isn't sure
06:54 r0nny omg
06:54 r0nny even if i feed it with about 60 lines it wont print
06:54 r0nny :(
06:55 r0nny does anyone know, how to change this ?
06:55 luqui ??
06:56 r0nny i used netcat, to sein a longer file to the script, but it handt printed anything till i closed the socket
06:56 r0nny send
06:56 r0nny hmm
07:00 r0nny anyone ?
07:01 nothingmuch has quit IRC (Read error: 113 (No route to host))
07:01 QtPlatypus Does split('',$string) use the PGE and is there an equiverlent that doesn't
07:01 * QtPlatypus doesn't.
07:02 r0nny omg
07:02 r0nny pugscc is broken
07:02 autrijus r0nny: it doesn't use PGE
07:03 QtPlatypus ?eval "text".chars
07:03 evalbot_7159 4
07:03 QtPlatypus autrijus: That was directed at me I assume?
07:03 autrijus er sorry yeah
07:03 autrijus r0nny: try doing a $*OUT.flush after the say
07:04 r0nny wont work
07:04 r0nny seems like the socket wants to read waaay to munch data
07:05 autrijus hm, nopaste the code?
07:05 autrijus also try
07:05 autrijus while ($_ = $bar.readline) { ... }
07:05 QtPlatypus ?eval "text".split('')
07:05 evalbot_7159 ('t', 'e', 'x', 't')
07:05 autrijus bad style of course
07:06 autrijus but worth trying
07:06 r0nny $bar.readline breaks it complete :/
07:07 autrijus nopaste the code somewhere :/
07:07 r0nny my $bar=8060.listen.accept;for =$bar { print; };
07:08 r0nny this is the minimal working version - any additions had no effect
07:08 r0nny (or bad effect)
07:10 autrijus my $bar = 8060.listen.accept;
07:10 autrijus while ($_ = $bar.readline) {
07:10 autrijus    print;
07:10 autrijus    $*OUT.flush;
07:10 autrijus };
07:10 autrijus this worksforme
07:14 r0nny $*OUT.flush isnt needed
07:14 nothingmuch has joined #perl6
07:14 r0nny thx :)
07:14 r0nny is there a short way of connec t, too ?
07:15 autrijus look at examples/network
07:15 autrijus my $hdl = connect("www.phreeow.net", 80);
07:15 autrijus (in http-client.p6)
07:15 r0nny hmm
07:15 r0nny i think im going to write a simple irc bot now
07:16 autrijus cool, there are quite a few in examples/network :)
07:16 nothingmuch morning
07:18 autrijus iblech: what do you think about Net::IRC-OO?
07:18 autrijus as it stands it can't be use'd
07:18 autrijus may I mv it to Net::IRC::OO?
07:19 r0nny lol
07:19 r0nny the wizard game is broken ;)
07:19 autrijus iblechbot: anyway I'm going ahead :)
07:19 autrijus r0nny: yeah, we are in bad need of releng help
07:19 nothingmuch autrijus: maybe we need yet another namespace
07:19 autrijus how is it broken?
07:20 nothingmuch Net::IRC <author> <version> <style>
07:20 nothingmuch where style is freeform, but not part of the module's name
07:20 autrijus mmm "style"
07:20 r0nny autrijus:  aafter i win vs the first opponent, it seems to stay in a endless loop
07:20 autrijus I think it's part of "version"
07:20 autrijus r0nny: weird because it worksforme
07:20 r0nny hmm
07:20 evalbot_7159 has quit IRC (Remote closed the connection)
07:21 evalbot_7160 has joined #perl6
07:21 * luqui thinks that that's part of the module name
07:21 luqui it's a different interface, compatible in no way
07:22 autrijus good observation.
07:22 nothingmuch aye
07:22 nothingmuch but Net::IRC is the "purpose"
07:22 luqui yeah, well Class::Multimethods is the purpose of Class::Multimethods::Pure
07:22 nothingmuch why should Net::IRC::OO be called Net::IRC::OO just because we didn't have OO back when Net::IRC was started
07:23 luqui Why should Net::IRC have style "OO" just because we didn't have "style" back when Net::IRC was started?
07:23 luqui I don't think you can ever get away from the problem
07:23 svnbot6 r7160 | autrijus++ | * Move Net::IRC-OO into Net::IRC::OO
07:24 luqui the only way I see style being useful is if it's really part of the same module by the same author
07:24 luqui and in that case, shouldn't it just be an import flag?
07:25 penk has quit IRC (Read error: 110 (Connection timed out))
07:27 luqui so autrijus (and nothingmuch too), any suggestions for new features Language::AttributeGrammar?
07:27 luqui I want to make it a nice, complete module, not a POC
07:29 autrijus ok, looking at it
07:30 luqui thanks
07:30 didie_p6 has joined #perl6
07:30 didie_p6 has quit IRC (Remote closed the connection)
07:33 larsen has joined #perl6
07:37 r0nny omg
07:38 nothingmuch ?
07:38 r0nny i think i found, why for =$bar wont work
07:38 luqui you did?
07:38 r0nny seems like there is no autoflush for sockets
07:43 nothingmuch luqui: i'll have a look at your AG later
07:43 * nothingmuch has to work
07:43 nothingmuch didn't notice your RFC
07:44 luqui RFC?
07:44 luqui ahh
07:44 luqui not formal :-)
07:44 luqui 'sok, I'm going to bed in 15 mins or so
07:44 nothingmuch RFC the acronym, not the document
07:44 nothingmuch aye
07:44 nothingmuch sleep well
07:44 luqui thank you, work well, er...
07:46 * nothingmuch feels like a mega-xp-whore on perlmonks
07:47 nothingmuch i've never gotten so many upvotes on such a pointless node ;-)
07:48 luqui which node?
07:48 nothingmuch in this thread: http://perlmonks.org/?node_id=494914
07:48 r0nny omg
07:50 luqui that was it?
07:50 luqui oh no, r0nny is talking about his god again :-)
07:53 r0nny the life example is slow as hell
07:54 luqui everything is slow as hell
07:54 luqui we're hoping we can optimize later :-/
07:55 r0nny i hope someone steals a time machine, and rips the final perl6 from the future
07:55 luqui well, apparently the parrot backend was pretty fast when it was working (on a very small subset of p6)
07:55 xinming_Beijing has joined #perl6
07:55 luqui r0nny, you can't wait?
07:55 luqui :-)
07:55 luqui there are better things to do if you can't wait then hope for the impossible
07:55 r0nny the language perl6 really __kicks__ ass
07:55 luqui one is to learn haskell and start hacking
07:55 luqui (or even don't learn haskell and hack on nonhaskell parts)
07:56 luqui r0nny, thank you.  we're all proud.
07:56 r0nny haskell is slow
07:56 luqui faster than perl5
07:57 luqui faster than java.  it ain't bad.
07:57 r0nny faster then java is easy ;)
07:57 luqui :-) agreed
07:57 r0nny well - cant beat good c/c++
07:58 luqui asm can.  but that just goes to reflect that fundamental tradeoff
07:58 luqui you want developer time or program time?
07:58 luqui (the tradeoff seems to break down when it comes to java)
07:59 luqui (where both times are large)
07:59 r0nny well - actually im writing some c++ libs trading compiler time for programmer time - but dont waste programm time
07:59 luqui a type inferencer?
08:00 r0nny type inferencer ?? - never heard of this
08:00 luqui that's one thing I can think of that doesn't waste program time but saves on developer time.  But it trades compilation time (which is cheap)
08:00 luqui type inferencer looks at your program and fills in all the "int" "MyClass" "vector<foo>", etc. for you
08:00 luqui by looking at how you use your variables
08:00 r0nny omg
08:01 r0nny this sucks
08:01 * luqui yawns
08:01 luqui I think it's time for bed
08:02 r0nny i combine different idoms, to make it allmost impossible to do bugs, if only using the libs
08:03 luqui I'll have to look at that
08:03 r0nny and with some operator overloading, polimorhism, template mixins and the other fun stuff it is very easy to write complex stuff in less time
08:03 r0nny the lib is very basic atm
08:03 r0nny since it will be part of the os i develop
08:03 luqui oh my
08:03 luqui Laziness, Impatience, and what?
08:04 r0nny later it will have templates for easy definig of devices, so huge parts of the drivers will write themself
08:04 luqui anyway, I got to get to bed
08:04 r0nny n8 then
08:04 luqui night
08:04 luqui has quit IRC ("Leaving")
08:08 lao has joined #perl6
08:09 jabbot has joined #perl6
08:19 autrijus luqui: nice job on L::AG.
08:19 autrijus looks most sane.
08:32 * Juerd wonders what L::AG is
08:33 autrijus Juerd: useful stuff.
08:33 autrijus http://pause.cpan.org/incoming/Language-AttributeGrammar-0.04.tar.gz
08:33 Juerd Oh, OS X, how I hate thy awful performance!
08:33 Juerd Thanks
08:33 nothingmuch aweful performance?
08:34 Juerd http://pause.cpan.org/incoming/Language-AttributeGrammar-0.04.tar.gz
08:34 Juerd I could learn a foreign language while waiting for the beach ball
08:34 Juerd Hmm
08:34 nothingmuch ah
08:34 nothingmuch not enough ram, perhaps?
08:34 dduncan how much RAM do you have?
08:35 dduncan I find its good to have a minimum of 512MB ... 768 or a Gig works great
08:35 * nothingmuch gets that on his laptop after he's opened about 20 terminal windows and 50 browser tabs, and something code goes out of control
08:35 dduncan certainly, if you only have 256M, making Pugs will crawl
08:35 nothingmuch after i kill the memory hungry progam, everything beachballs when swapping back
08:35 Juerd nothingmuch: I have 512 MB. I'm used to getting much better responsiveness out of machines with this amount of memory.
08:36 nothingmuch Juerd: well... the UI comes at a cost =/
08:36 nothingmuch you can run KDE/Gnome on it via fink though
08:36 dduncan its not the UI so much
08:36 Juerd Oh, don't worry
08:36 nothingmuch GNU Darwin can get you started
08:36 evalbot_7160 has quit IRC (Remote closed the connection)
08:36 Juerd This box will know Linux.
08:36 nothingmuch dduncan: that was my best guess... what is it?
08:36 Juerd And the more it'll beach ball, the sooner it will get Linux on it.
08:37 dduncan since Mac OS X 10.2, all that RAM used by the window manager is highly compressed, and hardware accelerated
08:37 evalbot_7161 has joined #perl6
08:37 dduncan so the issues are more elsewhere
08:37 Juerd In fact, when I'm done watching all movies stored on this hard drive, I'm migrating.
08:37 nothingmuch dduncan: not the graphics buffers... the book keeping data
08:37 Juerd Saves me a whole lot of money too
08:37 Juerd I had this thing for a few months, and then 10.4 came, with a huge price tage.
08:38 dduncan I'm still using 10.3.9 ... meets my needs
08:38 Juerd I hadn't considered the subscription element of proprietary software. I just wasn't used to that anymore.
08:38 svnbot6 r7161 | autrijus++ | * don't "use" String::Multibyte unless we actually use .grapheme
08:38 Juerd dduncan: Well, I want a newer Java, but Apple won't let me have it.
08:38 Juerd And I hate having to jump through lit hoops to get something working.
08:39 dduncan and that's unique to Mac OS X?
08:39 * nothingmuch thinks Juerd and OSX are not meant to be together
08:39 Juerd dduncan: So far, oh yes.
08:39 Juerd Perhaps Debian and Ubuntu spoiled me
08:39 Juerd But yes, I am used to getting much simpler installations and upgrades.
08:40 Juerd Oh, did I mention that in the lower left corner of my primary workspace, there is part of the window that has been requesting a reboot, since three weeks ago?
08:40 Juerd It has no close button, and I hate it too much to just kill it.
08:40 Juerd It has to suffer in the corner.
08:41 dduncan why is that dialog there, requesting reboot?
08:41 Juerd The reboot is for non-system upgrades, by the way.
08:41 Juerd Well, the upgrading software has installed newer software.
08:41 dduncan did it install additions to the OS?
08:41 Juerd Safari, Mail, and iTunes.
08:41 Juerd All three I can restart without restarting the entire system, thank you, OS X.
08:42 Juerd But even if it were a system kernel upgrade
08:42 nothingmuch you can subvert that
08:42 dada has joined #perl6
08:42 nothingmuch cmd+opt+esc
08:42 Juerd It would still allow me the choice to collect some uptime.
08:42 dduncan he didn't want to kill it
08:42 nothingmuch btw, softwareupdate from the command line asks you to rebooto manually
08:42 Juerd s/would/should/
08:42 Juerd no, MUST.
08:43 * nothingmuch thinks that it doesn't really matter
08:43 nothingmuch because most OSX users are like my mom
08:43 nothingmuch they don't know about the fundamental diff between an app and the syste,
08:44 nothingmuch and if safari ships a new shlib that everything else must be restarted for, then perhaps rebooting the machine is the simplest choice
08:44 dduncan since those Apps are made by Apple, it is possible that they bundle certain system service updates with the app updates, and/or the apps are composed partly of daemons, hence the restart request ...
08:44 Juerd It should optimize for "most" users
08:44 dduncan however, I haven't seen this when I've updated those apps
08:44 nothingmuch the fundamental issue is that OSX is not meant for people who like control
08:44 Juerd Not cripple other users.
08:44 dduncan I don't recall being asked to restart when upgrading an app, only a system component
08:44 nothingmuch well, too much choices are, at least in apple's opinion, adding unnecessary complexity
08:44 Juerd I have some applications open that I wish to keep open for weeks.
08:45 nothingmuch then you can ignore it
08:45 nothingmuch if you know what you are doing you can get around it
08:45 Juerd After restarting the system, it takes me at least 20 minutes to get the system back in its current state.
08:45 nothingmuch Juerd: what apps are those?
08:46 nothingmuch you can use launchd to get veyr right init.d like behavior
08:46 dduncan according to my current Software Update, the iTunes upgrade does *not* request a restart
08:46 nothingmuch per user or per system
08:46 dduncan however, Quicktime does
08:46 dduncan and they did appear around the same time
08:46 Juerd nothingmuch: It includes firefox with some tabs, the whole of applications, window arrangement
08:46 dduncan maybe that's what requested the restart
08:46 nothingmuch that's because quicktime is reused in around every app that makes noises or draws pictures
08:46 dduncan yes
08:46 Juerd nothingmuch: Some permanently open IM windows
08:47 Juerd I don't even know who they all are
08:47 Juerd But it sure does help me to have them open all the time.
08:47 Juerd dduncan: It's not a matter of simple requesting
08:47 nothingmuch Juerd: if you like that specific form of stability, then i think you should just refrain from updating.. there's nothing really important in there, I think
08:47 Juerd The window doesn't have a "I'll restart later, but thanks for the heads up" button
08:48 Juerd Even Microsoft Windows has this.
08:48 dduncan oddly enough, some of them do
08:48 nothingmuch Juerd: the issue is cultural
08:48 nothingmuch based on apple's assumptiont hat most people should not be given that choice
08:48 svnbot6 r7162 | autrijus++ | * At suggestion from dons, expose the main "runProgramWith" API
08:48 svnbot6 r7162 | autrijus++ |   of Pugs so it can turn a Pugs "Val" into anything at all,
08:48 svnbot6 r7162 | autrijus++ |   particularly something that hIDE or yi would accept, so perl6
08:48 svnbot6 r7162 | autrijus++ |   could become a scripting language for them.
08:48 Juerd This is the computer instructing me to do something
08:48 Juerd That is the wrong way around.
08:48 nothingmuch because it will cause behavior that is unpredictable to them
08:48 Juerd The computer is my slave, I am not its.
08:48 nothingmuch i'm not saying right or wrong, it's just that this is OSX, and it won't change
08:49 Juerd Unpredictable? Hell no, the changes aren't even active until you restart.
08:49 dduncan Windows PCs also have a physical button to eject floppy disks etc, so you can easily make them come out while the R/W heads are still working on them, potentially damaging the disk; Apple didn't give that option
08:49 nothingmuch right, but serve users who don't know how things work you need to cut some choices
08:49 Juerd dduncan: That's not a software issue, it's a hardware issue. It has nothing at all to do with either Microsoft or Windows.
08:50 dduncan it had to do with Apple not giving people certain choices, though
08:50 Juerd nothingmuch: Then give the dialog a red "close" button, for power users. It doesn't confuse any newbie, but provides me with enough of a hint to know that it'll rid me of this thing.
08:50 Juerd Anyway
08:50 Juerd I've had this Apple for half a year now
08:50 Juerd And it was nice to finally get to know OS X
08:50 Juerd It's not for me.
08:51 Juerd It's not even for my grandpa.
08:51 lao has quit IRC ("Leaving")
08:51 dduncan does your grandpa have Linux too?
08:51 nothingmuch fair enough =)
08:51 Juerd Both me and my digi-analphabetical grandpa are much better off with Ubuntu.
08:51 Juerd dduncan: Sure. Windows is too hard and complex for him.
08:51 Juerd The man's 81.
08:51 nothingmuch so what makes OSX bad for him?
08:51 Juerd He must have few choices, and a system that works.
08:52 evalbot_7161 has quit IRC (Remote closed the connection)
08:52 Juerd nothingmuch: The way too prominent difference between window and application.
08:52 evalbot_7162 has joined #perl6
08:52 Juerd It's hard for some people to grok that if you close a window, the application is still running.
08:52 nothingmuch ah
08:52 Juerd And he finds it hard to predict whether clicking a dock icon will or will not open a new window
08:53 Juerd This is the same problem, but in a different guise
08:53 dduncan and other people are confused by menu bars for multiple apps being visible at the same time, in their windows
08:53 dduncan on the Mac, the menu bar shows the front most application all the time
08:53 Juerd Oh, he's just confused by the weird menu bar that is attached to windows semantically, but not visually
08:53 dduncan and there aren't others to confuse them
08:54 Juerd dduncan: The concept of rectangles that serve to delimit windows works really well.
08:54 nothingmuch yes and no
08:54 nothingmuch these are issues which are completely cultural
08:54 Juerd The MacOS menu bar is outside the window, but part of it.
08:54 nothingmuch there is no point arguing over them
08:54 dduncan often, the same app has multiple windows
08:54 Juerd dduncan: Newbies are single taskers.
08:54 Juerd All of them are.
08:55 Juerd People needing simplicity don't have two windows visible at the same time. That's confusing.
08:55 integral sounds like the original Machintosh interface
08:55 dduncan what if they want to compare 2 documents side-by-side?
08:55 Juerd But when they do get multiple windows, not out of choice, it should be damn obvious which widget belongs to which window.
08:55 Juerd dduncan: They switch, via the task bar, or via expose.
08:55 dduncan integral, not quite
08:55 integral dduncan: they print them both out and use a red pen!  or at least my mum does
08:55 dduncan the original Mac OS only had one app running at a time
08:56 dduncan but any app could always have multiple windows, and the menu bar was separate on the screen
08:56 integral ok, the "multitasking" one after that, each app just used the full screen still
08:56 Juerd My dad mostly dislikes OS X because of its exaggerated font smoothing which he finds ugly...
08:56 Juerd But then, the guy dislikes Gnome too just because it's ugly
08:56 dduncan that can be turned off
08:57 Juerd dduncan: And then you get no smoothing at all
08:57 nothingmuch windows is pretty? ;-)
08:57 Juerd There should be something between the lightest smoothing and it turned off
08:57 dduncan I think smoothing can also be adjusted
08:57 Juerd This something is usually called anti aliassing in other systems
08:57 Juerd Apple's smoothing is just a blur, it's not actual anti aliasing.
08:57 nothingmuch uh, wtf?
08:58 dduncan oh really?
08:58 dduncan that would seem very unlike them
08:58 Juerd A straight line should not have a half-transparent line along it.
08:58 integral gnome's one's very good at getting that right
08:58 Juerd integral: That's not Gnome's, but X's
08:59 Juerd dduncan: If you happen to have Terminal.app open right now, with ssh running in it, look at the h
08:59 integral yes, but gnome insists on running on X
08:59 dduncan done
08:59 Juerd dduncan: There's a gray line left of it. It shouldn't be there - it hinders clarity.
08:59 Juerd The h in the title bar, by the way :)
08:59 * nothingmuch turns smoothing off for the terminal
09:00 dduncan I can see the grey line in the terminal window itself
09:00 Juerd nothingmuch: Yes, it's the title bar, not the terminal
09:00 dduncan where I type it in
09:00 nothingmuch but for everything else it has never made me sad
09:00 integral it's very noticable on the — in the title
09:00 Juerd nothingmuch: You like beach balls, then? :)
09:01 dduncan now that its running, in the title bar I can see a slight grey on the right side, but its more subtle ... easier to read
09:01 Juerd dduncan: Anyway, this is very bad for clarity and contrast. People with bad vision who are beyond what lenses can fix, really do notice these little things more than other people.
09:01 dduncan I have a flat panel display, if that makes any difference
09:01 dduncan mind you, I always operate at a lower resolution than my screen is capable of, due to *my* poor vision
09:01 Juerd AA is good, "smoothing" (blurring) is not.
09:01 dduncan in this case, 1024 rather than 1280
09:02 dduncan on 17"
09:02 Juerd At lower resolution, especially on a flat panel, it's very visible.
09:02 Juerd Has it never bothered you at all?
09:02 dduncan I can see it, but it doesn't really bother me
09:02 Juerd It bothers me because I know the theory, but isn't really a problem for me
09:02 Juerd (Like bad spelling)
09:03 Juerd Anyway
09:03 Juerd I'm looking forward to installing kubuntu on this machine
09:03 Juerd I still do like the hardware - a lot.
09:03 dduncan never heard of that one
09:03 Juerd Of what?
09:03 dduncan kubuntu
09:03 dduncan is the k a typo?
09:03 Juerd No
09:03 Juerd Ubuntu is ubuntu-base + ubuntu-desktop
09:03 svnbot6 r7163 | autrijus++ | * remove Pugs.External.C at prompt from SyntaxNinja.
09:03 Juerd Kubuntu is ubuntu-base + kubuntu-desktop
09:03 dduncan okay
09:04 Juerd Kubuntu uses KDE, not Gnome.
09:04 dduncan so KDE is not a desktop of its own?
09:04 dduncan how is that different from "kubuntu-desktop"?
09:04 Juerd kubuntu-desktop is a meta-package that depends on the packages that together make up kubuntu
09:05 Juerd One of these packages is KDE
09:05 Juerd It doesn't install the full KDE suite, though, because ubuntu, and thus kubuntu, do try to keep things simple and grokkable.
09:05 dduncan so then, "ubuntu-base" by definition does not include a desktop?
09:05 Juerd Indeed
09:05 Juerd It's the base system
09:05 Juerd Usable for both workstations and servers
09:06 dduncan I guess that base system only includes a command line
09:06 Juerd And for people who don't like meta packages :)
09:06 dduncan with the GUI desktop an optional install for a desktop
09:06 Juerd Well, the default installation for ubuntu is ubuntu-desktop
09:06 Juerd This is what you get if you don't choose "server"
09:07 Juerd You can then choose to install kubuntu-desktop, and remove ubuntu-desktop's packages manually
09:07 dduncan would you consider Ubuntu to be the most user-friendly Linux distro?
09:07 Juerd You can also get a Kubuntu cd instead of an Ubuntu cd, and then you get kubuntu-desktop out of the box
09:07 Juerd Yes, I do
09:07 evalbot_7162 has quit IRC (Remote closed the connection)
09:07 Juerd Kubuntu a little less so, but it's nicer to work with if you do know how a computer works.
09:07 evalbot_7163 has joined #perl6
09:08 dduncan I do prefer a system with a default config that "just works"
09:08 dduncan while I like to program, I do not like to spend time configuring a system
09:08 Juerd So do I
09:09 Juerd That's why I migrated from Debian to (k)Ubuntu for workstations
09:09 Juerd I do like to have the option of being in control when I want to quickly do something, though
09:09 dduncan is there a Ubuntu live cd?
09:09 Juerd There is
09:10 Juerd It even works on macs.
09:10 Juerd (If you get the ppc one, that is)
09:10 dduncan yes
09:10 Juerd Ubuntu ships free cd's, they are in two-cd cartons: live and installation
09:10 dduncan I'll try to remember the name Ubuntu, should I want to try running a Linux desktop some day
09:10 Juerd These are pressed cd's
09:10 chihchun has quit IRC ("Terminated with extreme prejudice - dircproxy 1.0.5")
09:10 Juerd It takes a month at least to get them
09:11 Juerd But you can get 10 or even 100 at a time.
09:11 Juerd To hand them out to friends.
09:11 integral the linux version of aol...
09:11 Juerd Don't bother trying to get a single one. It's cost ineffective :)
09:11 Juerd integral: It works well.
09:11 dduncan probably cheaper to download and burn it myself
09:11 Juerd dduncan: Cheaper than free?
09:11 integral Juerd: it does.  I use it on a couple of desktops
09:11 integral also dead easy to install for X terminals
09:11 Juerd dduncan: It'll certainly be faster to download and burn it, though :)
09:12 dduncan that's what I mean
09:12 dduncan but are you also saying you can get CDs in the mail at no cost?
09:12 Juerd Yes
09:12 dduncan I thought they always charged for those, to cover pressing and shipping costs to them
09:12 Juerd You get a box with in it, 10 or more cartons
09:12 Juerd No, it's completely free.
09:12 Juerd Canonical, the company behind Ubuntu, sponsors it.
09:13 dduncan I see
09:13 Juerd That's only for ubuntu, though
09:13 dduncan anyway, its 2:13 am here, so I'm about to nod off
09:13 dduncan I see
09:13 Juerd Good night
09:14 cognominal has joined #perl6
09:14 dduncan and I leaveth ...
09:14 rantanplan_ has joined #perl6
09:14 dduncan has quit IRC ()
09:20 masak strings aren't subscriptable by default, are they?
09:20 svnbot6 r7164 | autrijus++ | * import dons's Data.FastPackedString, an amazingly fast
09:20 svnbot6 r7164 | autrijus++ |   string manipulation library.  Currently just for PIL.Str,
09:20 svnbot6 r7164 | autrijus++ |   but thinking about moving it back to Pugs.Val as well.
09:20 masak what kind of overloading would it take to make them subscriptable?
09:20 Juerd masak: Not following current design, but I'd like them to be, .chars-wise.
09:21 masak Juerd: it would certainly help in a char-based algorithm
09:21 Juerd masak: class Array { method postcircumfix:<[ ]> ($index) { .chars[$index] } }
09:21 Juerd Or something like that.
09:21 masak class Array? why?
09:22 * masak is implementing boyer-moore in p6 right now, for fun
09:22 Juerd It'd be nice if strings would just coerce to arrays as character arrays, and [] provided Array context to its left.
09:22 evalbot_7163 has quit IRC (Remote closed the connection)
09:23 Juerd masak: Oh, then ignore me. I'm talking theoretical, have no good overview of what's current, andcertainly have no idea what's actually implemented
09:23 G2 has joined #perl6
09:23 evalbot_7164 has joined #perl6
09:23 masak Juerd: oki
09:23 masak I'll just try a few things and see what works
09:23 masak there is something to be said for writing things in the normal way, too
09:26 QtPlatypus ?eval ord("A")
09:26 evalbot_7164 65
09:26 Juerd I think there's mostly something to say for changing the normal way
09:27 QtPlatypus ?eval ord("Z")
09:27 evalbot_7164 90
09:27 QtPlatypus ?eval ord("a")
09:27 evalbot_7164 97
09:27 QtPlatypus ?eval ord("z")
09:27 evalbot_7164 122
09:29 masak Juerd: if maintainability is an issue, writing things as normally as possible is very desirable
09:35 masak dang, the pseudocode would be trivial to write in p6 if the indexing weren't one-based
09:35 masak which serious language uses one-based indexing nowadays?
09:35 * masak begins at 0, as god intended
09:36 integral lots of maths...
09:42 rantanplan_ has quit IRC (Remote closed the connection)
09:43 masak integral: not really, but lots of chances of making an off-by-one mistake
09:47 integral not really?
09:47 Juerd masak: Sure, but what's normal?
09:48 Juerd masak: Is normal when the programmer writes what he means in a certain syntax, or is normal when a programmer writes things that don't resemble what he means, but are needed because of the language's limitations?
09:48 Juerd The simplest example I can give is until/if not
09:48 Juerd I love that I can choose.
09:48 Juerd eh
09:48 Juerd unless/if not
09:48 Juerd and until/while
09:49 Juerd I do usually write if not, but I do feel limited in a language that has only if.
10:11 qwr unless not ;)
10:13 masak Juerd: I actually agree with you there, and it's one of the places where I feel that damian is being overly strict in his PBP book
10:14 masak (he doesn't like unless/until because they tend to be difficult to maintain)
10:14 masak & #lunch
10:30 QtPlatypus ?eval given "T" {when "A" le $_ le "Z" {say "Test"}}
10:30 evalbot_7164 undef
10:31 QtPlatypus ?eval given "T" {when "A" ge $_ ge "Z" {say "Test"}}
10:31 evalbot_7164 undef
10:31 QtPlatypus ?eval given "T" {when ("A" le $_ le "Z") {say "Test"}}
10:31 evalbot_7164 undef
10:32 QtPlatypus ?eval given "T" {when ("A" le $_ ) {say "Test"}}
10:32 evalbot_7164 undef
10:32 QtPlatypus ?eval given "T" {when (1) {say "Test"}}
10:32 evalbot_7164 undef
10:32 QtPlatypus ?eval given "T" {when /T/ {say "Test"}}
10:32 evalbot_7164 Error: cannot cast from VUndef to Pugs.AST.Internals.VCode (VCode)
10:32 QtPlatypus ?eval given "T" {when "T" {say "Test"}}
10:32 evalbot_7164 Test bool::true
10:39 Juerd masak: I still have to read that book
10:39 Juerd It feels weird to be spending money on a style guide though
10:40 Juerd And I think a style guide without "my" in its title is just wrong.
10:47 SamB has joined #perl6
10:49 * knewt builds ghc on his new laptop. gonna take a while methinks :)
10:50 svnbot6 r7165 | clkao++ | UTF8.PackedString is no more
10:51 clkao oops, that doesnt fix the build
10:53 evalbot_7164 has quit IRC (Remote closed the connection)
10:53 evalbot_7165 has joined #perl6
10:57 iblechbot has quit IRC (Connection timed out)
11:09 ruz has quit IRC (Read error: 110 (Connection timed out))
11:19 Odin- has quit IRC (Connection timed out)
11:23 pdcawley_ has joined #perl6
11:28 iblechbot has joined #perl6
11:28 wolverian QtPlatypus, 'when "A" le $_"' means 'if $_ ~~ "A" le $_' which probably doesn't make any sense
11:30 * QtPlatypus nods.
11:31 QtPlatypus wolverian: Is there anyway to do something like that?
11:40 spinclad ?eval given 't' { when 'a'..'z' { say 'in' }}
11:40 evalbot_7165 undef
11:41 Juerd QtPlatypus: when .ge("A"), perhaps
11:42 spinclad ?eval $*VERSION
11:42 evalbot_7165 \undef
11:42 spinclad ?eval $?VERSION
11:42 evalbot_7165 Error: Undeclared variable: "$?VERSION"
11:43 spinclad .oO (how's that go again?)
11:51 autrijus ?eval $?PUGS_VERSION
11:51 evalbot_7165 \'Perl6 User\'s Golfing System, version 6.2.9, August 3, 2005 (r7163)'
11:52 autrijus clkao: the build doesn't build for you?
11:53 spinclad (didn't build for 7164 either?)
11:53 G2 Wow, Autrijus is here.
11:54 evalbot_7165 has quit IRC (Remote closed the connection)
11:54 evalbot_7166 has joined #perl6
11:54 spinclad ?eval $?PUGS_VERSION
11:54 evalbot_7166 \'Perl6 User\'s Golfing System, version 6.2.9, August 3, 2005 (r7163)'
11:54 autrijus G2: hi.
11:55 G2 autrijus: Hi. This is quite cool. You're famous
11:55 * G2 ends suck up
11:55 autrijus er. hi. thanks.
11:55 iblechbot has quit IRC (Read error: 110 (Connection timed out))
11:56 G2 autrijus: nevermind. I just love the community, being able to say hi to the top people
11:56 svnbot6 r7166 | autrijus++ | * Merged Data.FastPackedString with Simon Marlow's FPS API.
11:56 svnbot6 r7167 | autrijus++ | * Fix build by supressing DrIFT.Binary's instance for PackedString
11:56 autrijus G2: thanks :) it's indeed a lovely place, this channel
11:57 xinming_Beijing has quit IRC (Client Quit)
11:57 G2 autrijus: Aye. On PM, you get to speak to Merlyn, Larry and all the other greats. Anyway, I'll leave you to do your thing. Keep up the great work (from ghenry on perlmonks.org - http://perlmonks.org/?node_id=386673)
11:57 Jooon G2: I usually just hang around and say "oooh" and "aaah" when I see something cool, which happens a lot :)
11:58 G2 Jooon: I know. It's very cool.
11:58 hlen has quit IRC (Read error: 110 (Connection timed out))
11:58 G2 Off for lunch. Catch you guys later.
11:58 * nothingmuch struggles with extern, c headers, and name mangling
11:58 autrijus nothingmuch: for $job or blondie or something else?
11:59 nothingmuch $job
11:59 autrijus nod
11:59 * nothingmuch spent the weekend doing charity sysadmining
11:59 nothingmuch indirectly
11:59 nothingmuch fixed up a catalyst site for a friend and wasted lots of time getting it to actually run
11:59 nothingmuch but Blondie is going to be top priority soon
11:59 nothingmuch i have big plans for the type inferrencing stuff
11:59 * spinclad deeply enjoyed backlogging the MM2 work tod^Wyesterday
12:00 nothingmuch spinclad: anything specific?
12:00 evalbot_7166 has quit IRC (Remote closed the connection)
12:00 evalbot_7167 has joined #perl6
12:01 spinclad no, just... being here when it got hammered out.  \camels++
12:03 spinclad been looking forward to it for a while, just bystanding
12:05 nothingmuch ah
12:13 spinclad nm: do you have any interest in CPS-transformations in Blondie?  such as would let you define both map and for in root terms...
12:13 svnbot6 r7168 | autrijus++ | * oops, wrong linking.
12:15 spinclad ?eval $?PUGS_VERSION
12:15 evalbot_7167 \'Perl6 User\'s Golfing System, version 6.2.9, August 3, 2005 (r7163)'
12:15 evalbot_7167 has quit IRC (Remote closed the connection)
12:15 evalbot_7168 has joined #perl6
12:16 spinclad ?eval $?PUGS_VERSION
12:16 evalbot_7168 \'Perl6 User\'s Golfing System, version 6.2.9, August 3, 2005 (r7163)'
12:16 spinclad heh
12:16 spinclad do we know the evalbot's really running a new-build pugs?
12:17 spinclad s/build/built/
12:17 autrijus disagreement would mean the build failed
12:17 nothingmuch not necessarily
12:17 nothingmuch sometimes i think that doc fixes etc would not cause a rebuild, no?
12:17 autrijus true...
12:18 nothingmuch spinclad: yes, i'm going to do CPS tranformations in Blondie
12:18 nothingmuch eventually i hope Blondie becomes somewhat like schemee
12:18 spinclad should still make a new pugs_version.h, and so a new pugs
12:18 nothingmuch hmm... good point
12:18 spinclad that would be most tasty
12:19 nothingmuch ... but with perlish semantics WRT pluggable runtimes and typing, and so forth
12:19 nothingmuch i have some pretty funky ideas about macro processing
12:20 nothingmuch where every macro is just a normal function
12:20 nothingmuch except that it's type signature is AST -> ... -> AST
12:20 nothingmuch so they can be defined in terms of higher order functions
12:20 nothingmuch whenever the Macro node happens at compile time, the invocant function is compiled, and it gets it's uncompiled parameters as ASTs
12:21 nothingmuch and then the return value is compiled in place of the Macro
12:21 nothingmuch but this is slightly special casey
12:21 nothingmuch what I would really like is an Immediate node
12:21 nothingmuch and a Delayed node.
12:21 kolibrie has joined #perl6
12:22 nothingmuch as well as compile and apply nodes
12:22 dudley_ has joined #perl6
12:24 * spinclad notes a schemish front-end as a possible TODO
12:25 * spinclad works on his tuit supply
12:27 autrijus spinclad++ # tuit boost
12:28 penk has joined #perl6
12:28 spinclad thanks, that helps
12:28 wolverian QtPlatypus, you could just use 'if'
12:37 iblechbot has joined #perl6
12:46 QtPlatypus wolverian: I have.  It ends up being a big ugly 6 line if {...} elsif {...} else {...} thingy
12:50 xinming_Beijing has joined #perl6
12:51 Juerd QtPlatypus: when condition ?? $_ :: "\0" { ... } ;)
12:51 Juerd eh
12:51 Juerd QtPlatypus: when condition ?? $_ !! "\0" { ... } ;)
12:54 wolverian oh, well. why isn't when "a".."z" working, anyway? :)
12:56 Juerd ?eval "a".."z"
12:56 evalbot_7168 ('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', 'z')
12:56 Juerd ?eval "b" ~~ ("a".."z")
12:56 evalbot_7168 bool::false
12:56 Juerd Because it's broken :)
12:56 wolverian right. :)
12:56 Juerd ?eval 3 ~~ (1..9)
12:56 evalbot_7168 bool::false
12:56 wolverian ?eval 3 ~~ [1..9]
12:56 evalbot_7168 bool::false
12:56 Juerd ?eval [1..9] ~~ 3
12:56 evalbot_7168 bool::false
12:58 wolverian ?eval (1..9) ~~ 3
12:58 evalbot_7168 bool::false
12:59 wolverian hm. where is &infix:<~~> defined?
13:03 nnunley has joined #perl6
13:03 icyvapor has joined #perl6
13:06 icyvapor anyone know about processing large-scale dataset using perl?
13:07 integral using perl6?
13:07 nnunley Junctions.
13:07 integral LINQ in c#, maybe
13:07 autrijus icyvapor: which shape of data?
13:08 icyvapor how can I check whether I'm using perl6
13:08 icyvapor flat text files
13:08 nnunley icyvapor: Unless you install perl 6, it's unlikely that you would be using it by accident.
13:09 icyvapor :) then I'm not, the perl I'm using comes with linux
13:09 integral icyvapor: well this channel, as indicated by the name, is specifically for perl6, Perl 6 and related bits
13:09 nothingmuch booya
13:10 mauke has quit IRC (Remote closed the connection)
13:10 integral there's #perl, #perlhelp, #perlcafe, ##perl, etc for more general stuff
13:10 mauke has joined #perl6
13:10 * nothingmuch had to make ufc_crypt from glibc compile on windows since CryptAcquireContext is so nice
13:10 joao has joined #perl6
13:12 icyvapor ok, thanks a lot guys
13:12 wolverian nothingmuch, what is that?
13:12 nothingmuch wolverian: on windows to do the normal crypt() style hashing you need to do some pretty weird stuff
13:13 nothingmuch namely be logged in with permissions that allow you to get a handle to a remote object in a secured process
13:13 nothingmuch via CryptAcquireContext
13:13 nothingmuch makes sense, doesn't it? =(
13:13 nothingmuch anyway, since there are many ways of instantiating it, and more ways in which it could fail, and we had yet another scenario where it doesn't work
13:13 nothingmuch i just integrated a plain C implementation
13:14 icyvapor has left
13:16 wolverian nothingmuch, I .. uh .. that fits my theme of browsing we.hates-software.com today :)
13:18 iblechbot has quit IRC (Connection timed out)
13:19 nothingmuch wolverian: anything specific?
13:19 nothingmuch btw, this is the second time CryptAcquireContetxt has pissed me off
13:19 nothingmuch the last time it couldn't compute a crypt of a password because the session i was logged in through (cygwin ssh server) was not interactive
13:20 nothingmuch we now have if (!CryptAcquireContext(...)) three times with different args
13:21 wolverian nothingmuch, just taking a break from being serious
13:21 nothingmuch ah
13:21 nothingmuch so how do you post to there?
13:22 Qiang has joined #perl6
13:24 nothingmuch huh... /me didn't even notice itunes changed look again
13:28 elmex has joined #perl6
13:29 wolverian nothingmuch, http://hates-software.com/
13:29 nothingmuch yeah, i found out by now ;-)
13:29 fglock has joined #perl6
13:29 * nothingmuch suspects he will be a regular poster
13:30 wolverian :)
13:36 autrijus mm this ongoing talk is going to make iblech happy
13:37 nothingmuch how come?
13:37 autrijus a new way to implement first class continuations on utterly uncooperating VMs
13:37 nothingmuch ah
13:37 nothingmuch brief us?
13:38 autrijus compile away continuations into stack inspection primitives.
13:38 autrijus starting from ANF
13:39 autrijus the URL is:
13:39 autrijus http://www.cs.brown.edu/~sk/Publications/Papers/Published/pcmkf-cont-from-gen-stack-insp/
13:39 jql has joined #perl6
13:39 nothingmuch that sounds a bit like continuations in C on the c2 wiki
13:40 nothingmuch basically it changes the base and stack pointers to point at a different stack each time
13:40 nothingmuch and it can swap these
13:42 * nothingmuch wishes he was next to autrijus right now
13:42 nothingmuch the talks sound so interesting
13:42 autrijus it is!
13:45 * nothingmuch is pissed
13:45 nothingmuch beyahad.info is responding like crap because it can't write to an old template cache file
13:46 nothingmuch Petal is so stupid about these things... It's a really nice language for when the template is very simple
13:46 nothingmuch but the implementation is just not good enough
13:46 * nothingmuch will prolly have to switch to TT soon
13:48 xinming_Beijing autrijus: hmm, May I know what does class continuation mean please?
13:49 autrijus xinming_Beijing: "first-class" continuation
13:49 autrijus first-class means you can pass it around as a value
13:49 autrijus so his insight is basically that a disciplined use of try/catch can implement ANF-plus-continuation-marks
13:50 autrijus which is then proved to be rich enough to carry full call/cc
13:50 autrijus it's quite nice, complete with working implementation targetting C#
13:50 coet has quit IRC (Read error: 104 (Connection reset by peer))
13:50 coet has joined #perl6
13:53 xinming_Beijing autrijus: hmmm, So, does it mean. The class will act as a object automatically?
13:53 autrijus his paper mentions Parrot :)
13:53 xinming_Beijing hmm...
13:54 autrijus "Parrot went to great length to become CPS based, but we proved that normal error-handling would serve their purpose just as well"
13:54 autrijus xinming_Beijing: er, no, it has nothing to do with class and objects
13:55 autrijus xinming_Beijing: this is about continuations. think of the dump() functions.
13:55 autrijus if you can put dump()'s result into a variable, _without_ stopping the program
13:55 autrijus and you can resume from that variable later
13:55 autrijus then you get first-class continuations
13:57 autrijus so this technique is as good as CPS, but does not rely on tail calls or trampolines
13:57 autrijus so it should fit well with javascript
13:58 integral and perl5?
13:58 xinming_Beijing hmm, So, In my understanding. It makes the snapshot of a "function" or a object into a variable. And you can use this "variable" later, right?
13:58 autrijus integral: I don't know how good eval{} and die" is in term of performance
13:58 autrijus xinming_Beijing: no, read "perldoc -f dump" :)
13:59 xinming_Beijing autrijus: So, Does this mean, We can hibernate the program, And run in that state again soon.
13:59 xinming_Beijing autrijus: Ok,
13:59 autrijus yup
13:59 integral there's labelled blocks and last too
13:59 autrijus it's exactly like hibernation.
13:59 autrijus integral: can't use that in interaction with normal perl5
14:00 autrijus hm, wait, maybe we can
14:00 integral oh, not without a shim layer, but you're using the metamodel stuff anyway, so lots of stuff doesn't look normal anyway...
14:00 autrijus right exactly
14:00 autrijus and
14:00 xinming_Beijing Oh, This idea will make well-written perl 6 program fastest without close... >_<
14:00 autrijus FOO: { ... call_complex_ffi(sub { last FOO }) }
14:00 autrijus this in p5 will dtrt?
14:00 integral oh, no it won't
14:00 autrijus no?
14:01 autrijus "last-outside-sub"
14:01 integral the FOO is searched in the dynamic scope at the time the last is executed
14:01 autrijus really.
14:01 integral sorry, I think I assumed the wrong question
14:01 xinming_Beijing autrijus: Ok, I think I know that idea now...
14:02 autrijus integral: it worksforme.
14:02 xinming_Beijing autrijus FOO: { ... call_complex_ffi(sub { last FOO }) }
14:02 xinming_Beijing is this example for me?
14:02 integral autrijus: I don't quite understand what the WTRT is wrt to though.  Capturing the stack when the closure is created?  last across a Perl->XS->Perl callback?
14:04 autrijus the speaker notes no prior art on trampolining or cps+try/catch on javascript vms
14:04 autrijus xinming_Beijing: no, that was for integral
14:06 * integral &
14:07 nothingmuch autrijus: help...
14:08 nothingmuch "the relation |-->* is the reflexive transitive closure of the transition function"
14:08 nothingmuch what does that mean?
14:10 autrijus closure means the return value can be used again as the argument
14:10 autrijus transitive means that if a|->*b  and b|->*c then a |->*c
14:11 autrijus reflexive means that forall a.  a|->*a
14:11 autrijus does that help?
14:11 nothingmuch somewhat
14:12 * nothingmuch will try to compile
14:12 wolverian what is the transition function?
14:12 wolverian :)
14:13 autrijus wolverian: turing machine, apparently
14:14 fglock has left "Fui embora"
14:14 nothingmuch CEK machine
14:14 fglock has joined #perl6
14:15 coet has quit IRC (Read error: 113 (No route to host))
14:15 nothingmuch http://research.compaq.com/SRC/personal/flanagan/papers/pldi93.ps
14:16 wolverian autrijus, oh.
14:16 wolverian nothingmuch, ah.
14:17 iblechbot has joined #perl6
14:18 alexe has joined #perl6
14:20 * nothingmuch wonders if you can take a course on reading mathy articles
14:21 * nothingmuch wants to be a certified lambdexic
14:21 nothingmuch lambdlexic, sorry
14:21 SM_ax has quit IRC ("Download Gaim: http://gaim.sourceforge.net/")
14:22 nothingmuch so what does "the relation" is the <adjectives> closure of the translation function mean"?
14:22 nothingmuch beh, quoting bugs
14:23 nothingmuch a relation is a closure of a function
14:23 elmex has quit IRC ("leaving")
14:23 nothingmuch closure means an ad-noun describing a property of the function
14:23 brentdax has quit IRC (Read error: 110 (Connection timed out))
14:24 nothingmuch autrijus: help more
14:24 nothingmuch ;-)
14:25 fglock how about "the set of things that can be both arguments or results of the function" (is this correct?)
14:25 nothingmuch hmm, that is begining to make sense
14:26 autrijus mm after this talk the all-exciting icfp contest result will be revealed
14:28 nothingmuch the robots contest?
14:28 autrijus yeah
14:28 nothingmuch sounds fun
14:29 autrijus one wonders which language will win this year... not perl6, that we are sure
14:29 autrijus acme played in it iirc
14:29 autrijus with perl5
14:29 nothingmuch atomic { nothingmuch.tuits += acme.tuits; acme.tuits = 0 };
14:30 nothingmuch AG made me think a bit
14:30 brentdax has joined #perl6
14:30 nothingmuch lots of things in perl 6 are hard to compile
14:30 nothingmuch atomic { } requires that everything is stmish
14:30 nothingmuch eval makes things tough
14:30 nothingmuch etc etc
14:30 nothingmuch what if blocks were attributing through callgraphs
14:31 nothingmuch e.g. atomic { } adds an attribute to all assignments that can be called from it
14:31 nothingmuch assignment adds the attribute to all lvalues
14:31 nothingmuch SSA Phi compression of lvalues allows to determine which containers need logging and which ones are never shared
14:31 Odin-LAP has joined #perl6
14:32 * nothingmuch will try to add leverage Language::AttributeGrammar in Blondie to make such conversions possible
14:32 hexmode has joined #perl6
14:32 nothingmuch compute how little work can be done for a given program
14:37 nothingmuch aha... our compsci guru explained it to me
14:39 justatheory has joined #perl6
14:40 kolibrie Juerd: I didn't quite grasp the full meaning of: K vnd grtz n btj mljk t lzn, n d z mt ntrljk n s zjn
14:40 kolibrie Ik vnd grtz een beetje moeilijk te lezen, ...
14:41 nothingmuch is that dutch with no vowels?
14:41 kolibrie believe so
14:42 kolibrie except for the grtz, which is a quote from an e-amil
14:42 mauke hmm, "mljk t lzn" could be "possible to read"
14:43 kolibrie oh, I can translate from dutch to english, I want the rest of the dutch
14:43 kolibrie "I find grtz a little difficult to read, ..."
14:43 mauke oops, didn't see your second line :-)
15:02 nothingmuch has quit IRC (Read error: 110 (Connection timed out))
15:09 joao has quit IRC ("This computer has gone to sleep")
15:11 joao has joined #perl6
15:18 Juerd kolibrie: en die z moet natuurlijk een s zijn
15:23 alexe has left
15:24 autrijus judges's prize: dylan hackers
15:26 autrijus 3rd: Combat Tantalin (haskell)
15:26 autrijus 2nd same as judge's prize
15:26 kolibrie Juerd: ah, thank you
15:28 autrijus the dylan hackers are giving the audience an overview of the language
15:31 acme__ pesky lack of winning
15:31 autrijus 1st prize: Wolfgang Thaller et al, Haskell, two years running
15:31 clkao poor team orange
15:32 acme__ they didn't use my much-prettier-maps
15:33 G2 has quit IRC ("oops")
15:37 acme__ eg http://icfpc.cs.uchicago.edu/reg/026/robber/ vs http://astray.com/icfp/game/robber_vs_randomcar/mark_wutka/10
15:43 stevan dylan++ # very cool language
15:43 stevan good morning boys and girls
15:47 GeJ has quit IRC ("Client exiting")
15:54 fglock hi stevan
15:55 nothingmuch has joined #perl6
15:55 stevan hey fglock
15:57 * nothingmuch doesn't get a greeting?
15:58 fglock hi nothingmuch!
15:58 stevan hi nothingmuch
15:58 nothingmuch and I thought you didn't love me for a second there ;-)
15:58 stevan nothingmuch: I will never stop loving you ;)
15:58 nothingmuch stevan: to prove it want to fly me to euro-oscon?
15:59 * stevan was messing around with lambda calc till very late last night
15:59 stevan nothingmuch: I suggest you start walking now,.. you might make it
15:59 * stevan would love to go to EURO-OSCON too
15:59 nothingmuch yep... the lebanese/syrian boundry is no problem what so ever
16:00 nothingmuch s/boundry/border/
16:00 stevan you could pass
16:00 stevan with your accent
16:00 nothingmuch hehe
16:00 mshiltonj has joined #perl6
16:01 nothingmuch also, amsterdam is pretty far away
16:01 nothingmuch 20 days is not much
16:01 stevan bike then
16:01 nothingmuch i could possibly bike there
16:01 nothingmuch yeah
16:02 mshiltonj is there a place i can find info on the status of perl6 rfc 89 -- http://dev.perl.org/perl6/rfc/89.html -- if it is being implemented, etc.
16:02 stevan take gaal with you and you can cut throught he turkish mountains
16:02 nothingmuch mshiltonj: i doubt you'll find a formal status
16:02 nothingmuch the de facto situation on typing is pretty advanced though
16:03 nothingmuch (not as advanced as it's going to be but we do have some things)
16:03 nothingmuch but the rfc itself seems pretty far from what will eventually get implemented
16:04 mshiltonj k. but data typing of some sort will be there?
16:04 nothingmuch yes
16:04 nothingmuch i'm off to dinner
16:04 mshiltonj cool beans. thanks.
16:04 nothingmuch but i can explain a bit later if you like
16:04 kolibrie stevan: If nothingmuch and gaal bike to EuroOSCON, could we bike too?
16:07 fglock mshiltonj: rfc 089 was rejected (A04)
16:11 mshiltonj didn't know that. I'm interested in whether or not there will be data typing in perl6. If it's in perl, the specific implementation can vary.
16:17 fglock mshiltonj: see the second half of S06 - it's about the Type system
16:20 GoCooL has joined #perl6
16:22 stevan kolibrie: we would need to swim a little too :)
16:30 nothingmuch evening
16:32 mshiltonj fglock: thanks!
16:36 fglock np
16:37 evalbot_7168 has quit IRC (Remote closed the connection)
16:38 evalbot_7169 has joined #perl6
16:38 svnbot6 r7169 | autrijus++ | * first step at cabalizing Pugs -- first try at PIL (which works)
16:38 nothingmuch cabalizing?
16:38 autrijus nothingmuch: the eqv of M::B in Hs
16:38 autrijus enables libPugs.a
16:38 rantanplan_ has joined #perl6
16:38 autrijus which means much easier linking
16:38 nothingmuch ah
16:38 autrijus nonbroken pugscc
16:39 autrijus and perl6 scripting for all hs programs
16:52 nothingmuch sounds fun
16:52 Limbic_Region has joined #perl6
16:54 gaal hola
16:55 Limbic_Region kumusta ma na?
16:55 Limbic_Region err kumusta ka na even
16:56 gaal Limbic_Region: what language is that? :)
16:57 DesreveR has joined #perl6
17:03 fglock Limbic_Region: Magandang hapon
17:04 nothingmuch autrijus: how much effort do you estimate Inline::GHC will take?
17:06 Limbic_Region gaal - tagalog
17:06 mshiltonj has left
17:07 Limbic_Region and good localtime to you too fglock
17:13 r0nny has quit IRC (Read error: 110 (Connection timed out))
17:14 joao has left "Leaving"
17:20 elmex has joined #perl6
17:29 gaal autrijus: do we have some sort of equivalent to p5's PL_compiling / PL_curcop?
17:35 * geoffb fades in
17:35 geoffb autrijus, ETOOMUCHCOMPRESSION: the eqv of M::B in Hs
17:35 integral the equivalent of Module::Build in Haskell.
17:35 geoffb Equivalent and Haskell decompressed, but M::B wasn't in the dictionary window
17:35 geoffb integral, thanks
17:36 nnunley has quit IRC (Read error: 113 (No route to host))
17:36 geoffb .oO( Man, it is *so* nice to see multiple commits per day from autrijus again )
17:38 * stevan agrees very much with geoffb
17:38 geoffb (leased laptops)--
17:39 Limbic_Region s/very much/vehemently/
17:39 fglock has quit IRC ("CGI:IRC (Ping timeout)")
17:43 geoffb tea++
17:43 geoffb Life is just better with a warm cup of tea.
17:44 DesreveR is now known as r0nny
17:57 nothingmuch luqui: ping
18:00 r0nny has quit IRC ("restart")
18:00 r0nny has joined #perl6
18:03 rafl has quit IRC (Read error: 60 (Operation timed out))
18:07 ruz has joined #perl6
18:08 nothingmuch did luqui ever mention if L::AG has a public repo somewhere?
18:28 Limbic_Region L::AG ?
18:28 Limbic_Region is that the module he just released?
18:30 Limbic_Region ah yes,  Language::AttributeGrammar
18:30 dada has quit IRC ("you said you wanted evolution, the ape was a great big hit")
18:33 kgftr|konobi stevan: does metamodel also help describe relationships between objects or just the objects and their structures?
18:34 stevan kgftr|konobi: it is the object model itself,.. which is both
18:35 pdcawley has quit IRC ()
18:52 kgftr|konobi stevan: hhhmmm... k
18:52 kgftr|konobi would that include things like data inheritance?
18:53 stevan yes sir
18:54 stevan kgftr|konobi: there is a MM version of Test::Buidler in the ext/ directory (under perl5/Perl6-MetaModel2.0/
18:54 kgftr|konobi =0o
18:54 stevan it is kind of ugly, but it demonstrates all the functionality needed by Test::Builder
18:55 kgftr|konobi then i'll wrap MetaModel with Data::Model to then wrap Class::DBI/Tangram/DBIx::Class objects
18:55 stevan LOL
18:56 stevan uhm
18:56 stevan I wouldnt suggest it with this metamodel
18:56 stevan it is optimized for theoretical correctness
18:56 stevan so it is largely unusable from any other p5 apps
18:57 stevan however, I am planning to make a p5 usable version of it in the very near future
18:57 stevan cause after this,.. I have no desire at all to return to the p5 OO world
18:58 Limbic_Region out of curiosity stevan
18:58 kgftr|konobi =0o
18:58 * stevan give Limbic_Region some of his curiosity
18:58 Limbic_Region using MM Objects in p5 will be how easy
18:58 Limbic_Region package Some::Object;
18:58 stevan Limbic_Region: not sure yet
18:58 Limbic_Region use Meta::Model;
18:58 Limbic_Region and then ?
18:59 Limbic_Region oh - well, nevermind then
18:59 Limbic_Region I will wait until you are finished
18:59 stevan Limbic_Region: I would like to make it as simple as possble
18:59 buu_ has joined #perl6
18:59 stevan Limbic_Region: see Class::C3 for an example
18:59 buu has quit IRC (Read error: 104 (Connection reset by peer))
18:59 Limbic_Region salutations buu
18:59 stevan that swaps the normal method dispatch for C3
18:59 Limbic_Region stevan - I am weeks, if not months away from needing it
19:00 stevan its kind of hackish,.. but it is totally transparent
19:00 kgftr|konobi stevan: we'll need to have a chat at some point... to avoid duplicating efforts... but need to wait for me to get my mac
19:00 stevan which is what I would like the meta-model to be
19:00 stevan kgftr|konobi: what are you doing?
19:03 Limbic_Region stevan - I have only pedagogal needs ATM - when I am at the point of discussing OO and how it got shoehorned into p5, I would like to be able to point to an alternative ;-)
19:03 kgftr|konobi stevan: providing standard interfaces and model descriptions for both persistent objects, etc and realtime objects, etc.
19:03 kgftr|konobi but in p5
19:03 * kgftr|konobi ducks
19:03 stevan nice # both of you guys
19:04 kgftr|konobi should help me a bit more to make objects of different types to be more interoperable
19:05 andynu has quit IRC ("leaving")
19:15 dduncan has joined #perl6
19:17 fglock has joined #perl6
19:19 rafl has joined #perl6
19:32 nnunley has joined #perl6
19:39 G2 has joined #perl6
19:44 buu_ is now known as buu
19:46 dudley_ has quit IRC ()
19:59 gaal has quit IRC ("leaving")
20:03 svnbot6 r7170 | stevan++ | just need to commit this for the other changes to work ...
20:04 evalbot_7169 has quit IRC (Remote closed the connection)
20:04 evalbot_7170 has joined #perl6
20:07 Odin- has joined #perl6
20:08 rantanplan_ has quit IRC (Remote closed the connection)
20:08 svnbot6 r7171 | stevan++ | Perl6::MetaModel 2.0 -
20:08 svnbot6 r7171 | stevan++ | * over 50% speedup from optimizing chaos.pl, old chaos.pl moved
20:08 svnbot6 r7171 | stevan++ |   to chaos_unoptimized.pl for reference
20:09 evalbot_7170 has quit IRC (Remote closed the connection)
20:09 evalbot_7171 has joined #perl6
20:18 autrijus    greetings
20:18 Limbic_Region salutations
20:18 stevan hey autrijus
20:19 Limbic_Region stevan - did that speed increase also lead to clarity in the code?
20:19 stevan Limbic_Region: LOL
20:19 stevan a severe decrease in fact
20:19 gaal has joined #perl6
20:19 Limbic_Region I wasn't trying to be facetious
20:19 stevan there are several functions which are called for each and every method call,..
20:20 autrijus if it becomes clearer, _unoptimized.pl wouldn't be there :)
20:20 Limbic_Region typically huge speed ups cause readability to change in either direction - or at least it has been my experience
20:20 stevan I optmized those to the point of unreadablility in some cases :)
20:20 Limbic_Region it either gets more readable because the crap has been removed and the elegance remains
20:20 stevan my personal favorite is WALKMETH
20:20 stevan I will nopaste the change so you can see
20:21 Limbic_Region or it gets fugly because the elegant solution just doesn't have what it takes to be fast
20:21 gaal golf physics, 100 years ago today: http://www.fourmilab.ch/etexts/einstein/E_mc2/www/
20:21 gaal (sorry, completely off topic, but hey.)
20:22 pasteling "stevan" at 67.186.136.119 pasted "changes in WALKMETH" (25 lines, 770B) at http://sial.org/pbot/13333
20:22 stevan Limbic_Region: as you can see,.. the optimized version is difficult to read at best
20:23 stevan but that one alone was worth about 30%
20:23 stevan so it was worth it
20:23 Limbic_Region yeah
20:23 Limbic_Region but a lines of comment could make it just as readable and fast
20:23 autrijus ...for some reason the optimized one is easier to read to me
20:23 autrijus right, what Limbic_Region said.
20:23 Limbic_Region I was thinking it actually became obfusucated
20:23 gaal autrijus, i'm unsure about where to carry pragma list state in the parser
20:24 autrijus gaal: where else then?
20:24 autrijus in Env as $* ?
20:24 gaal heh, no, i meant where in the parser to put it :)
20:24 stevan Limbic_Region: the worst obsfucations in it are when I avoid unpacking sub parameters and use the @_ index
20:24 stevan you would be supprised how much that shaves off if you do it enough
20:24 evalbot_7171 has quit IRC (Remote closed the connection)
20:25 svnbot6 r7172 | stevan++ | Perl6::MetaModel 2.0 -
20:25 svnbot6 r7172 | stevan++ | * fixing the mutli-methods test
20:25 stevan that and some of the if/else statements were converted to boolean expressions
20:25 evalbot_7172 has joined #perl6
20:25 Limbic_Region stevan - I wasn't talking about your particular optimization - just optimizations in general that lead to less readability - they become unreadable
20:25 Limbic_Region in your case it is easy to read
20:25 gaal the c patch does something i thought i grokked but now i think i don't, puts the list on the current cop; when a scope ends you automagically get the head pointer that was there just before the scope started
20:25 Limbic_Region but not necessarily easy to understand
20:25 Limbic_Region adding the comments achieves both
20:25 stevan Limbic_Region: very true
20:25 gaal there's absolutely no code that handles that explictly, that's beautiful
20:26 gaal i don't think we have an analogous structure though (so we?)
20:26 autrijus gaal: cop?
20:26 autrijus cop would just be annotation nodes.
20:26 autrijus but you are thinking of scope markers
20:26 autrijus and those would be Stmt boundaries I think
20:27 autrijus I have no problem of extending Stmts type
20:27 autrijus so you can write between the lines
20:27 autrijus so to speak
20:27 autrijus that will also get us goto LINE: support.
20:27 * gaal grins
20:27 autrijus Stmts Cop Exp Exp
20:27 autrijus not neccessarly with the Cop name
20:27 autrijus but you get the idea.
20:28 gaal not quite sure that i do yet :/
20:28 autrijus oh ok... the idea is that Stmts are cons for statements
20:28 gaal when i constuct a Stmts Exp, I need to copy the Cop from the last one.
20:28 autrijus Stmts exp1 (Stmts exp2 (Stmts exp3 Noop))
20:29 autrijus now we regulate that Cop are cumulative
20:29 gaal that part i do get
20:29 gaal oh
20:29 autrijus Stmts cop1 exp1 (Stmts cop2 exp2 (...))
20:29 * Limbic_Region notes the time and heads homewards
20:29 autrijus exp2 gets (cop2 . cop1)
20:29 gaal that point i *don't* :)
20:29 Limbic_Region stevan++ # good work
20:29 Limbic_Region has quit IRC ("tiuq\")
20:29 autrijus makes some sort of sense?
20:30 gaal where Cop is something like MkCop { Label :: String, Pragmas :: [Pragma], ... } ?
20:31 gaal why does it need to be cumulative?
20:31 gaal and, how does that get us scopy behavior?
20:31 autrijus it doesn't, to think about it. we can share structure
20:32 autrijus scope is easy; any Stmts cons belong to the same scope
20:32 gaal for lex prags, we do indeed share structure, as the list is immutable once installed; and *as* it's being installed it only gets prepended to.
20:32 autrijus because under scope boundary it will be wrapped in a App's body
20:32 gaal (if that's what you mean by share structure)
20:33 autrijus yes that's what I mean.
20:33 autrijus so it's just a ponter to a common heap pointer
20:33 brentdax has quit IRC (Remote closed the connection)
20:33 autrijus { 1; { 2 } }
20:33 autrijus the 2 is inside a App
20:33 autrijus of sub{2}
20:33 autrijus so we can easily see scope boundary
20:34 gaal ..so how do i know the latest value of the pragma list when i am consuming 2?
20:34 brentdax has joined #perl6
20:34 autrijus you ask your outer
20:34 autrijus aka envOuter
20:35 autrijus hm, bad idea
20:35 autrijus thinking
20:35 gaal it's not in the env...
20:35 autrijus I think it need to by in env.
20:35 autrijus envCop
20:37 gaal okay, so every Stmt constructed puts its Pragmas in envCop? yes, that is kinda analogous to p5
20:37 autrijus yeah, but that leaves us wondering why not just do it in VCode
20:37 autrijus ah, because of middle-of-scope starting points
20:38 autrijus so Stmts boundary is probably the way to go.
20:39 gaal so the cop carryover should probably go in ruleBlockBody?
20:39 autrijus right, that looks very sensible
20:39 autrijus but I'm not sure "cop" is a good name.
20:39 gaal Ann?
20:40 autrijus perhaps just "label" or "ann" or "pragmas"
20:40 autrijus depending on whethe we treat pragma as a kind of label
20:40 autrijus or a label a kind of a pragma
20:40 autrijus or neither as either and call it ann
20:40 gaal though you previously mentioned maybe "ann" for unifying Cxt and Pos
20:40 gaal :-)
20:40 autrijus indeed. so maybe Label?
20:40 gaal Desc? :)
20:41 autrijus your choice. :) Desc is oka
20:41 autrijus y
20:41 gaal okay, Label first and if I think of something better I'll change it.
20:42 autrijus nothingmuch: re Inline::GHC, I think with the new libGHC stuff it'd be downright trivial
20:42 autrijus sadly it's GHC 6.6, still 3 months away
20:42 autrijus happily we're not in any particular rush for that.
20:43 clkao agent tang, i was trying to play with pil-run, but it seemed slow and not evening running "say 'ok'"
20:44 autrijus I think you are in the middle of mm2 switch.
20:44 autrijus the version 36 hours ago should work much better :)
20:45 autrijus clkao:
20:45 autrijus $ ./pugs -BPerl5 -e'say "ok"'
20:45 autrijus ok
20:45 autrijus workedforme actually
20:45 gaal my fridge just croaked.
20:45 autrijus wow.
20:46 gaal too bad, i made a big food purchase two days ago :(
20:48 clkao autrijus: it works now, you still have account on my laptop haven't you :)
20:49 autrijus I see
20:53 fglock has quit IRC ("CGI:IRC")
20:55 fglock has joined #perl6
21:00 evalbot_7172 has quit IRC (Remote closed the connection)
21:00 svnbot6 r7173 | rafl++ | * Fixed shebang for examples/irclog.html.p6
21:00 evalbot_7173 has joined #perl6
21:01 clkao 0/win21
21:05 evalbot_7173 has quit IRC (Remote closed the connection)
21:05 svnbot6 r7174 | rafl++ | * Fixed shebang for ext/Algorithm-TokenBucket/t/test.t.
21:05 evalbot_7174 has joined #perl6
21:08 gaal if we now have Stmts Label Exp Exp, how to tweak mergeStmts not to lose label information?
21:09 gaal or should the envLabel-writing be done there?
21:12 kolibrie has quit IRC ("leaving")
21:13 mauke has quit IRC (Remote closed the connection)
21:13 mauke has joined #perl6
21:13 gaal zzzZZZZ &
21:26 G2 has quit IRC ("oops")
21:27 putter has joined #perl6
21:28 putter stevan: it looks like   use Perl6::MetaModel; print eval("my \$Foo = \$::Class->new('\$.name'=>'Foo'); \$Foo->superclasses([ \$::Object ]); \$Foo->new();");   is still printing #<AnonClass=(9)>, which seems unlikely to be correct.  Any thoughts?
21:34 putter 08_multi_methods.t is now all ok. :)
21:38 putter my, that _is_ sprightly.  stevan++
21:40 putter Current performance should be fine for the indefinite future.  Thanks stevan.
21:41 geoffb I take it the improvement was better than the 50% he claimed?
21:42 putter hopefully I'll be able to spend an hour or so this evening (t+5hr) debugging pilrun on mm2.
21:42 putter geoffb: not sure.  subjective impression is yes.  but you know how reliable those are.
21:42 geoffb nodnod
21:44 putter ok.  cheers &
21:44 putter has quit IRC ("Leaving")
21:44 geoffb cheers
21:51 evalbot_7174 has quit IRC (Remote closed the connection)
21:51 evalbot_7175 has joined #perl6
21:54 svnbot6 r7175 | fglock++ | * perl5/List - 'a'..'zzz' is now supported
21:56 evalbot_7175 has quit IRC (Remote closed the connection)
21:56 evalbot_7176 has joined #perl6
21:57 buu Yay!
21:57 buu Now what about "b"--
21:58 geoffb It's Perl 5, buu, go crazy.  :-)
21:59 fglock buu: "b" is a constant, but $a="b"; $a-- should work (it doesn't, currently)
21:59 iblechbot has quit IRC (Connection timed out)
21:59 svnbot6 r7176 | fglock++ | * perl5/Value - log update
22:03 nnunley has quit IRC (Read error: 111 (Connection refused))
22:03 hexmode has quit IRC (Read error: 104 (Connection reset by peer))
22:05 svnbot6 r7177 | fglock++ | * pil-run - fixed $a++ and $a-- when $a is a Str
22:06 evalbot_7176 has quit IRC (Remote closed the connection)
22:06 evalbot_7177 has joined #perl6
22:07 r0nny has quit IRC ("sleep")
22:11 svnbot6 r7178 | fglock++ | * pil-run - fixed prefix "++" and "--".
22:11 svnbot6 r7178 | fglock++ |   - postfix "++" and "--" are broken - need Scalar.clone() to work properly
22:11 evalbot_7177 has quit IRC (Remote closed the connection)
22:12 evalbot_7178 has joined #perl6
22:18 stevan has quit IRC ("Leaving")
22:19 stevan has joined #perl6
22:47 evalbot_7178 has quit IRC (Remote closed the connection)
22:47 svnbot6 r7179 | fglock++ | * pil-run - added very basic support for Junctions
22:47 svnbot6 r7179 | fglock++ |   - pil-run try to stringify Junction as a number (which is wrong)
22:47 evalbot_7179 has joined #perl6
22:51 pdcawley has joined #perl6
22:58 fglock has quit IRC ("CGI:IRC (Ping timeout)")
23:04 thomason has joined #perl6
23:05 GoCooL has quit IRC ("Leaving")
23:06 thomason_ has joined #perl6
23:06 thomason_ has left "Leaving"
23:06 thomason_ has joined #perl6
23:06 thomason_ has left "Leaving"
23:21 larsen has quit IRC ("later")
23:59 elmex has quit IRC ("leaving")

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

Perl 6 | Reference Documentation | Rakudo