github.com/moarvm/moarvm | IRC logs at irclog.perlgeek.de/moarvm/today
Set by moderator on 3 August 2013.
00:04 ggoebel joined 00:05 lizmat joined 00:06 crab2313 joined 00:18 woolfy joined 00:19 FROGGS joined 00:40 FROGGS joined 00:48 FROGGS joined 01:19 FROGGS joined 01:43 benabik joined 01:56 colomon joined 02:10 FROGGS joined 02:13 benabik joined 02:33 flussence joined
dalek arVM/libuv2: 0febb98 | jimmy++ | 3rdparty/libuv/src/ (2 files):
small fixes for libuv
 03:00
03:05 FROGGS joined
[Coke] ~. 03:07
dalek arVM/libvu3: 4acafef | jimmy++ | 3rdparty/libuv/include/uv- (2 files):
added uv_thread_yield
 03:32
arVM/libvu3: 258c470 | jimmy++ | 3rdparty/libuv/include/uv- (2 files):
added uv_thread_exit
arVM/libvu3: 56cd273 | jimmy++ | 3rdparty/libuv/ (3 files):
added uv_fs_flush function
arVM/libvu3: 1091745 | jimmy++ | 3rdparty/libuv/ (5 files):
added uv_fs_getstdin/uv_fs_getstdout/uv_fs_getstderr
arVM/libvu3: a35c3ae | jimmy++ | 3rdparty/libuv/src/win/fs.c:
small fixes for uv_fs_mkdir_p
 03:38
03:40 crab2313 joined 03:51 FROGGS joined 04:28 FROGGS joined 04:43 birdwindupbird joined 05:09 FROGGS joined
dalek arVM/libvu3: 3204eaa | jimmy++ | / (9 files):
removed some needless ops
 05:20
05:26 crab2313 joined
dalek arVM/libvu3: 829a60c | jimmy++ | / (7 files):
removed another needless ops
 05:27
JimmyZ segomos: ping? 05:34
05:45 FROGGS joined 06:40 FROGGS[mobile] left 06:42 woolfy left 06:46 FROGGS joined
FROGGS o/ 06:47
JimmyZ FROGGS: good morning 06:49
FROGGS: we can do some LHF apr => libuv now 06:50
FROGGS: see gist.github.com/zhuomingliang/6077134 update 06:51
FROGGS JimmyZ: k, sounds nice 06:54
JimmyZ ;) 06:55
the main problem is about uv loop 07:04
I think uv_default_loop should be in instance, other loops are per thread in tc 07:05
and about req ... I don't know yet 07:06
FROGGS I'm currently fighting some nqp<=>rakudo module loading problem... 07:29
08:15 not_gerd joined
not_gerd o/ 08:15
FROGGS hi not_gerd 08:17
not_gerd JimmyZ: I've got a branch that builds linenoise and dyncall on windows
JimmyZ: I'd like to add libuv to that as well
JimmyZ: I'm asuming libvu3 is the branch to use? 08:18
JimmyZ not_gerd: nice, and yes 08:22
jnthn: ping 08:23
not_gerd JimmyZ: merge conflict - where did the hostname op fo? 08:26
*go?
JimmyZ removed 08:27
08:35 crab2313 joined
not_gerd why is bank 6 missing an opcode 0? 08:37
JimmyZ not_gerd: cd nqp-cc && make 08:41
not_gerd JimmyZ: see src/core/oplist 08:42
JimmyZ not_gerd: no, it's my fault
not_gerd ;)
perhaps we should use lua - they start indexing from 1
JimmyZ not_gerd: we can update update_ops.p6 to auto ++ 08:43
and with it, we don't need do 0x01 ... manually 08:45
not_gerd well, keeping the opcodes explicitly might make a good reference documentation
but the tools should check consistency
JimmyZ oh, yeah
you're right 08:46
not_gerd update_ops.p6 should complain and tell the programmer to fix his mess...
JimmyZ +1 08:47
not_gerd github.com/gerdr/MoarVM/commits/ge...ntegration # my pull requests + dyncall + linenoise + libuv 08:58
currently only works on windows
JimmyZ not_gerd: how do you support libuv on *nix? not all src/unix/* need for linux 09:04
not_gerd JimmyZ: I'll probably have to keep a list of the platform-specific files in unix/ 09:07
JimmyZ not_gerd: Did you see content in uv.gyp? that file contains some good logic for integration 09:09
not_gerd JimmyZ: yup, I'll probably have to re-produce part of that logic in perl 09:12
however, HAVE_CONFIG_H might be a problem
ie we might need to shell out to the configuration script...
JimmyZ yeah... 09:13
some configure detection, ie HAS_REALINE 09:14
MVM_HAS_READLINE
and with it, we don't need compile linenoise ...
can stell same ideas from parrot though 09:15
steal some ...
not_gerd we should probably also maintain a proper fork of libuv as well... 09:18
JimmyZ maybe 09:19
09:21 cognominal joined 09:23 lizmat joined
nwc10 diakopter/jnthn: Each iTLB miss causes the processor to do a walk over the page table (called a "page walk"). Page walks can be particularly compute-intensive for large applications (like Facebook's site) where the page table is so big that it exceeds the capacity of the last-level cache (LLC). 09:24
from www.facebook.com/notes/facebook-en...0460698920
JimmyZ jnthn: I want to merge github.com/MoarVM/MoarVM/pull/45 and linenoise branch... 09:39
jnthn +1 to 45, provided it's building/passing tets fine 09:41
Not so comfortable with the linenoise thing, so that stays in a branch for now
JimmyZ ok
09:45 patspam joined 09:50 TimToady joined
dalek arVM: 27aa573 | (Gerhard R)++ | / (60 files):
keep forward declarations for all public structs and unions in src/types.h
 09:55
09:56 patspam joined
dalek arVM/libvu3: 3704e9b | (Gerhard R)++ | / (60 files):
keep forward declarations for all public structs and unions in src/types.h
 09:59
10:04 patspam joined
not_gerd bye, #moarvm 10:39
10:39 not_gerd left
JimmyZ hmm, why threads fails on windows with "can't GC Stables", but passes on linux 11:07
*threads.t
11:30 lizmat joined
JimmyZ jnthn: ping 11:32
FROGGS JimmyZ: it failed for me on linux from time to time too
JimmyZ oh
FROGGS at least till one week ago or so 11:33
my moarvm is not up-to-date
11:40 lizmat joined 11:51 colomon_ joined 12:00 BabsSeed joined 12:10 lizmat joined
JimmyZ BabsSeed: welcome 12:13
BabsSeed JimmyZ: Thanks 12:36
13:30 lizmat joined
BabsSeed JimmyZ: Wanting to get involved with development. 13:34
JimmyZ BabsSeed: good 14:07
BabsSeed JimmyZ: I'm just a little unsure as to what I should be doing, is there a 'to-do' list or anything? Are tasks allocated to contributors? 14:10
(I've only worked in closed source corporate environments or on 1-2 contributor open-source projects, never on a major free software project)
14:11 timotimo joined
BabsSeed Well, I've written patches for free software, but never major features or contributions 14:11
FROGGS atm there is some work done on the libraries moarvm uses, this is for example the transition from libapr to libuv 14:12
and moarvm ismost capable of selfhost NQP, diakopter++ is working on that 14:13
so there are maybe some LHF (low hanging fruits) for the library-work
JimmyZ++ and not_gerd++ are working on that mainly 14:14
JimmyZ I want to know why "WideCharToMultiByte( CP_UTF8 ..." outputs messy chars ...
FROGGS JimmyZ: output to what? terminal? 14:17
JimmyZ FROGGS: any file ...
BabsSeed Right now I'm finishing up a few projects, hoping to get properly involved later this month but will get familiar with the code before that. 14:19
JimmyZ WideCharToMultiByte( CP_ACP .. works well
but I want UTF8
BabsSeed I assume I need a github account :D 14:20
FROGGS BabsSeed: yeah, you need one :o)
JimmyZ: I cant find that WideCharToMultiByte, is that in a branch only? 14:21
JimmyZ FROGGS: in my local ..
FROGGS hold on, I'm coming :P
JimmyZ basically 'm doing GBK to Unicode by MultiByteToWideChar, and then Unicode to UT8 by WideCharToMultiByte 14:23
FROGGS ahh, this is a windows function?
JimmyZ yeah, 14:24
BabsSeed FROGGS: Got one part down then - github.com/BabsSeed
JimmyZ wchar is annoying
jnthn: I really want to import github.com/moriyoshi/libmbfl :(
libmbfl is the lib by using in PHP's mb_string 14:25
and ruby and sublime 14:26
I want to do it, because getenvhask is getting my GBK env chars 14:27
*getenhash
14:32 crab2313 joined 14:34 lizmat joined 15:15 colomon joined
dalek arVM: e4a10b0 | jimmy++ | src/io/procops.c:
fixed encoding issue in MVM_proc_getenvhash function, mostly appears in CJK windows environment.
 15:48
JimmyZ I got it fixed eventually ... 15:52
segomos JimmyZ: pong
JimmyZ segomos: how about push you commit to somewhere? 15:53
and good night
segomos JimmyZ: am planning to when i can compile it..i changed the fs stat function to work with uv_fs_t
err uv_stat_t ^
backend storage is offline in part of the libuv branch... 16:05
16:32 lizmat joined
segomos diakopter: can you give me push access to libuv branch? 17:00
17:03 not_gerd joined
not_gerd BabsSeed: if you're looking for things to do, you could help me with my build system work 17:06
(assuming you're not on windows)
17:13 crab2313 joined 17:17 benabik joined 17:41 crab2313 joined 17:45 not_gerd left 17:50 FROGGS joined 18:21 japhb_ joined
BabsSeed not_gerd 19:11
Oh he left :(
I have Windows & Linux boxes
19:18 not_gerd joined
not_gerd BabsSeed: it would be helpful if you could try to build github.com/MoarVM/MoarVM/pull/46 on your linux box and tell me how it blows up 19:19
20:32 colomon joined
BabsSeed not_gerd: Sure, sec lemme boot her up 20:32
not_gerd: I have Windows box, OBSD box and Linux box btw
not_gerd BabsSeed: linux would be helpful right now 20:34
you know how to locally checkout the pull request?
BabsSeed not_gerd: Yeah, have to install git fast 20:46
not_gerd: Any prereqs? 20:47
not_gerd BabsSeed: nope, any dependencies are bundled 20:49
jnthn develops on windows (ENOPACKAGEMANAGER)
on linux, `perl Configure.pl` *should* work 20:50
`make` will likely fail to link
BabsSeed not_gerd: I dev on Windows too 20:52
My excuse is I work with DirectX :P
not_gerd cross-compiles from cygwin to mingw64 20:57
BabsSeed not_gerd: OK I pulled the master repo, cant seem to pull this patch :D 21:04
Spot the github noob 21:05
not_gerd BabsSeed: help.github.com/articles/checking-...ts-locally
you need to add a line to your .git/config
then, you can `git checkout pr/46` 21:06
BabsSeed not_gerd: Thanks, running make now 21:10
(Have done perl Configure.pl)
not_gerd anything FAIL or everything OK when running Configure.pl? 21:11
BabsSeed All OK 21:13
No errors on 'make' either
not_gerd can you run ./moarvm ? 21:14
BabsSeed $ ./moarvm
Error: You must supply an
input file
Seems so
not_gerd curious 21:15
other mentioned linux needed an -luuid
I did not include it as I could not test it myself
BabsSeed I jHmm 21:16
I'm running Wheezy x64
And just ran what you said: perl Configure.pl; make
And seems to work :P
not_gerd well, to see if it *really* works, you'll have to build the nqp cross-compiler as well 21:17
BabsSeed I'm on it, need to get NQP first though, dont have it on this box
Oh nice, the config script will do it for me <3 21:18
How you know the script was written by perl devs: it allows for laziness
not_gerd well, the laziness is there and hubris probably as well 21:20
impatience, though?
that's not reallymy experience when building parrot/nqp/rakudo 21:21
BabsSeed nqp needs lreadline, licuuc and licudata
not_gerd I think one can disable those 21:23
BabsSeed I'll go install them
May try to compile moarvm on obsd later, anyone tried? 21:28
I have obsd on my 'router' (which is an i7 >_<) 21:29
Hmm, my issues all seem to be with nqp :D 21:30
make: *** [src/vm/parrot/ops/nqp_ops.so] Error 1
KnowHOWREPR.c:88:27: error: expected expression before 'void'
not_gerd BabsSeed: that's the ALIGNOF1 line? 21:37
BabsSeed not_gerd: That's what I see yeah 21:42
Sorry had to install emacs, my vi-fu is nothing special
spec.align = ALIGNOF1(void *);
not_gerd I suspect I figured out why it fails 21:46
you have an install/ subdirectory?
look for include/parrot/x.x.x/parrot/has_header.h 21:47
if PARROT_HAS_COMPILER_OFFSETOF_ALIGNOF is not set, ALIGNOF1 won't work with all types 21:48
you could probably make it compile by replacing ALIGNOF1(void *) --> ALIGNOF1(voidptr) 21:53
that might break the opposite configuration, though
21:54 colomon joined
BabsSeed Sec will check, sorry am cooking too 21:57
22:03 colomon joined 22:06 segomos joined
BabsSeed not_gerd: Issue was I had parrot installed from debian repo so it was using my ancient parrot instead of building new, I think 22:26
Trying with fresh parrot 22:28
22:31 segomos joined
not_gerd pushed configuration changes for *BSD ~10 minutes ago 22:32
BabsSeed not_gerd: Will try tomorrow, 00:30 here and have to be awake at 07:00
22:33 lizmat joined
BabsSeed not_gerd: BSD that is, I'll finish this compile for Linux before I sleep 22:34
22:35 lizmat joined
not_gerd also needs to get some sleep 22:37
BabsSeed: thanks for the help
doing cross-platform development on a single one is a bit academic
segomos i have a linux box if you just want me to compile.. 22:38
not_gerd segomos: if you have the time, try github.com/MoarVM/MoarVM/pull/46 22:39
BabsSeed not_gerd: No problem, I've got Xen running on one box so if we want to try specific archs (i386/x64) of *BSDs or Linuxes or specific kernel versions 22:41
Should all be available
segomos not_gerd: configure fails on osx for libuv.a 22:44
gist.github.com/tony-o/6216048 22:45
not_gerd segomos: that's expected
if you want to fix it yourself, see github.com/joyent/libuv/blob/maste...le.am#L153
BabsSeed segomos: Compiling nqp now
Its a slow box, just a VM
segomos should i just build libuv manually and make?
not_gerd build/Makefile.in and Configure.pl need to be patched accordingly
segomos: the whole point of that branch is to get the build system figured out 22:46
segomos ic
not_gerd bypassing it somewhat defeats the purpose ;)
BabsSeed not_gerd: At the weekend I'll set up a buildbox so we can automatically build & test stuff if that would help 22:52
I've got a 2x dual-core Opteron lying around
not_gerd segomos: just pushed some changes 22:54
should now compile on darwin, but not link
BabsSeed: that probably will help 22:57
you should get in touch with jnthn and/or diakopter
BabsSeed Will do
22:58 FROGGS joined
not_gerd that's it for me 22:58
good night o/
BabsSeed not_gerd: paste2.org/UdbEtvkI
My output from compiling nqp-cc 22:59
OK, catch you tomorrow, I'll hit the hay too
error there though
not_gerd might be sime of the stuff JimmyZ++ did
I suspect master doesn't doesn't have this problem 23:00
not_gerd really off now
23:00 not_gerd left
BabsSeed Linux cloudsdale 2.6.32-042stab078.28 #1 SMP Fri Aug 2 17:16:15 MSK 2013 x86_64 GNU/Linux 23:00
That's my VM info
If anyone knows why that may happen, let me know and I'll take another look tomorrow at lunch or compile break
23:52 cognominal joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!