CDM FileIOTest: Explictly test corrupted file.

CDM FileIOTest: Explictly test corrupted file.

Since Write() takes two (or more) steps to finish, it's possible for
CdmFileIOImpl to be destructed after SetLength() is called but before
WriteFile() is called. In that case, since we SetLength(data_size), we
could end up with a corrupted file.

This CL update FileIOTest to explictly cover this case.

BUG=410630
TEST=ECKEncryptedMediaTest.FileIOTest still passes.

Committed: https://crrev.com/f6b40dd68203de747ee018c8c913fe688c69e821
Cr-Commit-Position: refs/heads/master@{#294867}

[xhwang, 2014-09-12 05:05:46]

xhwang@chromium.org changed reviewers:
+ ddorwin@chromium.org

[xhwang, 2014-09-12 05:05:47]

This CL is split from 
https://codereview.chromium.org/562063002/

PTAL

[ddorwin, 2014-09-12 20:25:24]

https://codereview.chromium.org/563193002/diff/1/media/cdm/ppapi/cdm_file_io_...
File media/cdm/ppapi/cdm_file_io_test.cc (right):

https://codereview.chromium.org/563193002/diff/1/media/cdm/ppapi/cdm_file_io_...
media/cdm/ppapi/cdm_file_io_test.cc:78: // Generate |corrupted_big_data_|.
Do we really need a member for one test?

https://codereview.chromium.org/563193002/diff/1/media/cdm/ppapi/cdm_file_io_...
media/cdm/ppapi/cdm_file_io_test.cc:79: corrupted_big_data_.assign(kData, kData
+ kDataSize);
I don't think the naming is right. It is data but larger. Difficult naming is
another reason not to have a non-local variable.

[xhwang, 2014-09-13 01:19:30]

https://codereview.chromium.org/563193002/diff/1/media/cdm/ppapi/cdm_file_io_...
File media/cdm/ppapi/cdm_file_io_test.cc (right):

https://codereview.chromium.org/563193002/diff/1/media/cdm/ppapi/cdm_file_io_...
media/cdm/ppapi/cdm_file_io_test.cc:78: // Generate |corrupted_big_data_|.
On 2014/09/12 20:25:24, ddorwin wrote:
> Do we really need a member for one test?

We add test cases first then run all tests together. So the data to be expected
must be valid throughout the lifetime of this class.

Naming is hard...

https://codereview.chromium.org/563193002/diff/1/media/cdm/ppapi/cdm_file_io_...
media/cdm/ppapi/cdm_file_io_test.cc:79: corrupted_big_data_.assign(kData, kData
+ kDataSize);
On 2014/09/12 20:25:24, ddorwin wrote:
> I don't think the naming is right. It is data but larger. Difficult naming is
> another reason not to have a non-local variable.

Acknowledged.

[ddorwin, 2014-09-13 01:21:32]

lgtm

[xhwang, 2014-09-15 17:38:10]

The CQ bit was checked by xhwang@chromium.org

[commit-bot: I haz the power, 2014-09-15 18:02:18]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patchset/563193002/1

[commit-bot: I haz the power, 2014-09-15 19:33:29]

Committed patchset #1 (id:1) as 5564096ba6f9b171bd775b9e76052f19fc84164b

[commit-bot: I haz the power, 2014-09-15 19:37:35]

Patchset 1 (id:??) landed as
https://crrev.com/f6b40dd68203de747ee018c8c913fe688c69e821
Cr-Commit-Position: refs/heads/master@{#294867}