I found out that the -m option of s3cmd for automated mime-type-scanning doesn't work - correct?
Just add some lines to S3.py at line 178 (function object_put):
content_type = mimetypes.guess_type(filename)
if not content_type:
content_type = 'text/plain'
headers["content-type"] = content_type
You'll also have to include the mimetypes-python library at the beginning of the script:
Now, just recompile/install the S3-module width "python setup.py install" and s3cmd will use pythons mimetype-detection and set the correct "content-type" header for S3-contents.