[LIBCLOUD-612] Customize HTTP vendor prefix for Google Storage

[sfriesel]

https://issues.apache.org/jira/browse/LIBCLOUD-612

Google storage uses/expects custom headers like meta data
and storage class to be prefixed with "x-goog" instead of
"x-amz". This enables use of object tagging in the Google
Storage provider.

[sfriesel]

Unittests passed on 2.7 and I successfully uploaded an object with metadata and retrieved them again.

>>> bucket.upload_object("testfile", 'testfiles/foobar', extra={'meta_data':{'mytag': 'myvalue'}})
>>> bucket.get_object('testfiles/foobar').meta_data
{'mytag': 'myvalue'}

[sfriesel]

@Kami Is anything missing? I could provide the whole debug log of the interaction or maybe some credentials for testing if needed

[Kami]

@sfriesel I'll have a look asap.

[Kami]

I need to spend some more time testing this myself. The reason for this is that we are using S3 compatible API which requires you to send x-amzn headers.

From your pull request, it seems like you can use x-goog headers even if you use the S3 compatible API, but I need to confirm this myself.

[Kami]

I just tested your changes and verified it works correctly.

While testing the changes I also noticed meta_data is only returned when retrieving information for a single object (get_object), but not when listing objects (list_container_objects).

I will update documentation and add a note about this known limitation.

[Kami]

Merged into trunk. Thanks!