[Webkit-unassigned] [Bug 53529] [fileapi] Add support for filesystem: URI handling

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=53529





--- Comment #20 from Adam Klein <adamk at chromium.org>  2011-02-02 17:38:04 PST ---
(From update of attachment 80975)
View in context: https://bugs.webkit.org/attachment.cgi?id=80975&action=review

>> LayoutTests/fast/filesystem/filesystem-uri-origin.html:1
>> +<a style="display:none" href="filesystem:http://www.webkit.org:8080/temporary/a/b/c/file.txt">foo</a>
> 
> Here you have webkit.org as the filesystem origin.  In directory-entry-to-uri-expected.txt you have file:// as the origin.  In an iframe test I see http://localhost:8000.  Please make sure these tests will work both on the webkit bots and when run locally.

The file:/// URIs are there to avoid needing to be an http test. Would you rather I use http tests for those toURI() tests?

The iframe test uses localhost to distinguish it from 127.0.0.1 (where the hosted page is running).

The use of webkit.org:8080 here is purely ornamental, since this URI is never followed; it's simply being used to set up the dom so that the origin is available.  Changing this to localhost or similar wouldn't change anything...

>>> Source/WebCore/fileapi/Entry.cpp:95
>>> +    uriBuilder.append(m_fileSystem->asyncFileSystem()->type() == AsyncFileSystem::Temporary ? "temporary" : "persistent");
>> 
>> Do the strings "temporary" and "persistent" exist anywhere as constants?  It would be nice not to have them inline here.  At the very least they should be shared with the code that cracks the URI.
> 
> Same for the protocol name "filesystem"?
> 
> The changes in FileSystem callback chains to include type parameter lgtm.

Any suggestions about where these strings would live?  Right now, the URI cracking code is in Chromium.  Should it move into WebCore?

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.