But the C library does support it (and ones that don't will start - +nnnn is ISO format, and frankly, I don't know a C library that does not support it, though I'm sure some exist).

The question is different - as the original post shows:

>>> strftime("%a %b %e %H:%M:%S %Y %Z %z")
'Tue May 23 16:28:31 2006 IST +0100'
>>> strftime("%a %b %e %H:%M:%S %Y %Z %z", localtime())
'Tue May 23 16:28:31 2006 IST +0000'

Python has clearly used the C-library to get correct data in the first code path through strftime, but there is a second code path that prints wrong data.

Someone posted code analysis in comp.lang.python explaining this, maybe that can be used as a starting point for a patch to fix the two code paths in Python strftime to work the same?