Yeah s3cmd is a name of more than one thing.  That's probably what it was.

I have ruby 1.8.6, and use it on Windows 7.

Newest s3sync works with ruby 1.8.x.

I'm not sure if anyone got it working with the 1.9 ruby branch.  One of these days I'm going to get around to a rewrite that won't require ruby.

It could be a bug still.  I have been going by memory and not actually testing. 

I have been trying to get time to make a new s3sync that will be more easily extended and have less limitations.  One down side to having this plan is that making changes and fixes to the old code seems a waste of time, and I've been neglecting it.

Aha, I should have seen this from your command line.

I'm doing this from memory here, but:
When you don't end the S3 path in a slash, anything that starts with that "prefix" gets seen as a subdirectory.  I think this could be classified as a bug in the client code, since I don't think anyone ever intends that usage.  It's an artifact of how S3 does list operations.  Try trailing the s3 path with a slash?

You can PM it to me. (I don't think the admin account can get PMs) 

Unfortunately I haven't worked on this in a substantial capacity since before S3 supported the rename feature, so there's no "move" other than downloading out of S3 and re-uploading.  Which clearly is not optimal.

You have to re-set-up your certificate validation.. it doesn't know whether to trust the S3 certificate, so it's failing.

Sorry man I still have the 1.8.x ruby branch everywhere.  Anyone else?

This sounds like a serious issue, if it is caused by the sync.  I have been very busy and unable to work on this project at all for a long time (as you noted).

But if you are able to isolate some case where this corruption is occurring (say, with a specific file) and you can send me that file to test with, I'll have a look at what's going on.

My long term "plan" for s3sync is to rewrite it in C and Lua because the ruby runtime is just total crap and has been (indirectly) responsible for many of the complaints about the project.

I don't understand where you're coming up with your syntax.  s3cmd has no argument "sync".  There's no "S3:://".  I think you may have confused my code with someone else's tools....

Well first, you probably don't really want a / after the bucket name and colon.  The debug flag might give you more info.  But are you sure it's hanging?  It won't report as it's checking files that match up, only when copying new ones.  Maybe verbose just isn't verbose enough for you?

Actually it does not store the MD5 for files, that comes for free with S3.  But it won't have permission data, which .. I think is OK.  It will likely freak out about not having the right folder nodes.

I never designed it to play nice with other tools, and it usually doesn't. 

I'm sorry that the s3sync code is so fragile with respect to delimiters.  I mean to rewrite the entire project, but have not had time.