Tim Bray: Sam Ruby tells me there are some lingering corner-case bugs; report ’em and I bet they’ll fix ’em
Let’s find out:
link/@href: if xml:base is not present, the default base in effect is the URI of the document itself. Example.
author: these elements seem to be ignored — they used to be supported. Example.
source: various feed source elements (e.g., title) seem to override the equivalent entry elements. Example.
updated: this elements indicates when an entry was modified in a way the publisher considers significant Example. (compare the updated values in the feed to the updated values displayed by Bloglines)
Extra credit: any chance of supporting inline SVG? Example. MathML? Example.
I will say that it feels strange to find Bloglines more responsive than Google Reader, Google Sitemaps, and Blogger. (Hint, hint, nudge, nudge)
NNW Lite 2.1 also isn’t too happy with the lack of xml:base. It thinks the URL of this entry is “http://./2006/07/08/Bloglines-Edge-Cases”.
The extract method is basically done. I’m sure it could be improved a bit more, but it seems to be fairly effective. I added a few extra features beyond the original URI class’s capabilities, such as supplying a base uri to resolve...
If any of these tools can provide me with a credible schedule for when they will fix their bugs, I may be willing to accommodate. In fact, I did so before.
- link/@href: Not Fixed yet, but I expect it will be within a day or two.
- author: Fixed.
- source: Fixed.
- updated: Fixed.
- MathML / SVG Support: Not Yet. We are looking into it, MathML seems like it might be easier, but SVG will be difficult / not very useful unless we could better parse inline CSS chunks without causing security issues.
The example from Tantek’s Updates is also fixed.
For some of these, they might not be ‘fixed’ in old posts, due to how our parser works, but in new posts going forward, it should be correct.
The hard part of MathML and SVG isn’t CSS, but rather the requirement that the page produced must be well formed XHTML, and with modern browsers that means no document.write. To my knowledge, I am the only person who has ever attempted to support such a thing — using a combination of Planet, the UFP, and BeautifulSoup along with a number of patches. Perhaps someday MathML can be used inside of HTML tag soup, but today is not that day.
I’ve verified a number of the items you reported already; for others, I imagine I may need to wait until there is enough “fresh” content, but if you follow through with xml:base’s default (and perhaps look in the the issue with category), I’d be a happy camper.
Sam, I just double-checked that I’m on the latest NNW Lite, and I’m still seeing the buggy behavior. Some preliminary poking around on my lunch break didn’t reveal any obvious pointers to an actual bug database, but I’ll have a look around their forum when I get home from work.
I’ve given Bloglines a fair amount of grief over the past few months over their pathetic-at-the-time handling of Atom feeds. I’m not ego-centric enough to believe that I got them to change – at most, I may have increased awareness of...
[more]
The extract method is basically done. I’m sure it could be improved a bit more, but it seems to be fairly effective. I added a few extra features beyond the original URI class’s capabilities, such as supplying a base uri to resolve...