site_diff: Relative URLs with <base> tag not supported

[hlubek]

Some websites use a <base href="http://www.example.com/" /> tag to specify the base for relative URLs. The site_diff script currently generates wrong URLs in these cases (the relative path is appended to the current URL of the page).

[bslatkin]

Good catch! That's a tricky one. What's your expectation of that behavior?

I figure the way it would have to work is to set the prefix here:

https://github.com/bslatkin/dpxdt/blob/master/dpxdt/tools/site_diff.py#L141