[PATCH weston] tests: Skip buffer count test if wayland EGL extension isn't present

[Derek Foreman]

On 22/01/15 03:48 PM, Daniel Stone wrote:
> Hi,
> 
> On Tuesday, January 20, 2015, Derek Foreman <derekf at osg.samsung.com
> <mailto:derekf at osg.samsung.com>> wrote:
> 
>     This also skips the test when running on the headless backend.
> 
> 
> I like this part!
>  
> 
>     +       if (strstr(extensions, "EGL_WL_bind_wayland_display") == NULL)
>     +               skip("Wayland extensions not present.\n");
> 
> 
> But not this part. bind_wayland_display is the server-side extension,
> not the client.

Is the concern here that some EGL implementation wont return
EGL_WL_bind_wayland_display in the string to the client (despite the
server actually using it), and the test will be skipped when it shouldn't?

> I'd rather skip on:
>   - EGL_KHR_platform_base present but not platform_wayland (see Jonny's
> patch series)

EGL_KHR_platform base isn't in the list of extensions available here.  I
can't find any documentation for that extension, do you have a URL handy?

Which patch series are you referring to?  I can't find anything in
patchwork, is it older than that?

>   - failure to connect to Wayland display

The test is successfully connecting to the wayland display.  (shouldn't
it just fail if it can't do that?)

>   - failure to create context or surface

eglCreateWindowSurface() is succeeding, as is eglMakeCurrent()

> 
> That should eliminate basically all the false positives. Thanks for
> tackling this niggle though.
> 
> Cheers,
> Daniel