Allow ignoring some domain in URLs

[jayvdb]

Some sites support multiple languages , and therefore a URL should be different for each language.
A good example is Wikipedia, where a English Wikipedia link in the English source should be converted to a French Wikipedia link in the French translation.

Currently that is reported as W303: different html .

[jayvdb]

See https://travis-ci.org/jayvdb/localized_docs/builds/167009605 for many examples of this.

[willkg]

First off, I had no idea anyone was using Dennis. It's nice to find out someone is using it.

That's a really interesting problem. I'll have to think about how to support that. Lint rules don't have configuration, so there's no infrastructure for pass configuration through command line arguments or files or anything to the lint rule. I'll have to think about how to do that. I'm not sure when I'll have a chance to do that, though.

In the meantime, I think you have a couple of options to alleviate the problem until we can do something in Dennis:

1. ignore the warning either by being ok with it being in the output or by excluding that lint rule from being run
2. fork Dennis and fix it in your fork however you want to fix it and use that version

I thought I had enabled plugins in Dennis, but it looks like that's covered in #2 and not actually done, yet. 😞 That would have been a good option.

Sorry I don't have better news!

[willkg]

I looked at the errors again. That's not HTML--it's reST. It shouldn't be parsed as HTML and shouldn't be kicking up HTML errors at all regardless of the URLs.

Another thing I noticed is that in many/all of those issues the msgid is reST, but the msgstr are often HTML. Even if Dennis learned reST, I don't think it'd be able to naturally deal with format translations like that.

I don't think that negates this feature request. But in your case, I'd exclude the HTML-related lint rules altogether.