rework sentence stop detection, support blacklist configuration

[dirk-thomas]

This patch reworks the sentence stop detection. It addresses many tickets at once (with one "regression" though):

• Check for different sentence stop characters (., ?, !) : #6
• Only consider dots which are followed by a space
  • Fixes #3: file, URL
  • Fixes #8: Decimal numbers
  • Fixes #11: Jekyll tags
  • This fails one of the tests since it breaks the current behavior that a dot without a space afterwards is detected "correctly". I think the pros are more significant then the con. The behavior could also be exposed as an option?

• Support a blacklist option so anything can be explicitly excluded

  • Fixes #5: etc.
  • Fixes #7: i.e., .e.g.
  • Fixes #9: ...
  • Fixes #10: Mr., Mrs.
  • Example config in .remarkrc:

  "external": ["mdast-lint-sentence-newline"],
  "sentence-newline": {
    "blacklist": ["e.g.", "E.g.", "etc.", "i.e.", "vs.", "..."]
  },

  • Currently the comparison is case sensitive. Maybe it should be case insensitive?

[why-jay]

Thanks so much! I'll need some time to review this.

[why-jay]

I added tests against your algorithm, and everything looks good! Thanks for the work.

I think the pros are more significant then the con.

Definitely.

The behavior could also be exposed as an option?

Not for now - I'm too lazy to do the work and I imagine there's no demand. But if there's demand, I will add the option :)

Maybe it should be case insensitive?

Again, this could be an option. If there's demand :)