Replace uses of GR_DEBUG by SK_DEBUG.

Replace uses of GR_DEBUG by SK_DEBUG.

BUG=None
R=bsalomon@google.com,robertphillips@google.com

Committed: http://code.google.com/p/skia/source/detail?r=10978

[tfarina, 2013-08-23 02:07:18]

please, check if it is alright.

[bsalomon, 2013-08-23 12:54:23]

Am I wrong that we need to either:

a) Make SK_DEBUG be defined to 0 or 1

-or-

b) Change all the lines that read
#if GR_DEBUG
to
#ifdef SK_DEBUG
?

For the places where we stick GR_DEBUG in a struct (gmmain.cpp, benchmain.cpp) I
think if we went with b) above then we'd need something like:
#ifdef SK_DEBUG
static const bool kDebugOnly = true;
#else
static const bool kDebugOnly = false;
#endif

and then replace the instances of GR_DEBUG in struct declarations with
kDebugOnly.

https://codereview.chromium.org/23137022/diff/9001/bench/benchmain.cpp
File bench/benchmain.cpp (right):

https://codereview.chromium.org/23137022/diff/9001/bench/benchmain.cpp#newcod...
bench/benchmain.cpp:259: { SkBitmap::kARGB_8888_Config,  "Debug",        0,
kGPU_Backend,          GrContextFactory::kDebug_GLContextType,  SK_DEBUG },
Does this work? GR_DEBUG was always 0 or 1, but SK_DEBUG is just defined.

https://codereview.chromium.org/23137022/diff/9001/include/gpu/GrConfig.h
File include/gpu/GrConfig.h (right):

https://codereview.chromium.org/23137022/diff/9001/include/gpu/GrConfig.h#new...
include/gpu/GrConfig.h:131: *  e.g. it can specify SK_DEBUG/GR_RELEASE as it
please, change the BUILD
We should remove this first example since GR_USER_CONFIG_FILE is not the place
to set SK_ anything

e.g. it can change the BUILD target or supply its own defines for anything else
(such as GR_DEFAULT_TEXTURE_CACHE_MB_LIMIT)

[tfarina, 2013-08-23 23:26:17]

https://codereview.chromium.org/23137022/diff/9001/bench/benchmain.cpp
File bench/benchmain.cpp (right):

https://codereview.chromium.org/23137022/diff/9001/bench/benchmain.cpp#newcod...
bench/benchmain.cpp:259: { SkBitmap::kARGB_8888_Config,  "Debug",        0,
kGPU_Backend,          GrContextFactory::kDebug_GLContextType,  SK_DEBUG },
On 2013/08/23 12:54:23, bsalomon wrote:
> Does this work? GR_DEBUG was always 0 or 1, but SK_DEBUG is just defined.
just learned something new today :)

#if vs #ifdef.

If I understood this preprocessors of C right, #if is the macro equivalent of if
clause, and #ifdef is actually just testing if the macro was defined or not, it
doesn't care if the macro has a value or not.

[tfarina, 2013-08-24 00:16:48]

skia code is too dammit big/dense!!! that is another thing I learned today just
looking at a couple of files (e.g, gmmain.cpp) :(

Brian, could you take another look?

[bsalomon, 2013-08-24 02:28:35]

Hey Thiago, Thanks for doing this. I've noted a few more places where we need to
not treat SK_DEBUG as having a 0 or 1 value.

https://codereview.chromium.org/23137022/diff/25001/include/gpu/GrUserConfig.h
File include/gpu/GrUserConfig.h (right):

https://codereview.chromium.org/23137022/diff/25001/include/gpu/GrUserConfig....
include/gpu/GrUserConfig.h:17: #if 0
Let's just kill this whole block. I don't know why it's here.

https://codereview.chromium.org/23137022/diff/25001/include/gpu/gl/GrGLConfig.h
File include/gpu/gl/GrGLConfig.h (right):

https://codereview.chromium.org/23137022/diff/25001/include/gpu/gl/GrGLConfig...
include/gpu/gl/GrGLConfig.h:110: #define GR_GL_LOG_CALLS                        
    SK_DEBUG
Hmm... I don't think this will work. I think we need:

#if !defined(GR_GL_LOG_CALLS)
    #ifdef SK_DEBUG
        #define GR_GL_LOG_CALLS 1
    #else
        #define GR_GL_LOG_CALLS 0
#endif

and similar for GR_GL_CHECK_ERROR.

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGLProgramDesc.h
File src/gpu/gl/GrGLProgramDesc.h (right):

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGLProgramDes...
src/gpu/gl/GrGLProgramDesc.h:19: #define GR_GL_EXPERIMENTAL_GS SK_DEBUG
#ifdef SK_DEBUG
   #define GR_GL_EXPERIMENTAL_GS 1
#else
   #define GR_GL_EXPERIMENTAL_GS 0
#endif

This is one of the reasons I'd like to eventually convert all the boolean SK_
macros from #ifdef style to #if style. It also makes it possible to write things
like:

if (SK_SOME_FEATURE && runtime_check()) {
   do_something();
}

which can be handy at times.

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGpuGL.cpp
File src/gpu/gl/GrGpuGL.cpp (right):

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGpuGL.cpp#ne...
src/gpu/gl/GrGpuGL.cpp:892: #if 0 && SK_DEBUG
#if 0 && defined(SK_DEBUG)

[tfarina, 2013-08-24 23:25:49]

Brian, ptal.

https://codereview.chromium.org/23137022/diff/25001/include/gpu/GrUserConfig.h
File include/gpu/GrUserConfig.h (right):

https://codereview.chromium.org/23137022/diff/25001/include/gpu/GrUserConfig....
include/gpu/GrUserConfig.h:17: #if 0
On 2013/08/24 02:28:36, bsalomon wrote:
> Let's just kill this whole block. I don't know why it's here.

Done.

https://codereview.chromium.org/23137022/diff/25001/include/gpu/gl/GrGLConfig.h
File include/gpu/gl/GrGLConfig.h (right):

https://codereview.chromium.org/23137022/diff/25001/include/gpu/gl/GrGLConfig...
include/gpu/gl/GrGLConfig.h:110: #define GR_GL_LOG_CALLS                        
    SK_DEBUG
On 2013/08/24 02:28:36, bsalomon wrote:
> Hmm... I don't think this will work. I think we need:
> 
> #if !defined(GR_GL_LOG_CALLS)
>     #ifdef SK_DEBUG
>         #define GR_GL_LOG_CALLS 1
>     #else
>         #define GR_GL_LOG_CALLS 0
> #endif
> 
> and similar for GR_GL_CHECK_ERROR.

Done.

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGLProgramDesc.h
File src/gpu/gl/GrGLProgramDesc.h (right):

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGLProgramDes...
src/gpu/gl/GrGLProgramDesc.h:19: #define GR_GL_EXPERIMENTAL_GS SK_DEBUG
On 2013/08/24 02:28:36, bsalomon wrote:
> #ifdef SK_DEBUG
>    #define GR_GL_EXPERIMENTAL_GS 1
> #else
>    #define GR_GL_EXPERIMENTAL_GS 0
> #endif
> 
> This is one of the reasons I'd like to eventually convert all the boolean SK_
> macros from #ifdef style to #if style. It also makes it possible to write
things
> like:
> 
> if (SK_SOME_FEATURE && runtime_check()) {
>    do_something();
> }
> 
> which can be handy at times.

Done.

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGpuGL.cpp
File src/gpu/gl/GrGpuGL.cpp (right):

https://codereview.chromium.org/23137022/diff/25001/src/gpu/gl/GrGpuGL.cpp#ne...
src/gpu/gl/GrGpuGL.cpp:892: #if 0 && SK_DEBUG
On 2013/08/24 02:28:36, bsalomon wrote:
> #if 0 && defined(SK_DEBUG)

Done.

[tfarina, 2013-08-27 11:12:26]

Brian, Robert, ping??

[bsalomon, 2013-08-27 17:55:54]

https://codereview.chromium.org/23137022/diff/32001/include/gpu/GrConfig.h
File include/gpu/GrConfig.h (right):

https://codereview.chromium.org/23137022/diff/32001/include/gpu/GrConfig.h#ne...
include/gpu/GrConfig.h:94: #if defined(SK_DEBUG) && defined(GR_RELEASE)
I think this needs to be 
#if defined(SK_DEBUG) && GR_RELEASE

since GR_RELEASE is always defined.

[tfarina, 2013-08-28 03:27:46]

https://codereview.chromium.org/23137022/diff/32001/include/gpu/GrConfig.h
File include/gpu/GrConfig.h (right):

https://codereview.chromium.org/23137022/diff/32001/include/gpu/GrConfig.h#ne...
include/gpu/GrConfig.h:94: #if defined(SK_DEBUG) && defined(GR_RELEASE)
On 2013/08/27 17:55:54, bsalomon wrote:
> I think this needs to be 
> #if defined(SK_DEBUG) && GR_RELEASE
> 
> since GR_RELEASE is always defined.
you are right. arr, I didn't review it carefully, sorry. :(

Done.

[bsalomon, 2013-08-28 12:59:25]

On 2013/08/28 03:27:46, tfarina wrote:
> https://codereview.chromium.org/23137022/diff/32001/include/gpu/GrConfig.h
> File include/gpu/GrConfig.h (right):
> 
>
https://codereview.chromium.org/23137022/diff/32001/include/gpu/GrConfig.h#ne...
> include/gpu/GrConfig.h:94: #if defined(SK_DEBUG) && defined(GR_RELEASE)
> On 2013/08/27 17:55:54, bsalomon wrote:
> > I think this needs to be 
> > #if defined(SK_DEBUG) && GR_RELEASE
> > 
> > since GR_RELEASE is always defined.
> you are right. arr, I didn't review it carefully, sorry. :(
> 
> Done.

lgtm

[commit-bot: I haz the power, 2013-08-28 12:59:56]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/tfarina@chromium.org/23137022/60001

[commit-bot: I haz the power, 2013-08-28 14:17:51]

Change committed as 10978