How to use xmlrpc properly with Korean (non-ascii characters)

[Martin v. Loewis]

Skip Montanaro <skip at pobox.com> writes:

> Be my guest.  I got tired of Dave Weiner's stubborn stance on several issues
> and jumped off the xmlrpc mailing list over a year ago.  He controls what is
> called "XML-RPC" very tightly.  Last time I checked, there was no movement
> on his part to allow non-ASCII strings or anything like Python's None
> object.

I know. But would he actually enforce a copyright on "XML-RPC", on the
grounds that some implementation provides an extension? IOW: let him
follow his opinion, and follow your own technical views.

>     Martin> Admittedly, that may require changing the xmlrpc libraries being
>     Martin> used, rather than putting something on top of them. However,
>     Martin> that shouldn't be too difficult, assuming that those libraries
>     Martin> somehow use an underlying XML parser.
> 
> It's not a technical issue.  It's an interoperability issue.

That's my point: It is not. If you now use base64 for character data,
you gain no more interoperability than you would when using non-ASCII
in content. The other end has to process it properly, and it would
need to apply base64-decoding in one case, interpreting the resulting
binary data then accordingly, and would just need to apply standard
XML parsing in the other case.

The resulting protocol would not be strictly XML-RPC, anymore, but it
would still interoperate.

Regards,
Martin