Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2013-05-04

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | full mode

All times shown according to UTC.

Time Nick Message
06:30 raiph http://blogs.perl.org/users/perl_6_reports/2013/05/perl6-summary-for-april-25th---may-1st.html
06:52 moritz raiph: note that irclog.perlgeek.de is up again
07:37 dalek roast: bc4c9a5 | (David Warring)++ | S05-capture/caps.t:
07:37 dalek roast: tests separator captures RT #117831
09:10 dalek panda/panda2: 0bbb63b | tadzik++ | lib/Panda (2 files):
09:10 dalek panda/panda2: Make destdir easier to override
09:10 dalek panda/panda2: Bugfix, and make more real-world proof
09:23 labster What should I call the File::Spec module, now in the core?  I think it should be under IO, but it's really a path specification or file specification.  So, IO::Spec, IO::FileSpec, or IO::PathSpec?  (IO::Path::Spec would get confusing due to looking like IO::Path::Unix).
09:24 labster Second, should I change the names of the OS submodules?  Perl 5 uses ::Unix and ::Win32, but S16 qp// uses :posix and :win.  For comparison, python uses os.posixpath and os.ntpath.
09:24 moritz how about IO::Path::Spec and then IO::Path::Spec::Unix?
09:27 * lizmat was thinking more or less the same thing
09:28 labster As long as IO::Path::Spec::Unix and IO::Path::Unix aren't too confusing, we're ok.
09:32 labster then again, we don't want anyone really using the backend, so maybe that's okay.
09:34 * grondilu was reading http://rosettacode.org/wiki/Anonymous_recursion#PicoLisp and thought that "recurse" would be a nice keyword instead of &?ROUTINE or &?BLOCK.
09:51 dalek roast: 4ee1dab | (Brent Laabs)++ | S32-io/io-path.t:
09:51 dalek roast: fudge some IO::Path tests in preparation of windows support
17:25 dalek roast: 33746c6 | larry++ | S05-modifier/ignorecase.t:
17:25 dalek roast: [M] is not <[M]>
17:41 dalek std: 291c7f0 | larry++ | CORE.setting:
17:41 dalek std: add some missing function declarations
17:42 dalek roast: use '=== Nil', not '~~ Nil'
17:58 dalek roast: avoid ~~ Nil
18:12 dalek nqp: e3a67e4 | (Alessandro Ghedini)++ | 3rdparty/dyncall/dyncallback/dyncall_ (2 files):
18:12 dalek nqp: Fix build failure on PowerPC (closes #62)
18:12 dalek nqp: Fix build failure on SPARC
18:34 dalek roast: 220bf16 | larry++ | S05-capture/named.t:
18:34 dalek roast: STD warning suppression
18:34 dalek roast: class Foo does Role{} needs ws before {}
18:34 dalek roast: don't smartmatch against Nil
18:34 dalek roast: |$capture --> |capture
18:34 dalek roast: capitalize -> wordcase
18:34 dalek roast: $.foo form not allowed in BUILD
18:37 dalek std: b33d8e0 | larry++ | CORE.setting:
18:37 dalek std: more missing names
18:43 dalek nqp: 4934e64 | (Stéphane Payrard)++ | / (2 files):
18:43 dalek nqp: support parameters with dash
18:53 dalek nqp: a24051c | (Stéphane Payrard)++ | src/QAST/Operations.nqp:
18:53 dalek nqp: Remove stray blanks
18:53 dalek nqp: Rolled QAST::Operations.add_hll_unbox calls into loop
18:55 moritz question for the OS X users: is https://github.com/drKreso/nqp/commit/86a3b20d72aeaa2acf92fa0119b92d3edc4348a5 (the assumption that $^X eq 'darwin' implies 64bit) good?
18:59 geekosaur no
19:03 geekosaur anyway, the assumption is, as the comment says, imperfect but largely true for 10.6+ and most machines using Apple's perl5 (the earliest Apple Intel hardware did not support 64bit extensions)
19:05 geekosaur so anyway, that assumption is false on: PPC, 10.5.x, early Core1-based Intel hardware. do we care?
19:07 geekosaur and may be false on 32-bit builds of perl from other sources, although in many cases those will only be 32 bit in the case where 64 bit won't run anyway
19:08 moritz geekosaur: can you think of a better heuristic?
19:11 geekosaur not really, see 'absurdly difficult
19:13 geekosaur and just to make things more annoying, there is no difference between perl -V output on this machine with `arch -32` vs. `arch -64`
19:14 geekosaur apple went to a lot of trouble to hide as much as possible the difference between 32 and 64 bit on OS X
19:21 * moritz will just merge that pull request, and wait for complaints.
19:21 labster *sigh*.  go for it, moritz.
19:21 geekosaur yes, this is moderately insane
19:22 dalek nqp: 86a3b20 | drKreso++ | Configure.pl:
19:22 dalek nqp: OSX 64 bit target for NativeCall
19:24 dalek nqp: cae6f61 | (Gerhard R)++ | 3rdparty/dyncall/buildsys/gmake/tool/gcc.gmake:
19:24 dalek nqp: [dyncall] next shot at fixing the build on Strawberry Perl
19:31 * geekosaur pulls and builds on his OS X desktop
19:34 geekosaur although I don't think I've done any NativeCall stuff before, so not sure how I'd test any of this
19:34 geekosaur (well, ideally, it's already got tests and I can just run them)
19:37 dalek rakudo/nom: 5b53365 | (Brent Laabs)++ | / (7 files):
19:37 dalek rakudo/nom: added IO::Spec, started work on IO::Path
19:37 dalek rakudo/nom: add support for IO::Path::<os subclasses>
19:37 dalek rakudo/nom: rename `file-name-is-absolute` to `is-absolute` in IO::Spec
19:37 dalek rakudo/nom: IO::Spec.tmpdir now returns `.` on failure, trying $*TMPDIR out
19:37 dalek rakudo/nom: move IO to later in the build: lib.pm now works, regex works
19:37 dalek rakudo/nom: cleanup tmpdir, remove unused and already commented out code
19:37 dalek rakudo/nom: fix dir(), panda now runs \o/.  Also change splitpath to use :nofile param
19:37 dalek perl6-examples: 95cc309 | (David Warring)++ | parsers/CSSGrammar.pm:
19:37 dalek perl6-examples: removed null operator fudge
19:42 dalek rakudo/nom: 454c190 | moritz++ | docs/ChangeLog:
19:42 dalek rakudo/nom: update ChangeLog
19:47 dalek perl6-examples: b71f739 | (David Warring)++ | parsers/CSSGrammar.pm:
19:47 dalek perl6-examples: removed null combinator fudge
20:00 moritz labster++ # great work in rakudo
20:14 dalek rakudo/nom: 5ad46fd | moritz++ | src/Perl6/Grammar.nqp:
20:14 dalek rakudo/nom: start to switch **0..1 back to ?
20:38 timotimo in my EXPORT sub I can't access the "create_adt" symbol. it's defined as "our sub create_adt(...) is export"
20:41 timotimo https://github.com/timo/ADT/tree/EXPORT_sub - this branch here
20:47 lizmat r: module M { our sub a is export { "foo" } }; import M; say a
20:47 camelia rakudo 454c19: OUTPUT«foo␤»
20:48 lizmat so, the import allows you to not have to specify fully qualified names
20:48 FROGGS but then I wonder what is wrong with timotimo's code...
21:47 * moritz couldn't sleep, so he implemented 'recurse': https://github.com/perl6/specs/issues/37#issuecomment-17442034
22:15 * moritz found https://lwn.net/SubscriberLink/548560/26d15e832d21a483/ nice
23:55 labster r: say IO::Spec.splitpath("dir", :nofile).perl;  my $a = ("file", :nofile).item;  say IO::Spec.splitpath( |$a ).perl;
23:55 camelia rakudo 5ad46f: OUTPUT«("", "dir", "")␤Too many positional parameters passed; got 3 but expected 2␤  in method splitpath at src/gen/CORE.setting:10471␤  in method splitpath at src/gen/CORE.setting:10897␤  in block  at /tmp/qvdFInYXtl:1␤␤»
23:56 labster okay, how am I using method(|$a) wrong here?

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | full mode

Perl 6 | Reference Documentation | Rakudo