Fix up signed-integer-overflow warnings

Fix up signed-integer-overflow warnings

When checking whether a matrix was a pure scale, we subtracted
0x3f800000 from the diagonals; if the diagonal value was already
very negative, we'd underflow. Replace subtraction with XOR.

When dealing with repeating tiled bitmaps, when the bitmap was
very large, we'd multiply an offset by 65535, possibly causing
underflow. Throw in a cast to long (casting to unsigned also
silences the warning and wouldn't involve extension, but I can't
convince myself that it's correct).

BUG=skia:4635
R=mtklein@google.com

Committed: https://skia.googlesource.com/skia/+/d5c4265b49d212f7888d0516328f39c4d45d3058

[tomhudson, 2015-12-07 17:55:17]

There were only a couple of signed-integer-overflow warnings in the codebase.
PTAL?

[mtklein, 2015-12-07 17:58:16]

https://codereview.chromium.org/1504933002/diff/1/src/core/SkBitmapProcState_...
File src/core/SkBitmapProcState_matrixProcs.cpp (right):

https://codereview.chromium.org/1504933002/diff/1/src/core/SkBitmapProcState_...
src/core/SkBitmapProcState_matrixProcs.cpp:106: return SK_USHIFT16((long)((fx) &
0xFFFF) * ((max) + 1));
Let's do long -> int64_t?  long is funky across compilers/platforms.

[tomhudson, 2015-12-07 18:17:22]

tomhudson@google.com changed reviewers:
+ reed@google.com

[tomhudson, 2015-12-07 18:17:23]

Drew back to (unsigned), in more places, and added some SkASSERT()s as reed@
suggested. Now no warnings from dm or nanobench.

[mtklein, 2015-12-07 18:19:22]

lgtm

[reed1, 2015-12-07 18:21:48]

possibly we could try to share the expression in the bitmapshader, between the
macros and the function-ptr, but lgtm either way.

[tomhudson, 2015-12-07 18:22:20]

The CQ bit was checked by tomhudson@google.com

[commit-bot: I haz the power, 2015-12-07 18:22:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1504933002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1504933002/20001

[commit-bot: I haz the power, 2015-12-07 18:38:07]

Description was changed from

==========
Fix up signed-integer-overflow warnings

When checking whether a matrix was a pure scale, we subtracted
0x3f800000 from the diagonals; if the diagonal value was already
very negative, we'd underflow. Replace subtraction with XOR.

When dealing with repeating tiled bitmaps, when the bitmap was
very large, we'd multiply an offset by 65535, possibly causing
underflow. Throw in a cast to long (casting to unsigned also
silences the warning and wouldn't involve extension, but I can't
convince myself that it's correct).

BUG=skia:4635
R=mtklein@google.com
==========

to

==========
Fix up signed-integer-overflow warnings

When checking whether a matrix was a pure scale, we subtracted
0x3f800000 from the diagonals; if the diagonal value was already
very negative, we'd underflow. Replace subtraction with XOR.

When dealing with repeating tiled bitmaps, when the bitmap was
very large, we'd multiply an offset by 65535, possibly causing
underflow. Throw in a cast to long (casting to unsigned also
silences the warning and wouldn't involve extension, but I can't
convince myself that it's correct).

BUG=skia:4635
R=mtklein@google.com

Committed:
https://skia.googlesource.com/skia/+/d5c4265b49d212f7888d0516328f39c4d45d3058
==========

[commit-bot: I haz the power, 2015-12-07 18:38:07]

Committed patchset #2 (id:20001) as
https://skia.googlesource.com/skia/+/d5c4265b49d212f7888d0516328f39c4d45d3058