Sebastian Dröge
2018-03-26 08:01:52 UTC
Package: mesa
Version: 17.3.7-1
Severity: serious
Forwarded: https://bugs.freedesktop.org/show_bug.cgi?id=105328
Hi,
Currently it's impossible to include the GL and GLES headers
simultaneously on non-64 bit architectures, see e.g.
https://buildd.debian.org/status/fetch.php?pkg=gst-plugins-good1.0&arch=armel&ver=1.14.0-1&stamp=1521580828&raw=0
This causes
a) gst-plugins-base1.0 to only build with support for GL and not both
GL and GLES on the affected architectures,
b) causes gst-plugins-good1.0 to fail to build because Qt is including
the GLES headers, while GStreamer includes only the GL headers because
of the above
Currently gst-plugins-good1.0 fails to build on armel/armhf and can't
migrate to testing because of this bug. As a workaround I could force
gst-plugins-good to only build with GLES support on armel/armhf but
that's just that, a workaround.
This was reported upstream here
https://bugs.freedesktop.org/show_bug.cgi?id=105328
and upstream agrees that this is a bug, and it seems easy enough to fix
by guarding the type definitions with the preprocessor to not redefine
them if they were already defined.
Thanks!
Version: 17.3.7-1
Severity: serious
Forwarded: https://bugs.freedesktop.org/show_bug.cgi?id=105328
Hi,
Currently it's impossible to include the GL and GLES headers
simultaneously on non-64 bit architectures, see e.g.
https://buildd.debian.org/status/fetch.php?pkg=gst-plugins-good1.0&arch=armel&ver=1.14.0-1&stamp=1521580828&raw=0
This causes
a) gst-plugins-base1.0 to only build with support for GL and not both
GL and GLES on the affected architectures,
b) causes gst-plugins-good1.0 to fail to build because Qt is including
the GLES headers, while GStreamer includes only the GL headers because
of the above
Currently gst-plugins-good1.0 fails to build on armel/armhf and can't
migrate to testing because of this bug. As a workaround I could force
gst-plugins-good to only build with GLES support on armel/armhf but
that's just that, a workaround.
This was reported upstream here
https://bugs.freedesktop.org/show_bug.cgi?id=105328
and upstream agrees that this is a bug, and it seems easy enough to fix
by guarding the type definitions with the preprocessor to not redefine
them if they were already defined.
Thanks!