Sylvain ThÃnault wrote:
> done. However, does sections of rfc 2396bis match sections of rfc 3986 ?
Yes. There were only very minor editorial changes in the last drafts before
rfc2396bis became RFC 3986.
> I did found them. As I said I've run relevant tests again the restricted
> version of Uri.py and all of them pass.
Ah, OK. I wasn't sure what you meant at first.
> And I've also added the following modification to
> prepare_input_source since I send it here:
>
> @@ -510,7 +510,7 @@
> source = xmlreader.InputSource()
> source.setByteStream(f)
> if hasattr(f, "name"):
> - source.setSystemId(f.name)
> + source.setSystemId('file:%s' % f.name)
> if source.getByteStream() is None:
> sysid = absolute_system_id(source.getSystemId(), base)
> source.setSystemId(sysid)
I'm not sure without seeing it in action, but this does not look
right to me (the change, as well as its context). I need to look at
what it's doing more closely.
If you need to be lenient, be lenient with the base URI. When you
prepend 'file:' to something, you're making it be absolute, which
probably isn't what you wanted, and probably won't be ideal.
> did you take a look at those tests ?
Not yet, sorry. :) Busy.
_______________________________________________
XML-SIG maillist - XML-SIG@python.org
http://mail.python.org/mailman/listinfo/xml-sig
