mettli / maratis

Automatically exported from code.google.com/p/maratis
0 stars 0 forks source link

Cygwin exec do not launch anymore #4

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
1. build the latest trunk under cygwin
2. wait for the build
3. Try to run the player, editor or any other execs without success. No error 
code is returned. The program does not seem to enter inside the main().

What is the expected output? 
run

What do you see instead?
nothing (is nt there any logs ?)

Anael: 
1: have anyone changed any windows.py params ?
2: I did nt see any logs feature/function/classes indide MEngine. Would you 
mind if I had such ?

Original issue reported on code.google.com by skaiw...@gmail.com on 20 Oct 2012 at 11:49

GoogleCodeExporter commented 9 years ago
here is the output of gdb:
(gdb) run
Starting program: 
/cygdrive/g/devel/maratis/maratis/trunk/dev/prod/cygwin/release/Maratis/Bin/Mara
tisPlayer.exe
[New Thread 5572.0xcec]
[New Thread 5572.0x1820]

Program received signal SIGSEGV, Segmentation fault.
0x0028ff18 in ?? ()

How to build in debug ?

Original comment by skaiw...@gmail.com on 20 Oct 2012 at 2:00

GoogleCodeExporter commented 9 years ago
Here it is building Maratis in debug : python scons.py build_mode=debug
Waiting...

Original comment by skaiw...@gmail.com on 20 Oct 2012 at 2:17

GoogleCodeExporter commented 9 years ago
The debug version of maratis is not verbose:
Starting program: 
/cygdrive/g/devel/maratis/maratis/trunk/dev/prod/cygwin/debug/Maratis/Bin/Marati
sPlayer.exe
[New Thread 3944.0xf0c]
[New Thread 3944.0x1650]

Program received signal SIGSEGV, Segmentation fault.
0x0028ff18 in ?? ()

What the hell...

Original comment by skaiw...@gmail.com on 20 Oct 2012 at 3:49

GoogleCodeExporter commented 9 years ago
Hi, 

no I don't think anyone changed windows.py,
I just did some work on iOS after your last commit.
Did you change something in your system, external libs, etc ?

I actually never tested the debug compilation with scons, do you have the same 
crash in release ?

Original comment by anael.se...@gmail.com on 21 Oct 2012 at 9:26

GoogleCodeExporter commented 9 years ago
[deleted comment]
GoogleCodeExporter commented 9 years ago
Here it is I got it : before relinking I just added the crtdll.a library in 
order to link with __commit necessary/used by for npk_dev. But this lib seems 
to prevent for any exec to run. Consequently I still dont know how to link with 
npk_dev as it of course believes we are on a NPK_PLATFORM_WINDOWS and try to 
run _commit( handle ); :

NPK_RESULT npk_flush( NPK_HANDLE handle )
{
    if( handle != 0 )
    {
#ifdef NPK_PLATFORM_WINDOWS
    _commit( handle );
#else
    fsync( handle );
#endif
    }
    return NPK_SUCCESS;
}

Anael: any idea ?

Original comment by skaiw...@gmail.com on 24 Oct 2012 at 4:54

GoogleCodeExporter commented 9 years ago
ok, I see.

I guess we are forced to do a fix of npk
by simply forcing NPK_PLATFORM_LINUX using the __CYGWIN__ define (in npk_conf.h)

we should return the info to the npk team : http://code.google.com/p/npk/

Original comment by anael.se...@gmail.com on 25 Oct 2012 at 8:09

GoogleCodeExporter commented 9 years ago
Yes I have thought too of forcing NPK_PLATFORM_LINUX on cygwin as there is no 
NPK_PLATFORM_POSIX.
I have already created a ticket on npk website.
For the moment I have 'fixed' npk_dev.c directly by doing _commit if __CYGWIN__ 
not def.

Original comment by skaiw...@gmail.com on 26 Oct 2012 at 9:34

GoogleCodeExporter commented 9 years ago

Original comment by anael.se...@gmail.com on 1 Jan 2013 at 1:36