text/html with mode=xml in Atom 0.3
I've been seeing a number of feeds recently using Atom 0.3 with a content type of text/html and no mode attribute (i.e. the equivalent of mode=xml). However, the markup in that content is wrapped in a CDATA section, for example something like this: content type=text/html ![CDATA[div xmlns=http://www.w3.org/1999/xhtml;pContent goes here./p/div]] /content If it had been marked as escaped you would obviously unescape the CDATA before interpreting the markup. However, since the mode is technically xml, I was under the impression that it should be treated as inline XML and no unescaping was necessary. But that would result in the literal text div xmlns=http://www.w3.org/1999/xhtml;pContent goes here/p/div being displayed to the user which is obviously not what is intended. So is this a bug in the content generator (all the feeds I've seen appear to be using TypePad) or are you supposed to ignore the mode attribute when the content type is set to text/html and always treat it as escaped? I know Atom 0.3 is deprecated and I shouldn't be having to deal with this, but the reality of the situation is that there are a whole lot of Atom 0.3 feeds still out there (probably more than Atom 1.0) and I need to be able to support them. Some feeds where you can see the problem (not all entries though): http://feeds.feedburner.com/Flickrblog http://dilbertblog.typepad.com/the_dilbert_blog/atom.xml http://blog.cymfony.com/atom.xml Regards James
Re: text/html with mode=xml in Atom 0.3
* James Holderness [EMAIL PROTECTED] [2006-03-23 17:30]: So is this a bug in the content generator (all the feeds I've seen appear to be using TypePad) Yes. or are you supposed to ignore the mode attribute when the content type is set to text/html and always treat it as escaped? No. In 0.3, the `mode` attribute was the final arbiter for the form of the content. In Atom 1.0, its role was subsumed by switching on the `type` value because consumer developers reported that this sort of layering was unnecessarily hard to support and provided no discernible benefit. Regards, -- Aristotle Pagaltzis // http://plasmasturm.org/
Re: text/html with mode=xml in Atom 0.3
A. Pagaltzis wrote: So is this a bug in the content generator (all the feeds I've seen appear to be using TypePad) Yes. or are you supposed to ignore the mode attribute when the content type is set to text/html and always treat it as escaped? No. Thanks for the confirmation. I was beginning to think I was wrong. I tested this in 15 different aggregators and all but one ignored the mode and unescaped the content anyway. I have a horrible feeling I'm going to have to add code to emulate this behaviour. Regards James
Re: text/html with mode=xml in Atom 0.3
* James Holderness [EMAIL PROTECTED] [2006-03-23 18:40]: I tested this in 15 different aggregators and all but one ignored the mode and unescaped the content anyway. Good thing this rule was changed in Atom 1.0, then… What I really don’t get is what that `xmlns` attribute is doing there in the CDATA block of your data sample. Sometimes I wonder if CDATA should not have been left out of the XML spec; it seems to create far too much confusion to be worthwhile. Regards, -- Aristotle Pagaltzis // http://plasmasturm.org/
Re: text/html with mode=xml in Atom 0.3
A. Pagaltzis wrote: What I really don’t get is what that `xmlns` attribute is doing there in the CDATA block of your data sample. Sometimes I wonder if CDATA should not have been left out of the XML spec; it seems to create far too much confusion to be worthwhile. Well if you look at some of those feeds I listed, many of the entries are type=application/xhtml+xml with a namespaced div element as you would expect. It looks like they may have taken the exact same code (or template, or however it is they do this stuff) and reused it for type=text/html. Only with the html they decided they should wrap everything in a CDATA block just to be safe. Regards James