-
Notifications
You must be signed in to change notification settings - Fork 4
RSS links get served as text/html #26
Comments
For what it's worth, I think most web browsers will try to provide a navigable rendered UI for XML files these days. But probably only if they can tell it's XML. |
There's definitely a difference in Chrome between the two content types: visiting /feed on my private site shows XML source, whereas /feed on the public site shows a badly mangled rendering of the XML as though it were HTML. |
(Re-reading your comment, I now think you were saying that it's okay to serve up the content as XML. I'd agree.) |
I think the issue here is that our wget hack for generating static content is not smart enough to properly handle XML. |
I'm now tempted to rewrite that in Python. |
It's possible that wget has learned some new tricks in the past two years and now has a nice way to deal with this kind of thing. |
IMHO we should use a WordPress plugin that is specifically designed for turning a WordPress site static, such as: rather than rolling our own solution that will run into WordPress-specific subtleties that we will have to reinvent solutions for. |
Assuming that's the road we want to go down, I'm certainly willing and possibly able to make time to do it. @dwrensha what do you think? |
Yeah, it'd be great if we could get an existing plugin to do the hard work for us. Last time I tried to go that route, I think I ran into difficulties regarding directory structure, and maybe I also hit sqlite-vs-mysql problems. But it seems like there ought to be a way to make it work, so I encourage you to try. |
FWIW, I also personally think that MySQL would be a fine database for the Sandstorm WordPress package, to maximize compatibility. That's a separate topic to discuss on a separate issue perhaps. |
This issue was moved to sandstormports/wordpress-sandstorm#20 |
On my private site, if I visit /feed/, I get a response content-type of application/rss+xml. On the public site, that link gets rewritten as /feed/index.html and served as text/html. This seems weird. I can't tell if this is bad or not, aside from causing Chrome to try to render it.
The text was updated successfully, but these errors were encountered: