[Bf-committers] Linux build error: svn 35825

Richard Shaw hobbes1069 at gmail.com
Wed Mar 30 17:38:01 CEST 2011


On Wed, Mar 30, 2011 at 10:31 AM, Campbell Barton <ideasman42 at gmail.com> wrote:
> On Wed, Mar 30, 2011 at 3:03 PM, Richard Shaw <hobbes1069 at gmail.com> wrote:
>> On Mon, Mar 28, 2011 at 8:11 AM, Richard Shaw <hobbes1069 at gmail.com> wrote:
>>> On Mon, Mar 28, 2011 at 6:44 AM, Campbell Barton <ideasman42 at gmail.com> wrote:
>>>> This is used by a macro to check if python interpreter is active:
>>>> #define PYC_INTERPRETER_ACTIVE
>>>> (((PyThreadState*)_Py_atomic_load_relaxed(&_PyThreadState_Current)) !=
>>>> NULL)
>>>>
>>>> Py 3.1 works a bit differently, you could just check if
>>>> PyThreadState_GET() is not NULL, in py3.2 that aborts blender so I had
>>>> to look at pythons source.
>>
>> Campbell,
>>
>> That did the trick. My patch worked on the first try (which only
>> impresses me because I'm not a C programmer so I guessed at the
>> syntax).
>>
>> I got a good build but grep'ing through the build log I did find these warnings:
>>
>> Parsing file /builddir/buil/builddir/build/BUILD/blender-2.56.svn35889/source/blender/makesrna/intern/rna_ui_api.c:77:
>> Warning: include file UI_icons.h not found, perhaps you forgot to add
>> its directory to INCLUDE_PATH?
>> Prep/builddir/build/BUILD/blender-2.56.svn35889/intern/iksolver/intern/TNT/vecadaptor.h:254:
>> Warning: include file iostream not found, perhaps you forgot to add
>> its directory to INCLUDE_PATH?
>>
>> Anything to worry about?
>>
>> Thanks,
>> Richard
>
> Thats strange, if the file isnt found blender shouldnt build, the
> include is added
> ./source/blender/makesrna/intern/CMakeLists.txt:145
>
> Just built with gcc4.6 and it doesn't give any warnings about missing
> includes with  RNA.

Never mind. After investigating I'm pretty sure the error is from
doxygen and not the build process.

Thanks,
Richard


More information about the Bf-committers mailing list