svn switch --relocate svn.openfoundry.org/pugs svn.pugscode.org/pugs/ | run.pugscode.org | spec.pugscode.org | paste: sial.org/pbot/perl6 | pugs.blogs.com | dev.pugscode.org/
Set by putter on 11 February 2007.
TimToady heh, I'm famous for almost being famous, I gues... www.eweek.com/slideshow_viewer/0,12...31507EP35A 00:02
lambdabot Title: Top 100 Most Influential People in IT, Part 1 - Top 100 Most Influential People ..., tinyurl.com/2qhnnt
TimToady *guess 00:03
Aankhen`` LOL. Congrats.
TimToady or maybe almost not being famous, is more like...
geoffb Reading back to audreyt & TimToady's discussion about pads, I was confused until I realized they were refering to a 'padentry', not to 'pedantry'. 00:20
TimToady the two are not mutually exclusive 00:22
diakopter TimToady: is Conway's pod parser far off? 00:25
tene Heh. 00:28
Aankhen`` Does !~ have a particular meaning? 00:29
TimToady don't concatenate? er... 00:30
Aankhen`` I was just wondering why it can't have the same meaning as !~~
offby1 ~~~: <-- snake! Run. 00:31
TimToady Aankhen``: 'cuz I'm cruel.
Aankhen`` Ah, that explains it. :-D
TimToady diakopter: it's close, but he's had various distractions including father-in-law in the hospital with a heart attack, and generating copious feedback on current synopses 00:33
Aankhen`` Okay, quick, someone make a couple of commits so I can test my patch to svnbot. :-P 00:34
TimToady seriously, since we're making everyone change from =~ to ~~ it didn't seem like a problem to switch !~ to the more regular !~~ syntax at the same time
didn't seem worth it in the case of != or ne though 00:35
especially since != goes all the way back to C
decommuting & 00:36
Aankhen`` Makes sense; it's just that !~ is a whole character shorter. =) 00:37
TimToady heh, GoogleWiFi works pretty good in a car... 00:55
(not driving)
doesn't like sound walls though. 00:57
offby1 well, who does? 01:38
allbery_b ?eval 1 01:53
evalbot_r15727 1
allbery_b pushes the button... 01:54
svnbot6 r15728 | diakopter++ | Restoring XHTML validation for an html page. 01:58
r15728 | diakopter++ | And testing Aankhen``'s new svnbot.
Aankhen`` diakopter: I had to give up on that... couldn't get it working properly. 02:00
allbery_b @tell audreyb still getting lots of "does not export ..." messages with 6.7. giving up on it 02:12
lambdabot Consider it noted.
Khisanth audreyb? 02:13
allbery_b duh 02:14
@tell audreyt still getting lots of "does not export ..." messages with 6.7. giving up on it
lambdabot Consider it noted.
allbery_b still can';t type worth a damn :> 02:15
Khisanth has the same problem with no,now,not 02:18
allbery_b just does typoes a lot. some fatfingers, some he has no idea how they happen
allbery_b @tell audreyt the good news is, ext/HTML-Entities/t/entities.t no longer loops 04:05
lambdabot Consider it noted.
svnbot6 r15729 | bsb++ | r301@ogum: bsb | 2007-03-16 16:13:15 +1100 05:15
r15729 | bsb++ | started adding URLs to a_dependency_graph.dot
Alias__ audreyt: FYI, I just did a M:I release and added recursive tests 05:19
only dev release for now
svnbot6 r15730 | bsb++ | r310@ogum: bsb | 2007-03-16 16:55:17 +1100 05:57
r15730 | bsb++ | Instructions to make a clickable a_dependency_graph.html
audreyt interesting p5 tidbit: ${...} is a cheaper way to write do{...} 10:12
i.e. when you say
condition(...) or do { action; action; next };
it's equally valid to say
condition(...) or ${ action; action; next } 10:13
audreyt goes back cleaning up pads 10:16
allbery_b: I see, mtl has changed layout. I'll deal. 10:17
audreyt hm. perl5's \&z always works as \&THIS_PACKAGE::z and never padsv &z 10:47
so it already treats sigils nonuniformly 10:48
ah, that was because &z never lives in the pad.
dakkar subs are always package-level...
audreyt well, 5.10 has my $_, which can leads to my subs 10:49
but yeah.
the fun thing is that with pugs refactoring, subs are never looked up in package level unless you use &qualified::subs notation
which hopefully means even faster dispatch. we'll see 10:50
dakkar subs get imported into pads?
at compile time, I suppose...
audreyt yes, import is _always_ lexical
the only difference is whether you also put it into qualified scope, i.e. you can import to OUR 10:51
but it's always lexical.
dakkar ok, I used the wrong verb...
audreyt but that is fantastic as you can say
{ use CGI; ...a lot of symbols ... } ... not leaked ...
and they won't pollute YourPackage::symbol 10:52
since they don't, by default, enter into OUR scope.
dakkar sub stuff {...}; sub another {stuff()}; <-- «stuff» is looked for in «another»'s pad, then in the container pad?
audreyt correct.
and the lookup is calculated at compile time
at runtime we can use O(1) representation. 10:53
(as a consequence, "sub stuff{...}; sub another {stuff(); sub stuff{...}}" is error)
not merely warning but error
dakkar so how do we do that delightfully horrible thing of overriding a sub in someone else's packages?
audreyt if you switch the inner "sub stuff" with "stuff()" call, then it's merely warning
you do it explicitly...
&another_package::symbol := ... 10:54
dakkar and it writes into all the pads?
audreyt no
when you declare
our &foo
dakkar is getting confused
audreyt which is the default for "sub foo"
it creates a padentry
dakkar oh! we can't override a 'my' sub
audreyt it's stored in both pkg and lex pad 10:55
if you rebind it in one it's of course reflected in another
correct. you cannot modify somebody else's lexical pad.
not without that somebody's consent.
that somebody can declare consent saying
my sub foo (...) is context<rw> { ... } 10:56
but you can't introduce new symbols
nor can you rebind symbols in other pads, without some other route to reach its entry
e.g. by OUR:: or by CALLER:: navigation
which are both off-limits by default when you say "my sub foo". 10:57
which is good :)
dakkar yes, definitely
all this means that the powerful-but-don't-do-it things that we can do in Perl 5 will still be possible, but they won't happen by accident ;-)
audreyt yes :) 10:59
the imported-subs-conflicting-with-your-subs 11:00
and import-subs-masquerading-as-your-methods
and overrideen-methods-accidentally-destroys-imported-subs
problems are naturally gone
:)
[smash] audreyt: ping 12:37
audreyt [smash]: pong! 12:42
audreyt [smash]: have some cycles? 12:45
[smash] audreyt: cycles !? 12:47
audreyt i.e. time to work on Capture/Signature compatibleness checke 12:48
checker
[smash] not yet.. didn't you read the private messages i sent you ?
audreyt oh ok 12:51
got it now
fglock the 6on6 compiler seems to need first-class namespaces and pad-stacks; I wonder if is this due to a wrong choice of internal architecture 12:55
mj41 Google Summer of Code 2007 is on! code.google.com/soc 12:57
lambdabot Title: Google Code - Summer of Code - Google Summer of Code
fglock I mean, because first-class namespaces and pad-stacks are not in the p6 spec 12:58
integral mj41: No TPF this year though 12:59
fglock re Google - no perl?
audreyt no perl no pugs :) 13:03
<- signed up to be mentor at haskell.org side for people who'd work on pugs
integral there's some pugs ideas on the haskell.org ideas page fwiw
fglock re 6on6 - closures are currently created with eval(), which calls the compiler recursively 13:24
lichtkind morjn guys and audrey :) 13:30
audreyt had any notes? 13:31
fglock lichtkind: hi
lichtkind hi fglock 13:32
audreyt lichtkind: try adding in the spec.pugscode.org/ link somewhere :) 13:33
lambdabot Title: Official Perl 6 Documentation
lichtkind audreyt thats a nobrainer since there is an linklist below it wasnt because some unready textblocks were between text and linklist and i dont wanted to send you this 13:34
audreyt ok
Ann's name is Ann not Anne
fglock 6on6 actually calls eval( "...", :lang<vm-language>)
audreyt fglock: the compiler would need first class ns/pad objects 13:35
doesn't mean we need to expose it to userland
that sounds like not a problem
lichtkind Audreyt yes and btw spec.pugscode.org/ was already in :)
lambdabot Title: Official Perl 6 Documentation
fglock audreyt: ah, ok
audreyt lichtkind: I mean s/Anne/Ann/ in your article
oh cool.
lichtkind audrey yes been there done that :) 13:36
fglock audreyt: would you have a use for 6-on-java? 13:37
lichtkind audreyt i currently rewritten some parts and want to finish it today , you want it again then?
audreyt sure
lichtkind: it may be good to mention that pugs is no longer only a haskell-based interpreter
lichtkind it is ? *surprised* 13:38
its now pugs based ?
audreyt it's hybrid: the perl5 vm and GHC vm is linked together
so perl5 modules and haskell modules can both run natively
and we can link in other interpreters if needed (SpiderMonkey for JavaScript modules, JVM for Java modules) 13:39
so we glue VMs together (in the same process) to achieve multi language interoperability
moritz_ ... to produce pugs executables with arbitrary sizes ;)
audreyt rather than glue languages together into one VM
audreyt why yes but at least you asked for it :) 13:40
lichtkind thats great feature but what type are the vm files?
audreyt .so, .dll
dynamic libraries, usually
lichtkind ah
audreyt perl58.dll on windows, libperl.so on unix
so pugs is a compiler that targets multiple, maybe hybrid VMs
lichtkind yes thats ge mentioned in the second unwritten part of the pugs chapter 13:41
audreyt parrot is a vm that can be targetted by multiple, maybe hybrid compilers
each aims to achieve language interop mandated by p6, but in opposite directions
fglock: not atm :)
lichtkind so we get 2 platforms or there any pülans to let grow togehter? 13:42
audreyt pugs has a parrot backend but it only supports a very limited range of perl6
same applies to kp6/mp6
fglock audreyt: ok - just checking; there seems to be some interest on $job here
audreyt fglock: I've found Inline::Java to be too good :) 13:43
fglock: once you use its AUTOSTUDY and JNI together, the overhead and hassle becomes negligible
so probably won't look for another solution beyond that
fglock oh, nice - I should try it then 13:44
audreyt atm
it's surprisingly sane :)
(it's also been under development for 6 years.)
(more, if you count the ancestors such as JPL)
fglock this makes possible eval(..., lang:<Java>) 13:45
audreyt fglock: pugs now does that :)
fglock cool
audreyt it's just
use Inline ( Java => 'STUDY', AUTOSTUDY => 1, JNI => 1, STUDY => ['java.module.here'] ); 13:46
java::module::here->new(...)
moritz_ wow, that's way cool
audreyt is all.
moritz_ are there any examples/test cases in the repository?
audreyt (that's if you are in the main package. otherwise, YourPackage::java::module::here->new)
moritz_: no, the code was just in yesterday. want to write some?
moritz_: the current syntax to write tests for is 13:47
use java:java.lang.System;
say "Greetings from Java ", System.getProperty('java.version');
which should get
Greetings from Java 1.5.0_07
by all means write tests and/or examples :)
moritz_ audreyt: I'd like to, but my girlfriend is here for ~ 1 week, and I have nearly no time for such things ;)
audreyt oh. that's okay, code can wait :)
moritz_ audreyt: but when she's gone again I might write some
lichtkind audreyt: ehat please is kp6/mp6?
audreyt lichtkind: svn.pugscode.org/pugs/v6/v6-KindaPe...cs/FAQ.pod 13:48
lichtkind k 13:53
fglock++ 13:57
bye 13:59
moritz_ lichtkind: bye
rindolf Somebody talk to me! 14:18
;-)
moritz_ rindolf: foo()
rindolf moritz_: return undef;
moritz_ die "Can't use undef as numerical value in RINDOLF line 3" 14:19
svnbot6 r15731 | fglock++ | kp6 - FAQ: </vacations>
moritz_ ;-)
rindolf moritz_: OK, I'm not fluent in PerlSpeak. Shall we talk in English, instead? 14:22
moritz_: what's up?
moritz_ rindolf: not much..
I'm currently generating 12x12-Sudokus for my website :-)
rindolf moritz_: with "digits" up to 12? 14:23
moritz_ rindolf: with letters instead of digits
rindolf moritz_: but how are they divided?
moritz_ rindolf: in 4x3-blocks
rindolf moritz_: ah. 14:24
moritz_ rindolf: empty grids of various sizes are already available
rindolf: sudokugarden.de/en/info/flavours
lambdabot Title: About Sudoku - Different Flavours :: Sudoku Garden
moritz_ rindolf: no I'm working on a script that generates arbitrary sized sudokus into pdf files 14:25
rindolf moritz_: ISTR that Sudoko is NP-complete.
At least I read it in the wikipedia.
moritz_ rindolf: it isn't if you can assume that a single solution exists
rindolf: or at least that's what I read some time ago 14:26
rindolf Although typical Sudoku puzzles (with 9×9 grid and 3×3 regions) can be solved quickly by computer, the generalization to larger grids is known to be NP-Complete. Various optimisation methods have been proposed for large grids. 14:28
en.wikipedia.org/wiki/Sudoku
lambdabot Title: Sudoku - Wikipedia, the free encyclopedia
moritz_ rindolf: ok, the _generation_ is NP-complete 14:29
rindolf moritz_: no generali*sa*tion
moritz_ ?eval say 17/81 14:38
evalbot_r15731 OUTPUT[0.2098765432098765432098765432098765432099␤] Bool::True
Limbic_Region ooh neat 14:47
?evaly my
er
evalbot_r15731 Error: ␤Unexpected end of input␤expecting "=", formal parameter, context, ":" or "("
Limbic_Region ?eval my $pi = atan(2,2) * 4; $pi
evalbot_r15731 \3.141592653589793
Limbic_Region ? 14:48
?eval say atan(2,2) * 4.0
evalbot_r15731 OUTPUT[3.141592653589793␤] Bool::True
Limbic_Region feels cheated
moritz_ Limbic_Region: why? 14:49
Limbic_Region why should 17/18 get more digits than pi - everyone loves pi, I eat it when ever I can 14:51
but with so little pi to go around, I am not getting my fair share
moritz_ $ pi|wc -c 14:52
108
Limbic_Region oh, you mean it is just evalbot 14:53
hrm
or what? 14:54
moritz_ Limbic_Region: no, I have a program "pi" installed, that gives 106 digits per default
Limbic_Region: so if you love pi, use this program ;) 14:55
time pi 10000|wc -c
10010
real 0m0.095s
Limbic_Region oh, well - calculating pi isn't terribly difficult
I'm just confused as to why pugs would be giving different levels of precision 14:56
moritz_ Limbic_Region: well, atan2 is certainly more difficult than a fraction
svnbot6 r15732 | audreyt++ | * Pugs.AST.Internals: Remove the envClasses field as it does 15:34
r15732 | audreyt++ | not add to anything; we'll switch to MO-based type narrowness
r15732 | audreyt++ | dispatch anyway, so the notion of "global type-distance" can
r15732 | audreyt++ | be obsoleted now.
r15733 | audreyt++ | * regen instances.
Limbic_Region audreyt - don't know what the rest of the field looks like, but Win32 was about 412 failing tests last night (but nearly 90 min smoke time) 15:36
svnbot6 r15734 | audreyt++ | * Pugs.Run.Args: Repair -p and -n.
rindolf Hi audreyt 15:38
svnbot6 r15735 | audreyt++ | * Pugs.AST.Pos: Switch to a ByteString representation of
r15735 | audreyt++ | the source file name to avoid unneccessary heap use.
pasteling "evalbot_r15731" at 194.145.200.126 pasted "Pugs build failure" (406 lines, 21.8K) at sial.org/pbot/23530 15:40
svnbot6 r15736 | audreyt++ | * Pugs.Types: Memoize all type-distance calls since we no
r15736 | audreyt++ | longer have a Env-global class tree.
TimToady Limbic_Region: the fraction of pi is probably using rationals, and just cuts off the division arbitrarily at 40 digits, since it's a repeating fraction. atan2 is using floaters 15:41
svnbot6 r15737 | audreyt++ | * Pugs.*: Chase the envClasses change. 15:43
r15737 | audreyt++ | This concludes the envClasses batch.
Limbic_Region thanks TimToady. Is there anything in the spec that speaks to precision? 15:45
integral hmm, continued fractions
TimToady S02:549 talks about how division can be forced to return a rational by pragma, but doesn't specify the pragma 15:49
Num is assumed to be floating point. 15:50
which has built-in precision limitations
Limbic_Region ok 15:51
TimToady how precision is specified to the Rational class is underspecified
current Rat seems to like the number 40 for some reason
rindolf Hi TimToady
TimToady: what's up?
TimToady foo()
rindolf TimToady: you've started speaking in Perl too? 15:52
TimToady I dreamed in Perl once...
rindolf TimToady: so did my friend. 15:52
TimToady sorry to hear your friend is going insane
svnbot6 r15738 | audreyt++ | * util/yaml_harness.pl: Speed up concurrent smoking by
r15738 | audreyt++ | assigning interleaving tasks (instead of consecutive ones)
r15738 | audreyt++ | across childrens.
rindolf TimToady: he dreamed in Perl once. 15:53
I never dreamed in Perl, but I sometimes dreamed in gvim. 15:54
Thinking I was sitting next to gvim being able to type stuff on the keyboard.
Wow! I commited over 20 revisions today to the Test::Run repository.
TimToady hmm, shouldn't the noun for a rational number just be Ratio? 16:07
then we can snigger when the newbies pronounce it "rat I/O" 16:08
audreyt that's the name Haskell uses... 16:09
whew, smoke time back to <20min
TimToady Ratio?
or Rat?
audreyt yeah
Ratio
arbitrary-precision rationals (Ratio Integer) is type-synonymed to "Rational" 16:10
but you can have (Ratio Int) too if you want
which is fixed-precision
TimToady I'd be more inclined to alias Ratio to Rat
audreyt ok... 16:11
Perl 6, the first language to have built in Rat and Cat types.
TimToady where does H get the 40 digit precision from?
just hardwired somewhere? (somehow I doubt it)
audreyt Double I think 16:12
same as C
TimToady why would Ratio have anything to do with Double? 16:13
audreyt oh, nvm.
it's hardcoded alright
src/Pugs/AST/Utils.hs:135 16:14
and :137
forgot how/why we did it :)
TimToady well, we can probably allow $x / $y :rat(42) as a micropragma 16:15
audreyt mmm more multi variants to write
TimToady but that might slow it down
audreyt now we actually _have_ to allow named multi resolution...
TimToady do we require it for anything else yet?
ignoring my usage in STD.pm 16:16
we can probably just "use Rat :prec(582)" and avoid committing adverbial pragmacide. 16:17
audreyt well :p 16:26
I mean hash :p 16:27
not the smiley
Limbic_Region hrm, concurrent doesn't appear to be doing the right thing on Win32
TimToady pity we don't have a 3-letter word for Complex 16:32
TimToady course, whoever called them "complex numbers" in the first place had no clue about the notion of the problems of using too general of a word... 16:34
Limbic_Region use a synonym that can be crammed into 3 chars?
specbot6 r14352 | larry++ | The ratio type is now a noun, not an adjective. 16:35
r14352 | larry++ | No longer spec that / must pragmatically produce a Rat.
r14352 | larry++ | Instead, Int div Int is now guaranteed to return a Rat.
r14352 | larry++ | An integerized Rat is a truncated Rat.
Limbic_Region IIRC, I first hacked in concurrency support on Win32 using Parallel::ForkManager
TimToady Rpi "real plus imaginary" not pronouncable. There are cultural problems with "Real and imaginary", though. 16:36
Limbic_Region wonders why it doesn't work quite right now
TimToady Real Imaginary Number, a Rin. Now just need a Tin type to really go to the dogs... 16:38
jrockway maybe we should add a unicode symbol for these types :) 16:39
diakopter how about % 16:43
oh wait, we're done with ratio, on to complex; sorry
TimToady Both Imaginary and Real Data 16:45
rindolf TimToady: Rin, heh
TimToady: can we also have a Dolf?
TimToady Dolphin, maybe 16:46
rindolf Rindolphin?
TimToady maybe complex should be Dodo, for "double double". :) 16:50
Limbic_Region hrm - I suspect Parallel::ForkManager
TimToady but that's not 3 letters... 16:52
lessee. if the type actually tries to optimize non complex into a num, then we could call it a "Real Or Complex", that is, a Roc.
TimToady Complex has too many consonants to abbreviate to three letters, alas. 16:54
C5x would work. :) 16:55
TimToady biab & 16:55
jrockway Cmplx :) 17:01
TimToady Nim 17:08
Nums and Nims 17:09
geoffb imagines his declarations reading like Jabberwocky or Dr. Seuss
TimToady I wonder how you play Nim with imaginary numbers? 17:10
Limbic_Region I kind of liked c5x
TimToady looks like some kind of cable
geoffb TimToady: two piles, taking from either one or the other each turn?
Limbic_Region of course - it would only work for those who could s/5/omple/
geoffb TimToady: That would be c5e, of course. 17:11
Wow, c5e is 40 words in /usr/share/dict/words
TimToady I think having an unrelated digit in the middle of nimeric computations would be a mistake. 17:12
geoffb Interestingly
Oh, doh, stupid me
TimToady there's that 40 again.
geoffb It's 110 words. c5x, however, is only two: complex and chamoix 17:13
Limbic_Region geoffb - we could generalize it to c2x 17:14
TimToady 18 words here
geoffb I'm guessing the chamoix type does proactive garbage collection
Limbic_Region where instead of 2 being two chars
it is to
so starts with c and ends with x
TimToady cakebox carapax carefox casebox cashbox centrex chamoix chevaux choenix ciseaux coalbox coannex complex conflux corebox coteaux cow-pox curtlax
geoffb TimToady: clearly your dict is bigger than mine
TimToady er 17:15
geoffb Oh see, now there are several non-english words there
Unless carapax is actually english, hmmm
Yeah, looks to be German 17:16
TimToady I have some German blood in me--that might explain it.
geoffb Jokes aside, where did your dictionary come from? Is it system default, or something you installed separately? 17:18
TimToady fedora 6
geoffb hmm
geoffb raises his eyebrows: Apparently Google is using wikipedia as a dictionary now too ... it's giving it as responses to define: queries 17:20
TimToady @tell audreyt my smoke of 15738 only went from 64 min to 60 min. but at least that's the right direction. :) 17:23
lambdabot Consider it noted.
Limbic_Region TimToady - are you able to take advantage of concurrency in your smoke? 17:31
TimToady nope
Limbic_Region it took nearly 90 minutes last night - I set concurrency to 2 (as I have 2 procs) and I am testing again as of the most recent rev 17:33
TimToady Language question for the day: How come people talk about "serial killers" but not "parallel killers"?
Limbic_Region provided it doesn't "get stuck", I should have the results soonish
TimToady Hmm, if ratios were a generic type we could have the Rats of Nim(h) 17:36
Limbic_Region TimToady - and with built in cryptography, we could have the secrets too 17:38
TimToady shower power & 17:40
Limbic_Region @tell audreyt smoke concurrency intermittently breaks on Win32. I will look into it if I get some round tuits 17:57
lambdabot Consider it noted.
pmurias hi 18:46
offby1 ho
pmurias why dosn't TPF take part in the google summer of code? 18:49
TimToady nobody volunteered to mentor, basically. 18:49
pmurias is too young to apply this year, but hopes that things will be different next summer 18:52
spinclad TimToady: Cpx came to mind and stuck there 18:56
now i want to type some gaussian ints and rats: Z[i] and Q[i] ? 18:58
Complex of Int, Complex of Rat, Rat of Complex Int ... 18:59
spinclad TimToady: in r14352:S02:540: s/C<infix:/C<< infix:/ 19:00
TimToady: in r14350:S02:1521: I don't see what `, and so on transitively´ is about. 19:05
and in line 1525, s/^/visible / i think. 19:08
pmurias fglock: hi 19:28
specbot6 r14353 | larry++ | Fixes suggested by spinclad++. 19:33
fglock pmurias: hey 19:38
TimToady ?eval <good fast cheap>.pick(rand 3) 19:47
evalbot_r15738 ()
TimToady ah, the government solution
?eval <good fast cheap>.pick(rand 3)
evalbot_r15738 ("fast",)
TimToady extreme programming
?eval <good fast cheap>.pick(rand 3)
evalbot_r15738 ("cheap", "good")
TimToady there's Perl 6
?eval <good fast cheap>.pick(rand 3) 19:48
evalbot_r15738 ("good", "fast")
TimToady that'd be a Cray of some sort, I suppose 19:49
I love this language. :D 19:50
kolibrie :)
fglock :)
Limbic_Region yes, certainly not cheap 19:52
TimToady or maybe a PS3 19:53
rindolf Hi all! 20:00
A bit OT, but on-topic because it's TAP-related.
Is there a Ruby way to attach meta-data to certain methods, similar to Perl 5's attributes?
kolibrie I am creating an object ::Resource(url => '' ~ $domain ~ '/path'), and class Resource { has $.domain, $.url }, but domain is not assigned until after url is, so my url is incorrect. 20:41
how can I make this lazy
or do it a different way? 20:42
kolibrie ::Resource(url => lazy('' ~ self.domain ~ '/path')) 20:44
kolibrie does not think the self will point to the object he means 20:45
dmpk2k Are there any perl6 list summaries more recent than Feb 12, 2006? 21:30
tene dmpk2k: pugs.blogs.com/
lambdabot Title: Pugs
diakopter @tell audreyt you've got mail. the good kind.
lambdabot Consider it noted.
tene I don't know why they're not posted to dev.perl.org or whatever anymore 21:31
dmpk2k tene: oh, thanks. :)
dmpk2k was worried that had died off somehow.
tene Rather confusing to me, too, before I subscribed o the mailing lists
Thought that Perl 6 had died off.
dmpk2k Maybe someone should update dev.perl.org. For a long time I thought that perl6 had finally kicked the bucket. 21:32
It's not a hard conclusion to come to if you're on the outside.
tene dmpk2k: are you volunteering to maintain dev.perl.org if given access? 21:36
;)
dmpk2k Disaster! 21:37
lichtkind please what was the difference betweent PAST and POST again? 21:42
dmpk2k Auto-accessors are analogous to properties? 21:54
TimToady Properties are just methods. 22:01
Adding properties to an object is just mixins.
TimToady However, mixins are not just mixins. :) 22:02
pmurias lichtkind: the POST is the low-level AST representing parrot bytecode, PAST represents high-level language constructs
TimToady A for abstract, O for operators 22:03
lichtkind pmurias: so PAST ~~ PIR ; POST ~~ PASM ? 22:04
pmurias PAST ~~ perl6|python|ruby ; POST ~~ PIR 22:05
lichtkind thanks a lot 22:05
for what they need PASM now?
pmurias i'm not really a parrot guy, but PASM has mostly been replaced by PIR 22:07
i'm not sure if i'm looking in the right implementation , but POST is emitted to PIR at compilers/past-pm
the transformation process is i think source --> PAST --> POST --> PIR --> bytecode 22:09
makes any sense?
sleep& 22:14
lichtkind good night 22:15
thanks but some things i still saw in PASM but it looked like some constant definitions ans older benchmarkes 22:16
pmurias: yes your transform chart is correct 22:17
bye 22:31
meppl good night 22:55
rindolf TimToady: here? 23:36
avar rindolf: use the async powers of irc and @tell :) 23:45
rindolf www.frappr.com/freenodeperl/ 23:51