Add support for site.baseurl, workaround

[wgmyers]

When using html-proofer on a Jekyll site with non-empty baseurl such as /test, all internal links are reported as broken.

It would be great to be able to specify something like --internal_baseurl "BASEURL" on the command line, and then have the checker prepend the value of BASEURL to all internal links. This would also locate instances where BASEURL was incorrectly not prepended, and be extremely useful to Jekyll users.

Others have found a workaround with the url_swap option, but it does not allow replacing eg /test with the empty string or with /

$ htmlproofer --assume-extension --disable_external --url_swap "/test\:/" ./_site
htmlproofer 3.18.2 | Error:  undefined method `split' for /test\:/:Regexp

The workaround working for me is:

$ cd _site
$ ln -s . test
$ cd ..
$ htmlproofer --assume-extension --disable_external ./_site

[kylezervas]

ok thanks im going to try and farm this in to my on

[PeterJCLaw]

Not sure if this has since been fixed, but I'm finding that this url swap works for this case: --url-swap "^/website/:/". That's to line up with a Jekyll setting of baseurl: /website.

I think the issue in the original example is possibly the use of \: rather than : to separate the original from the replacement?

The docs on this could definitely be clearer though -- I figured out my url swap by debugging the code.

[wgmyers]

@PeterJCLaw Thank you. You are quite right: I was using duff syntax, and the solution you give works for me also.

I have no idea what I thought I was doing with --url_swap "/test\:/" , but --url_swap "^/test/:/" is the answer.