[XML-SIG] Re: Finding _xmlplus in Python 2.3a2

[James Kew]

Uche Ogbuji wrote:

> So does ditching _xmlplus even do the right thing?  Is there so little
> discssion about this here because so few care?

As a PyXML user, I'm +0 on leaving it as-is: it works for me. Forcing me to
import from somewhere else breaks my code.

I can't, as a developer, completely ignore _xmlplus, though: if I want to
sniff for presence and/or version of PyXML I have to explicitly look for
_xmplus.__version__: xml.__version__ in a base Python distribution exists,
and is higher, than xml.__version__ in a base distribution + PyXML.

At the moment, though, I only do this in my email-me-the-traceback exception
handler -- generally it's easier and better to simply attempt to import what
I need from xml and raise the user an "install PyXML" if the import fails.

Finally, I'd always assumed that the xml-aliases-to-_xmlplus business was a
trick to allow PyXML to eventually _become_ the standard Python xml library
with minimal impact on early adopters. Am I way off base here?

--
James Kew
james.kew@btinternet.com