Allow for no pdf rasterizers in gm.

Allow for no pdf rasterizers in gm.

Our poppler does not build properly with VS2013, but using
gyp_skia -Dskia_enable_poppler_enabled=0
also does not build due to kPDFRasterizers being statically empty.
This adds a do nothing, not run by default 'pdf rasterizer'.

R=epoger@google.com

Committed: https://code.google.com/p/skia/source/detail?r=14676

[epoger, 2014-05-09 14:35:30]

https://codereview.chromium.org/270263009/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/270263009/diff/1/gm/gmmain.cpp#newcode1333
gm/gmmain.cpp:1333: // The following exists so that this array is never zero
length.
I take it you investigated making the code properly handle a zero-length array,
but that was more trouble than it was worth?

https://codereview.chromium.org/270263009/diff/1/src/utils/SkPDFRasterizer.h
File src/utils/SkPDFRasterizer.h (right):

https://codereview.chromium.org/270263009/diff/1/src/utils/SkPDFRasterizer.h#...
src/utils/SkPDFRasterizer.h:21: static bool SkNoRasterizePDF(SkStream*,
SkBitmap*) { return false; }
Maybe put this in gmmain.cpp instead?  (Nobody else needs to ever see it,
right?)

[bungeman-skia, 2014-05-09 14:56:14]

https://codereview.chromium.org/270263009/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/270263009/diff/1/gm/gmmain.cpp#newcode1333
gm/gmmain.cpp:1333: // The following exists so that this array is never zero
length.
On 2014/05/09 14:35:30, epoger wrote:
> I take it you investigated making the code properly handle a zero-length
array,
> but that was more trouble than it was worth?

Well, it's technically against the c++ specification to declare a static/stack
array of length 0. GCC seems to allow it as an extension, but VC++ doesn't allow
it. The only other thing I can think of is to make this an SkTDArray or
something and then init it somewhere. However, that's a lot of extra work.

https://codereview.chromium.org/270263009/diff/1/src/utils/SkPDFRasterizer.h
File src/utils/SkPDFRasterizer.h (right):

https://codereview.chromium.org/270263009/diff/1/src/utils/SkPDFRasterizer.h#...
src/utils/SkPDFRasterizer.h:21: static bool SkNoRasterizePDF(SkStream*,
SkBitmap*) { return false; }
On 2014/05/09 14:35:30, epoger wrote:
> Maybe put this in gmmain.cpp instead?  (Nobody else needs to ever see it,
> right?)

Done. I suppose I put it here on the idea that this is where the others are, and
also at the time I didn't know if it was going to need a more complex
implementation somewhere.

[epoger, 2014-05-09 15:15:35]

lgtm

https://codereview.chromium.org/270263009/diff/1/gm/gmmain.cpp
File gm/gmmain.cpp (right):

https://codereview.chromium.org/270263009/diff/1/gm/gmmain.cpp#newcode1333
gm/gmmain.cpp:1333: // The following exists so that this array is never zero
length.
On 2014/05/09 14:56:15, bungeman1 wrote:
> On 2014/05/09 14:35:30, epoger wrote:
> > I take it you investigated making the code properly handle a zero-length
> array,
> > but that was more trouble than it was worth?
> 
> Well, it's technically against the c++ specification to declare a static/stack
> array of length 0. GCC seems to allow it as an extension, but VC++ doesn't
allow
> it. The only other thing I can think of is to make this an SkTDArray or
> something and then init it somewhere. However, that's a lot of extra work.

Wow, "The More You Know".  Thanks; I'm a little sheepish that I didn't know this
already!

[bungeman-skia, 2014-05-09 15:22:43]

Committed patchset #2 manually as r14676 (presubmit successful).