Parrot 6.1.0 "Black-collared Lovebird" | parrot.org/ | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 18 February 2014.
00:23 drift joined 00:45 kid51 joined 00:47 drift joined 01:05 bluescreen_ joined 01:12 kid51_ joined 06:13 Chirag joined 07:42 FROGGS joined 07:45 he joined 10:19 dngor joined 11:12 denis_boyun joined 12:32 bluescreen joined 12:37 bluescreen_ joined 12:40 bluescreen__ joined 13:18 Chirag joined 14:28 bluescreen joined 16:34 FROGGS joined 17:06 Chirag joined 17:10 Chirag_ joined 18:25 Chirag_ joined 18:28 Chirag joined
dalek rrot/rurban/pasm_out: 20e816d | rurban++ | / (27 files):
re-add PASM output to imcc, re-add imcpasm tests

one new IMCCompiler api method: set_to_pasm remove static output in imcc/instructions.c change Parrot::Test pir_2_pasm mthod to use -d1000, which is equivalent to pasm output (just to stderr) detect .pasm -o outfile extension add __builtin_expect LIKELY/UNLIKELY macros (WIP)		 18:40
18:43 rurban1 joined 19:13 denis_boyun joined
Util Pre-report: 19:29
# Done:
* Bought ticket for YAPC::NA::2014; plan to attend the pre-YAPC hackathon, too.
* (Nill else; $WORK)
# Doing: Giving Perl (5&6) talk at Atlanta.pm on Thursday. 19:30
.end
Hello! Who all is in attendance? 19:31
rurban Done: Re-enable old imcc flags for parrot and parrot_old, such as -On -a -v -y -E -dxx. [GH #1033]
Fixed imcc -O1 and -O2.
Wrote a blog post about it
Merged #1045 19:32
Work: rurban/pasm_out, valgrind leaks, smolder exitcode 19:33
helping GSOC with the method perf task (looks good)
Todo: check the old jit (it *is* scalable if you know how), mentor GSOC, more zefram tickets 19:34
.end
Util rurban++ 19:35
rurban forgot: I'll add __builtin_expect LIKELY/UNLIKELY macros
like in the linux kernel or perl5 19:36
and more benchmarking and profiling, esp. with rakudo
Util sounds good 19:37
rurban yes, I'm quite pleased with myself :) 19:38
Util That was impressive throughput!
rurban pasm_out is still a bit hairy, but it gets better. I started with mimicing -d1000 outputting proper pasm. this was okay. but the api restrictions are horrible 19:39
This is the GSOC overview for our student: wiki.enlightenedperl.org/gsoc2014/i...signatures 19:40
I still need to find the pcc revamp branch (=> nci) and tell him what to keep and what to avoid. 19:42
Util What branch is pasm_out in?
Coke rurban/pasm_out ?
rurban But then I'm tempted to write it by myself in 2 days and have the result earlier... 19:43
Util Doh! Thanks
(I had though that you were referring to a pasm_out function, which I of course did not see in the master branch)
rurban But then our student will not get his 5.000 and not get into the project
Util *and* we would lose the chance that the student would become a long-term contributor, as some past students have done! 19:44
rurban the func will be e_pasm_out() 19:45
Coke: yes, but still WIP. I broke O2
I don't like the names yet. set_to_pasm => set_emit_pasm maybe
or set_write_pasm 19:46
the ticket for ruirban/pasm_out is #989 19:47
19:55 Hunger joined
dalek rrot/rurban/pasm_out: c642a1d | rurban++ | / (8 files):
enable pasm_out in parrot2, todo failing imcpasm tests

opt2 fails with changed used_once semantics: I0 = 10 is side-effecting, but *can* be removed if I0 is used only once (GH #1036), ...		 20:05
Util hmmm. I just noticed that I started all of this in #parrot instead of #parrotsketch. Whoops! I will post a line to that effect in #ps. 20:09
rurban yes, noticed it also 20:10
Util Any other business before we adjourn?
rurban not for me
Util OK. Thanks to all! 20:11
Meeting adjourned.
20:16 bluescreen_ joined 20:20 Chirag joined 20:31 Chirag_ joined
dalek rrot/rurban/pasm_out: 960c532 | rurban++ | lib/Parrot/Test.pm:
[test] ignore TEST_PROG_ARGS -O2 for opt1.t tests
 20:54
rrot/rurban/pasm_out: a620924 | rurban++ | compilers/imcc/debug.c:
[cage] overlong line
rurban what would be a proper vim code for test_c.in files? "vim: expandtab shiftwidth=4 cinoptions='\\:2=2' :" misses the c mode 21:57
vim coda
ff=c maybe? 22:00
22:06 bluescreen joined
dalek rrot/rurban/expect: ea071dd | rurban++ | / (8 files):
[config] support __builtin_expect LIKELY/UNLIKELY branch optimizations

add EXPECT,LIKELY,UNLIKELY macros to feature.h as in the linux kernel and perl5 probe for __builtin_expect() support in a new auto::expect configure step
Usage: if (UNLIKELY(var == 0)) fail();
Add emacs mode:c support for test_c.in files, and relax the coda test for this. Should be used in all test_c.in files.		 22:06
rurban ft=c should be enough as I'm reading 22:12
dalek rrot/rurban/pasm_out: 98c4eaf | rurban++ | / (26 files):
re-add PASM output to imcc, re-add imcpasm tests

one new IMCCompiler api method: set_to_pasm remove static output in imcc/instructions.c detect .pasm -o outfile extension add and fix old t/compilers/imcc/imcpasm/*.t tests and the pir_2*_
   Parrot::Test methods
change Parrot::Test pir_2_pasm method to use -d1000, which is equivalent
   to pasm output, just to stderr.
use new LIKELY/UNLIKELY macros		 22:59
rrot/rurban/pasm_out: 6378792 | rurban++ | / (8 files):
enable pasm_out in parrot2, todo failing imcpasm tests

opt2 fails with changed used_once semantics: I0 = 10 is side-effecting, but *can* be removed if I0 is used only once (GH #1036), ...
rrot/rurban/pasm_out: 5ab56c7 | rurban++ | lib/Parrot/Test.pm:
[test] ignore TEST_PROG_ARGS -O2 for opt1.t tests
rrot/rurban/pasm_out: c0c7c5c | rurban++ | compilers/imcc/debug.c:
[cage] overlong line
rrot/rurban/pasm_out: 92aea5e | rurban++ | / (2 files):
[cage] fix MANIFEST.SKIP, .gitignore
rrot/rurban/pasm_out: f4e00a5 | rurban++ | compilers/imcc/ (3 files):
[codingstd] add docs, ...
23:32 rurban1 joined 23:33 rurban2 joined
dalek rrot/rurban/pasm_out: e846a2c | rurban++ | / (2 files):
[cage] imcc/embed.h

error: nonnull argument references non-pointer operand with llvm		 23:36

Powered by Cro and the Raku® Programming Language.

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

Thank you!