For server accesses (such as skins) $basepath will stay the same anyways.
However, when the skin tries to access its resources (CSS, images…) it will call skinbase()… this should probably become a method of the Blog object.

I think blogs should have basically two modes:

- NORMAL: the blog is INSIDE of the $baseurl . In this case, $Blog->siteurl is a subpath of $baseurl

- INDEPENDANT: the blog can be on any URL. In this case, $Blog->siteurl is a full URL.

The Blog object should handle all of the abstraction due to this.

Blog pages/skins should not rely on $baseurl anymore but on Blog object methods.

I was actually just discussion something like this with Jason this morning. All you need to do is generate the URLs off of HTTP_HOST and they can operate on completely different domains.

Of course, it might be a good idea to implement a way to keep track of which domains can mask over a given blog. That would keep people from redirecting to your site with some sort of man in the middle type security exposure.

in addition, just using HTTP_POST will generate cookie domain issues. We have learned from our mistakes that we need to make links to the canonical URL.

I’m all for this one. I run a small hosting business with a few personal blogs and it sure would be nice to be able to allow users to have their URL in the blog URL.

Thanks for a great app!

–Chris