Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2014-10-05

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:16 Hor|zon joined #perl6
01:23 Mso150 joined #perl6
01:35 gtodd left #perl6
01:49 FROGGS_ joined #perl6
02:02 noganex_ joined #perl6
02:03 vendethiel- joined #perl6
02:11 vendethiel joined #perl6
02:17 Hor|zon joined #perl6
02:27 _slade_ joined #perl6
02:51 Exodist joined #perl6
03:00 noganex joined #perl6
03:19 pat_js joined #perl6
03:25 gtodd joined #perl6
04:18 Hor|zon joined #perl6
04:39 rindolf joined #perl6
04:44 ggoebel1111110 joined #perl6
04:59 petercommand joined #perl6
05:22 fhelmberger joined #perl6
05:47 rurban joined #perl6
05:58 bjz joined #perl6
06:03 darutoko joined #perl6
06:19 Hor|zon joined #perl6
06:27 yeahnoob joined #perl6
06:59 SamuraiJack_ joined #perl6
07:02 masak alohamorning, #perl6
07:12 dalek book: 0f87d07 | masak++ | src/regexes.pod:
07:12 dalek book: fix quantifier modifier thinko
07:12 dalek book:
07:12 dalek book: Discovered by 宋文泰 in https://rt.perl.org/Ticket/Display.html?id=122904
07:12 dalek book: review: https://github.com/perl6/book/commit/0f87d07103
07:13 ribasushi joined #perl6
07:17 rurban joined #perl6
07:34 masak a quick skim of http://www.cs.cmu.edu/~aldri​ch/papers/objects-essay.pdf indicates that it'd be of interest to people here.
07:35 masak HN discussion: https://news.ycombinator.com/item?id=8411225
07:38 vendethiel o/.
07:47 * masak .oO( is that a ping-pong ball? don't crush it! )
07:49 rurban left a comment at HN
07:50 rurban defending functional over objects of course
07:51 masak uhm.
07:52 masak I really really like functional, but I'm not sure I believe it subsumes OO to the point of making it irrelevant.
07:52 masak in the end it comes down to different focus.
07:53 rurban sure it does. it's not irrelevant because its standard industry practice. worse is better. but is not inevitable. see e.g. CLOS
07:54 rurban langs without even a lambda or method dispatch on arg types are doomed to have only objects
07:55 rurban in potion objects are implemented as functions e.g. even variables are functions
07:58 spider-mario joined #perl6
08:02 kurahaupo joined #perl6
08:04 spider-mario joined #perl6
08:20 Hor|zon joined #perl6
08:34 psch joined #perl6
08:35 psch hi #perl6 o/
08:38 jack_rabbit joined #perl6
08:39 lizmat psch \o
08:40 jnthn o/ psch, lizmat
08:40 lizmat jnthn \o
08:41 psch jnthn, lizmat \o
08:41 psch i found the tests for -e :)
08:41 psch they still fail with my patch :(
08:45 gfldex joined #perl6
08:45 rurban got strange results for t/spec/S03-junctions/autothreading.t Failed 76/89 subtests but manually all tests pass with parrot and moar
08:45 lizmat rurban: on various builds I've seen various stable tests flap when doing spectest
08:46 lizmat I assume we still have a few heisengremlins in moar / libuv
08:46 jnthn lizmat: afaik, the only ones we have left in moar since the inlining/gc/continuations bug I fixed the other day are S17 ones...lemme know if you're seeing anything other than that.
08:47 lizmat jnthn: will do
08:47 jnthn (Because all I have here are S17 flappers, and a small number of repeatable Win32 issues.)
08:48 jnthn lizmat: Did you ever get to the bottom of the io-spec-cygwin.t issue, ooc?
08:49 lizmat eh, I think I did...it went away afaict
08:49 * jnthn pulls stuff and gets a new spectest run
08:50 lizmat fwiw, I'm about to unleash new chdir and indir :-)
08:53 jnthn Cool
08:54 vendethiel rurban: I strongly believe the only reason java has all these factories is because they didn't have lambdas :-)
08:54 vendethiel (also, when people say "dependency injection", I tend to correct by saying "argument passing")
08:55 jnthn It's partly cultural too.
08:55 jnthn Before C# got lambdas, I don't remember it having a huge culture of factory writing.
08:56 vendethiel jnthn: I think c# guys go with the "provider" stuff instead of factories?
08:56 jnthn There's something about the Java way of API design that seems to embrace/encourage putting pattern names in class names, and using plenty of them...
08:57 jnthn vendethiel: Go with all sorts; convention-based DI containers amongst them...
08:57 vendethiel heehhe https://github.com/EnterpriseQualit​yCoding/FizzBuzzEnterpriseEdition/ :-)
08:57 rurban well, "dependency injection" is manipulating @ISA, the method dispatch
08:58 rurban which forbids compile-time optimization of method dispatch
08:58 jnthn rurban: That isn't what it's meant on any codebase I've worked on.
08:58 rurban i.e. use oo :final forbidden in libs
08:58 vendethiel me neither
08:59 jnthn Typically I see it being used to refer to techniques like constructor injection
08:59 rurban interesting. how do you inject deps via argument passing?
09:00 jnthn And then arranging for said DI framework to be in charge of creating the "entry point" (for example, controllers in an MVC app).
09:00 rurban Ah, I see. java style
09:00 jnthn Much wider than that.
09:00 vendethiel rurban: you just pass you "db connection" as an argument instead of storing it on the object
09:00 rurban yes, I see. you need that for static languages.
09:00 jnthn vendethiel: Well, more important is that something outside of the object creates it
09:01 vendethiel jnthn: I just mean - DI is argument passing glorified.
09:01 nine Dependency injection is just "programming to an interface, not to an implementation" followed through.
09:02 jnthn vendethiel: Operationally, yes. The reasons I tend to spend time talking about it in classes is because, done right, you get looser coupling and much better testability.
09:03 jnthn nine: Yes, it's a concrete way of achieving that.
09:03 vendethiel jnthn: Compared to singleton/globals, sure. Compared to argument passing, no.
09:03 vendethiel If I give you your db connection through an argument or through some instance variable, it's the *same*
09:03 vendethiel actually, the argument one is less coupled, even :-)
09:05 jnthn vendethiel: Again, technically yes, but from an engineering point of view, making a decision once at object construction time when it will likely need to hold over the object's entire lifetime is likely a better design than relying on code interacting with the object to come up with an appropriate argument to pass each time.
09:05 FROGGS[mobile] joined #perl6
09:05 vendethiel jnthn: in an OO world, certainly :-). In a more functional world, you don't even have a choice
09:05 vendethiel jnthn: lots of language provide some kind of option to even ease that further, from monadic stuff to implicit arguments (scala and idris/agda-like)
09:06 kurahaupo joined #perl6
09:07 lizmat jnthn: non S17 flapper:
09:07 lizmat t/spec/S04-phasers/first.t                                  (Wstat: 65280 Tests: 1 Failed: 0)
09:07 lizmat Non-zero exit status: 255
09:07 lizmat Parse errors: Bad plan.  You planned 4 tests but ran 1.
09:07 lizmat Files=922, Tests=33883, 176 wallclock secs ( 9.15 usr  3.47 sys + 1068.72 cusr 158.26 csys = 1239.60 CPU)
09:07 lizmat runs fine by itself
09:08 lizmat but fails under make
09:08 jnthn Grr, I hadn't seen that issue for a long time...
09:08 lizmat reliably, so I guess technically, it's not a flapper
09:09 jnthn vendethiel: Agree, there are non-object ways to solve the same issue, but if you have objects to hand, might as well use them to their strengths. :)
09:10 dalek rakudo/nom: a1da1f5 | (Elizabeth Mattijsen)++ | src/core/IO/Path.pm:
09:10 dalek rakudo/nom: Add  IO::Path.chdir
09:10 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a1da1f5f1d
09:10 dalek rakudo/nom: de6cb3e | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
09:10 dalek rakudo/nom: Update chdir(), add indir()
09:10 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/de6cb3e38b
09:14 * jnthn needs to get himself back into MOP stuff again in prep for working on native arrays, typed array improvements, etc, and figures he'll take on http://perlcabal.org/syn/S1​2.html#Custom_Meta-objects :)
09:15 FROGGS eww, the following borks my terminal on moar: perl6-m -e 'my $fh = open(~$*EXECUTABLE, :w, :p); $fh.print: "say 42"; $fh.close'
09:16 jnthn Due to handles being inherited, mebbe?
09:17 FROGGS jnthn: the custom metaop sounds highly "interesting"
09:17 FROGGS jnthn: I guess so
09:17 FROGGS we probably inherit and therefore close our own stdout
09:18 jnthn FROGGS: I don't think it'll be so hard to do, which is why I'm taking it as a warm-up task ahead of diving into some deeper 6model extensions :)
09:19 FROGGS jnthn: if you say so... I'm just scared about all these things :o)
09:19 FROGGS seems like not to study CS does not pay off :o)
09:22 bartolin joined #perl6
09:27 dalek rakudo/nom: 7968c69 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
09:27 dalek rakudo/nom: Bring chdir up to spec
09:27 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7968c694ee
09:27 dalek rakudo/nom: 7dec9a9 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
09:27 dalek rakudo/nom: Make sure we don't lose failures
09:27 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7dec9a9540
09:33 rurban darwin had the same utf8 issues as the other bsd's. e.g. t/spec/S05-mass/charsets.rakudo.parrot. fixed right now
09:45 dalek roast: ada98bc | usev6++ | S14-roles/parameterized-type.t:
09:45 dalek roast: Add test for RT #112970
09:45 dalek roast: review: https://github.com/perl6/roast/commit/ada98bc5f1
09:45 synopsebot Link: https://rt.perl.org/rt3//Publ​ic/Bug/Display.html?id=112970
09:45 tinyblak joined #perl6
09:47 timotimo o/
09:48 FROGGS o/
09:49 lizmat $ perl6 -e '$CWD = 42; say $*CWD'
09:49 lizmat 42
09:49 lizmat this seems an unwanted side effect of no strict
09:49 lizmat FROGGS timotimo o\
09:50 jnthn lizmat: Programs start out in GLOBAL, and $*FOO lookup falls back to $GLOBAL::FOO
09:50 FROGGS lizmat: at least surprising, aye
09:51 lizmat aha... hmmm....
09:51 FROGGS I tripped over that in v5 too
09:51 lizmat jnthn: even if the var in GLOBAL is not marked as dynamic ?
09:51 lizmat $ perl6 -e '$CWD = 42; say $CWD.VAR.dynamic'
09:51 lizmat False
09:52 jnthn lizmat: Yes, dynamic is all about visibility through the lexpad
09:52 jnthn lizmat: But this is about package lookup, and packages are just glorified hashes...
09:52 lizmat hmmm..... yeah, gotcha
09:52 kaare_ joined #perl6
09:52 jnthn But yeah, it's a curious interaction of a few things that each make sense :)
09:53 lizmat maybe auto-defined vars should be in their own package, not visible to dynamic lookup ?
09:53 jnthn We could consider re-visiting if "our" scope is the right thing for undeclared variables.
09:53 lizmat :-)
09:53 virtualsue joined #perl6
09:53 jnthn Or if maybe they should be lexicals in UNIT or so
09:54 lizmat wouldn't that break a lot of opts ?
09:54 jnthn Not that I can see...are you thinking of anything in particular?
09:54 jnthn Note that it's the compiler that spots the variable is undeclared and goes off to declare it for us
09:54 lizmat true
09:55 lizmat or is it?
09:55 jnthn So it doesn't violate the "lexpad keys immutable at runtime" thing.
09:55 lizmat even if we're doing an EVAL inside a no strict block ?
09:55 jnthn Hmm
09:56 jnthn An EVAL has its own UNIT
09:56 lizmat ah, yup
09:56 psch re lax mode: S19:430 says -e doesn't default to lax mode
09:56 synopsebot Link: http://perlcabal.org/syn/S19.html#line_430
09:56 jnthn So it's true, but probably identifies the downfall of the approahc I suggested
09:56 FROGGS I really like to get a backtrace here:
09:56 FROGGS $ perl6-p
09:56 FROGGS > 1
09:56 FROGGS Null PMC access in get_string()
09:56 FROGGS :/
09:56 psch FROGGS, timotimo o/
09:56 FROGGS hi psch
09:56 jnthn FROGGS: Maybe comment out a CATCH somewhere in HLL::Compiler?
09:57 * FROGGS tries
09:57 jnthn lizmat: (To be clear, if we declare them in UNIT, then an undeclared variable inside of an EVAL evaluated in lax mode would not be visible outside of the EVAL)
09:58 jnthn (I'm also not sure if that's a bad thing... :)
09:58 lizmat that just sounds about right to me  :-)
09:58 lizmat otherwise we *would* be violating the immutability rule, right ?
09:59 jnthn lizmat: Yes. Well, the alternative is to leave things as they are today and declare them our-scoped, and get the surprise you just ran into. :)
09:59 dalek rakudo/nom: 3962ee5 | (Elizabeth Mattijsen)++ | src/core/IO/Path.pm:
09:59 dalek rakudo/nom: Make IO::Path.chdir also a class method
09:59 dalek rakudo/nom:
09:59 dalek rakudo/nom: For the so inclined, it's basically $*CWD.chdir
09:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3962ee51d9
10:00 jnthn Auto-declaration is generally surprising in some way, as I tend to find out each time I end up writing Python. :)
10:00 timotimo hm, do we currently do mixtures of multiple -e correctly in rakudo?
10:00 FROGGS timotimo: there is a PR about that
10:00 jnthn timotimo: No, which I think is why there's a bunch of discussion around it ATM :)
10:01 FROGGS $ perl6-m -e 'say 1' -e 'say 2'
10:01 FROGGS 1
10:01 psch i'm still on that
10:01 psch HLL::CommandLine::Parser is getting in my way a bit
10:01 timotimo more importantly, i think -e 'say "init"' -pe 'say "do something"' -e 'say "done"' is nicer than using phasers ...
10:01 psch -e'say 1;' -e'say 2;' e.g. works
10:01 kjs_ joined #perl6
10:01 psch but -e 'say 1;' -e 'say 2;' doesn't
10:02 psch because the space after e does something about how the rest is parsed
10:02 timotimo ah, i remember looking at that!
10:04 * nine finds it odd that the implementation of a language that takes parsing to whole new levels fails at parsing its command line...
10:06 vendethiel hahaha
10:09 jnthn If only we implemented the langauge in itself, then we could use its parsing technology to parse its command line... :P
10:10 timotimo we should just concatenate the options with spaces again, then we would be able to parse it in one go! :P
10:13 xinming joined #perl6
10:15 jnthn With spaces is a bad idea, but with some other sentinel (NULL) might even be feasible
10:15 tinyblak_ joined #perl6
10:15 timotimo does our parsing work properly with that kind of thing?
10:18 jnthn m: say "ab\0c\0de" ~~ / [\w+]+ % \0 /
10:18 camelia rakudo-moar 9c807f: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/Ra70SZ9ICOâ�¤Unrecognized backslash sequence: '\0'â�¤at /tmp/Ra70SZ9ICO:1â�¤------> [32msay "ab\0c\0de" ~~ / [\w+]+ % \[33mâ��[31m0 /[0mâ�¤    expecting any of:â�¤        postfixâ�¤        infix…»
10:18 jnthn m: say "ab\0c\0de" ~~ / [\w+]+ % \x0 /
10:18 camelia rakudo-moar 9c807f: OUTPUT«「ab␀c␀de」␤␤»
10:18 jnthn m: say "ab\0c\0de" ~~ / [(\w+)]+ % \x0 /
10:18 camelia rakudo-moar 9c807f: OUTPUT«「ab␀c␀de」␤ 0 => 「ab」␤ 0 => 「c」␤ 0 => 「de」␤␤»
10:18 timotimo ah. this could actually work.
10:20 denis_boyun_ joined #perl6
10:21 Hor|zon joined #perl6
10:21 jnthn Well, maybe. Just saying that if we really are struggling with scaling the current parsing, we might want to look into using the parsing tech we have :)
10:22 lizmat fwiw, I would not be in favour of allowing \0 in *code*
10:22 lizmat rurban has some pretty string ideas about that, I think
10:22 lizmat problem is that code with \0 may hide code after it in some (dumb) situations
10:22 lizmat because \0 is C's end of string marker
10:23 lizmat and would thus be an attack vector
10:23 slavik joined #perl6
10:23 timotimo this may be a good opportunity to make sure we're null-byte-clean
10:26 jnthn It's mostly an issue if you're passing strings off to C-land
10:26 timotimo aye
10:27 jnthn And yeah, there are some interesting attacks around that. But if anything we'd be reducing risk here, not increasing it... :)
10:27 jnthn My favorite attack of that kind was in combination with a badly written UTF8 decoder
10:27 timotimo IIRC, the Nintendo Wii got "pwned" because they were comparing the code's signature with strcpm instead of memcmp and the signature was "the binary" instead of "hexadecimal as ASCII"
10:27 jnthn Such that you could smuggle null bytes into places by giving the thing some invalid UTF8, which caused it to emit nulls as it decoded :)
10:28 timotimo so all you had to do was find a legit game that has a null byte near the beginning of its signature and you could brute force a colliding payload for the same signature
10:29 jnthn cute :)
10:32 jnthn Yay. My bag is packed for today's trip up to Stockholm, I've had lunch, and now I've got some free time to code :)
10:33 FROGGS sounds good :o)
10:37 tinyblak joined #perl6
10:50 lizmat m: say $*CWD.WHAT; say nqp::iscont($*CWD)   # seems to work ok
10:50 camelia rakudo-moar 3962ee: OUTPUT«(timeout)»
10:50 lizmat :-(
10:50 lizmat $ 6 'say $*CWD.WHAT; say nqp::iscont($*CWD)'
10:50 lizmat (IO::Path)
10:50 lizmat 1
10:51 lizmat $ 6 'say $*TMPDIR.WHAT; say nqp::iscont($*TMPDIR)'
10:51 lizmat (IO::Path)
10:51 lizmat 0
10:51 lizmat so why is one IO::Path a cont, and the other isn't ?
10:51 lizmat jnthn:  ^^^ any ideas ?
10:51 timotimo "is a cont" vs "is in a cont"
10:52 timotimo as in: the IO::Path itself isn't "the cont"
10:52 lizmat $ 6 'temp $*TMPDIR'
10:52 lizmat Can only use 'temp' on a container
10:52 lizmat $ 6 'temp $*CWD'
10:52 lizmat (works fine)
10:53 FROGGS I guess TMPDIR gets bound to PROCESS and CWD not?
10:53 rurban lizmat: do you really say "string ideas", not strong ideas?
10:53 lizmat rurban: sorry, strong ideas  :-)
10:53 FROGGS hehe
10:53 rurban or stringent idea
10:54 lizmat strong has the shortest StrDistance
10:55 jnthn lizmat: Probably just because one ends up getting bound into PROCESS or so, while the other is assigned.
10:55 jnthn lizmat: I doubt there's a strong design reason for those two...
10:56 lizmat aha... I see what you mean...
10:57 timotimo fwiw
10:57 timotimo timo@schmand ~/p/rakudo (nom)> perl6-m -e 'say $*CWD.VAR.WHAT'
10:57 timotimo (Scalar)
10:57 timotimo timo@schmand ~/p/rakudo (nom)> perl6-m -e 'say $*TMPDIR.VAR.WHAT'
10:57 timotimo (IO::Path)
10:58 dalek specs: 37874f2 | (Elizabeth Mattijsen)++ | S16-io.pod:
10:58 dalek specs: Re-introduce tmpdir() and homedir(), update chdir()
10:58 dalek specs:
10:58 dalek specs: The original idea behind unifying them into a single chdir() usage, becomes
10:58 dalek specs: very cumbersome in the implementation.  Therefore we're going back to these
10:58 dalek specs: functions.
10:58 dalek specs: review: https://github.com/perl6/specs/commit/37874f2cb4
10:58 dalek rakudo/nom: d033738 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
10:58 dalek rakudo/nom: Reintroduce tmpdir(), homedir(), update chdir()
10:58 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d0337384ac
11:03 bartolin rurban: this morning I tested rakudo on latest parrot from master on freebsd. the two non-ASCII tests from S19-command-line/dash-e.t fail with non-utf8 locale, but run fine otherwise. (same result as on netbsd and linux)
11:04 dalek nqp: b79b20c | (Tobias Leich)++ | src/vm/parrot/QAST/Operations.nqp:
11:04 dalek nqp: remove eof on parrot that breaks on streams
11:04 dalek nqp:
11:04 dalek nqp: This was meant to unify the behaviour on all backends so that we detect EOF already
11:04 dalek nqp: before we attempted to read past it. Sadly this does not work out here and we have to
11:04 dalek nqp: take care about that in rakudo again.
11:04 dalek nqp: review: https://github.com/perl6/nqp/commit/b79b20c4d6
11:07 dalek rakudo/nom: 58cd6f1 | (Tobias Leich)++ | / (2 files):
11:07 dalek rakudo/nom: nqp::eoffh on parrot does not detect EOF early
11:07 dalek rakudo/nom:
11:07 dalek rakudo/nom: This re-introduced code that checks if we are at EOF. Also bump nqp for nqp::eoffh fix.
11:07 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/58cd6f1a77
11:11 dalek rakudo/nom: 71b9822 | (Tobias Leich)++ | src/Perl6/Grammar.nqp:
11:11 dalek rakudo/nom: fix repl on parrot and jvm
11:11 dalek rakudo/nom:
11:11 dalek rakudo/nom: $?FILES can be NULLish and we have to check instead of strcmping it against something.
11:11 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/71b98224ad
11:11 lizmat FROGGS++
11:11 FROGGS pmichaud: that should fix the REPL
11:13 FROGGS will now run tests (again) and then respond to tickets/mails
11:15 FROGGS lizmat: about eof on parrot... please keep in mind that you have to check .DEFINITE like I did in order not not have additional empty strings at the end of slurped files...
11:16 lizmat is that a parrot only thing ?
11:16 FROGGS lizmat: it was a parrot and jvm thing before I hacked nqp::eoffh three weeks ago, then it was not a thing and now it is a parrot thing again :o)
11:17 lizmat so, .lines and .slurp would be affected, right ?
11:17 lizmat (and .words in the future)
11:17 tinyblak_ joined #perl6
11:17 lizmat or until we have nqp::readcharsfh on parrot, right ?
11:23 fhelmberger joined #perl6
11:25 bjz joined #perl6
11:26 FROGGS[mobile] joined #perl6
11:26 psch it occurs to me that we have two kinds of stoppers for cli args.  with -e we want to stop at the last one, with -- we want to stop at the first
11:26 lizmat m: 42
11:26 camelia rakudo-moar 3962ee: ( no output )
11:27 lizmat m: $*TMPDIR = 42
11:27 camelia rakudo-moar 3962ee: OUTPUT«Cannot modify an immutable IO::Path␤  in block <unit> at /tmp/etMnEm2meM:1␤␤»
11:27 lizmat m: $*TMPDIR; $*TMPDIR = 42'
11:27 camelia rakudo-moar 3962ee: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/zsJPeybRvAâ�¤Two terms in a rowâ�¤at /tmp/zsJPeybRvA:1â�¤------> [32m$*TMPDIR; $*TMPDIR = 42[33mâ��[31m'[0mâ�¤    expecting any of:â�¤        postfixâ�¤        infix stopperâ�¤        infix or m…»
11:27 lizmat m: $*TMPDIR; $*TMPDIR = 42
11:27 camelia rakudo-moar 3962ee: OUTPUT«Cannot modify an immutable IO::Path␤  in block <unit> at /tmp/P00tgYuGLq:1␤␤»
11:28 psch although i think that solves itself with the grammar i'm putting together anyway, as any stopper option is an option first...
11:28 dalek rakudo/nom: e5884bf | (Elizabeth Mattijsen)++ | src/core/Process.pm:
11:28 dalek rakudo/nom: First step in fixing $*TMPDIR
11:29 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e5884bf46e
11:30 lizmat jnthn timotimo: the $*TMPDIR problem is the lazy evaluation of it
11:31 lizmat once it is initialized, a subsequent attempt to change works
11:31 lizmat $ 6 '$*TMPDIR; $*TMPDIR=42; say $*TMPDIR.VAR.WHAT'
11:31 lizmat (Scalar)
11:36 lizmat jnthn timotimo: maybe some magic in Actions/make_variable_from_parts is needed ?
11:38 jnthn lizmat: I don't think so...that sounds like a level mixup..
11:38 jnthn Feels more like we're losing the container somewhere we ought not be
11:39 lizmat testing a fix
11:44 FROGGS lizmat: I only saw fallout in .lines... but I'm not sure if that means much
11:47 dalek roast: de7eb2d | jnthn++ | S12-meta/ (5 files):
11:47 dalek roast: Some initial tests for EXPORTHOW.
11:47 dalek roast: review: https://github.com/perl6/roast/commit/de7eb2d8ca
11:57 dalek rakudo/nom: 2b2bafa | jnthn++ | src/ (2 files):
11:57 dalek rakudo/nom: Implement EXPORTHOW::SUPERSEDE.
11:57 dalek rakudo/nom:
11:57 dalek rakudo/nom: Also detect and complain about invalid directives, and report DECLARE
11:57 dalek rakudo/nom: and COMPOSE as NYI so far. Keep existing things working for now; we
11:57 dalek rakudo/nom: can start issuing a deprecation warning in the future for a plain use
11:57 dalek rakudo/nom: of EXPORTHOW::class, and remove it a bit further into the future.
11:57 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2b2bafa4a1
11:57 dalek rakudo/nom: 2662e01 | jnthn++ | t/spectest.data:
11:57 dalek rakudo/nom: Run S12-meta/exporthow.t.
11:57 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2662e01c1c
11:59 jnthn train &
12:00 robinsmidsrod joined #perl6
12:00 FROGGS[mobile] jnthn: have a nice time in Stockholm :o)
12:03 Alula joined #perl6
12:18 cognome joined #perl6
12:22 Hor|zon joined #perl6
12:22 BenGoldberg joined #perl6
12:23 cognome joined #perl6
12:33 yeahnoob joined #perl6
12:38 * lizmat gives up on the $*TMPDIR container issue
12:38 lizmat the workaround is simple: just mention it before in your code, and it works
12:39 lizmat because that *will* initialize and make subsequent accesses work
12:39 FROGGS huh
12:39 FROGGS t/spec/S17-procasync/print.rakudo.moar                      (Wstat: 256 Tests: 16 Failed: 1)
12:39 FROGGS Failed test:  15
12:39 FROGGS TODO passed:   12-14
12:39 FROGGS was that me?
12:39 lizmat (the reason $*CWD already works, is that it currently is non lazily initialized)
12:39 lizmat no
12:39 lizmat that's a flap
12:39 bjz joined #perl6
12:39 FROGGS ohh, that one too
12:40 lizmat yup: all async features have flapping tests  :-(
12:40 FROGGS :/
12:40 lizmat meanwhile I'm bisecting parrot breakage
12:40 FROGGS t/spec/S16-filehandles/io.t                                   (Wstat: 139 Tests: 23 Failed: 0)
12:40 FROGGS Non-zero wait status: 139
12:40 FROGGS Parse errors: Bad plan.  You planned 87 tests but ran 23.
12:40 FROGGS this?
12:41 lizmat yup, a segfault
12:41 FROGGS lizmat: so you had that already before my commits today?
12:42 lizmat looks like it
12:42 FROGGS okay...
12:43 FROGGS I've got not much time right now, but I can help this evening, in case there is still some breakage that needs fixing
12:43 no-doz left #perl6
12:53 araujo joined #perl6
12:53 araujo joined #perl6
12:55 guru joined #perl6
12:57 [TuxCM] where can I find the docs about the difference between Str and Str:D and what it is used for
12:58 BenGoldberg Str is a type.
12:58 BenGoldberg :D is a modifier used in a parameter list.
12:58 lizmat it means a defined Str
12:58 lizmat m: say Str.DEFINITE
12:58 camelia rakudo-moar 2662e0: OUTPUT«False␤»
12:58 BenGoldberg The :D can be applied to any parameter type, not just Str.
12:58 lizmat m: say "foo".DEFINITE
12:58 camelia rakudo-moar 2662e0: OUTPUT«True␤»
12:59 BenGoldberg It's a way for the person writing the code to not have to worry about what to do if passed an undefined value.
13:00 laben joined #perl6
13:00 laben hai #perl6 o/
13:00 lizmat laben BenGoldberg [TuxCM]  \o
13:01 BenGoldberg o/
13:01 BenGoldberg :)
13:02 laben today is a good day for updates, just a little thing needs to be fixed
13:02 laben lizmat: could you give a look at the build of Inline::Perl5?
13:03 vendethiel joined #perl6
13:03 laben lizmat: seems like LibraryMake broke, i get this when building I::P5 => No such method 'chdir' for invocant of type Str
13:04 laben panda seems to work fine
13:04 lizmat this is probably somewhere a .path that needs to be replaced by a .IO
13:05 laben need to dig a bit deeper in my logs to find the whole error trace
13:06 lizmat something is setting $*CWD to a Str, rather than an IO::Path
13:08 laben lizmat: can you reproduce it? or is it only me?
13:08 lizmat reproduced
13:14 dalek rakudo/nom: 205b588 | (Elizabeth Mattijsen)++ | src/core/IO/Spec/Unix.pm:
13:14 dalek rakudo/nom: Add naive $*SPEC.extension for basenames
13:14 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/205b588121
13:14 dalek rakudo/nom: 762d947 | (Elizabeth Mattijsen)++ | src/core/CompUnit.pm:
13:14 dalek rakudo/nom: Properly use basename/extension
13:14 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/762d94750c
13:17 PotatoGim^Home joined #perl6
13:24 kjs_ joined #perl6
13:29 dalek panda: e3fb048 | (Elizabeth Mattijsen)++ | lib/Panda/Common.pm:
13:29 dalek panda: Temporary solution to Inline::Perl5 breakage
13:29 dalek panda: review: https://github.com/tadzik/panda/commit/e3fb048df3
13:30 dalek rakudo/nom: 66cd275 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
13:30 dalek rakudo/nom: Put a canary in chdir()
13:30 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/66cd275b01
13:30 lizmat laben: could you verify that the above panda update fixes your problem ?
13:32 nine Why am I trying to figure out how to get syntax highlighted code into a presentation program instead of just using vim for the presentation?
13:32 laben lizmat: on it! i've started rebuilding rakudo etc right now
13:33 vendethiel- joined #perl6
13:35 dalek rakudo/nom: f7dc3c3 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
13:35 dalek rakudo/nom: Fix typo
13:35 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f7dc3c39d8
13:36 awwaiid nine: well, you could also use vim to output html
13:36 awwaiid (if your presenter takes html)
13:38 timotimo that's right, with the :hc tool, no?
13:39 laben lizmat: everything fine now :-)
13:40 lizmat cool!
13:40 nine awwaiid: you'd think that after 15 years of using vim one would just assume that there's some feature like that available :)
13:41 lue joined #perl6
13:42 FROGGS[mobile] joined #perl6
13:43 timotimo nine: if you output to html, you can usually copy&paste the formatted stuff into many gui presentation tools, like google docs presentations for example
13:45 nine timotimo: but obviously not into LibreOffice Impress. At least not without a detour through Writer.
13:49 timotimo oh :\
13:53 dalek specs: d8869f8 | (Elizabeth Mattijsen)++ | S02-bits.pod:
13:53 dalek specs: Pod would be an IO::Handle, not an IO::Path
13:53 dalek specs: review: https://github.com/perl6/specs/commit/d8869f83c6
13:53 dalek specs: a5df488 | (Elizabeth Mattijsen)++ | S16-io.pod:
13:53 dalek specs: Add indir/tmpdir/homedir to overview
13:53 dalek specs: review: https://github.com/perl6/specs/commit/a5df488e95
14:02 denis_boyun___ joined #perl6
14:04 dalek rakudo/nom: e418ce8 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
14:04 dalek rakudo/nom: Correct defaults for tmpdir() and homedir()
14:04 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e418ce8f52
14:04 dalek rakudo/nom: c503e01 | (Elizabeth Mattijsen)++ | src/core/Main.pm:
14:04 dalek rakudo/nom: Remove bare IO::Spec calls from Main
14:04 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c503e01fb5
14:10 dalek rakudo/nom: bb0f9a3 | (Elizabeth Mattijsen)++ | src/core/IO/Handle.pm:
14:10 dalek rakudo/nom: Remove bare IO::Spec calles from IO::Handle
14:10 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bb0f9a39c4
14:10 lizmat afk for a bit
14:13 Ven_ joined #perl6
14:15 jnthn FROGGS[mobile]: Thanks. Going for a few days concurrency teaching :)
14:15 * jnthn is enjoying the autumn colors out the train window
14:17 FROGGS[mobile] jnthn: that sounds like fun and fun at the same time :o)
14:17 zakharyas joined #perl6
14:18 timotimo %)
14:19 dalek rakudo/nom: a4d933a | jonathan++ | Configure.pl:
14:19 dalek rakudo/nom: Find .exe/.bat in prefix on Windows.
14:19 dalek rakudo/nom:
14:19 dalek rakudo/nom: Means a Configure with --prefix is sufficient; before an explicit
14:19 dalek rakudo/nom: --backend was always needed.
14:19 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/a4d933a3ee
14:21 jnthn oh, *that's* why things were darn slow...a hung iTunes swallwing a while core...
14:22 anaeem1_ joined #perl6
14:22 FROGGS[mobile] joined #perl6
14:23 Hor|zon joined #perl6
14:27 FROGGS[mobile]2 joined #perl6
14:28 Ven_ jnthn: why is itunes so godamn terrible, I wonder /
14:29 jnthn Ven_: tbh, I don't normally have a big problem with it
14:29 jnthn Ven_: Just now and then it does...stupid things.
14:30 Ven_ jnthn: Really? Everytime I use it, it's awkward and slow. I'm not even talking about the bad UI, but just in itself, it's a big blob of wtf
14:30 jnthn I'm often a bit surprised how much memory it wants
14:30 grondilu m: say .elems given my @ = <foo bar>
14:30 jnthn But at least on the hardware I have it's snappy enough not to bother me...
14:30 camelia rakudo-moar 2662e0: OUTPUT«(timeout)»
14:30 nine Ven_: why do people use itunes I wonder
14:31 Ven_ nine: 'cause we have apple phones
14:31 timotimo grondilu: the result is 2 on my local box
14:31 jnthn nine: It's the lazy/easy way to shove stuff onto most beloved iPhone :)
14:32 Ven_ eh. this phone is the only one I ever had, and it still works pretty well after 4 years :)
14:33 grondilu m: say .elems given constant @a = <foo bar>
14:33 camelia rakudo-moar 2662e0: OUTPUT«2␤»
14:35 kaare_ joined #perl6
14:36 grondilu m: my @a = "a", <b c>; say @a.elems;  # here I expect 3
14:36 camelia rakudo-moar 2662e0: OUTPUT«3␤»
14:37 grondilu m: constant @a = "a", <b c>; say @a.elems;  # here I would *also* expect 3
14:37 camelia rakudo-moar 2662e0: OUTPUT«2␤»
14:37 timotimo oh, i think our constants don't do "proper" list assignment yet
14:37 grondilu ok, good to know.
14:39 Ven_ erm.
14:39 Ven_ m: say ("a", <b c>).elems
14:39 camelia rakudo-moar c503e0: OUTPUT«2␤»
14:39 Ven_ I think TimToady++ changed it not long ago
14:40 timotimo m: constant %foo = "a", "b", "hi", 2, "foo", 3; say %foo.perl; say %foo.elems;
14:41 camelia rakudo-moar c503e0: OUTPUT«("a", "b", "hi", 2, "foo", 3)␤6␤»
14:41 timotimo that's what i mean.
14:41 timotimo .o( put that in your pipe and smoke it )
14:43 lizmat S02:4690 : "    integer     int     Integral int" regarding integer item context
14:43 synopsebot Link: http://perlcabal.org/syn/S02.html#line_4690
14:43 lizmat feels like a fossil
14:44 lizmat or is this just a matter of Integral and int() being NYI ?
14:46 [TuxCM] joined #perl6
14:50 jnthn lizmat: Well, int(...) would be a type coercion if it exists
14:50 lizmat but do you think it is a fossil or not?
14:53 jnthn lizmat: tbh, I don't know.
14:53 jnthn lizmat: Worth checking with TimToady
15:01 dalek rakudo/nom: f9c4d07 | jonathan++ | src/Perl6/Grammar.nqp:
15:01 dalek rakudo/nom: Implement EXPORTHOW::DECLARE.
15:01 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f9c4d079b8
15:02 lizmat_ joined #perl6
15:02 dalek roast: c335b12 | jonathan++ | S12-meta/ (3 files):
15:02 dalek roast: Tests for EXPORTHOW::DECLARE.
15:02 dalek roast: review: https://github.com/perl6/roast/commit/c335b125bd
15:03 timotimo does that mean we can now export a "foobar" in EXPORTHOW::DECLARE and in the code using it have a foobar { ... } that creates a type with our given metaobject?
15:04 timotimo $<sym>=[$declarator] <.end_keyword> <package_def>  ← that certainly suggests it
15:05 jnthn Yes
15:05 jnthn See the test
15:05 timotimo token ::($meth_name) {  -  this is pretty cool, i wasn't aware we can do it just like that
15:05 timotimo is that available in perl6 as well as nqp?
15:06 jnthn That's how all the grammar additions work... :)
15:06 jnthn Not sure if it works in Perl 6 yet...I probably meant to do it then didn't get around to it, 'cus unlike with NQP I didn't have a pressing need for it :)
15:07 jnthn Oh...also the NQP way is a bit cheaty :)
15:08 pmichaud good morning, #perl6
15:08 pmichaud constants don't understand list assignment
15:08 jnthn Anyway, that's SUPERSEDE and DECLARE done. COMPOSE is the trickier one... :)
15:08 jnthn o/ pmichaud
15:08 pmichaud at least they don't appear to
15:09 pmichaud well... hrm.
15:09 Ven_ well, they have a special list assignment that doesn't flatten :P
15:09 jnthn pmichaud: The spec hasn't been too concrete about what sigils on constants mean afaik
15:09 jnthn Ven_: It's called "binding" :P
15:09 jnthn Trouble is,
15:10 jnthn TimToady seems to also want it so we can mention a constant on its RHS and have stuff work out
15:10 Ven_ conjecture: .flat.elems might be a bit counterproductive
15:10 jnthn And I *really* dunno how to make that one work out
15:10 Ven_ I mean, for the compiler. it shouldn't need to flatten it
15:10 Ven_ not sure what you mean?
15:10 cognome joined #perl6
15:10 Ven_ a constant on its RHS?
15:11 jnthn constant @foo = 1, @foo; # it wasn't exactly this, but there was *some* sequence that could be defined elegantly that way...
15:11 jnthn I don't actually see how that can be made to work, alas...
15:12 jnthn We could say that constant @foo = blah; really does blah.list
15:12 pmichaud I'm thinking the problem may be in the parse
15:12 jnthn And constant %foo = blah; really does blah.hash
15:12 Ven_ jnthn: ah, the haskell-like one...
15:12 pmichaud i.e.,   I'm wondering if   constant @ = 1, <b c>     is using "item assignment" or "list assignment"
15:12 jnthn pmichaud: Oh, we might parse it at the wrong precedence? That'd also not surprise me...
15:13 pmichaud (and truthfully, I don't like calling this "assignment" -- it's really "initialization")
15:13 Ven_ jnthn: maybe make constants immutable only after they've been init? I'm probably talking nonsense here
15:13 jnthn pmichaud: But if so that's only half the problem - and the easy half, at that...
15:13 jnthn Ven_: Yes, "make immutable" is, uh, rather under-specified.
15:13 Ven_ fair enough.
15:13 pmichaud jnthn: why would it be only half the problem?  (not disagreeing, wanting to know)
15:14 jnthn What constant REALLY means in Perl 6 is "evalute the RHS at BEGIN time"
15:14 jnthn pmichaud: If I do
15:14 mr-foobar2 joined #perl6
15:14 jnthn my @foo = 1,2,3;
15:14 jnthn I know what @foo.WHAT is (Array)
15:14 jnthn If I do
15:14 jnthn constant @foo = 1,2,3;
15:14 jnthn What is @foo.WHAT?
15:14 Ven_ Parcel or Array
15:15 pmichaud not Parcel.
15:15 jnthn Well, if it's Parcel that we don't flatten right
15:15 jnthn And Array is...very mutable :)
15:15 Ven_ "constant @foo = 1, 2, 3;" <- array before ";", parcel after :P
15:15 jnthn I can live with it being Array, and the "assignment" done at BEGIN time.
15:16 jnthn Ven_: That won't work out, since a Parcel isn't lazy
15:16 pmichaud jnthn: well, which part of Array being mutable are you concerned about?
15:16 pmichaud the individual elements, the array, or both?
15:16 jnthn pmichaud: I'm not sure. I'm well aware that List is mutable in its own way too.
15:16 pmichaud I mean, we can certainly come up wtih Lists that aren't mutable
15:17 jnthn pmichaud: So maybe I should stpo worrying about it. I just feel a bit odd with the keyword "constant" giving me an immediately mutable data structure.
15:17 brrt joined #perl6
15:18 mr-foobar2_ joined #perl6
15:18 pmichaud well, I think I'd prefer    constant  @foo = ....    to at least give @foo the proper contents (if mutable), rather than silently give the incorrect contents
15:18 jnthn I think users will expect flattening semantics of constnat @foo = ... to match those of my @foo = ... though.
15:18 Ven_ I guess so
15:18 pmichaud i.e., I'd prefer a bug report saying  "constant @foo isn't"   to a bug report saying  "constant @foo improperly initialized"
15:18 jnthn Which is not the case right now if you aren't very careful what you put on the right hand side.
15:19 jnthn pmichaud: Yes, good point.
15:19 jnthn Well, if that's what we want, then it's probably quite fixable.
15:19 pmichaud OTOH, I just proved that the problem isn't just a parsing one
15:19 jnthn "fixable" :)
15:19 pmichaud oh, wait, no I didn't.
15:20 pmichaud hrmmmmm.....
15:20 jnthn I...don't see constant doing much about $*LEFTSIGIL
15:20 pmichaud I'll look into it more sometime.  What's the grammar rule that handles the '=' in scope declarations?
15:20 pmichaud *scoped*
15:20 jnthn oh...
15:20 jnthn pmichaud: initializer
15:21 jnthn It's a proto iirc
15:21 pmichaud got it
15:21 jnthn Yes, and constant calls that
15:21 jnthn Thing is
15:21 jnthn It looks at $*LEFTSIGIL
15:21 pmichaud right
15:21 jnthn and constant calls variable, which sets $*LEFTSIGIL
15:22 jnthn But we don't declare it in constant itself
15:22 pmichaud oh, because 'constant' isn't a scope declarator
15:23 virtualsue joined #perl6
15:23 jnthn Right, it's a type declarator
15:23 pmichaud looking.
15:23 jnthn my constant foo = ...; # lexically scoped constant
15:23 jnthn m: constant @foo = 1,2,3; say @foo.perl
15:23 camelia rakudo-moar c503e0: OUTPUT«(timeout)»
15:23 pmichaud so <declarator>  ends up calling <type_declarator>
15:23 jnthn m: constant @foo = 1,2,3; say @foo.perl
15:23 camelia rakudo-moar c503e0: OUTPUT«(1, 2, 3)␤»
15:24 jnthn m: constant $foo = 1,2,3; say $foo.perl
15:24 camelia rakudo-moar c503e0: OUTPUT«(timeout)»
15:24 pmichaud uh oh  :)
15:24 jnthn m: constant $foo = 1,2,3; say $foo.perl
15:24 camelia rakudo-moar c503e0: OUTPUT«(1, 2, 3)␤»
15:24 jnthn *sigh* camelia is...
15:25 Ven_ tired
15:25 jnthn ...a little slow sometimes these days.
15:25 pmichaud anyway, I think that constant needs to become $*LEFTSIGIL aware
15:25 jnthn Agree
15:26 pmichaud it might be that we need a scope_declarator:sym<constant>  in addition to type_declarator
15:27 pmichaud token scope_declarator:sym<constant> { <?sym> <scoped('my')> }    might do it :-)
15:27 * jnthn doesn't immediately see why
15:27 pmichaud because "constant @foo = ".... is equivalent to "my constant @foo = ..."
15:27 jnthn Well, the defualt is our, iirc.
15:27 pmichaud i.e., when "constant" is the leading keyword, it's both a scope declarator and a type declarator
15:27 pmichaud okay, "our" then :-)
15:28 jnthn True, though I think term is open to both
15:28 pmichaud ??
15:28 jnthn constant @foo = ...; already parses today
15:28 pmichaud right, but it does so because   <declarator>   falls through to <type_declarator>
15:28 jnthn and it's the type_declarator that'd need to declare $*LEFTSIGIL, so far as I can see
15:29 jnthn Yes
15:29 jnthn Is that a problem?
15:29 pmichaud looking
15:30 jnthn oh, wow
15:30 jnthn token multi_declarator:sym<null> {
15:30 jnthn :my $*MULTINESS := '';
15:30 jnthn <declarator>
15:30 jnthn }
15:30 jnthn *that* is how we end up in declarator from term, iiuc
15:30 pmichaud I'm trying to avoid having type_declarator do the $*LEFTSIGIL setting... that feels very wrong.
15:31 jnthn setting, or declaring for other things below it to set?
15:31 JimmyZ joined #perl6
15:31 pmichaud declaring for other things below it to set
15:31 pmichaud but perhaps it's okay... still figuring out how these fit together :)
15:31 jnthn Oh...
15:31 mr-foobar2_ http://queue.acm.org/detail.cfm?id=1039534 ( Extensible Programming for the 21st Century )
15:32 jnthn The only thing I see declaring it is EXPR
15:32 jnthn and comp_unit
15:32 jnthn So maybe we don't need to, after all...
15:33 jnthn Then, why doesn't it already do the right thing...
15:34 jnthn (As in, it gets the list case right, but not the constant $foo case...)
15:34 pmichaud perhaps I'm wrong about it being an incorrect parse.
15:34 pmichaud it may just be that it's not flattening properly
15:34 jnthn Yes, the latter is certainly a problem
15:34 jnthn Because it just binds what it gets to the symbol
15:34 pmichaud m:  (constant @ = 1, <b c>).perl.say
15:34 camelia rakudo-moar c503e0: OUTPUT«(1, ("b", "c"))␤»
15:34 jnthn Rather than enforcing any context on it
15:34 pmichaud well, that didn't tell me anything :-(
15:35 pmichaud m:  (constant @ = 1, <b c>).WHAT.say
15:35 jnthn Well, contrast it to my
15:35 camelia rakudo-moar c503e0: OUTPUT«(Parcel)␤»
15:35 jnthn m:  (my @ = 1, <b c>).perl.say
15:35 camelia rakudo-moar c503e0: OUTPUT«Array.new(1, "b", "c")␤»
15:35 yeahnoob joined #perl6
15:36 pmichaud but I don't know if I'm getting a Parcel with an anonymous constant, or an anonymous constant that is a Parcel
15:36 jnthn Oh...
15:36 jnthn But a named constant will tell
15:36 jnthn m: constant @a = 1, <b c>; say @a.WHAT
15:36 camelia rakudo-moar c503e0: OUTPUT«(Parcel)␤»
15:37 jnthn m: constant @a = 1, <b c>; say @a.perl
15:37 camelia rakudo-moar c503e0: OUTPUT«(1, ("b", "c"))␤»
15:38 pmichaud so, what is it that binds @a to an Array in  "my @a" ?
15:39 pmichaud or "our @a"
15:40 jnthn For the my case, it's that the lexpad is set up with an Array instance, and configured to clone it per scope.
15:40 pmichaud the 'our' case is probably more apropos
15:41 jnthn our is the same plus code is generated to re-bind it
15:41 jnthn But for the case we have here, we really want to create an Array at compile time and call .STORE with the thing the RHS produced
15:41 pmichaud yup.  probably a similar issue for constant hashes
15:41 jnthn Right
15:41 pmichaud okay.
15:42 pmichaud anyway, I've convinced myself it's not a parsing issue, so definitely not an easy fix
15:42 jnthn Aye, though I think not a terribly hard one either
15:42 pmichaud agreed
15:42 jnthn The infrastructure for doing stuff at BEGIN time these days is quite decent :)
15:42 pmichaud sadly, I have to run other errands for a while, and prepare for a soccer match
15:43 jnthn k...guess my train journey ain't for much longer too. And when I get off I'll want to do dinner :)
15:43 pmichaud see you later, or on Thursday :)
15:43 pmichaud (or Friday)
15:43 Alula joined #perl6
15:43 jnthn Yeah, I get in on Thursday early afternoon, iirc :)
15:44 brrt joined #perl6
15:44 pmichaud me too, I arrive 13:00 ish
15:44 lizmat which is earlier than we will
15:44 jnthn Oh...
15:44 jnthn pmichaud: Uh, not because you're changing at Frankfurt, by any chance?
15:44 pmichaud I'm changing at Frankfurt, yes :)
15:44 jnthn Oh
15:44 pmichaud LH1104 FRA to SZG
15:44 jnthn Then we might see each other at the gate for that flight :)
15:45 pmichaud that would be very awesome
15:45 pmichaud likely, since I have 4 hrs in FRA
15:45 pmichaud well, depending on how close your connection is
15:46 jnthn Oh, I think I'm wrong...
15:46 jnthn oh no, I misread
15:46 pmichaud APW will be the first perl event I've attended in 12 years where I don't have to prepare a talk :)  I'm looking forward to it
15:46 jnthn My flight is the 12:50 one out of FRA
15:46 pmichaud that's mine
15:47 pmichaud LH 1104 departs FRA 12:50 arrives SZG 13:45
15:47 jnthn It's an SAS codeshare, thus my momentary confusion :)
15:47 pmichaud anyway, I'll look for you at the gate
15:47 jnthn But yeah, exact same times, and says operated by Lufthansa. So it's the same one.
15:47 jnthn I get in from CPH at 11:45
15:47 pmichaud I arrive FRA from DFW at 08:50
15:48 jnthn Though Terminal 3, and we go out of Terminal 1, so I guess I'll have a bit of walking to do.
15:48 pmichaud anyway, great to know we're on the same flight -- we can navigate from SZG to hotel together too :)
15:48 lizmat yup, you're in the same hotel  :-)
15:49 pmichaud lizmat++
15:50 pmichaud anyway, I'd better get my other tasks done or I will miss a deadline here.
15:50 jnthn Sure. See you somewhere in FRA on Thursday :)
15:50 pmichaud afk
15:50 jnthn lizmat: Are you flying in also, or driving it?
15:50 lizmat driving...  with a small detour  :-)
15:51 jnthn OK :)
15:51 * jnthn will see if there's a sane public transport option from airport to hotel for him and Pm
15:52 jnthn If not, I think the airport is close enough to town to taxi it at a sane enough price :)
15:57 lizmat jnthn: the airport is about 5km from the center
15:57 lizmat it is *really* close
15:57 lizmat (and not too large either, only 1 runway)
15:57 Ven_ joined #perl6
15:58 jnthn ah, cool
16:03 gtodd depending on luggage, weather and the surrounding environment, 5km is pleasant walk.
16:16 brrt joined #perl6
16:18 psch alright, the cli-opts grammar prototype seems to work, maybe i can get it into HLL-space during next week.  i'm out of brain for today :)
16:19 b2gills m: $*HOME
16:19 camelia rakudo-moar f9c4d0: OUTPUT«Unhandled exception: Dynamic variable $*HOME not found␤   at <unknown>:1  (/home/p6eval/rakudo-inst-1/languages/perl6/r​untime/CORE.setting.moarvm:throw:4294967295)␤ from src/gen/m-CORE.setting:13680  (/home/p6eval/rakudo-inst-1​/languages/perl6/runtime/…»
16:19 psch i will probably ask about ways to improve it as well, because i don't really feel particularly confident with grammars
16:19 psch but that's for somewhen next week as well o/
16:19 psch for now, ttyl #perl6
16:19 jnthn o/ psch
16:20 jnthn psch++ # working on getting command line stuff more in shape
16:20 timotimo mhhh. is line a shape?
16:21 lizmat hmmm...  seems dalek is not getting update info
16:21 lizmat just did: https://github.com/rakudo/rakudo/commit/​0c53bfea29971c6aa59993e13d0fbd45487a2183
16:23 dalek joined #perl6
16:23 [Coke] joined #perl6
16:24 Hor|zon joined #perl6
16:25 pmichaud joined #perl6
16:27 PerlJam joined #perl6
16:28 sergot joined #perl6
16:28 gtodd do any p6-istas have a REPL history mechanism worked out?
16:28 Util joined #perl6
16:29 gtodd I mean a method for preserving it to a file ...  I can arrow up and down through it but was wondering how I could access it as a list in the repl itself ... possible with nqp magic or something?
16:44 carlin "$*CWD is a Str, not an IO::Path!!!"
16:44 carlin 3 exclaimation points, must be serious
16:47 FROGGS[mobile]2 gtodd: that stuff happens at VM level, so it is not accessible
16:49 tadzik joined #perl6
16:50 FROGGS[mobile]2 lizmat: what's the hotel name the Perl 6 crew is staying?
16:50 lizmat define Perl 6 crew ?
16:52 FROGGS[mobile]2 I dunno... you, wendy, Pm and jnthn?
16:53 FROGGS I'm in hotel Plainbrücke
16:53 lizmat http://act.useperl.at/apw2014/​wiki?node=Who%20stays%20where
16:53 * timotimo is in "hotel guter hirte"
16:54 timotimo together with vendethiel
16:54 lizmat afaik, nwc10 is thare as well
16:54 lizmat *there
16:54 FROGGS well, we're are at last at the same venue :o)
16:54 vendethiel- yarr
16:56 psch joined #perl6
16:57 psch ...i can't not read the clog, and gtodd's question is intriguing :)
16:58 psch gtodd: as it is right now you'd have to patch nqp
16:58 timotimo we'll eventually want something much cleverer than just a little readline-based thingie (or in this case linenoise)
16:58 psch gtodd: more precisely, you'd have to add some check (like and env-var) for a filename and add "write the line of code to that file" in HLL::Compiler::interactive
16:59 psch s:1st/and/an/
17:00 psch timotimo: i have (very minimalistic) binding for gnu-readline on github (http://github.com/peschwa/readline-gnu, but "which readline" probably isn't the biggest thing to think about... :)
17:00 gtodd OK I had some amateurish notion that since nqp had "linenoise inside" I could somehow access whatever it was doing biut that it would be at a sort deeper/llower than the repl itsell
17:01 psch in my understand we want subparse for every line and... that's kind of as far as i thought about it :)
17:01 timotimo iirc moarvm just offers an "interactive readline" nqp:: op for you
17:01 gtodd :-)
17:02 pat_js left #perl6
17:02 gtodd so at this point it wouldn't be as easy as "loading a package" that writes to a file ...
17:02 pat_js joined #perl6
17:03 psch gtodd: no.  nqp reads the fh for the repl
17:03 psch line 70 in nqp/src/HLL/Compiler.nqp
17:04 pat_js are there some deltas (apart from the git logs) for rakudo? I haven't looked at perl 6 for some time now and want to know what the changes are implemented now and what crucial changes to the language have happend.
17:05 kjs_ joined #perl6
17:06 gtodd the easy approach of REPLHelper.pm or repl-trick for "shortcuts"  isn't going to work at this point so unless I dig into nqp ... :-)
17:06 psch pat_js: as for "what works now", http://perl6.org/compilers/features is probably most comprehensive
17:07 psch pat_js: wrt "what started working when", that's probably only via git
17:08 psch pat_js: oh, except for timotimo++'s http://p6weekly.wordpress.com although that only goes back about half a year or so i think
17:09 pat_js psch: thanks, but that's too short, I'm looking for a rundown of the last two years, then i'm looking into the git logs.
17:09 Mso150 joined #perl6
17:11 timotimo pat_js: what's your favourite part of perl6?
17:11 timotimo the last two years is basically all the time i spent with perl6 so far
17:11 timotimo performance improved in a big way
17:12 timotimo stage parse for the core setting used to be about 120 seconds on my machine, now it's below 30
17:12 timotimo in part thanks to moarvm, in part thanks to targeted optimizations all over the place
17:12 timotimo we got inlining to work on the AST level for all backends, too
17:13 colomon Threading
17:13 timotimo about 2 years ago, there were still loads and loads of tickets marked "nom regression"
17:13 colomon Sets and Bags
17:13 timotimo aye, the JVM and MoarVM backends appeared and got multithreading and asynchronous IO
17:14 timotimo niecza stopped being moved forward, pugs has been removed from the daily test runs
17:15 gtodd psch: of course running the repl in tmux or in emacs buffer offers a different way to save history ... but you can't really "load" the file ...
17:15 gtodd anyway thanks
17:15 timotimo rlwrap*
17:16 timotimo i mean: rlwrap is a workable alternative
17:16 pat_js timotimo: not sure, I really liked the full utf8 support and all those operators. method dispatch was pretty nice. the Meta-Programming posibilities had some nice touch to it.
17:17 timotimo our unicode support is still lacking; we'll still have to build NFG before that'll truly rock your socks off
17:17 timotimo but method dispatching, meta programming, the operators ... that's all in good shape; i think it was in good shape before, too
17:18 pat_js thanks a lot, that's some directions. Is the parcel/listy stuff now working?
17:18 [Sno] joined #perl6
17:18 psch it's up next, i think
17:18 psch well, the performance part of it
17:18 timotimo there'll still be the Great List Refactor that'll be a pretty huge deal when it comes to performance at least
17:18 timotimo a portion of that is going to happen during the APW2014 which happens next weekend
17:19 psch right, that reminds me, i should probably add that one use case i discovered when hacking m:g// to the meta ticket
17:19 timotimo yay
17:20 psch m: class A { method Str() { "foo" } method List() { (1, 2, 3) } }; for A.new { .say }
17:21 camelia rakudo-moar f9c4d0: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/kXO2S_REWTâ�¤Two terms in a rowâ�¤at /tmp/kXO2S_REWT:1â�¤------> [32mclass A { method Str() { "foo" } [33mâ��[31mmethod List() { (1, 2, 3) } }; for A.new[0mâ�¤    expecting any of:â�¤        pos…»
17:21 psch m: class A { method Str() { "foo" }; method List() { (1, 2, 3) } }; for A.new { .say }
17:21 camelia rakudo-moar f9c4d0: OUTPUT«A.new()␤»
17:21 * psch shakes fist at TTIAR
17:21 zakharyas joined #perl6
17:21 psch that case is arguably correct, but 'for m:g/$pat/ { ... }' likely wants .list implicitely...
17:22 psch welp, i'll add it to the ticket and let the smart people worry about it :)
17:23 molaf joined #perl6
17:23 fhelmberger joined #perl6
17:24 bartolin I've got a question regarding unicode: on a box with debian stable (7.6) I get spectest failures with rakudo on parrot in S05-mass/charsets.t and S05-mass/properties-general.t (the failures stem from the recently unfudged tests, cmp. https://github.com/perl6/roast/commit/60​6523b3e5d9ba10bc67bb7705b0530ff5ce049c)
17:24 bartolin I *think* I understand why those tests fail: debian stable comes with libicu version 4.8 which is used by parrot. libicu 4.8 only has the data from Unicode 6.0 -- but the failing tests check definitions from Unicode 6.1
17:24 bartolin can you confirm this interpretation?
17:26 denis_boyun_ joined #perl6
17:31 zakharyas joined #perl6
17:33 dalek specs: e9650c0 | (Elizabeth Mattijsen)++ | S32-setting-library/IO.pod:
17:33 dalek specs: All negated letters also in IO::Path.all
17:33 dalek specs: review: https://github.com/perl6/specs/commit/e9650c02f4
17:39 dalek rakudo/nom: 0d7664f | (Elizabeth Mattijsen)++ | src/core/IO/Path.pm:
17:39 dalek rakudo/nom: Implement IO::Path.all
17:39 dalek rakudo/nom:
17:39 dalek rakudo/nom: So now you can say "foo".IO.all: <!d r w x s> to get the size of something
17:39 dalek rakudo/nom: that is *not* a directory, but is readable/writable/executable.
17:39 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0d7664f7ff
17:40 ghostlines joined #perl6
17:42 Ven_ joined #perl6
17:42 dwarring joined #perl6
17:46 lizmat is it me, or is the spectest significantly faster after my latest IO::Path work ?
17:46 lizmat Files=923, Tests=33895, 173 wallclock secs ( 9.02 usr  3.39 sys + 1041.16 cusr 158.50 csys = 1212.07 CPU)
17:47 colomon is it that or JIT becoming mainstream?
17:48 * colomon has definitely seen a noticeable ~3% decrease in stage parse times this week...
17:49 lizmat colomon: there is temporarily more code in the setting while I move stuff around
17:50 lizmat IO:FileTestable is about to be removed :-)
17:50 colomon lizmat: sorry, didn't say that right.
17:50 colomon parse times this week are the fastest I've ever seen
17:50 lizmat ah, even better!
17:50 colomon 26.2 overnight.
17:50 colomon it was 27 or os just a few days ago.
17:51 colomon it had gotten down to something like 26.5 before a while ago, then retreated back to 27, but the last few days have been great.
17:52 dalek rakudo/nom: 1a0e4a0 | (Elizabeth Mattijsen)++ | src/core/IO/Path.pm:
17:52 dalek rakudo/nom: Optimize IO..Path.all: <e !e>
17:52 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1a0e4a0398
17:52 dalek rakudo/nom: 5065a05 | (Elizabeth Mattijsen)++ | docs/ChangeLog:
17:52 dalek rakudo/nom: Update ChangeLog with some new features
17:52 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/5065a05c45
17:52 colomon lizmat: I really don't have any idea whether it's your code or JIT -- my timings are all with the default Moar rakudobrew gives you.
17:53 colomon I had just assumed JIT because the timing seemed about right.
17:53 lizmat mine as well
17:53 lizmat then you have now JIT where you didn't have that before
17:53 colomon right
17:53 lizmat I've been working with JIT enabled for several weeks now
17:53 colomon that's why I assume JIT was responsible for the speedup.  :)
17:53 lizmat well, expect a little more on top of that for the spectest
17:54 lizmat by the looks of it now
17:54 colomon \o/
17:55 dalek rakudo/nom: 47f8a12 | (Elizabeth Mattijsen)++ | docs/ChangeLog:
17:55 dalek rakudo/nom: Add blurb about IO::Path.modified|accessed|changed
17:55 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/47f8a12ff4
17:56 gtodd bartolin: I hope someone confirms that deduction :-)
17:56 gtodd bartolin: that would mean failing parrot tests might magically start to pass !
17:59 flussence gtodd: not sure about that - I've seen S05 tests persistently fail even though I'm on a Gentoo with shiny new ICU versions...
18:17 dalek rakudo/nom: 6661df6 | (Elizabeth Mattijsen)++ | src/core/ (4 files):
18:17 dalek rakudo/nom: Bring $*SPEC.tmpdir up to spec, and $*TMPDIR init
18:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6661df6580
18:18 lizmat could be noise, but this is looking good: Files=923, Tests=33895, 170 wallclock secs ( 8.91 usr  3.39 sys + 1029.64 cusr 156.33 csys = 1198.27 CPU)
18:18 lizmat below 1200 CPU again
18:21 ivanshmakov joined #perl6
18:22 vladtz joined #perl6
18:24 virtualsue joined #perl6
18:25 Hor|zon joined #perl6
18:28 dwarring bartolin: S05-mass/chars.t is failing for me under unicode version 5.2
18:29 dwarring https://gist.github.com/dwa​rring/e60d5e134dd980a9aab2
18:30 dwarring I think this is unicode version dependant
18:30 lizmat bartolin stated earlier that the failing tests were for Unicode 6.1
18:31 dwarring can anyone reproduce that with 6.3 onwards?
18:32 bartolin I'm not really sure, all of them are from 6.1. (the comments for fudging said so)
18:32 bartolin dwarring: icu-config --unicode-version <- nice to know!
18:33 dwarring I think the comment could be wrong (I wrote it)
18:33 bartolin on debian stable it gives 6.0 and I got test failures. on freebsd 10.0 it gives 6.3 and I have a clean spectest
18:33 bartolin ;-)
18:34 * dwarring bisecting unicode versions
18:34 bartolin dwarring: at least for the test in S05-mass/charsets.t ('Punctuation chars') the added characters are from 6.1 (just looked that up).
18:35 dalek rakudo/nom: 603a994 | usev6++ | tools/lib/NQP/Configure.pm:
18:35 dalek rakudo/nom: Fix open command
18:35 dalek rakudo/nom:  could contain a space (e.g. when installing in dir with space)
18:35 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/603a994874
18:35 dalek rakudo/nom: 2e30a3b | (Tobias Leich)++ | tools/lib/NQP/Configure.pm:
18:35 dalek rakudo/nom: Merge pull request #316 from usev6/nom
18:35 dalek rakudo/nom:
18:35 dalek rakudo/nom: Fix open command
18:35 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2e30a3b8d2
18:36 bartolin http://www.unicode.org/versions/Unicode6.1.0/ mentions U+00A7 and U+00B6 and those appear in S05-mass/chars.t
18:36 dalek nqp: 52c4e54 | usev6++ | tools/lib/NQP/Configure.pm:
18:36 dalek nqp: Fix open command
18:36 dalek nqp:  could contain a space (e.g. when installing in dir with space)
18:36 dalek nqp: review: https://github.com/perl6/nqp/commit/52c4e54f02
18:36 dalek nqp: 0bd419f | (Tobias Leich)++ | tools/lib/NQP/Configure.pm:
18:36 dalek nqp: Merge pull request #188 from usev6/master
18:36 dalek nqp:
18:36 dalek nqp: Fix open command
18:36 dalek nqp: review: https://github.com/perl6/nqp/commit/0bd419f713
18:36 Mouq joined #perl6
18:36 dwarring bartolin: I'll refudge, this with a more accurate comment
18:36 lizmat FROGGS[mobile]2: should be up NQP_REVISION?
18:37 lizmat *we
18:37 * dwarring $?UNICODE would be nice to have - nyi
18:37 dalek doc: bdbc61e | Mouq++ | lib/Language/variables.pod:
18:37 dalek doc: Clarify some of the twigil docs
18:37 dalek doc: review: https://github.com/perl6/doc/commit/bdbc61ea4f
18:37 dalek doc: c986f43 | Mouq++ | lib/Language/variables.pod:
18:37 dalek doc: Stub Variable Declarator section
18:37 dalek doc:
18:37 dalek doc: Though perhaps declarators and/or scope deserve their own document(s)? At
18:37 dalek doc: the very least, though, they should be mentioned in variables.pod.
18:37 dalek doc: review: https://github.com/perl6/doc/commit/c986f43015
18:37 FROGGS lizmat: yes, but I don't think it is very urgent
18:38 bartolin lizmat: the fixes FROGGS++ just merged don't help without further changes
18:39 rindolf joined #perl6
18:39 lizmat ah, ok
18:39 bartolin lizmat: I probably shouldn't have created a pull request just for those.
18:40 lizmat no worries, I was just thinking about releasing early and often :-)
18:40 [Tux] joined #perl6
18:40 dalek doc: 08bc0fd | Mouq++ | lib/Language/variables.pod:
18:40 dalek doc: Call twigils "twigils" in headers
18:40 dalek doc: review: https://github.com/perl6/doc/commit/08bc0fd567
18:41 bartolin dwarring: We don't have a way to fudge according to an second criterion (apart from backend), do we? (say OS version)
18:43 kjs_ joined #perl6
18:45 dwarring bartolin: not that I know of
18:46 FROGGS bartolin: no
18:47 lizmat could it be that nqp::loadbytecode($path); creates an IO::Handle with a Str in its $!path  ??
18:48 bartolin maybe the tests could stay unfudged but get comments describing the issue?
18:48 lizmat if so, how can I change the Str $path in load_setting to an IO::Path ?
18:48 pepl joined #perl6
18:48 lizmat FROGGS:  any ideas?  ^^
18:49 FROGGS hmmm
18:49 lizmat I could fudge it for now in IO::Handle.open, but that seems the wrong way  :-)
18:51 FROGGS I don't think that nqp::loadbytecode can create a Perl 6 level class
18:52 lizmat aaaahhhh  FROGGS++  yup, I see what the problem is
18:53 lizmat open('-', :w)
18:56 dalek roast: 5dc7e0d | (David Warring david.warring@gmail.com)++ | S05-mass/charsets.t:
18:56 dalek roast: refudged unicode-version dependant test on parrot. Correct comments - need unicode 6.1 bartolin++
18:56 dalek roast: review: https://github.com/perl6/roast/commit/5dc7e0d098
18:56 FROGGS why don't we use nqp::getstdout/in there?
18:56 lizmat we do, deeper in the code
18:59 * lizmat is trying to not break too much at once
19:00 jnthn evening, #perl6
19:00 lizmat jnthn o/
19:01 FROGGS lizmat: btw, the segfault in lines.t (and perhaps others) is due to my repl fix... and I think I just solved that
19:01 FROGGS hi jnthn
19:01 lizmat FROGGS++
19:03 * jnthn is safely in Stockholm, ate a curry, and went for a walk :)
19:03 jnthn It's nice to be in a European city again :)
19:03 Mso150_h joined #perl6
19:04 lizmat I bet it's a bit cooler  :-)
19:05 jnthn Oh yes :D
19:05 jnthn It's actually comfortable to go for an evening stroll here :)
19:05 FROGGS jnthn: does that work on windows? perl6-p -e 'say shell("echo \"exit(42)\" | $*EXECUTABLE")'
19:06 FROGGS (I guess not)
19:07 jnthn C:\consulting\rakudo>perl6-m -e "say shell('echo ""exit(42)"" | ' ~ $*EXECUTABLE)"
19:07 jnthn exit(42)
19:07 jnthn Proc::Status.new(exit => 0, pid => Any, signal => 0)
19:08 jnthn Though note:
19:08 jnthn C:\consulting\rakudo>perl6-m -e "say shell('echo exit(42) | ' ~ $*EXECUTABLE)"
19:08 jnthn Proc::Status.new(exit => 42, pid => Any, signal => 0)
19:08 FROGGS hmmm
19:08 jnthn The first one actually includs the quotes in what it echos :D
19:08 moritz \o
19:08 lizmat moritz o/
19:09 jnthn C:\consulting\rakudo>echo "WHY, ffs!"
19:09 jnthn "WHY, ffs!"
19:09 jnthn :D
19:09 FROGGS bah
19:09 FROGGS *g*
19:09 FROGGS but I can live with that :o)
19:09 jnthn o/ moritz
19:09 FROGGS thanks
19:09 guru joined #perl6
19:11 dalek doc: ea84e6c | Mouq++ | lib/Language/variables.pod:
19:11 dalek doc: Attempt to explain lexical scope
19:11 dalek doc: review: https://github.com/perl6/doc/commit/ea84e6c4bb
19:13 pat_js is there a way to "apply" a parcel to a signature?
19:13 Ven_ joined #perl6
19:14 pat_js or create a subroutine with a given Signature as it's signature
19:14 Mouq pat_js: What do you mean by "apply"?
19:15 kjs_ joined #perl6
19:17 moritz you can interpolate it
19:17 moritz m: my @parcel := (4, 1); say 'abcdefghijk'.substr(|@parcel);
19:17 moritz m: my @parcel := (4, 1); say 'abcdefghijk'.substr(|@parcel);
19:17 camelia rakudo-moar 47f8a1: OUTPUT«e␤»
19:17 camelia rakudo-moar 47f8a1: OUTPUT«e␤»
19:17 jnthn In the unlikely even reading the question too literally actually helps...
19:18 jnthn m: sub swas($sig) { return sub (|c where $sig) { say 'oh yeah!!!' } }; my $c = swas(:(Int $a, Str $b)); $c(1, "omg"); $c(2, 69);
19:18 camelia rakudo-moar 47f8a1: OUTPUT«oh yeah!!!␤Constraint type check failed for parameter 'c'␤  in sub  at /tmp/2R4AbM7uGQ:1␤  in block <unit> at /tmp/2R4AbM7uGQ:1␤␤»
19:18 pat_js yeah thatnks, that's what I wanted.
19:18 jnthn ...what for?! :D
19:18 pat_js curiosity
19:18 jnthn Ah :)
19:18 jnthn As I was writing it I was thinking...what on earth could this be good for... :)
19:19 pat_js i thought if everything is an object
19:19 jnthn *nod*
19:19 jnthn Yeah, sigs are objects...
19:20 pat_js and Parcels are sort of the "keys" to the Signature "locks", this should work in some way
19:21 Mouq m: say (1, "omg") ~~ :(Int $a, Str $b)
19:21 camelia rakudo-moar 47f8a1: OUTPUT«True␤»
19:21 Mouq m: say (2, 69) ~~ :(Int $a, Str $b)
19:21 camelia rakudo-moar 47f8a1: OUTPUT«False␤»
19:22 pat_js but there is no way to do variable binding with it, right?
19:23 SamuraiJack_ joined #perl6
19:23 jnthn You mean, use it for destructuring?
19:23 pat_js yeah, I thought about some pattern matching
19:23 jnthn Hmm
19:23 jnthn Well
19:24 pat_js but it feels like leaking scopes.
19:24 jnthn m: sub foo() { return 12, "dugong"; }; my (Int $a, Str $b) = foo(); # there's this...
19:24 camelia rakudo-moar 47f8a1: ( no output )
19:24 jnthn m: sub foo() { return 12, "dugong"; }; my (Int $a, Str $b) = foo(); say $a; say $b;
19:24 camelia rakudo-moar 47f8a1: OUTPUT«12␤dugong␤»
19:24 dalek rakudo/nom: 2e4dca6 | (Elizabeth Mattijsen)++ | / (4 files):
19:24 dalek rakudo/nom: Add IO::Path::Special for STDIN/STDOUT/STDERR etc
19:24 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/2e4dca68fa
19:25 Mouq It doesn't really make sense to have first-class variable declarations :P IMO
19:25 pat_js yeah i have found that
19:25 kjs_ joined #perl6
19:26 jack_rabbit joined #perl6
19:26 jnthn One of the cute things you can do with the syntax is unpack object attributes
19:26 jnthn my (:$numerator, :$denominator) := 4.2; say $numerator; say $dominator
19:26 jnthn m: my (:$numerator, :$denominator) := 4.2; say $numerator; say $dominator
19:26 camelia rakudo-moar 47f8a1: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/_OoupbHq82â�¤Variable '$dominator' is not declared. Did you mean '$denominator'?â�¤at /tmp/_OoupbHq82:1â�¤------> [32m) := 4.2; say $numerator; say $dominator[33mâ��[31m<EOL>[0mâ�¤    expecting …»
19:27 jnthn uh, yes, I *did* mean...
19:27 jnthn m: my (:$numerator, :$denominator) := 4.2; say $numerator; say $denominator
19:27 camelia rakudo-moar 47f8a1: OUTPUT«21␤5␤»
19:27 dalek roast: ac5e8d4 | (Tobias Leich)++ | S19-command-line/repl.t:
19:27 dalek roast: add sanity test for the REPL
19:27 dalek roast: review: https://github.com/perl6/roast/commit/ac5e8d4e35
19:28 jnthn And can do them nested too
19:32 jnthn FROGGS: That test fails here :(
19:32 pat_js m: my(:$re,:$im) = 3+4i; $im.say
19:32 camelia rakudo-moar 47f8a1: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/CxX0G6tvtw�Variable '$re' is not declared�at /tmp/CxX0G6tvtw:1�------> [32mmy(:$re[33m�[31m,:$im) = 3+4i; $im.say[0m�»
19:32 jnthn oh, no wonder!
19:32 dalek roast: 18fc4b8 | jonathan++ | S19-command-line/repl.t:
19:32 dalek roast: Fix a tyop.
19:32 dalek roast: review: https://github.com/perl6/roast/commit/18fc4b8e6a
19:32 FROGGS ohh nose!
19:32 jnthn Now it passes just fine.
19:34 pat_js m: my (:$im) := 3+4i; $im.say
19:34 camelia rakudo-moar 47f8a1: OUTPUT«␤  in block <unit> at /tmp/CdfqrSeiYS:1␤␤»
19:34 FROGGS no wonder it passed here :o)
19:34 robinsmidsrod joined #perl6
19:35 lizmat jnthn: nqp::bindattr(self, IO::Handle, '$!PIO', $foo) is just $!PIO := $foo, no?
19:35 moritz if the current class is IO::Handle, yes
19:35 jnthn lizmat: Yeah
19:35 jnthn Well, what moritz said
19:35 lizmat that's what I thought
19:36 lizmat another cleanup coming :-)
19:39 Mouq m: my (:$re,:$im) = 3+4i; $im.say
19:39 camelia rakudo-moar 47f8a1: OUTPUT«(Any)␤»
19:39 Mouq m: my (:$re,:$im) := 3+4i; $im.say
19:39 camelia rakudo-moar 47f8a1: OUTPUT«4␤»
19:39 Mouq pat_js: You were missing a space between `my` and `(` that caused it to be treated like a function
19:42 Mouq masak: It would be awesome to be able to do something like: sub where ($v, *@subs) { my @whens = @subs.map({ quasi {when {{{$_}}}.signature { $_ ~~ {{{$_}}} } } }); quasi { given {{{$v}}} { {{{@whens}}} } } }
19:44 jnthn FROGGS: don't forget to add it to spectest.data, btw :)
19:45 MilkmanDan joined #perl6
19:47 timotimo m: say &my
19:47 camelia rakudo-moar 47f8a1: OUTPUT«[31m===[0mSORRY![31m===[0m Error while compiling /tmp/2v6oWMy48Gâ�¤Undeclared routine:â�¤    &my used at line 1â�¤â�¤Â»
19:47 FROGGS jnthn: will do after more testing
19:47 timotimo maybe it should output the worries after a sorry, too?
19:47 timotimo in that case it'd complain about the "undefined routine my used"
19:48 jnthn FROGGS: ah, cool :)
19:49 jnthn hm, why on earth do my laptop and phone have an hour's difference on what time it is...
19:50 jnthn All the same, I'm happy to discover I've an hour more of evening than I was thinking... :)
19:51 timotimo \o/
19:52 anaeem1_ joined #perl6
19:55 anaeem1 joined #perl6
19:56 dalek rakudo/nom: 94a64fb | (Tobias Leich)++ | src/core/IO/Handle.pm:
19:56 dalek rakudo/nom: fix .lines on parrot
19:56 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/94a64fba41
19:57 xenoterracide joined #perl6
20:00 anaeem1 joined #perl6
20:09 dalek doc: 0810e68 | Mouq++ | lib/Language/variables.pod:
20:09 dalek doc: Clarify special vars section
20:09 dalek doc: review: https://github.com/perl6/doc/commit/0810e68dca
20:09 dalek doc: ed0ba4d | Mouq++ | lib/Language/variables.pod:
20:09 dalek doc: Explain temp
20:10 dalek doc: review: https://github.com/perl6/doc/commit/ed0ba4d4e1
20:11 beastd joined #perl6
20:13 rmgk_ joined #perl6
20:14 anaeem1 joined #perl6
20:24 denis_boyun_ joined #perl6
20:26 Hor|zon joined #perl6
20:27 anaeem1 joined #perl6
20:27 _slade_ joined #perl6
20:36 lizmat S03:279 : so what is a 'signature literal' ??   :(Dog $self:)
20:36 synopsebot Link: http://perlcabal.org/syn/S03.html#line_279
20:40 jnthn m: say :(Int $x).WHAT
20:40 camelia rakudo-moar 94a64f: OUTPUT«(Signature)␤»
20:42 lizmat just a short way to create a signature ?
20:42 jnthn a way to create a standalone signature, independent of any sub, method, block, etc.
20:47 lizmat ack
20:47 japhb lizmat, FROGGS: Are we currently in a place where path/IO stuff is working, in that e.g. panda, REPL, etc. is unbroken?
20:47 lizmat it was earlier today, lemme check now
20:48 japhb I've been seeing enough bug fixes that I'm not sure if now is a good time to build a perl6 that I plan to keep around for a few days (I generally like to have at least one built that I can rely on working, even when HEAD builds are various levels of broken)
20:48 japhb lizmat: thank you.  :-)
20:49 jnthn japhb: You might want to hold off a little.
20:49 japhb jnthn: Oh?
20:49 jnthn japhb: If you're after less churn, at least.
20:50 japhb Was there any discussion of the merits of keeping nom a little less churn-y?
20:50 lizmat HEAD currently rebootstraps panda fine, even with Inline::Perl5 and its dependencies
20:50 japhb It seems like the past month or so, it's been "jungles of Borneo"
20:51 lizmat but we're getting a faster IO and filetests for it  :-)
20:52 japhb Oh sure, but it seems like enough change that, well, panda and the REPL need regular love.  I'm wondering if I'm alone in thinking nom should be "always shippable", so to speak, or if we've made the conscious decision that only monthly releases can be considered that reliable.
20:53 jnthn japhb: It's a difficult trade-off. For various reasons, a lot of people track HEAD rather than releases, so there's some demand to keep HEAD stable. OTOH, putting stuff off in branches is a great way to take a long time to discover breakage...
20:54 jnthn japhb: Given that doing so greatly reduces the amount of testing stuff gets.
20:54 japhb Yeah, I agree with that.
20:54 lizmat japhb: it also shows roast deficiencies
20:54 jnthn Right, that too.
20:54 lizmat last week, spectest was fine, but panda was severely borked
20:54 jnthn the daily roast for both spectests and modules uses nom
20:55 * japhb wonders if there's value in keeping a pointer (branch, moving tag, what have you) of not-quite-churny-but-close-to-HEAD.  Kindof like Firefox Aurora
20:55 jnthn I guess the issue is, a bunch of things are set up such that you get more value, as a developer, for comitting might-be-OK stuff to nom, than putting it in a branch...
20:56 japhb Well, I'm not suggesting putting new code in a branch, because I pretty much agree with the reasons that doesn't really work for us except for e.g. major refactors.
20:56 jnthn Aye. The "moving pointer" idea ain't a bad one.
20:56 japhb I'm thinking more having a way for people to get something between monthly and nom in both recency and risk.
20:57 jnthn I mean, we already have MOAR_REVISION/NQP_REVISION to give us some control there.
20:57 japhb Yeah, actually, that's kinda what I want, but for Rakudo itself.
20:57 jnthn Right.
20:58 jnthn It's an interesting idea. I'm not totally sure how to set things up so the incentive system is right to get it to have the right stability/recency trade-off...
20:58 lizmat but that would be the equivalent of putting all code after such a point in a branch and untested
20:59 japhb But since {MOAR,NQP}_REVISION exist in the next higher repo in the implementation stack, but Rakudo is already the top of the stack, you'd probably end up needing something like a git tag to get the same functionality in a usable way.
20:59 jnthn lizmat: Only if everybody suddenly jumped to using that branch insted of nom.
21:00 jnthn lizmat: and I'm pretty sure the idea is not to switch roasts, evalbot, etc over do it.
21:00 lizmat ah, ok the other way around
21:00 jnthn japhb: Do you think an automated approach may be viable?
21:00 lizmat but that could cause roast failures in a tagged "safe" version
21:01 japhb lizmat: I don't think so, because I feel like we've got several different risk/recency tradeoffs:  monthly (stable but as old as we go), branch (likely broken, or we'd merge it), nom (may be broken, but the best place for patches and such), and tag (probably not broken, may be a few days old, good for people working on features that don't want to get broken builds while they work)
21:01 jnthn japhb: That is, clean roast + panda verified as working in some sense = branch moved
21:02 jnthn Or some similar criteria.
21:02 pat_js m: sub infix:<←> (&f, $b){->(|c){f($b,c)}};my &add = *+*;my &add_1 = &add ← 1;say add_1(5)
21:02 camelia rakudo-moar 94a64f: OUTPUT«1␤»
21:02 pat_js why isn't that 6?
21:02 japhb jnthn: Yeah, I think that could actually work for us.  Sortof a "no unknown regressions" tag/branch
21:03 jnthn Also making it automatable/objective makes it rather predictable what it provides.
21:03 jnthn (And what the boundaries of what it can promise are.)
21:03 japhb Agreed, and agreed that's a good thing.
21:03 kjs_ joined #perl6
21:04 jnthn pat_js: Did you want |c in the call too, to flatten the capture there?
21:05 jnthn pat_js: Also I don't think you want parens around the signature of the pointy
21:05 pat_js jnthn: yeah, basically i wanted some kind of curry operator
21:05 timotimo pat_js: do you hate whitespace? :)
21:05 Mouq -> |c { f($b, |c) }
21:05 japhb [Coke]: Would you consider making that part of your roasting?  If the roast is clean, and panda can rebootstrap with some set of "expected always working" modules, then your scripts push the tag/branch pointer ahead?
21:05 pat_js Mouq: no, not at all
21:05 Mouq Orrrr, &f.assuming($b)
21:06 jnthn I was gonna say, try assuming :)
21:06 dalek rakudo/nom: 7c74f70 | jonathan++ | src/Perl6/Metamodel/BOOTSTRAP.nqp:
21:06 dalek rakudo/nom: Allow build to be set in Attribute.new(...).
21:06 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7c74f7067f
21:06 timotimo m: my &infix:<←> = { $^a.assuming($^b) }; my &add = * + *; my &add_1 = &add ← 1; say add_1(5); # ? :)
21:06 camelia rakudo-moar 94a64f: OUTPUT«No such method 'assuming' for invocant of type 'WhateverCode'␤  in block <unit> at /tmp/sQy1uAJXgq:1␤␤»
21:07 jnthn Use &[+] instead of *+*, I think
21:07 timotimo m: my &infix:<←> = { $^a.assuming($^b) }; my &add = &[+]; my &add_1 = &add ← 1; say add_1(5); # ? :)
21:07 camelia rakudo-moar 94a64f: OUTPUT«6␤»
21:07 timotimo m: my &infix:<←> = { $^a.assuming($^b) }; my &add = &infix:<+>; my &add_1 = &add ← 1; say add_1(5); # ? :)
21:07 camelia rakudo-moar 94a64f: OUTPUT«6␤»
21:07 timotimo i think we really want WhateverCode to be able to assuming, too
21:08 jnthn I'm not sure we do...
21:08 japhb Agreed.
21:08 japhb jnthn: Why not?
21:08 japhb Why wouldn't any Callable not be able to .assuming
21:08 Mouq m: say (my $= * + *).^mro; say &[+].^mro
21:08 camelia rakudo-moar 94a64f: OUTPUT«(WhateverCode) (Code) (Any) (Mu)␤(Sub+{<anon>}+{Precedence}) (Sub+{<anon>}) (Sub) (Routine) (Block) (Code) (Any) (Mu)␤»
21:08 japhb ?
21:08 jnthn Oh, wait... .assuming, not .wrap...
21:08 jnthn Yeah, sorry, I was just confusing things :)
21:09 jnthn No objection to .assuming working more widely.
21:09 timotimo how about Int.assuming? :)
21:09 Mouq .assuming is defined in Routine
21:09 jnthn m: say WhateverCode.^mro
21:09 camelia rakudo-moar 94a64f: OUTPUT«===SORRY!===␤Internal error: failed to remove block␤»
21:09 jnthn um...pardon?
21:10 timotimo we must factor it out into a routine named CanAssume? HasAssumptions? :)
21:10 jnthn m: say (my $a = WhateverCode).^mro
21:10 camelia rakudo-moar 94a64f: OUTPUT«(WhateverCode) (Code) (Any) (Mu)␤»
21:10 kjs_ joined #perl6
21:10 jnthn Hm, it'd have to go in Code to have it in a common base class...
21:10 pat_js hmm, thanks a lot
21:10 japhb .ask [Coke] What do you think of http://irclog.perlgeek.de/​perl6/2014-10-05#i_9460362 through http://irclog.perlgeek.de/​perl6/2014-10-05#i_9460432 ?
21:10 yoleaux japhb: I'll pass your message to [Coke].
21:11 Mouq m: say (Whatever + Whatever)(4, 5)
21:11 camelia rakudo-moar 94a64f: OUTPUT«9␤»
21:15 dalek modules.perl6.org: 6c2fce9 | asb-cpan++ | web/index.tmpl:
21:15 dalek modules.perl6.org: fixed width for badges column
21:15 dalek modules.perl6.org:
21:15 dalek modules.perl6.org: adding fixed width to badges column to avoid linebreaks between badges
21:15 dalek modules.perl6.org: review: https://github.com/perl6/modul​es.perl6.org/commit/6c2fce9ca5
21:24 fhelmberger joined #perl6
21:30 leont joined #perl6
21:32 dalek doc: f4353ee | Mouq++ | / (3 files):
21:32 dalek doc: Generalize definition detection
21:32 dalek doc:
21:32 dalek doc: Now allows more generalized forms. In the future, should also look for
21:32 dalek doc: X<infix,foo> or some such in the header.
21:32 dalek doc: review: https://github.com/perl6/doc/commit/f4353ee6e5
21:35 jnthn Mouq++ # working on doc \o/
21:36 jnthn 'night, #perl6
21:36 japhb o/ # night, jnthn
21:39 lizmat gnight jnthn
21:39 dalek rakudo/nom: bc6de57 | (Tobias Leich)++ | t/spectest.data:
21:39 dalek rakudo/nom: run S19-command-line/repl.t
21:39 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bc6de5743a
21:40 Mouq 'night jnthn! :)
21:40 FROGGS gnight all
21:40 leont Been mostly AFK this weekend. What was the way out of the panda issues mentioned on Friday? Upgrade panda or upgrade rakudo?
21:41 FROGGS leont: both
21:41 FROGGS &
21:41 timotimo .o( and don't forget to upgrade nqp, too )
21:42 PZt joined #perl6
21:55 pat_js joined #perl6
22:20 Mouq "Failed to open file: no such file or directory"
22:21 Mouq SUCH an annoying error that I've been getting recently. What did I fail to open?
22:24 lizmat Mouq: is that on HEAD ?
22:24 lizmat if so, how did you do that ?
22:26 Mouq Looks like I tried to spurt to "html/$/.html"
22:27 Hor|zon joined #perl6
22:28 dalek rakudo/nom: 4b78763 | (Elizabeth Mattijsen)++ | src/core/IO/Handle.pm:
22:28 dalek rakudo/nom: Cleanup Part #1 of IO::Handle
22:28 dalek rakudo/nom:
22:28 dalek rakudo/nom: - no longer does IO::FileTestable, has its own "handlers" now
22:28 dalek rakudo/nom: - added support for IO::Special paths
22:28 dalek rakudo/nom: - simplified a lot of unneccesary nqp::bindattr calls
22:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4b78763fb0
22:28 dalek rakudo/nom: 146c6f8 | (Elizabeth Mattijsen)++ | src/core/IO/Path/Special.pm:
22:28 dalek rakudo/nom: IO::Path::Special -> IO::Special
22:28 dalek rakudo/nom:
22:28 dalek rakudo/nom: No need to inherit from IO::Path, so this is a better name
22:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/146c6f8a44
22:28 dalek rakudo/nom: d748851 | (Elizabeth Mattijsen)++ | src/core/IO/Path.pm:
22:28 dalek rakudo/nom: Fix IO::Path.open
22:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d7488511ac
22:28 dalek rakudo/nom: 6e01187 | (Elizabeth Mattijsen)++ | src/core/io_operators.pm:
22:28 dalek rakudo/nom: Fix open() and initialization of $*IN|OUT|ERR
22:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6e01187f74
22:29 lizmat Mouq: earlier today there *was* a problem in HEAD for a while...
22:29 lizmat if I remember correctly...
22:29 Mouq lizmat: Ahh, you know what, I'm not on HEAD
22:29 lizmat ah... then all bets are off  :-)
22:29 Mouq :)
22:30 cosimo joined #perl6
22:31 dalek rakudo/nom: 8ca6460 | (Elizabeth Mattijsen)++ | / (5 files):
22:31 dalek rakudo/nom: Move IO::Special to the correct location
22:31 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/8ca6460e25
22:35 Mouq perl6 -e'spurt "html/$/.html", "uhhh"'
22:35 Mouq Failed to open file: no such file or directory
22:35 Mouq Which is LTA
22:35 lizmat yup
22:35 lizmat checking...
22:35 Mouq lizmat++ for all this IO work in any case :)
22:37 lizmat ah, does the "html" directory exist ?
22:38 Mouq lizmat: Oh, no, I did that in the Rakudo dir
22:39 carlin m: say "html/$/.html"
22:39 camelia rakudo-moar bc6de5: OUTPUT«html/.html␤»
22:40 Mouq carlin: Yeah..
22:40 lizmat carlin: still a valid file name
22:40 lizmat but if the "html" directory does not exist
22:40 Mouq lizmat: The issue is that $/ was interpolating where I meant literal <html/$/.html>
22:40 Mouq Either way, it's from the directory not existing
22:41 lizmat yup
22:41 lizmat LTA, but handling open() errors is already marked as "todo" in the code  :-)
22:42 Mouq lizmat: Okay :)
22:50 Mouq Of course, on HEAD doc/htmlify.p6 SEGFAULTS :(
22:51 * lizmat tries
22:51 dalek doc: bf302b8 | Mouq++ | htmlify.p6:
22:51 dalek doc: Fix for last commit; add syntactics to search
22:51 dalek doc: review: https://github.com/perl6/doc/commit/bf302b83da
22:51 dalek doc: fdf4522 | Mouq++ | / (2 files):
22:51 dalek doc: Fix issue with $/.html being attempted to be created
22:51 dalek doc: review: https://github.com/perl6/doc/commit/fdf45228dd
22:55 lizmat Mouq: I get: ===SORRY!=== Error while compiling ./htmlify.p6
22:55 lizmat Unsupported use of - as character range; in Perl 6 please use .. for range, for explicit - in character class, escape it or place it as the first or last thing
22:55 lizmat at ./htmlify.p6:538
22:56 Mouq O_o
22:57 tony-o who is carbin?
22:58 carlin tony-o: me, thanks for merging my pull req
22:59 lizmat Mouq: what VM are you on ?
23:00 Mouq lizmat: Moar, gimme a few minutes to recompile
23:01 lizmat I just found that I broke the parrot build
23:01 lizmat testing a fix now
23:02 lizmat .oO( don't you hate VM specific code in the setting )
23:09 lizmat hmmm..  doesn't fix it...
23:10 jack_rabbit joined #perl6
23:12 Mouq Oh, I'm dumb...
23:13 Mouq I need to rebootstrap Panda. I think it used to give a better error than just segfaulting, though
23:13 lizmat maybe a reconfig is in order as well
23:13 lizmat I added a file to the core
23:13 dalek rakudo/nom: 7c86547 | (Elizabeth Mattijsen)++ | src/core/IO/Handle.pm:
23:13 dalek rakudo/nom: Make sure standard handles have proper encoding
23:13 dalek rakudo/nom:
23:13 dalek rakudo/nom: And we only need one .open
23:13 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7c86547ef8
23:14 Mouq lizmat: yeah, `make install` complains and dies if you don't reconfigure after that
23:15 avuserow_ joined #perl6
23:15 lizmat if this doesn't fix the parrot build, I will have to look at it tomorrow
23:15 lizmat well, in ~ 8 hours from now
23:16 dalek doc: 8ea4be9 | Mouq++ | htmlify.p6:
23:16 dalek doc: Fix issues with .subst call
23:16 dalek doc: review: https://github.com/perl6/doc/commit/8ea4be93d0
23:18 Mouq Though, come to think of it I should probably just be using uri_escape
23:18 avuserow joined #perl6
23:18 lizmat building the restricted setting on parrot is broken atm
23:18 lizmat will look at it in ~ 8 hours
23:19 lizmat need some sleep now
23:19 lizmat good night, #perl6!
23:19 lizmat &
23:19 Mouq 'night lizmat++
23:25 BenGoldberg joined #perl6
23:31 cosimo joined #perl6
23:46 [Coke] timotimo: niecza also removed from daily runs
23:46 yoleaux 21:10Z <japhb> [Coke]: What do you think of http://irclog.perlgeek.de/​perl6/2014-10-05#i_9460362 through http://irclog.perlgeek.de/​perl6/2014-10-05#i_9460432 ?

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo