Fix a unit test failure. One of the tests sends an empty bitmap through this...

Fix a unit test failure. One of the tests sends an empty bitmap through this
function, which has no config, and the assertion fails. I just removed the
assretion, we assume ARGB8888 everywhere.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=17976

[brettw, 2009-06-09 21:06:06]

TBR

[Peter Kasting, 2009-06-09 21:10:45]

http://codereview.chromium.org/118456/diff/1/2
File skia/ext/image_operations.cc (left):

http://codereview.chromium.org/118456/diff/1/2#oldcode629
Line 629: DCHECK(bitmap.getConfig() == SkBitmap::kARGB_8888_Config);
Instead of removing this, the right fix seems like moving it below the "nop
case".  That will avoid DCHECKing unless we're actually going to do something
destructive, but still catch the bad case.  The unit tests would pass with that.

[brettw, 2009-06-09 21:12:49]

I am not concerned about this at all. We will fail in 1000 different wans for
non 8888 bitmaps, plus getAddr32 below will fail and the error will not be
obscure. I only added this at the last minute when Carlos mentioned it.

[cpu_(ooo_6.6-7.5), 2009-06-09 21:55:07]

I mentioned because I assumed that was somewhat standard practice to check the
type of bitmap. If it is not standard practice to set/check it then LGTM. 


On 2009/06/09 21:12:49, brettw wrote:
> I am not concerned about this at all. We will fail in 1000 different wans for
> non 8888 bitmaps, plus getAddr32 below will fail and the error will not be
> obscure. I only added this at the last minute when Carlos mentioned it.