»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:00 replore left 00:01 gbacon left 00:02 pnu left
dalek ecs: 6b021ee | larry++ | S03-operators.pod:
Take a first quack at defining "duck mapping"

In which we bias unary hypers toward a shallow interpretation if the node in question responds to the operator, for a strict definition of "responds to" not including Cool coercions.		 00:03
ecs: fa5e087 | larry++ | S12-objects.pod:
Merge branch 'master' of github.com:perl6/specs
00:04 pnu joined
TimToady oh, good grief, what happened to my edits? 00:08
diakopter I see stuff in the first url
[Coke] rakudo: for 1, 2, 3 { sub foo {}; .say }
p6eval rakudo 38907e: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in block <anon> at /tmp/smwPX3cVYX:1␤ in method reify at src/gen/CORE.setting:3752␤ in method reify at src/gen/CORE.setting:3657␤ in method reify at src/gen/CORE.setting:3657␤ in method gimme at src/…
TimToady any git experts know an easy way to get 6b021ee merged back in? 00:10
and what I did wrong?
[Coke] I blame emacs. 00:11
(sorry, I'm usually asking the git questions here. ;)
rakudo: for 1, 2, 3 -> { sub foo {}; .say }
p6eval rakudo 38907e: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in block <anon> at /tmp/7n6PMmUEgH:1␤ in method reify at src/gen/CORE.setting:3752␤ in method reify at src/gen/CORE.setting:3657␤ in method reify at src/gen/CORE.setting:3657␤ in method gimme at src/…
TimToady I did a 'git pull' like it said, and then a push, and it blows away my version 00:12
diakopter TimToady: I see your edits in S03
TimToady oh, that would explain it :)
[Coke] I typically do "git pull --rebase" to avoid any mergy bits.
TimToady feels more than unusually stupid right now... 00:13
diakopter oh, now I see what you are asking, I think
00:13 soh_cah_toa joined
ingy I got make and make install working for pugs now 00:13
TimToady diakopter: the git issue was just a PEBCAK
too many other things on my mind... 00:14
00:15 Chillance left
ingy pugs: say 3 00:15
p6eval pugs: OUTPUT«3␤»
[Coke] ZOMG IT'S ALIVE! 00:16
ingy not my doing...
maybe it never died afterall
[Coke] rakudo: class A {}; class B {}; subset C of A & B; say C ~~ A;say C ~~ B 00:17
p6eval rakudo 38907e: OUTPUT«Type check failed in assignment to '$v'; expected 'Mu' but got 'C'␤ in method REIFY at src/gen/CORE.setting:4391␤ in method reify at src/gen/CORE.setting:3657␤ in method reify at src/gen/CORE.setting:3657␤ in method gimme at src/gen/CORE.setting:4015␤ in metho…
ingy masak: I'm ready to start PugsClub™ when you are 00:18
[Coke] rakudo: my $a = 1; sub foo() { my $a = 2; our sub bar($x where $a) { say $x } }; bar(1); bar(2) 00:20
p6eval rakudo 38907e: OUTPUT«Could not find sub &bar␤ in block <anon> at /tmp/XCB2X1q5A9:1␤ in <anon> at /tmp/XCB2X1q5A9:1␤»
[Coke] rakudo: sub foo() { my $a = 2; our sub bar($x where $a) { say $x } }; bar(2);
p6eval rakudo 38907e: OUTPUT«Could not find sub &bar␤ in block <anon> at /tmp/hSS1YJsEDY:1␤ in <anon> at /tmp/hSS1YJsEDY:1␤»
[Coke] rakudo: { &^c("$^a, $^b") }.("morning", "perl6ers", &say) 00:21
p6eval rakudo 38907e: OUTPUT«morning, perl6ers␤»
[Coke] rakudo: say "\xFDD0"
p6eval rakudo 38907e: OUTPUT«===SORRY!===␤Invalid character for UTF-8 encoding␤␤»
[Coke] rakudo: say "\xFFFD" 00:22
p6eval rakudo 38907e: OUTPUT«�␤»
[Coke] rakudo: say "\xFFFE"
p6eval rakudo 38907e: OUTPUT«===SORRY!===␤Invalid character for UTF-8 encoding␤␤»
[Coke] see lue?
aloha, seen lue? 00:23
aloha [Coke]: lue was last seen in #perl6 2 days 2 hours ago saying "use Grammar; and use Actions;".
00:26 donri left 00:31 pnu left 00:32 pnu joined
[Coke] nom: say [+]<264 456 7 -72> 00:34
p6eval nom 38907e: OUTPUT«655␤»
00:38 pnu left, pnu joined 00:39 the-golem joined
[Coke] oh look, someone just has to close out 5 more tickets to get us to 650! 00:47
cognominal I have not yet read TimToady prose but I expect it to answer my concern about recursive dwim mapping. To get something like [1,[2,3]] mapped using * * 2 into [2,[4,6]] 00:55
[Coke] rakudo: my %h; sub h { say "srsly wtf" }; enum Foo %h;
p6eval rakudo 38907e: OUTPUT«Could not find sub &Foo␤ in block <anon> at /tmp/P0SEYNmvr_:1␤ in <anon> at /tmp/P0SEYNmvr_:1␤»
cognominal I need a serious read of S03 to start with. 00:56
[Coke] rakudo: role Maybe[::T] { role Just[T] {} }; say Maybe[Int].new()
p6eval rakudo 38907e: OUTPUT«Maybe<-8106445831223374807>␤»
[Coke] rakudo: role A { class B {} } 00:57
p6eval rakudo 38907e: ( no output )
[Coke] rakudo: our sub xyz($abc) { say $abc.WHAT; }␤ xyz(1); say "alive" 01:01
p6eval rakudo 38907e: OUTPUT«Int()␤alive␤»
[Coke] rakudo: ␤our sub xyz($x) { say $x }␤␤xyz 3
p6eval rakudo 38907e: OUTPUT«3␤»
[Coke] rakudo: ␤our sub xyz($x) { say $x }␤␤xyz 3
p6eval rakudo 38907e: OUTPUT«3␤»
[Coke] whoops.
ingy pugs: say "seen audreyt" 01:03
p6eval pugs: OUTPUT«seen audreyt␤»
ingy awww
diakopter ingy: YAML::XS needs Test::More to install, but cpan doesn't try to install Test::More as a dependency 01:04
ingy diakopter: Perl 4?
diakopter erm
the-golem Hello. I just finished setting up an ilbot to log a couple of my channels. However, the process to set up the web interface has me completely buggered. Can anyone help me out?
[Coke] rakudo: sub a( %h?) { say %h.HOW }; a();
p6eval rakudo 38907e: OUTPUT«Method 'gist' not found for invocant of class 'Perl6::Metamodel::ClassHOW'␤ in sub say at src/gen/CORE.setting:4957␤ in sub a at /tmp/W657PMmq2T:1␤ in block <anon> at /tmp/W657PMmq2T:1␤ in <anon> at /tmp/W657PMmq2T:1␤»
diakopter ingy: perl 5.12.4 on fedora 15 01:05
ingy diakopter: j/k. I thought Test::More was core for forever
how could you be missing it on 5.12? 01:06
[Coke] ingy: since 5.6.2
diakopter fedora must not include it in the distro perl 01:07
[Coke] diakopter: a broken system perl?
[Coke] too slow.
diakopter although actually I had the same problem on cygwin last week
ingy sounds like a bigger prob than YAML::XS
[Coke] rakudo: sub foo($x where 1 --> Int) {} 01:08
p6eval rakudo 38907e: OUTPUT«===SORRY!===␤Missing block at line 1, near ") {}"␤»
[Coke] std: sub foo($x where 1 --> Int) {}
p6eval std 580b69a: OUTPUT«Potential difficulties:␤ $x is declared but not used at /tmp/dS2AwKhGYy line 1:␤------> sub foo(⏏$x where 1 --> Int) {}␤ok 00:01 122m␤»
ingy yaml::xs requires 5.8.1 btw
should be my problem, but I'd be glad to help 01:09
maybe ask on #toolchain
er
diakopter on magnet?
ingy shouldn't be my problem...
aye
[Coke] rakudo: sub twigil:<@> {}; 01:10
p6eval rakudo 38907e: ( no output )
[Coke] rakudo: sub twigil:<@> {}; say "alive!"
p6eval rakudo 38907e: OUTPUT«alive!␤»
ingy diakopter: I don't declare core deps. maybe I need to :\ 01:11
[Coke] ingy: probably safest.
rakudo: say :16<10> 01:12
p6eval rakudo 38907e: OUTPUT«16␤»
[Coke] rakudo: say :16<10>.WHAT
p6eval rakudo 38907e: OUTPUT«Num()␤»
[Coke] should that be an Int() and not a Num() ?
rakudo: :16<10> === 0x10 01:13
p6eval rakudo 38907e: ( no output )
[Coke] rakudo: say( :16<10> === 0x10 )
p6eval rakudo 38907e: OUTPUT«Bool::False␤»
[Coke] (That's RT#77034 if you have an opinion) 01:14
rakudo: Any.substr(0, 1) 01:15
p6eval rakudo 38907e: OUTPUT«Method 'substr' not found for invocant of class 'Any'␤ in block <anon> at /tmp/vrf7tHkzbo:1␤ in <anon> at /tmp/vrf7tHkzbo:1␤»
[Coke] rakudo: substr(Any, 0, 1)
p6eval rakudo 38907e: OUTPUT«Method 'substr' not found for invocant of class 'Any'␤ in sub substr at src/gen/CORE.setting:1601␤ in block <anon> at /tmp/RxkC_OtZn2:1␤ in <anon> at /tmp/RxkC_OtZn2:1␤»
[Coke] nom: say [+]<262 458 7 -75> 01:17
p6eval nom 38907e: OUTPUT«652␤»
[Coke] rakudo: class Forest { class Frog { method speak { say "ribbit ribbit" } }; has Frog $.frog; method new() { my Frog $frog .= new; self.bless(*, :$frog) } }; Forest.new.say; 01:19
p6eval rakudo 38907e: OUTPUT«Forest<3317213220548048558>␤»
[Coke] rakudo: multi sub f(0) {}; multi sub f($n) { say $n; callwith($n - 1) }; f(3)
p6eval rakudo 38907e: OUTPUT«3␤»
ingy trivia question: Why is #perl6 on freenode?
[Coke] nom: say [+]<262 458 7 -77> 01:21
p6eval nom 38907e: OUTPUT«650␤»
[Coke] \o/
zzz 01:22
ingy trivia answer: #perl6 is on freenode, because that is where ingy first /join'd it :)
o/ [Coke]
01:28 bluescreen10 left
colomon ingy++ 01:42
01:43 Josh4all joined, whiteknight left 01:44 Josh4all left 01:55 plobsing joined
ingy colomon: can you add ingydotnet to perl6 org in github? 02:12
please 02:13
colomon ingy: I don't know if I can or not. 02:14
I don't think I can. Looks like jnthn, moritz, pmichaud, sorear, and TimToady are the ones with admin rights. 02:15
02:15 am0c joined
colomon sorry 02:15
ingy jnthn, moritz, pmichaud, sorear, and TimToady: ^^ :) 02:16
jnthn, moritz, pmichaud, sorear, and TimToady: I'm pretty sure I mailed in my release form way back when....
colomon ingy: what projects do you want to be able to change? I think you only need a release form for rakudo itself... 02:18
sorear i've added ingydotnet to perl6 02:19
colomon sorear++ 02:20
ingy sorear++ 02:21
colomon sorear: I'm still stumped by the TreeViewColumn constructor.
my $renderer = CellRendererText.new; 02:22
my $column = TreeViewColumn.new($title, $renderer);
ingy au told me to fork pugs into that organization
sorear colomon: what is $title?
colomon sorear: for <Tune Snippet Comment>.kv -> $i, $title 02:23
sorear niecza: say <Tune Snippet Comment>.kv.perl # I have my doubts this will work at all
p6eval niecza v10-40-ga57ed3d: OUTPUT«(0, "Tune", 1, "Snippet", 2, "Comment").list␤»
02:23 tyatpi joined
sorear ....ok apparently it does. 02:24
colomon I just added say $title.WHAT; and it says GatherIterator()
which seems not worky
sorear ...
that shouldn' 02:25
that shouldn't even be possible.
colomon that certainly seems to be the root of my current problem, whatever it is.
ingy perl6 org now has a member with 200+ repos! ;)
I just noticed I had 201
sorear niecza: for <Tune Snippet Comment>.kv -> $i, $title { say $title.WHAT }
p6eval niecza v10-40-ga57ed3d: OUTPUT«Potential difficulties:␤ $i is declared but not used at /tmp/iWGnZPq2Sd line 1:␤------> for <Tune Snippet Comment>.kv -> ⏏$i, $title { say $title.WHAT }␤␤GatherIterator()␤»
sorear ingy: want me to add gitpan as a collaborator? :D
ingy puts his daily trash on github....
sorear niecza: for <Tune Snippet Comment>.kv -> $i, $title { say $title }
ingy haha
p6eval niecza v10-40-ga57ed3d: OUTPUT«Potential difficulties:␤ $i is declared but not used at /tmp/FLkyLTPSg7 line 1:␤------> for <Tune Snippet Comment>.kv -> ⏏$i, $title { say $title }␤␤GatherIterator.new(...)␤»
sorear double wonky 02:26
02:26 am0c left
ingy sorear: can I have permission to create a new repo in the org? 02:31
sorear: maybe easier to just fork github.com/audreyt/Pugs.hs for me
02:34 colomon_ joined, colomon left, colomon_ is now known as colomon
sorear ingy: I don't know 02:34
colomon niecza: for <Tune Snippet Comment>.kv -> $i, $title { say $i } 02:35
p6eval niecza v10-40-ga57ed3d: OUTPUT«Potential difficulties:␤ $title is declared but not used at /tmp/Rs93HdW1jX line 1:␤------> for <Tune Snippet Comment>.kv -> $i, ⏏$title { say $i }␤␤0␤»
ingy sorear: you don't know what?
colomon niecza: my @a = <Tune Snippet Comment>; for @a.kv -> $i, $title { say $title.WHAT } 02:36
p6eval niecza v10-40-ga57ed3d: OUTPUT«Potential difficulties:␤ $i is declared but not used at /tmp/ZVA07cwl4F line 1:␤------> <Tune Snippet Comment>; for @a.kv -> ⏏$i, $title { say $title.WHAT }␤␤Str()␤Str()␤Str()␤»
sorear ingy: what permissions are required to create repos
niecza: for [<Tune Snippet Comment>].kv -> $i, $title { say $title.WHAT }
p6eval niecza v10-40-ga57ed3d: OUTPUT«Potential difficulties:␤ $i is declared but not used at /tmp/2BgxkCsPH3 line 1:␤------> for [<Tune Snippet Comment>].kv -> ⏏$i, $title { say $title.WHAT }␤␤Str()␤Str()␤Str()␤»
ingy can you just fork that repo, sorear?
colomon sorear: using @titles.kv got me around my issue. odd bug, though... 02:37
sorear ingy: I forked the repo and gave perl6 access 02:38
02:40 aindilis left, aindilis joined 02:41 khisanth_ joined 02:42 Khisanth left 02:43 am0c joined 02:47 tyatpi left
colomon sorear: Success! www.harmonyware.com/pictures/gtk-treeview.png 02:48
02:48 bluescreen10 joined
sorear \o/ 02:51
ingy how do we make pugs commits show up here?
thx btw sorear 02:52
araujo colomon, using perl6?
colomon araujo: yup. niecza.
araujo colomon++
colomon should be sorear++, he's been a huge help. (Not to mention his creating niecza, of course.)
ingy installs niecza for fun...
araujo sorear++ 02:53
colomon source code is at www.harmonyware.com/pictures/gtk-treeview.png 02:55
02:55 preflex left, preflex_ joined
colomon it's still messy, I started with the shell of another perl6 Gtk script and have just kept hacking away at it 02:55
02:55 preflex_ is now known as preflex
colomon afk # bedtime 02:56
sorear bye 03:00
03:07 the-golem left
ingy sorear: time mono run/Niecza.exe -e '6**7**7' 03:10
less than 3 times lower than ruby :)
19 secs on niecza, 7 on ruby
not bad 03:11
sorear ingy: try again 03:13
ingy um, can I say something first
sorear the first time you run niecza it has to compile the setting
sadly
you may say something
ingy pugs.... 0m0.163s 03:14
03:14 khisanth_ is now known as Khisanth
ingy holy shit 03:14
sorear pugs++
ingy I use 6**7**7 as a general benchmark with ruby and others
sorear oh 03:15
ingy used to use 6**6**6 but things are faster these days
sorear I misread that as 6*7*7
so I thought 19s could only be explained as first-run startup time
anyways, pugs is going to be the fastest for that one because it's using GMP 03:16
ingy ingy@thinkingy:~/src/niecza$ time pugs -e 'say 6**7**7'|wc 1 1 640843
real0m0.867s
640K digits
the 19 sec niecza time was after a couple runs 03:17
sorear ingy: this is what Schönhage-Strassen multiplication looks like!
written in assembly by people who actually know how the Intel instruction scheduler works 03:18
ingy :D 03:19
just read up on GMP
sorear niecza is using Rodrigo Kumpera's C# bignum library. Which is optimized for being fairly small and doesn't even support Karatsuba
ingy is that built into haskell?
sorear it's built into the Glasgow Haskell Compiler
ingy yeah, sorry, meant ghc
sorear: you mean The Glorious Glasgow Haskell Compilation System ;0 03:20
it's not a compiler, man. it's a COMPILATION SYSTEM 03:21
sorear yes, ghc is more than a compiler. it also includes runtime libraries and an REPL. 03:22
ingy is gonna try to get pugs working against the modern test suite. 03:24
masak++ (you hear me? :)
sorear: can niecza use GMP? 03:25
one time I tried 6**7**7 with ruby on my native machine and on a VM 03:27
03:27 envi_ joined
ingy the VM was over 4 times faster 03:27
it turned out the VM was 64bit, and native OS was 32
so the 19sec time is on 32 bit, sorear 03:28
I wonder how fast pugs is on 64
sorear faster, I'm sure 03:31
ingy: niecza cannot use gmp
ingy sorear:
ingy@thinkingy:~/src/niecza$ time pugs -e 'say 8**8**8'
Inf
oops
looks like something is wrong 03:33
03:34 soh_cah_toa left 03:55 tewk joined 03:58 tewk left 04:00 tewk joined 04:06 replore_ joined 04:13 satyavvd joined 04:17 bluescreen10 left 04:23 daniel-s joined 04:27 drbean left 04:30 drbean joined 04:32 am0c left
dalek ecza/serialize: b1088dc | sorear++ | lib/ (3 files):
implement serialization for SubInfo, LAD, LexInfo
 04:35
04:36 icwiener left 04:50 am0c joined 04:51 Reaganomicon left 04:52 Reaganomicon joined 05:13 replore_ left, yath left, _sri left, dukeleto left, _ilbot left, jdhore left, tty234_ left, spacebat_ left, imarcusthis left, arnsholt left, jesk left, lumi_ left, Helios left 05:14 replore_ joined, yath joined, _sri joined, dukeleto joined, _ilbot joined, jdhore joined, tty234_ joined, spacebat_ joined, imarcusthis joined, arnsholt joined, jesk joined, lumi_ joined, Helios joined, tty234_ left 05:15 tty234 joined 05:18 drbean left, SHODAN joined 05:24 drbean joined 05:29 lue left 05:33 sftp_ joined, sftp left
moritz hugme: add ingydotnet to perl6 05:43
05:43 hugme left 05:46 hugme joined, ChanServ sets mode: +v hugme
moritz phenny: tell soh_cah_toa that source_url in raw.github.com/soh-cah-toa/p6-dige.../META.info points to the old repo name (without the p6- prefix) 05:55
phenny moritz: I'll pass that on when soh_cah_toa is around.
05:58 Khisanth left 06:01 dual joined 06:02 GlitchMr joined 06:10 ggoebel left 06:12 ggoebel joined 06:23 kaleem joined
dalek ecza/serialize: d91ecdf | sorear++ | lib/ (5 files):
Serialization for objects of type Rat, FatRat, Complex, BigInteger, Int32, Double, String, P6any[], Variable[], VarDeque, VarHash
 06:24
sorear wonders to what extent Frame objects should be freezable 06:32
if they are freezable, it should be in a way that doesn't chase CALLER, me thinks
o/ GlitchMr 06:33
moritz so when you return from one, it's to your new caller? 06:34
sorear I don't know
the use case is constant @foo = 1, 3, 5 ... *; 06:35
that creates a List which wraps a GatherIterator which wraps a Frame
but I don't want it saving a copy of the compiler by chasing up the call tree :D
I'm too tired to be certain atm
sorear sleep 06:39
dalek ecza/serialize: ab35238 | sorear++ | lib/ (3 files):
Tidy up STable fields a bit, add TiedVariable serialize
 06:40
06:50 am0c left 06:57 the-golem joined
the-golem Moritz, my friend :-) 06:57
moritz: If you've the time, I could use some assistance in getting the front-end to ilbot working 07:01
i've been dumbing my way through it for a while, and i'm just about at my wit's end. It's likely something brainless that my dull wits cannot grasp 07:03
07:03 grondilu joined
grondilu perl6: say ?(0|1 == 0&1); say ?(0&1 == 0|1) 07:05
p6eval pugs: OUTPUT«1␤1␤»
..rakudo 38907e: OUTPUT«Bool::False␤Bool::True␤»
..niecza v10-40-ga57ed3d: OUTPUT«Bool::True␤Bool::True␤»
grondilu amazing stuff. The '==' operator is not commutative with junctions. Just as it would be in quantum mechanics :) 07:06
but only according to rakudo
tadzik b: say ?(0|1 == 0&1); say ?(0&1 == 0|1) 07:07
p6eval b 1b7dd1: OUTPUT«Bool::True␤Bool::True␤»
07:07 am0c joined
tadzik seems NYI in the new rakudo 07:07
grondilu what's b: ? 07:08
rakudo say ?(0|1 == 0&1), ?(0&1 == 0|1) 07:09
rakudo: say ?(0|1 == 0&1), ?(0&1 == 0|1)
p6eval rakudo 38907e: OUTPUT«Bool::FalseBool::True␤»
masak mroing, #perl6
grondilu I think rakudo's behavior is better. Much more interesting anyway.
But it gives a left associativity to operators that normaly should be chained, and not left associatives. 07:12
07:13 baest joined
masak 0|1 == 0&1 means any(0,1) == all(0,1) means all(any(0,1) == 0, any(0,1) == 1) means all(any(0 == 0, 1 == 0), any(0 == 1, 1 == 1)) 07:13
which is true, AFAICS. 07:14
masak submits rakuodbug
tadzik lol, masak
masak rakudobug*
tadzik good morning
masak :)
seems I showed up just in time :P
grondilu: commutativity should hold with junctions in Perl 6. 07:15
grondilu I disagree.
It should not.
masak I may be wrong on this. it would be easier to work from an example.
but given the ways junctions are evaluated, it feels like commutativity still holds. 07:17
grondilu 0|1 is always equal to any boolean. 0&1 is always not equal to a boolean. Thus, there is a conflict between 0&1 and 0|1 and there is some kind of incertitude principle. It nicely emerges from rakudo as the result of == depends of the order of the arguments. 07:18
As a result we get non-commutativity, just as in QM. I think it's great. 07:19
07:19 GlitchMr left
masak sure, but you seem to be ignoring the order of evaluation described in S09. 07:19
with it in mind, there is no conflict, and no incertitude. 07:20
grondilu oh. I need to read this. Sorry.
masak S09:996 07:21
07:22 mberends joined
masak mberends! \o/ 07:22
grondilu is 996 a line number?
(I get to a notice about autoindex when I get there)
mberends \o/ masak, $customer.network allows IRC :) 07:23
mberends backlogs 07:24
masak perl6: say Array ~~ Cool 07:26
p6eval rakudo 38907e, niecza v10-40-ga57ed3d: OUTPUT«Bool::True␤»
..pugs: OUTPUT«*** No such subroutine: "&Cool"␤ at /tmp/nrhmKBFeq8 line 1, column 5 - line 2, column 1␤»
masak grondilu: "The results of these separate calls..." 07:27
the duck mapping looks cool, but I'm a little concerned about the violence it perpetrates on ordinary MMD. 07:30
grondilu MMD? 07:31
seriously it's a bit more than a duck mapping. Again, 0|1 is always true with any boolean, and 0&1 is always false. So == should not priviledge any False or True result when being applied to 0|1 and 0&1 in the same time. 07:33
07:34 drbean left
masak my duck mapping comment wasn't related to our junctions discussion. 07:34
but it seems you either didn't read S09:996, didn't understand it, or don't agree that it should work that way. 07:35
07:35 domidumont left
grondilu Ok ok. I need to read S09 indeed. 07:35
07:35 grondilu left 07:41 drbean joined 07:43 sftp_ left
moritz TimToady: in your latest specs patch, + $node.foo // $node.list: { .».foo } 07:45
should that be $node.map: ... or so? 07:46
I don't think .list takes a closure argument
masak moritz++
masak frowns at rt.perl.org/rt3/Ticket/Display.html?id=78232 07:48
moritz rakudo: for 1, 2, 3 { sub foo {}; .say }; say "alive" 07:49
p6eval rakudo 38907e: OUTPUT«Too many positional parameters passed; got 1 but expected 0␤ in block <anon> at /tmp/cicnSsXJMb:1␤ in method reify at src/gen/CORE.setting:3752␤ in method reify at src/gen/CORE.setting:3657␤ in method reify at src/gen/CORE.setting:3657␤ in method gimme at src/…
moritz rakudo: for 1, 2, 3 { .say }; say "alive"
p6eval rakudo 38907e: OUTPUT«1␤2␤3␤alive␤»
masak oh, and moritz, you didn't provide enough steps-to-reproduce on rt.perl.org/rt3/Ticket/Display.html?id=76692 , it seems. 07:50
rt.perl.org/rt3/Ticket/Display.html?id=73344 needs reformulating because 'our' subs don't work that way anymore. 07:52
(not addressing moritz++ with that one, but the general public) 07:53
not sure the same behaviour can be had with the new 'our' sub semantics.
rakudo: { &^c("$^a, $^b") }.("morning", "perl6ers", &say)
p6eval rakudo 38907e: OUTPUT«morning, perl6ers␤» 07:54
masak \o/
moritz masak: just return sub bar from sub foo
masak moritz++: I'll try that
rakudo: my $a = 1; sub foo { my $a = 2; sub bar($x where $a) { say "called with $x" }; &bar }; foo()(2); foo()(1) 07:55
p6eval rakudo 38907e: OUTPUT«called with 2␤Constraint type check failed for parameter '$x'␤ in sub bar at /tmp/MRp2dEpxLI:1␤ in block <anon> at /tmp/MRp2dEpxLI:1␤ in <anon> at /tmp/MRp2dEpxLI:1␤»
moritz rakudo: my $a = 1; sub foo() { my $a = 2; sub bar($x where $a) { say $x } }; my &bar := foo; bar(1)
p6eval rakudo 38907e: OUTPUT«Constraint type check failed for parameter '$x'␤ in sub bar at /tmp/Ob_vsJ7vrl:1␤ in block <anon> at /tmp/Ob_vsJ7vrl:1␤ in <anon> at /tmp/Ob_vsJ7vrl:1␤» 07:56
moritz rakudo: my $a = 1; sub foo() { my $a = 2; sub bar($x where $a) { say $x } }; my &bar := foo; bar(1)
p6eval rakudo 38907e: OUTPUT«Constraint type check failed for parameter '$x'␤ in sub bar at /tmp/9fraezIidG:1␤ in block <anon> at /tmp/9fraezIidG:1␤ in <anon> at /tmp/9fraezIidG:1␤»
masak rakudo: my $a = 1; sub foo { my $a = 2; our sub bar($x where $a) { say "called with $x" }; &bar }; foo()(2); foo()(1)
p6eval rakudo 38907e: OUTPUT«called with 2␤Constraint type check failed for parameter '$x'␤ in sub bar at /tmp/7MnZBwhjCT:1␤ in block <anon> at /tmp/7MnZBwhjCT:1␤ in <anon> at /tmp/7MnZBwhjCT:1␤»
masak all indications point to it working nowadays. 07:57
masak adds that to the ticket
moritz huh?
masak no?
moritz my example failed with 2
masak oh!
masak looks again
moritz nom: my $a = 1; sub f($ where $a) { }; f 1
p6eval nom 38907e: ( no output )
moritz nom: my $a = 1; sub f($ where $a) { }; f 2
p6eval nom 38907e: OUTPUT«Constraint type check failed for parameter '(null)'␤ in sub f at /tmp/VqsIIv0Mvb:1␤ in block <anon> at /tmp/VqsIIv0Mvb:1␤ in <anon> at /tmp/VqsIIv0Mvb:1␤»
moritz so in general it works
masak please run the failing one again. 07:58
I don't see it above; it's so messy.
moritz rakudo: my $a = 1; sub foo() { my $a = 2; sub bar($x where $a) { say $x } }; my &bar := foo; bar(2)
07:58 koban joined
p6eval rakudo 38907e: OUTPUT«2␤» 07:58
moritz oh
masak right.
moritz maybe it was the mess :-)
rakudo: my $a = 1; sub foo() { my $a = 2; sub bar($x where $a) { say $x } }; my &bar := foo; bar(1)
p6eval rakudo 38907e: OUTPUT«Constraint type check failed for parameter '$x'␤ in sub bar at /tmp/u9nPOhDfn6:1␤ in block <anon> at /tmp/u9nPOhDfn6:1␤ in <anon> at /tmp/u9nPOhDfn6:1␤»
masak so it works.
moritz aye
07:59 domidumont joined
masak I think rt.perl.org/rt3/Ticket/Display.html?id=67942 can be closed, but it's the requestor's (moritz) call. 08:04
moritz I still think it's wrong 08:05
we shouldn't reparse a declarator as a listop when the parse later on goes wrong 08:06
masak agreed. 08:07
but it's a NPMCA ticket, and the NPMCA is gone.
moritz std: my %h; enum Foo %h;
p6eval std 580b69a: OUTPUT«===SORRY!===␤Illegal redeclaration of symbol 'Foo' (see line 1) at /tmp/pcfy2Yu9ky line 1:␤------> my %h; enum Foo⏏ %h;␤Undeclared routine:␤ 'h' used at line 1␤Check failed␤FAILED 00:01 122m␤»
moritz ok, we can close it, and file a new one :-)
and a std ticket to TimToady too
masak I don't like saying "something else is still wrong" about a ticket without changing the stated expectations.
it makes it hard for RT robots like [Coke] and me to reason about tickets. 08:08
moritz you're an RT robot? :-)
masak you didn't know? :-)
I'm the ticket-producing kind. 08:09
dalek ast: c0c24ab | moritz++ | S02-types/subset.t:
tests for RT #73344, where-constraint picked up the wrong outer lexical in rakudo
 08:17
moritz TimToady: again re latest spec patch, either you simply that there is a Hash.map which returns a Hash (and of which I don't know anything), or that .deepmap turns hashes into lists 08:20
masak imply* 08:26
moritz correct 08:27
masak turning hashes into lists sounds more like the Perl I know. 08:31
by the way, I'm having a Perl 6 day today. 08:36
I have two goals: blog a bit, and hack on the macros branch.
moritz wanted to blog about the three-way balancing act of the smartmatch operator, and that it's certainly at least one too much 08:38
masak I can see you pulling off such a blog post well. 08:40
ooc, what are the three ways?
moritz 1) the actual smart matching 2) topicalization 3) invoking regexes 08:41
note that 'abc' ~~ m:g/./ does not fit into the "does the $lhs match the $rhs?" type of question
(because a boolean answer doesn't need :g) 08:42
masak the topicalization bit is fairly new, innit? 08:43
ISTR it was introduced in discussions during YAPC::EU 2010.
moritz like, 2 years or so?
no, before
masak oh, ok.
moritz I remember discussing some fallout of that decision at YAPC::EU 2010 though 08:44
masak oh! that was it. 08:45
yes, please make such a blog post.
I'll try to write one about macros.
08:48 Khisanth joined 08:49 dakkar joined 08:53 wtw joined 09:03 wk joined
masak ooh! there could be twigil macros! 09:19
09:28 orafu left, orafu joined 09:42 Mowah_ joined, drbean left 09:44 mattp_ left, mattp_ joined
dalek kudo/nom: e0cc816 | masak++ | NOMMAP.markdown:
[NOMMAP.markdown] typography fixes
 09:48
09:49 drbean joined
masak TimToady: re panic: I don't actually think I've panicked for real so far. most of my complaints and kvetches stem either from a genuine appreciation and caring for the Perl 6 project, or from wanting to focus more on coding and less on arbitrary official release ceremonies. 09:55
if what I am doing counts as panic, I want to point out that I'm not finished with my first and only one yet :P 09:59
moritz: fwiw, I've totally come around to TimToady's view that "I want a thunky parameter" means you want a macro but maybe don't know it yet. 10:01
moritz masak: well, I actively don't want a macro, because I wouldn't know how to evaluate the AST that the macro places for me in the parameters 10:02
masak the 'is ast' trait won't fly, I think, since a routine either is a macro or isn't. marking up some of the params and not others doesn't make much sense.
moritz masak: I don't want to deal with that complexity. I just want a thunk that I can invoke
masak moritz: we probably need an AST evaluate at some point. 10:03
s/evaluate/evaluate function/
`leave` happening at unwinding time, not at throw time -- does that mean that the CATCH block has "magic" access to lexicals in the block that threw the exception? 10:14
i.e. is the "effective" lexical scope disconnected from the "textual" lexical scope in the CATCH block until the unwind happens? 10:16
and if so, how does the CATCH block have access to its own declared lexicals?
mls_ lexical lookup has nothing to do with call frames 10:17
(hi perl6)
masak oh, true.
yeah, I'm mixing up lexpads and call frames.
my bad :)
so what does it mean, in effect, that the call stack isn't unwound at the point of the CATCH? 10:18
moritz that LEAVE hasn't happened yet
and that you might resume
mls_ well, you have access with the callframe() function
niecza: { warn() ; CONTROL { default {} } } 10:23
p6eval niecza v10-40-ga57ed3d: OUTPUT«␤Unhandled Exception: No value for parameter $str in CORE warn␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (CORE warn @ 0) ␤ at /tmp/rFYiqWnBJv line 1 (MAIN C1_ANON @ 2) ␤ at /tmp/rFYiqWnBJv line 1 (MAIN mainline @ 1) ␤ at /home/p6eval/niecza/lib/C…
mls_ niecza: { warn("foo") ; CONTROL { default {} } }
p6eval niecza v10-40-ga57ed3d: ( no output )
mls_ niecza: { LEAVE { warn("foo") } ; CONTROL { default {} } } 10:24
p6eval niecza v10-40-ga57ed3d: ( no output )
10:24 ab5tract joined
mls_ niecza: { LEAVE { warn("foo") ; warn("bar") }; warn("hello") } 10:25
p6eval niecza v10-40-ga57ed3d: OUTPUT«hello␤ at /home/p6eval/niecza/lib/CORE.setting line 613 (CORE warn @ 2) ␤ at /tmp/74b5BBg9I3 line 1 (MAIN C1_ANON @ 2) ␤ at /tmp/74b5BBg9I3 line 1 (MAIN mainline @ 1) ␤ at /home/p6eval/niecza/lib/CORE.setting line 2133 (CORE C1024_ANON @ 2) ␤ at /home…
mls_ niecza: { LEAVE { warn("foo") ; warn("bar") } }
p6eval niecza v10-40-ga57ed3d: OUTPUT«foo␤ at /home/p6eval/niecza/lib/CORE.setting line 613 (CORE warn @ 2) ␤ at /tmp/whyRLnCvEv line 1 (MAIN C2_ANON @ 2) ␤ at line 0 (ExitRunloop @ 0) ␤ at /tmp/whyRLnCvEv line 0 (MAIN C1_ANON @ 0) ␤ at /tmp/whyRLnCvEv line 1 (MAIN mainline @ 1) ␤ at /h…
snarkyboojum dude - these errors are so dense :D 10:27
that was supposed to be quoted, but c'est la vie 10:28
masak snarkyboojum! \o/ 10:31
snarkyboojum masak: dude! o/
masak Excellent. 10:33
snarkyboojum masak: did your macros grant idea get off the ground eh?
masak snarkyboojum: well, I'm working on D1 and it's going fine. haven't heard back from TPF yet. 10:34
snarkyboojum masak: ah - awesome :) 10:36
10:38 fhelmberger_ left
masak snarkyboojum: if you want to try macros right now, 'git checkout macros' in the rakudo repository. 10:41
snarkyboojum masak: it's well beyond me, but it sounds fun :)
masak snarkyboojum: well, using them is something anyone can do. 10:46
snarkyboojum: macro foo { say "OH HAI" }; say 1; foo; # OH HAI\n1\n
hm, but that would probably result in an error. it needs an empty quasi at the end.
snarkyboojum: macro foo { say "OH HAI"; quasi {} }; say 1; foo; # OH HAI\n1\n 10:47
snarkyboojum feels compelled to make some macrocosmic response :) 10:49
masak but you see what's happening there. syntactically it's just a subroutine, except it gets called when the parser finds it, not when the runtime finds it. 10:50
snarkyboojum so it's like you're extending the parser or the parse phase? 10:53
masak yes, I guess. 10:54
a macro is a parse-time subroutine.
and the return value of the macro (the quasi block) gets re-incorporated into the code and compiled, too.
snarkyboojum what's to stop you extending the parse phase into something unparseable?
masak the fact that you have to obey normal syntactical rules inside the quasi block. 10:55
well, normal to a first approximation ;)
you're allowed to have "holes" with bits of AST parameters in the quasi block, similar to how you can interpolate variables into qq strings in Perl. 10:56
macro add($a, $b) { quasi { {{{$a}}} + {{{$b}}} } } # $a and $b contain bits of Perl 6 code
so anywhere the parser saw 'add(40, 2)' it'd replace it by '40 + 2'. 10:57
snarkyboojum so you have to follow normal syntactic rules which are defined by the parser/grammar, which you're extending via macros? :D
masak yup :)
snarkyboojum that sounds like circular fun to me
masak you've put your finger on a delicate point there.
it's technically possible to call a macro from within itself. 10:58
there's actually two cases there. recursing in the macro but outside of the quasi, and recursing inside the quasi.
the latter case is prefectly fine, as long as you bottom out.
the former case -- I'm not sure it's wise, or even technically possible.
problem is, at that point, the compiler is still trying to come to grips with the macro definition. and then it calls itself. not good. 10:59
snarkyboojum it seems that the parser would need to know ahead of time what is syntactically OK and what isn't, but you can extend what that is, so how does the parser know? 11:00
masak oh, that works just like with sub calls and operators. 11:02
there are defaults.
the spec has a bit of wording on that. 11:03
if you go 'foo 1, 2, 3, 4, 5', the parser assumes that &foo is a listop, even if it hasn't seen the declaration of &foo yet.
snarkyboojum but you can change the meaning of a listop in a macro? 11:04
11:04 drbean left 11:10 drbean joined
masak well, yeah. macros are all about changing the meaning. 11:11
but so are subs, only in a different phase of the program. 11:12
11:12 replore_ left
snarkyboojum subs change meaning? in a different phase? 11:15
masak a sub, in a very real sense, is something that transforms a subcall into a return value. 11:16
snarkyboojum subs change meaning during runtime, but macros change meaning when you're building the semantics of a program?
masak yes! \o/
achievent unlocked!
snarkyboojum so you can have macros which change the semantics of subs :D 11:17
masak sure.
snarkyboojum but you can't transform a runtime thing into a macro thing? :D 11:18
11:20 fhelmberger joined 11:21 Alias_ joined 11:23 the-golem left 11:25 jaldhar_ is now known as jaldhar
masak snarkyboojum: not sure what you mean by that. it sounds like it would fly in the face of causality. 11:26
moritz you could inject macros into an eval(), which compiles at run time again
masak right, but most of the program will already have been parsed at that time. 11:27
snarkyboojum I guess I'm wondering if a macro can cause a program to be reparsed?
masak not as far as I know. 11:28
snarkyboojum that would seem to make things interesting
:)
[Coke] nom:say 723-77 11:30
rakudo: class Foo does Positional[::T] {}; say Foo ~~ Positional 11:32
p6eval rakudo e0cc81: OUTPUT«===SORRY!===␤Could not locate compile-time value for symbol T␤»
[Coke] (RT #68134) - is that still the right way to write that?
masak think so. 11:33
[Coke] rakudo: my $a = Multi.new; $a.push(sub (1,2,3) { }); $a.push(sub (1,2) { }); say join ", ", $a.candidates
p6eval rakudo e0cc81: OUTPUT«Could not find sub &Multi␤ in block <anon> at /tmp/6TeqtEsvyz:1␤ in <anon> at /tmp/6TeqtEsvyz:1␤»
[Coke] commutes 11:34
masak [Coke]: that one has fossilized. there's no Multi type any more. 11:35
moritz nom: proto sub foo($) { }; say &foo.WHAT 11:36
p6eval nom e0cc81: OUTPUT«Sub()␤»
11:37 Psyche^ joined 11:39 satyavvd left
flussence yay, I just got to rewrite something using event-sourcing at $dayjob and a bunch of bugs disappeared 11:40
11:40 Patterner left, Psyche^ is now known as Patterner
moritz nice 11:41
masak flussence: woot! 11:43
flussence masak++ # teaching us magic tricks :) 11:44
masak it's been almost a year, and I'm still stoked on the idea of event sourcing. :)
11:45 benabik joined
flussence the funny thing is, our main dataset vaguely resembles that already. It's just this one little bit that got it horribly wrong 11:45
masak flussence: did I mention that github.com/masak/farm is event-sourced? 11:46
flussence I think you did :)
really cool stuff now that I've figured out how to do it 11:47
masak :D
errand &
dalek kudo/nom: 3fd3534 | moritz++ | t/spectest.data:
run mro.t
 11:56
snarkyboojum "Macros may return either a string to be reparsed, or a syntax tree that needs no further parsing." 11:58
does that mean macros can cause a program to be reparsed, or am I reading that wrong? 11:59
moritz only the returned string is reparsed
and it's reparsed at compile time, because macros are run at compile time 12:00
snarkyboojum "returned string"? 12:01
returned from the macro
isn't a program a big-arse macro? :D
benabik No. The programmer is a macro (no comment on size of arse), and the compiler parses the string returned by them. 12:03
snarkyboojum benabik: so by definition, the compiler parses programmers.. I like it 12:04
12:06 HarryS left
benabik snarkyboojum: They say what languages you know/use affects the way you think. ;-) 12:10
12:10 ab5tract left
snarkyboojum benabik: linguistic relativity? :D 12:12
12:12 HarryS joined
snarkyboojum i.e. the sapir-whorf hypothesis? 12:12
benabik Knew there was some fancy name for it. Yes, although programming languages probably have a smaller effect. :-) 12:14
dalek ast: dbcee47 | moritz++ | S12-methods/defer-call.t:
test for RT #69756
 12:16
snarkyboojum programming languages being a subset of human languages perhaps ;)
benabik Programming languages having a smaller problem domain than human language. Most people don't try to program their everyday life. 12:17
12:23 ab5tract joined
snarkyboojum benabik - a completely different abstraction imo :) 12:24
12:39 replore joined 12:48 m6locks left 12:59 grondilu joined 13:01 am0c left 13:09 BudiantoWang joined
moritz blogs.perl.org/users/rafael_garcia-...uture.html 13:13
13:16 GlitchMr joined
PerlJam I see your link and raise you www.modernperlbooks.com/mt/2011/10/...ption.html 13:26
13:31 im2ee joined
im2ee Hello 13:35
tadzik hello im2ee
colomon sorear, mberends: Having trouble figuring out how to wire up my "button toggled" callback to the CellRendererToggle object. Any notions? 13:37
13:39 Trashlord left
dalek gs.hs: f5d5f76 | (Ingy dot Net)++ | Pugs/Makefile:
make clean
 13:39
gs.hs: 9a11832 | (Ingy dot Net)++ | Pugs/Makefile:
Makefile basics

builds pugs by default. runs cabal configure when needed.
support 'make install'
gs.hs: b1c99f1 | (Ingy dot Net)++ | Pugs/Makefile:
Account for all .hs sources
moritz (that was the "Test hook" button) 13:40
13:42 envi_ left
ingy moritz++ 13:42
13:46 envi_ joined 13:49 SHODAN left 13:55 koban left
mls_ aloha: seen jnthn 13:58
aloha mls_: jnthn was last seen in #perl6 1 days 2 hours ago saying "(Due to the gradual typing.)".
13:59 Trashlord joined
moritz irclog.perlgeek.de/phasers/2011-10-05#i_4526983 13:59
14:01 kaleem left
mls_ thanks moritz (I need him to look at my exceptions/enter/leave commits) 14:02
moritz mls_: try to tell him via phenny 14:03
phenny: tell mls_ like this
phenny moritz: I'll pass that on when mls_ is around.
mls_ I know phenny. But we probably need to do some live chat about design decisions 14:05
phenny mls_: 14:03Z <moritz> tell mls_ like this
moritz ok 14:06
14:07 Exodist_ is now known as Exodist
moritz tests the branch 14:07
14:18 wtw left 14:19 domidumont left, ggoebel left, yath left, _sri left, dukeleto left, _ilbot left, jdhore left, spacebat_ left, imarcusthis left, arnsholt left, jesk left, lumi_ left, Helios left 14:20 _sri joined
moritz that does feel nice 14:23
throwing exception in a CATCH handler, and it's not re-caught
14:23 jesk joined
moritz working ENTER and LEAVE 14:23
mls_++
mls_ I'm not sure about the enter/leave implementation, though. It's currently done with no special parrot patches. 14:24
14:24 domidumont joined, ggoebel joined, yath joined, dukeleto joined, _ilbot joined, jdhore joined, spacebat_ joined, imarcusthis joined, arnsholt joined, lumi_ joined, Helios joined
mls_ It might turn out that we need to do it in a different way, e.g. for language interop 14:24
moritz wow, I can even put a CATCH block inside a CATCH block to catch the exception from the outer CATCH :-) 14:25
14:25 risou_awy is now known as risou
mls_ also exceptions in LEAVE blocks need some tinkering 14:25
benabik mls_: If patching parrot makes things cleaner, it's not impossible. :-) #parrot is always interested in how we can make things better. 14:27
That said, if it works without it, then who cares? ;-)
mls_ thanks. But we first need to know what we need from parrot.
benabik Clear goals always helps. :-) 14:28
mls_ ;)
grondilu I tried once to write a perl6 CGI with no success. I get a Method 'exists' not found for invocant of class 'Hash'
moritz benabik: I think mls_ even has a commit bit for parrot these days
grondilu: I remember seeing a ticket about that. I think it's closed in nom 14:29
nom: say %*ENV.exists('HOME')
p6eval nom 3fd353: OUTPUT«Bool::True␤»
moritz nom: say %*ENV.exists('NOTTHERE')
p6eval nom 3fd353: OUTPUT«Bool::False␤»
moritz looks better, eh? :-)
grondilu doesn't know what 'nom' is
moritz b: say %*ENV.exists('NOTTHERE')
p6eval b 1b7dd1: OUTPUT«Bool::False␤»
moritz grondilu: it's the current development branch of rakudo
grondilu ok 14:30
moritz if you use the branch or the 2011.09 release of rakudo, that's what you get
grondilu my version is '2011.09-69-g3fd3534' 14:31
that's not 'nom', is it?
moritz that's nom. 14:32
grondilu Well, I still get the 'exists' error :-/ 14:33
PerlJam grondilu: show the code that gives the error
grondilu hang on
moritz mls_: a few tests fail in your branch, but I guess many spectests are bogus or outdated when it comes to exceptions
14:34 pmurias joined
mls_ yes, I think four or five fail 14:34
moritz for example the last one in try.t is bogus, it never marks $! as handled, nor does it smartmatch successfully against $_ 14:35
pmurias sorear: how can i see the serialisation branch?
grondilu it's nested in a bash script as this is the only way I could see the error message: pastebin.com/NhyGCMRJ
mls_ yes, it needs a default { ... }
pmurias git branch only shows master?
grondilu it's just a perl6 -e ';' and it works fine outside of CGI
moritz pmurias: 'git branch' only shows local branches by default. Try git branch -a for more
grondilu: wow, that's scary 14:36
PerlJam maybe something with %*ENV?
moritz yes, but what?
PerlJam that it's not there? :)
grondilu there is probable a CGI env var that perl6 doesn't like.
moritz grondilu: could you please change that to perl6 --ll-exception -e ';'
yes, I suspect so 14:37
grondilu: and tell us what output that produces
grondilu oh my god
hang on
pmurias moritz: so how do i switch to a branch?
moritz pmurias: git checkout otherbranch 14:38
grondilu pastebin.com/qayiZBxB
moritz grondilu: thanks
grondilu I'm using thttpd under Debian if this is of any help 14:39
pmurias moritz: works 14:40
moritz: thanks
colomon sorear, mberends: Aha! add_Toggled is the key. of course, now I need to make my DoneToggled sub actually work. But that will come after exercise. 14:41
moritz that backtrace is ... weird
while loading the setting, a concatenation triggers an undef warning (thats' the 'note', line 4) 14:42
sorear good * #perl6
moritz which seemingly dies while trying to look up the $*ERR variable 14:43
colomon sorear: oooo, good to see you
moritz o/ sorear
colomon sorear: decided to try one more thing before exercising. public bool GetIterFromString (out TreeIter iter, string path_string); -- is it possible to get the out iter there? 14:44
grondilu anyway how is it possible that perl6 doesn't find a method 'exists' for a 'Hash' class??
PerlJam At first blush I though that was a parrot error btw.
moritz grondilu: there's a parrot Hash which doesn't have an exists method
sorear grondilu: either Hash.exists doesn't exist, or it's not the 'Hash' you think it is 14:45
moritz nom: pir::new__PS('Hash').exists('foo')
p6eval nom 3fd353: OUTPUT«Method 'exists' not found for invocant of class 'Hash'␤ in block <anon> at /tmp/vTbVojSoSY:1␤ in <anon> at /tmp/vTbVojSoSY:1␤»
moritz nom: Hash.new.exists('foo')
p6eval nom 3fd353: ( no output )
colomon I thought Hash.exists was obsolete?
sorear colomon: it is, but that's beside the point
moritz so, the question is, why does GLOBAL.WHO or PROCESS.WHO end up being a parrot hash?
colomon: we still use it in rakudo until we can parse the :exists adverb 14:46
PerlJam not enough boxing
moritz nom: say GLOBAL.WHO
p6eval nom 3fd353: OUTPUT«().hash␤»
moritz nom: say PROCESS.WHO
sorear colomon: what's the trouble with getting the out iter? it should Just Work, although you'll need to preset the variable to TreeIter to make MMD work
p6eval nom 3fd353: OUTPUT«("\$IN" => IO<7868443299104802926>, "\$OUT" => IO<7868443299104231782>, "\$ERR" => IO<7868443299112620663>, "\$TZ" => $*TZ, "\@ARGS" => [], "\$ARGFILES" => ArgFiles<7868443299111605347>, "\%ENV" => {"PERLBREW_VERSION" => "0.12", "USER" => "p6eval", "MAIL" => "/var/mai…
moritz PerlJam: if it were that simple, one of the above lines would fail 14:47
pmurias sorear: hi 14:48
colomon sorear: Ah, I didn't know. Hmmm... how to preset the variable to TreeIter....
mls_ Hey, I can reproduce it: env - ./perl6 -e ';'
PerlJam mls++ 14:49
grondilu mls++ indeed
grondilu will try to remember about the env command :) 14:50
sorear colomon: my $x = TreeIter;
colomon sorear: that doesn't work
(error message in a moment) 14:51
mls_ it's triggered by a missing $HOME
14:51 sftp joined
colomon sorear: gist.github.com/1278289 14:51
grondilu indeed: env - HOME=foo /usr/local/bin/perl6 -e ';'
does not crash
mls_ so it's line 30 in src/core/terms.pm 14:53
grondilu so I guess a quick fix would be to setup my HOME in my CGI shebang, right?
mls_ yes, just set it to /
moritz though we really shouldn't die that way 14:54
I'll come up with a fix
sorear colomon: interesting, it looks like an actual bug
mls_ yeah, we just need a // ''
sorear colomon: does it work with my $x = Any; ?
moritz mls_: no, we shouldn't set up that path if neither of HOME/HOMEDRIVE is set
mls_ yes, you're right of course 14:55
colomon sorear: nope, then it doesn't even dispatch
if it will help, I can push my broken code to github
grondilu oh #!HOME=/ /usr/local/bin/perl6 is not reckognized as shebang :-(
moritz just do an
export HOME=/
in you shell wrapper 14:56
or do you want to get rid of it?
maybe
/usr/bin/env HOME=/ perl6
or so
grondilu yes I got rid of the shell wrapper
sorear colomon: github.com/sorear/niecza/issues/66
grondilu thus the shebang pb
moritz tests a patch 14:57
colomon sorear: ah, that does seem to golf it nicely. :)
pushed my version anyway. now for exercise. sorear++
grondilu I guess I should put a sh shebang, and an HOME=/ exec perl6 just after 14:58
or something like that
sorear colomon: oh! TreeIter is a value type.
colomon: try my $x = TreeIter.default; or TreeIter.Zero; 14:59
colomon sorear: oh, right, that's why I was trying to figure out how to m... oo
15:00 sftp_ joined
colomon sorear: TreeIter.default gets me the same dispatch fail as Any did 15:00
15:00 dukeleto- joined
colomon ditto for .Zero 15:00
15:00 sftp left, dukeleto left, imarcusthis left 15:01 dukeleto- is now known as dukeleto
sorear colomon: stupid question do you have a good String argument? 15:01
colomon sorear: not in the least. sorear++ 15:02
I thought I'd be getting a string argument when my callback was called, but instead it's a CLR::Gtk::ToggledArgs
pmurias sorear: hi
15:03 spacebat joined
pmurias sorear: i'll be giving a presentation for uni about niecza ;) 15:03
colomon sorear: well, that makes the callback not crash. it doesn't quite work properly yet, but that's a post-exercise mission. sorear++ 15:04
15:05 baest left, imarcusthis joined, jdhore left, spacebat_ left, Helios left 15:07 Helios joined
moritz pmurias: so what will you tell them? 15:07
15:07 jdhore joined
pmurias moritz: not yet sure 15:08
moritz what's the context of the talk? compiler construction? or "talk about your favorite open source project" ? 15:09
pmurias moritz: "advanced .net techniques" 15:11
moritz: so i suppose it should be a gentle Perl 6 introduction 15:12
15:15 pmurias_ joined
masak did anyone submit a rakudobug about $HOME? 15:16
15:17 pmurias left, pmurias_ is now known as pmurias
pmurias moritz: what thing in Perl 6 would be interesting to non Perl programmers? 15:18
15:24 z-b is now known as zb
sorear leaves 15:24
15:31 grondilu left
masak pmurias: grammars, I think. niecza is good at that. 15:36
pmurias: also, somewhat more nebulously, the sliding scale between extreme expressiveness/brevity and large-scale structure formation that Perl 6 is on. 15:37
"from scripts to applications", that kind of thing.
are they programmers-but-not-Perl6-programmers, or non-Perl-programmers-and-maybe-not-programmers-at-all? :) 15:38
s/6//
15:39 BudiantoWang left
colomon sorear: any tips on dealing with CLR::System::Boolean 15:42
? 15:43
15:45 simcop2387 left, simcop2387 joined, simcop2387 left, simcop2387 joined 15:47 kaare_ joined
pmurias masak: they are fellow computer science students 15:47
masak: so a lot of them don't qualify as programmers
masak CS people, hm. 15:49
sounds like gradual typing could interest them. things like where clauses, subtypes, roles...
TimToady tell 'em it's Lisp done right :)
masak felher: I'm in the midst of backlogging, saw you asking about type synonym declarations.
rakudo: constant Relative := Int; say Relative.WHAT
p6eval rakudo 3fd353: OUTPUT«===SORRY!===␤Malformed constant at line 1, near ":= Int; sa"␤» 15:50
15:50 simcop2387 left
masak rakudo: constant Relative = Int; say Relative.WHAT 15:50
p6eval rakudo 3fd353: OUTPUT«Int()␤»
masak constant := should be allowed, right?
I mean, that's basically what happens, even if I write '=' :)
moritz: thanks for the link with parsing blog posts. will peruse as time permits. 15:51
dalek ecs: f798397 | larry++ | S03-operators.pod:
braino spotted by moritz++
15:53 simcop2387 joined 15:57 KyleHa joined
cognominal I am not sure I understand $node.foo // $node.map: { .».foo } 15:59
if .foo is not defined for a list invocant, how can one call it? 16:00
TimToady well, perhaps it's a .?foo there
and it's also assuming the .map attempts to return an object resembling the type of $node, which plain .map might not actually do 16:01
16:02 KyleHa left, KyleHa1 joined
tadzik pmurias: will there be streaming or so>? 16:03
16:03 KyleHa1 left
pmurias tadzik: no 16:03
cognominal may I have a commit bit on github perl6?
pmurias tadzik: we have so called seminar classes, where everyone who signed up has to give a talk on a given topic 16:04
tadzik: usually on something he has no idea about
tadzik interesting
pmurias most of the seminars end up horrible 16:05
dalek ecs: 694d6ef | larry++ | S03-operators.pod:
clarify unary hyper semantics for cognominal++
 16:06
ingy seen fglock 16:08
aloha fglock was last seen in #perl6 154 days 19 hours ago leaving the channel.
ingy what became of fglock?
pmurias ingy: he can't use the irc channel anymore 16:09
ingy he lost his fingers?
or he was a bad boy?
pmurias got a job at booking?
ingy: heh 16:10
ingy booking disallows irc?
pmurias dunno
TimToady I didn't think slavery was allowed in the Netherlands...
ingy is lost
pmurias ingy: he visits the channel occasionally 16:11
ingy ok
pmurias very rarely
ingy I want to visit his couch
if he and his couch are in Buenos Aires
TimToady pictures fglock as a psychiatrist... 16:12
ingy haha
talk about your childhood, Ingy
TimToady doubtless a retired Nazi...
pmurias ingy: fglock is still hacking on mp6, which is quite similiar to your acmeish stuff 16:13
ingy: a small p6 subset and loads of backends
ingy wow, thx
I had no idea
now it's a must
TimToady now we just have to compile real Perl 6 down to mp6 :) 16:14
pmurias considered that option before
TimToady I will bet there will be...impedance mismatches... 16:15
cognominal ingy, I noted you can't do without your laptop. At your age you should be able to spend time away from that transitional object :)
TimToady cognominal: y'know, some of us can't see the fine print anymore 16:16
ingy TimToady: impedance challenges
TimToady must admit he just got a smartphone though
but it's really hard to type fast on a smartphone
ingy cognominal: I am only tied to $laptop
specific one is not important 16:17
TimToady it's a common noun, not a proper noun :)
ingy I worked had on that during the last year.
although I do like touching a thinkpad 16:18
cognominal ingy, so I am worse than you because I am tied to $laptop.MacBooK :)
ingy I have a macbook still, somewhere. it runs ubuntu
TimToady far be it from me to criticize another (dead) man's cult 16:19
ingy like every other throwaway laptop i have
TimToady: I'll gladly do it for you :)
cognominal a MacBooK being a bif financial investment easily become an emotional one 16:20
*big
16:20 mberends left
ingy I used mac for 10 years. I decided to try not using it for a weekend last PPW, and never used it again 16:21
TimToady
.oO(funny how "okay" means 'yes' but "that's okay" means 'no'...)
ingy TimToady: that's ok
flussence screwed up the rootfs on my good netbook yesterday, been stuck on an eee701 today. feels pretty weird going back to such a tiny keyboard...
TimToady
.oO(except when it doesn't...)
ingy that's okay by me 16:22
TimToady masak: I don't know how you get commutativity out of what S09 says; it's specifically designed to work like English, which is not commutative in that spot 16:36
so use of junctions within a chained comparison is likely a DIHWIDT 16:37
well, a single junction is fine, but two different adjacent terms with junctions is asking for it 16:38
so 1 < $all < 10 is fine 16:41
[Coke] IWBNI if aloha linked to the logs. 16:45
(when reporting "seen")
16:45 GlitchMr left 16:50 daniel-s left
TimToady masak: re macros and ASTs, we don't have to require all the params to be ASTs; anything that's not declared AST could be autoevaluated to a normal value with the AST eval function proposed, and a macro param declared as a thunk could certainly auto-compile to a Callable 16:50
and macros are rare enough that I don't feel like defaulting params to ASTs makes much sense huffmanly speaking 16:51
we could go so far as to say that macros must be explicitly typed 16:52
TimToady waits with bated breath to see which of masak's hot buttons that pushes, the one that likes clean semantics, or the one that dislikes arbitrary restrictions :) 16:54
16:54 plobsing left 16:56 plobsing joined, GlitchMr joined
TimToady the difference between a macro and an immediate sub is that the macro knows it's feeding the parser, and the immediate sub doesn't 16:58
which is maybe not much difference if you're returning an AST, but a profound difference if you're returning a Str 16:59
16:59 sftp_ left
TimToady the 'macro' declarator biases Str to be reparsed, while the 'sub' declarator would presumably work like a constant fold 17:00
17:00 donri joined
TimToady or maybe we call them both 'macro', but distinguish two types of return string, Str vs CodeStr or some such 17:01
then the declarative weight would not be on the 'macro' but on the --> CodeStr
--> DirtyMacroString :) 17:02
we can discourage text macros if we like :)
masak no buttons pushed so far over here.
I'm just glad you're thinking about macros ;) 17:03
TimToady that's because I'm supposed to be thinking about 5 other things instead
masak I must confess I kinda like the 'macro' keyword. I think it's a keeper.
TimToady it means 'immediate', basically
I think everything else can be declaratively controlled with types 17:04
masak and yes, maybe we do need a way to declare sub params as AST-ish, or macro params as non-AST-ish.
we definitely need ways to convert between "values" and ASTs.
also, note that my grant proposal explicitly leaves out text macros. I have no idea where to even start with those, and I'd be glad if someone else decided to tackle them. ;) 17:05
they don't interest me that much, frankly.
TimToady they're pretty trivial, except for the tricky bits :) 17:06
17:06 mberends joined
masak I realize textual macros are a "safety valve", just like BEGIN and eval are, when the language falls short and something radical is needed. 17:06
TimToady they need a mechanism for diverting the parser to look at them, and a mechanism for reporting error messages not inaccurately 17:07
other than that, they're conceptually simple, if unhygienic
masak even clean macros probably need to report an error in a sort of before/after manner. 17:08
17:08 dakkar left, ab5tract left
masak I speculated about that in one of my YAPC::EU gists. 17:08
TimToady for that you can treat them almost as a separate callframe
17:09 sftp joined
TimToady and give a micro traceback 17:09
s/micro/delimited/ 17:10
not much point in giving a traceback of the whole recursive descent when the user is unlikely to be interested...
though that would be an interesting debug option
(for grammars) 17:11
17:11 Trashlord left
masak I don't get how applying a macro can be seen as a separate callframe. 17:11
I meant something like this: macro add($a, $b) { quasi { {{{$a}}} + {{{$b}}} } }; add(40, foo()) 17:12
17:12 Trashlord joined
masak something wrong happens when calling foo(). error shows both '40 + foo()' and 'add(40, foo()' 17:13
colomon sorear, mberends: I've actually got the toggle updating the model now, but the view doesn't change. Seems like I need some sort of redraw command, but I haven't been able to find it yet...
masak TimToady: it will be increasingly important to do that as more layers of macros are applied on some bit of code. 17:15
TimToady: cf C++'s incomprehensible template error messages.
mberends colomon: try $drawingarea.QueueDraw; or $drawingarea.QueueDrawArea(0,0,$windowSizeX,$windowSizeY); # fsvo $drawingarea
TimToady masak: sure, and this probably coincides with my wish to see long, well-formatted error messages dumped to html or some such for later browsing 17:17
colomon mberends: doesn't blow up when I do that, but it doesn't change, either. 17:18
TimToady generally, if we have a place to put such messages, we should just put a short message to the screen that says where the full traceback/parse/rewrite can be viewed
colomon mberends: actually, now I wonder if I set it up incorrectly
masak TimToady: yes, I like the sound of that. no idea about the specifics, though :) 17:19
TimToady we need a concept of STDERRDIR or some such :)
masak TimToady: since it's disruptive, it needs to be pulled of extra well not to make people reject it outright.
colomon mberends: bingo! I needed to say the toggle was active.
TimToady I think people dislike terminal spews enough that just fixing that will be somewhat disruptive 17:20
mberends colomon: you can sprinkle 'say' statements in your code to ... \o/
17:20 aindilis left
TimToady and as soon as you don't have to worry about spew, you could keep a complete cockpit recorder of information, as long as it's well structured 17:20
17:21 kaleem joined
TimToady 'course, the Unixoids will just say, that's what 2>file does 17:22
but that neglects the "well structured" aspect
17:23 Trashlord left
TimToady btw, I believe one of the Apocalypses speculated about recursive macros requiring lazy parsing semantics or some such 17:25
"there will be an AST here, but I don't know what it is yet"
part of the macro's BEGIN semantics are to apply itself recursively when it can 17:26
17:26 Trashlord joined
TimToady oh, the other thing that's different between text macros and AST is that a text macro always stays in the same parse state before reparsing, while an AST macro has to know what the next state is (via term:, infix: etc.) 17:27
but I think that's specced somewhere 17:28
or apo'd
17:36 Trashlord left
TimToady my backlogging turned into far to much frontlogging today... 17:40
*too
sjohnson :3 17:45
masak frontlogg'd! 17:56
18:01 PacoLinux_ left
dalek osystem: 53fd42a | (Timothy Totten)++ | META.list:
Added MIME::Types, used by WWW::App for basic mime-type detection.
 18:04
18:10 supernovus joined
supernovus Any news on grammars/regex progress in nom? Getting the JSON::Tiny and Exemel libraries working again would be awesome, but both rely heavily on grammars. :( 18:13
[Coke] still waiting on pmichaud, I believe. 18:14
felher masak: Thnx for your answer to type-synonyms. :) This is a very nice way to create real synonyms. I wanted to use them in multiple dispatch, though. Like two classes, that are exactly the same, just have different names. :) 18:23
masak felher: so... you want them to be different? 18:27
rakudo: class Relative { has Int $.cargo }; say Relative.new(:cargo(42)).cargo 18:28
p6eval rakudo 3fd353: OUTPUT«42␤»
18:28 risou is now known as risou_awy 18:29 GlitchMr left 18:30 envi_ left 18:31 kaleem left 18:33 pyrimidine joined
masak felher: either composition (above) or subclassing would do what you want. 18:39
rakudo: class Relative is Int {}; multi foo(Int) { say "Int" }; multi foo(Relative) { say "Relative" }; foo(42); foo(Relative) 18:40
p6eval rakudo 3fd353: OUTPUT«Int␤Relative␤»
felher masak: Yeah, i tried that first, with the following Problem: pastebin.com/3mjiK4uE 18:41
masak: how do i get values into that subclass? 18:42
18:42 wamba joined
felher masak: i guess the solution is pretty easy, i'm just blind :) 18:42
masak not sure it's possible, actually. 18:43
felher is very happy he isn't _that_ blind :)
masak rakudo: role Relative {}; multi foo(Int) { say "Int" }; multi foo(Relative) { say "Relative" }; foo(42); foo(42 but Relative) 18:44
p6eval rakudo 3fd353: OUTPUT«Int␤Ambiguous dispatch to multi 'foo'. Ambiguous candidates had signatures:␤:(Int)␤:(Relative)␤␤ in sub foo at /tmp/Lnr6tVGfvO:1␤ in block <anon> at /tmp/Lnr6tVGfvO:1␤ in <anon> at /tmp/Lnr6tVGfvO:1␤»
masak rakudo: role Relative {}; multi foo(Int) { say "Int" }; multi foo(Int $ where Relative) { say "Relative" }; foo(42); foo(42 but Relative)
p6eval rakudo 3fd353: OUTPUT«Int␤Relative␤»
masak felher: that works.
felher masak: hm, nice, that solution works for me. Big thanks :) 18:46
masak++
masak \o/ 18:47
PerlJam Isn't that what subsets are for? 18:49
japhb ingy, ping 18:50
18:50 supernovus left
masak subsets would tie in the MMD. 18:50
hm, maybe not.
but I got the feeling that in felher's example, the types aren't distinct in any other way than that felher say they are. 18:51
that's a typical Value Object pattern to me.
says*
PerlJam masak: they have different names, ergo they're distinct ;) 18:52
masak well, since subtypes basically carve out values from their nominal type, that's not enough.
rakudo: subset FunkyInt of Int; multi foo(Int) {}; multi foo(FunkyInt) {}; foo(42) 18:53
p6eval rakudo 3fd353: ( no output )
PerlJam Is there a manual way to affect who wins in the event of a tie in MMD?
masak rakudo: subset FunkyInt of Int; multi foo(Int) {}; multi foo(FunkyInt) { say "FunkyInt" }; foo(42)
p6eval rakudo 3fd353: OUTPUT«FunkyInt␤»
masak oh right, subsets are narrower. 18:54
so foo(Int) will never get called.
ooh, we could detect such things statically as well! \o/
PerlJam: yes, there's 'is default'.
PerlJam: I would consider using it to be a fairly strong code smell. 18:55
felher masak: that value-object: c2.com/cgi/wiki?ValueObject ?
masak felher: aye.
PerlJam masak: I meant something more ... detailed than than. But you're right .. code smell.
masak felher: Probability is a good example. one could just use a Num. but using Probability makes it clear what kind of value one is dealing with. 18:56
felher: and you could put nice methods on the role too, like .combinedWith(Probability $b) which calculates the product of self and $b 18:57
typically, a domain model has some value objects and some entity objects. 18:58
c2.com calls entity objects "business objects" or "reference objects".
felher masak: yeah, thats (object value) exactly what i wanted. It just had to work in multiple dispatch. I wanted 'one' jump method. Not jump-relative and jump-absolut, because the jump-value would carry the information with it anyway :). 19:01
masak oh, then you can do 'given $value { when Relative { ... }; when Absolute { ... } }' 19:02
if you need to, I mean.
'given/when' is sort of the dual of MMD, by the way ;)
felher masak: yes, if i needed to. :) It's en.wikipedia.org/wiki/Double_dispatch, isn't it? 19:03
masak ummmmm 19:04
felher masak: oh, well, no, it isn't :) i mixed something up :)
masak "Double Dispatch" is one of those concept I've never really learned because the languages I've used so far haven't been brain-damaged enough to need it :P
concepts*
std: multi foo() is default {}; multi foo() is default {} 19:07
p6eval std 580b69a: OUTPUT«ok 00:01 120m␤»
PerlJam they're *all* default.
masak TimToady: is it allowed for more-than-one multi variant in the same scope to have 'is default'?
masak guesses "yes"
either way it doesn't affect the compiler's ability to detect ties at compile time. 19:08
felher masak: :). Well, in the end i wanted to be able to write my @jumps = (30 but relative, 20 but absolute, ... ); for @jumps { $m.jump($_) }. Which only works well with multiple dispatch :) (or given/when - but thats not 'well' )
PerlJam my @jumps = (30 but meters, 20 but feet, ... ); # Would you do that too? 19:09
masak nice! 19:11
rakudo: role Length {}; role Meters is Length {}; role Feet is Length {}; my @jumps = 30 but Meters, 20 but Feet; multi jump($x where Meters) { say "$x meters" }; multi jump($x where Feet) { say "$x feet" }; jump($_) for @jumps 19:13
p6eval rakudo 3fd353: OUTPUT«P6opaque only supports type changes where the MRO of the original type is a suffix of the MRO of the new type␤ in mixin at src/gen/Metamodel.pm:918␤ in sub infix:<but> at src/gen/CORE.setting:7376␤ in block <anon> at /tmp/PlYluHQIBN:1␤ in <anon> at /tmp/PlYluHQ…
masak hm.
felher PerlJam: in principle. Why not?
masak rakudo: role Length {}; role Meters does Length {}; role Feet does Length {}; my @jumps = 30 but Meters, 20 but Feet; multi jump($x where Meters) { say "$x meters" }; multi jump($x where Feet) { say "$x feet" }; jump($_) for @jumps
p6eval rakudo 3fd353: OUTPUT«30 meters␤20 feet␤»
masak \o/
felher, PerlJam: you're both doing superstitious parentheses on array initialization, by the way. I'm guessing you're both good at Perl 5. ;) 19:16
PerlJam aye ... I'm coding in perl 5 right now ... hard to context switch. :) 19:17
felher masak: thanks for pointing that out. I forget about it all the time :)
masak PerlJam: at $client, I do a lot of Perl 5. I'm considering putting a piggy bank on the desk and put a krona in for every p6ism I don't catch before compiling my program. ;) 19:18
PerlJam: had I done that from the start in Feb, I'd probably be up to about 30 SEK now.
there's probably another 25 SEK or so in p6isms that I caught before compilation by reviewing my own diffs ;) 19:19
but given how much code I write in both languages, I don't consider 55 SEK to be much at all. I usually find it quite easy to context-switch. 19:22
19:22 benabik left, benabik joined
PerlJam sometimes it's easy, but when my brain is working p5 problems, it doesn't quickly generate p6isms. 19:23
masak we probably come at it from different default directions. 19:24
I'm a Perl 6 developer, occasionally working in Perl 5.
PerlJam aye, and I'm the other way around basically :)
masak my Perl 5 is often some kind of pidgin Perl 6. 19:25
19:25 zb left, zb joined
masak I'm guessing it's loads easier to automatically translate to Perl 6 than most other people's Perl 5 code. 19:25
flussence Mu.pm:132 is causing me much grief right now :( 19:28
flussence changes it to a die locally just to try and get a stacktrace 19:30
ingy hi japhb 19:36
masak oh right. macros. 19:50
20:00 benabik_ joined 20:04 Alias_ left, Alias_ joined 20:06 wooden joined, wooden left, wooden joined, alvis left 20:14 soh_cah_toa joined 20:19 alvis joined
moritz oh hai 20:26
masak moritz! \o/ 20:27
20:28 espadrine joined 20:47 kaare_ left
colomon sorear: any easy way to get today's date in niecza? 20:48
masak niecza: say time 20:53
p6eval niecza v10-40-ga57ed3d: OUTPUT«1318366385.55574␤»
masak I declare it possible, if not easy ;)
colomon masak++
I could actually work with that, I think, in a pinch
masak well, DateTime in Rakudo should be a good guide.
colomon: maybe you'll end up implementing DateTime and Date in Niecza :) 20:54
colomon I'd really rather not. I'm just looking for a simple way to get the calendar day. :)
masak 's interest is piqued 20:55
colomon now that I have 90% of a UI for my current project, getting the date is the last major conceptual hurdle
masak rakudo: say DateTime.now.day-of-week
p6eval rakudo 3fd353: OUTPUT«2␤»
masak rakudo: say (DateTime.now.get-daycount + 2) % 7 + 1 20:56
p6eval rakudo 3fd353: OUTPUT«2␤»
colomon masak: not day of the week, I mean like 2011-10-06 20:57
masak oh.
well, have a look at ymd-from-daycount in Temporal.pm 20:58
20:58 skangas left
colomon it's mostly for temporal sorting, so the time above would work, but it wouldn't be very human user friendly 20:58
colomon dances, as he just got The Dragons of Babel in the mail as a late birthday present 20:59
im2ee Good night! Dobranoc o/
masak 'branoc, im2ee!
diakopter colomon: ./run/Niecza.exe -e 'say CLR::("System.DateTime").Now.Day;'
masak colomon: happy belated birthday!
diakopter colomon: msdn.microsoft.com/en-us/library/sy...etime.aspx for the other struct members 21:00
masak diakopter++ # pragmatism
21:00 im2ee left
mberends perl-sharp strikes again! 21:00
colomon diakopter++ # that'll do, for sure 21:01
[Coke] needs a webdev minion, if anyone is job hunting. ;) 21:03
diakopter perlesque: my $now = System::DateTime.Now; say $now.Day; say $now.Month; say $now.Year
p6eval perlesque: OUTPUT«11␤10␤2011␤»
diakopter colomon: you can experiment with a similar CLR interop using perlesque
perlesque isn't locked down in p6eval though :) 21:04
s/though// 21:05
colomon niecza: say 5.fmt("%02d") 21:07
p6eval niecza v10-40-ga57ed3d: OUTPUT«␤Unhandled Exception: Unable to resolve method fmt in class Int␤ at /tmp/aGGkGlxJXv line 1 (MAIN mainline @ 1) ␤ at /home/p6eval/niecza/lib/CORE.setting line 2133 (CORE C1024_ANON @ 2) ␤ at /home/p6eval/niecza/lib/CORE.setting line 2134 (CORE module-COR…
colomon oh
niecza: say sprintf("%02d", 5)
p6eval niecza v10-40-ga57ed3d: OUTPUT«05␤»
21:09 packetknife joined
ingy packetknife is a great nick 21:19
masak especially if you're Swiss. 21:20
21:22 s1n1 left 21:25 PacoLinux_ joined 21:26 localhost joined 21:28 benabik_ left 21:29 ponbiki left, ponbiki joined
packetknife ingy: Thanks.. no, sorry, not Swiss though masak. ;-) 21:30
masak dang :)
dalek kudo/macros: 991ade7 | masak++ | src/Perl6/Actions.pm:
decontainerize quasi AST before attr lookup

Many nqp ops will decontainerize for you, but getattr is not one of them. Occasionally the quasiquote would be hidden inside a container layer and the getattr would fail. This is now fixed.		 21:33
masak I invite people to play around with the macros branch and give feedback. 21:34
the coding bit of D1 is now complete, to the best of my knowledge. 21:36
next up is a bit of spectest coverage and a blog post or two.
diakopter masak: are there spectests for macros yet? 21:37
masak not by me. 21:39
there are a few, but they're old and not very decisive :)
dalek gs.hs/bigexp: 2866550 | (Ingy dot Net)++ | Pugs/src/Pugs/Prim/Numeric.hs:
Make this per au instuctions:

00:11 <au> I see. it can be fixed in Pugs/Prim/Numeric.hs 00:11 <au> in op2Exp to do a (^) instead of a (^^) if the right hand side is non-negative
masak I have quite a few ideas on what to test.
whoa, a Pugs commit :)
ingy :D
masak ingy++ 21:40
ingy masak++
21:40 y3llow_ joined, y3llow left, pothos left, pothos_ joined
masak thanks, though I have no idea why you're giving me karma... :P 21:41
21:41 y3llow_ is now known as y3llow
ingy masak: let's try to get pugs testing against current stuff 21:41
masak: for getting me into pugs
pugs++
21:42 pothos_ is now known as pothos
ingy masak: I got make and make install working 21:42
so I think we are set to begin PugsClub™
masak oh wow 21:43
I haven't gotten that far yet.
ingy moritz: can you make the irc bot use current pugs master
masak ingy: were the instructions au pointed to enough to get make working?
ingy I worked with au to get it going 21:44
masak PugsClub™ \o/
ingy pull the latest from github.com/perl6/Pugs.hs
masak oki
ingy on debian is trivial
like 2 or 3 steps
masak I'm on Ubuntu, so it should be trivial-ish. 21:45
ingy me too. it is trivial
i'm on natty
masak ok, I checked out Pugs.hs 21:46
what's next? 21:47
21:47 huf left 21:49 benabik left, localhost left
dalek gs.hs: 565cdac | (Ingy dot Net)++ | HACKING:
Add hacking instructions.
 21:52
ingy masak: ^^
:D
masak \o/
mberends, [Coke]: ^^
ingy: hm, it won't actually be 'cd Pugs' unless you rename the Pugs.hs directory... 21:53
ingy where's my umlöt? :\
masak: no. inside Pugs.hs
Pugs.hs/Pugs
masak oh. 21:54
ingy I assumed you were in there. but yeah that's confusing
I might add a top level Makefile at some point
sudo make -C Pugs install 21:55
mberends gets Pugscited
21:55 Mowah_ left
masak this all coincides with me going to bed. I think I'll have to do the INSTALL stuff tomorrow. 21:55
yeah. $dayjob. it interferes with my work. :P 21:56
anyway, ingy++!
colomon hmmm.... is it open $filename, :w, :a ? 21:58
masak :a implies :w, methinks. 22:01
but I've yet to see an implementation that supports :a
well, this is my stop. 'night, #perl6. 22:04
I hope to get to blogging and Pugs tomorrow. 22:05
22:09 Woodi left 22:11 molaf left, Woodi joined 22:12 dual left 22:13 huf joined 22:20 pmurias left
colomon sorear: ping? 22:24
22:26 packetknife left, replore left
colomon sorear: I need to modify twriter_open to call the StreamWriter constructor which takes a string and a bool, so that I can properly handle open :a. I'm really not clear on the syntax to do that, and I've got to run to preschool open house in a moment, so massive experimenting is not practical. Danke. 22:27
diakopter perlesque: my $wr = System::IO::StreamWriter.new("hihi", true); $wr.Write("blah"); $wr.Close(); # colomon I know this doesn't help you but I'm just experimenting 22:33
p6eval perlesque: ( no output )
tadzik map -> $n { ((-1) ** $n * $x ** (2 * $n)) / [*] 1..(2 * $n) }, 0..* # taylor series for cosine 22:49
gotta love infinite series 22:50
22:56 aloha left
tadzik down to 42 characters: map {(-1)**$_*$x**(2*$_)/[*]1..2*$_},0..* 22:56
22:58 aloha joined
tadzik shame that map (-1)****$x**(2**)/[*]1..(2 * *), 0..* caused a parsefail :) 22:59
23:04 ZapZ_ joined 23:07 soh_cah_toa left 23:10 whiteknight joined
tadzik perl6: my @a = 1, 2, 3, 4; @a».++ 23:11
p6eval pugs: OUTPUT«decodeUTF8': bad data: '\187'␤»
..rakudo 3fd353: OUTPUT«Nominal type check failed for parameter '$op'; expected Any but got Mu instead␤ in sub METAOP_HYPER_POSTFIX at src/gen/CORE.setting:7657␤ in block <anon> at /tmp/NIbBlYzLZh:1␤ in <anon> at /tmp/NIbBlYzLZh:1␤»
..niecza v10-40-ga57ed3d: ( no output )
tadzik I suppose rakudo's error message could've been better 23:12
23:20 supernovus joined
supernovus Just a quick question, is there a way to determine if a scalar contains an object instance or a type? My basic desire is method load ($something) { my $object = $something; if is_type($something) { $object = $type.new; } ... # so something with $object } 23:23
tadzik .defined?
nom: $_ = ($^n.defined ?? $^n !! $^n.new); say .(5), .(Int) 23:25
p6eval nom 3fd353: OUTPUT«===SORRY!===␤Cannot use placeholder parameter $^n in the mainline at line 1, near ".defined ?"␤»
tadzik nom: $_ = { $^n.defined ?? $^n !! $^n.new }; say .(5), .(Int)
p6eval nom 3fd353: OUTPUT«50␤»
supernovus tadzik: Thanks! That will work great :-) 23:28
tadzik I'm glad :)
23:31 benabik joined
supernovus Although I have to use ng to do it, I'm refactoring my Flower library into a much cleaner and more extendable codebase. So instead of being just a TAL/METAL implementation, it will be an XML Application Language Toolkit (with a TAL/METAL implementation included as an example XML Application Language...) 23:32
23:45 yves_ left
tadzik huh 23:46
23:46 wamba left

Powered by Cro and the Raku® Programming Language.

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

Thank you!