[omniORB] omniORBpy: implicit str->wstring & unicode->string conversion

[Luke Deller]

Hi,

Do you think it is a good idea for omniORB to be a bit more relaxed about str/unicode mismatches?

Python itself will happily converts implicitly from str to unicode [eg 'hi' + u'there'] and from unicode to str [eg sys.stdout.write(u'test'), using the default encoding].  OmniORBpy however will throw a BAD_PARAM exception when trying to marshal a python str to a CORBA wstring, or a python unicode string to a CORBA string.

I can suggest a patch to make omniORB relax in this regard... but before I do so, can somebody tell me if it's a bad idea?

Thanks,
Luke.