[oe] [meta-qt5][RFC] qtbase: fix build for gles2 enabled by PACKAGECONFIG

Andreas Müller schnitzeltony at googlemail.com
Tue Oct 29 08:14:09 UTC 2013


On Tue, Oct 29, 2013 at 8:07 AM, Martin Jansa <martin.jansa at gmail.com> wrote:
> On Tue, Oct 29, 2013 at 12:06:15AM +0100, Andreas Müller wrote:
>> While playing around with QtQuick I needed gles2 support. Doing so I faced tons
>> of error messages as mentionet in patch. It is caused by include chain:
>>
>> EGL/egl.h -> EGL/eglplatform.h -> X11/Xlib.h -> X.h
>>
>> In X.h we find definitions like
>>
>> #ifndef None
>> #define None                 0L       /* universal null resource or null atom */
>> #endif
>>
>> This causes building Qt to crash when using these kind of symbols.
>>
>> The patch was sent as RFC because:
>>
>> 1.  I would like ro discuss this at interest at qt-project.org and/or
>>     development at qt-project.org but it was not possible to subscribe. I sent
>>     a mail to listmaster but no response either.
>> 2.  qtbase_git might need same (and fails with the patch)
>> 3.  This is not a proper solution: It is fragile and will return for future
>>     changes. Proper soultion would be renaming the symbols Qt but I cannot
>>     discuss due to 1.
>> 4.  Nobody reported a build error for meta-qt5 so far.
>
> We had similar problem, but fixed gles2 provider instead, in our case we
> were using libhybris and the fix was to prevent loading X11 headers when
> we want to use just gles2 from it.
>
> Similar changes were added to mesa lately, can you do the same with
> gles2 provider on your platform?
It is the egl part causing trouble and the headers we use are the
latest from khronos. I don't understand what the workaround exactly
looks like - could you provide further hints?
>
> I must admit that I haven't read the patch yet and haven't finished
> morning coffee :).
>
Andreas



More information about the Openembedded-devel mailing list