Fix uninitialized array in color management.

Fix uninitialized Storage

BUG=387843
R=palmer@chromium.org

Committed: https://pdfium.googlesource.com/pdfium/+/456cde9

[palmer, 2014-07-10 18:54:17]

https://codereview.chromium.org/380293002/diff/1/core/src/fxcodec/lcms2/lcms2...
File core/src/fxcodec/lcms2/lcms2-2.6/src/cmslut.c (right):

https://codereview.chromium.org/380293002/diff/1/core/src/fxcodec/lcms2/lcms2...
core/src/fxcodec/lcms2/lcms2-2.6/src/cmslut.c:1308: cmsFloat32Number
Storage[2][MAX_STAGE_CHANNELS] = {0.0f};
I don't think you can initialize a 2-dimensional array with a 1-dimensional
array-literal. Does this compile on your system?

Probably easier to just use memset:

    memset(Storage, 0, sizeof(Storage));

[jun_fang, 2014-07-10 19:18:55]

On 2014/07/10 18:54:17, Chromium Palmer wrote:
>
https://codereview.chromium.org/380293002/diff/1/core/src/fxcodec/lcms2/lcms2...
> File core/src/fxcodec/lcms2/lcms2-2.6/src/cmslut.c (right):
> 
>
https://codereview.chromium.org/380293002/diff/1/core/src/fxcodec/lcms2/lcms2...
> core/src/fxcodec/lcms2/lcms2-2.6/src/cmslut.c:1308: cmsFloat32Number
> Storage[2][MAX_STAGE_CHANNELS] = {0.0f};
> I don't think you can initialize a 2-dimensional array with a 1-dimensional
> array-literal. Does this compile on your system?
> 
> Probably easier to just use memset:
> 
>     memset(Storage, 0, sizeof(Storage));

It can fix the issue raised in this bug. My original concern is that it has
better efficiency than memset. Anyway, it's a safer way to use memset.

[palmer, 2014-07-10 19:36:03]

> > I don't think you can initialize a 2-dimensional array with a 1-dimensional
> > array-literal. Does this compile on your system?

Apparently, C99 compilers reject it, but C++ compilers accept it. OK then.

> It can fix the issue raised in this bug. My original concern is that it has
> better efficiency than memset. Anyway, it's a safer way to use memset.

Since it works (?!), use the array literal syntax. It's cleaner than memset,
which matters.

(Do you have a benchmark showing that memset on small objects is a performance
concern?)

[palmer, 2014-07-10 19:36:10]

LGTM

[jun_fang, 2014-07-10 20:37:01]

Committed patchset #1 manually as r456cde9 (presubmit successful).