Based on @nacx and @andrewgaul's comments, I think the best course of action for now is:
* leave this PR open for now * try to get hold of @adriancole to determine the origin of the "weird encoding". If it was added only for an S3 issue as @nacx's search suggests, then it should be easier to try to determine whether we need it or not, since @andrewgaul and @diwakergupta will be making plenty of connections against S3 * if we can't get hold of Adrian, potentially merge this for 1.7.0-SNAPSHOT and ask users to test the snapshot. That gives us a chance to revert if necessary without creating a DOA release. I'm fine with leaving https://github.com/jclouds/jclouds/pull/78 in because it does not affect existing behaviour. Based on @diwakergupta's [comment](https://github.com/jclouds/jclouds/pull/82#issuecomment-21988773) I'm unclear on whether we can expect that feature to _work_, or not, but in the worst case 1.6.2-incubating will have a bug in JCLOUDS-200 that we can hopefully address in the next release --- Reply to this email directly or view it on GitHub: https://github.com/jclouds/jclouds/pull/82#issuecomment-22005263
