strip trailing dots in goto/datatip (and rename)

[aviatesk]

While implementing #203, I had to implement the logic to remove trailing dots and found that this is better to be merged as a separate feature, since the logic can be used for goto/datatip.

In short, if we have a word like

Atom.SYMBOLSCACHE.age

and then, we can get goto and datatip for:

• Atom when our cursor is anywhere on Atom
• Atom.SYMBOLSCACHE when our cursor is anywhere on SYMBOLSCACHE
• no datatip if our cursor is on age (the field)

(c.f.: The current behaviour: we can't get any goto and datatip at all)

---

Requires JunoLab/atom-julia-client#651

[codecov]

Codecov Report

Merging #208 into master will increase coverage by 0.19%.
The diff coverage is 96.96%.

@@            Coverage Diff             @@
##           master     #208      +/-   ##
==========================================
+ Coverage   40.03%   40.22%   +0.19%     
==========================================
  Files          40       40              
  Lines        1856     1857       +1     
==========================================
+ Hits          743      747       +4     
+ Misses       1113     1110       -3

Impacted Files	Coverage Δ
src/utils.jl	86.2% <ø> (ø)	⬆️
src/display/methods.jl	86.66% <100%> (ø)	⬆️
src/goto.jl	81.81% <100%> (+1.02%)	⬆️
src/datatip.jl	79.68% <90.9%> (+3.32%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 66270cc...9d80e2f. Read the comment docs.

[aviatesk]

@pfitzseb any objection on this idea ? I will make changes so that the striping would be done in the frontend, though.

[pfitzseb]

Makes sense to me, I think. Might be somewhat unintuitive, but we can probably add some clarity by highlighting the part we're using (if that makes sense).

[aviatesk]

okay, so the striping task is now all done in the frontend (basically by using different regex for the beginning and end of the word), and the awkward fullword stuff here are all removed.

Might be somewhat unintuitive

yes, maybe especially when we get goto/datatip for Mod when a cursor is on Mod.foo or something.
(But at least there are corresponding underline/highlights)

---

Still I'm not sure which behaviour do people prefer (not so many of users mayn't even care about the difference tho)

[aviatesk]

The last two commits refactor the logic in gotoglobalsymbols and update tests according to that.

Once the CI passes, I will merge this PR, since now trailing dots are stripped all in the frontend (JunoLab/atom-julia-client#651) and the changes in this PR only do the trivial refactors.

If we find the stripping trailing dots really unintuitive, we can discard/revert JunoLab/atom-julia-client#651, but the changes in this PR wouldn't get affected by that at all.