»ö« | perl6-projects.org/ | nopaste: sial.org/pbot/perl6 | evalbot: 'perl6: say 3;' | irclog: irc.pugscode.org/ | UTF-8 is your friend!
Set by Tene on 14 May 2009.
skids and a more overarching utilization of the word "constrain" comes to mind for that. 00:01
e.g a sub that wants a Num generally wont try to use something exclusive to Int on it. 00:03
jnthn Right.
sjohnson will anyone mind if i do a /version on the channel 00:05
00:05 ElectricHeavyLan left 00:09 meppl left
jnthn OK, sleep time for me...night all. 00:10
wayland76 'night 00:11
sjohnson: What does that do?
sjohnson just tells me what irc client people use 00:12
and also hints at possible active perl users
by that i mean ActivePerl / win32 perl 00:13
for example, wayland76, i take it you use a KDE version for linux of some sort? 00:14
wayland76 Well, Kinda
I'm using Fedora 10, and it has both gnome and KDE 00:15
So I'm using Konversation, but I lean towards Gnome stuff as a general rule
sjohnson you can run K* programs in gnome without a hitch mostly? 00:16
sjohnson didn't know this fact
i was worried that their software only worked on KDE desktops 00:17
wayland76 I don't use either desktop
I use Enlightenment :)
sjohnson i remember when that was pretty spectuclar about 10+ years ago
and i can only imagine it still use back when i was playing with slackware linux
wayland76 Well, E17 is pretty cool :). But it's still not claiming to be production-ready 00:18
That said, I've had few problems with the version I'm using
The main problem is that the maximise button doesn't work, and you have to right click on the title bar and choose maximise from the menu. Other than that it's fine
sjohnson wayland76, do you have any bookmarked links handy for a good "perl 6" page that explains all the cool tricks you can do? 00:19
wayland76 Oh, but I have both desktops installed, so any libraries, etc, are there. They're just not running (but the libraries are no doubt called upon by the relevant programs)
sjohnson i read the wiki page and learned some stuff, but i don't believe it was exhaustive 00:20
wayland76 sjohnson: Not sure that I do; have you checked out www.perl6-projects.org/ ?
sjohnson i think this is what i am looking for, ty 00:21
wayland76 That links to the specs, and the specs can be quite readable (ie. compared to W3C or IETF stuff)
Also, if you're after more tutorial-format, readable stuff, we don't have any official stuff yet, but JDlugosz stuff is quite readable 00:22
JDlugosz's stuff: www.dlugosz.com/Perl6/web/ 00:23
sjohnson is the Perl 6 butterfly an official mascot? or just some guys cute drawing
skids s/some guy/larry/ 00:24
wayland76 sjohnson: The disadvantage of not running KDE or gnome is that it makes your programs load slower
yes, official mascot for the specs, the STD grammar, and the test suite, but not the implementations
sjohnson it is cute 00:25
wayland76 I'd be unsurprised, though, if implementation logos incorporated Camelia somehow (that's the name of the butterfly)
skids thinks larry should "quietly" open a cafe press account :-) 00:27
wayland76 googles cafe press 00:28
idea++ :) 00:29
sjohnson looks like t-shirt custom design page 00:30
00:30 fridim_ left 00:32 ElectricHeavyLan joined 00:38 lambdabot left
sjohnson rubs his hands together.. perl6! 00:38
00:40 lambdabot joined
sjohnson question: will Perl 6 have mutator methods like ruby's exclamation mark methods? 00:42
for example... @array_to_be_shuffled.pick!(*);
or something
skids Not sure exactly what those do... 00:43
sjohnson they affect the object calling it 00:44
in ruby: a = [1, 2, 3].shuffle 00:45
b = [1, 2, 3]
b.shuffle!
will be the same idea
the exclamation mark will change the object calling it
where as: a[1,2,3].shuffle # will not shuffle a permenantly 00:46
shuffle is a cute word, btw
especially with it has a ! at the end of it 00:48
skids I don't think there will be a consistent syntax throughout for it, but methods can modify their invocant and any parameter (as long as they prototype it as "is rw") 00:50
And they can return an lvalue, too.
Also, there are attributes: 00:52
rakudo: class A { has $.a is rw }; my $a = A.new; $a.a = 4; $a.a.say
p6eval rakudo 23718a: OUTPUT«4␤»
wayland76 sjohnson: Basically, yes. S12 (section 12 of the Perl 6 specifcation) has a Mutating Methods section
And that seems to do what you want
eg. @array .= sort; 00:53
00:56 DemoFreak left, H1N1 joined
skids Oh right, but it's not really a "mutating method" perse but an operator that invokes a normal method and assigns the return value. Hrm, though could you MMD-specify that as a variant somehow for true in-place? 00:56
wayland76 Well, I'm sure the optimiser will do it in-place :) 00:57
skids Eh. I never put much faith in optimizers, butam always willing to be impressed. :-)
00:57 lichtkind joined
wayland76 I never write anything big enough to worry about optimisation :) 00:58
If it can't be written in a month, it won't happen :)
skids e.g. maybe S12 should specify it to search for ($self is rw:) first...
rakudo: my $b = 5; sub foo { $b }; foo() = 4; $b.say # shouldn't this require sub foo is rw? 01:01
p6eval rakudo 23718a: OUTPUT«4␤» 01:02
01:04 H1N1[A] left 01:05 M_o_C left
skids sjohnson: anyway, you were asking about %{@{%h{f}}} messiness... attributes pretty much mitigate the need for that. 01:05
01:08 H1N1[A] joined
JDlugosz I like the use of "constrain". 01:08
lambdabot JDlugosz: You have 1 new message. '/msg lambdabot @messages' to read it.
sjohnson skids, how pretty can it look? 01:09
skids Well, assuming what you were using them for was class-like storage: $a.b<foo>.c[4] 01:11
sjohnson hmm, ok here's another question
anyway to pretty this up? 01:13
sub foo {
# accepts an array and an int
my @array = @{shift(@_)};
my $int = shift;
}
01:16 H1N1 left
skids rakudo: sub foo (@rray of Int is rw, $i of Int is rw) { $i = @rray.shift; }; my @a = <a b c>; my $stuff; foo(@a, $stuff); @a.perl.say; $stuff.perl.say 01:16
p6eval rakudo 23718a: OUTPUT«["b", "c"]␤"a"␤»
01:17 sjohnson left 01:24 eternaleye joined 01:25 cognominal left 01:27 cotto left 01:28 cotto joined 01:45 gabiruh joined 01:46 Eevee left 01:47 rewt left 01:54 rewt joined, nbrown joined 01:58 rewt left 01:59 Eevee joined 02:04 sjohnson joined 02:06 jonathanturner left
sjohnson perl6: say 'hi'; 02:19
p6eval elf 26914, pugs, rakudo 23718a: OUTPUT«hi␤»
sjohnson rakudo: print 5.^methods 02:21
p6eval rakudo 23718a: OUTPUT«ScalarabsabsACCEPTSperlpredsuccWHICH»
sjohnson rakudo: print 5.times { "hi" }
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "{ \"hi\" }"␤in Main (src/gen_setting.pm:0)␤»
sjohnson rakudo: print 5.times { "hi"; } 02:22
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "{ \"hi\"; }"␤in Main (src/gen_setting.pm:0)␤»
sjohnson rakudo: print "moose".^methods
p6eval rakudo 23718a: OUTPUT«Null PMC access in inspect_str()␤in Main (src/gen_setting.pm:3166)␤»
eternaleye rakudo: say "'hi' x 5
p6eval rakudo 23718a: OUTPUT«say requires an argument at line 1, near " \"'hi' x 5"␤in Main (src/gen_setting.pm:2400)␤»
eternaleye erg
02:22 clintongormley left
eternaleye rakudo: say 'hi' x 5 02:22
p6eval rakudo 23718a: OUTPUT«hihihihihi␤»
skids grumbles something about "stop installing avahi-daemon on my system!"
sjohnson rakudo: print "moose".uc 02:23
p6eval rakudo 23718a: OUTPUT«MOOSE»
eternaleye rakudo: say "\c[WHITE SMILING FACE]" 02:24
p6eval rakudo 23718a: OUTPUT«☺␤»
sjohnson rakudo: if (0) { print "no"; } else { print "yes";}
p6eval rakudo 23718a: OUTPUT«yes»
eternaleye sjohnson: In p6, the idiom has evolved into leaving out the parentheses on if, for, etc 02:25
skids rakudo: multi sub foo (Int $a) { "integer" }; multi sub foo (Num $a) { "number" }; my Int $a = 3; $a.foo; my Num $n = 5.4; $n.foo
p6eval rakudo 23718a: OUTPUT«Method 'foo' not found for invocant of class 'Int'␤»
eternaleye rakudo: if True { say "yes" } else { say "no" }
p6eval rakudo 23718a: OUTPUT«yes␤»
skids rakudo: multi sub foo (Int $a) { "integer" }; multi sub foo (Num $a) { "number" }; my Int $a = 3; $a.foo.say; my Num $n = 5.4; $n.foo.say;
p6eval rakudo 23718a: OUTPUT«Method 'foo' not found for invocant of class 'Int'␤»
sjohnson rakudo: if (1) print "hi"; 02:26
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "print \"hi\""␤in Main (src/gen_setting.pm:0)␤»
sjohnson i wish that would be implemented
eternaleye sjohnson: Curlies are mandatory now
sjohnson they always were
but i can see why
eternaleye If you want to omit them, you need the modifier form
rakudo: if 1 { say "hi" } 02:27
p6eval rakudo 23718a: OUTPUT«hi␤»
eternaleye rakudo: say "hi" if 1
p6eval rakudo 23718a: OUTPUT«hi␤»
skids rakudo: multi sub foo (Int $a) { "integer" }; multi sub foo (Num $a) { "number" }; my Int $a = 3; foo($a).say; my Num $n = 5.4; foo($n).say;
p6eval rakudo 23718a: OUTPUT«integer␤number␤»
sjohnson how can you get a list of all the STDLIB methods and such for Perl 6
so i can see all the cool stuff it can do 02:28
rakudo: print < 1 2 undef 3>.collapse
p6eval rakudo 23718a: OUTPUT«Method 'collapse' not found for invocant of class 'List'␤»
sjohnson rakudo: print < 1 2 undef 3>.methods
skids Well, a good amount of that is in flux, but it is Synopsis 32 (S32)
p6eval rakudo 23718a: OUTPUT«Method 'methods' not found for invocant of class 'List'␤»
sjohnson rakudo: print < 1 2 undef 3>.^methods
p6eval rakudo 23718a: OUTPUT«invoke() not implemented in class 'ResizablePMCArray'␤in Main (src/gen_setting.pm:3166)␤»
02:29 Whiteknight left
skids What's collapse? 02:29
eternaleye say <1 2 undef 3>.grep: { $_ !~~ undef } 02:30
rakudo: say <1 2 undef 3>.grep: { $_ !~~ undef }
p6eval rakudo 23718a: OUTPUT«12undef3␤»
skids simon says :-)
eternaleye rakudo: say <<1 2 { undef } 3>>.grep: { $_ !~~ undef }
p6eval rakudo 23718a: OUTPUT«Use of uninitialized value␤123␤»
sjohnson rakudo: say <1 2 undef 3>.grep { $_ !~~ undef }
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "{ $_ !~~ u"␤in Main (src/gen_setting.pm:0)␤»
eternaleye sjohnson: It doesn't work because it's the string "undef", not an undef value 02:31
And the colon is necessary
sjohnson rakudo: say <1 2 undef 3>.grep (defined)
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "(defined)"␤in Main (src/gen_setting.pm:0)␤»
sjohnson it was worth a shot :)
eternaleye rakudo: say <1 2 undef 3>.grep: { $_ !~~ "undef" }
p6eval rakudo 23718a: OUTPUT«123␤»
sjohnson what does the second squiggle do?
skids say <<1 2 { undef } 3>>.grep { .defined } 02:32
rakudo: say <<1 2 { undef } 3>>.grep { .defined }
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "{ .defined"␤in Main (src/gen_setting.pm:0)␤»
eternaleye ~~ is smartmatch, ! negates it
=~ does not exist in P6
skids rakudo: say <<1 2 { undef } 3>>.grep: { .defined }
eternaleye Nor !~
p6eval rakudo 23718a: OUTPUT«Use of uninitialized value␤123␤»
sjohnson oh i see
does it work the same way?
or is it totally different
eternaleye Yes, but more intelligent
skids arguably too dwimmy for its own good :-) 02:33
eternaleye so `my Dog $rover; say $rover ~~ Dog` checks ISA
Or 'does' really
sjohnson rakudo: say ("moose" =~ m/oo/); 02:36
p6eval rakudo 23718a: OUTPUT«Unable to set lvalue on PAST::Val node␤in Main (src/gen_setting.pm:3166)␤»
sjohnson rakudo: say ("moose" ~~ m/oo/);
p6eval rakudo 23718a: OUTPUT«oo␤»
eternaleye std: say ("moose" =~ m/oo/); 02:37
p6eval std 26914: OUTPUT«##### PARSE FAILED #####␤Obsolete use of =~ to do pattern matching; in Perl 6 please use ~~ instead at /tmp/bOdZQh9SJR line 1:␤------> say ("moose" =~ m/oo/);␤FAILED 00:04 36m␤»
sjohnson rakudo: say ("m00se" ~~ m/00/);
p6eval rakudo 23718a: OUTPUT«00␤»
sjohnson rakudo: if ("m00se" ~~ m/00/) { say 1 } else { say 0 }
p6eval rakudo 23718a: OUTPUT«1␤»
sjohnson rakudo: if ('00') { say 1 } else { say 0 } 02:38
eternaleye sjohnson: It returns a Match object, which stringifies to the matched text and boolifies to whether the match succeeded
p6eval rakudo 23718a: OUTPUT«1␤»
sjohnson rakudo: if ('0') { say 1 } else { say 0 }
p6eval rakudo 23718a: OUTPUT«0␤»
sjohnson rakudo: if ('000') { say 1 } else { say 0 }
p6eval rakudo 23718a: OUTPUT«1␤»
skids rakudo: if (?'00') { say 1 } else { say 0 } 02:39
sjohnson that's good to know
p6eval rakudo 23718a: OUTPUT«1␤»
eternaleye rakudo: say ("moose" ~~ m/oo/).WHAT
sjohnson what does the ?'00' do?
p6eval rakudo 23718a: OUTPUT«Match()␤»
skids rakudo: if (+'00') { say 1 } else { say 0 }
p6eval rakudo 23718a: OUTPUT«0␤»
eternaleye sjohnson: Coerces it to a boolean value
skids sjohnson: boolean context
sjohnson is this new to perl?
i never came across that in 5 :(
eternaleye sjohnson: '+' is for numbers, '~' is for strings
skids too tired to look up the str boolification spec.
eternaleye sjohnson: Yes, it's new 02:40
Only in Perl 6!
rakudo: say (?'00').WHAT, ' ', (+'00').WHAT, ' ', (~'00').WHAT 02:41
p6eval rakudo 23718a: OUTPUT«Bool() Num() Str()␤»
skids rakudo: say "0002" ~| "1234"; say "0002" +| "1234";
sjohnson is there anything Perl 6 can't do?
p6eval rakudo 23718a: OUTPUT«1236␤1234␤»
skids rakudo: say "0" ?| "1234"; 02:42
p6eval rakudo 23718a: OUTPUT«1␤»
eternaleye sjohnson: Anything that isn't implemented yet can't be done currently; I'm not sure if there's anything (besides pimc, pifl, and pire) that the final result of the spec can't do
c2.com/cgi/wiki?PimcPiflPire 02:43
Well, it can't solve the Halting Problem, but nothing else ever will be able to either ;D
sjohnson rakudo: exit(); 02:44
p6eval rakudo 23718a: ( no output )
sjohnson rakudo: say "im still running!" 02:45
p6eval rakudo 23718a: OUTPUT«im still running!␤»
eternaleye sjohnson: p6eval writes the command to a temporary file, and run()'s that.
TimToady sjohnson: the answer to your earlier question is simply sub foo (%array, $int) {...}
02:46 orafu left
TimToady you use signatures instead of all that fancy shifting 02:46
02:46 orafu joined
skids TimToady: up late? s/%/@/ :-) 02:47
TimToady it's always too late for me
except when it's too early
skids I feel EXACTLY the same way :-)
TimToady but then you can simply call it foo(@x, 42) 02:48
02:48 hercynium joined
TimToady no need for \@x anymore 02:48
rakudo: @a = 1..10; @a.=pick(*); say @a 02:49
p6eval rakudo 23718a: OUTPUT«Symbol '@a' not predeclared in <anonymous> (/tmp/9ZThr2qVSa:1)␤in Main (src/gen_setting.pm:3166)␤»
TimToady rakudo: my @a = 1..10; @a.=pick(*); say @a
p6eval rakudo 23718a: OUTPUT«29673541810␤»
TimToady there's your mutator
sjohnson TimToady: was saying (%array, $int) a mistake? is % == hash? 02:50
02:50 kirillm left
TimToady @array 02:50
lambdabot I'll crush ye barnacles!
eternaleye rakudo: sub foo( @bar, $baz --> Str) { return ([~] @bar) x $baz; }; my @arr = <qux corge grault>; say foo( @arr, 3 );
p6eval rakudo 23718a: OUTPUT«quxcorgegraultquxcorgegraultquxcorgegrault␤»
skids about that, can you MMD a variant on the method end somehow for in-place, or must you override .=? 02:51
sjohnson with all this new technology it looks like my Perl 5 book i got work to pay for will be used as a doorstop
the one written by Larry Wall and friends
TimToady rakudo: sub foo (@array, $int) { say @array[$int] }; foo('a'..'z', 15)
p6eval rakudo 23718a: OUTPUT«p␤»
eternaleye sjohnson: Perl 6 isn't (quite) ready for replacing Perl 5 yet 02:52
sjohnson rakudo: print ('a'..'c').WHAT 02:53
TimToady Perl 6 is trying to turn all the other books into doorstops; Perl 5 isn't special that way :)
p6eval rakudo 23718a: OUTPUT«Range()»
sjohnson rakudo: print ('a'..'c')[1]
p6eval rakudo 23718a: OUTPUT«b»
skids Plus, you can embed p5 regexe syntax in P6, so you'll still need it to debug code where other people punted.
sjohnson is frothing at the mouth for the new Perl 6
sjohnson snarls loudly 02:54
are block comments too much to ask for instead of pods in Perl 6 that work even if there it whitespace before them?
TimToady use #{.....} 02:55
or #(...) or #[....] etc
sjohnson do those work in Perl 5?
TimToady or #<<<...>>>
no
sjohnson heheh
the camel book was a very entertaining and funny read 02:56
i have never laughed so hard reading a manual in my life
my brother doesnt seem to get the jokes tho
skids or #«
TimToady maybe your brother should program in Python 02:57
sjohnson he's a C programmer mostly
i showed him this new Perl 6 stuff after he was whining about a lot of things in Perl he didn't like
he seemed pretty impressed
rukado: say `uname -a` 02:58
TimToady rakudo: say qx/uname -a/ 02:59
p6eval rakudo 23718a: OUTPUT«Linux timtowtdi 2.6.18-6-686 #1 SMP Mon Aug 18 08:42:39 UTC 2008 i686 GNU/Linux␤␤»
sjohnson back ticks a thing of the past?
TimToady yes, too unreadable
also, reserved for user-defined syntax
skids sjohnson: reserved for user-defined uses.
JDlugosz Hi Larry. I saw your sub foo. 03:00
lambdabot is still not giving me my messages though.
TimToady case (in)sensitivity?
JDlugosz Even when I copy/paste from its instructions to me. 03:01
sjohnson TimToady: a faster way to do: @technology = split ('\n', qx/ps aux/); ?
err.."\n" i mean
skids Awww F'd didn't get into the spec yet? :-) 03:02
03:02 cosimo left
TimToady space not allowed before that (, by the way 03:03
sjohnson who are you referring to?
TimToady you
split( not split (
sjohnson that must be new, right?
03:03 cosimo joined
TimToady unless you want it parsed as a listop 03:03
solves the p5 problem with print (1+2)*3 03:04
skids One of those areas where parsing/unambiguity trumps convenience.
sjohnson Question: Perl 6 have anything like basename() or pathname()? 03:05
TimToady IO specs still developing; ask wayland76 03:06
sjohnson rakudo: say basename("/etc/passwd");
p6eval rakudo 23718a: OUTPUT«Could not find non-existent sub basename␤»
sjohnson rakudo: say base("/etc/passwd");
p6eval rakudo 23718a: OUTPUT«Could not find non-existent sub base␤»
sjohnson *scratches head*
rakudo: my $x = if 1 { 10 } else { 15 } }; say $x 03:08
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "{ 10 } els"␤in Main (src/gen_setting.pm:0)␤»
sjohnson rakudo: my $x = if 1 { 10 } else { 15 } ; say $x
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "{ 10 } els"␤in Main (src/gen_setting.pm:0)␤»
sjohnson rakudo: my $x = (if 1 { 10 } else { 15 }) ; say $x
p6eval rakudo 23718a: OUTPUT«10␤»
sjohnson wow 03:09
thats neat
sjohnson cracks open a beer
JDlugosz TimToady I saw your example foo.
TimToady you keep saying that 03:10
sjohnson heh
JDlugosz Had me worried, with no examples of that in the current document.
sjohnson 'foo or foo()?
skids wait. isn't if supposed to return the conditional?
TimToady that's p5
p6 doesn't, so you can write list comprehensions
JDlugosz So I'm on the right track, that actual arguments are taken one at a time to match with the positional parameters, but THEN it is taken as list context to bind it. 03:11
TimToady huh?
JDlugosz <www.dlugosz.com/Perl6/web/passing_e...s.html>
TimToady it switches to pulling slurpy terms out of the positions when it's done binding mandatory and optional 03:12
JDlugosz Working it through, I ran into a problem with Nil and "item context".
I know, slurpy evaluates arg list in list context.
TimToady there is no list context until slurpy-binding starts
JDlugosz sub foo (@y) { ... } 03:14
foo(Nil);
We want @y to be bound to an empty list, not wonder how to bind to undef. Nil gives up the empty list in list context, not item context.
sjohnson rakudo: @a = <1 2 3>; print $a[0].WHAT, @a[0].WHAT;
p6eval rakudo 23718a: OUTPUT«Symbol '@a' not predeclared in <anonymous> (/tmp/lwUunSC1az:1)␤in Main (src/gen_setting.pm:3166)␤»
sjohnson rakudo: use lenient; @a = <1 2 3>; print $a[0].WHAT, @a[0].WHAT; 03:15
p6eval rakudo 23718a: OUTPUT«Can't find ./lenient in @INC␤in Main (src/gen_setting.pm:438)␤»
sjohnson rakudo: my @a = <1 2 3>; print $a[0].WHAT, @a[0].WHAT;
p6eval rakudo 23718a: OUTPUT«Symbol '$a' not predeclared in <anonymous> (/tmp/tm9ZLfbDcQ:1)␤in Main (src/gen_setting.pm:3166)␤»
TimToady the meaning of Nil is not precisely driven by item/list, I think
sjohnson rakudo: my @a = <1 2 3>; print $a[0].WHAT; @a[0].WHAT;
p6eval rakudo 23718a: OUTPUT«Symbol '$a' not predeclared in <anonymous> (/tmp/tlhbtjBVoP:1)␤in Main (src/gen_setting.pm:3166)␤»
sjohnson rakudo: my @a = <1 2 3>; print @a[0].WHAT;
p6eval rakudo 23718a: OUTPUT«Str()»
TimToady binding Nil to @y should still bind () 03:16
sjohnson rakudo: my @a = 1, <2 3>; print @a.perl;
JDlugosz TimToady: it should be... that's the point of having it. Nil as an object replaces "undef in item context, () in list context" exactly in the description of what some construct does.
p6eval rakudo 23718a: OUTPUT«[1, "2", "3"]»
sjohnson rakudo: my @a = 1, (<2 3>); print @a.perl;
p6eval rakudo 23718a: OUTPUT«[1, "2", "3"]»
JDlugosz The machiery of contextualizer DOES THAT. Now you ignore that and come up with another way.
TimToady Nil is really just (), and undef is only if forced to mean something non-arrayish 03:17
sjohnson rakudo: say Nil.WHAT
p6eval rakudo 23718a: OUTPUT«Nil()␤»
sjohnson rakudo: say undef.WHAT 03:18
p6eval rakudo 23718a: OUTPUT«Failure()␤»
sjohnson rakudo: say (0 == 1).WHAT
JDlugosz It gets far worse with flattening Captures with extra levels of nesting. The context thing works for that, as desgined. It just needs to get invoked at the right moment.
p6eval rakudo 23718a: OUTPUT«Bool()␤»
sjohnson rakudo: say qx/w/; 03:19
p6eval rakudo 23718a: OUTPUT« 03:19:10 up 262 days, 14:24, 0 users, load average: 0.89, 0.32, 0.20␤USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT␤␤»
TimToady if it is at all possible to interpret Nil as as (), it does so
skids rakudo: (1,2,3,Failure,4).say
JDlugosz I've identified where the passing description needs to be updated to go along with the context/Capture stuff, and it's elegant enough. You just can't mix up "non-slurpy" with "item context". Contexualizers (now) tell delayed objects what they should be.
p6eval rakudo 23718a: OUTPUT«123Failure()4␤»
sjohnson TimToady: is it true that pick(*) is the shuffle() function for Perl 6? 03:20
TimToady sjohnson: how many times do you need to be told before you believe it?
sjohnson i know it works, but i was wondering if there was... a TimToady way to shuffle as well planned
JDlugosz So, knowing that foo (@x, @y, $z) {...} foo(@list1,@list2,$x); is supposed to work, unless I'm missing something big here, I'd like to "run with that".
TimToady that is the TimToady way to shuffle 03:21
sjohnson i was just wondering if it was the only TimToady way tho
JDlugosz It is the contextualizer "step" that makes Nil become () as opposed to undef.
wayland76 rakudo: say qx/ls/
p6eval rakudo 23718a: OUTPUT«build␤Configure.pl␤CREDITS␤docs␤LICENSE␤Makefile␤parrot␤perl6␤perl6.c␤perl6.o␤perl6.pbc␤perl6.pir␤perl6_s1.pbc␤rakudo_revision␤README␤src␤t␤Test.pir␤Test.pm␤tools␤␤»
sjohnson other than writing your own function
skids rakudo: for (1,2,3,Failure,4) { say $_ if $_; } 03:22
p6eval rakudo 23718a: OUTPUT«1␤2␤3␤4␤»
JDlugosz So even if it is wrong to say the arg is "in list context", it needs to call that same step to trigger that.
skids Oh, Failure is the class. duh.
sjohnson TimToady: i will take your silence as a yes 03:23
TimToady more like "being used as a list/Positional" than list context, but yes
certainly a lot of the Any methods are in a similar state 03:24
you're using the invocant as a particular type, but this is implied by the method itself, as part of the langauge
JDlugosz Discusions with ruoso beat into me that the contextualizer is not the same as expressing the desired interface.
TimToady item/list context are not the only kinds of context
wayland76 sjohnson: The philosophy of Perl 6 is to get as many functions out of the core and into libraries as makes sense. I suspect that basename et. al. will be some of those
sjohnson rakudo: <b c a a b c c c a b a>.uniq.sort.say 03:25
p6eval rakudo 23718a: OUTPUT«abc␤»
wayland76 There will be multiple Perl6 "distros", one that packages Perl 6 with lots of modules for doing web stuff, one with modules for sysadmin, etc
TimToady but yes, basically, which is why a Positional interpretation of Nil should result in () even in "item" context
JDlugosz invocant as a particular type... I don't know what you're getting at.
wayland76 one of those distros will presumably include basename
sjohnson wayland76: if you are in charge of the basename stuff, please name a basedir too *putty dog eyes*
skids heads in for precious 4 hours sleep. Gah. Planned outage windows suuhuhuck. 03:26
TimToady rakudo: say "foo".join
p6eval rakudo 23718a: OUTPUT«foo␤»
TimToady join is not a method of Str
wayland76 sjohnson: I'm not really in charge. I just write the specs, which are then ignored or rewritten by the implementors
TimToady nevertheless, we use it as a list
JDlugosz That's where I was a couple days ago... context same as choosing an interface role.
TimToady and it treats it as a list of 1 elem
sjohnson wayland76: well thanks for reading my request / question seriously
wayland76 Basically, I'm completely unqualified to do anything to the Specs except rearrange them (which I did), but I provoke others into writing them by doing a really bad job of writing something :) 03:27
sjohnson rakudo: my @a = 1,2,3; @a.pick(*); @a.say;
p6eval rakudo 23718a: OUTPUT«123␤»
TimToady you need .=
sjohnson rakudo: my @a = 1,2,3; @a .= pick(*); @a.say;
p6eval rakudo 23718a: OUTPUT«231␤»
sjohnson just making sure :) 03:28
wayland76 is your suggested "basedir" the same as "dirname" in File::Basename (Perl 5)?
JDlugosz Hmm, wouldn't it have to be a method of some base class (Any is the ultimate base, right)?
So the body of that decides what to do with "self".
TimToady correct
sjohnson if FIle is a CPAN module you need to download, then yeah... i hate having to do basic good wholesome things with CPAN
cause i'm usually not a sudo user on many computers i dick around on
TimToady the Any version of join makes it a language feature, not just an OO feature :)
wayland76 Well, on my system, File::Basename comes with Perl 5
skids Some out of the box distros are lean on modules. 03:29
sjohnson slaps himself
wayland76, thanks tho
it is likely what i was looking for
JDlugosz In terms of explaining what goes on under the hood, the Smalltalk model is simple. Method call. Simple dispatch. "self" is easy. <G>
wayland76 skids: Yes, but File::Basename was first released with perl 5 03:30
sjohnson: Just "use File::Basename" :)
skids: In other words, it comes with Perl 5, and they'd have to remove it to get rid of it
TimToady yes, but it would be redundant to reinvent Smalltalk
sjohnson TimToady: are you a vi(m) kind of guy? 03:31
TimToady I use vim, though I think its API beyond the vi bits are rather ad hoc
JDlugosz Are we agreed that in principle, as things are now with Contexts and Captures, that Nil is a fragile object that turns into undef (I guess Nill is an interesting instance of undef) or () when needed. Right?
One way of triggering that is with the contextualizers, @(...) etc. 03:32
TimToady for some def or other
sjohnson vim stands for VI, a Man's editor
JDlugosz So parameter binding would draw upon the same behavior of the object. Or any other object that has context-sensitive behavior.
wayland76 I want something like Emacs, but with Perl6 instead of Elisp
(and of course, supporting VI keyboard interface as an option :) ) 03:33
Well, actually, $Parrot_based_language instead of Elisp :)
JDlugosz On the next computer over, my wife is talking to her folks in China on the webcam. It's a connected world. Where is everyone now on this channel? I'm in Texas. 03:34
sjohnson Question: in pick(*), what is the astrisk? is that a high-level globbing? or a special reserved keyword that means all elements or something
wayland76 JDlugosz: Australia 03:35
sjohnson is in Canada and is 27 years old
03:36 nsh- joined
TimToady is in San Diego this weekend 03:36
but usually in Mountain View, CA 03:37
wayland76 The "*" character as a standalone term captures the notion of "Whatever" (quoted from S02)
sjohnson say *.WHAT
rakudo: say *.WHAT
p6eval rakudo 23718a: OUTPUT«Whatever()␤»
sjohnson rakudo: say *.WHAT.^methods
p6eval rakudo 23718a: OUTPUT«perlACCEPTSScalar␤»
TimToady * means the operator gets to pick a "globby" default answer
JDlugosz Hmm, I'm thinking, TimToady, that Nil is too simple of an example. It =can= express a Positional role and be interesting undef all at the same time. Ruoso's examples are more challanging. 03:38
wayland76 JDlugosz: You're not in the Dallas/Fort Worth area are you?
JDlugosz Yes I am. Allen to be exact. You?
(my neice is still scared of me. Webcams have their downsides too)
wayland76 JDlugosz: Australia, as I said. I was just thinking that there are already at least 3 other Perl6ers in the DFW area 03:39
JDlugosz I mean, you have a connection with D/FW?
I noticed notices on the mailing list today... didn't notice them before. 03:40
wayland76 No, other than that I hold US citizenship as well as Australian :)
Btw, sjohnson, if you're really slavering with anticipation and unable to restrain yourself, you could redirect some of that unrestraint into one of the P6 implementations 03:41
TimToady sjohnson: many of the uses of * actually indicate an indefinite/infinite number of elements
03:41 orafu left
wayland76 Rakudo or SMOP 03:41
03:42 orafu joined
TimToady so 1..* is an infinite range 03:42
and "yes" xx * is an infinite number of "yes" list elements
but to other operations it can mean "all the elements" or "the number of elements" 03:44
03:46 mikehh left
JDlugosz Oh, question on the svn and synopses. 03:47
03:47 skids left
JDlugosz Is incrementing the Version and Last Modified automated? 03:47
TimToady no 03:48
I answered policy on that somewhere... 03:49
okay to incr for typos but not necessary
03:52 nsh left
sjohnson wayland76: how can i help? the only way i can think of helping is suggesting quick and easy things that might not be obvious, but would bring substantial peace of mind to Perl programmers 03:58
JDlugosz You can read my essays and tell me if anything needs improving re tutorials. 03:59
eternaleye JDlugosz: phenny is an option that isn't lambdabot. 04:00
phenny: tell JDlugosz phenny is helpful
phenny eternaleye: I'll pass that on when JDlugosz is around.
JDlugosz I'm not the one leaving messages for me.
phenny JDlugosz: 04:00Z <eternaleye> tell JDlugosz phenny is helpful
eternaleye JDlugosz: No @clear'ing :D
sjohnson p6 makes p5 look like stone-age technology. i am really glad i came here! 04:01
and i mean that without trying to offend anyone, but i cant wait till a solid working Perl 6 implementation comes out 04:02
so i will do look over your things JDlugosz 04:03
well, here's a suggestion. could $string.trim! be an alias for $string .= trim? 04:04
ala Ruby
wayland76 sjohnson: What are your skills? Documentation? C Programming? Any Programming necessary (ie. willing to learn new language to help)? Packaging?
sjohnson: That's a question for TimToady, but even if he says no, you can define a slang that lets you do that if you want 04:05
sjohnson by the ! mark i mean that p6 would recognize the ! as a .= of that method invocation, so you don't to define 100 more methods with the same name, except a ! after it 04:06
JDlugosz sjohnson: consider that a challenge. Write a macro for postfix:<!> that does that. Learn enough to take a stab at it. 04:07
(See my APL and Perl 6 essay to get ideas)
sjohnson let's say I figured it out, someone stick it on the p6-bot so i could test it out? 04:08
s/someone/could someone/
eternaleye rakudo: sub postcircumfix:<. !>( $var, &method ) { $var = $var.&method.() }; my $str = " foo "; say $str.trim! 04:09
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "!"␤in Main (src/gen_setting.pm:0)␤»
eternaleye pugs: sub postcircumfix:<. !>( $var, &method ) { $var = $var.&method.() }; my $str = " foo "; say $str.trim!
JDlugosz I think tht would be a bit much for any existing implementation to handle. And not a one-liner.
p6eval pugs: OUTPUT«*** ␤ Unexpected "!"␤ expecting term postfix, operator, ":" or ","␤ at /tmp/exv7I77nol line 1, column 106␤»
sjohnson looks like eternaleye might be close to figuring it out
eternaleye sjohnson: But postcircumfix custom operators are NYI. 04:10
sjohnson wayland76: skills 04:11
1. Documentation / explaining difficult in a very informal yet informative way.
2. Programming in high-level languages. Perl 5 and PHP (I hate PHP but don't mind that it has many functions out of the box).
3. Anally-retentive / perfectionist.
wayland76 sjohnson: You can download your own copy of Rakudo if you want
sjohnson: Sounds like you need to help with documentation and maybe spec then
sjohnson oh and 4. Finding easier ways of doing the same old thing that everyone else has been doing for 15 years. 04:12
though it seems that a lot of my ideas are already included in Perl 6 04:13
wayland76 sjohnson: U4X might be a good place. svn.pugscode.org/pugs/docs/u4x/ and use.perl.org/~masak/journal/38279
U4X = Userdocs for Christmas 04:14
When people ask "When is Perl 6 coming out", we say "Christmas", but we don't say which year.
Hence the tradition of referring to Perl 6.0.0 as "Christmas"
masak (Carl Masak) is in charge of U4X 04:15
and we expect some of JDlugosz's stuff to go in there too
literal rakudo: say "hi"
p6eval rakudo 23718a: OUTPUT«hi␤»
wayland76 (I think he said he was cool with the idea, anyway) 04:16
JDlugosz Yes, I'm cool. I want generations of newbes to ponder how to say my name.
sjohnson J. D. Lugozi 04:17
is my guess
sounds like Bela Lugosi
TimToady I have a pretty good notion how it sounds, if you pronounce it the olde way :)
the initial consonant cluster will, of course, stump most native English speakers 04:19
JDlugosz No sjohnson, The L is actually a Ł, but that's hard to type.
eternaleye sjohnson: No, he's John M Dlugosz
JDlugosz It's not a consonant at all.
TimToady well, laterals are always a liitle of both
JDlugosz <www.dlugosz.com/Me/index.html> 04:20
04:20 sjohnson is now known as DoGuess
DoGuess this would help others out :) 04:20
JDlugosz I like to eat (and make) pierogi, but I can't properly pronounce any of the consonants in it. 04:21
04:21 DoGuess is now known as sjohnson
sjohnson TimToady: do you have any thoughts on the ! ruby exclamation mark self-invocation idea? 04:21
TimToady I prefer to unify with the other mutating assignment operators 04:23
and reserve ! for factorial :)
04:23 mikehh joined
JDlugosz Hmm, I need a favicon for my web pages. 04:23
TimToady rakudo: sub postfix:<!> ($n) { [*] 1..$n }; say 5!
sjohnson $string .= trim; => $string.trim!
p6eval rakudo 23718a: OUTPUT«120␤» 04:24
TimToady so I think the Ruby notation is a dead end
sjohnson wow you sure write really short code 04:25
sjohnson looks at the line with a stern eye
wayland76 sjohnson: That's what happens when you have the specs in your head
The [] puts the * operator between all of the 1..$n 04:26
eternaleye wayland76: He no need can haz spec, he iz spec!
wayland76 (in this case)
04:26 nbrown_ joined
sjohnson is that like a eval(join type thing? 04:26
wayland76 eternaleye: Well, that's true, but I was trying to indicate to sjohnson that he might be able to achieve short code too :)
eternaleye sjohnson: It's the reduction operator, like foldr in haskell IIRC
wayland76 sjohnson: Well, kind of, if you want to think of it in Perl 5 terms 04:27
TimToady it's supposed to work the same as eval join but without the eval or the join
eternaleye [op] list, for any op, means to apply 'op' infixed between all items of the list
sjohnson rakudo: say <a b c>.[uc()] 04:28
p6eval rakudo 23718a: OUTPUT«No applicable candidates found to dispatch to for 'uc'␤in Main (/tmp/0LyEso4V08:1)␤»
eternaleye so if [<] @list checks if @list is in strict ascending order
sjohnson im probably going to break the compiler one of these days with bad syntax
wayland76 sjohnson: [] only works on operators. There's a function called "reduce" that does the same thing on functions
say <a b c>.reduce(uc()) 04:29
rakudo say <a b c>.reduce(uc())
eternaleye rakudo: say <a b c>.map( *.uc )
p6eval rakudo 23718a: OUTPUT«ABC␤»
wayland76 thanks
rakudo: say <a b c>.reduce(uc())
p6eval rakudo 23718a: OUTPUT«No applicable candidates found to dispatch to for 'uc'␤in Main (/tmp/dAcq78kugB:1)␤»
wayland76 rakudo: say <a b c>.reduce(*.uc)
p6eval rakudo 23718a: OUTPUT«Cannot reduce() using a unary or nullary function.␤␤»
sjohnson rakudo: say <a b c>.reduce(uc) 04:30
p6eval rakudo 23718a: OUTPUT«No applicable candidates found to dispatch to for 'uc'␤in Main (/tmp/Z9RFmyDxgp:1)␤»
eternaleye rakudo: say <a b c>.reduce( cmp )
p6eval rakudo 23718a: OUTPUT«Could not find non-existent sub cmp␤»
eternaleye rakudo: say <a b c>.reduce( <=> )
p6eval rakudo 23718a: OUTPUT«Parameter type check failed; expected something matching Code() but got something of type Str() for $expression in call to reduce␤in method Any::reduce (src/gen_setting.pm:93)␤called from Main (src/gen_setting.pm:3166)␤»
eternaleye rakudo: say <a b c>.reduce( &infix:<cmp> )
wayland76 rakudo: say <a b c>.[cmp]
TimToady you can' t hang bare operators out there as terms
p6eval rakudo 23718a: OUTPUT«Null PMC access in find_method()␤in Main (src/gen_setting.pm:3166)␤»
rakudo 23718a: OUTPUT«Could not find non-existent sub cmp␤»
TimToady cmp is non associative 04:31
eternaleye rakudo: say <a b c>.reduce( &infix:<lt> )
p6eval rakudo 23718a: OUTPUT«Null PMC access in find_method()␤in Main (src/gen_setting.pm:3166)␤»
TimToady that would be a bug
say <a b c>».uc 04:32
rakudo: say <a b c>».uc
p6eval rakudo 23718a: OUTPUT«ABC␤»
sjohnson shit my utf-8 isnt working
i cant see what he did
Infinoid sjohnson: >>
sjohnson (<A B C>) >>.uc 04:33
TimToady utf-8 is pretty much a necessity on this channel
sjohnson rakudo: say (<A B C>) >>.uc
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near ">>.uc"␤in Main (src/gen_setting.pm:0)␤»
TimToady can't have a space before a postfix!!!
sjohnson rakudo: say (<A B C>)>>.uc
p6eval rakudo 23718a: OUTPUT«ABC␤»
TimToady rakudo: say <a b c>>>.uc
p6eval rakudo 23718a: OUTPUT«ABC␤»
Infinoid nice.
sjohnson oops
i forgot to make the letters small
this perl 6 stuff is something else 04:34
eternaleye rakudo: sub infix:<ö>( $l, $r ) { say "$l »ö« $r"; }; <Camelia Rose> »ö« <butterfly bumblebee> 04:35
p6eval rakudo 23718a: OUTPUT«Camelia »ö« butterfly␤Rose »ö« bumblebee␤»
TimToady rakudo: my $x = "uc"; say <a b c>>>."$x"
p6eval rakudo 23718a: OUTPUT«ABC␤»
eternaleye Wait, when did unicode hypers start working? :D
TimToady rakudo: my $x = "uc"; my @x = <a b c>; @a>>.="$x"; say @a
p6eval rakudo 23718a: OUTPUT«Symbol '@a' not predeclared in <anonymous> (/tmp/d8cGXp9I4z:1)␤in Main (src/gen_setting.pm:3166)␤» 04:36
TimToady rakudo: my $x = "uc"; my @a = <a b c>; @a>>.="$x"; say @a
p6eval rakudo 23718a: OUTPUT«ABC␤»
TimToady there's one reason to use .= instead of !
the ! would get lost after the quote
sjohnson i think i can agree with you there 04:38
hmm 04:39
what if you knew you wanted to mutate it with a !, wouldn't you just make "uc!" string, and call it with a period still/
?
TimToady that changes the types 04:40
eternaleye buubot: spack Junction
buubot eternaleye: S02-bits.pod:5 S03-operators.pod:3 S06-routines.pod:1 S09-data.pod:4 S29-functions.pod:13
TimToady you really shouldn't be calling a method with such different semantics invisibly 04:41
sjohnson i think i will be just as happy with .= anyway
TimToady it is much more extensible
most of the places where Ruby has "cute" syntax tend to interfere badly with extensibility 04:42
the spots where Perl 6 looks slightly klunkier are that way for a good reason, in general 04:43
04:43 nbrown left, nbrown_ is now known as nbrown 04:44 justatheory joined
sjohnson thank godness perl is free 04:44
as i use it a lot for day to day things
and it's on every unixy system i go to 04:45
04:45 sri_kraih joined
sjohnson anyone here like shooting pool? 04:47
wayland76 Was Hæl Larry!
TimToady nah, the bullets slow down too fast in water
wayland76 (I was thinking of dead fish)
sjohnson heh 04:48
wayland76 It kinda takes the sport out of the fishing, though :)
sjohnson or young kids with goggles
TimToady not nearly as much as dynamite
wayland76 sjohnson: Yes, but I only do it once every 5-10 years
Hmm. Have we just designed a new computer game? :) 04:49
TimToady my wife is sure that game exists already embedded in some other game
sjohnson alright guys thanks for chatting 04:53
gonna go play a game of pool with my shotgun
CYA!
04:54 sjohnson left 05:01 cognominal joined 05:22 cognominal left 05:23 davidad joined 05:49 sjohnson joined
sjohnson pugs: my $x = ( if (1) { 5 } else { 10 } ); print $x; 05:51
p6eval pugs: OUTPUT«*** ␤ Unexpected "$x"␤ expecting "=", expression, ")", context, ":" or "("␤ postfix op␤ at /tmp/4rBlMEvu0J line 1, column 4␤»
sjohnson whats the one with the r again
should be in the topic maybe
rutaku or something
wayland76 Rakudo? 05:54
We had a really long topic, so we replaced it with perl6-projects.org/
sjohnson rakudo: my $x = ( if (1) { 5 } else { 10 } ); print $x;
p6eval rakudo 23718a: OUTPUT«5»
sjohnson i wonder why pugs cant do that line 05:55
05:55 justatheory left, ashizawa joined
sjohnson rakudo: @list = 1,2,3; print @list[1]; 05:59
p6eval rakudo 23718a: OUTPUT«Symbol '@list' not predeclared in <anonymous> (/tmp/yeMfwAIWky:1)␤in Main (src/gen_setting.pm:3166)␤»
sjohnson rakudo: my @list = 1,2,3; print @list[1];
p6eval rakudo 23718a: OUTPUT«2»
sjohnson rakudo: say qx/whoami/; 06:02
p6eval rakudo 23718a: OUTPUT«evalenv␤␤» 06:03
06:04 lichtkind_ joined
sjohnson is TimToady in bed? 06:07
TimToady I'll never tell 06:09
pugs: my $x = do if 1 { 5 } else { 10 }; print $x 06:11
p6eval pugs: OUTPUT«5»
TimToady the () syntax is more recent
sjohnson TimToady: can you field this question for my brother who programs in C 06:15
(11:04:36 PM) [email@hidden.address] make it so you dont have to have { } 06:16
(11:04:38 PM) [email@hidden.address] after an if
(11:04:44 PM) [email@hidden.address] it really annoys me.
(11:04:55 PM) [email@hidden.address] just wanna be able to say
(11:05:08 PM) [email@hidden.address] if $x print $var;
his response to just reversing it and putting the "if" after he knows, but says it's not what he wants
if ($var) print "hi!"; 06:19
06:21 lichtkind left
wayland76 sjohnson: There are many other short ways to do this kind of thing in perl 06:24
rakudo: 1 and print "hi"; 06:25
p6eval rakudo 23718a: OUTPUT«hi»
wayland76 rakudo: 0 and print "hi";
p6eval rakudo 23718a: ( no output )
sjohnson i take it Larry knows about this and probably has his own opinion as to why that wont work without the { } 06:28
thanks wayland76 tho, i didnt know about that trick 06:29
wayland76 You can also do ...
rakudo: 0 or print "hi"
p6eval rakudo 23718a: OUTPUT«hi»
wayland76 rakudo: 1 or print "hi"
p6eval rakudo 23718a: ( no output )
sjohnson rakudo: 1 && print "hi" 06:30
p6eval rakudo 23718a: OUTPUT«hi»
wayland76 I learned it from open(FILE, "filename") or die "message";
sjohnson wayland76: are && and "and" the same? or do they have some other differences
wayland76 different predence level
"and" is very low, "&&" is the same as C
I just always use "and", these days, because it's usually what I mean anyway 06:31
sjohnson if i use ($var) && "print hi";
for simple things like that, it's okay?
cause i will likely forget if i should use && or and 06:32
wayland76 Well, if in doubt, you can always put parentheses in
sjohnson i usually always do
wayland76 Yeah, && works well most of the time
sjohnson i can never remember the order of operations
wayland76 About the only time it would make a difference is if you did something like $b = $a && $c = 1 06:33
and yes, those are assignments there
sjohnson you can do that in perl 6? 06:34
wayland76 I just think "and" is more readable, and more likely to be what I want, so I always use that
sjohnson assign two vals in one statement?
wayland76 Well, that's Perl 5
I don't know if you can in Perl6
sjohnson oh
i see what you are doing
wayland76 It's probably a bad way to program
sjohnson you are essentially doing if ($b = a$a) { $c = 1 } 06:35
wayland76 but it's an example of where there'd be a difference
sjohnson err, $b = $a
something like that
wayland76 Well, the example I gave would be $b = ($a && $c = 1)
Whereas you might want ($b = $a) and ($c = 1) 06:36
The whole thing would be in an if statement or something
(those two commands I wrote on the previous lines are the same with and without ()
That first should maybe have been $b = ($a && $c) = 1 06:37
sjohnson hopefully i will never have to be that expressive when i program myself
wayland76 which would be really confusing.
That first should maybe have been $b = (($a && $c) = 1 )
Well, it's probably a bad idea all around, but we needed an example :)
06:38 frew|work joined
wayland76 might also make a difference in if($a, $b and $c, $d) {...} 06:38
Oh, wait, I think I remember who it's bad
or good 06:39
$a = $b && $c # does what you want here
$a = $b and $c; # always sets $a to $b
I think, anyway
So it doesn't always do what you want :) 06:40
sjohnson in that case 06:43
ill use english words "and" and "or" to do the die() stuff
06:43 Psyche^ joined 06:44 Patterner left, Psyche^ is now known as Patterner
sjohnson thanks for the answers wayland76 06:48
you seem to be interested in perl 6 as much or more than i am 06:49
06:52 ElectricHeavyLan left 06:54 FurnaceBoy left
sjohnson i wish i could do something to help, but i will probably just keep asking q's until i find something good that isn't in perl 5 06:58
07:00 azawawi joined
wayland76 sjohnson: Well, if you're not so keen on the Userdocs 4 Christmas approach... 07:00
azawawi good localtime() 07:01
wayland76 You could write some actual working code, and submit any bugs you find back to the coding team
Of course, you'd need to install one of the versions of Perl 6 for that :)
07:02 DemoFreak joined
wayland76 azawawi: hi :) 07:02
sjohnson what is the best one that is executable ? pugs?
wayland76 depends what you mean by "executeable"
But if I was going to try one, it would be Rakudo
sjohnson ... closest to bytecode
wayland76 sjohnson: remind me; what OS?
sjohnson probably linux since thats what i do most of my perl in 07:03
wayland76 Anyway, I'd avoid Pugs, and go with either SMOP or Rakudo
Rakudo is on the Parrot Virtual Machine (and is the one I lean towards), and SMOP is, IIRC, in native C 07:04
Which Linux? 07:05
Fedora? Ubuntu? Gentoo? etc
sjohnson old ass ubuntu 07:06
mostly run these on work computers
as i mostly use perl at work
wayland76 Oh, ok
I don't know what the Debian/Ubuntu packaging status is for any of these projects 07:07
sjohnson oh
i would just compile it if possible
it's so old that not even the packages work anymore
i have to compile most things from src
thankfully, most of these things work with a ./configure && make
wayland76 :) 07:08
07:08 azawawi left
wayland76 Rakudo has a built-in build script that will download and install Parrot for you. I don't use it myself, though (I'm slowly working on making an RPM of it) 07:08
sjohnson does it look like SMOP might end up being the fastest one? 07:09
in Perl 6, what was the better way again for writing this? ${%hash{symbol}} 07:11
%hash.symbol ?
TimToady do you mean symbol to be a string?
%hash<symbol> if so 07:12
%hash{symbol} would always call the symbol function
sjohnson yeah like if it was a key
${%hash{key}} is what i should have said
TimToady there's no ${} anymore 07:13
sjohnson thank god
TimToady or @{} %{}
sjohnson though i realize that was a bad example
$hash{key} would get you the stuff in p5
now my brother will be happy with perl
and myself too 07:14
well, i always am happy with it
TimToady since the .<> subscript is based on qw, it automatically gives you slices
sjohnson but im excited that he will be excited
TimToady %hash<foo bar> is %hash{'foo','bar'}
sjohnson %hash.foo won't work for foo only?
TimToady that only works if there's a foo method 07:15
sjohnson hmm... since <> is like qw
is there a different way to do it (out of curiosity only)?
TimToady to do which?
sjohnson just reference what's in the foo key in that %hash
wayland76 Yes. Is it %hash{foo}{bar} 07:16
or do I need a . in there somewhere?
TimToady no, you never need the dot except on alphanumeric methods
on all other postfixes it's optional 07:17
sjohnson wayland76: but i dont want the bar thing
TimToady %hash<foo><bar> if they're strings
sjohnson just %hash{foo} and $hash<foo> are the only ways to get this method?
TimToady what method?
sjohnson s/method/var
sorry too many beers tonight
TimToady {'foo'} has to be quoted to be equivalent 07:18
sjohnson unlike p5, correct?
TimToady yes
07:18 iblechbot joined
sjohnson ... and that's it? just {} and <>? no other special trick? 07:19
TimToady well, everything else is longer
sjohnson and it's not considered overkill to use a <> type thing on only one key ?
i think the answer is "yes"
TimToady %hash.postcircumfix:<{ }>('foo') or some such also works 07:20
sjohnson ok i am happy with that
thanks larry
TimToady why would that be overkill?
sjohnson just cause i normally use qw() in perl 5 to do more than 1 item in a list
TimToady it makes the autoquoting very visible
sjohnson please note that i also understand why it's perfectly fine 07:21
because it is syntatically perfect
TimToady well, a real <foo> as a qw would generally not be distinguishable from a 'foo'
so there's no point, unlike with subscripts
but you can't say %hash'foo'
sjohnson is <foo> not an array though?
or does it not even matter in 99% of the cases
likely because the Hash object can take in that array etc 07:22
TimToady the qw form is defined as equivalent to ('foo')
sjohnson is my guess
TimToady the subscript form is defined as equivalent to {'foo'}
sjohnson rukado: say ('foo').WHAT
damn i can never remember the name of this bot
07:23 ssm left
TimToady all you have to remember is that rakuda is "camel" in Japanese 07:23
sjohnson rakudo: say ('foo').WHAT
p6eval rakudo 23718a: OUTPUT«Str()␤»
sjohnson rakudo: say <foo>.WHAT
p6eval rakudo 23718a: OUTPUT«Str()␤»
sjohnson rakudo: say <foo bar>.WHAT
p6eval rakudo 23718a: OUTPUT«List()␤»
sjohnson i see.. the TRUTH now
07:24 ssm joined
sjohnson i think camel is "rakuda" 07:24
in japanese
TimToady that's what I said
sjohnson hahah 07:25
you are funny
heheheh
i can see why i liked the camel book so much
hmmm... 07:27
i cant think of another way to improve perl 5 07:28
i had a giant list of things too
well there is one thing:
string manipulation without regex. things like $food = 'cake';
$food[0] = 'C';
TimToady we might be able to support something like that someday 07:29
sjohnson it's as old as the hills... and it sure beats splitting it into a @food_chars list
TimToady it kinda depends on whether we can make this thing called NFG work out 07:30
sjohnson i found it particularily helpful i was programming in the wee hours a perl project to do that oxford hype thing
i shall show you
Lrray Wlal! Tkhans for wrintig Prel 5 and sonupitprg and enicoranugg and diisscnsug the frtuue of Prel 6, its soscucser.
TimToady My eyes! I should go to be for real 07:31
*bed 07:32
wayland76 'night :)
TimToady gotta get up in 6 hrs...
zzz &
sjohnson thanks hope to see you again soon
wayland76 sjohnson: You know this is all permanently recorded, and available on the 'Net? :)
Oh, I don't think you could keep him away from here :) 07:33
sjohnson so far i haven't said anything bad, have i?
or that i'm a super nerd
wayland76 Well, no, but that was some fairly beery typing :)
sjohnson lol
it's called... obfuscated English 07:34
araujo be careful, they might think perl6 is being developed by beery people
wayland76 ...and Beeri, as everyone knows, was a Hittite
(which is probably the Old Testament equivalent of a Klingon or something)
(Beeri was the father of one of Esau's wives in the Old Testament) 07:35
sjohnson is it possible to read the Perl 6 test that must be passed to be considered a Perl 6 official-approved implementation?
wayland76 Yes
But I forget where it is. They refer to them as spectests, I think 07:36
Tip: the link labelled "The official test suite" on perl6-projects links to the SVN for the spectests 07:37
07:37 revdiablo left, dukeleto left, diakopter left
wayland76 But be warned that they're incomplete 07:37
sjohnson time to write an ASM Perl 6 interpretor 07:38
wayland76 There is one. If you call Parrot's Assembly langugage an ASM :)
07:38 revdiablo joined, dukeleto joined, diakopter joined
sjohnson hmm 07:40
rakudo: for (1..10) { print "hi!"; }
p6eval rakudo 23718a: OUTPUT«hi!hi!hi!hi!hi!hi!hi!hi!hi!hi!»
sjohnson rakudo: print qx/date/; for (1..1000) if (0 == 1) { print "buy a new computer"; } print qx/date/; 07:42
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "{ print \"b"␤in Main (src/gen_setting.pm:0)␤»
sjohnson rakudo: print qx/date/; for (1..1000) { if (0 == 1) { print "buy a new computer"; } } print qx/date/;
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "print qx/d"␤in Main (src/gen_setting.pm:0)␤»
sjohnson rakudo: print qx/date/; for (1..1000) { if (0 == 1) { print "buy a new computer"; } } ; print qx/date/;
p6eval rakudo 23718a: OUTPUT«Sat May 23 07:42:47 UTC 2009␤»
sjohnson is waiting... 07:43
07:44 kate21de joined
sjohnson rakudo: print qx/date/.chomp; for (1..1000) { if (0 == 1) { print "buy a new computer"; } } ; print qx/date/; 07:47
p6eval rakudo 23718a: OUTPUT«Sat May 23 07:47:25 UTC 2009»
sjohnson does the pugs creator hang around here too? 07:51
wayland76 sjohnson: Not any more, I think. 08:01
Unless I just don't know their handle
Audrey Tang was the name 08:04
08:05 nbrown_ joined
sjohnson it is impressive that is a girl around my age that is doing Pugs 08:06
not something you see everyday
08:11 abra joined 08:22 nbrown left, nbrown_ is now known as nbrown 08:28 davidad left 08:32 davidad joined 08:54 kst left 08:55 kst joined 09:24 sbp is now known as sbp_, sbp_ is now known as sbp, phenny left 09:25 phenny joined 09:32 tulcod joined 09:43 iblechbot left 09:45 M_o_C joined 10:06 araujo left 10:10 |MoC| joined 10:13 cognominal joined 10:16 M_o_C left 10:17 cognominal left, frew|work left, cognominal joined 10:22 skids joined 10:32 Whiteknight joined 11:09 nsh joined 11:14 nsh- left 11:19 fridim_ joined 11:21 meppl joined 11:26 nsh- joined 11:35 payload left, abra left 11:43 tulcod left 11:54 ssm left 11:56 nsh_ joined, nsh left, nsh- left, nsh_ is now known as nsh 12:06 justatheory joined, nsh- joined, nsh_ joined 12:07 nsh left, nsh_ is now known as nsh 12:12 justatheory left 12:20 CMA joined 12:37 ZuLuuuuuu joined 12:44 mizioumt joined 12:55 pmurias joined
pmurias ruoso: hi 12:55
pugs_svn r26915 | pmurias++ | [re-mildew] added -x m0ld option to accept m0ld code 12:58
r26916 | pmurias++ | [re-smop] added .perl to idconst
r26917 | pmurias++ | [re-smop] added Hash.keys
r26918 | pmurias++ | [re-smop] added a extremely crude way of listing leaks
r26919 | pmurias++ | [re-smop] added a non-spec $OUT.describe for debugging purposes
pmurias ruoso: what i'm considering is to always return the result from a method calls by setr 13:09
13:09 ilbot2 left
pmurias * i'm considering always returning the result from a method call by setr 13:10
13:20 hercynium left 13:23 ilbot2 joined 13:33 justatheory joined 13:44 abra joined 13:45 icwiener joined 13:47 justatheory left 13:55 bacek joined
bacek hi there 13:55
rakudo: say log10(1i) 13:56
p6eval rakudo 23718a: OUTPUT«0␤»
13:56 nsh- left
bacek rakudo: say (2*log(10) 13:56
p6eval rakudo 23718a: OUTPUT«say requires an argument at line 1, near " (2*log(10"␤in Main (src/gen_setting.pm:2400)␤»
bacek rakudo: say (2*log(10))
p6eval rakudo 23718a: OUTPUT«4.60517018598809␤»
bacek rakudo: say 1i * pi 13:57
p6eval rakudo 23718a: OUTPUT«0+3.14159i␤»
bacek rakudo: say 1i * pi / (2*log(10)
p6eval rakudo 23718a: OUTPUT«Statement not terminated properly at line 1, near "/ (2*log(1"␤in Main (src/gen_setting.pm:0)␤»
bacek rakudo: say 1i * pi / (2*log(10))
p6eval rakudo 23718a: OUTPUT«0+0.682188i␤»
bacek rakudo: say abs(1i * pi / (2*log(10))) 13:58
p6eval rakudo 23718a: OUTPUT«0.682188176920921␤»
bacek rakudo: say abs(log10(i1) - (1i * pi / (2*log(10))))
p6eval rakudo 23718a: OUTPUT«Could not find non-existent sub i1␤»
bacek rakudo: say abs(log10(1i) - (1i * pi / (2*log(10))))
p6eval rakudo 23718a: OUTPUT«0.682188176920921␤» 13:59
14:03 ZuLuuuuuu left
bacek rakudo: say log10(1i) - (1i * pi / (2*log(10))) 14:04
p6eval rakudo 23718a: OUTPUT«0-0.682188i␤»
bacek rakudo: say abs(log10(1i) - (1i * pi / (2*log(10)))) 14:05
p6eval rakudo 23718a: OUTPUT«0.682188176920921␤»
bacek moritz_: ping? 14:07
pugs_svn r26920 | pmurias++ | [re-mildew] fix test count 14:09
r26921 | pmurias++ | [re-mildew] started working on RoleHOW
14:10 amoc joined
bacek msg moritz_ Can you check S32/log.t why "is_approx(log10(1i), 1i * pi / (2*log(10)), 'got the log10 of i');" is passing? Because it shouldn't... 14:12
14:14 moritz_ left, ilbot2 left
pmurias bacek: they are supposed to be skipped on rakudo 14:26
?
bacek Not all of them... 14:27
Ah. Yes.
AFAIU all of them should be skipped.
perl6: my $a = [^0.1]; say $a.shift 14:29
p6eval elf 26921: OUTPUT«Undefined subroutine &GLOBAL::prefix__94 called at (eval 124) line 3.␤ at ./elf_h line 5881␤»
..pugs, rakudo 23718a: OUTPUT«0␤»
pmurias rakudo: say log10(1i)
p6eval rakudo 23718a: OUTPUT«0␤»
bacek perl6: say log10(1i) 14:38
p6eval elf 26921: OUTPUT«Undefined subroutine &GLOBAL::log10 called at (eval 124) line 3.␤ at ./elf_h line 5881␤»
..rakudo 23718a: OUTPUT«0␤»
..pugs: OUTPUT«0 + 0.6821881769209206i␤»
bacek rakudo is so wrong.., 14:39
Infinoid bacek: I just left you a purl-message in #parrot. tt452 saved 5.8 minutes of CPU time doing spectest 14:55
ruoso pmurias, the result of any block happens in terms of setr 15:04
but if you mean return as the function, then the control exception is actually a requirement
15:05 Muixirt joined
pmurias ruoso: what i mean is that we should stop returning values using the Cish return 15:06
ruoso this dual-life form of SMOP_DISPATCH is important for bootstrap purposes
but as soon as it evolves, less and less code will return directly from SMOP_DISPATCH...
but basically...
Muixirt hi, I wonder whether there is a 'exists' function in perl6 15:07
ruoso you should only use "setr" if you have a continuation change
if you don't touch the current continuation, then you should just return from MESSAGE
pmurias when do we depend on dual-life form SMOP_DISPATCH
15:08 bacek left
pmurias ruoso: or is it just a bootstrapping shortcut? 15:09
15:11 aindilis left, Psyche^ joined
pmurias shortcut meaning a way to write less code 15:12
ruoso: continuation.eval could just call a interpreter.finished method instead of calling setr 15:14
ruoso: most of the places where we depend on SMOP_DISPATCH returning a value we break encapsulation as we assume that a given method is implemented in a stackless manner 15:15
15:21 rindolf joined, kst left, kst joined 15:23 Muixirt left 15:24 Patterner left, Psyche^ is now known as Patterner 15:27 meppl left
rindolf Hi all. 15:29
15:30 rindolf left 15:31 rindolf joined 15:33 Whiteknight left 15:56 kane_ joined
pmurias ruoso: one think we should keep in mind that we need to have a way for cyclic references to be collected 16:09
ruoso: and we have to do it automatically like python as we can't expect the user do weaken references explicitly as other implementation are trace gc 16:12
TimToady you should also delay destruction a random amount of time so people don't depend on it. :) 16:14
16:14 Lorn left
TimToady * .25 16:14
16:16 iblechbot joined 16:20 payload joined 16:25 skids left 16:28 meppl joined 16:49 nsh left 16:52 nihiliad joined 16:59 |MoC| left 17:00 CMA left 17:01 M_o_C joined 17:04 root joined
root woot woot woot!! I got perl 6 compiled on my Mac!! 17:04
heh
17:04 root is now known as dextius
dextius heh, I just installed an irc client, forgot to change my name ;-) 17:04
17:06 nsh_ joined, nsh_ is now known as nsh
dextius I have a question... Why does the function call "time" continue to return 5 digit precision when the underlying gettimeofday call has higher precision? (In perl5, I could print the higher precision time by doing the following) printf("%f!\n", scalar(gettimeofday())); 17:06
17:07 FurnaceBoy joined, frew|work joined 17:18 ElectricHeavyLan joined 17:22 ZuLuuuuuu joined 17:25 nsh_ joined 17:29 ab5tract joined
ab5tract hmm, just trying to compile rakudo from the github repo. perl Configure.pl --gen-parrot produces the following: 17:31
"Reading configuration information from parrot/parrot_config ..."
"Died at Configure.pl line 104." 17:32
any thoughts? 17:33
jnthn ab5tract: I'm guessing it somehow failed to build Parrot. Do you have a parrot/parrot binary?
ab5tract yeah
just put it into PATH?
jnthn But I guess no parrot/parrot_config binary?
Well, it needs parrot_config rather than just parrot
But it's kinda weird you've got a parrot binary but no parrot_config... 17:34
ab5tract sorry i thought you meant do i have a separate parrot compiled somewhere
there actually is a parrot_config binary in the rakudo/parrot dir
17:36 nsh left
ab5tract so i'm not sure what the problem here is 17:37
jnthn Me either. :-S 17:38
One thing to try 17:39
perl Configure.pl --parrot-config=path/to/parrot_config
I'm guessing that's just parrot/parrot_config
ab5tract right, unless i use the newest svn revision that i just compiled outside of the rakudo dir. but i figured it made more sense to use the one that rakudo is "tied" to 17:40
atm
jnthn Yeah, it tends to
You can try your other revision too.
ab5tract weirdly enough that solved it
jnthn Oh, hmm.
jnthn is confused as to why.
ab5tract quite strange indeed
jnthn Most strange is 17:41
Line 104 is just
close $PARROT_CONFIG or die $!;
Why on earth does a close die. :-S 17:42
(or more specifically, why in only one case...)
17:42 Andronic left
ab5tract bizarre 17:43
do you get the same effect ?
jnthn dextius: I'm not sure if the issue with time() is that the number it returns doesn't have the precision or if the output of it is rounding it... 17:44
ab5tract: No, I haven't seen that issue.
I was just looking at what's on the line number in the configure script that was mentioned in the error you pasted. 17:45
s1n jnthn: i'm getting a method not found for invocant error, but it doesn't indicate where it comes from
is there anyway of getting that information?
ab5tract it might be interesting to try a cd /tmp; git clone git://github.com/rakudo/rakudo.git; cd rakudo; perl Configure.pl --gen-parrot
to see if this is specific or general
jnthn s1n: In latest Rakudo?
(or one just released, at least)? 17:46
s1n jnthn: well, latest as of early this week :/
was there a change to report those errors better?
jnthn s1n: Yeah, I sneaked a fix in just before the Parrot release, either late Monday or early Tuesday. 17:47
s1n jnthn: okay, i'll update and see what it says about my bad code :)
jnthn That particular fix did deal with some error reports not having a line number and backtace.
So it's possibly you've found another bug, but my first guess is that you're missing the fix.
rakudo: my $x; $x.this-does-not-exist; 17:48
p6eval rakudo 23718a: OUTPUT«Method 'this-does-not-exist' not found for invocant of class 'Failure'␤»
jnthn hmm
ab5tract hehe
jnthn rakudo: my Int $x; $x.this-does-not-exist;
s1n that's the error :)
p6eval rakudo 23718a: OUTPUT«Method 'this-does-not-exist' not found for invocant of class ''␤»
s1n that one exactly
it doesn't have the class name
jnthn oddness
s1n: nod 17:49
Hmm
s1n i've been doing a bunch of stuff and i don't know which change i did broke it
and i haven't committed so i can't bisect :/
jnthn s1n: OK, stuff that certainly did have a backtrace pretty recently - after I put a fix in - seems to have lost it. 17:51
jnthn wonders if somebody decided to "improve" his commit 17:52
s1n jnthn: which commit (sha1) added backtraces?
jnthn s1n: It was a Parrot level fix. 17:53
s1n: However, I don't see any changes to the files I changed so hmm. 17:54
s1n: Things that I know worked a couple of days back now seem not to be though.
s1n jnthn: it worked with the version i had build on like monday or so, i didn't update/rebuild and that one showed up 17:55
jnthn: you might want to see if other backtraces work
jnthn s1n: Yeah, if you explicitly die they do work.
TimToady dextius: I suspect time should actually be returning a Rat, or a Num that can do Rat internally 17:56
jnthn It seems to be in other cases...
erm, seems not to be
s1n jnthn: there were several times the backtraces helped save me hours of debugging, so they were definitely working
jnthn s1n: Right.
Do you know how to pull a specific git revision?
Or to switch to one? 17:57
s1n jnthn: not yet :)
jnthn: what do you have in mind?
jnthn s1n: Trying to work out where things broke.
github.com/rakudo/rakudo/commit/9d2...96550282c7 is where I bumped Rakudo's revision up. 17:58
Parrot revision, that is.
So that's when I'm fairly happy it was working.
It'd help greatly to try that revision against current Parrot to see if something in Parrot changed or something in Rakudo did. 17:59
s1n jnthn: i can just knock parrot's revision back down
jnthn Of course, I can go debugging to find out too. :-)
s1n: You maybe could try that.
s1n jnthn: building rakudo master w/ parrot r38795 18:00
jnthn s1n: thanks 18:04
jnthn is a little distracted at the moment from looking
erm
well yes, that too by s/from looking/by cooking/
18:05 araujo joined
jnthn s1n: It seems to be that the .backtrace() method hands back something with zero elements... :-S 18:07
s1n jnthn: sorry, i got distracted (i'm trying to figure out flight plans to yapc with my wife lol) 18:08
18:09 skids joined
s1n jnthn: it won't build anymore with that version of parrot 18:10
ab5tract s1n, what error do you get? 18:11
jnthn, very strange thing: i did the perl Configure.pl --gen-parrot over again (after rm -rf parrot) and it worked fine... 18:13
jnthn ab5tract: on noes it's haunted!
ab5tract hehe
s1n jnthn: error:imcc:syntax error, unexpected '[', expecting '(' ('[') in file 'src/classes/Object.pir' line 230 18:15
it scrolls for quite a while
jnthn s1n: Yes, yes, that Parrot will be Too Old 18:16
OK, 39071 has the problem.
So it was before then.
s1n: I'm curious - if you have a moemnt to try it - if the release version had the issue. 18:22
Just tried 39030 and also seeing the issue. 18:23
s1n jnthn: yeah gimme a minute (booking my trip now)
jnthn s1n: Trying 39007 here which is furthest we can go back without being unable to build, afaict. 18:24
s1n: No hurry. :-)
18:26 shinobi-cl joined
shinobi-cl hi 18:26
18:26 frew|work left
shinobi-cl is there a way to use case-insensitive keys for perl6 hashes? 18:28
jnthn rakudo: my CIHash { has %storage handles *; method postcircumfix:<{ }>($key) { return %storage{$key.lc} } }; my CIHash $x .= new; $x<a> = 42; say $x.keys; $x<A> = 42; say $x.keys; 18:31
p6eval rakudo 23718a: OUTPUT«Malformed declaration at line 1, near "CIHash { h"␤in Main (src/gen_setting.pm:0)␤»
18:31 eternaleye left
jnthn rakudo: class CIHash { has %storage handles *; method postcircumfix:<{ }>($key) { return %storage{$key.lc} } }; my CIHash $x .= new; $x<a> = 42; say $x.keys; $x<A> = 42; say $x.keys; 18:32
p6eval rakudo 23718a: OUTPUT«The use of a %hash with the handles trait verb is reserved␤in Main (/tmp/bQUbUd1fjY:1)␤»
jnthn oh, gah.
rakudo: class CIHash { has $storage handles * = Hash.new; method postcircumfix:<{ }>($key) { return $storage{$key.lc} } }; my CIHash $x .= new; $x<a> = 42; say $x.keys; $x<A> = 42; say $x.keys;
p6eval rakudo 23718a: OUTPUT«No such attribute '$!storage'␤in method CIHash::postcircumfix:{ } (/tmp/vb58ZfsioC:1)␤called from method CIHash::postcircumfix:{ } (/tmp/vb58ZfsioC:1)␤called from Main (/tmp/vb58ZfsioC:1)␤»
jnthn rakudo: class CIHash { has $!storage handles * = Hash.new; method postcircumfix:<{ }>($key) { return $storage{$key.lc} } }; my CIHash $x .= new; $x<a> = 42; say $x.keys; $x<A> = 42; say $x.keys; 18:33
p6eval rakudo 23718a: OUTPUT«Symbol '$storage' not predeclared in postcircumfix:{ } (/tmp/eaoPYbEgVy:1)␤in Main (src/gen_setting.pm:3166)␤»
jnthn rakudo: class CIHash { has $!storage handles * = Hash.new; method postcircumfix:<{ }>($key) { return $!storage{$key.lc} } }; my CIHash $x .= new; $x<a> = 42; say $x.keys; $x<A> = 42; say $x.keys;
p6eval rakudo 23718a: OUTPUT«No such attribute '$!storage'␤in method CIHash::postcircumfix:{ } (/tmp/EHYfPdpRK3:1)␤called from method CIHash::postcircumfix:{ } (/tmp/EHYfPdpRK3:1)␤called from Main (/tmp/EHYfPdpRK3:1)␤»
jnthn wtf
18:33 nsh_ left
jnthn shinobi-cl: yes, if you can make it work ;-) 18:33
shinobi-cl thanks :) i got the idea 18:34
jnthn oh, better way maybe
shinobi-cl i though that can be done maybe using traits
like a case-insensitive trait
jnthn oh, no, we can't do that yet...
shinobi-cl: Yeah
shinobi-cl: That'll probably be the answer in the long run. 18:35
Rakudo is nearly but not quite there yet with making it possible to do it with a trait.
18:36 nsh_ joined 18:38 Kisu left 18:40 lisppaste3 left 18:43 abra left
TimToady more likely a Hash with a declared key type of a canonicalized string 18:44
s1n jnthn: okay, what is it you need me to do? 18:45
jnthn s1n: Build the latest Rakudo release (Stockholm) along with the Parrot it wanted and see if backtraces work. 18:46
perl6 -e 'foo()' # should give one if it does
s1n build specifically the stockholm release and not git master? 18:47
18:49 lisppaste3 joined
jnthn s1n: yes 18:50
s1n: At least that gives us a relativley recent point where it (hopefully!) worked.
ok, my dinner is ready, I'll be back in 20 mins or so
18:55 dextius left
s1n jnthn: is there a way to switch to that git revision (using sha1)? 19:02
pugs_svn r26922 | lwall++ | [Temporal.pod] prefer Rat from time() 19:05
19:08 mikehh_ joined
pugs_svn r26923 | lwall++ | [Temporal.pod] finish previous patch, urgh 19:10
19:17 icwiener_ joined
jnthn back 19:18
s1n: Probably but I couldn't work out how.
19:23 mikehh left 19:26 shinobi-cl left, rindolf left 19:29 nbrown_ joined 19:30 ZuLuuuuuu left 19:33 icwiener left 19:47 nbrown left, nbrown_ is now known as nbrown 19:49 sjohnson left 19:50 sjohnson joined 19:52 mikehh_ is now known as mikehh 20:11 nsh_ left 20:16 cotto left 20:17 cotto joined 20:19 PacoLinux left 20:22 PacoLinux joined 20:26 ElectricHeavyLan left 20:40 nsh_ joined 20:45 icwiener_ left 20:46 icwiener joined 21:01 JDlugosz left 21:02 nsh_ left, bacek joined, iblechbot left 21:18 ejs joined 21:37 tulcod joined 21:51 ejs left 21:53 PZt left 21:58 kate21de left 21:59 VX64Z joined 22:03 kst left, kst joined 22:05 PZt joined 22:06 ZuLuuuuuu joined
literal so, what if I want to define a sigil for my new role? 22:07
22:11 bacek left 22:13 amoc left
ruoso pmurias, it's not a shortcut in that sense... but the general idea is that it returns a meaningfull value *or* talks to the interpreter... this is not *just a shortcut*.. it's a fundamental concept on how we bootstrap the type system 22:20
literal all sigils are tied to roles, right? $ is Object, @ is Positional, % is Associative & is Routine, etc
ruoso literal, to define new sigil, you need to extend STD 22:21
thus you have a new language
literal ok
ruoso then you can do whatever you want in your new language 22:22
literal I was hoping for a more pedestrian way of adding a sigil :)
like with new ops 22:23
22:24 nsh joined 22:31 nacho_ joined 22:32 bacek joined 22:35 justatheory joined
sjohnson ruoso: hey you are working on the C implentation SMOP correct? 22:41
22:41 nsh left
ruoso sjohnson, yes 22:41
sjohnson it is it coming along? 22:42
s/^it/how/
ruoso well... we have a pretty solid base already 22:43
and we're starting to get things running in Perl 6 level on top of it
we still need to deal with some memory leaks 22:44
sjohnson i went to the perlfoundation.org site for SMOP, but i have to say that the layout of the page is very strange, and I don't really see where it describes any overview of what SMOP is, and what i's progress is
i's => its
ruoso yeah... and I kinda left the wiki a bit outdated... my bad...
because I started writing some more concise documentations
sjohnson, do you have pugs checked out? 22:45
(SMOP is inside the pugs repo)
sjohnson i checked out the Pugs website and found that more like something was looking for... ie, soemthing that describes the project
22:45 ZuLuuuuuu left
ruoso sjohnson, I mean the pugs svn 22:46
sjohnson so forgive me if i ask obvious questions like.. what is SMOPS goal, etc
not sure what SVN stands for
22:46 Chillance joined
ruoso sjohnson, ok... basically, SMOP is a runtime library that implements the semantics needed by Perl 6 22:46
it's not a Virtual Machine in the same sense as Parrot is 22:47
but if you think in p5 as a VM, then SMOP is a VM too...
SMOP is in the same level as p5, that is..
and not in the same level as parrot...
literal sjohnson: svn is subversion
22:47 pochi joined
literal svn.pugscode.org/pugs/ 22:47
ruoso sjohnson, for instance... Parrot will probably be able to implement ACL to memory access 22:48
and that is something SMOP won't ever be able to do
sjohnson ruoso: so after the runtime lib is done, then you will write something simple that interfaces with it, if you havent done so already?
ruoso well... starting from the beggining 22:49
SMOP, conceptually, is really just a .h file
but then there are implementations that follow that .h file
sjohnson, are you familiar with C programming?
sjohnson a tiny bit
i have K&R's C book 22:50
but find myself having a hard time reading thru it, not because it's boring or hard, but because of a lack of ideas for what I would use it for
ruoso svn.pugscode.org/pugs/v6/re-smop/lo...1_base.pod
the above link specifies what the base of SMOP is
everything else is built on top of that
sjohnson now, is SMOP the name for your imp. of Perl 6? or is SMOP something like Parrot in the sense that it can be used for othings that aren't related to Perl 6 22:51
as the title says "Simple Meta Object Programming / Simple Matter Of Programming" with no mention of Perl in it 22:52
i was a bit confused by that too
ruoso sjohnson, SMOP is just a runtime library, it's not really a Perl 6 implementation... but we are already working on a Perl 6 implemenation that uses this runtime, which is mildew
sjohnson is SMOP then going able to share it's "goods" for other uses other than Perl 6? is that the idea?
ruoso it's not the primary goal, but it should be useable by other stuff.. ys 22:53
*yes
22:53 tulcod left
sjohnson ok i c 22:53
ruoso sjohnson, take a look at the link i posted...
sjohnson because it's written in C, is the goal so that your Perl 6 runs fast?
i am looking at the link, a lot of it is over my head 22:54
22:54 wknight8111 joined
ruoso sjohnson, it's written in C because, well, there's nothing better to use... I never thought about using anything else... 22:54
and it gets low-level enough to build a runtime library
any other thing would include some runtime support 22:55
C is bare-bones
sjohnson isn't it tough programming in C though? or is it really easy once you know what you are doing
ruoso C is really like Perl
sjohnson i am really fast at programming Perl I can say 22:56
so I am happy to hear that it's like C
ruoso you just have no introspection
so, data is just data
and you have to know how to use it when you use it 22:57
sjohnson is it not possible to write functions / methods that can determine the type of data being presented to them?
ruoso no...
sjohnson or are you saying that objects can't have methods that self-identify
ruoso you can *try* to *guess*
sjohnson or structs
ruoso you can define something in the struct 22:58
but the struct itself doesn't provide any label to tell you which type of struct it is
(besides in the source code, of course)
sjohnson you cant make a struct { chat* name; } in there ? 22:59
( i realize to you i must sound like a complete idiot, so forgive me)
ruoso you can... but only the structs that have that char there will have the char there
and if, eventually, you receive something that isn't that struct
you might end up causing a segfault
sjohnson oh i c 23:00
what would you do in perl that solves this problem?
skids rakudo: my $a = 1; my $b = 2; (($a,$b) xx 1)[1] = 3; ($a, $b).say;
23:00 icwiener left
p6eval rakudo 23718a: OUTPUT«12␤» 23:00
skids rakudo: my $a = 1; my $b = 2; (($a,$b) xx 2)[1] = 3; ($a, $b).say;
ruoso sjohnson, well... in Perl you can use "ref" to know what it is
23:00 icwiener joined
p6eval rakudo 23718a: OUTPUT«3Use of uninitialized value␤␤» 23:00
ruoso sjohnson, besides a lot of other introspection
sjohnson, and predefined ways to access the data
sjohnson you sound like you really know what you are doing 23:01
ruoso it's been 2 years already
(sort of)
23:01 bacek left
sjohnson the best thing i've done in perl is write my own "management system" that interfaces with multiple git trees at work to make my life easy 23:01
all in perl
ruoso sounds cool
sjohnson but the stuff you are talking about here sounds really advanced
ruoso not really "advanced", just "low-level" and "abstract" 23:02
sjohnson how long have you been writing in C / when or how did you learn?
ruoso SMOP was my first real C project
sjohnson i take it you are learning a lot at the same time 23:03
ruoso I learn by coding
but there isn't much to learn in C
sjohnson yeah as Ritchie says, "C is not a complex language"
did you read that book too? 23:04
or learn on the internets
skids ruoso: except it's close enough to the metal to get into some pretty complicated OS nuances e.g. alignment, etc. 23:05
ruoso skids, yeah... sure... but you can get a long run without hitting that
sjohnson, I really meant it when I say that I learn by coding... I recall having read half of "practical C" several years back 23:06
sjohnson are there not times when you say "damnit, I wish C had Perl's regex stuff.." at all?
ruoso sjohnson, when I come to that point, I usually use libpcre 23:07
sjohnson i was writing a thing in C++ for my boss but i ended up just using C calls to do the work... but it was text manuipulation, and it got really tedious
i succeeded though with brute force character array tricks
ruoso: is C still as popular as it always was and has libraries for it? 23:08
i like C better than C++ in a lot of cases
as the C++ std lib is bloated as hell
23:09 nsh joined
ruoso sjohnson, well... C is... C... there's no substitute for that... 23:09
when I need C, it's because nothing else fits
skids spits upon hearing C++
not a fan 23:10
sjohnson skids: can you write in C too?
ruoso I never had a reason to even look at C++
skids Yeah, I do a bit of kernel hacking now and then in C.
ruoso (besides trying to patch an app written in C++ with half of it using plain C
the thing I don't understand is... why using C++ if half of your app is still going to need to be plain C? 23:11
sjohnson it can be all C++ but most programmers find it easier to just use C stuff
i talked with the guy who wrote uTorrent on efnet asking him this kind of stuff 23:12
and he said he wrote uTorrent in like 95% C code and some C++ stuff to make his life a bit easier with object stuff
skids The thing I don't understand is how a cantankerous OO system like that can be so attractive to use that it spawns 5x the useless redundant classes in a typical application than are needed.
sjohnson that is a good question 23:13
probably simply hype and more readily available books a the library talking about C++ as .. the NEW C
would be my guess
and a lot of cases at school teach it and job employers look for it. vicious cycle i suppose
have anyone of you heard of the game called Cave Story? 23:16
skids Now I can understand a lot of classes in aneasy to use OO language, mind you. I ilike the way Perl6 anonymizes and hides classes that are trivial mods of other classes, though.
sjohnson i think it was written in C
skids Can't say I have.
sjohnson skids: can it be said that Perl 6 is like Ruby in that "Everything is an object" saying? 23:17
skids To most extents, I think. 23:18
sjohnson rakudo: say 3.WHAT
p6eval rakudo 23718a: OUTPUT«Int()␤»
sjohnson seems to be an Int object
skids but not to say everything is an Object :-) 23:19
sjohnson what isn't one?
rakudo: {}.WHAT
p6eval rakudo 23718a: ( no output )
sjohnson rakudo: { 5; }.WHAT
p6eval rakudo 23718a: ( no output )
ruoso rakudo: say {}.WHAT
sjohnson i got that to work somehow
p6eval rakudo 23718a: OUTPUT«Hash()␤»
sjohnson oh
forgot the say
ruoso rakudo: say {5;}.WHAT
p6eval rakudo 23718a: OUTPUT«Block()␤»
skids sjohnson: I'll leave that question to jnthn, but I'm sure there are examples. 23:20
sjohnson is jnthn a developer alongside Mr. Wall t?
?
as he and one other guy have Ops in this channel
skids For all intents and purposes 2 is an Int (which is an Object) but behind the scenes it is "autoboxed". 23:21
So if you have an Array of Int, it is supposed to be packed in memory with each value stored consecutively. But @a[*] will still behave as Int objects. 23:22
jnthn and pmichaud are doing a great share of the class system work. 23:23
(among many other things) 23:25
sjohnson ruoso: is it just you working on SMOP? or do you have some friends / community members also helping
ruoso sjohnson, pmurias is a co-author of smop/mildew 23:26
skids ruoso: which one of you is batman, and which one robin, anyway? :-)
ruoso meh... bad analogy... 23:27
I started SMOP, pmurias started mildew
we both hack them inside out
23:29 cotto left
sjohnson has SMOP programming been fun or frustrating at times? 23:29
ruoso sometimes fun, sometimes frustrating... 23:30
but always with a line-of-sight to the end of the tunnel
23:30 cotto joined
skids ruoso: you're out of "refactor" mode now I take it? 23:31
ruoso skids, yeah... the refactor is over... we still need to merge re-smop and re-mildew back to their original places... but that's less importatnt 23:32
sjohnson ruoso: are you using HydraIRC? 23:33
23:33 nsh left
ruoso no... 23:34
23:35 nsh joined
sjohnson oh xchat that is right 23:35
i used to use that
ruoso it's not that bad... it has some quirks... but it is useable 23:37
and works nicely with several networks
sjohnson the one i use, weechat-devel, is not widely known
but i think it is the best irc client i have ever used, far better than irssi 23:38
plus the author hangs out on freenode and actually accepted 2 of my suggestions to make this my "dream client"
weechat (devel version) i really mean
23:39 frew|work joined 23:45 fridim_ left
sjohnson has anyone here tried ruby? have they concluded that Perl 6 would be better for them? 23:48
i am feeling that way right now, but curious to hear what others think
23:50 justatheory left 23:51 orafu left 23:52 sri_kraih left 23:53 eternaleye joined
ruoso sjohnson, I'm not that familiar with ruby 23:54

Powered by Cro and the Raku® Programming Language.

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

Thank you!