RFC: strip trailing dots

[aviatesk]

No description provided.

[aviatesk]

@pfitzseb can you make a quick review on this ?

This PR extends our getWord (and its internal -- getWordRangeAtBufferPosition) functions so that we can use different regexes for the beginning and the end of a word.

These lines were adapted from getBeginningOfCurrentWordBufferPosition and getEndOfCurrentWordBufferPosition respectively -- all the purpose is we can do the same things as those functions for a buffer point.

So the now the behavior of getWordAndRange (renamed from getWord) is like:

• if both beginWordRegex and endWordRegex options are given, use them to detect each word boundary
• if not, use wordRegex (default to our wordRegex): this is the previous behavior

As a result, we can extract head from head.field etc, which would help goto/datatip/docs, and also JunoLab/Atom.jl#203.

[pfitzseb]

Sorry, I was kinda out of commission for the last couple of days.

Tbh, I don't like the

getWordAndRange (editor, options = {
  bufferPosition: undefined,
  wordRegex: wordRegex,
  beginWordRegex: undefined,
  endWordRegex: undefined
})

API. Do you anticipate any need for user defined regexps there?
Imho it would be cleaner to just split the current "word" on .s and indicate in which the cursor is. Or maybe return a

[
  {
    text: 'Foo',
    range: [0, 2]
  },
  {
    text: 'bar',
    range: '[4, 6]
  }
]

for Foo.bar.

[aviatesk]

Sorry, I was kinda out of commission for the last couple of days.

No worries :)

Tbh, I don't like the

getWordAndRange (editor, options = {
  bufferPosition: undefined,
  wordRegex: wordRegex,
  beginWordRegex: undefined,
  endWordRegex: undefined
})

API. Do you anticipate any need for user defined regexps there?

yeah, fair enough. I made such a complex API since I thought it would be more extensible, but I'm not coming up with the use cases of the user-defined regexes other than this case for now, to be honest.

Imho it would be cleaner to just split the current "word" on .s and indicate in which the cursor is. Or maybe return a

[
  {
    text: 'Foo',
    range: [0, 2]
  },
  {
    text: 'bar',
    range: '[4, 6]
  }
]

for Foo.bar.

And so this would be better solution -- stripping trailing dots would be fairly easy with this approach, too. I will try to change the function APIs then :)

[aviatesk]

okay, e9e9845 hopefully simplifies the logic.