Suggestion regarding XHTML handling
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Might I suggest actually doing basic namespace handling when a page is delivered using an XML MIME type? For example, this page: <http://aggregator.singpolyma.net/feeds/denver/> contains both XHTML and embedded (namespaced) Atom, but instead of seeing the Atom as something it cannot process or otherwise handling it, the rendering engine just turns all the tags into inline flow content and shows the text contents in what is essentially line noise. I know that in a "normal" HTML page you want the heuristics to do this sort of thing in case some bozo has dumped something that maybe looks like a namespace into the tag soup (and namespaces aren't officially supported in HTML anyway), but if the MIME-type is an XML one, the UA should be able to process it as XML and ignore or otherwise handle namespaces it doesn't understand. - -- Stephen Paul Weber, @singpolyma See <http://singpolyma.net> for how I prefer to be contacted edition right joseph -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iQIcBAEBCAAGBQJPbUV7AAoJENEcKRHOUZze+aMQAKDSNe3QZzzsIFYZkH2fvWE+ 03gRZy7XtwzqQMBQENAuZgK1orUGR8KO8t69re7f+GFTed36lE5QrqtY2tKPmtrl BD0W7d0bbZpAzIk0akS44l9+ICy2X+iKMxhdzeELcEABxlRLqfgYwS3zZ+XjvOyK VGOj8ejDs1n4G2SO6IC28DXJxa8cijkhfN7QzAO1rUpHSlPV8xO6vCIOultr05Nc VoybKhBTetDV4FsRX4hpwd5/ToAl6JYWvkwaSEny7Pc30H8EXthIzASnf+K2rPwh v31aNWJgrf6DwXhnX1SZLYYFkePpuiLzOAWGbljacJlnkI9r6guXDUIq0U0mCdLD G1fuvnaLCSv78yz7MwDS/hDwxx7rElXk3w7BbwzA71tbuJocBTq3clX4Ho+dL84L +pYCXWjWbt5BPiIgHDiPzj+48AS4kvW9y4a/yCPoAJFExZDdoNdSlFGUvGmc2xwa 9Rcvz7lX6DQ5iZXXBLpVi5IrwMFEaXKd+0dUbsaXD30a6n6r+8o6sDMSh+CIDu20 dVhB58ff7j7cwVsPSQtHZmkjq4x3KxMiIm/gtYijkxh48lsUwMiyIBpw76gUZDPp N22Q4c5bGiHWRUhYHTwuu1C04Dx0wHeWoaKDGaPV9YyEyPq1EiZZTz2Vzd7AMbUT K1x3M4Fep0SUuLsVusle =ldnT -----END PGP SIGNATURE-----
On Fri, Mar 23, 2012 at 10:54:35PM -0500, Stephen Paul Weber wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
Might I suggest actually doing basic namespace handling when a page is delivered using an XML MIME type? For example, this page: <http://aggregator.singpolyma.net/feeds/denver/> contains both XHTML and embedded (namespaced) Atom, but instead of seeing the Atom as something it cannot process or otherwise handling it, the rendering engine just turns all the tags into inline flow content and shows the text contents in what is essentially line noise.
I know that in a "normal" HTML page you want the heuristics to do this sort of thing in case some bozo has dumped something that maybe looks like a namespace into the tag soup (and namespaces aren't officially supported in HTML anyway), but if the MIME-type is an XML one, the UA should be able to process it as XML and ignore or otherwise handle namespaces it doesn't understand.
I don't see what could be done differently here. Firefox renders it pretty much the same way as Dillo does. Am I missing something? Cheers, Johannes
Somebody claiming to be Johannes Hofmann wrote:
On Fri, Mar 23, 2012 at 10:54:35PM -0500, Stephen Paul Weber wrote:
Might I suggest actually doing basic namespace handling when a page is delivered using an XML MIME type? ... instead of seeing the Atom as something it cannot process or otherwise handling it, the rendering engine just turns all the tags into inline flow content and shows the text contents in what is essentially line noise.
I don't see what could be done differently here. Firefox renders it pretty much the same way as Dillo does. Am I missing something?
Yes, but that doesn't mean dillo can't be better than firerox :) Stuff from a namespace we understand (right now just XHTML, but in the future maybe SVG, MathML, whatever. or not) should get processed according to the rules of that namespace. Namespaces we don't understand, and for which we have no processing rules, should be ignored completely. And, I feel that if a server is serving with an XML MIME-type, we can trust that they expect XML-style processing rules like this, since tagsoup sites are not served with such a MIME. -- Stephen Paul Weber, @singpolyma See <http://singpolyma.net> for how I prefer to be contacted edition right joseph
participants (2)
-
Johannes.Hofmann＠gmx.de -
singpolyma＠singpolyma.net