- The client loader creates a new object on every invocation. This makes possible working with multiple clients having different internal states (such as: API endpoints, SQS queue paths, etc).
- Forces dependencies that include bugfixes which affect aws2js.
- [BREAKS COMPAT] The automatic S3 path escaping implemented in v0.5.0 won't escape anymore the ? char. There's no proper technical solution for solving the ambiguities regarding this char, therefore, if it's part of the S3 file name / S3 file path, it has to be manually escaped. This situations happens due to the fact that this char is used for seding query params as well. This version enables the query params for all resources.
- Deprecated the usage of s3.escapePath(). Use JavaScript's encodeURI() instead.
- Sending query parameters was broken for the S3 client into the previous v0.5.x releases due to automatic escaping.
- Fixes the broken Content-Length header if Content-Type was manually specified for the S3 PUT operation.
- Added Amazon CloudFormation support.
- sqs.setQueue() in favor of sqs.queue(). Tells the story better if you look at how the aws2js API looks like.
- sqs.queue() helper for calling a specific SQS queue.
- s3.del() allows calling it without calling s3.setBucket(). If you really want to use that pesky path style addressing.
- Throws an Error() if the credentials aren't set. Wrapping the error handling into a callback is not practical here since it's clearly an user error.
- Throws an Error() if the request / response body handlers are invalid.
- Revoked the ability to override Content-MD5 for s3.putFile().
- The body handlers are more forgiving by returning the Error() as the error argument of the callback insteaf of throwing it.
- Added Amazon SQS support.
- [MIGHT BREAK COMPAT] The S3 paths are now escaped. The chars that normally aren't part of an S3 path are now URL encoded. The s3.escapePath() is called automatically. Exposed as helper in order to be able to know exactly the input for the S3 REST API.
- The 'closed' events of the HTTPS response are treated as errors. The AWS APIs should end the request cleanly. Normally they do. This fixes possible hangs of the HTTPS support.
- Avoids the node.js issue #1399 without my own http.js bundle.
- Adds the client.getApiVersion() method in order to indicate which is the default or defined API version. The query APIs support this feature. This is an elegant way of wrapping client.query.Version which may be an arcane methodology for outsiders. Usually useful for debugging.
- Adds the client.setApiVersion() method for setting the API version. The query APIs support this feature.
- Adds s3.setEndPoint() helper for the S3 client.
- Adds Amazon CloudWatch support.
- Adds Amazon ElastiCache support.
- Enables the client.getEndPoint() for all the API clients.
- Updates the EC2 API client to default to version 2011-07-15.
- Updates the ELB API client to default to version 2011-08-15.
- Updates the AutoScaling API client to default to version 2011-01-01.
- The input query for the query argument of the query APIs takes precendence over the query parameters that are configured into the client itself. This allows per API call custom configuration (eg: Version - indicates the API version).
- Integrates with mime-magic to provide the automatic MIME type detenction when the Content-Type header for the S3 PUT operation is undefined. This method does a bit of I/O, it is slower than the previous method for computing the MIMEs, but the results are more reliable. The libmagic functionality returns the MIME type by reading the file itself instead of doing a dumb file extension lookup.
- Deprecates query.call() in favor of query.request() for the query APIs.
- The query.request() method makes the query argument to be optional.
- Changed the internal structure of the library.
- Deprecates s3.putObject() in favor of s3.putFile().
- Implements new GET response handlers: buffer - the response contains a buffer key with the buffer contents; stream - returns the HTTPS response itself which implements node.js's Readable Stream interface.
- Adds a new s3.putStream() helper for PUT'ing streams to S3.
- Adds a new s3.putBuffer() helper for PUT'ing buffers to S3.
- Goes fully async by removing the only blocking call, fs.statSync(), when using PUT with file request body handler.
- Unit testing coverage.
- Fixes a possible race condition that could appear into the fsync(2) wrapper.
- Proper support for the fsync(2) wrapper. If s3.get() is used for downloading the objects to the disk, the ENOENT errors that could happened from time to time should now be gone.
- Adds s3.renameObject().
- Fixes the error reporting. When an AWS API didn't return XML as response body, the callback wasn't called.
- Fixes the broken handling of error reporting of the XML parsing.
- [BREAKS COMPAT] Removed the response argument in case of error. If there's an error document returned by the AWS API itself, it is exposed as error.document. This may break some code that expects the error document to be returned as the response argument. This change unifies the error reporting that won't expect the result argument anymore.
- [MIGHT BREAK COMPAT] Returns the error argument as null in case of success in order to follow the node.js convention instead of undefined. This may break some code if the evaluation was made against 'undefined'.
- Exposes client.getEndPoint() method if the client.setRegion() method is available.
- Calls fsync(2) after each downloaded file via s3.get() in order to make sure that the application has a consistent behavior when the s3.get() callback is called.
- Adds again the backport-0.4 dependency, v0.4.10-1, that targets issue #1399 from node v0.4.10. This release fixes a rare race condition that may appear when doing S3 PUT requests with bodies that are streamed from files.
- Drops the backport-0.4 dependency. node.js v0.4.10 finally came around with the desired fixes. Only node 0.4.10 and above is supported.
- Adds support for Amazon Auto Scaling.
- Exposes the client.setMaxSockets() method for changing the https.Agent.defaultMaxSockets property.
- Adds backport-0.4 as module dependency in order to properly fix the broken request.abort() support. This version is crappy workaround-free.
- Adds support for Amazon Identity and Access Management (IAM).
- If the WriteStream fails when the file response handler is in use by the GET request, the HTTPS request itself is aborted. Previously it was continued, therefore it might waste a lot of bandwidth, that you're going to pay for. This task was not trivial as a bug in node.js complicates the implementation of this feature. See #1085.
- Changes file the GET response handler to receive an object indicating the file path instead the file path itself in order to introduce more flexibility. Unfortunately this introduces a slight backward incompatibility. Hate doing it, but it's a must.
- Fixes the acl checker that did not accept a false value in order to go with the default 'private'.
- [BREAKS COMPAT] Client loader. Previously all the clients were loaded when aws2js was required. Now a specific client is loaded when executing the exported load() method. Unfortunately, this introduces backward incompatibility.
- [BREAKS COMPAT] Removed the client.config() method as it may break more stuff than it introduces.
- The README relies on Wiki pages in order to provide the docs.
- Amazon S3 support.
- Amazon ELB support.
- Made the Amazon RDS API version to be the latest 2011-04-01 by default.
- Made the Amazon SES API version to be the latest 2010-12-01 by default.
- Adds mime as dependency due to mime/type auto-detection for S3 uploads.
- Updates the libxml-to-js dependency to v0.2.
- Fixes the client.setRegion() call as it is currently broken.
- Disables client.setRegion() for SES.
- Implements the Amazon Simple Email Service (SES) client.
- Migrated to a cleaner XML to JS implementation (my own libxml-to-js wrapper).
- Initial public release with versioning and npm support.
- Initial version, featuring Amazon EC2 and Amazon RDS support.