[Mesa-dev] [PATCH 13/18] GROSS HACK: anv: add dummy MESA_GIT_SHA1 define

[Emil Velikov]

On Friday, 27 May 2016, Jason Ekstrand <jason at jlekstrand.net> wrote:

> On Fri, May 27, 2016 at 7:35 AM, Emil Velikov <emil.l.velikov at gmail.com
> <javascript:_e(%7B%7D,'cvml','emil.l.velikov at gmail.com');>> wrote:
>
>> From: Emil Velikov <emil.velikov at collabora.com
>> <javascript:_e(%7B%7D,'cvml','emil.velikov at collabora.com');>>
>>
>> Otherwise we'll error out if build from a release tarball.
>>
>> Cc: Jason Ekstrand <jason.ekstrand at intel.com
>> <javascript:_e(%7B%7D,'cvml','jason.ekstrand at intel.com');>>
>> Cc: Kristian Høgsberg Kristensen <krh at bitplanet.net
>> <javascript:_e(%7B%7D,'cvml','krh at bitplanet.net');>>
>> ---
>> This patch isn't going to fly obviously but is just enough to get things
>> building ;-)
>>
>
> You're right.  It won't. :-)
>
> I think what we need is to ship something that provides MESA_GIT_SHA1 in
> the tarball.  The purpose of this is to ensure that any shader caches get
> 100% invalidated if the mesa version changes at all.  When Timothy lands
> his shader cache stuff, we'll need this in the GL driver as well so we
> might as well find a solution now.
>
> Indeed. I was looking at how GL does it only to find it isn't wired up
yet. And yes shipping git_sha1.h sounds like a good solution. Question is
do we want to print the sha in other places (GL_VERSION iirc) for release
builds ? I'd go with yes on that one as well.

I'm on the move(train) atm so it might be a little late in the day for
anyone to review 'v2'.


> All of the other (not this one) anv/isl patches in the series are
>
> Reviewed-by: Jason Ekstrand <jason at jlekstrand.net
> <javascript:_e(%7B%7D,'cvml','jason at jlekstrand.net');>>
>

Thanks.
Emil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20160527/7b65bb94/attachment-0001.html>