Improve fish#Indent().

[Raimondi]

This PR addresses some problems with the current fish#indent() function:

• Fix: multiline strings should not be modified.
• Fix: "end" is always decreased by one.
• Fix: "case" and "else" are always decreased by one.
• Fix: continued lines are not treated differently.
• Fix: "if", "while", etc followed by "begin" should increment next line
  by two instead of by one.
• Fix: "case" should be one more than "switch".
• Fix: multiline strings and continued lines are considered in the
  calculations for the next line.

In this gist you can compare the results of using gg=G in a really bad indented fish file with the current and the proposed fish#indent().

[ElijahLynn]

May I ask what your improvement is? There aren't any comments and the code isn't that self documenting.

[Raimondi]

My bad, I was too tired when I finished it and created the PR. I have added comments to the code, a more explicit commit message and some text to the PR comment.

[dag]

I'm trying to work through the PR queue but this is a bit much for me to review at the moment (not having touched vim scripting for years and not having vim, git, fish or vim-fish currently set up).

However, if I remember correctly the idea was you use gq to format larger blocks of code, which will use the native fish_indent program that should probably handle every case correctly. fish#Indent() is meant to handle manual line-by-line input efficiently; is that working for you?

[dag]

See also the comments on #20. Does your changes work with endwise?

[amiller27]

@dag Is there any way this could get merged? I tested with endwise and it seems to work. The indent functionality on master is broken for me and this fixes it.

[MindTooth]

Could this be merged if @Raimondi rebased on master? Having the end living its own life is a bit annoying.