On 23 February 2017 at 17:34, Matt Turner wrote:
> s/flacky/flakey/ in the title.
>
> On Thu, Feb 23, 2017 at 9:13 AM, Emil Velikov
> wrote:
>> From: Emil Velikov
>>
>> Note that all the symbols/ABI checks in tree do so :-\
>>
>> Signed-off-by: Emil Velikov
>> ---
>> Mostly an RFC since the es*api code might be going soon.
>
> Tell me more :)
>
Should have said "soon (tm)" ;-)
GLVND effectively makes these dead code/deprecated/unwanted. I've got
the EGL one locally, but I need to do some testing on it.
>> ---
>> src/mapi/es2api/ABI-check | 4
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/src/mapi/es2api/ABI-check b/src/mapi/es2api/ABI-check
>> index e0bf3c8314..e338408c7e 100755
>> --- a/src/mapi/es2api/ABI-check
>> +++ b/src/mapi/es2api/ABI-check
>> @@ -5,6 +5,10 @@
>> # GL_EXT_multi_draw_arrays
>> # GL_OES_EGL_image
>>
>> +# FIXME: checking only for T symbols in incomplete/wrong.
>
> s/in/is/ ?
>
> What's wrong with checking for T symbols? I don't know, and I think it
> should be in the commit message.
In general we want to track all symbols (but the platform specific
_edata and friends) since those are the ABI we expose. For example:
The DRI modules expose "B __driDriverExtensions", while glapi "B
_glapi_tls_Context" and "D _glapi_tls_Dispatch" amongst others.
Not to mention that libEGL/libgbm provides fluid ABI based on the
configure toggles. I.e. wl_drm_interface is always exposed by libgbm,
for libEGL things vary. That has some Wayland roots which I'm also
trying to resolve as well, just need some convincing.
Obviously special casing libGLES* doesn't bring much, even if they
expose only T (barring the platform specific of course) symbols.
Thanks
Emil
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev