|
|
Posted by Alan J. Flavell on 03/24/06 15:52
On Fri, 24 Mar 2006, Jim Higson wrote:
> > On such a browser-like object, if you send it
> > application/xhtml+xml with a .xhtml filename extension, it offers
> > to download it; but if you send it with a .html extension (still
> > with the XHTML content-type), it does its best to render it as
> > HTML. Sigh.
>
> Just out of interest, what does it do if you send
> application/xhtml+xml with an extention like ".php"?
Just how "like" .php do you need it to be? (I couldn't - at least not
within a limited amount of fiddling - persuade our server not to parse
the file for PHP if I gave it that filename extension.)
But I wouldn't have expected PHP to have any special meaning to
MessIE. If I try some oddball filename extension, which I've also
configured in our server to be sent as the XHTML content-type,
application/xhtml+xml, then MessIE (XP SP2 IE6) offers to download
this "unknown file type".
http://ppewww.ph.gla.ac.uk/~flavell/tests/karlsruhe.karl
However, if in Windows I create[1] a new "File Type", with filename
extension .karl, and associate it with a web browser, say Firefox,
then MessIE offers the option to open the file. If I take that
option, it is evident that MessIE has downloaded a temporary copy to
file, and has then invoked Firefox to open the temporary file.
This is almost staggering towards the behaviour that the web
interworking specifications require, and what www-compatible browsers
have always implemented - what is wrong with IE is that it's going via
the filename extension - when, according to RFC2616, it is required to
honour the MIME type, without consideration of any "filename
extension" which it might discern in the URL.
h t h
[1] e.g Windows control panel> Folder options> File Types
Navigation:
[Reply to this message]
|