[Mesa-dev] [PATCH 16/23] readpix: raise priority of FBO completeness error

[Ian Romanick]

On 01/09/2013 12:47 PM, Jordan Justen wrote:
> On Tue, Jan 8, 2013 at 9:21 AM, Jordan Justen <jljusten at gmail.com> wrote:
>> On Mon, Jan 7, 2013 at 1:19 PM, Ian Romanick <idr at freedesktop.org> wrote:
>>> On 01/04/2013 06:41 PM, Jordan Justen wrote:
>>>>
>>>> GTF/gles3 test suite wants this error to have higher priority
>>>> than the type checking.
>>>
>>>
>>> If the ReadPixels call in the test has multiple errors, the test should be
>>> changed.  There is no requirement to generate a specific error when multiple
>>> problems exist.  Is my understanding of the test case correct?
>>
>> Okay, it sounds like it is a test bug. I'll drop this patch and look
>> at the test instead.
>
> Looks like I'm wrong about the test. It seems like the fb-inc error is
> the only valid error.
>
> But, in patch 12, I added:
>     rb = _mesa_get_read_renderbuffer(ctx, internalFormat);
>     if (rb == NULL) {
>        _mesa_error(ctx, GL_INVALID_OPERATION,
>                    "glCopyTexImage%dD(read buffer)", dimensions);
>        return GL_TRUE;
>     }
>
> This is done because we need to look at the read-buffer format type
> for gles3. But, this code will also cause an error when the
> framebuffer is not complete, only I had it return invalid-op.
>
> So, I think this patch is needed to check the framebuffer completeness
> before this code tries to get the read-buffer. I propose a new log
> message:
>
> readpix: check FBO completeness before trying to access the read-buffer

Ah... that makes sense.

Reviewed-by: Ian Romanick <ian.d.romanick at intel.com>

> -Jordan