Handle the fact that MIMETypeRegistry::getMIMETypeForPath returns...

Handle the fact that MIMETypeRegistry::getMIMETypeForPath returns
application/octet-stream when it can't figure out the MIME type for a path.

BUG=26253
TEST=LayoutTests/http/tests/navigation/reload-subframe-object.html

[Victor Wang, 2009-10-30 18:44:28]

http://codereview.chromium.org/344030/diff/3001/3002
File webkit/glue/webframeloaderclient_impl.cc (right):

http://codereview.chromium.org/344030/diff/3001/3002#newcode1288
Line 1288: if (mime_type.isEmpty() || mime_type == "application/octet-stream")
what if the path is a binary file. getMIMETypeForPath returns
"application/octet-stream" if it is a binary file or file extension is not
recognized. The old code just sets ObjectContentFrame if the extension is not
recognized. This patch changes the behavior, should we return ObjectContentFrame
if the path points to a binary file?

[dglazkov, 2009-11-10 16:21:56]

Can we get this moving again?

[dglazkov, 2009-11-10 16:22:15]

Can we get this moving again?

http://codereview.chromium.org/344030

[dglazkov, 2009-11-10 16:28:16]

Can we get this moving again?

http://codereview.chromium.org/344030

[abarth-chromium, 2009-11-10 16:54:55]

Yes.  Currently the Sify issue is my top priority, but this is second.

[abarth-chromium, 2009-11-10 16:55:15]

Yes.  Currently the Sify issue is my top priority, but this is second.

http://codereview.chromium.org/344030

[abarth-chromium, 2009-11-10 17:01:15]

Yes.  Currently the Sify issue is my top priority, but this is second.

http://codereview.chromium.org/344030

[abarth-chromium, 2009-11-10 17:11:22]

Yes.  Currently the Sify issue is my top priority, but this is second.

http://codereview.chromium.org/344030

[abarth-chromium, 2009-11-12 10:57:17]

This moved upstream with the WebKit API:
https://bugs.webkit.org/show_bug.cgi?id=31398.  I'm going to close this code
review.