Parrot 4.8.0 "Spix's Macaw" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 20 September 2012.
whiteknight aha! 00:29
Creating a new thread proxies interp->HLL_info, but Parrot_hll_get_HLL_type tries to access the internals of that PMC directly 00:30
We don't want to proxy it, so instead we probably want to create a local clone of that array
...but fixing that seems to expose a different problem, a segfault somewhere in NQP's lexpad code 00:43
dalek rrot/threads: c73ae65 | Whiteknight++ | src/pmc/parrotinterpreter.pmc:
In clone_interpreter, fix an issue where HLL_info was being proxied, but Parrot_hll_get_HLL_type expects a FixedIntegerArray that can be poked into directly. Leave a note about this change. This fixes one issue in NQP, but exposes another segfault later.
00:47
Coke anyone else notice that chromatic cancelled parrotsketch on the google calendar? (for the best, I think, since no one but rurban is attending) 00:51
whiteknight I did not notice
I haven't been attending lately, so my opinions on the matter are moot 00:52
Okay, this next nqp issue appears to be from NQPLexPad poking directly into a proxied lexinfo pmc 01:01
Coke ah, good catch 01:08
whiteknight so much code was written as if Parrot did not have threads 01:09
which wasn't wrong per se, just not very forward-thinking
fixing that proxy error, tadzik's example program no longer segfaults. Instead it exits with error code 1 01:13
We'll pretend that's progress
Coke I just did a fresh build on an old OSX box, and it was very confused by the already installed old parrot. 01:16
Coke blows that away and tries again. :| 01:17
whiteknight I'm out of time for the night. I'll stare at this again tomorrow
goodnight, Coke
Coke good white knight. 01:19
Coke just gives up and moves back to the much faster laptop. 01:20
dalek rtcl-nqp/nqp2: ea6caae | coke++ | src/Partcl.pm:
Fix command line parsing.

now "./partcl foo.tcl" works again
01:43
03:14 travis-ci joined
travis-ci [travis-ci] parrot/parrot#625 (threads - c73ae65 : Whiteknight): The build was broken. 03:14
[travis-ci] Change view : github.com/parrot/parrot/compare/b...3ae654132a
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/2692564
03:14 travis-ci left 03:20 benabik joined 07:21 Mike-PerlRecruiter_ joined
dalek kudo/nom: e52612f | jnthn++ | src/core/operators.pm:
Fix an exception throw.
07:24
p: f44a799 | jnthn++ | src/QRegex/P6Regex/Grammar.nqp:
Toss :PIR{{...}} parsing in P6Regex.

Legacy, and didn't even have an action method so couldn't possibly have been being used.
08:48
p: e349cb6 | jnthn++ | src/ (4 files):
Basic implementation of :dba(...).
kudo/nom: 59fc70a | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for :dba(...).
kudo/nom: 88a9ba4 | jnthn++ | src/Perl6/Grammar.pm:
Uncomment and add some :dba(...).
p: 30b8883 | moritz++ | src/ops/nqp_bigint.ops:
[bigint ops] s:g/in PMC/invar PMC/
09:17
p: 47593ca | jnthn++ | src/QRegex/P5Regex/Actions.nqp:
Respect (?i) when compiling charclasses.
09:22
kudo/nom: 46b8eb8 | moritz++ | src/Perl6/Actions.pm:
sprinkle some native types into Actions.pm
10:02
p: 8b5a7a0 | jnthn++ | src/QAST/Compiler.nqp:
Fix subcapture compilation.

We have to ensure $!pos gets updated appropriately, since it is used by
  !cursor_capture. Fixes much in m:P5/.../ spectests, but could also be a
bug in normal Perl 6 regexes too.
10:07
p: a1c54d8 | jnthn++ | t/p5regex/rx_captures:
Untodo a P5Regex test.
kudo/nom: afb8a63 | jnthn++ | tools/build/NQP_REVISION:
Bump NQP_REVISION for regex improvements.
10:18
kudo/nom: e644792 | jnthn++ | t/spectest.data:
Run S05-modifier/perl5_4.t.
kudo/nom: d3de02f | moritz++ | src/Perl6/Actions.pm:
more native types for Actions.pm
10:32
11:27 whiteknight joined 12:10 PacoAir joined
whiteknight good morning, #parrot 12:39
dalek Heuristic branch merge: pushed 30 commits to nqp/kill-nqpattr by jnthn 14:00
14:15 PacoAir_ joined
dalek p/kill-nqpattr: 8a899d4 | jnthn++ | src/NQP/Actions.pm:
Simplify some code.
14:48
p/kill-nqpattr: 7db86ab | jnthn++ | src/core/NQPRoutine.pm:
Flesh out NQPRegex a bit.

Gives it methods and attributes that will store the captures and NFAs.
p/kill-nqpattr: e212a96 | jnthn++ | src/QRegex/NFA.nqp:
Add a way to persist an NFA to a data structure.

This will (eventually) be used to serialize the NFA instead of using the nqpattr hack.
14:52 Psyche^ joined 15:41 benabik joined
benabik ~~ 15:45
whiteknight hello benabik 16:25
benabik o/ whiteknight
whiteknight how are you doing today? Is it raining out your way? 16:26
benabik Not raining at the moment, but there's some in the forecast.
I'm watching a clever program with some amusement. One of my classes assigned cracking a password list for homework. I wrote a program that's checked 38 trillion passwords so far. 16:27
16:56 brrt joined
brrt hi #parrot 16:56
whiteknight hello brrt 17:26
brrt hi whiteknight 17:27
good that you are here, I'm busy designing the calling conventions for parakeet :-) and i was hoping you could help a bit 17:28
in short, should stacks have a root stack? 17:30
17:47 brrt1 joined
dalek p: 9e503bd | jnthn++ | src/QRegex/P5Regex/ (2 files):
Toss some P6Regex code that's ported.
18:28
p: 5591588 | jnthn++ | src/QRegex/P5Regex/Actions.nqp:
Handle quantification of vanishing things.
p: 3442c18 | jnthn++ | src/QRegex/P5Regex/ (2 files):
Implement lookbehind.
p: b8ad051 | jnthn++ | src/QRegex/P5Regex/Grammar.nqp:
Parse 's' mod (though NYI).
whiteknight brrt: What do you mean, "root stack"? 18:35
18:40 brrt joined 18:50 contingencyplan joined 19:21 Mike-PerlRecruiter_ joined
dalek p: f982b88 | jnthn++ | src/QRegex/P5Regex/ (2 files):
Revise whitespace and comment handling.

This makes whitespace significant by default, requiring (?x) to make it insignificant again, as is the Perl 5 default. Also implements the (?#xxx) comment syntax.
19:21
p: 08e45ac | jnthn++ | t/p5regex/rx_metachars:
Update tests for whitespace semantics.
p: 28bf11c | jnthn++ | src/QRegex/P5Regex/Actions.nqp:
Implemnt (?m).
19:22
p: cd58c13 | jnthn++ | src/QRegex/P5Regex/Actions.nqp:
Give . Perl 5 semantics; honor (?s).
p: 5238d7b | jnthn++ | t/p5regex/rx_metachars:
Correct a test for Perl 5 semantics.
kudo/nom: cc92def | jnthn++ | tools/build/NQP_REVISION:
Get latest P5Regex improvements.
19:26
kudo/nom: 870d189 | jnthn++ | t/spectest.data:
Run S05-modifier/perl5_5.t.
19:38 PacoAir joined 19:40 PacoAir joined 19:51 PacoAir joined 20:01 PacoAir joined 20:21 benabik joined 20:30 MikeFair joined 20:56 perlite_ joined 21:39 rurban_home joined
rurban_home I could reproduce kid51's darwin thread problems with embed. wrong linker flags 23:31