[PATCH v4 09/19] Android manual state tracking: add integer,pointer maps

[José Fonseca]

On Mon, Jul 16, 2012 at 2:26 PM, Imre Deak <imre.deak at intel.com> wrote:
> Hi,
>
> On Sat, 2012-07-14 at 08:08 +0100, José Fonseca wrote:
>> Imre,
>>
>> One of these "Android manual state tracking ..." patches is causing
>> failures in apitrace test suite [1]:
>>
>> $ make test
>> [...]
>> The following tests FAILED:
>>         7 - app_gl_varray_drawarrays (Failed)
>>         8 - app_gl_varray_elements (Failed)
>>         9 - app_gl_varray_interleaved (Failed)
>>        10 - app_egl_gl_tri (Failed)
>>        13 - app_egl_gles1_tri (Failed)
>>
>> $ cat Testing/Temporary/LastTest.log
>> [...]
>> 7/35 Testing: app_gl_varray_drawarrays
>> 7/35 Test: app_gl_varray_drawarrays
>> Command: "/usr/bin/python"
>> "/home/jfonseca/projects/apitrace/tests/app_driver.py" "--apitrace"
>> "/home/jfonseca/projects/apitrace/apitrace" "--a
>> pitrace-source" "/home/jfonseca/projects/apitrace" "--api" "gl"
>> "--ref-dump" "/home/jfonseca/projects/apitrace/tests/apps/gl/varray_drawarrays.
>> ref.txt" "--" "/home/jfonseca/projects/apitrace/tests/apps/gl/varray"
>> "pointer" "drawarrays"
>> Directory: /home/jfonseca/projects/apitrace/tests/apps/gl
>> "app_gl_varray_drawarrays" start time: Jul 14 08:07 BST
>> Output:
>> ----------------------------------------------------------
>> /home/jfonseca/projects/apitrace/scripts
>> /home/jfonseca/projects/apitrace/tests/apps/gl/varray pointer drawarrays
>> /home/jfonseca/projects/apitrace/apitrace trace --api gl --output
>> /home/jfonseca/projects/apitrace/tests/apps/gl/varray_drawarrays.trace
>> -- /ho
>> me/jfonseca/projects/apitrace/tests/apps/gl/varray pointer drawarrays
>> apitrace: tracing to
>> /home/jfonseca/projects/apitrace/tests/apps/gl/varray_drawarrays.trace
>> apitrace: redirecting dlopen("libGL.so.1", 0x102)
>> /home/jfonseca/projects/apitrace/apitrace dump --color=never
>> /home/jfonseca/projects/apitrace/tests/apps/gl/varray_drawarrays.trace
>> FAIL (missing call `glColorPointer(size = 3, type = GL_FLOAT, stride =
>> 12, pointer = blob(72))` (found `glColorPointer(size = 3, type =
>> GL_FLOA
>> T, stride = 12, pointer = NULL)`))
>> <end of output>
>> Test time =   0.73 sec
>> ----------------------------------------------------------
>> Test Failed.
>>
>>
>> So it seems that pointers to user arrays are wrongly becoming NULL.
>
> Ok, I admit I haven't run these tests :/ I will try to do so now and
> track down the above issue.

Thanks. What I often do in such cases is to run astyle to the
generated .cpp file, before and after the change, and visually inspect
how the generated code changed.

Jose