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.
Coke ugh, that sucks. 01:17
01:40 mtk joined 02:07 Khisanth joined 02:45 Jashwanth_ joined 03:01 benabik joined 03:04 mtk joined
dalek p: 82452e6 | jonathan++ | / (4 files):
Implement (?i) modifier.
04:28
06:38 he joined 07:27 brrt joined 07:34 Psyche^ joined 09:15 mvorl joined 10:48 Jashwanth__ joined 10:53 Jashwanth joined
dalek p: 94286df | jonathan++ | src/QRegex/P5Regex/Actions.nqp:
Perl 5 capture/alternation semantics.
12:17
kudo/nom: fb11f13 | jonathan++ | t (2 files):
Run S05-modifiers/perl5_3.t.

Also bumps NQP_REVISION to get (?i) support and alternation fix.
12:27
12:58 PacoAir joined
rurban morning #parrot 13:39
I think we have a minor problem
the vms port guy sent me a huge patch, which needs a parrot copyright
github.com/parrot/parrot/pull/854 is only the first small part of it 13:40
ah, mvorl is even in the channel, hi 13:41
mvorl Hi everyone!
@rurban: sorry for sending the patch. After all I found out how to work with git ;-) 13:42
rurban mvorl: no problem. butg i think you need to send a signed copyright form. I'm looking for it. the patch is too big
here it is: www.parrot.org/sites/www.parrot.org...ot_cla.pdf can you send such a signed CLA? 13:43
you can also fax it. 13:44
it can by found on www.parrot.org/foundation/legal
mvorl Yes I can. Will do. 13:45
Coke mvorl++ 13:46
rurban super.
mvorl Right now I'm trying to figure out a parrot crash that I /think/ is caused by va_list args passed around (nut sure yet) 13:48
rurban oh yeah. I know this. I started with a branch to fix it but my branch made it even worse, rurban/va_arg_print_n 13:49
mvorl Unfortunately it happens deep into winxed.pbc so it'll probably take some time to nail it down
@rurban: Forget about the big patch. I'll fax the CLA, and then submit the patches via git pull request 13:50
rurban I'm smoking your pull request right now 13:51
mvorl I haven't seen ptrdiff_t anywhere in parrot, but it's a C standard type 13:53
dalek rrot/vms-usize_t-gh854: fedbe5e | (Martin VorlƤnder)++ | src/hash.c:
Fix a bug that causes parrot to fail on platforms where size_t is an unsigned type.

Signed-off-by: Reini Urban <rurban@cpanel.net>
mvorl Anybody has some hints how to go about debugging parrot crashes? It happens because the data element of a PMC is 0 (and then dereferenced) 13:54
rurban do you have a gdb or such? 13:56
I like windows msvc most for debugging
Windows is also stricter than liunux and catches more bugs 13:57
mvorl :I have a VMS debugger (which is quite nice). Unfortunately, the last port of gcc to VMS is ages old, so no gdb 13:58
Coke ~[ ā—_ā— ]~ . o O (ooh, more unimotes) 14:00
mvorl I tried to --trace but my parrot assembly language knowledge is really non-existent
rurban If you get a crash in the debugger it's pretty easy, because you just need to run it, and step upwards
--trace is a bit noisy. normally c-level debugging is easier 14:01
14:01 benabik joined
rurban but if you want to see the valuus of all internal parrot registers (SREG[]) you can add -t 14:01
what is your floatval size? 14:02
single floats are hairy with varargs
mvorl That I can do. The PMC is created in extend:Parrot_ext_call by Parrot_pcc_build_call_from_varargs
rurban long double ditto 14:03
mvorl nv=double, numvalsize=8, doublesize=8, longdoublesize=16
rurban double=8 is good.
mvorl It's just a "Pi->S" AFAICT 14:04
rurban Pi should be easy, hmm 14:05
And the crash happens in VTABLE_push_pmc(interp, signature, pmc_arg) then? 14:06
mvorl Nope. It's in Parrot_pcc_set_continuation(interp, call_object, ret_cont) 14:08
when it tries to CONTEXT_STRUCT(ctx)->current_cont = _continuation; 14:10
rurban ctx is empty?
mvorl (the above in src/call/pcc.c and src/call/context_accessor.c, respectively) 14:11
rurban in which test is this? 14:12
mvorl CONTEXT_STRUCT(ctx) == ctx->data == 0 14:13
It's in the building process still
rurban ah the ctx->data only. 14:14
so you got a miniparrot already? or before
mvorl I've got a parrot.exe already; running winded.pbc is it's first job 14:15
s/winded/winxed/
rurban here? $(PARROT) winxed.pbc -c --noan $(DEV_TOOLS_DIR)/pbc_to_exe.winxed 14:17
mvorl Yep. That's the one
rurban Did you change prt0.winxed? 14:19
mvorl Nope. (I wouldn't change code I don't understand ;-)
dalek p: 2d70bfa | jonathan++ | src/QRegex/P5Regex/ (2 files):
Lookahead, positive and negative.
14:22
rurban maybe run the winxex pir instead. parrot -t ext/winxed/driver.pir -c --noan tools/dev/pbc_to_exe.winxed
mvorl That one runs right through. 14:26
rurban that's ia64 little-endian, right? 14:27
is tools/dev/pbc_to_exe.pir created correctly?
mvorl Yes. (Actually IA64 is bi-endian, but VMS uses it little-endian) 14:28
Yes, tools/dev/pbc_to_exe.pir is there. Correctly? Dunno.
rurban you can look at it if looks like pir code 14:29
mvorl Looks okay to me.
rurban strange, since the frontend/parrot2/prt0.pbc worked correctly 14:30
that was also from a .winxed
ptrdiff_t looks good on cygwin, linux, msvc, solaris (native cc) 14:31
14:31 bluescreen joined
dalek Heuristic branch merge: pushed 25 commits to parrot by rurban 14:32
rurban merging now two smoked branches 14:33
dalek rrot: fedbe5e | (Martin VorlƤnder)++ | src/hash.c:
Fix a bug that causes parrot to fail on platforms where size_t is an unsigned type.

Signed-off-by: Reini Urban <rurban@cpanel.net>
rrot: 174ce54 | rurban++ | src/hash.c:
Merge branch 'vms-usize_t-gh854'
mvorl Got to go (my girlfriend isn't a programmer :-) Thanks a lot, Reini. I'll dig into it. 14:34
rurban interesting bug, bye
dalek rrot: 29ec1b0 | rurban++ | ChangeLog:
[GH #854] ChangeLog line
14:35
14:37 mvorl left
ttbot Parrot 87ac5ceb MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/99176 14:52
14:58 dmalcolm joined
rurban bacek: around? 15:08
msg bacek I revived your opsc_llvm branch in my rurban/ops2c_llvm branch. what was the problem? 15:09
aloha OK. I'll deliver the message.
rurban Oh, solaris found a mime_base64 regression... 15:10
Coke wonders if we can get all our coding bots back at once.
rurban lobsinger also? 15:12
and cotto
15:13 brrt joined
brrt hi #parrot 15:14
brrt has shirt!
rurban you are not naked?
brrt and a certificate, all messed up by customs people 15:15
no, i mean, a gsoc shirt :-)
rurban ah! a google shirt. I also got one, looks good
brrt yes, it is a pretty shirt
i also figured out a way for the parrot foundation to finally make some money 15:16
rurban uh?
brrt just say you are an alternative implementation of ruby
or, better yet
rurban and then ?
brrt a 'scalable' implementation of ruby 15:17
money flows in
profit
'it almost runs rails'
you might attempt the same with node.js-on-parrot, but i don't think people will buy into it 15:18
ttbot Parrot 29ec1b05 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/99256
dalek rrot: 48a9db2 | rurban++ | t/library/mime_base64.t:
[GH #855] utf8 comparison problem
15:20
brrt but in seriousness, a bit better marketing might help a lot 15:25
15:46 zby_home joined 15:57 travis-ci joined
travis-ci [travis-ci] parrot/parrot#620 (master - 48a9db2 : Reini Urban): The build was fixed. 15:57
[travis-ci] Change view : github.com/parrot/parrot/compare/2...a9db235434
[travis-ci] Build details : travis-ci.org/parrot/parrot/builds/2651726
15:57 travis-ci left
ttbot Parrot 48a9db23 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/99404 16:02
16:24 contingencyplan joined 16:55 tuxit joined 17:14 sivoais joined 17:51 jlaire joined 18:00 benabik_ joined
dalek rrot/rurban/llvm-gh853: c4ac9e8 | rurban++ | config/auto/llvm.pm:
[GH #853] Less failing llvm-gcc noise
18:53
rrot: 8632cf0 | rurban++ | config/auto/gcc.pm:
[cage] silence @noinline@ warning on unknown compilers

e.g. solaris cc
19:11
19:56 benabik_ joined 22:26 whiteknight joined
whiteknight good evening, #parrot 22:48
23:00 Psyche^ joined
sorear good evening whiteknight. 23:01
23:02 benabik joined
whiteknight hello sorear 23:02
23:25 mvorl_ joined 23:27 mvorl__ joined