When uploading an OpenOffice text document ("foo.sxw") to a File object ("foo.sxw"), it's content type prop is set to "application/x-zip-compressed" !!!
When downloading it with IE or Mozilla (Windows) it renames the file "foo.sxw.zip". The user has to rename it :-(.
Is there a way to set the correct content type automatically at upload time (dunno which) and prevent renaming the file ?
* Upload through Python scripts that detects those strnga behaviours and set the content type according to the file extension ?
* Add a plugin to the File class to check the right type and set content type accordingly ?
* Other ?
In the File class, in zope/lib/python/OFS/Image.py, the content-type is set by the supplied 'content-type' header or by content_types.guess_content_type, which uses the Python mimetypes library (plus a bit of extra logic for circumstances the library doesn't deal with). My guess is that you are not providing a type and that the mimetypes library's guess_type function is looking at the contents of your OO file to guess the type. Since they are, I think, zipped XML files, it's not a bad guess. Your first fix would work, though it's a bit messy. Better would be to fix it in either the Zope core or the Python library. Zope is easy enough: you can add your file extension to the '_addtypes' tuple in zope/lib/python/OFS/content_types.py directly, or if you want to be less invasive you can do a monkeypatch/hotfix. Then submit the change to the collector for inclusion in future versions, at which point you can discard your site-specific fix. The fix to Python will require going to the maintainer, perhaps with a patch. I can't say what the policy on changes to that library is. --jcc -- "My point and period will be throughly wrought, Or well or ill, as this day's battle's fought."