I expect a -beta2 or maybe even call it 1.5.0 "real soon now".  Please download github.com/s3tools/s3cmd master branch and verify that it still works for your use cases, and file bugs at github.com if not.  It's a good as I can make it for now.


On Sat, Mar 15, 2014 at 8:08 PM, Matt Domsch <matt@domsch.com> wrote:
On Mon, Feb 10, 2014 at 4:59 PM, Matt Domsch <matt@domsch.com> wrote:
I think 1.5.0 is getting close to release.  There are a few issues I'd still like to see addressed before another beta/rc release though.  Any input to these would be appreciated.

Open Pull Requests that I think are close to being mergeable, but need a little love:

* https://github.com/s3tools/s3cmd/pull/249
Fix incorrect content-header assignment and charset bug

A fix is in upstream master now for this.

Other Open Issues:

* Generically: unicode fixups.
It's not as simple as using LANG=en_US.UTF-8, but that's a good start, and is the root cause of a lot of the bug reports recently.

I've found a few more of these.  Tests and reports of success/failure will be appreciated.
* https://github.com/s3tools/s3cmd/issues/212
Regression: --exclude '*.*' --rinclude '^one_file#39; excludes all files

I just committed a fix for this.  To exclude a directory, use --exclude=somedir/  (trailing forward slash).

* https://github.com/s3tools/s3cmd/issues/37
s3cmd modify command
needs a patch, not just a bunch of +1s asking for the feature.

No change.

* https://github.com/s3tools/s3cmd/issues/269
s3cmd (1.5.0-beta1) on Windows and full file path with drive letter when uploading a file
Windows support has atrophied significantly lately.  I've fixed up a few spots, adding abstractions to noop for code paths that don't exist on Windows, but I'm amazed when it works at all.  The test suite fails.

I haven't done anything about this.  As no one else has stepped up, I say we release as-is and let users complain.  Or point them at other tools that do work on Windows.