CHANGELOG

head:

0.6.3:

- Make bucket names globally unique [Jeffrey Hardy jeff@37signals.com]
- Add S3Object.update for updating things like the content-type without re-uploading the file [Jeffrey Hardy jeff@37signals.com]
0.6.2:

- Use custom __method__ in Ruby versions *prior* to 1.8.7 not *up to* 1.8.7.

0.6.1:

- Rename Kernel#memoize to Kernel#expirable_memoize so that it doesn't conflict with memoize method in ActiveSupport which has an entirely different API and semantics. Reported by [Florian Dütsc (mail@florian-duetsch.de)].

0.6.0:

- Full 1.9 compatibility (all tests passing against 1.9 & 1.8.6). Thanks to [David (dvdplm@gmail.com), Cyril David (cyx.ucron@gmail.com)]

0.5.1:

- For now just redefine __method__ to take arguments so we don't break 1.8.7 use today

0.5.0:

- Bug #17458 fixed. Normalize the :expires_in option to always be an integer even if the actual object passed in is a proxy to an integer, such as is the case with 2.hours from ActiveSupport which is actually an instance of ActiveSupport::Duration. Reported by [Steve Kickert steve@riverocktech.com]

- Bug #19158 fixed. Don't prepend leading slash onto bucket name when deleting a bucket with the :force => true option.

- Bug #17628 fixed. Don't ignore :use_ssl => false in url_for when the connection is established over ssl. Reported by [Tom Fixed (tkat11)]

- Bug #13052 fixed. Appease some proxies by always including a Content-Length with all requests. Reported by [James Murty (jmurty)]

- Bug #13756 fixed. Attributes that are false should not raise NoMethodError in Base#method_missing. Fixed by [Scott Patten]

- Bug #19189 fixed. No longer reference Date::ABBR_MONTHS constant which was removed in Ruby 1.8.6. Reported by [Khurram Virani (kvirani)]

- Bug #20487 fixed. If a request fails and is retried, only escape the request path the first time. Reported by [anonymous]

- Replace ad-hoc S3Object.copy method with newly support built in API call.

- Do not make connections persistent by default. This "feature" causes far more broken pipes than it is worth. Use with caution.

0.4.0:

- Various adjustments to connection handling to try to mitigate exceptions raised from deep within Net::HTTP.

- Don't coerce numbers that start with a zero because the zero will be lost. If a bucket, for example, has a name like '0815', all operation it will fail. Closes ticket #10089 [reported anonymously]"

- Add ability to connect through a proxy using the :proxy option when establishing a connection. Suggested by [Simon Horne <simon@soulware.co.uk>]

- Add :authenticated option to url_for. When passing false, don't generate signature parameters for query string.

- Make url_for accept custom port settings. [Rich Olson] 

0.3.0:

- Ensure content type is eventually set to account for changes made to Net::HTTP in Ruby version 1.8.5. Reported by [David Hanson, Stephen Caudill, Tom Mornini <tmornini@engineyard.com>]

- Add :persistent option to connections which keeps a persistent connection rather than creating a new one per request, defaulting to true. Based on a patch by [Metalhead <metalhead@metalhead.ws>]

- If we are retrying a request after rescuing one of the retry exceptions, rewind the body if its an IO stream so it starts at the beginning. [Jamis Buck]

- Ensure that all paths being submitted to S3 are valid utf8. If they are not, we remove the extended characters. Ample help from [Jamis Buck]

- Wrap logs in Log objects which exposes each line as a Log::Line that has accessors by name for each field.

- Various performance optimizations for the extensions code. [Roman LE NEGRATE <roman2k@free.fr>]

- Make S3Object.copy more efficient by streaming in both directions in parallel.

- Open up Net:HTTPGenericRequest to make the chunk size 1 megabyte, up from 1 kilobyte.

- Add S3Object.exists?

0.2.1:

- When the bucket name argument (for e.g. Bucket.objects) is being used as the option hash, reassign it to the options variable and set the bucket to nil so bucket inference + options works.

- Don't call CGI.escape on query string parameters in Hash#to_query_string since all paths get passed through URI.escape right before the request is made. Paths were getting double escaped. Bug spotted by [David Hanson]

- Make s3sh exec irb.bat if on Windows. Bug spotted by [N. Sathish Kumar <nsathishk@yahoo.com>]

- Avoid class_variable_(get|set) since it was only recently added to Ruby. Spotted by [N. Sathish Kumar <nsathishk@yahoo.com>]

- Raise NoSuchKey if S3Object.about requests a key that does not exist.

- If the response body is an empty string, don't try to parse it as xml.

- Don't reject every body type save for IO and String at the door when making a request. Suggested by [Alex MacCaw <maccman@gmail.com>]

- Allow dots in bucket names. [Jesse Newland]

0.2.0:

- Infer content type for an object when calling S3Object.store without explicitly passing in the :content_type option.

0.1.2:

- Scrap (overly) fancy generator based version of CoercibleString with a much simpler and clearer case statement. Continuations are really slow and the specific use of the generator was leaking memory. Bug spotted by [Remco van't Veer]

0.1.1:

- Don't add the underscore method to String if it is already defined (like, for example, from ActiveSupport). Bug spotted by [Matt White <stockliasteroid@gmail.com>]

0.1.0:

- Initial public release