Add a tool to dump HTML5 filesystem from CUI

Add a tool to dump HTML5 filesystem from CUI

BUG=N/A
TEST=./out/Release/dump_file_system -l /tmp/hamaji/test-profile

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=194367

[kinuko, 2013-04-16 10:59:48]

Looks very handy, thanks for doing this!

Mostly nit-picky comments

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
File webkit/fileapi/dump_file_system.cc (right):

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:43: static void ShowMessageAndExit(const
std::string& msg) {
nit: no need of static?

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:48: static void ShowUsageAndExit(const
std::string& arg0) {
ditto

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:53: }  // anonymous namespace
nit: '// namespace' is fine in chrome

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:58: static bool g_opt_temporary;
can these be inside the anonymous space?

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:64: }
Append("t") for g_opt_temporary==true?

Can we use ObfuscatedFileUtil::GetDirectoryNameForType(type) here? (It's a
static method)

Just for our own convenience do you mind adding one more option (like '-s') to
show SyncableFileSystem contents?

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:71: " is not a filesystem origin directory");
nit: Please put { } for multi-line body (here and elsewhere)

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:76: " is not a consistent file system");
Probably you don't need these lines, or maybe add another flag to force calling
this?
(IsFileSystemConsistent() performs fsck like full scan)

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:182: ShowMessageAndExit(profile_dir.value() +
" is not a profile directory");
nit: "is not a directory" ?

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:185:
profile_dir.Append("Default").Append("File System"));
can you use
fileapi::SandboxMountPointProvider::kFileSystemDirectory
for "File System"?

Reg: "Default": In multi-profile environment the path may be different. Since
it's just an extra handy tool it's ok as is but it might be nicer to have
another option so that user can specify the profile user name (and if not
specified just use "Default")

[hamaji, 2013-04-16 12:25:12]

I also fixed (maybe) component build. Not sure mac build was fixed by this.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
File webkit/fileapi/dump_file_system.cc (right):

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:43: static void ShowMessageAndExit(const
std::string& msg) {
On 2013/04/16 10:59:48, kinuko wrote:
> nit: no need of static?

Done.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:48: static void ShowUsageAndExit(const
std::string& arg0) {
On 2013/04/16 10:59:48, kinuko wrote:
> ditto

Done.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:53: }  // anonymous namespace
On 2013/04/16 10:59:48, kinuko wrote:
> nit: '// namespace' is fine in chrome

Done.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:58: static bool g_opt_temporary;
On 2013/04/16 10:59:48, kinuko wrote:
> can these be inside the anonymous space?

Done.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:64: }
On 2013/04/16 10:59:48, kinuko wrote:
> Append("t") for g_opt_temporary==true?
> 
> Can we use ObfuscatedFileUtil::GetDirectoryNameForType(type) here? (It's a
> static method)
> 
> Just for our own convenience do you mind adding one more option (like '-s') to
> show SyncableFileSystem contents?

Done.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:71: " is not a filesystem origin directory");
On 2013/04/16 10:59:48, kinuko wrote:
> nit: Please put { } for multi-line body (here and elsewhere)

Done.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:76: " is not a consistent file system");
On 2013/04/16 10:59:48, kinuko wrote:
> Probably you don't need these lines, or maybe add another flag to force
calling
> this?
> (IsFileSystemConsistent() performs fsck like full scan)

Removed.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:182: ShowMessageAndExit(profile_dir.value() +
" is not a profile directory");
On 2013/04/16 10:59:48, kinuko wrote:
> nit: "is not a directory" ?

Done.

https://codereview.chromium.org/14294005/diff/1/webkit/fileapi/dump_file_syst...
webkit/fileapi/dump_file_system.cc:185:
profile_dir.Append("Default").Append("File System"));
On 2013/04/16 10:59:48, kinuko wrote:
> can you use
> fileapi::SandboxMountPointProvider::kFileSystemDirectory
> for "File System"?
> 
> Reg: "Default": In multi-profile environment the path may be different. Since
> it's just an extra handy tool it's ok as is but it might be nicer to have
> another option so that user can specify the profile user name (and if not
> specified just use "Default")

Done.

[kinuko, 2013-04-16 12:36:31]

lgtm

[tzik, 2013-04-16 15:09:35]

lgtm

[hamaji, 2013-04-16 16:14:39]

Committed patchset #6 manually as r194367 (presubmit successful).