»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.pugscode.org/ | UTF-8 is our friend! | Rakudo Star Released!
Set by moderator on 29 July 2010.
Ken-Guru i don't like installs ... 00:00
:D
jnthn Ken-Guru: huh...
Well, uninstall is just in Add/Remove programs.
pmichaud exodist: if there is one, it'd be on perl6.org somewhere. 00:01
gfldex exodist: www.perlfoundation.org/perl6/index....dex_tablet
sorear Why is it called Add/Remove programs if removing is the more important thing it does?
jnthn sorear: Why do I have to click the "Start" button to shut down? :-)
Ken-Guru 2(jnthn2): jupp. the linux way its ok. but having only used windows all life long its annoying ;-) ... you know this damn little thing called registry. uargh.... 00:02
gfldex little? :->
Ken-Guru japp, thats the problem.
exodist gfldex, well there goes my plan to make one :-P thanks 00:03
gfldex exodist: you might want to extend it
Ken-Guru never the less i'm a little win adict. don't ask me why, i don't know. 00:04
pmichaud exodist: we can always use more guides or improvements to existing ones.
exodist What I really want to write is a perl5 equivilent in perl6 topic guide, example topic: *Other::Package = sub { ... }, then how to do it in perl 6. 00:07
sorear exodist: that's the hard way. The easy way, in perl 5 and perl 6, is our sub Other::Package { ... } 00:09
the hard way in p6 is &Other::Package := sub { ... } btw
exodist sorear, our sub Other::Package { ... }, does that really work in perl5? I have never seen it done.. 00:10
sorear, that specific topic is because I want to write a module that exports subs from other modules into the module that usus it, like a bunile if you will. 00:11
use A; # like use B; use C; use D;
*budle
*bundle damnit 00:12
*and uses
masak listen up, #perl6 -- I haz a challenge for you. 00:18
I just pushed github.com/masak/yapsi/blob/master/...ce/2010.08
waste sorear: thank you
masak please help fill in the blanks for tomorrow's realease. it's not hard. anyone can help. 00:19
if you're unsure about the desired tone, look at previous announcements.
jnthn masak: lol!!
masak large amounts of silliness are encouraged. thank you.
jnthn I never suspected there was a template!
masak there wasn't until just now :)
00:20 Axord joined
masak I reserve the right to push things around a bit to keep people like jnthn on edge :P 00:20
jnthn Damm...only slightly inappropriate? I can only think of wildly inappropriate things. 00:21
dalek psi: 4051102 | masak++ | doc/announce/2010.08:
[doc/announce/2010.08] added

can help!		 00:22
masak it should be somewhere between "nodnod, he's happy about announcing stuff" and "eeh, can you really say that... in a release announcement?" :)
jnthn ooh, I wonder what satisfaction will be at those URLs! 00:27
masak nono, it's "Get Satisfied!". Mr Satisfied has escaped from the local prison, you see.
jnthn Oh noes! 00:28
dalek psi: 2c149f4 | masak++ | doc/announce/2010.08:
[doc/announce/2010.08] fixed petty typo
pmichaud masak: are you giving a talk on yapsi? 00:29
waste rakudo: $:a; say "alive";
p6eval rakudo 9b6189: OUTPUT«Lexical '$a' not found␤ in main program body at line 2:/tmp/jfcE3Mi3Tt␤»
jnthn At YAPC? 00:30
masak pmichaud: no. maybe I should lightning-mention it.
pmichaud yes, of course :)
I was thinking it deserves a lightning talk :)
jnthn masak: I think you should dedicate these release to Europe.
masak (because it's so blazingly fast)
jnthn *this 00:31
masak jnthn: "this release is dedicated to Europe, because it was written, well, inside it"
pmichaud I really should see if I can come up with a good hack talk for rakudo/nqp
masak jnthn: "you're free to call this release 'Europe' if you want. of course, you're free to call this release 'Johann Sebastian Bach' too, if that floats your boat" 00:32
pmichaud can I call it "The Village People"?
jnthn Not until it can run...the ASCII dance code...
pmichaud "Culture Club"?
masak you can call it "The Village People With a Pope Hat"...
pmichaud "Wham!"? 00:33
"The release formerly known as Yapsi"?
masak "Wham Sebastian Bach"
jnthn: just for you, I can introduce a special SIC opcode that does the ASCII dance... :P 00:34
jnthn "Run-time DMC"
avuserow We can ASCII Dance if we want to. 00:35
masak It's my party and I ASCII Dance if I want to. 00:36
jnthn: "this release is dedicated to Europa, the famous Jupiter moon, for its unyielding support of Perl 6" 00:37
see you at YAPC::Europa! 00:41
masak waves a translucent tentacle 00:42
masak decides to regenerate for a while
'night, #perl6
00:51 linz joined
kthakore moritz_: ping? 00:53
01:01 Axord joined 01:17 xinming joined
sorear exodist: I don't think rakudo implements this yet, but you shound use be able to do use Other::Module :EXPORT<:ALL>; 01:17
01:20 LaVolta joined
jnthn nqp: loop { } 01:22
p6eval nqp: OUTPUT«Confused at line 1, near "loop { }"␤current instr.: 'parrot;HLL;Grammar;panic' pc 552 (src/cheats/hll-grammar.pir:205)␤»
sorear I guess the binder overhaul needs to be a priority. :( 01:32
01:44 hugme joined 01:47 orafu joined 01:57 flup joined 02:09 edgar joined
edgar hi, is anybody still awake? 02:09
sorear "still"?
jnthn Depends on your definition of awake. :-)
sorear this is a global channel.
edgar in the netherlands it is 4 o'clock in the morning
but you are probably in the VS. 02:10
jnthn Yay stroopwafelland!
sorear I'm in the US, jnthn is in Iceland
jnthn sorear: Sweden!
:-)
sorear well you move every two weeks
jnthn Slight exaggeration. :P
edgar Yeah, we have them here. The same timezone in Sweden probably.
jnthn But you ahve a point. :-)
edgar: Yes. :-) 02:11
edgar: I visit NL now and then. :-)
edgar It is a good country Sweden. Not too warm. I will move to there at one day. It is 27 degrees here. 02:12
02:12 TiMBuS joined
jnthn Uff, that's too hot. 02:12
edgar I had a question about perl6. I try to implement some monads. I think Roles are suitable to replace typeclasses. But how do I access a simple attribute? 02:13
It is too hot for this place. But we have a heat wave at the moment. 02:14
jnthn As in, if you declare role Foo { has $!bar }, how do get at $!bar?
*to
Or maybe show me an example of what you're trying to do and how it's not working...
edgar No, more like if you declare class Just is Maybe (which has role Monad) { has $.a ... } 02:15
Maybe is a role too. Just is a class, and Nothing is a class 02:16
jnthn If you're adding a role to a class, you'd normally use "does" rather than "is"
"is" is for inheritance rather than role composition. 02:17
flup A thanks. I didn't know that.
edgar The webclient didn't work very well, so I switched to irssi. But it seems to work. 02:18
jnthn You can do "is" with a role but really it means "I want to generate a class that does the role and inherit from that class". So it gives you rather different semantics (especially with regard to attribute visibility). 02:19
02:23 Italian_Plumber joined
flup That sounds strange: I want to generate a class that does the role and inherit from that class. It is certainly not what I mean. Is there a good book on perl6 yet? 02:24
jnthn flup: Various of us are working on one, at github.com/perl6/book/ 02:25
flup: The roles coverage ain't great yet though. :-(
flup: svn.pugscode.org/pugs/docs/Perl6/Sp...-types.pod is the spec on roles 02:26
It's a little terse in places, but has many of the details on how stuff works.
flup Oh that is ok. I will figure it out. I am reading the apocolyps on that one.
tnx. I am from haskell, are Roles a bit similiar to type classes? 02:27
jnthn I'm not familiar enough with Haskell to really know, sorry. 02:28
I've heard people talk about them as having some similarities.
But I'm not sure how deep they go.
flup They are. A typeclass is sort of a interface. 02:29
jnthn A role is sort of an interface too, but can carry a default implementation. 02:30
But if those are all stubs, then they become requirements, so it degenerates to an interface (in the Java-y sense of interface)
flup A typeclass can have a default implementation too, so I think they are quite similiar. 02:31
One of the first implementations of perl6 was written in haskell, I think they used some of the ideas there 02:32
But haskell doesn't have the hyper oprators 02:33
I like those
Too drunk to program, time too sleep. Thanks for the links. I will be able to produce something now. 02:34
goodnight 02:35
cxreg hvae you considered reading about Moose?
flup Moose>
?
cxreg there's more docs on that, and the concepts are very similar
Moose is a project that implemented a good bit of perl 6's object model on perl 5
jnthn flup: Goodnight, and do return and let us know how you get on. :-)
flup: Oh, and hope you were drinking something good. :P
flup I will look at it, i am using perl5 at work sometimes. So it maybe useful. I will post something, if I get somewhere. Lol, just beer. But not the alcoholic type of beer. So it was good. 02:37
jnthn :-) 02:38
flup bye, and have a nice night/day, where ever you are. 02:39
jnthn night o/ 02:40
03:00 mmmpork joined
jnthn -> sleep 03:11
03:26 isBEKaml joined
isBEKaml hi #perl6! :) 03:26
whee hola 03:27
03:30 Guest46026 joined
sorear Hello! 03:33
isBEKaml (backlogging) offerkaye was probably confused by my message. I merely pointed him to another location for u4x docs (link is not shown on perl6.org, though) 03:34
03:51 araujo_ joined 03:52 jaldhar joined
frooh hey guys, for some reason a coderef is saying I passed it too many arguments; 03:52
I'm passing it a single string
but it says I passed it *two* things
I'm not even really sure how to debug this 03:53
03:53 araujo joined
frooh I'll nopaste the code to see if anyone has any ideas 03:53
03:53 snarkyboojum joined, snarkyboojum left
frooh paste.scsys.co.uk/47435 03:54
03:55 snarkyboojum joined 03:59 snarkyboojum joined 04:00 gfx joined 04:02 snarkyboojum left, snarkyboojum joined 04:05 snarkyboojum left
sorear frooh: the error is occuring at $.coderef("...") 04:05
frooh: this is short for self.coderef("...")
04:06 snarkyboojum joined
sorear frooh: which is short for (self.^!GET-METHOD("coderef"))(self, "...") 04:06
frooh: the coderef method only wants one argument. you passed it two.
frooh: the correct code is $.coderef.("...")
TiMBuS which causes another problem 04:07
invoke() not implemented in class 'Any'
frooh huh
ok then :-)
and yeah
TiMBuS: got the next error
sooo... is there soemthing else I can do? 04:08
TiMBuS $.coderef.WHAT is Any. I have no idea
frooh star: class Foo { has Code $.bar; method callbar { $.bar.() } }; my $x = Foo.new( bar => sub { say "lolz" } ); $x.callbar; 04:10
p6eval star 2010.07: OUTPUT«lolz␤»
frooh fjdkls;a
TiMBuS :|
frooh star: class Foo { has Code $.bar; method callbar { $.bar.() } }; my $x = Foo.new( bar => sub { say "$^a" } ); $x.callbar('shucks howdy');
p6eval star 2010.07: OUTPUT«Too many positional parameters passed; got 2 but expected 1␤ in 'Foo::callbar' at line 22:/tmp/PTbpFFa8y_␤ in main program body at line 22:/tmp/PTbpFFa8y_␤» 04:11
frooh star: class Foo { has Code $.bar; method callbar($x) { $.bar.($x) } }; my $x = Foo.new( bar => sub { say "$^a" } ); $x.callbar('shucks howdy');
p6eval star 2010.07: OUTPUT«shucks howdy␤»
04:11 masonkramer joined
frooh well then. 04:11
and I can't seem to be able to .perl.say coderefs
and I can't seem to be able to .perl.say coderefs 04:12
woops, sorry
waste Any is not a coderef; 04:16
frooh waste: agreed. but sub {...} should be 04:17
and that's what I passed in
waste the assignment does not work for some reason
TiMBuS passing in anything makes it an Any
setting it to a sub in BUILD does not set it
frooh so I need to set it... after construction?
it worked above in the chat right?
TiMBuS i guess so 04:18
strange bug
frooh now it says it's readonly
did I unknowingly make it ro?
TiMBuS i think they're ro by default 04:19
frooh well then
bizarre
that fixes it indeed
TiMBuS yep yep 04:20
the bug might be triggered by your add_method use..
frooh I wonder.. 04:21
ok 04:22
so it looks like it's my BUILD
when I leave coderef out of the signature, it doesn't get passed in
I thought that if you left something out it would automaticaly handle it
TiMBuS but i set $.coderef inside BUILD and it still didnt work 04:25
plus the default should still be set?
frooh TiMBuS: no, I set it and it wroks 04:27
well, yeah, something's wrong there for sure
but if I add :$coderef to BUILD and do $.coderef = $coderef it works 04:28
waste rakudo: class X { has Code &.c = sub (*@_) { @_ }; submethod BUILD () {nextsame} }; say X.new.c.(1)
p6eval rakudo 9b6189: OUTPUT«1␤»
frooh so I neext to nextsame in build?
TiMBuS .. oohhh 04:29
waste for the automatic behavior
frooh hm.
I'm too used to Moose I guess 04:30
so now I am getting an arror about .is-debug not being a method on invocant Str
but when I do .perl.say to the object 04:31
it's clearly not a Str
04:34 araujo joined 04:39 araujo joined
frooh dangit 04:39
invoke not implemented on resizable pmcarray
*facepalm* 04:41
frooh wonders why rakudo thinks that his object is a Str 04:43
04:44 plobsing joined
frooh man, I don't know what I changed, but now it works 04:47
it's like those times when your pc won't work and you take it apart and put it back together and it works again
waste for the record; S12:"Whether you write your own BUILD or not, at the end of the BUILD, any default attribute values are implicitly copied into any attributes that haven't otherwise been initialized." 04:53
04:54 adhoc left
frooh so it *should* have worked then 04:56
waste frooh: i agree
frooh I thought I read that
04:57 alester joined
alester My rakudo won't build and I am sad. 04:58
the make wants to "cd rakudo" and there is no rakudo dir 04:59
05:20 araujo joined 05:25 wooden joined
sorear Given: sub foo($x, $y) { ... } 05:37
Is this legal? : foo(:y(42), 24)
pmichaud sorear: yes, it's legal. 05:43
star: sub foo($x, $y) { say "$x, $y"; }; foo(:y(42), 24);
p6eval star 2010.07: OUTPUT«24, 42␤»
05:52 plol joined, snarkyboojum joined 05:53 kevinspacey joined 05:54 snarkyboojum left 05:55 sawyer_ joined
sorear star: sub foo($x, $y) { say "$x, $y"; }; foo(:y(42), 24, 12); 05:55
p6eval star 2010.07: OUTPUT«Too many positional parameters passed; got 2 but expected 2␤ in 'foo' at line 22:/tmp/y1CyhO3Lur␤ in main program body at line 22:/tmp/y1CyhO3Lur␤»
frooh star: sub foo($x, $y, $z) { say $z }; foo(1,2,3); my @bar = <a b c>; foo(@bar);
isBEKaml rakudo: my $num = 2000; my $foo = ($num ~~ /^(\\d{2}+)/); $foo.say;
sorear that is a lovely error message. 05:56
p6eval star 2010.07: OUTPUT«3␤Not enough positional parameters passed; got 1 but expected 3␤ in 'foo' at line 22:/tmp/VNveUUMpQJ␤ in main program body at line 22:/tmp/VNveUUMpQJ␤»
rakudo 9b6189: ( no output )
frooh star: sub foo($x, $y, $z) { say $z }; my @bar = <a b c>; foo(*@bar);
p6eval star 2010.07: OUTPUT«===SORRY!===␤Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 22␤»
frooh how do I make @bar "explode" (as a rubyist would say) into three params?
sorear isBEKaml: \\d{2}+ is a pretty useless pattern
frooh: |@bar
isBEKaml gee, that went into a loop in my machine. :D
frooh star: sub foo($x, $y, $z) { say $z }; my @bar = <a b c>; foo(|@bar); 05:57
p6eval star 2010.07: OUTPUT«c␤»
isBEKaml sorear: I just wanted to see what happens when I throw it out to rakudo. :)
frooh excallent
excellent*
sorear isBEKaml: {2} is an embedded block, it matches the null string while evaluating 2 in void context
05:57 snarkyboojum joined
sorear isBEKaml: {2}+ matches the null string infinitely many times 05:57
you know what 05:58
isBEKaml sorear: right, yes.
sorear this would be trivial to add a warning for
frooh thinks that this module is going to be less useful than it was in perl 5, due to the complexity of subroutine signatures
sorear does so
isBEKaml sorear++ 05:59
exactly what I wanted to see.
06:00 snarkyboojum left
frooh star: say 1 if <foo bar baz> ~~ <foo bar biff>; say 2 if <foo bar baz> ~~ <foo bar baz>; 06:01
p6eval star 2010.07: OUTPUT«2␤»
frooh is there a more succinct way to do: 06:02
oh wait
star: $f = sub { say @_[1] }; $f.(1,2,3); 06:03
p6eval star 2010.07: OUTPUT«===SORRY!===␤Symbol '$f' not predeclared in <anonymous> (/tmp/bbMCfz3_Si:22)␤»
frooh star: my $f = sub { say @_[1] }; $f.(1,2,3);
p6eval star 2010.07: OUTPUT«2␤»
frooh perfect
06:04 Khisanth joined 06:09 snarkyboojum joined
frooh star: my @foo = (sub { say wantarray })(); 06:09
p6eval star 2010.07: OUTPUT«Could not find sub &wantarray␤ in <anon> at line 22:/tmp/YyBEPB9nms␤ in main program body at line 22:/tmp/YyBEPB9nms␤»
frooh is there a wantarray distinction in perl 6? 06:10
sorear no 06:11
in Perl 6, functions are not context sensitive; the values themselves are
frooh interesting.. 06:14
but that means that *somewhere* down the line it has to get decided 06:15
isBEKaml frooh: you can have a look at S08
frooh pulls that up
06:15 finanalyst joined, Alias_ joined
frooh needs to just re-read all the Synopses 06:17
finanalyst if I create a class XX, is there a way to associate a constant (eg 'a') with that class? eg., XX('a') or XX::'a', though neither of these seem to be right. 06:18
06:19 kevinspacey left
sorear it's not working... 06:21
finanalyst or say I want a constant (eg. 1) to be explicitly Rat rather than Int?
sorear <1/1>
I don't think there's a general mechanism
1.Rat may work
pmichaud eventually, Rat(1) 06:23
1.Rat ought to work
star: say 1.Rat.perl
p6eval star 2010.07: OUTPUT«1/1␤»
pmichaud 1.0 also produces a Rat, iirc 06:24
star: say 1.0.WHAT
p6eval star 2010.07: OUTPUT«Rat()␤»
melte why is it called Rat
LaVolta rational? 06:26
pmichaud rational, yes.
melte ...that would make sense
pmichaud and trying to preserve the 3-character names for many of the scalar builtins 06:27
i.e., Any Int Str Num etc.
06:37 goksie joined
finanalyst can this be extended to subtypes, eg. odd integers? 06:39
06:39 snearch joined
avuserow star: say 1.can(0) 06:46
p6eval star 2010.07: OUTPUT«␤»
06:46 Guest23195 joined 07:01 goksie joined 07:02 drrho joined 07:07 redicaps joined, goksie joined 07:21 Su-Shee joined 07:22 snarkyboojum left
isBEKaml is there an equivalent of rounding to Int here? 07:22
rakudo: (82/4).say 07:23
p6eval rakudo 9b6189: OUTPUT«20.5␤»
isBEKaml rakudo: (82/4).Int.say
07:23 snarkyboojum joined
p6eval rakudo 9b6189: OUTPUT«20␤» 07:23
sorear isBEKaml: your warning is in. 07:27
std: /x{2}+/ # I wonder if p6eval has updated yet
p6eval std 31883: OUTPUT«ok 00:01 115m␤»
isBEKaml sorear: great! p6eval in the middle of build? 07:31
rakudo: (200 ~~ /^\\d{2}+/).say 07:32
p6eval rakudo 9b6189: ( no output ) 07:33
isBEKaml std: /\\d{2}+/
p6eval std 31883: OUTPUT«ok 00:01 115m␤»
sorear isBEKaml: p6eval doesn't actually pay attention to pushes. It rebuilds once an hour regardless of activity.
isBEKaml sorear: ok.
07:35 jfried joined 07:36 Ross joined
isBEKaml Can anyone tell me how I can speed this up ? gist.github.com/503068 07:38
sorear lose the my Int 07:43
type constraints require a run-time type check 07:44
on second thought, so does my == my Any
nevermind
lines 44 and 46 should use := to avoid duplicating the hash 07:45
56 is quite ugly, Just do it incrementally 07:46
isBEKaml I'm initialising the hash everytime in the block
incrementally as in?
sorear $count++ if $days-so-far %% 7; $days-so-far += $days-this-month; 07:47
07:48 wamba joined
isBEKaml we are just getting the first of every month in a given year. so I don't think the second statement is necessary. 07:49
yeah, I could do include the first one. :)
07:50 finanalyst left
sorear isBEKaml: if you do the second one, then almost all of the rest of your code becomes unneccessary. 07:51
isBEKaml sorear: oh, boy... I was simply replicating the algorithm shown here: en.wikipedia.org/wiki/Calculating_t...f_the_week 07:53
:/
sorear: I think you're doing it based on every 7th day. in 100 years, there will be around 25 leap years. It doesn't seem to factor leap years. 07:54
sorear: if Sunday is taken to be 0, then every 7th day is a Sunday. On that cound, we could simply do 1200/7 and strike gold. :) 07:55
sorear: the above is because, this is exactly in the same century. if the 100 years were in different centuries, like say: 1905 - 2004, this would not work. 07:56
07:56 kjeldahl joined, kjeldahl_ joined 07:57 Mowah joined
sorear pastie.org/1069380 is my solution 07:58
outputs 170 after ~30 seconds
isBEKaml sorear: mine shows 171 which is taken as the right answer. 07:59
sorear 18 seconds by time
isBEKaml Infix !% is deprecated in favor of infix %% at line 9, near " 7) && ($y" 08:01
this is what I get here when I run yours..
sorear I'm running a rakudo from before the spec change 08:02
isBEKaml oh..
08:04 molaf joined 08:06 tommyd joined
isBEKaml std: /\\d{2}+/ 08:09
p6eval std 31884: OUTPUT«[31m===[0mSORRY![31m===[0m␤Unsupported use of {2} as general quantifier; in Perl 6 please use X**2 at /tmp/nNrb9w8G7c line 1:␤------> [32m/\\d{2}[33m⏏[31m+/[0m␤Check failed␤FAILED 00:01 115m␤»
isBEKaml sorear++ :)
08:12 [1]Casan joined 08:15 avuserow joined 08:20 wamba left 08:21 wamba joined, wamba left, isBEKaml joined 08:24 wamba joined 08:28 envi^home joined
isBEKaml sorear: I found an off-by-one error in your code: my $sofar = 0; # a Monday in Jan 1900 08:28
this works in producing 171 which is taken as the right answer on projecteuler.
sorear: thank you and I really need to go back to books to prevent myself from writing bloated code. :) 08:31
08:32 Tom__ joined 08:35 maasj joined 08:37 maasj left, masak joined
masak oh hai, #perl6! 08:37
isBEKaml masak! :) 08:38
masak: aren't you at yapc yet?
masak isBEKaml: will be flying down there tomorrow. 08:39
isBEKaml: I'm only at Yapsi so far :P
isBEKaml masak: xD. So how's that coming through?
masak well, I once again extend my challenge from yesterday; help me fill in the release announcement for today \\o/ 08:40
isBEKaml I just found that I can't write code. :| 08:41
masak github.com/masak/yapsi/blob/master/...ce/2010.08
I haven't backlogged, so I don't know how many submissions I've reveived.
I got one via email so far. :) 08:42
isBEKaml :)
masak from someone I can't map to a nick -- exciting! 08:43
isBEKaml no pull requests.
08:44 isBEKaml_ joined 08:48 avuserow joined
masak re how it's going with the Yapsi refactor: the runtime is Mostly Done. I need to equip it with binding so we don't need to regress on a feature. I think I might be able to do binding better than last time. 08:50
then I basically need to write a new parser. I should be able to take the old parser and write new action methods for it.
isBEKaml_ are we still going through the lex scopes in a tree like pattern? [parent->child] 08:51
masak yes.
but the new runtime has a direct link to a scope and the correct variable in that scope.
so the lookup part has been pushed back to compile-time, which is a win. 08:52
isBEKaml_ I see.
08:54 avuserow joined
sorear isBEKaml: um, what? 0 is sunday, how can 0 possibly be right 08:58
08:58 meppl joined
isBEKaml_ sorear: that's what we want. count of all sundays. 08:59
sorear isBEKaml_: $sofar % 7 is the "current" day of the week. Before the big loop, it's 1 Jan 1900; didn't the comment say it was Monday?
09:01 isBEKaml joined
sorear isBEKaml_: $sofar % 7 is the "current" day of the week. Before the big loop, it's 1 Jan 1900; didn't the comment say it was Monday? 09:02
hello masak!
masak sorear: hi!
sorear: today is release day. :)
sorear Anything new happen to yapsi?
masak well, I wrote this new runtime in the beginning of July. 09:03
its chief trait is that it handles `my $a = 2; while --$a { say my $b; $b = 42 }` correctly. 09:04
LaVolta rakudo: my Int @nums = 1 .. 10;
masak (the current runtime doesn't)
p6eval rakudo 9b6189: ( no output )
LaVolta rakudo: my @nums = 1 .. 10; 09:06
p6eval rakudo 9b6189: ( no output )
LaVolta rakudo: my @nums = 1 .. 10; @nums.perl.say
p6eval rakudo 9b6189: OUTPUT«[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]␤»
sorear LaVolta: why do you torment our poor p6eval so 09:07
LaVolta sorear: sorry if I did something wrong
i got "maximum recursion depth exceeded" for my first paste 09:08
09:08 avuserow joined
sorear LaVolta: right. it's listed as the 3rd not working item in the rakudo star "don't even bother to try this" list 09:09
LaVolta oh my my my ... it's all my fault, my bad 09:10
sorear I know p6eval has resource limiters installed, but I'd still like to steer people away from running infinite loops on it
maybe this is pointless and it really does no harm.
LaVolta sorry, I will stop doing those things...
masak I'm more worried about people who try to get into the file system of the host computer. 09:11
breakfast nom &
sorear oops. I'm not *trying* to scare you
09:12 alinbsp joined, rurban joined, rhebus joined
LaVolta let's do that to our local copy... 09:12
instead
rurban The rakudo-star tests do nothing. prove -e ./perl6 -r modules/<name>/t eh? "for d in modules/*; do cd $d; test -d t && test -d lib && (cd lib; prove -e perl6 -r ../t); cd ${B}; done" comes close. How to add -Ilib to perl6?
isBEKaml sorear: well, yeah. I have modified your code - the problem asked for number of sundays that fall on the first of the month in years 1901 to 2000. So...
rurban @*INC is ~/.perl6/lib /usr/lib/parrot/2.6.0/languages/perl6/lib and ., but we need lib to be able to test a module before installing it...
09:13 Ross joined
sorear isBEKaml: yes, but the problem description doesn't give us the weekday for 1 Jan 1901 09:13
it gives us the weekday for 1 Jan 1900
so I do the first year, but without updating $sundays
isBEKaml sorear: right. so, I modified days-so-far to point to 1 Jan 1901 and iterated from 1901 to 2000. 09:16
isBEKaml can't code.. :( 09:17
sorear pugssvn, where art thou 09:18
isBEKaml lying dormant on feather. :D 09:19
.oO(sleeping under the feather)
09:21 ReiniUrban joined
rurban rakudo-star test results nopaste.snit.ch/22503 09:24
szabgab is there any %*ENV or similar in Rakudo? 09:27
rurban Well I would need a PERL6LIB env 09:28
09:36 tadzik joined
tadzik morning #perl6 09:39
masak szabgab: yes. it's spelled like you guessed -- %*ENV 09:40
far as I know, it works.
rurban Ok, PERL6LIB works okay, patch coming in my github star fork 09:41
tadzik hmm, what's so awesome about rakudo alpha that some thing still need it?
masak tadzik: hardly anything nowadays. but that's changed in the last month or so. 09:43
tadzik: two things are holding me back with certain projects: (a) inertia. (b) it's kinda nice to code against something that doesn't change all the time :) 09:44
tadzik I see that YAPSI needs alpha, IIRC november too.
masak correct.
tadzik inertia?
masak fixing the former omission today. :)
tadzik oh, ok
masak inertia is the property of large objects not to be easily moved. 09:45
isBEKaml masak: technically correct, but useless answer. :P
rurban For pmichaud: github.com/rurban/star/commit/cdaf0...33eac97ea5
masak isBEKaml: no, it carries over as a metaphor for software, too :)
isBEKaml masak: that's applicable only to MOP. ;) 09:46
masak moving a large-ish code base from alpha to ng/star is a mostly mechanical, but time-consuming task.
gfldex masak: does that mean you can't stop perl6?
isBEKaml (don't ask me what MOP is ;)
tadzik meta-object protocol? 09:47
09:47 Ross joined
isBEKaml tadzik: right. :) 09:47
masak gfldex: I was talking about code running Perl 6, but sure. you can't stop Perl 6. :)
gfldex \\o/
sorear out.
09:49 hanekomu joined
masak hanekomu! \\o/ 09:50
hanekomu hi
09:50 goksie joined
masak std: my $a = 0; loop while $a < 10 { say $a++ } 09:50
p6eval std 31885: OUTPUT«[31m===[0mSORRY![31m===[0m␤Missing block at /tmp/daks7PYql3 line 1:␤------> [32mmy $a = 0; loop [33m⏏[31mwhile $a < 10 { say $a++ }[0m␤Parse failed␤FAILED 00:01 117m␤»
masak I kinda wish STD'd say "no, it's 'repeat while'" or something. 09:51
isBEKaml std: /x{3}+/ 09:53
p6eval std 31885: OUTPUT«[31m===[0mSORRY![31m===[0m␤Unsupported use of {3} as general quantifier; in Perl 6 please use X**3 at /tmp/H9Mf4GgMog line 1:␤------> [32m/x{3}[33m⏏[31m+/[0m␤Check failed␤FAILED 00:01 115m␤»
isBEKaml std: /x{0,3}+/
p6eval std 31885: OUTPUT«[31m===[0mSORRY![31m===[0m␤Unsupported use of {0,3} as general quantifier; in Perl 6 please use X**0..3 at /tmp/VKMdXIsBJZ line 1:␤------> [32m/x{0,3}[33m⏏[31m+/[0m␤Check failed␤FAILED 00:01 115m␤»
isBEKaml std: /x**3{2}+/
p6eval std 31885: OUTPUT«[31m===[0mSORRY![31m===[0m␤Unsupported use of {2} as general quantifier; in Perl 6 please use X**2 at /tmp/fZe5MAFj8y line 1:␤------> [32m/x**3{2}[33m⏏[31m+/[0m␤Check failed␤FAILED 00:01 117m␤»
isBEKaml sorear++ :)
std: /x{0,3)/ 09:54
p6eval std 31885: OUTPUT«[31m===[0mSORRY![31m===[0m␤Unable to parse statement list; couldn't find right brace at /tmp/0zwMjhRGgu line 1:␤------> [32m/x{0,3[33m⏏[31m)/[0m␤ expecting any of:␤ POST␤ infix or meta-infix␤ null term␤ postfix␤ postfix_prefix_meta_operator␤
..statement modifier loop␤Parse…
frettled hanekomu: o/
isBEKaml uhh
hanekomu hm?
rurban Am I the first who packages rakudo-star? fedoraproject.org/wiki/Features/Rakudo_Star but nothing concrete
isBEKaml rurban: me and tadzik++ have been at it, but stalled due to DESTDIR issue. :) 09:55
rurban Ok, DESTDIR is solved (and easy)
tadzik rurban: I was trying to packagage it half an hour before it was released :)
rurban: is it?
rurban testing is partially solved, see my github fork 09:56
frettled rurban++
tadzik rurban: got tarball?
frettled std: /x{0,3}/
rurban Not yet, still testing
p6eval std 31885: OUTPUT«[31m===[0mSORRY![31m===[0m␤Unsupported use of {0,3} as general quantifier; in Perl 6 please use X**0..3 at /tmp/SbwTxXyIZk line 1:␤------> [32m/x{0,3}[33m⏏[31m/[0m␤Check failed␤FAILED 00:01 115m␤»
isBEKaml rurban: the dist was tarballed? with all modules?
rurban Just ufo and blizkost were misbehaving
I package as star just dependencies to an updated rakudo 2010.07-47 and the extras (libs, docs, tools) 09:57
isBEKaml that reminds me. I still have to submit packages to slackbuilds.org. Su-Shee++ for raising this.
bbiab 09:58
rurban my descr is "rakudo-star is parrot plus rakudo (a perl6 implemention on parrot)
plus some perl6 libraries, docs and libraries and blitzkost, a parrot
language which targets perl5. Contrary to the upstream rakudo-star release
for the masses, this does not include the external parrot or rakudo
releases. The external rakudo release matches the rakudo star release,
that's why it has the version 201007_47, not just 201007."
cxreg blizkost (not blitzkost) is not a "parrot language" 10:00
it's a perl 6 module which embeds libperl 10:01
rurban it's in parrot languages, so its just another parrot language, just as perl6
hanekomu where does the name "blizkost" come from? It does indeed sound like "blitzkost", which would be a new German word for something like "lightning-fast food"
rurban solvak
slovak,sorry
hanekomu so it might have the same connotations 10:02
rurban "closeness", "contiguity" and "familiar"
hanekomu oh
rurban README - Oh, and why the name?
masak hanekomu: it's Slovak, IIRC.
rurban: jnthn++ tends to give his projects unpronouncible Slavic names :P 10:03
rurban isBEKaml: I'm just preparing a star distro for cygwin
cxreg rurban: i guess technically it is implemented that way, but the way it's written it sounds more like Ponie than what it really is
(and i was wrong, it doesn't appear to depend on rakudo)
szabgab rakudo: say %ENV.perl 10:04
rakudo: say %*ENV.perl
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤Symbol '%ENV' not predeclared in <anonymous> (/tmp/CSQ5kpJvxp:22)␤»
rakudo 9b6189: OUTPUT«Method 'key' not found for invocant of class 'String'␤ in <anon> at line 4669:CORE.setting␤ in 'Any::join' at line 1␤ in 'EnumMap::perl' at line 4675:CORE.setting␤ in main program body at line 22:/tmp/apQGkcYZsb␤»
10:04 briang joined
szabgab masak: I was also hoping for that but it does not like me :-( 10:04
masak szabgab: small example? 10:05
rurban Ok. I wrote now "blizkost, a perl5 parrot language which embeds libperl5"
szabgab rakudo: say %*ENV{HOME}
p6eval rakudo 9b6189: OUTPUT«Could not find sub &HOME␤ in main program body at line 22:/tmp/2QUMQvKuMZ␤»
szabgab rakudo: say %*ENV.perl
p6eval rakudo 9b6189: OUTPUT«Method 'key' not found for invocant of class 'String'␤ in <anon> at line 4669:CORE.setting␤ in 'Any::join' at line 1␤ in 'EnumMap::perl' at line 4675:CORE.setting␤ in main program body at line 22:/tmp/C6nZQsbS5h␤»
masak szabgab: .{HOME} is a p6ism.
szabgab: read the error message.
szabgab rakudo: say .{HOME} 10:06
p6eval rakudo 9b6189: OUTPUT«Could not find sub &HOME␤ in main program body at line 22:/tmp/XFtRAQc1_J␤»
masak szabgab: I watched your hashes screencast. you know this! :)
szabgab rakudo: say %*ENV{"HOME"}
p6eval rakudo 9b6189: OUTPUT«/home/p6eval␤»
masak right.
szabgab crap
but still
rakudo: say %*ENV.perl
masak no, just a literal string :)
p6eval rakudo 9b6189: OUTPUT«Method 'key' not found for invocant of class 'String'␤ in <anon> at line 4669:CORE.setting␤ in 'Any::join' at line 1␤ in 'EnumMap::perl' at line 4675:CORE.setting␤ in main program body at line 22:/tmp/VPjUwQ3k5M␤» 10:07
masak that looks like Parrot things leaking through.
rurban code.google.com/p/cygwin-rurban/sou...tail?r=114 contains my star packaging (spec and patches)
masak submits rakudobug
szabgab hides
10:08 redicaps joined
rurban isBEKaml: are the slackware builds already accepted? 10:10
masak another small challenge to people here on the channel: if someone would go through Yapsi and list all of the Perl 6 features it uses, we make a roadmap for bootstrapping Yapsi out of that list.
I think if I were to give a lightning talk, I would show such a list, too. so this is a chance for fame and glory. 10:11
s/we/we can/
rurban I need some more rakudo-star distros so that it will be easier to get accepted at cygwin. Otherwise I need to go through voting
cxreg cygwin depends on it being in other distros? 10:12
masak rakudo: say "a", "c" ... "m" 10:13
p6eval rakudo 9b6189: OUTPUT«acdefghijklm␤»
rurban It is just faster
I was the first anyway who packaged parrot
(in an installable way)
cxreg debian's rakudo is ancient 10:14
cxreg wonders what Ryan Niebur is up to 10:15
that appears to be Ryan52 on irc
rurban: maybe bug him tomorrow :) 10:16
isBEKaml rurban: not yet. I have yet to submit them. I'm currently running them through on my machine. (My own custom version works, though) 10:21
masak does an array my @a[100] somehow link to 100 scalar containers? 10:23
isBEKaml rurban: you can find my custom version here: github.com/svatsan/RakudoStar-Slackbuilds 10:25
rurban: Note, when I submit them to slackbuilds, it won't be an R* dist version. It will be separate builds for parrot and rakudo each. 10:27
rurban I'll have 3 seperate packages: parrot (plus parrot-devel), rakudo, and now rakudo-star where -star just has the additional files 10:31
10:31 whiteknight joined
isBEKaml rurban: right. rakudo-star is just a custom version I'm making. It's currently unviable pushing for an install when they are actually two different packages/dependencies(however you see it) due to how slackware treats installs. 10:32
has anyone noticed this is a trending repo on github today: github.com/kartick18/inception ? 10:33
due to the awesomely popular inception! :D
snarkyboojum isBEKaml: linky no worky 10:35
masak snarkyboojum! \\o/
isBEKaml: it's been trending for a few days at least. 10:36
snarkyboojum masak: \\o/
cxreg rurban: maybe rakudo should be rakudo-compiler
masak snarkyboojum: release day today!
snarkyboojum masak: w00t!
cxreg and rakudo could be a meta-package that does both?
isBEKaml snarkyboojum: sorry, github.com/karthick18/inception
masak snarkyboojum: I'm hard at work getting the new runtime to do all that the old one does.
rurban nono, it's okay like that. It's a language interpreter
snarkyboojum masak: sweet - ready for tardis soon I take it :) 10:37
rurban I need stable names downstream
masak snarkyboojum: yes. there should be time (no pun intended) for that today as well.
snarkyboojum masak: oh wow - productivity ahoy! 10:38
masak :)
isBEKaml masak: :)
cxreg isBEKaml: /kick Fischer
isBEKaml cxreg: he never wakes up. :(
masak tries to make scalar containers invisible to '=' but visible to ':=' 10:39
snarkyboojum isBEKaml: wow, the whole concept behind that project is way over my head. Looks like I'll have to do some googling :) 10:40
cxreg needs more ascii art 10:42
isBEKaml snarkyboojum: I found the code easier to understand than the movie. :P
isBEKaml hides
snarkyboojum hasn't seen it 10:43
cxreg isBEKaml: i thought they were too forthcoming with explanation in the movie
snarkyboojum would reading the code/running it count as a spoiler for the film? :) 10:44
cxreg probably
isBEKaml cxreg: the talk about levels of sleep and drugs kept me thinking about how deep they were and how close they are to their goal. This was the first movie I found engaging. :)
masak TimToady: it's too bad that $obj.?nosuchmethod returns Nil and that Nil is defined. it would be very convenient to do $obj.?nosuchmethod // $some-default 10:45
rurban I've not seen it yet, so please don'r spoiler!
isBEKaml snarkyboojum: No, don't read it! Contains too many spoilers! :)
snarkyboojum isBEKaml: I'll park it for another day then ;)
isBEKaml snarkyboojum: :)
masak snarkyboojum: good movie. I enjoyed it. 10:46
definitely a movie for programmers. I can see why they created that github repo.
cxreg masak: it's no Primer, but it was decent 10:47
snarkyboojum masak: excellent :) might go see it this week
masak cxreg: oh, then I have to watch Primer, too :)
10:47 molaf joined
cxreg yes, do! 10:47
isBEKaml cxreg: I haven't seen Primer. Might watch it sometime.
masak I think I just got binding to work again in the new runtime.
cxreg primer is on netflix streaming, if you have that 10:48
masak yes, I did! \\o/
rakudo: say 42.?nosuchmethod // "default never reached :(" 10:49
p6eval rakudo 9b6189: OUTPUT«␤»
rurban I see the rakudo-star package *MUCH* smaller than perl5 (where I already include a lot of vendor_modules)
rakudo: 2.1MB star: 387KB parrot 1.1MB perl5 10MB
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤Confused at line 22, near "2.1MB star"␤»
masak p6eval: :)
isBEKaml rurban: star's just 387kB ? You're not packaging rakudo/parrot along with it but just the docs/extras? 10:51
10:53 avuserow joined
rurban yes 10:54
and the modules 10:55
and blizkost
using shared libs to libparrot
really tiny
masak rakudo: my $a = "fool!"; $a ~~ / (foo) /; say $0 10:57
p6eval rakudo 9b6189: OUTPUT«foo␤»
masak rakudo: my $a = "fool!"; $a !~~ / (foo) /; say $0
p6eval rakudo 9b6189: OUTPUT«Any()␤»
masak submits rakudobug 10:58
just because I want it not to match doesn't mean I don't want it to capture. :)
10:58 Italian_Plumber joined
masak oh well, easy/obvious workaround. 10:58
rakudo: my $a = "fool!"; $a ~~ / (foo) { say "OH HAI" } / 10:59
p6eval rakudo 9b6189: OUTPUT«OH HAI␤»
masak rakudo: my $a = "fool!"; $a !~~ / (foo) { say "OH HAI" } /
p6eval rakudo 9b6189: OUTPUT«OH HAI␤»
masak interesting. so the error seems to be in !~~ 11:00
the lack of assigning to $/, that is.
rakudo: my $a = "fool!"; $a ~~ / (foo) { say $0 } / 11:02
p6eval rakudo 9b6189: OUTPUT«foo␤»
masak rakudo: my $a = "fool!"; $a !~~ / (foo) { say $0 } / 11:03
isBEKaml rurban: nice. packaging it everytime is why I kept star dist custom to my own version.
p6eval rakudo 9b6189: OUTPUT«foo␤»
isBEKaml masak: Interesting...
masak interestinger and interestinger...
isBEKaml rakudo: my $a = "boo FAAA"; $a !~~ / (FAA) { say $0 } / 11:04
p6eval rakudo 9b6189: OUTPUT«FAA␤»
isBEKaml blah!
rakudo: my $a = "boo FAAA"; $a !~~ / (boo) (FAA) { say $1 } / 11:05
p6eval rakudo 9b6189: ( no output )
masak that shouldn't match :)
(no matching on the space)
isBEKaml $0 seems to be the default match for everything.
11:05 redicaps joined
isBEKaml rakudo: my $a = "boo FAAA"; $a !~~ / (boo) \\s (FAA) { say $1 } / 11:05
p6eval rakudo 9b6189: OUTPUT«FAA␤» 11:06
isBEKaml :)
masak isBEKaml: 'default match for everything'. I didn't understand what you meant there.
isBEKaml masak: never mind.
masak: I was thinking along the lines of awk's $0 11:07
isBEKaml blinks.
masak ah.
no, $0 desugars to $/[0]
11:08 colomon joined
isBEKaml masak: sometimes, I forget what I'm doing and see totally unrelated things. ;) 11:08
colomon o/
masak colomon: \\o
isBEKaml masak: can you explain the last p6eval output. !~~ doesn't work right there.. 11:09
masak isBEKaml: the more you do Perl 6, the less that'll happen :)
isBEKaml: the last p6eval output looks right to me.
isBEKaml: what did you expect?
isBEKaml masak: I didn't ask it to match. !~~ should not match what I ask to match, right?
masak isBEKaml: there's a confusion of levels in that question. 11:10
isBEKaml masak: IOW, if $a doesn't match <this pattern> do this <blocK>
masak isBEKaml: start by thinking of !~~ as 'not ~~'
isBEKaml: that is, do '~~' and then negate it.
isBEKaml: we still do exactly the same things inside the regex.
including binding to $0, $1 etc. 11:11
which was the point of my rakudobug submission.
isBEKaml 16:38 < isBEKaml> masak: IOW, if $a doesn't match <this pattern> do this <blocK>
so this is wrong?
masak yes.
it's a confusion of levels.
colomon masak: did I miss any excitement yesterday? 11:12
masak the regex is like a subroutine. the insides of it are not negated, only the eventual result.
isBEKaml masak: I see you speaking of Inception. ;)
"confusion of levels"
masak isBEKaml: confusions of levels aren't restricted to Inception :P
isBEKaml masak: so are they not restricted to dreams. :P 11:13
masak colomon: one always misses excitement by being absent from #perl6 :)
levels? dear me, no.
11:13 snearch joined
isBEKaml masak: I think I understand what you're trying to say, but I'll have to code some what better than the bloated piece I came up with this morning. :D 11:14
to really understand it.
masak isBEKaml: think of it like this: infix:<~~> means "match on this regex, return whether it matched". infix:<!~~> means "match on this regex, return whether it failed". apart from the return value, everything else remains the same. 11:15
isBEKaml rakudo: my $a = "Boo FAA"; if $a !~~ / (Boo)\\s(FA) { say $1 } / { say "I'm dead!"; }; 11:17
p6eval rakudo 9b6189: OUTPUT«FA␤»
isBEKaml riiight!
masak++ 11:18
colomon rakudo: say ("hello" ~~ /e/).WHAT 11:19
masak pmichaud: we don't have <?after ...> yet in nqp-rx? o.O
p6eval rakudo 9b6189: OUTPUT«Match()␤»
isBEKaml masak: I now see what you meant by "confusion of levels" - I'd say "confusion of blocks" :) 11:20
masak fair enuf.
colomon can one of you try that last snippwt of mine with !~~ 11:21
tt
isBEKaml akudo: say ("hello" ~~ /e/).WHAT 11:22
rakudo: say ("hello" !~~ /e/).WHAT
p6eval rakudo 9b6189: OUTPUT«Bool()␤»
colomon typing p6 on my phone is very hard
isBEKaml++
masak colomon: ah!
colomon++ 11:23
still doesn't explain why $/ doesn't get set.
11:23 hudnix joined, avuserow joined
colomon when you use !~~ ? 11:24
scope?
is !~~ using standard !op metaop?
masak colomon: I just submitted a Rakudobug about it. rt.perl.org/rt3/Ticket/Display.html?id=76892 11:25
colomon: I don't know. 我不知到 11:26
gah, wrong DAO. 道 11:27
colomon masak++
rurban This was my star packaging proposal for cygwin cygwin.com/ml/cygwin-apps/2010-08/msg00002.html
isBEKaml DAO ? 11:28
masak isBEKaml: I'm typing phonetically. both 到 and 道 are pronounced 'dao'. 11:29
isBEKaml: but it's the former that's used in the idiom "I don't know". it's the same character as the one in 'daoism', by the way. very fitting.
literally, 我不知道 translates to "I do not know the way". 11:30
but due to the way verb-object pairs work in Mandarin, no-one thinks "know the way" when they read that, only "know".
isBEKaml masak: I don't know chinese/japanese. All I see are four ???? 11:31
11:31 gbacon joined
masak isBEKaml: then you need more fonts. :) 11:32
isBEKaml mimo no comprende. :)
masak: Switching between X and console. ;)
11:33 avuserow joined
masak rakudo: my $s = "foo\\{foo\\}"; $s ~~ / <?{ $/.CURSOR.pos > 1 && $s.substr($/.CURSOR.pos - 1, 1) eq "\\{" }> foo /; say $/.from 11:37
isBEKaml masak: I'm not good with languages at all.
p6eval rakudo 9b6189: OUTPUT«4␤»
masak workaround for lack of <?after ...> :)
too bad the Cursor won't tell me what target string it's matching on. then I wouldn't have to save it into $s before. 11:40
cognominal hum how I can get a MyGrammar.parse( "...", :target<parse>) to get the same effect (dumping parse tree) as perl6 --target=parse ? 11:42
11:44 rurban joined
masak cognominal: augment the appropriate class with a new parse multi that requires :$target! = "parse" 11:44
I guess that'd work, at least.
11:44 barika joined
masak hm, :$target! where "parse" 11:45
cognominal say (A.parse('ab', :target<parse> )).perl; # dumps but in a different format.
masak rakudo: "foo" ~~ /<?after o> o/; say $/.from # should say 2 11:46
p6eval rakudo 9b6189: OUTPUT«Method 'after' not found for invocant of class 'Cursor'␤ in <anon> at line 22:/tmp/r91PDvy9VJ␤ in 'Cool::match' at line 2377:CORE.setting␤ in 'Regex::ACCEPTS' at line 5635:CORE.setting␤ in main program body at line 10:/tmp/r91PDvy9VJ␤»
11:55 azert0x joined
masak ah, it's good to be back to coding again. 11:56
12:09 x3nU joined
x3nU hi 12:09
masak x3nU: hi! 12:11
x3nU i've got problem with perl6 and/or SCGI module 12:26
i'm trying to run script from SCGI readme
(github.com/supernovus/SCGI/blob/master/README)
and that's what i get: 12:27
Unable to parse postcircumfix:sym<( )>, couldn't find final ')' at line 3
i'm using latest rakudo perl version on win32
masak sounds like there's a ')' missing.
masak looks at the code in the README
12:28 redicaps1 joined
masak x3nU: example looks fine to me. did you type it in or copy it in? 12:29
x3nU i copied
masak x3nU: what does line 3 contain in your script? 12:31
oh!
there's a comma missing in the README script.
should be one right after :addr<localhost>
(not necessary by spec IIRC, but by current Rakudo) 12:32
x3nU thanks, additionaly there was "return $return" and i changed this to "return $response" 12:34
and it started working
masak x3nU++ 12:36
phenny: tell supernovus that in github.com/supernovus/SCGI/blob/master/README, you need a comma on line 3 and 'return $response' last. x3nU++ discovered this. 12:37
phenny masak: I'll pass that on when supernovus is around.
snarkyboojum rakudo: package A {}; say A.WHAT 12:39
p6eval rakudo 9b6189: OUTPUT«Method 'WHAT' not found for non-object␤ in main program body at line 22:/tmp/QBY5vDgnzO␤»
snarkyboojum rakudo: module A {}; say A.WHAT
p6eval rakudo 9b6189: OUTPUT«Method 'WHAT' not found for non-object␤ in main program body at line 22:/tmp/tfSUuUY7RF␤»
masak snarkyboojum: that one is known and reported, IIRC. 12:40
snarkyboojum masak: ah cool - just trying to work it out - way over my head though :)
12:40 isBEKaml joined 12:41 umphy joined
masak snarkyboojum: well, the cause seems clear even if the fix isn't. 12:41
snarkyboojum: modules and packages are probably Parrot-y things, so they're too different from ordinary Perl 6 objects.
rakudo: say Module.WHAT; say Package.WHAT
p6eval rakudo 9b6189: OUTPUT«Could not find sub &Module␤ in main program body at line 22:/tmp/dDgx8iRMBG␤»
masak huh. 12:42
std: Module; Package;
p6eval std 31885: OUTPUT«ok 00:01 114m␤»
snarkyboojum rakudo: say Perl6::Compiler::Module.WHAT 12:43
p6eval rakudo 9b6189: OUTPUT«Perl6::Compiler::Module()␤»
masak snarkyboojum++
isBEKaml rakudo: module Foo; Foo.WHAT.say; 12:45
p6eval rakudo 9b6189: OUTPUT«Method 'WHAT' not found for non-object␤ in main program body at line 22:/tmp/RaCKtT4B4h␤»
isBEKaml rakudo: module Foo { ... }; Foo.WHAT.say;
p6eval rakudo 9b6189: OUTPUT«Failure()␤» 12:46
isBEKaml uhh, Failure() ?
rakudo: module Foo { module Bar { } }; (Foo::Bar).WHAT.say; 12:48
p6eval rakudo 9b6189: OUTPUT«Can not find sub Foo::Bar␤ in main program body at line 1␤»
12:49 colomon_ joined
isBEKaml rakudo: module Foo {}; Foo.WHAT.say; 12:50
p6eval rakudo 9b6189: OUTPUT«Method 'WHAT' not found for non-object␤ in main program body at line 22:/tmp/xmMm7ZJqLk␤»
isBEKaml rurban: why does the fedora project wiki state Rakudo Star as the first production release of Rakudo? 12:54
fedoraproject.org/wiki/Features/Rakudo_Star 12:55
masak is S05:3939 inaccurate? should it say .parsefile instead of just .parse? 12:58
by the way, what's so great about .parsefile that .IO.slurp.parse doesn't do?
oh well, maybe it's a reasonable shortcut. 12:59
12:59 Ross joined, Kodi joined
Kodi Should numeric types have a .Duration method to cast them to a Duration? If not, what would be the right way to create an arbitrary Duration, short of DateTime.new(...).Instant - DateTime.new(...).Instant? 13:01
masak maybe one could also create a Duration from a DateTime::Duration (not spec'd yet) 13:03
Kodi masak: DateTime::Duration? I've never even heard of that. What's it supposed to do? 13:04
masak Kodi: it's supposed to be a port of the corresponding CPAN module.
Kodi: would allow adding and subtracting DateTimes.
feel free to author that part of the Temporal spec :) 13:05
Kodi masak: Oh, yes, I've seen that module. I'm not sure that kind of thing belongs in the core.
masak: Then again, we didn't think leap seconds belonged in the core. :)
masak I think people might miss it, since adding and subtracting Date objects already work. 13:06
but yes, it's a slippery slope.
13:06 HarryS joined
Kodi We *could* get useful DateTime arithmetic much more easily by just making $dt - $dt equivalent to $dt.Instant - $dt.Instant, and so on. 13:07
masak I think autarch's CPAN module makes good arguments for why one doesn't want that.
specifically, when I say "one month later", I mean something context-dependent,, not a fixed time difference. 13:09
13:09 kid51 joined
Kodi masak: Yeah, there'll always be a place for fuzzy DateTime arithmetic like that. I guess the question is whether the built-in arithmetic should do that, or it should be provided in temporal-utils. 13:10
masak I'm fine with the latter. 13:11
m6locks omg i just noticed the OFFICIAL Rakudo Star release, building it now
Kodi masak: Good. I like it better, I think. It means it can be developed independently of Rakudo, for one thing. 13:12
m6locks with all the delicious treats and modules and stuff that didn't work sometime ago
okay
some error
find rakudo-star-2010.07 -name '.*' -prune -o -type f -printf '%P\\n' >rakudo-star-2010.07/MANIFEST
masak m6locks: yeah, it's pretty rad. 13:13
m6locks find: -printf: unknown option
make: *** [rakudo-star-2010.07/MANIFEST] Error 1
wuts dat
masak m6locks: you're on... Unix?
snarkyboojum m6locks: ah yeah.. on OS X?
m6locks mac os x
masak m6locks: I think ingy got that one too.
m6locks aye
snarkyboojum -printf isn't understood by find on OS X
i.e. not an option
m6locks its 10.6.4
snarkyboojum it's a GNUism or Linuxism I think 13:14
m6locks ok, so like can i hack that
edit it out
isBEKaml aye, printf works on linux
Kodi masak: Would you object to my adding Duration.new() to S02?
snarkyboojum I was going to, but didn't get around to it - it just strips off the first directory in the listing I think
m6locks MANIFEST as a file is zero lines, so can't do much about it :PP :PP 13:15
masak Kodi: insufficient data. a patch would help.
frettled snarkyboojum: GNUism, yes.
masak I probably won't object, though.
m6locks so does this mean that I don't get to play with it >:/ 13:16
isBEKaml it gets all filenames and writes to MANIFEST
m6locks hmph
masak m6locks: not without manually tinkering a bit, at least.
snarkyboojum m6locks: well not the star distro which *creates* star releases :), but shouldn't be too hard to work around :)
m6locks: you can still install and play with Star :) 13:17
frettled m6locks: -printf '%P\\n' removes the initial ./ from the results of find
m6locks: that is, it removes the pathspec
Kodi masak: A patch to the spec, you mean? It would be easy as adding the sentence "A C<Duration> can also be constructed with C<Duration.new>, which takes a number of seconds as its only argument." or so.
snarkyboojum I think it removes 'rakudo-star-2010.07' doesn't it?
m6locks thanks guys 13:18
it's fetching it now, i'll see if it works by just removing the -fprint '' part 13:20
13:20 colomon_ joined
masak Kodi: doesn't sound like it belongs in S02, actually. more like S32. 13:21
Kodi: and it's almost, but not quite, self-evident that it does have such a .new constructor. 13:22
Kodi masak: In fact, probably most of the stuff about Instants and Durations belongs in S32.
masak aye.
Kodi masak: The reason it didn't strike me as anywhere near self-evident is that it's really just supposed to be a kind of number, and we don't have, e.g., Int.new.
snarkyboojum m6locks: it'll create a manifest if you remove the -printf option, but the path to the files in the manifest will contain the name of the release directory, i.e. rakudo-star-2010.07/
m6locks: so that's all that needs to be worked around on OS X :) 13:23
Kodi (Or do we? It doesn't work in an obvious way in Rakudo, at least.)
m6locks snarkyboojum: aye, it'd be nice to mention that one in the README file I guess for all the OS X users 13:24
13:24 LaVolta joined
snarkyboojum m6locks: yeah - and/or find a command that works on all envs :) 13:24
m6locks aye ;P
isBEKaml I think -print would work. 13:25
13:25 am0c joined
isBEKaml but it doesn't remove the rakudo-star part, it would just write to MANIFEST along with it 13:26
Kodi gets a phone call.
snarkyboojum isBEKaml: works without -print as well, the problem is removing the release dir
"problem"
13:27 tommyd joined
m6locks hmm not working, getting lots of tar errors 13:27
tar: rakudo-star-2010.07/rakudo-star-2010.07/skel/README: Cannot stat: No such file or directory
isBEKaml snarkyboojum: yes, -print is _not_ needed in most cases.
13:27 mjk joined
snarkyboojum isBEKaml: yep 13:27
isBEKaml rakudo-star/rakudo-star? that doesn't seem right. what did you do?
frettled run it on an untested system ;) 13:28
(non-GNU)
isBEKaml yeah, mac :)
m6locks i removed the -fprint 'blah' part from the makefile
isBEKaml :q
hugme hugs isBEKaml, good vi(m) user!
frettled Wherever that find-line is, shoot the -printf '%P\\n' bit horribly, and pipe it through sed instead:
isBEKaml blah!
m6locks well that one above frettled just said 13:29
frettled | sed -e 's,^[^/]*/,,'
That ought to remove the first part of the path, up to and including the first slash
m6locks hmm I'll add it 13:30
13:30 colomon_ joined
m6locks bbl to report results, gotta run now 13:31
snarkyboojum I just piped it through perl and got it going, but yeah, sed or awk would be better :)
13:31 amkrankruleuen joined 13:32 Ross joined
frettled but you have to avoid gawking on a Mac ;) 13:33
isBEKaml macgawkin' 13:34
:D
13:41 colomon__ joined
snarkyboojum find $(DISTDIR) -name '.*' -prune -o -type f | sed -e 's|^[^/]*/||' >$(DISTDIR)/MANIFEST seems to work a treat 13:43
frettled++ # nicer than my perl incantation :)
err.. invocation ;)
frettled You're very welcome! 13:49
frettled prefers using commas rather than pipes, because a misplaced single quote can lead to, ahem, interesting side effects :) 13:50
isBEKaml prefers using + than , cos they aren't in english! ;) 13:54
13:55 tadzik joined 13:56 amkrankr1leuen joined
moritz_ is at the YAPC::EU venue 13:57
frettled moritz_: oooh. How is it? 14:00
moritz_ hm 14:01
odd
expensive :-)
not enough perl programmers around yet 14:02
masak moritz_: will the hackathon on Tuesday be at the venue somewhere?
moritz_ masak: I suspect in the lobby
masak: there's free wifi here 14:03
masak \\o/
moritz_ it's slow, and high ping times
but it works
masak excellent.
I'm not sure when I'll show up in Pisa tomorrow. probably late.
I suspect we won't meet until on Tuesday. 14:04
snarkyboojum potential patch for the rakudo star Makefile on OS X - gist.github.com/503392 (unless there is a better way of doing it) - I haven't tested it on linux ;) 14:08
isBEKaml snarkyboojum: that should work on linux. I haven't had the time to build it as I'm now looking to package for slackware. :) 14:10
but yeah, I could just run a sample command. :) 14:11
snarkyboojum isBEKaml: cheers - that's what I thought ;) but should be tested eh
isBEKaml++ :)
x3nU if i create perl 6 module where should i submit it? (just asking, currently i'm learning perl6 ;))
tadzik x3nU: publish them on github 14:12
isBEKaml x3nU: github is the unofficial source repo for p6 (even rakudo is hosted on github) :)
x3nU: it's fine if you host it elsewhere too, just let us know! :) 14:13
14:13 slavik joined
slavik there is no more pugs? 14:14
masak slavik: it's there, as always. 14:15
slavik ahh
it's not mentioned in the subject :(
masak pugs: say "OH HAI"
x3nU well too bad that github uses git ;/
p6eval pugs: OUTPUT«OH HAI␤»
masak x3nU: too bad? why? we think it's too good! :) 14:16
太好了!
isBEKaml x3nU: what do you want to use? svn?
snarkyboojum: as said, the command works. 14:17
snarkyboojum++
snarkyboojum w00t!
tadzik x3nU: github supports svn too ;)
snarkyboojum :)
tadzik (if you are a masochist) 14:18
isBEKaml tadzik: no, tell that to cvs folks. ;) 14:19
x3nU i was thinking about mercurial ;f
tadzik isBEKaml: but they are extinct, aren't they? :) 14:20
isBEKaml tadzik: yes, folks are still around to use it. :D 14:21
14:21 Axius joined
Axius hi 14:22
masak Axius: hi!
bye... 14:23
14:23 Axius joined
isBEKaml masak: no, wait. No Yapsi release today? 14:23
masak isBEKaml: yes Yapsi release today. I'm coding full speed here. :) 14:24
isBEKaml productivity++
masak isBEKaml: finishing up the refactor by providing the new runtime with a new compiler :)
essentially, Yapsi will get a whole new Yapsi this month.
Axius rakudo: say 53 **2 14:25
isBEKaml masak: great! happy hackin' :)
p6eval rakudo 9b6189: OUTPUT«2809␤»
masak is very happy
14:25 clintongormley joined
slavik btw, will/does rakudo allow dynamic loading of modules? or would that have to be part of parrot? 14:28
14:28 synth joined
masak it's the third time or so that I write a Match tree traverser. this time I'm really, really happy about it: gist.github.com/503417 14:28
I use it in the new Yapsi to find all the variable declarations in a given block. 14:29
slavik pmichaud: if you're around. where can I read about benchmarks of parrot?
pmichaud good morning, #perl6
slavik sweet :D
frettled good moaning!
slavik morning :)
masak morning, pm!
pmichaud ...benchmarks of parrot? 14:30
slavik pmichaud: www.pmichaud.com/2008/pres/oscon-pc...b.html#end
pmichaud: looking for some numbers or something :) (out of interest)
pmichaud those were all hand generated pir versus p5 equivalents 14:31
we didn't store them anywhere
frettled ooh, nice comment on my blog: www.blogger.com/comment.g?blogID=7...0528415524
slavik pmichaud: I see. :( 14:32
14:33 dual joined
slavik pmichaud: www.mail-archive.com/perl6-all@perl...27196.html 14:34
jnthn o/ (briefly) 14:35
pmichaud slavik: yeah, that's an example of one of the benchmarks that was used
masak jnthn! \\o/
slavik pmichaud: do those numbers look fine to you?
pmichaud jnthn: o/
slavik: depends on what you're looking for.
2003 was a long time ago.
slavik pmichaud: just wondeirng about parrot's performance relative to C/python :P 14:36
jnthn masak: Don't panic! :-)
slavik pmichaud: would you say parrot got faster?
masak panics a bit less
pmichaud slavik: I wouldn't say much of anything regarding parrot's performance anymore. (more)
beyond that, the parrot of 2003 looks nothing like the parrot of today
masak jnthn: actually, I'm having quite a good time with Yapsi today. for a little moment I've forgotten about GSoC and my slides... 14:37
slavik how come?
masak panics again
pmichaud slavik: because a lot has changed since that early version (more)
frettled Relax, everything is under full panic.
pmichaud I suspect that the PIR versions that were benchmarked then were doing everything in I registers
and there probably wasn't any subroutine call overhead 14:38
or, if there was, it would've been using the old parrot calling conventions, which didn't involve any PMC creation for parameters or contexts
jnthn masak: GSoC tormentor is tormented.
slavik pmichaud: would you say that it's worth doing some benchmarks now?
pmichaud slavik: sure, one could some benchmarks now. 14:39
masak jnthn: I'll make sure to get a bit of hacking and blogging in later today, when the Yapsi stuff feels stable.
slavik pmichaud: I am kind of interested in how Perl6 can perform at the moment compared with Perl5
jnthn masak: :-)
slavik pmichaud: according to that page, even though it is from 2003, parrot performance is very good for 'general' code 14:40
jnthn masak: I'm here for 5-10 mins more now, will be back later on though and around until I go to the airport, if you need any GSoC mentorization. :-)
slavik anyone here ever program with MPI?
pmichaud slavik: based on my experiences since then, I don't think I would agree with the 'general code' comment, at least not for Parrot of today
masak jnthn: our latest privmsg session still sustains me, so I'm fine for now.
jnthn masak: \\o/ 14:41
pmichaud at the time I think the parrot designers were making some extreme assumptions about the kind of code that could be produced from a dynamic language
slavik pmichaud: how come? are you saying that parrot has gotten slower?
Khisanth sounds more like assumptions have changed :)
jnthn slavik: The issue was more that the things they were saying "oh hey, Parrot can do this real fast" were not the kind of things we needed to do to run Perl 6 code.
slavik I see 14:42
pmichaud need an example?
slavik jnthn: but parrot is not primarily built for rakudo, was it?
pmichaud: yes please. :)
pmichaud okay
in a benchmark like the primes example, they likely make use of the 'mod' opcode
i.e., they can do $I0 = $I1 mod $I2
slavik right
pmichaud which does a modulo operation between $I1 and $I2, storing the result into $I0 14:43
in Perl 6, the equivalent statement would be
my $remainder = $v % $div
slavik remainder?
oh right 14:44
pmichaud well, whatever
slavik sorry
pmichaud anyway, that code translates to the following
slavik right
same as C
pmichaud $P0 = find_lex '$v'
$P1 = find_lex '$div'
$P2 = '&infix:<%>'($P0, $P1)
$P3 = find_lex '$remainder'
'&infix:<=>'($P3, $P2) 14:45
slavik even x86 is more efficient in modulo
what exactly is find_lex? (variable lookup?)
pmichaud i.e., it involves three lexical lookups and two subroutine calls
slavik should the 2 subroutine calls simply be assignments? 14:46
or is it for overloading ...
it seems that parrot is more like writing java than writing the jvm
pmichaud only if your compiler is smart enough to detect that overloading won't occur, and that type checking isn't in play, and that $remainder isn't readonly, and ...
slavik pmichaud: is it safe to assume that rakudo will be able to figure such things out? 14:47
pmichaud slavik: no.
slavik optimizer?
jnthn Depends.
On how much info is available.
pmichaud slavik: let's suppose that rakudo could figure it out
jnthn (at compile time)
pmichaud then we'd be down to
14:48 pascal__ joined
pmichaud and let's suppose that rakudo could avoid the extra fetches 14:48
then at best we would be down to
slavik my int $a = 5; my int $b = 3; my int $c = $a % $b;
that gives it more ifo
info*
jnthn Gotta run - back later
pmichaud in that limited case, rakudo might be able to optimize down 14:49
but your original question was about 'general code' 14:50
14:40 <slavik> pmichaud: according to that page, even though it is from 2003, parrot performance is very good for 'general' code 14:51
slavik which would involve arithmetic operations :)
pmichaud no, in this case it involves arithmethic operations of explicitly declared native types that aren't being passed or used in loops or subroutines 14:52
I don't think that's "general code"
slavik pmichaud: but from Larry's apocalypses, Perl6 allows specification of types for vars which should allow the compiler to optimize the code a bit. 14:53
pmichaud a bit, yes.
masak phew -- I thought I had discovered another case of blocks not being cloned properly, but it was just me not cloning an array properly :)
pmichaud all the way down to the primse.pasm example.... I'm doubtful.
*primes.pasm
masak pmichaud: I missed <?after ...> today. filed a rakudobug about it. 14:54
pmichaud masak: okay.
masak pmichaud: I found a workaround using $/.CURSOR.pos and the sacrifice of a chicken, though :)
slavik pmichaud: there is always a trade off when using a higher level language. 14:56
pmichaud slavik: no argument there. :)
slavik pmichaud: my next question. :) when rakudo compiled Perl6 code, does it go straight to parrot bytecode or to parrot assembly?
pmichaud: if Perl6 is 10% slower than parrot, then I'll take it :D
pmichaud currently to PIR, then to bytecode. Parrot doesn't have a mechanism to go directly to bytecode for anything except PIR. 14:57
masak rakudo: my &foo = { next }; for 1..3 -> $n { foo if $n == 2; say $n } 14:58
p6eval rakudo 9b6189: OUTPUT«1␤3␤»
masak \\o/
next from within a closure :)
sadly, I think I just found a legitimate use case for that... 14:59
15:03 aCiD2 joined
aCiD2 I'm hacking on adding "getprotobyname" to rakudo. I've added it to parrot as a dynop library, how do I get this into a perl 6 subroutine? 15:03
(I've got src/core/socket.pm which is like src/core/system.pm) 15:04
pmichaud pir::getprotobyname__<args> would seem to be easiest
aCiD2 what is the args bit? I've seen "time__n" but I don't know what that means
and how do I load the library? In pir I'd have to do .loadlib "socket_ops" first
pmichaud example: pir::xyz__PsPi 15:05
the first character is the type of register to store the result in
the remaining characters are the types of values to use as arguments
P = PMC register, S = string register, I = int register, N = num register
aCiD2 ah, that makes sense
pmichaud s = string value or string register, i = int value or int register, n = num value or num register 15:06
aCiD2 So my op takes a single string, so I'd use getprotobyname__is (returns an int, takes a string)
pmichaud uppercase I
aCiD2 ok
pmichaud because the result has to go into a register
aCiD2 makse sense. lets see if this compiles 15:07
pmichaud you may be better off starting out by creating a separate module for sockets than trying to bundle it into rakudo
aCiD2 why's that?
pmichaud because I'm not sure what part of sockets will end up being 'core' 15:08
aCiD2 Hm, this was core perl 5, so I assumed it would be core perl 6
it's also part of core python. but of course, maybe bad assumptions
pmichaud at one time would've totally agreed with that, but enough things have changed between p5 and p6 that I'm no longer sure such assumptions are valid 15:09
15:09 cognominal joined
aCiD2 nods 15:09
pmichaud anyway, if done as a module, it's easier to prototype and experiment
because you aren't constantly recompiling rakudo :)
aCiD2 I implemented the getprotobyname as part of core parrot - do you think that was the right choice?
pmichaud that'd have to be a question for the parrot folks (more) 15:10
in general, they've been moving away from specialized opcodes and towards methods on objects
aCiD2 yea, I've given patches for both :)
pmichaud so they might want to see it as a PMC method instead of an opcode
aCiD2 I did it once as a pmc method, and once as a dynamic op code
pmichaud I'd think the PMC method would be more likely
aCiD2 it seems wasteful as a pmc method though, because you end up creating a socket object then throwing it away straight away 15:11
(well, that's how I'm assuming the generated perl 6 pir would look)
snarkyboojum pmichaud: this gets star distribution building working on OS X gist.github.com/503392 - not sure if you want to use it or do something different for OS X 15:14
15:15 kiffin joined
pmichaud snarkyboojum: that looks.... wrong. 15:16
snarkyboojum pmichaud: ok :)
pmichaud is the 'sed' command stripping all of the leading path information?
15:16 kiffin joined
snarkyboojum yeah 15:16
pmichaud yeah, that's wrong. %P simply strips the leading directory prefix of whatever was passed to find 15:17
not all paths
in this case, it simply removes the leading $(DISTDIR) from each filename
snarkyboojum %P strips say 'rakudo-star-2010.07/' for example
pmichaud right 15:18
snarkyboojum because it's used as the path to find
pmichaud oh, wait, I misread
okay, that sed command might work
snarkyboojum so using sed doesn't relate to two as nicely, but has the same effect
the*
pmichaud it doesn't strip "all of the leading path information", it simply strips the first directory
snarkyboojum sorry, yep
either way, that's the problem on OS X, which doesn't support the -printf option for it's find 15:19
pmichaud right
patch looks good to me, I can apply it 15:20
snarkyboojum sweet
15:23 Axius joined
pmichaud oh, I'll need a better patch or pull request -- what's in the gist is malformed (whitespaces/tabs) 15:25
15:25 kiffin left
pmichaud afk for a bit 15:26
15:29 clintongormley joined
snarkyboojum pmichaud: really? seems to apply cleanly here 15:29
or perhaps not :| 15:31
pmichaud snarkyboojum: the gist paste has characters like » in it 15:32
snarkyboojum yeah.. just fixing that :O
pmichaud you can just email me the patch if you want <pmichaud at pobox.com> 15:33
snarkyboojum pmichaud: well I've updated the gist with the correct one, and it applies cleanly now 15:37
pmichaud snarkyboojum: Applied, thanks 15:49
snarkyboojum pmichaud: cheers 15:50
16:11 masonkramer joined 16:22 jaldhar joined
frettled pmichaud: the bonus of fixing it that way, is that it will work with Solaris, AIX, HP-UX, FreeBSD, NetBSD and OpenBSD ;) 16:25
well, it is more likely to, anyway 16:26
isBEKaml frettled: GNUism doesn't work with *BSDs?
pardon my ignorance, I have never used any of the BSDs.
frettled isBEKaml: if it did, it would have worked with MacOS X, which uses BSD tools. 16:27
In general, you can assume that GNUisms work with Linux, period.
isBEKaml frettled: yes, sure. But then, MacOSX forked BSD long ago and are into a separate stream with their own versions.The BSDs have changed a lot since, so I thought MacOSX were a different treat. 16:28
aCiD2 Is there a defined way in Perl 6 to have modules export constants? I'm hacking on my Sockets module and it'd be nice to be able to export things like PROTO_UDP and PROTO_TCP
isBEKaml well, no issues. I'm just curious. 16:29
frettled isBEKaml: Apple is, if anything, rather slow on taking up new NIH-features, just like most large companies.
aCiD2 I can get the constants through parrot, this just seems a little unfriendly
frettled isBEKaml: besides, Unix ecology wise, the fork from BSD isn't that long ago. :) 16:30
isBEKaml frettled: that was my point. They don't seem to take in any new features in the *BSD lines.
s:2nd/in/from/
16:30 Kodi left
frettled isBEKaml: and BSD folks are rather unhappy with the GNU license, so they don't seem to take in many new features from the GNU ecology. 16:30
isBEKaml well, they are an unhappy lot. :) 16:31
frettled Nah, not really.
isBEKaml I got a question for ya all, I'm about to submit the slackbuild scripts to slackbuilds.org. What should I tell them so they can accept it? 16:32
frettled Assumption is the mother of all ...., but who's the father?
isBEKaml vengeance? :P
frettled heh
Well, enough OT babbling from me, I have to run. *idle* 16:33
isBEKaml o/
EDIT:the slackbuild scripts of Rakudo and parrot.
16:35 tommyd joined 16:39 tylercurtis joined
tylercurtis Good morning, #perl6. 16:39
aCiD2 GOOD MORRRRNNNNNNNING TYLERCURTISS!! </radio-voice> 16:41
Revolve_ mornin
wait, where are you, for it to be morning?
if ya dont mind me asking 16:42
tylercurtis Tennessee, though 'tis only barely morning.
aCiD2 west coast?
Revolve_ cool
<-- manc, uk
19:42
16:43 tadzik joined
aCiD2 Revolve_: then why aren't you going to nw.pm meetups! 16:43
snarkyboojum 2:43 the next day here :)
aCiD2 at least I never saw you there when I was in lancaster
Revolve_ they have those here?
I was looking just the other day for stuff like linux user groups and perl users meetups in the area
aCiD2 Revolve_: yes, nw.pm is fantastic in my opinion
Revolve_ okay. that's in lancaster? 16:44
aCiD2 it's the north, i'll give you the link
Revolve_ I don't drive but I've no problem using trains
thanks :D
aCiD2 northwestengland.pm.org/
Revolve_ yeah man
aCiD2 We had socials in Manchester, Lancaster and Liverpool while I was at uni there
Revolve_ this is the one I was looking at
cool
can get to liverpool for a fiver, not including bus far to the station
16:44 pheuter joined
Revolve_ fare* 16:45
aCiD2 cool, well manchester seems to be most popular, so you're in luck there :)
are you at uni or working there?
Revolve_ neither. 18, getting into college after my first placement ended up being too far. looking for work in the meantime :) 16:46
aCiD2 oh cool, would have been nice having someone my age there :P mst goes sometimes who is probably closest to my age, I think
aCiD2 needs to get his ass down to some birmingham meetups now
Revolve_ how old are you guys?
lol
aCiD2 well I'm 21 and 5 days :P the rest are between 25-40 I think 16:47
16:48 justatheory joined
aCiD2 (if you meant nw.pm, i have no idea about this channel) 16:48
Revolve_ yeah, nw.pm 16:49
ohwait
it's not 19:42. it is where my shell provider is. but that's two hours ahead of where I am. 16:50
squeeky ooh. nobody has made a brew formula for rakudo? I might have a crack at this. 16:51
16:52 macroron joined
aCiD2 squeeky++ 16:53
16:53 isBEKaml joined
squeeky It'll be utter irony - writing Ruby that runs Perl 5 to install Perl 6. 16:53
isBEKaml slackbuilds.org/pending/ #slackbuild scripts submitted \\o/. Will have to wait until they(admins) come back. 17:00
Su-Shee: ^^
squeeky could someone confirm the md5 of the tar.gz on github is 3be6d0f4f9d8d1143b6c25768e8a7342 ? 17:02
isBEKaml which one?
squeeky cloud.github.com/downloads/rakudo/s....07.tar.gz
pmichaud pmichaud@orange:~$ md5sum rakudo-star-2010.07.tar.gz 17:03
3be6d0f4f9d8d1143b6c25768e8a7342 rakudo-star-2010.07.tar.gz
squeeky thanks.
17:11 pheuter joined 17:12 pascal__ left, wamba joined
drrho g'arvo #perl6 ! 17:19
17:19 kid51 joined
drrho quick question Re: blizkost: is there a list of things which work and which don't ? 17:19
17:20 Ross joined
drrho amazingly enough, a large package of mine works out of the box with blizkost. 17:20
these things make me always nervous ;-) 17:21
edgar Hi all, I translated the state monad to perl6. But it is not very handy without do notation. Has anyone an idea how to implement such thing?
just a hint is fine. I am relative new to perl6 17:22
so I haven't got a clue where to look.
Oh, this will help probably: en.wikibooks.org/wiki/Haskell/do_Notation 17:24
17:26 supernovus joined, Axius joined
supernovus Is there a way to make Hashes interpolate in regexes? 17:27
phenny supernovus: 12:37Z <masak> tell supernovus that in github.com/supernovus/SCGI/blob/master/README, you need a comma on line 3 and 'return $response' last. x3nU++ discovered this.
supernovus x3nU: Thanks for reporting the typos! Now fixed. 17:31
ingy greetings 17:33
tadzik drrho: I don't think so. Maybe it's time to create one? :)
ingy: o/
drrho: is there any thing that you know that doesn't work with Blizkost? 17:34
lue hello o/
isBEKaml hello, ingy!
drrho tadzik: no, everything I tried seemed to work!
tadzik drrho: so we can maybe assume that everything works, and anything that fails is considered a bug and must be vaporized :)\\ 17:35
tylercurtis tadzik: I know that "use Data::Dumper:from<perl5>; Dumper(1.5);" doesn't work. Same for arrays and hashes. Didn't try with non-builtin classes.
drrho of course, I consider to move stuff into Perl6, so I am also looking for a migration path.
;-)
tadzik tylercurtis: maybe. But .perl is the way, Data::Dumper is obsolete I'd say 17:36
we also have YAML::Dumper
ingy star: use YAML; say dump [1..5] 17:37
p6eval star 2010.07: OUTPUT«---␤- 1␤- 2␤- 3␤- 4␤- 5␤...␤␤»
tadzik :
:)
tylercurtis tadzik: I know. That was just the most obvious Perl 5 core module I knew of to test marshalling to Perl 5.
x3nU supernovus: np :)
drrho From my first experiments, I have to say that this Rakudo thingie rocks. 17:39
is there also a way to integrate C code? 17:40
gfldex yes, but it's rather indirect
drrho is it via parrot? 17:41
supernovus x3nU: I don't think anyone had noticed as prior to Rakudo Star, I don't think many people looked at the SCGI library. But if you're doing web apps with Perl 6, SCGI is much much faster than CGI.
17:41 pheuter_ joined
gfldex drrho: yes 17:42
tylercurtis drrho: check out Zavolaj.
drrho googling 17:43
tylercurtis drrho: github.com/jnthn/zavolaj/
drrho: it's included in Rakudo * iirc.
drrho just found it, yes.
x3nU supernovus: i wished to play with SCGI but mod_scgi doesn't work properly for me 17:45
;(
supernovus x3nU: Hmm, that's not cool. There is also a proxy_scgi module that comes with Apache 2.2. I haven't tried that one yet, but plan to try it out shortly (once I get my CMS working under the latest Rakudo, which may take some time.) 17:48
mod_proxy_scgi
17:48 eternaleye joined
supernovus httpd.apache.org/docs/2.2/mod/mod_p..._scgi.html 17:49
drrho looking at NativeCall.pm: do I understand it correctly, that I can write PIR code inside Perl6? 17:51
tylercurtis drrho: in Rakudo, yes, but please don't unless you really absolutely need to. Otherwise, your code will never be portable to other implementations. Also, PIR is ugly. :) 17:52
drrho I see. 17:53
Still sexy :-) 17:55
17:56 am0c joined 18:08 jedai joined 18:13 ambs joined
ambs bow 18:13
18:25 justatheory joined
supernovus So, anyone interested in a FastCGI library for Perl 6? It's a lot more complex than SCGI, but I was thinking of making it my next pet project. 18:26
Su-Shee WHAT QUESTION IS THAT? OF COURSE MAN. :) 18:27
how else is going any web framework going to work?
supernovus Su-Shee: Well, you could try using plain old CGI... but I'll tell ya, it's horrifically painful (which is why I wrote the SCGI library in the first place...) 18:28
m6locks ummm
Su-Shee scgi and fastcgi are the way to go. web server like nginx doesn't even support cgi anymore.
18:28 sekimura1 joined
m6locks is that buildall.pl supposed to build the whole thing? 18:28
jfried supernovus, a fastcgi library would be very cool :) 18:29
m6locks it says
Installing to '/Users/thebbo/star/build/build'
Can't chdir to rakudo: No such file or directory at buildall.pl line 15.
18:29 cbk joined
m6locks so it's not working, but I think I got the makefile 18:30
sorear good * #perl6
tadzik o/ 18:32
18:35 Axius joined, pheuter_ left
supernovus Well, FCGI is officially my next personal project. After that, maybe PSGI support in masak++'s Web framework (yeah, I'm in a webish mood lately.) 18:36
perigrin you know if you did PSGI support first you could in theory get FCGI support for free 18:38
18:41 am0c joined 18:42 kiffin joined
tylercurtis If I make a subset of OpaquePointer to make the signature of the functions clearer to read, will Zavolaj still generate the correct NCI signatures? 18:44
e.g., "subset Foo of OpaquePointer; sub Something() returns Foo is native('libsomething') {...}"? 18:45
japhb Who picks the content for the Rakudo CafePress store? 18:47
squeeky ..... cafepress? argh. 18:50
x3nU supernovus: ok i tested and it seems that mod_proxy_scgi works fine althrough i had to compile itself because there's no win32 version
it myself*
thanks for help 18:51
Tene pmichaud: the can_handle method of an exception handler is what is used to determine if the EH wants to handle a given exception or not.
x3nU hmm... i think that i should publish mod_proxy_scgi binaries on my site
supernovus x3nU: Great to hear! I may change the README to suggest mod_proxy_scgi instead of mod_scgi. 18:52
18:55 KenGuru joined 18:56 kiffin left 18:58 snearch joined
Casan supernovus: FCGI or indeed PSGI for Rakudo/Parrot/Perl6 would be xmas all over again. 19:03
19:06 cono joined
supernovus Casan: Well, I'm going to work on FCGI first. Then if it works well, it can be one of the available backends for masak++'s Web framework, which intends to include a PSGI/Plack compatible interface (hopefully with CGI, SCGI, FCGI, mod_perl6 and HTTP::Daemon backends.) 19:06
Casan supernovus: bravo. will follow your work with great interest. 19:08
19:09 jaldhar joined
supernovus Casan: If you're doing any web stuff with Perl 6 at the moment, I'd recommend trying the SCGI library available in proto/pls. It's WAY faster than standard CGI (mostly because your not loading a new rakudo on every page request...) 19:10
19:11 Ross joined, melte joined 19:15 alanstorm joined
tadzik supernovus: mberends++ is writing http::server::simple::psgi on github, you might be interested 19:16
alanstorm Hello all, I'm not sure if this is the right venue for this. I'm not involved in the perl6/raduko project. I'm just trying to download the recent raduko-star release and monkey around with perl6. I'm having trouble getting it to compile/install on my OS X 10.6 laptop. 19:18
When I run "perl Configure.pl --gen-parrot", it chugs along for a bit, but then fails with
Reading configuration information from install/bin/parrot_config ... ===SORRY!=== Parrot revision r48225 required (currently r0) 19:19
Is there anyone here who could help me figure out what's going on, or point me towards somewhere where I could help myself?
moritz_ alanstorm: want to come to the lobby?
alanstorm moritz_: sure, which channel is that?
moritz_ alanstorm: sorry, I misread - thought you meant the YAPC venue 19:20
moritz_ confused
supernovus tadzik: Awesome, I'm sure that will integrate with Web well. Perl 6 will make Web applications Fun again! :-)
moritz_ alanstorm: do you have any parrot installed?
whee moritz_: I want what you're on ;D
alanstorm moritz_: I'm not sure? I downloaded the rakudo-star-2010.07-0-gac1dc94.tar.gz tar ball that was recently released 19:21
moritz_ alanstorm: did you ever install a parrot before? 19:22
gfldex :D
supernovus --gen-parrot should install parrot automatically...
alanstorm used that to make a distribution with make VERSION=2010.07
and then tried to Configure.pl that distribution, and that's when I got teh above Sorry
moritz_: no, I've never installed "a parrot" before 19:23
whee alanstorm: do you have svn?
alanstorm I'm a little unclear on the terminology, my apologies for that
whee: yes, I have svn
m6locks can't build the whole thing with buildall.pl, i was gonna ask if there was some sort of build all those modules and stuff type of command or script
gfldex alanstorm: the error you get means that parrot wasnt build correctly
parrot is the runtime environment rakudo needs to do anything 19:24
alanstorm gfldex: ok, that makes sense
gfldex so that has to be build before
the real error message from the failed parrot build should be a few lines about the sorry
moritz_ my internet connection is too slow to reasonably chat 19:25
gfldex o.0
alanstorm there's no error message above the sorry, just a
Finished install_dev_files.pl 19:26
19:26 offerkaye joined
alanstorm Reading configuration information from parrot-2.6.0/parrot_config ... 19:26
===SORRY!===
and above that a listing of files 19:27
gfldex are you afraid of git?
alanstorm heh, not afraid, but a little wary.
offerkaye Hi
moritz_ hm, looks like the version compaarisn code is a bit fragile 19:28
m6locks can't build the whole thing with buildall.pl, i was gonna ask if there was some sort of build all those modules and stuff type of command or script
gfldex lets put git aside for a bit then
have a look if you can find a parrot_install dir
and check if you can find the binary parrot_install/bin/parrot 19:29
if you dont, you aint got no parrot
moritz_ not quite
R* installs parrot into build/ and not parrot_install 19:30
alanstorm I've got a
./install/bin/parrot
moritz_ what do you get when you run ./install/bin/parrot_config revision ?
alanstorm $ ./install/bin/parrot_config revision 19:31
0
moritz_ and if you query it for 'versino' instead of 'revision'?
sorry, version 19:32
alanstorm moritz_: liks this?
./install/bin/parrot_config version
no such key: 'version'
moritz_ hm
try VERSIION please 19:33
moritz_ a bit unconcentrated, sorry
alanstorm ah, that's better
./install/bin/parrot_config VERSION 2.6.0
./install/bin/parrot_config VERSION
2.6.0
moritz_ seems like your parrot built just fine
alanstorm heh, no sorries mortiz_, I'm the newbie forcing myself on all of you
sorries=worries 19:34
if it's any help, I found this IRC log
irclog.perlgeek.de/parrot/2009-05-11
around 13:51 someone seems to have a similar problem
gfldex there where osx build problems discussed a few hours back
2009-05-11 is quite some time ago 19:35
alanstorm yeah, quite a long time ago, but it's one of the few things that came up
are there IRC logs I can checkout to review the discussion from a few hours ago you mentioned? 19:36
tylercurtis alanstorm: did you download the rakudo-star-2010.07.tar.gz tarball from github.com/rakudo/star/downloads or the tgz under the 2010.07 tag on that page? If the latter, that may be causing you some problems, since, iiuc, that git repo holds the scripts used to create the distribution rather than the distribution itself.
19:38 m`` joined
m`` hello perlers 19:38
offerkaye Hi m``
pmichaud I'm thinking of getting rid of that tag.
alanstorm tylercurtis: I downloaded a file from that gitrepo, I can't remember which link, but the file was named
rakudo-star-2010.07-0-gac1dc94.tar.gz
pmichaud that's the wrong file. 19:39
gfldex alanstorm: if you dont get happy with the (all time first, not so well tested) R* release you can try to fetch rakudo via git, as described on www.rakudo.org/how-to-get-rakudo
alanstorm then from that folder ran make VERSION=2010.07
which made a distribution
pmichaud right. that's not the star tarball.
alanstorm ah ha, got it 19:40
m`` github.com/downloads/rakudo/star/ra....07.tar.gz
pmichaud I'm going to delete that tag, it's just too confusing.
alanstorm pmichaud: thanks, I'll give that a try 19:41
gfldex: and thanks for the how to link, that should prove equally useful
one last question: Is there a document/resource that covers "Perl 6 for people who haven't been paying attention, but wrote perl programs five to six years ago and want to write perl6 programs" 19:42
m`` five to six years ago?...
moritz_ btw I think I found a specbug wrt smart matching 19:43
m`` I really dont think so
alanstorm I get a little confused by what parrot, perl, raduko, etc. all are, which makes researching problems on my own difficult
m`` alanstorm: but you might look at "Perl6 for perl5 programmer"
gfldex alanstorm: perlgeek.de/en/article/5-to-6
alanstorm: www.perl6.org/documentation/
moritz_ alanstorm: perlgeek.de/en/article/5-to-6 for the language
m`` Parrot - virtual machinve build for perl6 but can run other languages
Rakudo - perl6 for parrot vm
"Perl 6 is anything that passes the official test suite." 19:44
tylercurtis alanstorm: the book included in Rakudo * as doc/UsingPerl6-draft.pdf looks helpful, although I haven't gotten around to reading all of it yet.
gfldex i have one too: Perl6 != Perl5+1
m`` So there is not "official" perl6 version
gfldex: same
unlike Perl5/Python/Ruby etc
gfldex is UsingPerl6-draft.pdf somewhere on the interwebs? 19:45
alanstorm that all looks like a good starting point 19:46
thanks again for all your help
gfldex yw
m`` github.com/rakudo/star/blob/master/...-draft.pdf
moritz_ github.com/perl6/book/downloads the latest download is basically the same
pmichaud tag deleted 19:47
moritz_ what should say 'foo' ~~ .uc print?
FOO or False?
tylercurtis moritz_: what's in $_? 19:48
offerkaye star: say ('foo' ~~ .uc)
p6eval star 2010.07: OUTPUT«0␤»
tylercurtis or does ~~ topicalize?
m`` perl6: say 'foo' ~~ .uc
pmichaud ~~ topicalizes
moritz_ it topicalizes
p6eval pugs: OUTPUT«1␤»
..rakudo 9b6189: OUTPUT«0␤»
moritz_ let's extend the question 19:49
what should say ('foo' ~~ /o/) print?
sorry
what should say ('foo' ~~ m/o/) print? 19:50
pmichaud 'o' followed by a newline.
offerkaye pmichaud: I just sent an email to perl6-lang about Perl6 user documentation. I mentioned you, hope you don't mind ;)
pmichaud offerkaye: no problem
moritz_ pmichaud: it can't, by current spec
pmichaud moritz_: ....because? 19:51
moritz_ because m/o/ is an immediate call
so there's a Match object on the RHS
pmichaud at one time, smart match overrode that particular item.
szabgab rakudo: say %*{PATH}
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤Non-declarative sigil is missing its name at line 22, near "%*{PATH}"␤»
moritz_ on which we call .ACCEPTS
szabgab rakudo: say %*ENV{PATH}
moritz_ (by spec, if I interpret it correctly)
p6eval rakudo 9b6189: OUTPUT«Could not find sub &PATH␤ in main program body at line 22:/tmp/dgotIgadbY␤»
pmichaud i.e., there was a time when smart match would change the way that the rhs was interpreted.
still is, afaik.
szabgab this thing tripped me in the morning to my emberassment 19:52
moritz_ it just topicalices
avuserow rakudo: say %*ENV<PATH>
p6eval rakudo 9b6189: OUTPUT«/bin:/usr/bin:/usr/local/bin␤»
szabgab but I wonder if that case could have a better error message for the pure perl 5 programmers?
moritz_ the syntactic cases seem to be gone
pmichaud moritz_: it topicalizes before the immediate match occurs.
moritz_ from the smart-match table
szabgab so recognize when bare-word is used as a %hash{subscript}
moritz_ pmichaud: yes, and then calls .ACCEPTS on the result 19:53
szabgab and tell the user s/he is not on perl 5
pmichaud moritz_: ah, I see.
moritz_: that's a question TimToady will have to answer.
moritz_ there's no rule that preveneents that call
aye
pmichaud at one point, it was very syntactic 19:54
moritz_ maybe "if $_ is used on the RHS" or so
I'm writing this up as an e-mail
pmichaud fair enough
szabgab pmichaud: pls look at my question above, about the error message
moritz_ I discovered it when I implemented it correctly 19:55
:-)
ie according to current spec
pmichaud std: say %*ENV{PATH}
p6eval std 31886: OUTPUT«[31m===[0mSORRY![31m===[0m␤Undeclared name:␤ 'PATH' used at line 1␤Check failed␤FAILED 00:01 116m␤»
pmichaud szabgab: that's about the best we'd ever be able to do.... in order to do that, we'll have to have a way at compile-time to know all of the lexically defined subs
(which we don't have in Rakudo yet) 19:56
I suppose we could say that barewords aren't allowed in {...}
tylercurtis moritz_: I don't think "$foo ~~ m/o/" is your best example. After all, it's trivial to do that as "$foo ~~ /o/". "$foo ~~ .uc" is a much more interesting example.
szabgab can't you catch that exception and then look at the code to see if came from such a contruct? 19:57
pmichaud szabgab: we don't have quite that much introspection available, no.
we might be able to do that someday.
moritz_ tylercurtis: it's the one example that makes it obvious that, no matter how much bike-shedding you do it, it's not what we want
szabgab ok, I was just hoping :)
moritz_ even the spec uses $wtr ~~ m/.../ in quite many places all over S05 19:58
pmichaud moritz_: at one point, the philosophy was that special cases were part of what made smart matching 'smart'
i.e., it didn't all involve .ACCEPTS
tylercurtis moritz_: ah, true. 19:59
pmichaud I don't know what TimToady's current leanings are about ~~
tadzik star: say %*ENV<PATH>
p6eval star 2010.07: OUTPUT«/bin:/usr/bin:/usr/local/bin␤»
pmichaud szabgab: fwiw, I think it might be entirely reasonable to say that .{bareword} should always produce at least a compile-time warning, if not be an error outright. There are other ways to express the meaning of .{bareword} if that's what is intended 20:00
one possible exception might be constants, however.
i.e., .{pi} probably should be allowed.
moritz_ I think it's reasonable for .{not-a-symbol-in-scope} to eventually produce a customized error messages 20:01
OTOH I don't want to clutter Perl 6
with p5 backcompat error messages
for perfectly reasonable code 20:02
like sub a { 'x' }; %h{a}
pmichaud yes, but %h<a> and %h{a} are so visually similar that I think it's reasonable to say the latter should be written as %h{a()} 20:03
regardless of the p5 history
moritz_ finds the difference between <> and {} quite striking
much more than '...' and "..." 20:04
or () and {}, for that matter
gfldex rakudo: say "foo {call_a_sub}"; say %h{call_a_sub()};
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤Symbol '%h' not predeclared in <anonymous> (/tmp/1iS_Yjl4TS:22)␤»
gfldex i would prefere to have call_a_sub the same in both cases 20:05
pmichaud gfldex: fair point
afk for a bit 20:06
btw, I really wish I could figure out how to get the book.pdf to build on my system. it'd be nice to work on it on the plane.
20:07 ashleydev joined
pmichaud afk 20:08
moritz_ pmichaud: try to build the a4 version 20:09
szabgab moritz_: stackoverflow.com/questions/3383339...ll-on-os-x < typo "raduko"
moritz_ szabgab: corrected (though not my typo) 20:12
m6locks hmm gotta read that one 20:13
i've had some problems too
20:15 masonkramer joined
jnthn back 20:16
szabgab moritz_: I know, thanks
moritz_ jnthn: when will you arrive in Pisa? 20:17
jnthn moritz_: You travelling to YAPC tomorrow too? :-)
moritz_ no, I'm already there
szabgab where it the perl 6 hackathon in pisa?
jnthn moritz_: Oh!
moritz_: You at the hotel?
moritz_: er, the conference hotel?
szabgab I'll arrive tomorrow night and would like to come by
jnthn Or staying elsewhere?
moritz_ jnthn: yes, at the venue 20:18
szabgab on tuesday
jnthn moritz_: OK, I'm staying there too.
moritz_ and tehre's free wifi in he lobby, so that's where you can find me most of the time :-)
jnthn moritz_: Depending on how well I can do in terms of airport -> train, I hope to get to Pisa railway station at either 14:30 or 15:00
moritz_: But before I move on from there, I want to buy my train tickets for leaving Italy on the Saturday (I'm going back home on train, via Switzerland) 20:19
moritz_ jnthn: ah, i might be downtown with some other German pm folks at that time, but should be back in the evening, and available on the day after
jnthn moritz_: OK 20:20
moritz_: I think Pm expects to get to the hotel at 8pm on Mon
moritz_ \\o/
jnthn moritz_: I plan to meet him there and then nom.
moritz_: Good to know there is free wifi in the lobby. 20:21
moritz_: Is it a hackathonable area, or do you think getting some meeting room would be better?
moritz_ jnthn: should work
jnthn moritz_: OK 20:22
moritz_ -> gone 20:24
pmichaud moritz_: I did try to build the A4 version -- same error.
phenny: tell moritz_ I did try to build the A4 version -- same error.
phenny pmichaud: I'll pass that on when moritz_ is around.
pmichaud (well, same error except with .A4 instead of .Letter) 20:25
moritz_ pmichaud: I fear I can't help much, I can build it only on my home computer, on my laptop the texlive installation is too old
phenny moritz_: 20:24Z <pmichaud> tell moritz_ I did try to build the A4 version -- same error.
pmichaud moritz_: okay. Maybe PerlJam++ can help out -- he's been able to get it to build on his systems.
wow, my flight to LHR tonight looks rather empty
I count about 35 empty seats at the moment 20:26
(based on the online checkin)
at least 6 rows with nobody in them
jnthn Wow
Plenty of room to space out. :-)
pmichaud well, they may get taken before then 20:27
jnthn Troo
20:27 Ken-Guru joined
jnthn No idea how mine will be 20:27
But it's < 2 hour flight.
pmichaud so far I'm not giving up my aisle seat. I'll check it again when I'm ready to check in (at the airpot)
*airport
been watching it since yesterday -- lots of seats moving around
jnthn Probably about 7 - 7.5 horus until I leave for the airport from here. :-) 20:28
Tene debates between working on exceptions and gaming.
pmichaud work on exceptional gaming
jnthn work on gaming exceptions 20:29
gfldex what game would it be?
jnthn It's one where you throw stuff and people die.
20:29 eternaleye joined
pmichaud unless one of the people catch whatever is thrown. in that case, they can't die, or else we get a segmentation fault and the universe halts. 20:31
cxreg jnthn: would that be horseshoes, or handgrenades? 20:32
gfldex i think we lost Tene to the game already :) 20:33
jnthn cxreg: In the Perl 6 version, it's mugs.
gfldex and the enemies are walls :->
whereby the walls are tempted to throw some mugs too 20:34
Tene rakudo: die Mug.new() for 1..20;
p6eval rakudo 9b6189: OUTPUT«Could not find sub &Mug␤ in main program body at line 22:/tmp/2mHmB483_j␤»
cxreg heh 20:35
class Mug does Motivation 20:36
20:37 eternaleye_ joined 20:50 tommyd joined
x3nU w/in 3 21:05
oops
21:09 Italian_Plumber joined 21:10 synth joined 21:15 PhatEddy joined 21:17 KenGuru joined
PhatEddy rakudo: class fee { has $.fie is rw; method fie { return $!fie ~ "x" } }; my $a = fee.new; $a.fie = "fum"; say $a.fie 21:19
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤A method named 'fie' already exists in class 'fee'. It may have been supplied by a role.␤»
PhatEddy Can you override a method supplied by a role without subclassing/inheritance?
21:21 plobsing joined
tylercurtis PhatEddy: your issue there is that you declare an accessor fie for $!fie and then a method named fie. 21:22
21:22 masak joined
jnthn panicitsmasak! 21:23
masak ahoy, #pe...
masak panics
tylercurtis Although it seems like you should be able to override the accessor...
masak actually, the Yapsi refactor is going well. I probably won't make it before midnight, but I expect to be finished not long after.
the new code is smaller, cuter, and generally more p6ish. :) 21:24
PhatEddy From S12: You may write your own accessors to override any or all of the autogenerated ones. ?? 21:25
masak I think I've finally figured out how to profitably traverse the $/ structure.
jnthn You figured out step 2?! 21:26
tylercurtis rakudo: class fee { has $.fie; method fie { $!fie ~ "x" } }; my $a = fee.new(:fie<fum>); say $a.fie;
jnthn PhatEddy: For now, if you want to write your own accessor, declare just the attr, not an attr and an accessor.
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤A method named 'fie' already exists in class 'fee'. It may have been supplied by a role.␤» 21:27
jnthn has in, has $!fie; method fie
tylercurtis jnthn: If $.fie is rw, should one be able to the reading accessor but not the writing accessor when overriding accessors is implemented? 21:29
jnthn tylercurtis: I'm not sure how that'd work. How would the dispatcher differentiate between them? 21:31
tylercurtis jnthn: that's what I thought, but I wasn't sure. It would make sense with a attribute trait that generated chainable writers, but not with is rw. 21:34
jnthn tylercurtis: That'll be do-able in a module (soonish). 21:35
tylercurtis jnthn: ooh. custom attribute traits? 21:37
jnthn Yes. :-) 21:39
tylercurtis jnthn: on an unrelated note, does Zavolaj's signature introspection work properly if I use a subset of OpaquePointer/Int/Str/etc. in the signature. For example, "subset Foo of OpaquePointer; sub DoSomething(Foo $f) returns Int is native('libb') { ... }"?
jnthn Probably not. 21:40
Why'd you want to do that, ooc?
Oh, actually
Maybe it should
I think it smart-matches
And Foo ~~ OpaquePointer
tadzik oh, by the way
tylercurtis jnthn: to make it look prettier.
jnthn So probably it does work.
Ah. :-)
tadzik jnthn: in Zavolaj, the void* or struct pointers are represented by OpaquePointer, right?
jnthn Right.
For now. 21:41
tadzik any way to get some fields out of these structs
?
tylercurtis jnthn: although "subset Void of Mu;" wouldn't work since you have to use === for that. :)
jnthn tadzik: No
tadzik I was planning to write some nice OO interface to loudmouth, but I'll have to look into those GError structs for proper error handling
jnthn tadzik: OpaquePointer will go away in the future.
tadzik: Well, it might not
tadzik jnthn: being replaced by?
jnthn tadzik: But most uses of it will be mappable to a compact struct. 21:42
(as per S09)
PhatEddy Sorry wasn't sure it was an issue but it's known - rt.perl.org/rt3/Ticket/Display.html?id=68996 - rakudo seems to behave better now though ...
masak jnthn: step 2 isn't anything complicated. it's actually term:<...>, not something elided as people usually thing. as you execute the step, you get a warning, but it's catchable. :P 21:43
jnthn tadzik: And that will let you get at the various fields.
tadzik cute
jnthn masak: :P :P
Cute is what we aim for.
tylercurtis jnthn: another couple of Zavolaj questions. Is there a way to tell if an OpaquePointer is NULL? As well, if I do "sub Something(Str $s) is native('blabla') {...}" is there a way to pass NULL as the C string? 21:44
cxreg pir::null__P() will give you a C null 21:45
21:45 xinming joined
masak / 'if' <expression> <block> [ 'else' <else=block> ]? /; will $<block> contain a Match or an Array of Match? 21:46
jnthn tylercurtis: I think it just maps it to an undef
tylercurtis: Or an undef will do the right thing
tylercurtis: I'd have to check the code to be sure.
cxreg (undef? Mu?)
jnthn masak: Match 21:47
masak jnthn: then I have a rakudobug.
jnthn masak: Oh, wait
cxreg: Maybe Mu yeah
masak: <else=block> is not the same as <else=.block> now
masak: Used to be. :-)
masak which is which?
oh!
jnthn masak: <else=.block> would not capture
masak right. 21:48
jnthn well, it'd capture into else
not block
masak so I should add the dot and then $<block> should contain a Match?
jnthn That'd make sense to me. :-)
What the grammar engine thinks is another matter. :-)
masak :)
jnthn I'd pipe it through pmichaud to be sure. :-)
masak no time... 21:49
cxreg it's odd asking questions about a programming language, for which nobody actually knows or has an answer
this truly is Perl Trek
tylercurtis jnthn: I'll try that at some point. I've been diverted from the project that inspired these questions by realizing that even with Zavolaj, wrapping sizable libraries is tedious. So I need to write a C parser with which to parse header files and generate Zavolaj-based wrapper modules for the functions in the headers. :)
jnthn cxreg: It's not that nobody has one, just nobody here right now. :-)
cxreg :)
21:49 PhatEddy left
masak cxreg: what jnthn said. I'm not sure I'm too fond of the term 'Perl Trek'... 21:49
jnthn Keep in mind that when it comes to people knowledgable about grammars, I'm a little way down the list. :-) 21:50
cxreg masak: it's healthy to be able to laugh at onesself
Perl 6 Is Srs Bsns 21:51
gfldex tylercurtis: it would be nice if you could shout if you got that working
pmichaud 21:46 <masak> / 'if' <expression> <block> [ 'else' <else=block> ]? /; will $<block> contain a Match or an Array of Match?
masak cxreg: sure, it's not that. I'm just not sure I agree that what we're doing is a 'trek'.
pmichaud array of match
masak pmichaud: and <else=.block> fixes it?
jnthn pmichaud: and if it was <else=.block> ?
pmichaud it'd be just a single match
jnthn Answers! We all want them!
masak \\o/
jnthn \\o/
cxreg this channel cracks me up 21:52
pmichaud it's a lot more cracked up than it used to be 21:53
jnthn <innocent look>
masak keep cracking up, lurkers!
jnthn heh heh 21:54
From p6l: "I would like for what Carl said to be true, but TSa seems to be contradicting it."
gfldex tylercurtis: trac.parrot.org/languages/browser/c...src/c99.pg might be helpful
jnthn Can I reply, "I'd trust the guy who's writing Perl 6 code every day." :-)
pmichaud I hate to say it, but I tend to simply skip TSa's messages. 21:55
oops, forgot the :) 21:56
jnthn I tend to simply skip much of p6l. :/
I almost might as well unsubscribe for the distraction/information factor I get from it.
pmichaud yes, I've been skipping p6l a fair bit also
jnthn Apart from every so often there's something interesting.
masak thanks for the vote of confidence. I'll read the messages after the Yapsi release. :P
tylercurtis gfldex: thanks. I'd forgotten about that. 22:05
22:07 nimiezko joined
masak while I finish up the last things, recall the challenge I posed last night: filling in the blanks of github.com/masak/yapsi/blob/master/...ce/2010.08 :) 22:09
the problems I run into at this late hour of refactoring are exquisite. I only have time to glance at them quickly and then move on. 22:17
need to change location. will be back shortly. 22:18
shirtless Are there any UI bindings for rakudo?
tadzik blizkost has tk example :) 22:19
maybe Gtk will work too
lue hello o/
tadzik o/
lue
.oO(For some reason, I feel like fixing broken dwimmery without knowing if there is any.)
 22:20
shirtless thats kinda cool
for now I guess i'll stick to HTML
tadzik what are you writing? 22:21
22:21 am0c joined
lue once (rakudo|parrot) allows using foreign code, you can use GTK, Qt, etc. 22:22
rakudo: use libc:from<C>; 22:24
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤"load_language" couldn't find a compiler module for the language 'C'␤» 22:25
lue not yet :/
tylercurtis lue: there's some support for this already. Go look at Zavolaj. And smile. Then try to actually manually wrap any sizable library with it and realize that no matter how pretty an individual native function declaration is in, wrapping a whole library by hand is monstrous. Then help me write a C parser in order to generate wrappers for native libraries. :) 22:26
22:29 masak joined
lue "How hard could that C parser possibly be?" 22:29
masak hard as diamond.
lue
.oO(cue jinx)
lue ponders a C -> P6 converter, but keeps these crazy thoughts to himself. Dang it. 22:31
tylercurtis lue: trac.parrot.org/languages/browser/c99/trunk 22:32
lue: not quite the same, but it is C99 on top of Parrot, though I don't know how complete it is.
avuserow int main(int argc, char* argv<::>) <% printf("Yeah, how hard could it be?\\n"); %>
though I'd argue that any library using digraphs is not worth wrapping directly :) 22:33
jnthn masak: gist.github.com/503838 22:34
lue
.oO(How hard can it be to turn your GameBoy Advance into an oscilloscope?)
avuserow well, considering the distance between me and my GBA is bigger than walking distance, pretty difficult 22:35
masak jnthn: wow, cool. :)
jnthn masak: ETOOSIC ;-)
masak jnthn: bit of a wonky sentence close to 'a manure heap that I want to announce'... :)
jnthn Oops. :-)
Maybe use a comma. :-) 22:36
masak it's "Kama Sutra" without the 'r' :)
jnthn I don't know about these things. :P
masak 'ancient' -- heh :)
jnthn Oh yeah, you didn't offer that para up for editing, did you... :P
masak you'll be glad to know that this release will run under Rakudo Star :)
jnthn Oh!
masak I hadn't made up my mind when I drafter the announcement. 22:37
jnthn
.oO( the announcements get drafter and drafter )
lue Karma Sutra? .oO(Very wacky announcement)
masak jnthn: you win my total appreciation of the official/complete paragraph. congratulations. :) 22:38
I find that paragraph increasingly difficult to write. this strikes the right tone.
I... I won't have puke in my release announcement. sorry. :P 22:39
jnthn lue: Yapsi gets increasingly well positioned to penetrate the Perl 6 compiler market.
masak: Aww!
:P
masak: spew? 22:40
lue .oO[ positive temperatures on which temperature scale? Hopefully you're not talking about Kelvin :) ]
masak jnthn: you make me realize I'm a tad more conservative in my humour than I thought :)
jnthn Bwahaha! 22:41
masak: You called it SIC. :P
lue
.oO(can't…stop…pun…) Sounds like one sic compiler.
masak jnthn: true. though until now I haven't related that to puking.
jnthn: I'm not a native. :)
jnthn masak: Oh. I did the first time I saw it.
Lucky you. :P 22:42
.oO( No SIC in .Net, it just gets IL )
masak I have three failing tests and a strange heisenbug left. I say ship it.
during the refactor, Yapsi lots the ability to detect when variables are used before they're declared. 22:43
jnthn I'm truely sorry for your lots.
masak it's probably not difficult to add it back in. I just don't see it now.
jnthn: :)
jnthn masak: skip is fine - the dugong is hungry again anyway. 22:44
22:45 sekimura1 joined
tylercurtis Is there a magically concise way to say "I want to match this thing, that thing, and this other thing, but you can leave off either the first or last thing, but not both?" 22:46
tylercurtis hopes no.
In regexes, rather.
masak tylercurtis: match and then <?{ ...
}>
kthakore wow
I thought perl6 was quite
then I scrolled down
hehe
doh'
22:47 rbaumer joined
kthakore How can I call a parrot pmc (SDL.pir) from perl6 22:47
jnthn <first_thing> <middley> [<last_thing> || <?{ $<first_thing> }>]
kthakore err rakudo
lue I am continually amazed at all the little things P6 has easy support for.
kthakore lue: me too!
jnthn tylercurtis: ^^ something like that
There may be a neater way but I think that'll work
kthakore can't wait till it gets optimized
lue rakudo: say 0 nand 0; # just a hunch 22:48
p6eval rakudo 9b6189: OUTPUT«===SORRY!===␤Confused at line 22, near "say 0 nand"␤»
tylercurtis jnthn: how would that get to "|| <?{ $<first_thing> }>" if <first_thing> didn't match?
lue rakudo: say 0 !and 0;
jnthn tylercurtis: oh
p6eval rakudo 9b6189: OUTPUT«0␤»
jnthn tylercurtis: sorry, meant a ? after <frist_thing>
tylercurtis lue: rakudo: multi infix:<nand> (Mu $a, Mu $b) { $a !and $b }; say 0 nand 0; # lue 22:49
rakudo: multi infix:<nand> (Mu $a, Mu $b) { $a !and $b }; say 0 nand 0; # lue 22:50
p6eval rakudo 9b6189: OUTPUT«1␤»
kthakore rbaumer: say 0.33333333333333 = 1/3
ops
rakudo: say 0.33333333333333 = 1/3
p6eval rakudo 9b6189: OUTPUT«0.33333333333333␤Cannot modify readonly value␤ in '&infix:<=>' at line 1␤ in main program body at line 22:/tmp/D2XFVcDwOp␤»
jnthn Cannot assign to readonly variable
kthakore rakudo: say 0.33333333333333 == 1/3
p6eval rakudo 9b6189: OUTPUT«0␤»
jnthn damm, I was close!
kthakore but why/
oh ok
yayay! Floating point is super awesome!
jnthn kthakore: Because they're not quite the same ;-) 22:51
kthakore I know
that is why I love you guys
first lang that did it that right
:D
tylercurtis jnthn, masak: thanks. I think I'll just stick with the more obvious but slightly more repetitive explicit way, though.
kthakore: Common Lisp beat us to rationals by quite a while. 22:52
kthakore tylercurtis: really?
kthakore has to try lisp
masak tylercurtis: I probably would, too.
kthakore tylercurtis: is it just common lisp?
tylercurtis: how does it compare in arithmetic calcs to C? 22:53
tylercurtis kthakore: I'm not sure. Some other Lisps might, but I just know for sure that CL has rationals.
kthakore cool
tylercurtis: but perl5 gets it wrong for sure 22:54
tylercurtis Although they call them ratios(with rational being the supertype of ratio and integer).
kthakore so does C if you are not careful
java too
This is the most sense ever made in a sentence"rational being the supertype of ratio and integer"
I really have to see common lisp
btw ... will perl6 have web REPL like learnhaskell.com or learnruby.com ? 22:55
that would be sweet
is anyone doing it?
I can try to halp
masak: jnthn ??
masak kthakore: people are on it, yes. 22:56
tylercurtis kthakore: github.com/moritz/try.rakudo.org
masak it would indeed be sweet.
tylercurtis kthakore: it's not complete yet, though, I don't think.
kthakore moritz_++
tylercurtis kthakore: ash_++, he's done a significant chunk of the work.
kthakore tylercurtis: I will go give it a shot and drop code where I can
tylercurtis: ah ok
masak moritz_ has been instrumental in the project. other people have written the actual HTML/CSS/code.
22:57 robjmj joined, lue joined
kthakore masak: ah ok 22:57
masak: I will talk to one of them to see where to help
they have no tickets 22:58
lue rakudo: my $a = 3*(1/3); my $b = 3/3; say "$a, $b, {$a == $b}"
p6eval rakudo 9b6189: OUTPUT«1, 1, 1␤»
kthakore back to SDL perl lair 22:59
tylercurtis: masak jnthn thanks be back soon
we have release comming soo
soon
lue $a should _technically_ be 0.999999…, but at least it equals $b :)
kthakore lue: true! 23:00
a != b
well in mathematical proofs
tylercurtis lue: 0.9999... is also equivalent to 1.
kthakore lue: actually 3 * 1/3 = 1 23:01
it is like 3/3 * 1 23:02
cxreg right, that's exactly why .999 repeating == 1. it's a number base issue.
masak lue: en.wikipedia.org/wiki/0.999...
lue 1/3 = 0.3…; 0.3… * 3 = 0.9…; there's why 3*(1/3) == 0.9… 23:03
kthakore what would be interesting to see is this: 3 * (1/3 - 1/10**10000 )
masak: how do I do power in perl6 ^ or **
rakudo: say 2**3
p6eval rakudo 9b6189: OUTPUT«8␤»
kthakore cool
tylercurtis kthakore: 1 / 10**10000 probably overflows to Num.
masak rakudo: say 2 ^ 3
p6eval rakudo 9b6189: OUTPUT«one(2, 3)␤» 23:04
kthakore rakudo: say ( 3 * ( 1/3 - 1/10*100 ) )
p6eval rakudo 9b6189: OUTPUT«-29␤»
kthakore rakudo: say ( 3 * ( 1/3 - 1/(10*100) ) )
p6eval rakudo 9b6189: OUTPUT«0.997␤»
kthakore hehe
rakudo: say ( 3 * ( 1/3 - 1/(10**100) ) )
p6eval rakudo 9b6189: OUTPUT«1␤»
kthakore there!
is that a bug or ... just too picky? 23:05
lue
.oO(It was just a test! I know this already)
kthakore rakudo: my $a = ( 3 * ( 1/3 - 1/(10**100) ) ); say $a == 1
jnthn rakudo: sub postfix:<...>($n) { $n.ceiling }; say 0.999...
p6eval rakudo 9b6189: OUTPUT«1␤»
lue .oO[ I actually hope to someday prove the number 0.0…1 (point zero recurring one) exists and is equal to 0 ]
kthakore rakudo: my $a = 1; say $a.ceiling; 23:06
p6eval rakudo 9b6189: OUTPUT«1␤»
kthakore whut?
jnthn: what is limit of a number
lue rakudo: say (1/3).fmt('%x') 23:07
p6eval rakudo 9b6189: OUTPUT«0␤»
lue O.o
kthakore lol what?
no way
rakudo: printf( "%x", (1/3) ); 23:08
p6eval rakudo 9b6189: OUTPUT«0»
kthakore rakudo: printf( "%x", (1/2) );
p6eval rakudo 9b6189: OUTPUT«0»
tylercurtis lue: "Let 0.0...1 refer to the result of subtracting the real number represented as 0.999.. with the "9" repeating infinitely from 1. Since 0.999... is equivalent to 1, 0.0...1 is equivalent to 0. Q.E.D".
lue rakudo: say (22).fmt('%x') # let's see if it's even working
p6eval rakudo 9b6189: OUTPUT«16␤»
kthakore rakudo: printf( "%x", (8) );
p6eval rakudo 9b6189: OUTPUT«8»
kthakore let me see perl5 23:09
jnthn kthakore: of the Num type?
Guess it's a double
kthakore jnthn: yea
tylercurtis "x an unsigned integer, in hexadecimal"
kthakore tylercurtis: right! oops
hehe
1/3 is 0 int
lue Integers Only? that's lame :/
kthakore lue no that is how format is set
lue rakudo: say :16<0.8> 23:10
p6eval rakudo 9b6189: OUTPUT«0.5␤»
kthakore rakudo: printf( "%g", (8) );
p6eval rakudo 9b6189: OUTPUT«8»
kthakore rakudo: printf( "%g", (1/3) );
p6eval rakudo 9b6189: OUTPUT«0.333333»
kthakore cool
bah ... this is too much fun 23:11
tylercurtis lue: wouldn't it be weird if sprintf had a totally different meaning for the same exact inputs in C, Perl 5, and Perl 6?
jnthn goes to cook some noms to keep him going until it's airport time.
kthakore tylercurtis: that would be insanity ... please never ever change printf
23:11 cognominal joined
tylercurtis I'm not saying it wouldn't be worthwhile. Just that it would trip people up all the time. 23:12
lue tylercurtis: I know, that doesn't mean I have to like how it is :)
grr. "Just like Perl 5" doesn't mean you can !explain it. 23:14
rakudo: printf("%x", (16)); printf("%x", (22)); 23:15
p6eval rakudo 9b6189: OUTPUT«1016»
dalek psi: 62ff73d | masak++ | (4 files):
[Yapsi] big compiler/runtime refactor

Rakudo master for the first time. Three tests fail in t/compiler.t. Some odd effect cause t/runtime.t to fail when nothing is printed from Yapsi.pm.
psi: e8bd9d2 | masak++ | doc/ChangeLog:
[doc/ChangeLog] updated
kthakore lue: the format for printf has been around for ages man
masak \\o/
kthakore masak: \\o/
lue Now that printf does what I thought it did, why not implement a sayf ? 23:16
masak++ \\o/
kthakore lue: make a module then and use format any which way you want to
tylercurtis rakudo: sayf "%x", 16;
p6eval rakudo 9b6189: OUTPUT«Could not find sub &sayf␤ in main program body at line 22:/tmp/Q7BBOHtjJ2␤»
tylercurtis .u frowning face
phenny U+2639 WHITE FROWNING FACE (☹)
masak I will now take comments on github.com/masak/yapsi/blob/master/...ce/2010.08 for a while. 23:17
after that, I will go into release mode.
lue I can't help but wonder why there's no easy way to convert from decimal like we can convert to decimal. (printf, as we found, is not good enough) 23:18
masak lue: the truth is probably that there's not a terribly big need. 23:19
dalek psi: e4c7930 | masak++ | doc/announce/2010.08:
[doc/announce/2010.08] de-draftified
 23:21
psi: 895382b | masak++ | doc/announce/2010.08:
[doc/announce/2010.08] commas comment
23:22 HarryS joined
lue yeah, the dissymmetry is annoying me though :D 23:22
masak lue: any comments on the release announcement before it goes out? :)
lue masak: good job, I, being me, happen to, of course, like it :) 23:26
jnthn masak: lol, the dugong made it in. :) 23:27
masak jnthn: twice!
jnthn doublegong 23:28
lue
.oO(darn. All our special ink was made from the diamonds of Midnight, and no-one will go there anymore.)
 23:29
masak jnthn: I added it in a third time in lieu of being serious. :) 23:30
tylercurtis masak: please leave the comma comment in the announcment. 23:31
announcement, rather.
masak tylercurtis: I plan to. :)
tylercurtis Is a negated character class in Perl 6 "<![abdkl r245]>"? 23:32
masak tylercurtis: or <-[...]>
tylercurtis: former is zero-width. 23:33
tylercurtis masak++ yay! I don't want zero-width.
masak that's what the pet dugong said!
dalek psi: fa5237a | masak++ | doc/announce/2010.08:
[doc/announce/2010.08] more dugong
masak jnthn++ # dugong! 23:34
masak switches into release mode
lue steps back
shirtless gasps 23:36
interesting list comprehension behavior
should I file this as a bug? say 'A'..'z'; 23:37
masak shirtless: not before you've read S03 at least once. :)
shirtless I will have to take alook
in the meanwhile try it, it might be good for generating entropy
tylercurtis non integer Ranges make me do this
masak shirtless: perlcabal.org/syn/
tylercurtis .u frowning face
phenny U+2639 WHITE FROWNING FACE (☹)
shirtless masak: thanks for th elink 23:38
masak you're quite welcome.
jnthn dug\\o/ng!
masak yes! I got the date 2010-08-01 at github.com/masak/yapsi/downloads \\o/ 23:42
lue rakudo: say 256 * .125
p6eval rakudo 9b6189: OUTPUT«32␤»
masak mail sent to p6c. blog post coming up. 23:45
frooh if I want to add a method to a package, how do I do that exactly? 23:49
I tried my &{"foo"} = sub (...) is export { ... }; 23:50
also our
but neither of them exported (and maybe they didn't get defined?)
masak frooh: packages don't have methods. classes and roles do. 23:51
frooh sorry, I know 23:52
I mean a sub anyway
and still used the sub keyword
masak oh, ok.
the syntax you use looks p5-inspired. it doesn't work (and won't ever work) in p6. 23:53
frooh yeah...
I thought patrick told me to try this
becuase I'd tried *"foo" = ...
because *that's* the p5 syntax
masak but doing 'our &method-name = sub (...) { ... }' should work.
thing is, I suspect you want 'method-name' to vary with the contents of some string. 23:54
frooh exactly
hence the &{"..."} :-)
masak frooh: may I suggest eval? 23:55
frooh :-(
srsly?
masak what you're trying to do is evil hackery anyway, so might as well :P
frooh no it isn't
access to the symbol table?
that's like... perl
masak I won't argue there. 23:56
frooh lemme try again with the our
I didn't get an error
masak frooh: I have written a grammar engine in Perl 6, which extensively uses the "add a sub with some name only known at runtime" pattern, just like you do. I ended up using eval. I didn't regret it. 23:57
frooh masak: is that due to missing features or extra power? 23:58
23:58 Psyche^ joined
masak I see it this way. the 'sub somename (...) { ... }' idiom is for when you know the name of your sub at authoring-time, which is like 99% of the cases. 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!