These methods don't issue any HTTP requests so they are much faster than their existing equivalents. The downside is that you need to know the bucket name and object key. But if you do these methods will be much faster. Signed-off-by: Tomas Carnecky <email@example.com>
otherwise Ruby will think is a different file and complaint about a double declaration of S3::VERSION
Fixes an issue with Paperclip calling TempFile.read internally just before flushing writes, resulting in s3_paperclib.rb uploading an empty file. See http://github.com/thoughtbot/paperclip/commit/2cdeb39a514242df2fdc8a75e193dcecde5a74d5
The S3::Request subclass of Net::HTTPGenericRequest was causing problems when used alongside anything that uses the http_connection gem. Since the http_connection gem also fixes chunked encoding, defer to it instead of doing our own thing.
You can now pass an options hash to Bucket#save and it will pass the options through to Connection#request. This is useful for setting headers (like :x_amz_acl) on the bucket. Backwards compatibility is maintained from the previous version that just accepted a location parameter.
This method generates a complete URL to an S3 resource. It accepts the same parameters as generate_temporary_url_signature. Both methods have been updated to accept optional :method and :headers parameters to generate temporary signatures for POSTS, PUTS, and DELETEs.