-Ofun: xrl.us/hxhk | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6
Set by apple-gunkies on 11 November 2005.
Smari Regarding infix operators; is it not possible to define an infix for two arrays? 00:27
clkao sure it's possible 00:33
i just did it
Smari Care to show me? Mine's not working as it should. 00:34
Probably just me being stupid, but hey.. :)
clkao our &infix:<»+«> := -> Array @a, Array @b { __hyper(\&infix:<+>, @a, @b) };
sub __hyper (Code $op, Array @a, Array @b) {
... }
Smari Hmm. 00:35
obra it's sad when one can't actually paste perl code to IRC.
Smari multi sub infix:<·>(Array @a, Array @b) returns Int is commutative { [+] (@a »*« @b); }
.. why won't this work?
clkao it works for me. if i name the operator XO 00:37
Smari Aye. Works here too when I call it that. 00:38
Incorrectly defined function though. Haha. 00:40
clkao Smari: works for me if i name it ♨
Smari Why doesn't · work?
Bug? 00:41
clkao see if hotspring works
Smari hotspring?
clkao the symbol i just pasted
Smari Yup, it works.
obra cries. I want unicode irc 00:42
clkao obra: it's almost year 2006, get your software working! 00:43
obra gnome-terminal, screen and irssi. These things should all work
xinming want the the console utf-8 in utf-8 encoding by default. :-) 00:45
oopos.
I just wish console will support utf-8 by default without starting X. :-P
Smari Hah! 00:46
Now it works.
It was failing because for some reason I was using the ISO-8859-1 · instead of the UTF-8 one.
paran01d hi im having trouble getting my ext tests to run they seem to be looking for pugs in a scripts dir within blib that isnt there any ideas ? 01:56
taking my first steps into pugs after a tlk from audrey last week at sydney perlmongers 01:58
dduncan paran01d, are you talking about the pre-existing ext tests, or new ones you are making? 02:25
gaal paran01d: how are you running your tests? does export HARNESS_PERL=`pwd` help? 05:20
audreyt gaal: hey 05:35
audreyt wakes up and goes out to collect laptop
anything I should look at when on the road wrt your lexical patch? 05:36
gaal audreyt: just that it's not lexical :-/ 05:47
I ended up punting and doing SGlobal in the caller's context
it at least solves the double-use problem for now
audreyt OKIE 05:48
er, sorry, capslock
okie.
do you have your half formed lexical patch somewhere commented out in th esrc?
gaal I haven't read that article about mixins and stuff yet -- busy weekend unfortunately over --
audreyt: well, look in Parser.hs 849 05:49
compare with 467
audreyt k
gaal ...re: that article, I do want to read it hoping it clarifies what some exports might be. Eg. I still don't undersatnd when you might want to import something into an env or state var. 05:50
audreyt *nod*
gaal Reason I punted on lexicals was that I just wasn't sure how to combine Pad transofmers. You'll see that SGlobal exports are much simpler (void context map in Haskell! haha) 05:52
audreyt ok:) 05:53
yup, I see it. will hack it some more on the taxi
s/on/in/ # or whatever the correct term is
gaal hopes the duct tape holds out :-) 05:54
(I mean on the temp laptop)
$Larry also mentioned something looking a bit like new syntax for the importation spec, too 05:55
let me find it...
oh, dang, I just remembered thinging it needed more discussion but I was too excited by the rest of the email to follow up: 05:56
my $hownow is export(:state); # override implicit :my default
but if that works, how do you set tagsets? 05:57
what it *does* imply though is that the export hash needs another layer of metadata, something like
%INC<MyModule><exports> = { '&f' => { bind => sub{...}, mode => 'lexical' }, '%s' => { bind = ??, mode => 'state' } } 05:59
but that's easy to add.
(s/mode/Scope/ too)
audreyt yup, got it. 06:02
I wonder what doe this mean 06:03
sub foo { my $bar is export; ... }
gaal eep :) 06:04
I hope it means a compiler error!
s/compiler/compile-time/
audreyt how is it diff from 06:05
sub foo { sub bar is export { ... } }
or even
{ my $foo; sub bar is export { ... } }
the last one is obviously correct
gaal hhmm. 06:06
but you can close on lexicals, not on subs.
audreyt but the block is conceptually a closure as well 06:08
because you can reenter it with &?BLOCK(...)
from within itself
gaal but you can't export to yourself 06:09
audreyt indeed that's not sane 06:10
but consider: 06:11
Foo.pm: my $foo is export;
vs
Foo.pm: { my $foo is export; }
that's what I was thinking about
namely how far the exports hoist themselves from inner blocks
gaal conceptually it already is in an inner block even in the first place, no? 06:12
audreyt it is, but it's the _first_ inner block
gaal package Foo; my $x === package Foo { my $x }
yes. hmm.
audreyt whilst in the second one, the "export" need to be hoisted to the package closure
because the package closure will not know about $foo
and hence can't provide it if we are to request it from the closure 06:13
gaal on a possibly related note, what does this mean? 06:14
my $foo is export(:GLOBAL)
or
our $foo is export(:MY)
etc.
audreyt that's easy 06:16
it affects symbol creation on the importation site
gaal you mean, there's just yet another binding of some arbitrary scope 06:17
but what about closures then?
package Foo; sub mk_counter { my $color = @_ is export; return { "<font color='$color'>" ~ @_ ~ "</font>" } } 06:18
apart from hoisitng 06:19
this especially doesn't make sense because $color is lexical and closed on.
well, drats. my($color) = @_, and I can't believe I fell for that once again. 06:20
audreyt my $color = @_; # this has a very well defined meaning of \@_ :) 06:21
but anyways.
I'll think about it more; for now it sufficies to rule out all but toplevel symbols as export candidates 06:22
we can always p6l if ambiguity arises
yay, the taxi is here. will be back in ~2hr 06:23
gaal does the parser readily know if a symbol is top level?
ok i'm going to $work. good luck
audreyt gaal: the idea is we collect it perclosure
and if it's not a package closure, something is very wrong.
this belongs to the compiler phase not parser 06:24
but the parser at importation site should just query the export list from the module -- we'll see how it turns out :)
enjoy $work :)
audreyt waves &
gaal waves back, er, away &
asavige paran01d: re running tests, I set env vbles HARNESS_PERL=./pugs PERL6LIB=ext/Test/lib 06:29
Then run test with prove -v t/fred.t I'm a bit rusty, is that still recommended way to do it?
gaal PERL6LIB=blib6/lib possibly even better
and HARNESS_PERL=`pwd` so you can cd around. 06:30
dduncan new commit pending ... 07:37
push of new file SQL::Routine::Language done 07:46
svnbot6 r8289 | Darren_Duncan++ | r1389@Darren-Duncans-Computer: darrenduncan | 2005-12-17 23:45:43 -0800
r8289 | Darren_Duncan++ | /ext/Rosetta-Incubator : added start of new file lib/SQL/Routine/Language.pod, the authoritative Rosetta VM language specs
dduncan oh, that's up too 07:47
avar 'Migration is important. The perl interpreter will assume that it is being fed Perl 5 code unless the code starts with a "class" or "module" keyword, or you specifically tell it you're running Perl 6 code in some other way, such as by:' 12:59
-syn01
so code like:
use strict;
class Foo { ... won't be considered perl6 code? 13:00
(doesn't start with class)
xinming avar: No, It will, It means, if the current file which has some signature that perl 6 has, It will be automatically turn running into perl 6 13:08
Just like, If you have a package which contain a class, then the file is perl 6 source, 13:09
hmm, for example, package Pkg1; bla...bla..bla; class Cls1; bla..bla..bla; 1; This will be considered perl 6, as there is a class. 13:10
certainly for some other signature... Just like `for @a -> $v { ... }` 13:11
hmm, maybe, the if without the parenthesis in test statement, will also be considered perl 6 13:12
avar sure there's lots of things, syn01 is just a bit unclear on it... 13:13
Another big thing to check for is Str/Num/Int/Array..
xinming It's understandable. Str is an object, and str is a bit like char *s = "bla...bla...bla..." 13:14
and str can be promoted automatically
hope what I say is understandable. :-P 13:15
audreyt stevan_: ping 13:36
stevan_: I see this 13:40
$nothead++ && last if exists $in_tail{$cand};
in your C3
shouldn't it be ++$nothead ?
otherwise the && won't take effect
or am I missing some very clever use of postincrement?
avar isn't that equivalent to ($nothead++) && last if exists $in_tail{$cand}; since ++ is above && in the precedence table? 13:43
audreyt yeah 13:44
but that will make the "last" not run on the first time of $nothead++
which sounds wrong to me
(this is not seen in the lisp or python implementation of C3)
avar is that perl5/Perl6-MetaModel/lib6/metamorph.pl ? 13:47
stevan_ audreyt: no cleverness involved 13:48
bad coding instead :)
audreyt avar: no... Class::C3 on CPAN and perl5/Perl6-ObjectSpace/lib/Perl6/MM/C3.pm
stevan_ however, it passes all the (Python, Dylan) tests
audreyt stevan_: sure, because it will loop to the end anyway 13:49
it just consumes some more CPU
stevan_ ah
13:49 stevan_ is now known as stevan
stevan I will adjust them then :) 13:49
audreyt: I looked for a Haskell impl of C3 last night, all I found was Scheme, so I went back to reading Algorithms :) 13:51
svnbot6 r8290 | iblech++ | t/builtins/lists/minmax.t: Added tests for min and max with literals (e.g.
r8290 | iblech++ | max(1,2,3)); bug in PIL2JS spotted by clkao++
stevan but its time for breakfast right now
stevan &
audreyt stevan: anyway, I coded it up in 3 mins :) 13:55
not sure about correctness tho.
testing
avar $ pugs -e 'say 5^5' 13:56
one() <- mm?
strange output;)
audreyt ^ is infix one()
avar ah, ;)
avar still learning all the operators 13:57
gaal yo a.* 13:58
audreyt stevan: it looks like this
doMerge res seqs
| seqs'@(_:_) <- filter (not . null) seqs
, (cand:_) <- [ s | (s:_) <- seqs', all (not . elem s . tail) seqs']
= doMerge (cand:res) [ if s == cand then rest else full | full@(s:rest) <- seqs' ]
doMerge res _ = res
gaal: hey... I'll have to wait until tomorrow afternoon to get my laptop back... they overflowed on weekend 13:59
gaal no worries :) i hope their overflow isn't liquid though 14:00
audreyt prolly not :)
gaal I should work on the other side of this anyway, too: parsing the import requests. 14:04
svnbot6 r8291 | autrijus++ | * C3 MRO lands to Minilang... would love some more tests.
gaal and exporting vars.
audreyt cool! 14:05
stevan: I've committed a bunch of work on C3 MRO and objs
svnbot6 r8292 | autrijus++ | * replace objects in minilang from simple hashes into
r8292 | autrijus++ | opaque objects with fetch/store/freeze/thaw virtual
r8292 | autrijus++ | methods. this should make autoboxing infinitely easier
r8292 | autrijus++ | as an autoboxed Int can use a haskell Integer without
r8292 | autrijus++ | exposing it to the externals.
audreyt sanity checks welcome
also for the [].apply(->{}) form, can you use (->{}).do_for([]) instead? 14:06
masak how does one find out if a reference holds a scalar in perl 6? ref() seems to be too specific, returning Int, Str and Bool 14:08
audreyt $var.does(Scalar) 14:11
bbl... *sleep*
audreyt prays tomorrow will yield a working laptop 14:12
&
masak audreyt: thx
ingy audreyt: yo 14:14
ingy alwayz missez audreyt by minutez 14:15
Juerd audreyt: Consider IBM laptops :) They don't break quite as easily 14:23
avar multi sub infix:<≠> (Int $x, Int $y) returns Bool { $x != $y } 14:24
say "five & ten are not equal" if 5 ≠ 10;
works with <z> and if 5 z 10;
xinming bought an Asus Laptop because of audreyt who also has one... Hope mine won't break so easy. :-/
avar limitation in pugs or perl6?
xinming avar: what do you mean? 14:25
Juerd xinming: Hm. Asus is one of the regular brands, on the cheaper side of things. They're great for normal users, but not for almost continuous use... 14:26
It's a shame laptop manufacturers don't specify the duty cycle their products were designed for.
clkao morning
Juerd Afternoon 14:27
clkao *yawn*
xinming Juerd: I ever wish to choose IBM series. But the problem is, It's too expensive... for my A3E (1.73G with 1G ram) It took about $1200, But for IBM, the price is oever $2000 Maybe. :-/
clkao Juerd: i like my asus quite alright, it's been solid and light :) 14:28
Juerd xinming: I was parsing "for my ... (...)" as Perl 5 and expected a { ... Wow.
xinming: Anyway, the trick is to be less picky about specs.
clkao xinming: very few laptops can pass the 'audrey test'
Juerd If you work remotely, the speed of your laptop is much less relevant.
And it's cheaper often to buy a very simple speed machine that you connect to a permanent internet connection, than to buy an equally much faster laptop. 14:29
While the simple speed machine has additional benefits.
I guess most depends on the availability of fast internet connections. 14:30
xinming Juerd: Yeap, I agree, But for a person who really loves computer, how will he think of the speed of computer. escpecially the faster the better. :-), at least, to me, I really hope my computer as good as possible.
Juerd xinming: I disagree about that. I absolutely love computers, and adore my laptop, but speed is not that important. How computers are built, and how they interact with me (comfort level) are much more important. 14:31
"good" should be established through a lot more than the hardware specs.
A good casing, comfortable keyboard (hey, you type HOW MANY keys a day?), good touchpad or trackpoint, battery life, screen brightness (and the ability to dim it enough), etcetera 14:32
xinming I ever got a laptop which contains C3 Via CPU, and 256M, Ram, If you make me compile pugs on that machine, I would rather choose a much cheaper computer instead of laptop.
jethr0 hi, i am trying to "darcs get" pugs from the repository, but it won't work :(
integral you'd only want to compile pugs locally if you didn't have inet...
Juerd If I can't hold the laptop by its screen, because it has weak hinges or a weak cover, I won't want the laptop. 14:33
If its keyboard is uncomfortable, I don't want it. If the battery only lasts an hour, no way.
xinming Juerd: thanks for sharing the experience. :-)
Juerd integral: Yes, but there's a balance to be found.
I think a centrino 1.5 GHz is fast enough to compile occassionally. 14:34
That's much cheaper than an 1.7
Of course, I'm not saying you should go for a C3 :) That's not nice for programmers.
jethr0 the ibm x40 is small, light and has up to 11h battery life with extension accu.
Juerd jethr0: Nonsense. It's 6 to 7 hours.
xinming hmm, by the way, for a machine which is running at home as server, to you all, is the paying for Adsl more than electicity?
.... 11h? >_< 14:35
jethr0 there's an additional (huge) accu
which fits exactly under the notebook
Juerd jethr0: Yes, the 8 cell one. I sell those things.
jethr0: And the laptops.
gaal hey Juerd, resident lang moose. Is this a pugsbug?
?eval <a b c d>.kv
Juerd What, that there is no response?
No, that's just because evalbot isn't here. 14:36
Can't assign that bug to pugs :)
gaal ahem. no, that's an evalbot hiccup :-)
jethr0 Juerd: ok, but it's still quite nice even with the 8cell accu attached. rather cute :)
gaal it'll fail on VUndef being ro or some such weirdness.
Juerd jethr0: Certainly.
jethr0: Just don't spread expectations that are too high, because that results in disappointment.
xinming hmm, I wish to setup a home server to myself, It's a bit expensive, which cost about $250 per year for electicity, and $120 for Adsl
jethr0 ok
gaal should it result in <0 a 1 b 2 c 3 d> ?
(i think it should). 14:37
Juerd jethr0: 6 to 7 hours is impressive for a 1.8 kg laptop; there isn't even need for exaggeration :)
jethr0 :)
Juerd gaal: It depends. Is .kv an array method, or a hash method only?
gaal on arrays it does work.
Juerd If it's a hash method ONLY, then the list is probably coerced to an array, and that again to a hash, via an implicit list. Which would be madness, by the way. 14:38
gaal @l = <a b c d>; @l.kv => what I said above.
Juerd What *works* I find irrelevant when answering these questions. What do the specs say? :)
gaal nono, it's even devined as List::kv.
the specs say it gives these indices.
Juerd Lists, when used with methods, are coerced to arrays, so there should be no difference between @foo = <>; @foo.bar and <>.bar 14:39
afaik.
gaal so there's certainly a bug then.
Juerd My back hurts
gaal afayk.
masak I encountered this line in some code I'm porting to perl6: 14:40
my $pat = quotemeta ($opt);
what does it do?
is there a perl6 equivalent?
gaal masak: probably not needed
masak oh, good 14:41
gaal it escapes hings like * and \b and so on
Juerd masak: Instead of /$pat/, just use /$pat/ :)
gaal so that when you interpolate $pat, they don't become metachars
Juerd masak: Instead of /$pat/, just use /$opt/ :)
gaal: It's not interpolation, and *because of that*, there are no metachars.
masak Juerd: oki
gaal there's a way to explicitly ask $opt to be interpolated 14:42
Juerd Oh, you were still discussing Perl 5. Nevermind.
gaal Juerd: yes, I know.
Juerd gaal: s/interpolated/interpreted as a rule/
gaal Juerd: true
gaal returns to $work
Juerd Just as $bar isn't interpolated in $bar()
masak has `foreach` been deprecated in perl6? 14:47
pugs doesn't recognize it
?eval for 1..5 { say $_ }
oh. evalbot is sleeping, it seems
nvm. anyway, `foreach` is not a recognized keyword in pugs, it seems 14:48
intentional?
xinming masak: It's deprecated, since in perl 6, list are lazy by default, So, No need to use foreach 14:49
for is enough
maybe It's deleted from perl 6 14:50
Juerd masak: In Perl 5, for and foreach are aliases. The list iterating thing is just "for" in Perl 6, and the C-style loop is "loop".
avar there's no foreach in p6
Juerd That was more or less implied :)
masak I interpret your answers as there no longer being any use for `foreach` in p6 14:51
fair enough
since they are synonyms anyway, I mean :)
Juerd I don't like this change. I think the alias is nice to have.
Even though I always write "for".
masak Juerd: me too. but you can create a macro if you want
at least as soon as we have macros in pugs :) 14:52
Juerd masak: my &statement:<foreach> ::= &statement:<for>;
Not a macro. An alias! :)
masak Juerd: of course. my bad.
clkao &infix:{'+'} 16:04
doesn't seem to parse
grr, currying broken: 16:12
t/subroutines/currying.t
pugs: malformed named arg: Syn "named" [Ann (Cxt (CxtItem (mkType "Str"))) (Val (VStr "y")),Val (VInt 2)]
avar Have there been any reports on what filesystem functions will be like in perl 6 ? 16:13
I noticed that there isn't a synopsis / apocolypse for that yet
clkao s29. very sparse
gaal avar: look in Prelude.pm for a link to some discussion. 16:20
svnbot6 r8293 | clkao++ | sym should be local variable. 16:33
clkao dogh 16:43
13902 pugs 44.3% 3:53.08 1 13 777 669M+ 6.37M 423M- 796M
svnbot6 r8294 | clkao++ | Failed tests for currying subs with array arguments. 16:54
r8295 | clkao++ | Hyperoperators coming to pil2js backend - 16:57
r8295 | clkao++ | We have >>+<< now, was hoping to do everything with currying,
r8295 | clkao++ | but it doesn't work yet, left the currying implementation in
r8295 | clkao++ | comments.
r8296 | clkao++ | a bit more hyper-operators so we can pass some more tests. 17:06
stevan_ audreyt: ping 17:12
avar clkao: does $x »++ work now?
eh
@x »++
r0nny re 17:23
stevan_ audreyt: all tests are passing now, I am going to try and write some more MRO tests (port them from Class::MRO, etc) 17:46
clkao avar: no. i didn't touch it. i was working on the js backend
svnbot6 r8297 | stevan++ | PIL.Native.Eval.hs
r8297 | stevan++ | - changed mro_merge to merge against obj, not cls (see NOTE)
r8297 | stevan++ | PIL.Native.Bootstrap.pil
r8297 | stevan++ | - uncached ::Class.MRO, I think we need to get
r8297 | stevan++ | attr.instantiate_container() working in order to
r8297 | stevan++ | make this work correctly
r8297 | stevan++ | - all tests in t/pil/bootstrap.pil work now
gaal stevan_: you hacking haskell now? cool! 17:48
TWD is closer by the ci :-) 17:49
clkao TWD? 17:50
gaal total world domination
clkao grins
now if only people can sort out weirdness in currying
it's totally broken with haskell backend. semi broken for js 17:51
t/subroute/currying.t
subroutine
gaal looks 17:52
urrr, didn't audreyt revive inlined source prelude? 17:56
clkao er?
gaal when you "make unopimized", precompiling the prelude typically fails 17:57
so for those builds, we wanted to bring back the old way of doing it, not precompiling the prelude but storing its p6 text source inside pugs, to be evalled every load.
clkao i was thinking we probably need to define different tiers of functions, so it's easily to do bootstrapping 17:58
gaal weird, i do get a prelude now. must see how that's done :-)
aaahhh, I get it. audreyt++ # smart 17:59
stevan_ gaal: if you consider deleting one word hacking Haskell :) 18:02
stevan_ still can't seem to grok Haskell
gaal well, I see what the problem is (an annotation node messing up patmatch), but I'm not sure yet about how to fix this. 18:03
stevan_ my brains just doesn't work that way,.. no matter how hard I try
clkao Oh, bootstrap.pil is beautiful # stevan++
gaal stevan_: we have one of those untranslatable words in hebrew (comes from Arabic really): "alek"
loosely, it means, "yeah right"
stevan_ clkao: thanks,.. audreyt deserves much credit too... it's his mini-lang :) 18:04
gaal I'm sure you can moose it!
just ignore monads for the while :-)
stevan_ gaal: I think if I had a few weeks to dedicate to it, maybe,.. but the problem is the $work distractions bring my mind back to Perl 18:05
gaal tell me about it :( I just hopped from straight c to C# 18:08
stevan_ I like C# actually :)
better than Java ;)
gaal oh, I didn't say I don't like it 18:09
but the context switches are jarring
stevan_ yeah I bet
Odin-LAP C# has a public domain spec.
(Which, admittedly, doesn't mean a thing, but it's still an interesting point.) 18:10
stevan_ Odin-LAP: that is because MS want's it standardized
standardized == Manager-freindly 18:11
gaal Open. But PD?
Odin-LAP gaal: Yes.
gaal: All ECMA publications are public domain.
stevan_ The CLR spec if public domain as well IIRC
gaal walla? (Arabic's the language du jour for interjections)
Odin-LAP stevan_: Yup. Both C# and CLI were submitted to ECMA for standardisation. 18:12
Hardcopies are available for free, too. Which is quite unusual. 18:13
gaal clkao: svk pull 18:15
svnbot6 r8298 | gaal++ | unbreak named args
gaal three tests still fail, but that's another problem.
clkao the last one fails? 18:16
that's the one i want to get it work :/ with arrays
as i want to implement hyperop in jsbackend with curryed function
gaal I think it passes with no message
clkao whoot 18:17
gaal add one? :)
svnbot6 r8299 | clkao++ | add name to currying array arg test. 18:21
gaal shrieks at L<S06/Currying>
"The result of a use statement is a (compile-time) object that also has an .assuming method"
hmmm, that section no longer stipulates .assuming only takes named args, so one of the tests is obsolete. but does that mean currying is now OK with positional args? 18:24
clkao no idea
gaal clkao: __hyper in that test looks wrong. I don't know if it's the same implementaion you're using elsewhere 18:29
it thunks elements from the long list when the short one is exhausted
but (I think) <1> >>-<< <1 2 3> => <0 -2 -3> 18:30
this is the same problem as the reduce identity issue (that got a looooong thread on p6l)
FWIW, main pugs gives <0 -1 -2>, which is weird in this case but just means someone chose 1 as the default identity 18:32
brb 18:34
clkao i thought it's doing that if it's a list. was going to do expanding if any of them is single 18:35
gaal oh, you mean take the last element? let me see 18:40
where's evalbot? :( 18:41
you're right!
./pugs -e 'say <5> >>-<< <1 2 3>' => 432
clkao but: 18:42
423
./pugs -e 'say <5 4> >>-<< <1 2 3>'
gaal looks like it takes the last element 18:43
oh. no
clkao no
gaal what does that mean then?
clkao only when it's with one ele
otherwise it takes the missing one
gaal ahhh ok. 18:45
gaal goes to look at the spec
avar boo!
%h{str} doesn't work anymore;/
gaal the spec doesn't say anything :/
Juerd Note: do not use lists like 1 2 3 or 0 1 2 in testing attempts :) 18:46
gaal avar: that's because pure {} means only a closure now
avar: %h<str> or %h{'str'}
Juerd You want to be sure it's using your value, not an internal counter.
gaal Juerd: point. I often start at 42
avar woo
Juerd Some newbies think @foo returns the last element, because after @foo = 1..4, @foo == 4. 18:47
avar is there something like .inspect in ruby to find out how a datastructure is laid out?
Juerd Don't get bitten by the same thing :)
gaal: I usually don't use regular intervals
gaal: My favourite numbers are 42, 15 and 123.
avar $ ruby -e 'x = {"a" => %w{a b c}}; puts x.inspect' 18:48
{"a"=>["a", "b", "c"]}
gaal which reminds me how just last week I was tuning some code where I wanted a rollover counter but didn't know what a good value would be. I ran some experiments and 42 turned out to be good, so I used that.
Juerd Hehe
So I'm not the only one using $foo % 42 ;) 18:49
gaal Juerd: I'd never use something like that because it'd mean I have to like, do arithmetic to check the result, and I'd probably be wrong cause I can't count :-)
Juerd Though a few months ago, a client asked me why his web based database application thingy paged results per 42 :)
gaal the correct answer is of course "why any other number?"
let's see you make an appointment on an hour that's != (mod 15). 18:50
Juerd Hours are so weird
gaal != 0 that is 18:51
Juerd But I've stopped trying to round appointment times at nice epoch times.
gaal explicitly mixed-radix system with an implicit further radix.
Juerd Which fluctuates in a fixed pattern 18:52
A pattern which again fluctuates per a fixed pattern which fluctuates per a fixed pattern.
gaal epoch times are, of course, just as arbitrary :)
Juerd Sure. 18:53
But at least one can easily calculate with them.
gaal true dat.
Juerd A base-60/base-60/base-24/base-(28|29|30|31)/base-12/base-10 number is hard. 18:54
clkao how to you create an array with same elements? like map { 5 } 1..10 ? 18:55
there gotta be something simpler 18:56
gaal 5 xx 10
clkao cool
araujo Hello. 18:58
xinming x is enough. :-)
clkao i thought x is for making the string?
avar xinming: x makes a string
*repets a string
xinming oops, sorry, misunderstood. :-/
gaal off to see a film. bye! :) & 18:59
avar mm 19:05
I can't do something like for 1..10.mysub(arg) ? 19:06
I was going to write a method like ruby's step() so I don't have to use loop()
for 1..0.step(2) -> $i { say $i }
1..10
Juerd Has feather been unreachable for long? 19:09
(by ssh)
wolverian heh, nice load average on feather :) 19:35
webmind Juerd, quite 19:39
maybe nice is an idea
hmm 19:40
it was niced
and now we're banned from perlhelp
:)
wolverian what happened? :) 19:41
(and I'm not banned.)
webmind my guess is a big load
mode/#perlhelp [+b *!*@feather.perl6.nl] 19:42
Juerd webmind: Thanks for the SMS 19:52
webmind np 19:53
didn't know if you where there already or simething
something
Juerd unbanned. 19:55
webmind: Nah - still eating :)
webmind k 19:56
thanks
Juerd 20:56 -!- mode/#perlhelp [-oo flang perl-d] by Juerd 19:57
20:56 -!- mode/#perlhelp [-b *!*@feather.perl6.nl] by Juerd
20:57 <@Juerd> Sorry about the bot de-op, but providing access should be a primary concern, and they
currently prevent legitimate users from accessing.
Try again.
lafk 19:58
s/l//
Dec 18 20:34:39 feather kernel: Swap cache: add 37775494, delete 37775494, find 16367583/20409750, race 5+29
Dec 18 20:34:39 feather kernel: Free swap = 0kB 19:59
Dec 18 20:34:39 feather kernel: Total swap = 987896kB
Load issue indeed.
Dec 18 20:34:39 feather kernel: Out of Memory: Killed process 18385 (ghc-6.4.1).
Someone forgot to ulimit.
(grrrr!)
afk
r0nny yo 20:33
anyone knows the state of audreyt's work on the new object system ? ? 20:36
gaal hey r0nny, you can see the commit log here: rt.openfoundry.org/Foundry/Project/.../pugs/log/ 21:19
clkao: is the problem with use fixed? if not where can I find a test case? 21:20
avar mm, anyone have an idea as to my previous question? 21:32
i.e. is there something like ruby's .inspect in p6?
gaal avar: I don't know ruby. Does .perl do what you want? 21:36
try: ./pugs -e 'class Foo; has $.x; my $foo = Foo.new(x => 42); $foo.perl.say' 21:37
avar woo, perfect
yeah, that's it
gaal cool 21:38
avar $ ruby -e 'a = [[1..2], [3..4]]; puts a.inspect'
[[1..2], [3..4]]
$ pugs -e 'my @a = ([1..2],[3..4]); say @a.perl'
[[1, 2], [3, 4]]
;)
gaal last I checked reflection wasn't specced very much, though. I imagine ruby has that. 21:39
avar well actually that 1..2 didn't create a range in ruby..
reflection?
gaal you know, "what is the signature of all the methods in this class" 21:40
but then, stevan and audreyt are working on that stuff.
...at least for OO code. I don't know about non-method routines in general. 21:41
stevan_ r0nny: re: the object system,.. I would say it will be about a week before we have anything working which can be used through Pugs 22:56
r0nny: it is a slow process, some of the stuff we are doing is being done for the first time (at least as far as I know, and have googled) 22:57
avar, gaal: re: reflection, the OO reflection should carry over to non-OO code since all things in p6 (at some level) are OO 22:59
however, some of those things might be more useful than others
22:59 evalbot_8018 is now known as evalbot_8301
stevan_ dinner & 23:00
Juerd I won't be having much time to finish the fears responses. 23:12
If you have ideas, please don't wait for me :)
clkao gaal: sorry? what problem with 'use'? 23:21

Powered by Cro and the Raku® Programming Language.

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

Thank you!