»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg p6eval perl6: ... | irclog: irc.perl6.org/ | UTF-8 is our friend!
Set by sorear on 4 February 2011.
00:02 whiteknight joined 00:04 raiph left 00:06 benabik_ joined, benabik left, benabik_ is now known as benabik 00:09 Chillance left
TimToady wonders if the chunk() routine from rosettacode.org/wiki/Floyd%27s_triangle#Perl_6 should be a built-in 00:57
also, the formatter map would be improved if we had a fmt infix that we could say Zfmt with 00:59
dinner & 01:01
pmichaud Zfmt -- interesting! 01:02
01:05 scott_ joined 01:07 xinming joined 01:10 xinming_ left
colomon TimToady: seems to me chunk ought to go in List::Utils. ;) 01:14
sorear Z[&sprintf] ? 01:17
colomon sorear: are you proposing that any sub which can take two arguments ought to be usable as an " 01:26
"operator" for a metaoperator? 01:27
colomon kind of likes that idea....
colomon likes that idea much better than making a fmt operator
sorear colomon: github.com/perl6/std/commit/bd9fad832c 01:28
benabik It works well for Haskell... x `func` y == fund x y
sorear someday I'm going to do something about the backlog of STD changes that haven't made it into niecza yet :/ 01:30
TimToady you're supposed to be able to say [&fmt] as an infix, which I guess is already specced 01:41
er, wait, fmt is a method
colomon sorear, TimToady: That's awesome 01:42
sorear I am not sure that ever made it into the synopses proper 01:43
fwiw
TimToady doesn't look like it 01:44
std: 42 [&atan2] 1
p6eval std 30385f8: OUTPUT«ok 00:00 44m␤»
TimToady I guess it was just added to STD
n: 42 [&atan2] 1 01:45
p6eval niecza v19-1-gbd1bd16: OUTPUT«===SORRY!===␤␤Unable to parse bracketed infix at /tmp/dmhgZx_Zhk line 1:␤------> 42 [⏏&atan2] 1␤Couldn't find final ']'; gave up at /tmp/dmhgZx_Zhk line 1:␤------> 42 [&⏏atan2] 1␤␤Parse failed␤␤»…
colomon sorear: how hard / disruptive would it be for, say me, to just start importing changes from STD to niecza?
sorear colomon: the hardest part is keeping track of which changes need to be imported. :) 01:46
TimToady or exported :)
colomon sorear: I was figuring on starting with a diff...
sorear colomon: importing that particular change should be quite trivial.
pmichaud ...deported?
sorear colomon: i'll just say that on the actions side you want to create an $OperatorFunction with arity 2 01:47
and no preargs/postargs
hmm, apparently '$' is silent to me 01:48
01:52 stephenlb left
TimToady std: "%3s" [&sprintf] 42 01:54
p6eval std 30385f8: OUTPUT«ok 00:00 44m␤»
pmichaud btw, typos on STD.pm6:4967-4970?
token metachar:sym<(?= )> { '(?:' <.obs("(?= ... ) for lookahead", "<?before ... >")> }
should be 01:55
'(?:' -> '(?='
(before the <.obs)
std: / (?:xyz) / 01:56
p6eval std 30385f8: OUTPUT«Use of uninitialized value $x in pattern match (m//) at STD.pm line 66584.␤Use of uninitialized value $x in concatenation (.) or string at STD.pm line 66623.␤===SORRY!===␤Unsupported use of (?: ... ) for grouping; in Perl 6 please use [ ... ] at /tm…
pmichaud std: / (?=xyz) /
p6eval std 30385f8: OUTPUT«===SORRY!===␤Quantifier quantifies nothing at /tmp/jwa8BBOJT5 line 1:␤------> / (?⏏=xyz) /␤Unrecognized regex metacharacter = (must be quoted to match literally) at /tmp/jwa8BBOJT5 line 1:␤------> / (?=⏏xyz) /␤U…
TimToady I guess so
pmichaud fixes
dalek d: fd2647b | pmichaud++ | STD.pm6:
Typo fixes in obsolete P5 regex syntax detectors.
 01:57
02:00 tyatpi left
[Coke] is there any way to introspect a character for unicode properties? 02:01
pmichaud in a regex, or otherwise?
I guess introspect implies "otherwise". I'm not aware of one.
[Coke] what about charname? 02:03
pmichaud don't find 'charname' in the synopses 02:04
*I don't
02:05 tomaw_ joined, tomaw left
pmichaud and I don't think Parrot exposes a charname interface. We could add one to NQP, though. 02:05
sorear n: use Unicode::UCD; say getprop('x', 'na') 02:06
p6eval niecza v19-1-gbd1bd16: OUTPUT«Unhandled exception: Unable to locate module Unicode::UCD in @path␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1402 (die @ 5) ␤ at /home/p6eval/niecza/src/NieczaPathSearch.pm6 line 23 (NieczaPathSearch.load_module @ 17) ␤ at /home/p6eval/niecza/sr…
sorear n: use Niecza::UCD; say getprop('x', 'na')
p6eval niecza v19-1-gbd1bd16: OUTPUT«===SORRY!===␤␤Q:CgOp not allowed in safe mode at /home/p6eval/niecza/lib/Niecza/UCD.pm6 line 4:␤------> { (ucd_get_ranges {$table} {$matcher}) ⏏}␤␤Q:CgOp not allowed in safe mode at /home/p6eval/niecza/lib/Niecza/UCD.pm…
pmichaud ...sigh, and then we have to ask "which name"? ;-) 02:08
sorear 'na', not 'na1' 02:09
evalbot control restart 02:11
02:11 p6eval left, p6eval joined, ChanServ sets mode: +v p6eval
sorear n: use Niecza::UCD; say getprop('x', 'na') 02:11
p6eval niecza v19-1-gbd1bd16: OUTPUT«===SORRY!===␤␤Undeclared routine:␤ 'getprop' used at line 1␤␤Unhandled exception: Check failed␤␤ at /home/p6eval/niecza/boot/lib/CORE.setting line 1402 (die @ 5) ␤ at /home/p6eval/niecza/src/STD.pm6 line 1147 (P6.comp_unit @ 37) ␤ at /h… 02:12
02:13 ajs_home left, ajs_home joined
sorear n: use Niecza::UCD; say value('na', "\x03BB") 02:14
p6eval niecza v19-1-gbd1bd16: OUTPUT«GREEK SMALL LETTER LAMDA␤»
sorear n: use Niecza::UCD; say value('na1', "\x03BB")
p6eval niecza v19-1-gbd1bd16: OUTPUT«GREEK SMALL LETTER LAMBDA␤»
sorear pmichaud: what did you mean by "which name"?
pmichaud ICU offers several choices: UNICODE_CHAR_NAME, UNICODE_10_CHAR_NAME (obviously not this one), EXTENDED_CHAR_NAME, CHAR_NAME_ALIAS 02:15
so, for example, CHAR_NAME_ALIAS is the one that is corrected from NameAliases.txt
sorear n: use Niecza::UCD; say value('NameAlias', "\x03BB")
p6eval niecza v19-1-gbd1bd16: OUTPUT«Unhandled exception: NAMEALIAS does not exist as a UCD table␤ at /home/p6eval/niecza/lib/Niecza/UCD.pm6 line 8 (get_value_num @ 5) ␤ at /home/p6eval/niecza/lib/Niecza/UCD.pm6 line 20 (value @ 5) ␤ at /tmp/Je6PjubW9D line 1 (mainline @ 4) ␤ at /home/p6ev… 02:16
sorear n: use Niecza::UCD; say value('Name_Alias', "\x03BB")
p6eval niecza v19-1-gbd1bd16: OUTPUT«Unhandled exception: NAMEALIAS does not exist as a UCD table␤ at /home/p6eval/niecza/lib/Niecza/UCD.pm6 line 8 (get_value_num @ 5) ␤ at /home/p6eval/niecza/lib/Niecza/UCD.pm6 line 20 (value @ 5) ␤ at /tmp/GaBmuA5lBf line 1 (mainline @ 4) ␤ at /home/p6ev…
sorear I think NameAliases.txt might only be supported in the reverse direction
pmichaud could be 02:17
u_charName() returns the "modern" name of a Unicode character; or the name that was defined in Unicode version 1.0, before the Unicode standard merged with ISO-10646; or an "extended" name that gives each Unicode code point a unique name.
so yes, it looks like the corrected name isn't available (through this function)
but we can get either the "modern" name or the "extended" one
02:20 orafu left, orafu joined 02:24 tokuhirom left, tyatpi joined 02:30 awwaiid left 02:31 awwaiid joined
colomon sorear: hate to admit it, but I've got no idea how to add the action for [&function]. NieczaActions.pm6 method infixish? I'm off to bed in a moment, but any hints would be welcome. (Porting the grammar bit from STD seemed completely straightforward, on the other hand.) 02:42
sorear yes, in infixish 02:44
colomon I'm guessing $Operator.funop will be involved?
sorear yes
colomon would that part be as simple as $Operator.funop($/, function-name, 2); 02:45
sorear check if $<infix> is actually a variable, then do_variable_reference and check_variable it
colomon like this bit? 02:48
make $Operator_Function.new(function =>
self.do_variable_reference($/, $<variable>.ast));
self.check_variable($<variable>);
sorear Yes.
well, $<infix>
because the parser fragment is <infix=.variable> 02:49
colomon instead of $<variable>
sorear Yes
colomon sigh. sorry for asking you to basically write it, but how do you check if $<infix> is a variable? 02:50
sorear $<infix>.reduced eq 'variable' 02:52
or $<infix>.ast !~~ $Operator
colomon my first stab gave me this: 02:55
Excess arguments to Cursor.O, used 1 of 2 positionals
I presume that's something to do with $¢.O(%additive)
colomon should have gone to bed fifteen minutes ago... 02:59
sorear yeah, should be <O(|%additive)> 03:02
colomon bother, different problem now. Will try to sort it out in the morning. 03:05
sorear++
sorear does that ++ mean you got it working
03:29 jaldhar_ left, jaldhar_ joined 03:50 jaldhar_ is now known as jaldhar 04:02 tyatpi left
moritz \o 04:18
04:21 telex left 04:22 kaleem joined 04:23 telex joined 04:33 telex left 04:39 telex joined 04:59 fhelmberger joined 05:27 dukeleto left 05:31 leprevost left 05:40 birdwindupbird joined 05:42 dukeleto joined 05:44 cog_ joined 05:47 cog__ left 05:48 pupoque left 05:54 wtw joined 05:57 simcop2387 left 05:59 dukeleto left 06:00 simcop2387 joined 06:13 dukeleto joined 06:15 tyatpi joined 06:19 tyatpi left 06:20 pupoque joined 06:49 birdwindupbird left 07:04 larks left 07:06 alvis` joined, bruges_ joined, revdiablo left, bruges left, driador left, alvis left, yves_ left, rsimoes left, apejens left, Tene left, driador joined 07:09 apejens joined, revdiablo joined 07:10 brrt joined 07:11 rsimoes joined, Tene joined 07:17 birdwindupbird joined 07:20 yves_ joined
dalek kudo/nom: ed46bdb | pmichaud++ | src/ (2 files):
METAOP_REDUCE -> METAOP_REDUCE_LEFT
 07:22
kudo/nom: 46dacb9 | pmichaud++ | src/Perl6/Actions.pm:
Refactor determination of reduction type for base operator.
kudo/nom: 433956d | pmichaud++ | src/ (2 files):
Refactor to pass reduction type to CROSS and ZIP metaoperators.
kudo/nom: b639af6 | pmichaud++ | src/core/metaops.pm:
Update METAOP_ZIP to honor associativity of the base operator.
kudo/nom: 5e3dfe9 | pmichaud++ | src/core/metaops.pm:
Update METAOP_CROSS to honor associativity of the base operator.
07:24 kresike joined
kresike good morning all you happy perl6 people 07:25
07:25 [particle]1 joined
moritz \o kresike 07:25
kresike moritz, o/ 07:26
07:26 larks joined 07:29 [particle] left
brrt kresike: i like your greeting 07:30
kresike brrt, then you must be a happy perl6 person :) 07:32
brrt i'm a happy parrot person at the moment, but close enough :-)
07:33 pochi left
moritz :-) 07:33
kresike the most important thing is that one's happy about something, the reason is secondary.
to be honest, I borrowed that greeting from Droopy :) 07:35
jnthn good morning o/
kresike hello jnthn o/
07:35 dylanwu left
jnthn ooh, pmichaud++ has been apatchin' 07:35
kresike jnthn, what does the Q stand for in QAST, QRegex ... ? 07:37
jnthn P++
;)
moritz Quantum!
jnthn r: say "P"++
p6eval rakudo 1ffd6c: OUTPUT«Cannot assign to a non-container␤ in sub postfix:<++> at src/gen/CORE.setting:1344␤ in block <anon> at /tmp/Kea0XOGr6c:1␤␤»
jnthn oops
moritz r: say 'P'.succ
p6eval rakudo 1ffd6c: OUTPUT«Q␤»
jnthn r: say "P".succ
p6eval rakudo 1ffd6c: OUTPUT«Q␤»
jnthn kresike: You can see it as a prefix to let you know "this is fairly new" :)
kresike jnthn, I figured as much, but does it involve some new idea or algorithm, or it's just a new and better implementation ? 07:39
moritz there are lots of new ideas that make it a new and better implementation :-)
kresike hmmm, that came out redundant ...
moritz it's not a whole new paradigm
it's mostly a rework to incorporate stuff that was missing in the previous model, and hard to add later on 07:40
and it takes advantage of 6model, our newish object storage system
07:40 pochi joined
moritz which, we hope, will reduce memory usage 07:40
jnthn kresike: PAST (which QAST is replacing) is one of the oldest pieces of our compiler toolchain. It's served us pretty well, but we've learned a lot in those years, and have some things we simply didn't know would exist back then.
Also PAST is written in PIR. Maintaining complex stuff in an assembly language is not so productive. 07:41
So it's a port to NQP plus improvements plus making it better integrate with 6model and bounded serialization.
kresike I see
jnthn hmm...open doesn't take :$enc for what encoding to read the file in... 07:42
07:42 tomaw_ is now known as tomaw
moritz jnthn: shouldn't be too hard to change 07:42
there's already a call to $!PIO.encoding
jnthn: we have a sub PARROT_ENCODING that maps from normal names to parrot encoding name 07:43
s
trying a patch now 07:45
jnthn++ # dddx blog post
blog.edument.se/2012/06/26/insights...12-part-1/ for those who don't twitter. (Not Perl 6 related, just general software engineering practise) 07:46
jnthn: the part about delaying decisions kinda parallels the reasons for bottom-up building -- it delays the higher level design decisions 07:47
07:47 thou left
jnthn Curious to see my other life leaking into #perl6 ;) 07:50
My reason for asking about $encoding is I'm faced with a bunch of latin-1 files here at $dayjob...
moritz well, there's another leak :-) 07:51
phenny: en nb "a year"? 07:52
phenny moritz: "et år" (en to nb, translate.google.com)
moritz now I have a test case for writing latin-1 stuff :-)
moritz on a system without a compose key :/
07:52 brrt left 07:53 brrt joined
moritz gah, I only patched the method form of open(), not the sub :/ 07:54
jnthn moritz++ 07:58
meeting &
moritz phenny: U+00E5 08:02
.u U+00E5
phenny moritz: Sorry, no results
moritz .u E5
phenny moritz: Sorry, no results for 'E5'.
dalek kudo/nom: 88a9d69 | moritz++ | / (2 files):
add :enc/:encoding support to &open
 08:06
moritz jnthn: there you go; my initial test was succesful
to all the lurkers that have feature requests: voice them. Tell us. Some of them are easy to do 08:07
08:13 sudokode left
kresike we would, but we're afraid :) 08:16
moritz don't be. I promise we won't eat you.
jnthn omnomnom users
moritz feeds jnthn with patches instead 08:18
jnthn another meeting... will try it in a moment
moritz no hurry 08:19
08:40 dakkar joined 08:57 dukeleto left
dalek ast: 5d689e8 | moritz++ | S06-signature/slurpy-and-interpolation.t:
test arglist interpolation of ranges, RT #113804
 08:59
ast: c8b1ef4 | moritz++ | S04-declarations/constant.t:
RT #112116, constant, hashes and map
 09:05
09:10 dukeleto joined
jnthn eek, I was using dir... 09:14
though
for dir($path) -> $name {
given "$path/$name" {
when :f {
I wonder if that gets simpler now...
moritz for dir($path) -> $io { when .f { ... } } 09:15
dalek ast: d1355fd | moritz++ | S06-operator-overloading/sub.t:
fix two operator overloading tests, unfudge one for rakudo
 09:17
masak so it does get simpler. that's a relief. 09:18
jnthn No 09:20
Everythign else gets worse :(
I actually only want the unqualified path part of the name in a bunch of plaes
And to do regex matches and .index on it and so on
And guess what? IO ain't Cool. 09:21
So I have to .Str everywhere.
So for my use case, dir returning IO is vastly worse.
moritz that's ugly
jnthn Having looked at a real word use of dir, my only feeling is, "revert please".
It's theoretically righter.
But practically wronger.
*real world 09:22
moritz but for other use cases (like file handles) it would be wrong to make IO Cool
jnthn Right.
moritz ok, let's revert
jnthn Well, I'm happy to hear countr-arguments.
But my gut feeling is, "ouch"
hmm, though one place gets a bit nicer 09:24
Hmm
Hre's a before/after example on one snippet 09:27
gist.github.com/4870f14c5f384e663daa
09:28 daxim joined
jnthn And really ugly 09:28
my $fh = open($dir_item.Str, :enc('latin-1'), :r);
I have to .Str it there
.O
o.O even
moritz hm, we could make the filename optional in IO.open 09:29
jnthn $dir_item.open(:enc...) didn't work
moritz yes, but could be made to work
jnthn
.oO( I don't get to be a whiney user very often...this is fun! :P )
 09:30
moritz now if we had a class IO::File is IO is Cool { } returned from &dir, and you could call .file on it to get the file name only (not full path), and which also .Str's to that, and you could call .open on it... 09:31
jnthn Right.
09:31 immortal left 09:32 erkan joined
jnthn It feels quite clear to me now that we need to separate out these two concepts. 09:32
Actual file handles being Cool feels wrong. But whatever dir returns really wants to be.
moritz moritz.faui2k3.org/tmp/0001-Revert-...-dir.patch # if we don't come up with something better
right, maybe it shouldn't inherit from IO at all 09:33
jnthn Or that.
moritz otoh it feels very convenient to be able to call .f, .s etc. on it 09:34
so, common role?
IO::FileTests, and both IO and IO::File do that role?
jnthn That's starting to feel better 09:35
moritz r: say $*IN.f
p6eval rakudo 88a9d6: OUTPUT«False␤»
moritz r: say $*IN.s
p6eval rakudo 88a9d6: OUTPUT«False␤»
moritz r: say ".gitignore".IO.s
p6eval rakudo 88a9d6: OUTPUT«True␤»
moritz erm
r: say ".gitignore".IO.size 09:36
p6eval rakudo 88a9d6: OUTPUT«No such method 'size' for invocant of type 'IO'␤ in block <anon> at /tmp/mdPL7aDefK:1␤␤»
moritz r: say ".gitignore".IO.bytes
p6eval rakudo 88a9d6: OUTPUT«No such method 'bytes' for invocant of type 'IO'␤ in block <anon> at /tmp/oUM5s7D5xp:1␤␤»
moritz thing is 09:37
whatever objects &dir returns, we should think hard about what its .lines method should return
if it's Cool, people will expect it to do the same as Str.lines
so we shouldn't allow reading from it directly 09:38
but allow an .open(*%opts) which returns an IO from which you can read
jnthn moritz: :$enc helped lots, btw. \o/ 09:43
09:43 odoacre left
moritz jnthn: good to hear 09:46
jnthn Well, just updated the big script with it and dir now...let's see how that works out 09:49
09:51 odoacre joined 09:56 mucker joined 10:05 cognominal joined, cog_ left 10:09 sergot joined
sergot hi o/ ! 10:09
kresike hello sergot o/ 10:11
masak sergot! \o/ 10:13
jnthn++ moritz++ # I like where this is going
10:22 scott_ left 10:30 JimmyZ_ joined 10:50 tyatpi joined
tadzik \o/ 10:56
jnthn Yeah, le other more complex script is working out far better now too.
Oh, apart from I did a silly bug that makes the analysis useless...
moritz l'other or l'autre please :-)
jnthn :P 10:57
10:57 snearch joined
jnthn corrects it and sets it off again 10:58
11:01 cog_ joined
moritz jnthn: do paths like ./foo/bar.txt work under windows? 11:03
ie if I feed them to open()
parrot's open, that is :-)
11:03 cognominal left
dalek kudo/io: cc780c2 | moritz++ | src/core/IO.pm:
move file tests to IO::FileTestable
 11:22
moritz l'experiment avec &dir based on this branch: gist.github.com/3003456 11:23
jnthn, masak, pmichaud: ^^ see above, feedback welcome
masak (l'other)++ # :P
moritz it tries to to DWIM without exaggerating 11:24
masak looks, interested
moritz maybe I should add an .IO coercion method
masak "IO::File<$.path>" # nice
moritz r: say Buf.new(1, 2, 244) 11:25
p6eval rakudo 88a9d6: OUTPUT«Buf:0x<01 02 f4>␤»
moritz same idea
masak waitwait, so then we would have IO and IO::File?
moritz yes
masak hmmmrm.
that might be the only detail I dislike slightly.
kresike moritz, while you're at it maybe it would be nice to have an IO::Dir class too
this way masak can have two things to dislike 11:26
moritz kresike: maybe
kresike: i'll probably experiment with it
masak: maybe we should rename IO to IO::Handle or so
kresike it makes sense because files can be opened while dirs cannot.
among other things 11:27
moritz kresike: one could also just add an .opendir or .dir method to IO::File
masak moritz: +1
moritz kresike: note that IO::File has an .open method even thought it could be a directory right now
*though
kresike moritz, yes, I think that's wrong 11:28
unless you think of a directory like vim does, and when you open it, gives you a list of files contained within as content 11:29
moritz kresike: I agree that it's wrong from a theoretical point of view
kresike: but in practise we often have methods that cause an error when used
compare (-2).sqrt for example
so, the question is, does it make sense from a user's perspective? 11:30
I could further imagine that we have an IO::Path does IO::FileTestable
kresike moritz, the question is, are files and directories different enough to create a new class for each of them ?
moritz and then IO::File and IO::Dir, which both do IO::Path 11:31
kresike: correct
I guess I'll experiment with it
bikeshedding didn't get us very far until now; I think experiments and use cases are the best approach 11:32
kresike IO::Path feels a little overkill to me, I'd add IO::File.basename, IO::File.path and IO::File.name
moritz well, /if/ we separate file and dir into separate types, then we should also have a common type for path handling, which IO::Path would be 11:33
if you think that having all three of Path, File and Dir is too much, we should probably just have one of them
kresike: IO::File.file is what you would call basename
maybe it's better to rename file to basename though 11:34
kresike well to me basename means the file name because I'm a *nix user, doesn't mean I'm right about it
moritz on a different tangent, one reason for having 'IO' instead of 'IO::Handle' was to have short file tests a la $str.IO ~~ :e 11:35
kresike you are right about the Path, File and Dir, Path should be an ancestor of the previous two
s/previous/other/
moritz if we rename it to IO::Handle, then we have to think of another way to do that
kresike: agreed (though I'm currently inclined to do with roles instead of classes; not sure if/how much that matters) 11:36
kresike iiuc roles are a kind of abstract class 11:37
moritz not quite 11:38
roles and classes are much alike, except that role composition forces you to resolve method name conflicts
whereas with classes and inheritance, one of them just wins
eiro masak, jnthn did you sent me the price of your flight ? 11:39
kresike I see
eiro (i'm afraid i lost the info)
kresike then I don't think it makes sense to use roles, since the Path of a File is very similar to that of a Dir
moritz kresike: you *can* get role semantics similar to Java's interfaces, but that's not how I mean to use them
a more interesting question is: will you ever instantiate an IO::Path without making it a File or a Dir? 11:40
the answer is probably "yes", because you might want to create path names without looking at the file system to see if it's a plain file or a directory 11:41
that means it should probably a class
arnsholt We'll definitely want to be able to represent a file path without considering the existence or type of that file, I think 11:42
kresike moritz, think of a filesystem as a cloud, that backs up making Path a class too
frettled Sounds nebulous to me. ;) 11:44
kresike needs to look up the differences between roles and classes in p6 11:47
11:49 cj left 11:52 k0ala joined 11:59 k0ala left
moritz gist.github.com/3003456#file_dir_split_up.pl # another version 12:04
12:11 k0ala joined
moritz that feels a bit more "engineered", I'm having trouble to decide whether that's a plus or a minus 12:15
pmichaud good morning, #perl6 12:17
kresike moritz, on second thought is there a use case when either File or Dir would override any of Path's methods ? 12:18
moritz kresike: well, IO::Dir might override method d { True } 12:19
kresike: or some such
kresike but that's Filetestable's job
12:19 snearch left
kresike I mean Dir is in fact overriding a method of Filetestable 12:20
masak eiro: we're going by train, at least on the way there. jnthn has the details.
12:20 not_gerd joined
kresike oh wait, Filetestable is a role 12:20
not_gerd hello, #perl6
moritz kresike: currently I'm not aware of any overriding, but I wouldn't trust it to stay that way
\o not_gerd
kresike my bad :) 12:21
moritz kresike: but that doesn't matter too much; the role is alreday flattened out in IO::Path
good am, pm
not_gerd moritz: S32/IO already specs things like File, Directory, Path
I take it the draft is obsolete? 12:22
moritz pmichaud: you might be interested in gist.github.com/3003456 (two different approaches to &dir that both dwim, with varying degrees of engineering behind it; needs the 'io' branch of Rakudo for IO::FileTestable)
not_gerd: I should read S32/IO again, but I usually mistrust that synopsis
eiro masak, train ? from sweden ? 12:23
moritz not_gerd: it's not a coherent picture. People have added their favorite parts to some roles/classes without considering the big picture, and others have split up classes in multiple classes and roles without thinking through how it all works out 12:24
eiro train+boat then ?
masak eiro: aye.
no, not boat.
we have a bridge.
moritz isn't there a bridge (over troubled water)?
kresike moritz, my point is, I can't think of a reason why Path cannot be a role, since it has to do the same things for File and for Dir too
masak the water ain't so troubled.
eiro OMFG! it must be a giant one
pmichaud moritz: looking 12:25
moritz kresike: the class/role distinction is really minor at this point, and easily changable the other way if we get it wrong now. I'd rather not spent too much time on this point right now
12:25 kaleem left
kresike moritz, right 12:25
eiro masak, jnthn are your at Strasbourg Sunday? i'm planning a "flamekuche party" at hope 12:26
not_gerd moritz: it's not all bad, but indeed there are some rough edges (Readable.read returns a new Buf on each read oO)
eiro home
not_gerd it lies as well: C89 does not specify integer file descriptors 12:27
moritz also it uses camelCase method names, which no other part of Perl 6 does 12:28
arnsholt wonders how floating point fds would work
moritz arnsholt: fractionally
arnsholt Random selection of bytes from the integer numbered files, weighted by the position between the two perhaps ^_^ 12:29
12:29 Pleiades` left
moritz not_gerd: more to the point, it misses some crucial bits of information 12:30
not_gerd: like, what kind of object does open() return?
not_gerd moritz: IO? 12:31
moritz =head2 IO
The base role only tags that this is an C<IO> object for more generic
purposes. It doesn't specify any methods or attributes.
sounds like not specific enough for me
if it returns an IO in the sense of S32/IO, you can do *nothing* with it 12:32
not_gerd I'm confused... - what about all the methods under that heading?
moritz they are in IO::Redable, IO::Writable etc.
not_gerd moritz: not according to what I'm reading
moritz and that's the real problem: there are scattered bits and pieces, but they don't fit together 12:33
pmichaud I'll have to catch up on the &dir and IO discussion in a little while; still doing morning household chores here
moritz not_gerd: then maybe we're reading different documents?
anyway, decommute now, will be available for discussion later 12:34
&
12:35 dukeleto left
pmichaud r say (1,2,3 Z, 4,5,6 Z, 7,8,9).perl # testing RT #113800 12:35
r: say (1,2,3 Z, 4,5,6 Z, 7,8,9).perl # testing RT #113800
p6eval rakudo 88a9d6: OUTPUT«((1, 4, 7), (2, 5, 8), (3, 6, 9)).list␤»
pmichaud \o/
afk, errands
12:42 Pleiades` joined 12:43 kaleem joined 12:48 dukeleto joined 12:54 odoacre left 12:57 PacoAir joined 12:58 JimmyZ joined 12:59 JimmyZ_ left 13:01 k0ala is now known as thilp 13:02 mucker left 13:07 cognominal joined 13:10 cog_ left, Psyche^ joined 13:11 Patterner left, Psyche^ is now known as Patterner 13:12 tyatpi_ joined 13:14 [particle]1 left 13:15 [particle] joined 13:16 kaleem left 13:17 bluescreen10 joined 13:18 kaleem joined 13:24 driador left, driador joined
eiro masak, jnthn i have to know for sunday: i'll go buy the stuff in a hour 13:33
13:35 atrodo joined 13:38 thilp left
masak that's a bit of a short notice. I'd give you the details but I don't have them. 13:45
I know we'll be going home on the Sunday but I don't know when. 13:46
probably early-ish.
hope that's enough.
not_gerd bye, #perl6 13:48
13:48 not_gerd left 13:55 thilp joined
eiro masak, ok i'll buy enought for you, hope you'll be there 13:55
13:57 fgomez left 13:59 JimmyZ left 14:00 JimmyZ joined
masak I hope so to, but I think it might be too time-constrained. 14:02
eiro i understand 14:07
14:22 fgomez joined 14:24 kaleem left 14:27 arlinius joined 14:29 fglock left, hoelzro joined 14:30 fglock joined
jnthn I don't have the times to hand for our train, but I know we'll be heading of relatively early on Sunday. 14:30
14:33 mikemol joined
TimToady dislikes camelCase 14:36
14:36 thilp left, Slacky joined 14:37 thilp joined 14:41 vmspb joined
jnthn masak: The boat bit is true. Our train drives onto a boat, and then is transported from Denmark to Germany. They're gonna build a tunnel there by 2017 or something... 14:44
arnsholt Laziness truly is a virtue: thecodelesscode.com/case/29 14:47
14:49 thou joined 14:57 wtw left
jnthn decommute & 14:57
15:00 fgomez left 15:13 szabgab joined
szabgab hi 15:13
moritz \o szabgab
Slacky hi szabgab
moritz jnthn: I remember that boat trip. Train on a ferry is quite a nice thing IMHO :-)
szabgab I am trying to get away from most of IRC but it is quite hard with perl 6 :)
sisar arnsholt: the whole series is fun to read :) 15:14
moritz iirc Puttgarden is one of the endpoints
szabgab I wonder if it would be ok with you guys if I started to post question on the mailing list ?
and if it is ok, which one should I use?
moritz szabgab: sure, and I'll try to answer them
arnsholt sisar: Yeah, I just read the whole thing
szabgab [email@hidden.address] ?
arnsholt The testing one especially struck me as Perlish 15:15
moritz szabgab: depends on your questions, but [email@hidden.address] if you don't know for sure
sisar arnsholt: thanks for bringing it to my notice.
moritz perl6-compilers when asking about rakudo or niecza
szabgab ok, I'll start with perl6-users as that's my POV and let me know if some question would be better discussed on the other perl6-complers list, I think I am on both 15:16
moritz szabgab: ok; both are pretty low volume, so it doesn't matter too much 15:17
15:17 birdwindupbird left
szabgab Let's hope that generating some volume there is good for perl6 :) 15:19
15:20 hoelzro is now known as hoelzro|away
szabgab bye& 15:20
15:20 szabgab left 15:26 alester joined 15:28 JimmyZ left 15:29 mikemol_ joined 15:31 mikemol left, dukeleto left
jnthn home 15:36
15:36 fglock left
jnthn moritz: Yes, Lubeck to Puttgarden :) 15:36
dalek kudo/parrot-iocleanup1: d0dc2d3 | moritz++ | src/core/IO/Socket/INET.pm:
explicitly set encoding on sockets
 15:40
15:45 dukeleto joined 15:48 skids left 15:49 cog_ joined, spider-mario joined
dalek ecs: 10d772f | larry++ | S03-operators.pod:
Document [&func] form of infix
 15:50
vmspb Hello 15:51
prn: say [-](); say [-](1)
p6eval rakudo 88a9d6, niecza v19-1-gbd1bd16, pugs: OUTPUT«0␤1␤»
vmspb [-](i) should be equal to (i) or (0 - i)? It falls out of the system.
prn: say [~]() == ''; say [*]() == 1; say [+]() == 0; say [-]() == 0;
p6eval rakudo 88a9d6, niecza v19-1-gbd1bd16: OUTPUT«True␤True␤True␤True␤»
..pugs: OUTPUT«1␤1␤1␤1␤»
15:51 cognominal left
vmspb prn: say [~](1) == '' ~ 1; say [*](2) == 1 * 2; say [+](3) == 0 + 3; say [-](4) == 0 - 4; 15:51
p6eval rakudo 88a9d6, niecza v19-1-gbd1bd16: OUTPUT«True␤True␤True␤False␤»
..pugs: OUTPUT«1␤1␤1␤␤»
TimToady while an argument could be made for such behavior, reduction on non-commutative ops is rather non-sensical to begin with 15:53
and you can always add the 0 explicitly if you want the other behavior 15:54
dalek ar: 5138e08 | moritz++ | skel/docs/announce/2012.06:
fiddle a bit with the release announcement
15:54 sergot left
moritz ok, if nobody objects I'll release star tomorrow 15:55
TimToady plus returning the first arg when there's only one works for reductions on operators that have no meaningful identity value
15:56 fgomez joined 15:57 jaldhar left 15:59 Slacky left
kresike bye all 16:00
16:00 kresike left 16:04 brrt left 16:11 sergot joined 16:13 MayDaniel joined
sorear good * #perl6 16:18
TimToady o/ 16:20
sorear: is there any way to get a list of available UCD tables?
unidata is kinda opaque... 16:21
sorear TimToady: not at the moment from within niecza; you may have better luck with github.com/sorear/multi-language-u...ucd-xml.pl 16:26
TimToady thanks
cxreg as someone who never really learned BNF in depth, is there a post any where that compares it to perl 6 grammars? 16:28
diakopter what's a domirrror
cxreg if not, i may have to learn enough to write one, because i think that would be useful 16:29
16:32 mikemol_ left
dalek ast: 9ab47bb | pmichaud++ | S03-metaops/zip.t:
Add tests for proper list associativity with Z, (RT #113800).
 16:33
moritz r: say (<a b c> Z <d e f> Z <g h i>).perl
p6eval rakudo 88a9d6: OUTPUT«(("a", "d", "g"), ("b", "e", "h"), ("c", "f", "i")).list␤»
sirrobert Do perl6 modules need to end true? (1;)
moritz no 16:34
masak sirrobert: no.
sirrobert great =)
moritz it was even an anachronism in p5, IMHO
16:34 thilp left, daxim left
masak "was"? :) 16:34
moritz if you wanted the compilation to fail, you simply die()d
well, it was 10 years ago
and stil is :-)
so, "has been rather"
sirrobert btw, someone (moritz?) requested I blog about adopting perl6 for my business. You can see it (in it's 1-post glory) at 6ing.wordpress.com/ 16:35
masak what's the use case for wanting the compilation to fail?
sirrobert++
moritz masak: I dunno, maybe if you want to load a data file at compile time, and it can't find that file, die?
but then it dies anyway
sirrobert++ 16:36
masak "design for semantic compression at the cost of syntactic compression" # nice! 16:39
sirrobert =) 16:40
moritz sirrobert: thank you for that blog post, it's really nice
sirrobert you're welcome =)
moritz subscribes to the RSS feed 16:41
sirrobert What's the irc command to do that?
* sirrobert whatever.
benabik It's `/me whatever.` 16:42
benabik whatever.
sirrobert ok:
sirrobert is curious how his CTO will respond when he comes back from vacation =)
moritz sirrobert: to you blogging, or to you chosing p6? :-) 16:43
sirrobert p6
masak sirrobert: really curious about hearing more about your adventures in Perl 6 land.
sirrobert me too. I posted a bug report the other day
a problem parsing grammars.
16:44 mucker joined
masak nice. 16:44
moritz ah, the one about implicit <.ws> mishandling?
sirrobert apparently: rule { ^ <foo> $ } parses differently than: rule {^ <foo> $}
yeah
masak yeah, that's wrong.
jnthn That used to be how it was spec'd to.
sirrobert huh
jnthn Until fairly recently.
pmichaud fixing that one is on my todo list :) 16:45
jnthn totally agrees with the spec change though :)
moritz wonders if it's as easy adding an s* somewhere
jnthn No :)
moritz :( 16:46
jnthn Well, maybe, but they you discover that NQP's and Rakudo's grammars rely on the current semantics.
sirrobert I tried a lot of variations of \s*
moritz sirrobert: I mean, insides the regex parser
which is, I might add, surprisingly short 16:47
sirrobert oh
pmichaud I can work on that one now, a bit.
sirrobert thanks =) 16:48
TimToady r: constant @t = [\+] 1..5; say (0,@t Z^.. @t).perl;
p6eval rakudo 88a9d6: OUTPUT«(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15).list␤»
TimToady that one still doesn't produce sub-parcels
pmichaud: ^^
jnthn r: my @t = [\+] 1..5; say (0,@t Z^.. @t).perl;
p6eval rakudo 88a9d6: OUTPUT«(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15).list␤»
jnthn OK, and constant ain't to blame this time :) 16:49
masak submits rakudobug
pmichaud TimToady: okay, looking.
TimToady probably Range behaving insufficiently like a Parcel somewhere
16:50 kaare_ joined
pmichaud r: my @t = [\+] 1..5; say @t.perl 16:50
p6eval rakudo 88a9d6: OUTPUT«Array.new(1, 3, 6, 10, 15)␤»
pmichaud r: my @t = [\+] 1..5; say (0,@t Z~ @t).perl
p6eval rakudo 88a9d6: OUTPUT«("01", "13", "36", "610", "1015").list␤»
pmichaud r: my @t = [\+] 1..5; say (0,@t Z, @t).perl 16:51
p6eval rakudo 88a9d6: OUTPUT«((0, 1), (1, 3), (3, 6), (6, 10), (10, 15)).list␤»
pmichaud r: my @t = [\+] 1..5; say (0,@t Z.. @t).perl
p6eval rakudo 88a9d6: OUTPUT«(0, 1, 1, 2, 3, 3, 4, 5, 6, 6, 7, 8, 9, 10, 10, 11, 12, 13, 14, 15).list␤»
pmichaud okay, I can look at it.
16:51 aloha left 16:52 thilp joined
TimToady n: constant @t = [\+] 1..5; say (0,@t Z^.. @t).perl; 16:53
p6eval niecza v19-1-gbd1bd16: OUTPUT«(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15).list␤»
TimToady n: constant @t = [\+] 1..5; say (0,@t Z^.. @t)[3].perl;
p6eval niecza v19-1-gbd1bd16: OUTPUT«6^..10␤»
TimToady niecza has a similar difficulty somewhere 16:54
also, doesn't have .lol yet
16:54 sudokode joined
TimToady r: constant @t = [\+] 1..5; say (0,@t Z^.. @t).lol[3].perl; 16:56
p6eval rakudo 88a9d6: OUTPUT«Any␤»
TimToady and rakudo's .lol doesn't lol it right either
pmichaud it looks like some premature flattening in the Z metaop, to me. 16:57
16:57 GlitchMr joined
sirrobert lunch 16:57
pmichaud r: say ((1,2) Z.. (3,4)).perl
p6eval rakudo 88a9d6: OUTPUT«(1, 2, 3, 2, 3, 4).list␤»
TimToady but then why would lol return Any?
16:58 aloha joined
pmichaud not sure 16:58
TimToady r: constant @t = [\+] 1..5; say (0,@t Z^.. @t).lol[0].perl;
p6eval rakudo 88a9d6: OUTPUT«ListIter.new()␤»
TimToady hates it just about every time he sees ListIter.new()... 16:59
pmichaud agreed; I haven't figured that one out yet either.
TimToady eager doesn't seem to help with that either
pmichaud basically, ListIter.new() is saying "something lazy here"
also, I think when it shows up it's been itemized
TimToady r: constant @t = [\+] 1..5; say (0,@t Z^.. @t).lol[0].eager.perl;
p6eval rakudo 88a9d6: OUTPUT«(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15).list␤»
TimToady well, that time it helped
pmichaud I can at least get ListIter.perl to show when it's an item 17:00
TimToady but .perl should probably consider itself eager, unless it can reproduce the ..* ish ness
17:00 dakkar left
masak +1 17:00
pmichaud I haven't figured out what ListIter.perl should look like in that case, though. 17:01
I suppose I could get ListIter to perlify it's generator 17:02
*its
sisar r: say $::; 17:08
p6eval rakudo 88a9d6: OUTPUT«===SORRY!===␤ResizablePMCArray: index out of bounds!␤»
sisar oops :)
n: say $::;
p6eval niecza v19-1-gbd1bd16: OUTPUT«Any()␤»
sisar std: say $::; 17:09
p6eval std fd2647b: OUTPUT«ok 00:00 40m␤»
jnthn Hmm, I thought I fixed that issue once... 17:12
I've no idea what it actually means :) 17:13
sisar too :)
masak jnthn: "index out of bounds" means the index was either larger than the last index of the array, or negative. now you know. :P
jnthn
.oO( You can't just type shit and expect it to do something sensible! )
sisar lol
jnthn masak: :P
masak is so helpful 17:14
jnthn Some things never change. :)
17:14 immortal joined, immortal left, immortal joined, immortal left, erkan left
pmichaud Looks like the Z^.. is an issue with gather/take 17:19
sisar can someone add sirrobert++'s blog to planetsix.perl.org ?
or teach me how to do that ? 17:20
pmichaud gist.github.com/3005502 # take is flattening the range
17:21 erkan joined, erkan left, erkan joined
pmichaud either that or the list managing the gather is flattening it. 17:21
gist.github.com/3005523 # more detailed example 17:22
jnthn Hmm...gather/take and flattening sounds familiar... 17:24
(As in, I think we've had issues in this space before.)
pmichaud TimToady: ping 17:25
jnthn hmm... rt.perl.org/rt3/Public/Bug/Display...?id=106986
But it's marked resolved.
pmichaud jnthn: that one involves scalar arrays
the scalar part wasn't being preserved.
jnthn *nod* 17:26
17:28 vmspb left
TimToady pmichaud: pong 17:32
pmichaud okay, I see where the problem with Z^.. is, and I'm not quite sure how to address it.
Here's the fundamental issue: 17:33
r: my @a = <x y>; my @l := gather { take @a; take @a; }; say @l.perl
p6eval rakudo 88a9d6: OUTPUT«("x", "y", "x", "y").list␤»
pmichaud the results of the take's are being flattened
TimToady yes, each take should result in a parcel, methinks 17:34
p: my @a = <x y>; my @l := gather { take @a; take @a; }; say @l.perl 17:35
p6eval pugs: OUTPUT«\("x", "y", "x", "y")␤»
TimToady n: my @a = <x y>; my @l := gather { take @a; take @a; }; say @l.perl
p6eval niecza v19-1-gbd1bd16: OUTPUT«("x", "y", "x", "y").list␤»
TimToady everyone else has the same problem
pmichaud if we do that, then take 1; take 1; would result in a list like ((1,), (1,))
I don't know that this would be a problem... but it would seem... weird.
TimToady well, scalars are allowed to be their own parcel in general 17:36
pmichaud it's actually not an issue of scalarness (more)
because 1..5 is a scalar
17:36 dukeleto left
pmichaud unless you mean "scalar" as "non-interpolating" 17:36
TimToady but something is treating it slurpily and flattening it
pmichaud sure, because Arrays and Ranges are Iterable, thus they interpolate in lists 17:37
TimToady 1..5 works as a single arg when binding to a non-slurpy, so what's different here?
17:37 sirrobert left
TimToady npr: sub foo($x) { $x.say }; foo(1..5) 17:37
p6eval pugs: OUTPUT«12345␤» 17:38
..rakudo 88a9d6, niecza v19-1-gbd1bd16: OUTPUT«1..5␤»
pmichaud as a single arg we're referring to its Positional nature (which doesn't have anything to do with flattening)
(if binding to a positional)
I'll put it another way -- slurpiness isn't entering into it at all. 17:39
TimToady r: my @a = <x y>; my @l := gather { take @a; take @a; }; say @l.DUMP
p6eval rakudo 88a9d6: OUTPUT«List<1956749320>(:items(▶Mu), :nextiter(ListIter<1956749344>(:reified(▶Mu), :rest(RPA<1956749364>(GatherIter.new())), :list(List<1956749320>))))␤»
pmichaud add an 'eager' @l first
to flatten out the GatherIter
TimToady r: my @a = <x y>; my @l := gather { take @a; take @a; }; say @l.eager.DUMP
p6eval rakudo 88a9d6: OUTPUT«List<-1931207599>(:items(RPA<-1931174844>(▶"x", ▶"y", ▶"x", ▶"y")), :nextiter(▶Mu))␤»
pmichaud r: my @a = <x y>; my @l := gather { take @a; take @a; }; @l.eager; say @l.DUMP 17:40
p6eval rakudo 88a9d6: OUTPUT«List<-135657117>(:items(RPA<-135607960>(▶"x", ▶"y", ▶"x", ▶"y")), :nextiter(▶Mu))␤»
pmichaud the list resulting from the gather ends up looking like (@a, @a)
when then interpolates the @a's whenever you start doing anything with them 17:41
(because they're iterable)
same thing for ranges -- the List resulting from the gather ends up being (1..4, 1..4)
and since Range is iterable, it interpolates.
TimToady which is why the old version of S07 distinguished get from getarg 17:42
pmichaud just to make sure we're on the same page, note that "interpolate" ne "flattens"
TimToady it needs to be possible to iterate a lol without flattening things like these 17:43
pmichaud oh!
TimToady okay, interpolating
pmichaud so, gather needs to return a LoL, then?
that doesn't feel right either
because it won't flatten.
TimToady a parcel of parcels/things that can behave like parcels
what we used to call "arguments"
pmichaud flattening is what happens (or doesn't happen) to parcels 17:44
Parcel !~~ Iterable, thus it doesn't interpolate by default, interpolation of a Parcel is "flattening"
TimToady how does the binder avoid this problem?
pmichaud the binder never has this problem -- it doesn't ever involve interpolation or flattening except when doing slurpies 17:45
TimToady it still has to "iterate" the argument list somehow
what that is doing is what we need to do for take and return values when fed to things like .lol or .perl
"getarg" in the old-fashioned terms 17:46
pmichaud I think it's the gather that is interpolating, not the take
the take is correctly returning things 17:47
basically, each take is like a "getarg"
17:47 aj7504 joined
pmichaud if I say "take 1..5" it gives me a single range 17:47
(unflattened, uninterpolate)
gather then has to put that into a structure
TimToady so maybe gather is not being quite primitive enough 17:48
pmichaud and that structure is what ends up interpolating
TimToady r: say (1..5, 2..7).perl
p6eval rakudo 88a9d6: OUTPUT«(1..5, 2..7)␤»
pmichaud r: say (1..5, 2..7).list.perl 17:49
p6eval rakudo 88a9d6: OUTPUT«(1, 2, 3, 4, 5, 2, 3, 4, 5, 6, 7).list␤»
pmichaud that's the difference. :-)
the fact that the return from gather is a List means it interpolates.
TimToady yes, so gather needs to de-listify itself somehow, I suppose
pmichaud ...but it can't really be a Parcel, unless we start adding laziness to Parcels.
17:50 dukeleto joined
TimToady well, we just need a kind of list that doesn't flatten subparcels then 17:50
pmichaud ....no
we need a list that doesn't flatten Iterables
we already have a list that doesn't flatten subparcels
r: say ((1..5,), (2..7,)).list.perl 17:51
p6eval rakudo 88a9d6: OUTPUT«((1..5,), (2..7,)).list␤»
pmichaud I have to leave in 2 mins to pick up kid, but will be back in 15
TimToady so maybe the real problem is better described as Iterables don't properly behave as subparcels?
pmichaud if we make Iterables behave as subparcels, we lose all sorts of flattening behavior that we currently expect
e.g., (1..5, 2..7).list.elems would end up being 2 17:52
as would (@a, @b).list.elems
afk for 15
TimToady only if list is the same as parcel-list
later
or maybe we've confused .list with .flat (yet again) 17:55
17:56 aj7504 left 18:02 mikemol joined
TimToady in any case we seem to have de-lazified interpolation too much such that the recipient of the whatever-it-is is not able to make the decision 18:02
18:02 thilp left 18:03 thilp joined
TimToady if we have to make .elems flattening by default, and add a .args to get the other value, we could do that 18:03
18:05 skids joined
TimToady or a .rawelems, or whatever 18:08
sisar r: print 1..10; 18:11
p6eval rakudo 88a9d6: OUTPUT«12345678910»
sisar r: print (1..10).list;
p6eval rakudo 88a9d6: OUTPUT«12345678910»
sisar r: print (1..10).WHAT;
p6eval rakudo 88a9d6: OUTPUT«use of uninitialized variable $v of type Range in string context in block <anon> at /tmp/kISW4FfZ3Q:1␤␤»
sisar huh ? 18:12
r: print (1..10).WHAT.gist;
p6eval rakudo 88a9d6: OUTPUT«Range()»
sisar r: print (1..10).WHAT.perl;
p6eval rakudo 88a9d6: OUTPUT«Range»
pmichaud back again 18:15
18:15 arlinius left 18:16 Je joined
sisar nr: say (1, 2, 3, 4, 5).Range; #ooc, can the compiler figure out a Rnage which will generate this list? Expected output here would be "1..5" ;-) 18:16
p6eval niecza v19-1-gbd1bd16: OUTPUT«Unhandled exception: Unable to resolve method Range in type Parcel␤ at /tmp/QtmVnXOV6_ line 1 (mainline @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3917 (ANON @ 3) ␤ at /home/p6eval/niecza/lib/CORE.setting line 3918 (module-CORE @ 562) ␤ at /hom…
..rakudo 88a9d6: OUTPUT«No such method 'Range' for invocant of type 'Parcel'␤ in block <anon> at /tmp/zxXEIY5Hlx:1␤␤»
pmichaud well, it seems to me that .list has to interpolate, in order for @(1..5) to do what we want.
as things stand now, .list interpolates Iterables, while .flat interpolates Iterables and Parcels 18:17
18:17 cognominal joined 18:18 thilp left
pmichaud what we really might want is something between .list and LoL, or (equally likely) that LoL needs a redefinition 18:18
then we'd have
we'd want potentially four levels, then: 18:19
.flat -- interpolate Iterables and Parcels
.list -- interpolate Iterables (but not Parcels)
.xyz -- don't do any interpolation
.uvw -- itemize everything 18:20
.uvw is what we currently do for .lol and LoL
i.e., (1..5, 6..10).lol results in a list where the ranges have been put into scalars
(so they can't be interpolated any further by any subsequent operator)
18:21 cog_ left
pmichaud r: say (1..5,6..10).lol.perl 18:21
p6eval rakudo 88a9d6: OUTPUT«LoL.new(1..5, 6..10)␤»
pmichaud r: say (1..5,6..10).lol.list.perl
p6eval rakudo 88a9d6: OUTPUT«LoL.new(1..5, 6..10)␤»
pmichaud hmmm
18:22 Je left
pmichaud r: my @a = 1,2; say (@a, @a).lol.[0] 18:22
p6eval rakudo 88a9d6: OUTPUT«1 2␤»
colomon phenny: tell sorear I don't have [&atan2] working, just wanted to thank you for your help (and if you're patient enough, the help that I still require).
phenny colomon: I'll pass that on when sorear is around.
pmichaud r: my @a = 1,2; say (@a, @a).lol.[0].perl
p6eval rakudo 88a9d6: OUTPUT«[1, 2]␤» 18:23
pmichaud (note non-flattening)
TimToady what is the use case for interpolating Iterables but not Parcels?
pmichaud ummmm 18:24
"that's what distinguishes .flat from .list" (yes, I know that's not a good answer here... still thinking)
18:24 sergot left
pmichaud I'm certain there was a case at one time -- that's how we ended up with .flat and .list. Let me review the core and see if I can recall which case it's handling. 18:27
TimToady has to run errands now...
18:27 fhelmberger left, snearch joined
pmichaud the distinction may have had to do with with flattening/non-flattening versions of .map 18:28
which .map always flattens now, so might not be relevant anymore
18:29 arlinius joined
sorear o/ colomon 18:30
phenny sorear: 18:22Z <colomon> tell sorear I don't have [&atan2] working, just wanted to thank you for your help (and if you're patient enough, the help that I still require).
colomon \o
pmichaud I guess I can turn .list into .flat and see what breaks :-) 18:31
18:31 thilp joined 18:32 adu joined
colomon sorear: gist.github.com/3005906 is what I have at the moment 18:32
pmichaud mainly I think that .list versus .flat is what makes Parcel do what it does -- a Parcel is something where the interpolating decision is postponed. 18:33
colomon when I try it, I get "Cannot use hash access on an object of type Str"
sorear: that's after infixish returns, as far as I can tell
18:35 sergot joined
sorear colomon: which line do you get that error on? 18:38
colomon sorear: I'm getting it from the REPL when I try 3 [&atan2] 4 18:39
sorear colomon: might be more useful to try from -e so that you have a stack trace. also, NIECZA_FAIL_FAST=1 to get stack traces from sorries 18:40
colomon sorear: gist.github.com/3005936 18:41
hmmm... my $inprec = $inO<prec>; 18:42
looks like it is the problem.
18:42 benabik left
sorear colomon: $O = $<infix><O> // <O(|%additive)>; # spot the problem here 18:44
I think // %additive; would work 18:45
colomon hmmm.... we're creating a precedence level for the new operator, I get that. but I don't particularly get what's wrong with <O(|%additive)> ... but then, I don't understand what it means, either. 18:46
pmichaud r: say (gather { take 1,2,3; take 4,5,6 }).elems
p6eval rakudo 88a9d6: OUTPUT«2␤»
pmichaud r: say (gather { take 1,2,3; take 4,5,6 }).flat.elems
p6eval rakudo 88a9d6: OUTPUT«6␤»
pmichaud that's one difference between .flat and .list
flussence rn: regex { <[a b c #]> } 18:47
p6eval rakudo 88a9d6: ( no output )
..niecza v19-1-gbd1bd16: OUTPUT«===SORRY!===␤␤Unrecognized regex metacharacter < (must be quoted to match literally) at /tmp/hE1bMAVFBP line 1:␤------> regex { <⏏[a b c #]> }␤␤Unhandled exception: Unable to resolve method ast in type Any␤ at /home/p6e…
18:49 dukeleto left
colomon sorear: errr, that worked! 18:50
pmichaud but I guess that difference is exactly the error we've been dealing with.
colomon Wynne:niecza colomon$ NIECZA_FAIL_FAST=1 mono run/Niecza.exe -e "say 3 [&atan2] 4"
0.64350110879328437
pmichaud s/error/problem/
sorear colomon: <O(|%additive)> is a string literal.
colomon: that's what's wrong with it.
colomon what is %additive?
sorear %additive is the precedence level of + and - 18:51
the default precedence level for newly defined infixes
colomon sorear: I understand that, but I don't understand why it's a hash variable
pmichaud colomon: the hash variable defines precedence, associativity, and other things for operators at that level 18:52
constant %additive = (:dba('additive') , :prec<t=>, :assoc<left>, :pure);
which says that operators at the %additive level are precedence "t=", left associative, and "pure"
colomon oh! so it's like a simple old-style p5 class.
pmichaud++, sorear++ 18:53
I'm assuming we don't have any tests for [&func] in roast? 18:54
btw, Wynne:niecza colomon$ NIECZA_FAIL_FAST=1 mono run/Niecza.exe -e "say 3 R[&atan2] 4"
0.92729521800161219
pmichaud probably not; especially if it hasn't been put into the synopses
18:54 sisar left 18:57 sisar joined
sisar pmichaud: TimToady++ just specced it ([&func]) today github.com/perl6/specs/commit/10d772fdfd 18:58
pmichaud sisar++ # thanks 18:59
pmichaud takes a walk to think more about .flat and .list
19:01 MayDaniel left 19:02 mikemol left, brrt joined 19:04 eric left 19:14 sirrobert joined
felher sirrobert++ #nice blogpost and using perl6 :) 19:16
sirrobert =) thanks
What's the bit of code that installs perl6 modules (such as prerequisites that ufo finds)? 19:17
Or, rather, I mean: How can I specify where perl6 modules are installed to? 19:18
pmichaud r: my @a = 1..10; say ((1,2,3), @a).elems; 19:20
p6eval rakudo 88a9d6: OUTPUT«13␤»
colomon sorear: ye gads, we're failing spectests left and right today. (nothing to do with my code, I stashed my changes to check on things.)
pmichaud r: my @a = 1..10; say ((1,2,3), @a).list.elems; 19:21
p6eval rakudo 88a9d6: OUTPUT«11␤»
sorear colomon: when I tested before the release I saw about a dozen failures, all seemingly new test 19:22
19:22 mikemol joined
brrt good afternoon #perl6 19:22
19:23 sisar left
pmichaud TimToady: I'm still playing with things a bit... but what would be the expected answer of...: my @a = 1,2,3; say (gather { take @a; take @a; }).elems 19:23
brrt i was warned that there would be some magic involved in getting nqp and the rest of parrot to interopterate
19:23 benabik joined
brrt so i'm here with full curiosity as to what that magic entails 19:24
pmichaud Rakudo doesn't like dealing with bare Parrot PMCs, because they don't fit into Rakudo's type hierarchy or dispatch system.
brrt thats fair enough 19:25
but i do need to run a script with rakudo as the compiler
colomon sorear: yeah, I do suspect that's it. I just hate checking in with a lot of failing spectests
brrt or nqp, that will be a start
pmichaud I don't understand "run a script with rakudo as the compiler". 19:27
(at least not in this context.)
brrt yes, that is vague
long story short, what happens is that i get a request, i determine based on configuration which compiler should compile the file that is requested, i compile the file, and return the results 19:28
rakudo (should be) one of those compilers 19:29
as winxed is now
and nqp-rx probably works, too
19:29 rodent_ joined
pmichaud so, you want Rakudo to be able to work with load_language, yes? 19:30
brrt yes
and compreg() and the like
19:30 GlitchMr left
pmichaud I'm pretty sure it already participates with compreg() 19:30
brrt :yes:
19:31 bluescreen10 left
jnthn r: pir::compreg__Ps('perl6').eval('say 42') 19:31
p6eval rakudo 88a9d6: OUTPUT«42␤»
brrt oh, nice
pmichaud and "make install" already installs Rakudo into languages/perl6 19:32
so you should just be able to do "load_language 'perl6'", and compreg and have everything "work".
pmichaud tests
19:32 bluescreen10 joined
brrt i'm going to try that out 19:32
however, embarassingly, i do not have p6 installed right now
colomon n: say [max] Any, Any, 2 19:33
p6eval niecza v19-1-gbd1bd16: OUTPUT«-Inf␤»
colomon r: say [max] Any, Any, 2 19:34
p6eval rakudo 88a9d6: OUTPUT«2␤»
colomon rn: say Any cmp 2
p6eval niecza v19-1-gbd1bd16: OUTPUT«Increase␤»
..rakudo 88a9d6: OUTPUT«use of uninitialized value of type Any in string context in block <anon> at /tmp/aQDVVXrGdU:1␤␤Increase␤»
pmichaud colomon: I think the problem is that 'max' in niecza treats a list with an undefined first element as being empty
n: say [max] 3, Any, 4 19:35
colomon rn: say Any max 2
p6eval niecza v19-1-gbd1bd16: OUTPUT«4␤»
rakudo 88a9d6: OUTPUT«2␤»
..niecza v19-1-gbd1bd16: OUTPUT«-Inf␤»
pmichaud max in niecza treats a list with an undefined first value as being empty, and thus returns '-Inf'
(without checking any other elements)
colomon pmichaud++ # think that's it 19:36
colomon was surprised to learn that infix:<max> takes a list! 19:38
colomon is scared to check blame and see if he actually wrote it that way... 19:39
pmichaud infix:<max> takes a list -- yes -- it has %tight_or precedence, which has :assoc<list>
constant %tight_or = (:dba('tight or') , :prec<k=>, :assoc<list>); 19:40
colomon is Any cmp 2 actually well-defined by the spec?
brrt is cooling his computer for the next few minites, bye &
19:40 brrt left
[Coke] sirrobert++ sirrobert++ 19:40
pmichaud since the two are of differing types, it defaults down to string context (or at least did at the time that Rakudo's implementation of cmp was last worked on :)
19:44 thilp left, birdwindupbird joined
colomon I don't see any hint of that in the Comparison semantics portion of the spec. But then, I don't see any support for the Niecza implementation, and I did heavy hacking on that one. Gack, I regret the fact that everything I did all winter is just a blur now.... 19:45
dalek ast: 3bddd1b | (Solomon Foster)++ | / (3 files):
Fudge for niecza.
 19:46
pmichaud brrt / jnthn: It appears that "load_language 'perl6'" is not sufficient; I suspect that perl6.pir's mainline Sub never gets executed.
19:46 cognominal left 19:48 cognominal joined
pmichaud it may be possible to fix that more easily now if Parrot has added the new sub tagging features (i.e., better introspection on whatever the Eval PMC has now morphed to :-) 19:49
19:49 szabgab joined, snearch left
pmichaud afk for a bit 19:49
szabgab and I just wanted to ask pmichaud about a compilation failure :) 19:50
after Stage post it is getting killed : make: *** [CORE.setting.pbc] Error 137
it is an ubuntu 12.04 32 bit 19:51
anyone maybe knows what to do?
sirrobert What does this mean (trying to use panda): Missing or wrong version of dependency 'src/gen/Metamodel.pm' 19:54
besides the obvious ;) Any ideas how to fix it? 19:55
szabgab sirrobert: I think I installed panda again
sirrobert I tried that, but got the same error ;)
ahh, because the script uses "panda install panda" 19:57
heh
20:00 fridim_ joined
szabgab hmm, it seem I ran out of memory and swap space 20:03
20:04 brrt joined
sirrobert szabgab: I hit that before... needed to use more than the RAM I had available on my server 20:05
(VPS with 1GB RAM running Ubuntu server)
szabgab sirrobert: and di you have any solution?
was 1Gb enough?
sirrobert nope
szabgab oh
sirrobert temporary solution: develop locally =)
szabgab yes, I can build Rakudo at home and upload to the server 20:06
sirrobert Since I have a VPS I might take a disk image, boot it up on my computer to compile, then pass it back
colomon pugs: say 2 max Any 20:08
p6eval pugs: OUTPUT«*** ␤ Unexpected "max"␤ expecting operator, ":" or ","␤ at /tmp/ZfygxIXHhk line 1, column 7␤»
20:11 brrt left
szabgab I'll try to build at home 20:12
20:14 bluescreen10 left, brrt joined 20:16 szabgab left
dalek ast: 677362a | (Solomon Foster)++ | S03-operators/minmax.t:
Fudge for niecza, add note doubting the correctness of the tests.
 20:16
ast: 518ae1d | (Solomon Foster)++ | S06-signature/slurpy-and-interpolation.t:
Fudge for niecza.
 20:17
ecza: b5cc8a1 | (Solomon Foster)++ | t/spectest.data:
Turn off dir.t test.
pmichaud colomon: I can't find it in the docs, but I'm pretty sure that min and max are to ignore undefined values in their computation 20:18
thus my $x; my $y; $x min $y min 2 returns 2 20:19
which is entirely unrelated to the result of Any cmp 2 20:20
colomon pmichaud: that certainly sounds reasonable-ish.
pmichaud similarly: 20:21
r: my @a; @a[2] = 20; say [max] @a
p6eval rakudo 88a9d6: OUTPUT«20␤»
tadzik good evening
pmichaud r: my @a; @a[2] = 20; say [mi] @a
p6eval rakudo 88a9d6: OUTPUT«===SORRY!===␤Confused␤at /tmp/p0UOjBvlq0:1␤»
pmichaud r: my @a; @a[2] = 20; say [min] @a
p6eval rakudo 88a9d6: OUTPUT«20␤»
20:22 brrt1 joined 20:23 brrt left
colomon pmichaud: though I don't see it anywhere in the spec either, and it seems like it interacts a bit strangely with the min and max methods. 20:23
pmichaud colomon: ?
how so
? 20:24
colomon (to be more precise... if there is a user defined ordering, do we do still skip undefineds?)
pmichaud rakudo does that, yes.
20:27 bluescreen10 joined 20:28 larks left, larks joined, mikemol left
pmichaud although it wouldn't bother me if Any cmp 2 returned "Mu" or some other value to indicate "I don't know the ordering" 20:29
then min/max could simply skip elements where the comparison came back as "Mu" 20:30
that would enable someone to write a custom comparator that handled undefined elements 20:31
diakopter 15:21 < pmichaud> r: my @a; @a[2] = 20; say [mi] @a
sirrobert Has anyone else encountered the panda problem ("missing or wrong version of Metamodel.pm") and know how to fix it?
pmichaud sirrobert: whenever I've gotten that error message, it's either because I didn't install perl6 or because my PERL6LIB environment variable isn't set. 20:32
sirrobert I just reinstalled perl6, so it's not that...
tadzik sirrobert: yes
jnthn Generally it means there's a pre-compiled module hanging around compiled for a previous Rakudo build.
PerlJam sirrobert: and I've gotten it when the perl6 I'm running isn't the one I thought it was.
sirrobert looks like it might be $PERL6LIB
tadzik sirrobert: if you've reinstalled perl6, you need to reinstall panda as well
20:33 adam7504 joined
sirrobert is in another terminal a min... 20:33
20:34 bluescreen10 left
pmichaud also, it would be really helpful if that error message could be split up to tell us which case has occurred; i.e. "missing Metamodel.pm" versus "wrong version of Metamodel.pm" 20:36
and if it got the wrong version, it would be helpful to let us know which version it got and which one it was looking for :)
tadzik I don't think anything is missing, it's just precompiled against another version of Metamodel.pir 20:37
PerlJam Seems like that could easily become a FAQ in the future. 20:38
tadzik it's been a FAQ in the past too :)
pmichaud which is why I'm saying the error message can help to avoid the FAQ :)
tadzik and in the present
jnthn Which is why I keep saying that Rakudo should manage the pre-compilation cache.
tadzik +1
sirrobert $PERL6LIB was unset. I set it to ~/.perl6/lib and got a different error for panda:
Missing or wrong version of dependency 'src/stage2/QRegex.nqp' 20:39
20:39 arlinius left
tadzik fun 20:39
moritz sirrobert: that always means that you're loading a .pir file that was compiled with an older rakudo
tadzik anyway, I'm quite convinced the reason is the same
sirrobert Ok... so a resolution would be
jnthn tadzik: Does panda have a rebuild-all-the-things option yet?
tadzik jnthn: nope
sirrobert remove ~/.perl6 and re-install everything? 20:40
moritz sirrobert: try RAKUDO_MODULE_DEBUG=1 perl6 yourscript.pl
tadzik the bootstraping problems :)
moritz sirrobert: yes
sirrobert ok =)
tadzik maybe one could implement that in BEGIN, hmm
before module loading
jnthn figures that we should have all the needed people on hand at YAPC::EU to try and hammer out a good solution for all of this.
PerlJam sirrobert: Just remove all of the .pir files (they'll be made anew from the .pm files) 20:41
tadzik I see a way, and it's going to be SOO dirty hack :)
PerlJam: they won't, no :(
but things will work
sirrobert What's .pir, btw?
PerlJam tadzik: oh, they're only made on install?
tadzik on build, to be exact
pmichaud I'm still quite uncomfortable with the "rakudo manages the cache" approach; it gives me that tension in my shoulders that says "wrong wrong wrong". 20:42
PerlJam sirrobert: Parrot Intermediate Representation or somesuch. It's the language of the underlying Parrot VM.
pmichaud certainly I have trouble with "the" cache :)
sirrobert ah, got it
jnthn pmichaud: Caches :P
pmichaud: I don't love it either, fwiw.
pmichaud but yes, I think yapc::eu is a good place to explore it.
diakopter BOOTArray->st is null
ww
pmichaud tadzik: are you perchance going to make it to yapc::eu ? 20:43
pmichaud hopes "yes"
tadzik pmichaud: absolutely
pmichaud \o/
jnthn \o/
tadzik \o/
pmichaud would you be able to attend a workshop/summit a couple of days before yapc::eu ?
tadzik (and I hope I can make it out without giving a talk, so I would be able to slack off all the time :P)
pmichaud ...no talk?!? are you kidding me?! :-) 20:44
tadzik pmichaud: possibly. Anything in particular that you have in mind?
tadzik checks the dates
well, maybe something :)
talks still make me a lot nervous
pmichaud tadzik: that's why you should do more, then. :) 20:45
tadzik and to get rid of... exactly :)
sorear it is at this point looking highly likely I will be at yapc::eu
i'd love to talk about the kinds of talks people want to hear from me 20:46
i think there is overlap here with the concept of 'what would I use a blog for'
20:46 bluescreen10 joined
PerlJam sorear: you need to figure out how to recruit more devs :) 20:48
pmichaud sorear: tbh, any talk you give will likely be well received. (more)
PerlJam sorear: so ... whatever talk makes people go "I want to do that!" :)
moritz sorear: make a talk "how to make Perl 6 fast"
pmichaud Lots of people know about Rakudo, I suspect many are unaware of Niecza or what it is/can do
PerlJam moritz++
pmichaud so if nothing else, a talk about niecza, how it works, how it compares to rakudo, etc would likely be very well received 20:49
moritz sorear: and talk about your experiences about what worked well for niecza (speed-wise), and what not
pmichaud if you don't want to talk about niecza itself (or solely about that) -- well... any topic dealing with Perl 6 is likely of interest to all o fus :)
*of us
for that matter -- anyone have a specific talk they'd like to see me give? 20:50
I can certainly do part 2 of my lists/iterators talk (that I didn't give last year at yapc::eu, and that I gave at yapc::na this year).
PerlJam pmichaud: how to make Perl 6 fast :-)
pmichaud PerlJam: that's jnthn++'s talk :)
sorear mmh. Later when I have more time I'd like to elaborate on this. maybe start coming up with notes that we can refine 20:51
i have never given a tech talk.
PerlJam Will YAPC::EU be videoed as well as YAPC::NA?
(or at all?)
pmichaud sorear: I'll be happy to help any way that I can. I might also be able to forward you a copy of Damian Conway's "Presentation Aikido" notes that describe ways to give good talks :) 20:52
brrt1 hi all, whom can i pester about nqp & load_language in relation to winxed
and a segfault to boot :-)
pmichaud PerlJam: although, I suppose I could give a lightning talk on "how to make Perl 6 fast", which basically consists of a slide saying "recruit jnthn++ and stay out of his way" :-)
jnthn ;) 20:53
pmichaud oh, and "make sure he's supplied with beer"
jnthn YAPC is in Germany this year.
tadzik :P
jnthn Beer.
:)
Not that the place in Riga that served 30 different types of beer on tap will be easy to beat :P
PerlJam "How to turn beer into Perl 6"
pmichaud PerlJam: lol 20:54
brrt1 would like to go to YAPC::EU
jnthn I suspect my main problem in the coming months will actually be to make sure I do enough $dayjob than it will be to make sure I do enough Perl 6. :)
sorear jnthn: I will need mentoring on that too, I am also a complete beer novice :p
jnthn Oh! 20:55
Well, Germany is a good first place :)
brrt1 land of the reinheitsgebot ;-)
20:57 adam7504 left
pmichaud sorear: if you apply for a talk at yapc::eu (+1) you could also request that it be on the 2nd or 3rd day. That would give you time to see what other presenters do and what it looks like. 20:57
jnthn needs to get around to submitting.
I'm planning a talk on roles. 20:58
pmichaud although you've likely already seen some of the videos from yapc::na. Generally, talks are low-key, not terribly formal (but you still want to have good preparation), and if you're interested in your topic then the audience almost certainly will be as well.
roles +1. Very timely for yapc::eu, for a number of reasons.
jnthn And maybe a Rakudo "Achievements and Plans" talk
tadzik Plans Unlocked!
jnthn :D
20:59 Je joined
jnthn I may start out by going back to my "plans" slides at the end of my Rakudo talk last YAPC::EU and pointing out how we did All The Things, pretty much. :) 20:59
sorear watching the videos from yapc::na is on my agenda :D
pmichaud jnthn: you could do that as a lightning talk, perhaps. Although "All the Things" is hard to fit in five minutes. 21:00
sorear whatever I decide to do, I'm probably going to use sandiego.pm as a guinea pig
21:00 brrt1 left
jnthn pmichaud: Well, one doesn't preclude the other. :) 21:00
pmichaud Although I did cover "Most of the Things" in my lightning talk at yapc::na. Only took me about 3 seconds. :-)
PerlJam Someone could do a "state of IO" talk. 21:01
pmichaud PerlJam: only if we get a "state of Europa" to go with it :-P
PerlJam :-P
jnthn lol
tadzik I don't get it
PerlJam tadzik: moons of jupiter
pmichaud tadzik: moons of Jupiter
jnthn PerlJam: That's just trolling somebody into doing conference driven development on IO :P
tadzik oh :)
pmichaud plus, the state of Europ(e|a) is a little unsettled at the moment. 21:02
s/unsettled/interesting/
jnthn Last I noticed, so was the state of IO :P
pmichaud _exactly_ 21:03
PerlJam jnthn: Do the fish bite even if you tell them it's bait?
jnthn PerlJam: Only if it's really tasty bait.
The state of &dir made me grrrr at $dayjob today.
Curiously, that's the first thing I ran into as problematic while actually using Rakudo to Get Stuff Done. 21:04
tadzik heh, a lightning talk: "&dir makes me $grrr"
PerlJam sounds Dr. Seussian 21:05
21:05 kaare_ left
pmichaud "I do not like Perl and CPAN, I do not like them, Tim I am." 21:06
(Mark Keating)++
PerlJam mdk++ indeed. Of all the YAPC talks I've seen so far, that one is my favorite. 21:07
pmichaud (www.youtube.com/watch?v=NUAOns_WAXA in case you haven't seen it yet)
pmichaud sadfaces at the pause in video uploads to youtube.com/yapcna
sirrobert removed everything and re-installed rakudo, panda, etc. from scratch. Now Panda is working again =) 21:08
gotta go to a meeting. thanks all =)
21:08 sirrobert left
jnthn pmichaud: A PAUSE is pretty common in uploading related things in the Perl world... 21:09
pmichaud *ouch*
nterpolation of Parcel differently from Iterable. I think I would be happy to s 21:11
erggg
ee those unify. We still need to flesh out .list versus .flat a bit (working on that)
phenny tell TimToady I can't recall or find the use cases that led me to treat interpolation of Parcel differently from Iterable. I think I would be happy to see those unify. We still need to flesh out .list versus .flat a bit (working on that) 21:12
phenny: tell TimToady I can't recall or find the use cases that led me to treat interpolation of Parcel differently from Iterable. I think I would be happy to see those unify. We still need to flesh out .list versus .flat a bit (working on that)
phenny pmichaud: I'll pass that on when TimToady is around.
pmichaud obviously I need help with the irc interface today :-/
"How does one spell 'IRC', anyway?"
pmichaud goes back to scribbling ideas for List and Parcel 21:13
atrodo pmichaud: I was curious, at YAPC::NA, I heard you talk about getting rakudo to run on other vm's and not just parrot. I wondered, if m0/lorito was implemented and ran on jvm/clr/etc, would that be a desirable? Would you still want rakudo to be able to target them directly?
21:14 birdwindupbird left
pmichaud I think being able to run in more environments is always a plus (more) 21:14
21:14 brrt joined
sorear pmichaud: why did uploads pause? 21:15
pmichaud Whether Rakudo would still target other machines directly would depend almost entirely on the cost/benefit of doing so. If there's little additional benefit to be gained, we probably wouldn't do it.
sorear: I don't know. I just know that it's been a day or two since any new yapc::na videos have been uploaded.
sorear have you asked anyone? 21:16
pmichaud sorear: no; I know that it can take some time. The blog post on the yapcna blog says they expect to have them all "by the end of the week", though.
sorear pmichaud: ah, ok. 21:17
masak good night, brave and glorious #perl6 21:18
pmichaud masak: what about the rest of us? ;-)
PerlJam Why can't they dub mithaldu's videos official?
pmichaud PerlJam: well, I'm kind of hoping they don't do that, if only because the video of my talk has blackouts in it and I'm hoping the copy they have doesn't have the blackouts. 21:19
masak pmichaud: there's no escaping bravery and glory :)
tadzik they have blanks :(
masak 'n
tadzik 'n
PerlJam pmichaud: make his official but only as a first approximation then?
jnthn atrodo: Given project velocities at the moment, my guess would be that Rakudo will get to targeting other things before m0/lorito is ready to support that use case. I could be wrong.
21:19 Je left
atrodo pmichaud: So, in theory, using m0 as a compatibility shim on top of other environments would be a good thing? 21:20
jnthn: I'm talking in much more curiosity terms than reality. I would tend to agree.
pmichaud atrodo: s/would/could/. There's sometimes a long distance between theory and reality. :)
tadzik hmm, szabgab's mail seem to indicate that HTML::Template is alive and well
PerlJam jnthn: I don't know ... whenever I do pay attention to #parrot, there seems to be lots happening on the m0 and m1 fronts
sorear pmichaud: Why are there blackouts?
pmichaud sorear: I don't know. Something happened either in the recording or transmission of the video 21:21
tadzik they seem to be deliberate
atrodo pmichaud: Gotcha. That makes sense. Thatnks
pmichaud I'm hoping it was in transmission (in which case there's a chance they're not blacked out in the archival copies)
it's also possible there was a signal mismatch between my system and whatever was handling the video feed 21:22
it might not have liked the smaller black text on all-white screen
tadzik that rings a bell
pmichaud but... I didn't find out about the blackouts until I saw the mithaldu videos myself. 21:23
(i.e., well after the conference)
sorear pmichaud: ah. in San Diego English blackout usually refers to a deliberate contract thing "we aren't authorized to share the video from 4:30 to 5:10, so we didn't"
pmichaud oh, I mean blackout in the sense of "power dropped for some technical reason beyond our control" :-) 21:24
what would be the San Diego English for the type of blackout where signal is lost unexpectedly? ;-)
ggoebel sorear: good advice on presentations, even when it isn't a keynote: thenextweb.com/lifehacks/2012/05/13...t-keynote/
tadzik did we get to know what happened to mpo? 21:25
pmichaud oops, have to run home quickly -- bbiaw 21:26
tadzik (modules.perl6.org)
PerlJam doesn't remember any blackouts while watching pmichaud's lists talk on the live streaming.
(I don't know how they captured the video, but if it was the same source as the streaming, this might bode well for complete videos) 21:27
tadzik: isn't it still broken? 21:28
tadzik PerlJam: it is
PerlJam I would think that if somoene figured out what happened, they'd fix it :) 21:29
tadzik luck and tuits given, we may be able to build something awesome with sergot which might be able to steal some of its glory :)
sergot \o/ 21:30
21:35 stephenlb joined
pmichaud PerlJam: yeah, that's why I'm hoping it was some sort of transmission issue and not a recording one. 21:45
I know that my wife said she kept getting blackouts during the live stream of my talk 21:46
jnthn Time for some rest...train leaves kinda early tomorrow 21:52
(Heading to FPW)
tadzik heh, I still have an exam before the flight tomorrow 21:53
PerlJam pmichaud: I was doing work while watching, so I may have just missed the blackout times.
jnthn tadzik: Good luck!
tadzik thanks :)
jnthn tadzik: What time do you arrive?
tadzik jnthn: 23-ish 21:54
jnthn tadzik: Ah, ok
tadzik: I guess I see you on the Friday then
tadzik when do you arrive?
jnthn (There's a pre-workshop beer)
20:15
tadzik hmm, huh
workshop is saturday?
jnthn Friday and Saturday
tadzik and beer both friday and pre-workshop? 21:55
jnthn I guess there's beer every night
I was asking when you arrived in case you would be at the pre-workshop beer meetup
tadzik that doesn't surprise me
oh
I see, ok 21:56
jnthn But given your arrival time I guess, "no"
tadzik well, if you overstay at the bar... :)
jnthn masak and I should make it...by some stroke of luck our hotel is exactly half way between the train station and the meet.
Me? Overstay at a bar? ;-)
tadzik and I'll manage to semi-prepare my talk by then... ;)
jnthn ;-)
I'm cheating and giving the same one I did at GPW :) 21:57
Apart from thanks to recent improvements I could make the code cuter. :)
tadzik okay, guess it's time for me to get some sleeps
jnthn Yup, me too 21:59
'night all o/
tadzik \oo
21:59 PacoAir left 22:00 bluescreen10 left 22:03 adam7504 joined 22:07 moukeddar joined, moukeddar left
sergot good night ! 22:13
22:13 sergot left 22:14 pupoque left 22:15 brrt left 22:16 adu left 22:21 alester left 22:23 HarryS left 22:32 crazedpsyc is now known as DrQwest 22:34 dorlamm joined 22:37 DrQwest is now known as crazedpsyc 22:46 fridim_ left 22:48 adam7504 left 22:49 dorlamm left 22:54 cognominal left 22:55 cognominal joined 23:03 spider-mario left
TimToady pmichaud: I think I'm okay with .elems saying 6 there as long as we have something like .objects that says 2 23:07
phenny TimToady: 21:12Z <pmichaud> tell TimToady I can't recall or find the use cases that led me to treat interpolation of Parcel differently from Iterable. I think I would be happy to see those unify. We still need to flesh out .list versus .flat a bit (working on that)
23:08 orafu left
TimToady that is, .elems migh thave an implicit .flat like .map and friends 23:08
flussence r: say [+] 1..500000 23:10
p6eval rakudo 88a9d6: OUTPUT«(timeout)» 23:11
flussence r: say [+] 500000..1
p6eval rakudo 88a9d6: OUTPUT«0␤»
flussence r: gist.github.com/3007551 23:22
p6eval rakudo 88a9d6: OUTPUT«125000250000␤»
23:40 thou left

Powered by Cro and the Raku® Programming Language.

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

Thank you!