On Tue, 21 Mar 2000 20:35:27 GMT,
chl(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk (Charles Lindsey) said:
If your system sees "application/xml" and has been configured to hand it
off to a program "XML" (you can certainly configure Mozilla in that way)
does the "XML" program get to see the complete message, including the
Content-Type header (or does it just get to see the body).
Unfortunately, you're allowed to make approximately zero assumptions
regarding what an MUA has been written to do algorithm-wise. I'd be quite
certain however that almost *no* MUAs would work this way. If an attachment
is an image/jpeg, *exactly* a valid JPEG file is passed to the viewer.
if it's a text/html, only the HTML is passed to the renderer.
I know of no way you could feed the headers to the XML program as
part of the input and still have valid XML. THe best you could hope
for is that the MUA (Mozilla, what have you) is able to do its OWN parsing
of the headers, and provide appropriate hints via command line switches
(similar to how, for instance, suggested filenames are passed for a
Content-Disposition: header).. This however involves code hacking on
both the MUA and the XML processor. THis is further complicated
by the fact that your XML processor may very well be a generic handler
whos source you may not control, and who's vendor may not be interested
in cluttering up the command switches for all 397 different programs
that want to pass in special tweaks... I know I'm limited to some
extent in what I can pass to Netscape if I ask it to open a window
and render HTML that I'm providing. I'll have to check back after
I actually get Mozilla to build under AIX whether the same openUrl()
hack is supported for XML...