source: http://blog.sina.com.cn/s/blog_57f61b4901015g56.html相隔短短两月，Openstack的对象存储组件Swift 1.7.5版于本月14号正式release.（2012/11/14）作为Swift的粉丝，当然要赶快跟进下，以下是对新版本的信息解读： 新版本的源码仍然托管于Github，可以从以下链接获取 from github.版本的升级仍然可以在线进行，换而言之，不用停机，不会中断已经在线的服务。在代码贡献者方面，这一版本有27位贡献者参与，其中有9位新人出现，目前Swift的代码贡献者总人数达到84名，详细清单可以在此查看Swift代码贡献者名单.1.7.5版修复了一些BUG，推出了一些新特性，（这里可以查看 完整的changelog）其中比较重要的几个新特性如下： CORS 支持CORS header意味着更广泛的前端支持，HTML5和JavaScript这下都没问题了。 Swift now supports setting CORS headers on your data. This allows HTML5 forms, browser uploads, and JavaScript to work between your website and your Swift storage. Users can set X-Container-Meta-Access-Control-Allow-Origin, X-Container-Meta-Access-Control-Max-Age, and X-Container-Meta-Access-Control-Allow-Headers on their containers to allow CORS support. Custom log handlers 产品化的一个重要特征是完备的log处理，你现在可以很方便的调用Swift的log，和自己的log处理系统进行集成了 Custom log handlers allow depoyers to more easily integrate with external log processors. This mechanism provides the ability to add your own handler to the log messages that Swift generates. The full docs are here. Multi-Range GETsMulti-Range的对象获取目前只限于标准对象，大型对象将在下一版本支持。 Swift has always supported single range requests, but the HTTP spec allows for requesting multiple ranges of a resource at once. Swift now supports this feature on standard objects, and support against large object manifests should be available in a later release. Configurable constraints 在安装部署时候，可以设置对客户端的限制，包括最大对象size等各种参数和行为，使得Swift很好管理。 Deployers can set the client limits in a config file. These limits include the maximum object size, the maximum number of items returned in listings, and the limits on metadata. You can read the fully-commented sample config file in the source code. Replicators can run against a specific partition or device 副本控制器 Replicators可以指定分区和设备进行工作，意味着可以人工干预副本的机制，有助于处理故障节点和进行特殊情况下的针对性优化 Replicators can now replicate a specific partition or device. This feature allows deployers to prioritize data replication for data that may not be fully replicated. For example, although Swift detects missing replicas automatically, a deployer can choose to replicate a specific set of data before the automatic systems detect the error.