Add final part of IPC plumbing for FileSystem API

Add final part of IPC plumbing for FileSystem API.

BUG=32277
TEST=none; to be added when we have complete implementation.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=57915

[kinuko, 2010-08-27 00:15:33]

On 2010/08/27 00:07:17, Kinuko Yasuda wrote:
> 

Right now I'm abusing WebKit structures in filesystem IPC.  I'm willing to
replace some of them if it doesn't look good.

[Eric U., 2010-08-27 00:35:33]

Some of the param_traits stuff is a bit hazy to me; I'd like someone who's
comfortable with it to comment as well.

http://codereview.chromium.org/3208007/diff/1/2
File chrome/browser/file_system/file_system_dispatcher_host.cc (right):

http://codereview.chromium.org/3208007/diff/1/2#newcode110
chrome/browser/file_system/file_system_dispatcher_host.cc:110: bool
for_directory) {
Instead of for_directory, perhaps is_directory [for OnExists as well, and in
other related files]?

http://codereview.chromium.org/3208007/diff/1/5
File chrome/common/file_system/file_system_dispatcher.h (right):

http://codereview.chromium.org/3208007/diff/1/5#newcode64
chrome/common/file_system/file_system_dispatcher.h:64: WebKit::WebFileInfo
file_info);
const WebKit::WebFileInfo&, perhaps?  This may get bigger, and you've only got a
forward declaration above.

http://codereview.chromium.org/3208007/diff/1/5#newcode68
chrome/common/file_system/file_system_dispatcher.h:68: void DidFail(int32
callbacks_id, WebKit::WebFileError);
Are you planning to use DidSucceed/DidFail for the true/false return of Exists? 
Or is there the concept of a successful, "Yes it doesn't exist"?

I'm not saying we need anything separate--just clarifying how the API works.

http://codereview.chromium.org/3208007/diff/1/9
File chrome/common/webkit_param_traits.h (right):

http://codereview.chromium.org/3208007/diff/1/9#newcode353
chrome/common/webkit_param_traits.h:353: struct ParamTraits<WebKit::WebString> {
The comment at the top of the file specifically calls out WebString as something
we don't want to serialize.

[kinuko, 2010-08-27 18:26:15]

http://codereview.chromium.org/3208007/diff/1/2
File chrome/browser/file_system/file_system_dispatcher_host.cc (right):

http://codereview.chromium.org/3208007/diff/1/2#newcode110
chrome/browser/file_system/file_system_dispatcher_host.cc:110: bool
for_directory) {
On 2010/08/27 00:35:33, Eric U. wrote:
> Instead of for_directory, perhaps is_directory [for OnExists as well, and in
> other related files]?

Done.

http://codereview.chromium.org/3208007/diff/1/5
File chrome/common/file_system/file_system_dispatcher.h (right):

http://codereview.chromium.org/3208007/diff/1/5#newcode64
chrome/common/file_system/file_system_dispatcher.h:64: WebKit::WebFileInfo
file_info);
On 2010/08/27 00:35:33, Eric U. wrote:
> const WebKit::WebFileInfo&, perhaps?  This may get bigger, and you've only got
a
> forward declaration above.

Changed this to file_util::FileInfo and made it const reference.
(We only need mod_time for now but if we're going to include file size in
Metadata sending back file_info won't be a bad idea)

http://codereview.chromium.org/3208007/diff/1/5#newcode68
chrome/common/file_system/file_system_dispatcher.h:68: void DidFail(int32
callbacks_id, WebKit::WebFileError);
On 2010/08/27 00:35:33, Eric U. wrote:
> Are you planning to use DidSucceed/DidFail for the true/false return of
Exists? 

Yes I am.  It returns "Yes, it exists  == success" or "No, it doesn't exist ==
failure".    (Actually it may have been better to subsume this one in
ReadMetadata WebKit API... it was suggested by a few people)

> I'm not saying we need anything separate--just clarifying how the API works.

http://codereview.chromium.org/3208007/diff/1/9
File chrome/common/webkit_param_traits.h (right):

http://codereview.chromium.org/3208007/diff/1/9#newcode353
chrome/common/webkit_param_traits.h:353: struct ParamTraits<WebKit::WebString> {
On 2010/08/27 00:35:33, Eric U. wrote:
> The comment at the top of the file specifically calls out WebString as
something
> we don't want to serialize.

Removed this trait and changed the code not to use WebKit's entry structure.

[michaeln, 2010-08-27 19:05:39]

lgtm

http://codereview.chromium.org/3208007/diff/13001/14012
File chrome/common/webkit_param_traits.h (right):

http://codereview.chromium.org/3208007/diff/13001/14012#newcode333
chrome/common/webkit_param_traits.h:333: struct
SimilarTypeTraits<WebKit::WebFileSystem::Type> {
ah... i like this trait!