gdata: Remove a werid behavior from GDataCacheMetadata::UpdateCache()

gdata: Remove a werid behavior from GDataCacheMetadata::UpdateCache()

Though it's not documented in the function comment, UpdateCache()
had ability to remove an existing cache entry. This is more like a
surprising behavior. This patch removes the behavior.

Along the way, rename the function to be more descriptive.
Also rename RemoveFromCache() to RemoveCacheEntry().

BUG=136921
TEST=out/Release/unit_tests --gtest_filter=GData*

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=146251

[hshi1, 2012-07-12 00:27:48]

http://codereview.chromium.org/10690154/diff/1/chrome/browser/chromeos/gdata/...
File chrome/browser/chromeos/gdata/gdata_cache_metadata.cc (right):

http://codereview.chromium.org/10690154/diff/1/chrome/browser/chromeos/gdata/...
chrome/browser/chromeos/gdata/gdata_cache_metadata.cc:187: } else {  // Resource
exists.
Can you DCHECK here to make sure the cache state is not CACHE_STATE_NONE?

[satorux1, 2012-07-12 00:35:31]

http://codereview.chromium.org/10690154/diff/1/chrome/browser/chromeos/gdata/...
File chrome/browser/chromeos/gdata/gdata_cache_metadata.cc (right):

http://codereview.chromium.org/10690154/diff/1/chrome/browser/chromeos/gdata/...
chrome/browser/chromeos/gdata/gdata_cache_metadata.cc:187: } else {  // Resource
exists.
On 2012/07/12 00:27:48, hshi1 wrote:
> Can you DCHECK here to make sure the cache state is not CACHE_STATE_NONE?

CACHE_STATE_NONE will be gone, once we change the representation from a bitmask
to a set of bools.

In theory, we could write a function that checks if all bools are false, but
it's probably not worth doing it.

[hshi1, 2012-07-12 00:36:15]

I see, LGTM, thanks!.
On 2012/07/12 00:35:31, satorux1 wrote:
>
http://codereview.chromium.org/10690154/diff/1/chrome/browser/chromeos/gdata/...
> File chrome/browser/chromeos/gdata/gdata_cache_metadata.cc (right):
> 
>
http://codereview.chromium.org/10690154/diff/1/chrome/browser/chromeos/gdata/...
> chrome/browser/chromeos/gdata/gdata_cache_metadata.cc:187: } else {  //
Resource
> exists.
> On 2012/07/12 00:27:48, hshi1 wrote:
> > Can you DCHECK here to make sure the cache state is not CACHE_STATE_NONE?
> 
> CACHE_STATE_NONE will be gone, once we change the representation from a
bitmask
> to a set of bools.
> 
> In theory, we could write a function that checks if all bools are false, but
> it's probably not worth doing it.