[Bf-committers] [BUG]Blender requires large file support on 32-bit plattforms

Giuseppe Ghibò ghibo at mandriva.com
Thu May 20 12:30:37 CEST 2010


Jochen Schmitt wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hallo,
>
> Unfortunately i have got several bug reports where people complaints
> crashes
> of blender-2.49b because they have files with more then 2G on there disc.
>
> this crashes are occures because off_t has a size of 4 bytes on a
> 32-bit system if
> you con't explicit request large file support during compile time.
>
> So I want to notfy you about this issue, so you can file Blender-2.49b
> and the
> upcoming 2.50 release if necessary.
>
> Best Regards:
>
> Jochen Schmitt
>
> Bugs:
> https://bugzilla.redhat.com/show_bug.cgi?id=585668
> https://bugzilla.redhat.com/show_bug.cgi?id=587343
> https://bugzilla.redhat.com/show_bug.cgi?id=589031
>
> Patch:
>
> diff -up blender-2.49b/config/linux2-config.py.org
> blender-2.49b/config/linux2-config.py
> - --- blender-2.49b/config/linux2-config.py.org   2009-09-01
> 17:23:25.000000000 +0200
> +++ blender-2.49b/config/linux2-config.py       2010-05-19
> 20:08:30.599484197 +0200
> @@ -177,10 +177,16 @@ CXX = 'g++'
>  ##ifeq ($CPU),alpha)
>  ##   CFLAGS += -pipe -fPIC -funsigned-char -fno-strict-aliasing -mieee
>  
> - -CCFLAGS = ['-pipe','-fPIC','-funsigned-char','-fno-strict-aliasing']
> +CCFLAGS = ['-pipe','-fPIC','-funsigned-char','-fno-strict-aliasing',
> '-D_LARGEFILE_SOURCE', '-D_FILE_OFFSET_BITS=64']
>  
>  CPPFLAGS = ['-DXP_UNIX']
> - -CXXFLAGS = ['-pipe','-fPIC','-funsigned-char','-fno-strict-aliasing']
> +CXXFLAGS = ['-pipe','-fPIC','-funsigned-char','-fno-strict-aliasing',
> '-D_LARGEFILE_SOURCE', '-D_FILE_OFFSET_BITS=64']
> +
> +import os
> +RPM_OPT_FLAGS = os.environ['RPM_OPT_FLAGS'].split()
> +CCFLAGS += RPM_OPT_FLAGS
> +CXXFLAGS += RPM_OPT_FLAGS
> +
>  REL_CFLAGS = ['-O2']
>  REL_CCFLAGS = ['-O2']
>  ##BF_DEPEND = True
> diff -up blender-2.49b/extern/SConscript.org
> blender-2.49b/extern/SConscript
> - --- blender-2.49b/extern/SConscript.org 2009-09-01 17:23:25.000000000
> +0200
> +++ blender-2.49b/extern/SConscript     2010-05-19 20:01:52.052609611
> +0200
> @@ -2,8 +2,6 @@
>  
>  Import('env')
>  
> - -SConscript(['glew/SConscript'])
> - -
>  if env['WITH_BF_GAMEENGINE']:
>         if env['WITH_BF_SOLID']:
>                 SConscript(['qhull/SConscript', 'solid/SConscript'])
> diff -up blender-2.49b/SConstruct.org blender-2.49b/SConstruct
> - --- blender-2.49b/SConstruct.org        2009-09-01 17:23:49.000000000
> +0200
> +++ blender-2.49b/SConstruct    2010-05-19 20:01:52.053609534 +0200
> @@ -223,7 +223,7 @@ if env['OURPLATFORM'] == 'linux2' :
>  
>                 def CheckFreeAlut(context,env):
>                         context.Message( B.bc.OKGREEN + "Linux
> platform detected:\n  checking for FreeAlut... " + B.bc.ENDC )
> - -                       env['LIBS'] = 'alut'
> +                       env['LIBS'] = 'alut, openal'
>                         result =
> context.TryLink(mylib_test_source_file, '.c')
>                         context.Result(result)
>                         return result
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/
>
> iJwEAQECAAYFAkv0Qj0ACgkQZLAIBz9lVu+TKwQAsCCvmngQDd4bLjmypD5Z21/I
> JU4/o5/iJxkq7YXb+Diqvbbe/kpFgMP7qOLqkRDbJ7Hr4Qy1g29mfHtFpMlIk5P6
> YzDuq4atO+b2UmHZbl1ZwNAPyGi6/qAorg0xoLxniYMg85c0iHcU8Pzgnz90+0qL
> fsDxknjcZa8YEmi1kVk=
> =CH7m
> -----END PGP SIGNATURE-----
>
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers
>
>   
I don't understand why this comes up again. This was discussed here:

http://lists.blender.org/pipermail/bf-committers/2008-February/020248.html

and Campbell committed a subset of my patch for bug #6811 (see 
patchtracker),

http://svn.mandriva.com/cgi-bin/viewvc.cgi/packages/cooker/blender/releases/2.44/6.20070912.1mdv2008.0/SOURCES/blender-2.44-bug6811.patch?revision=84806&view=markup

long time ago, so why it needs again the forcing of 64bit offset?



More information about the Bf-committers mailing list