Multi-line arrow functions

[ksmithbaylor]

Currently, when the body of an arrow function that returns an immediate value is split over multiple lines, the second line (the value) is aligned even with the line above:

const someVeryLongIdentifier = (a, b, c) =>
anotherVeryLongExpressionThatWillNotFitOnASingleLineNicely();

nextLine();

In my opinion, it should be indented one level more than the line on which it started, but still allow the following lines to stay at the same level as before:

const someVeryLongIdentifier = (a, b, c) =>
  anotherVeryLongExpressionThatWillNotFitOnASingleLineNicely();

nextLine();

This should also apply to arrow functions defined as function arguments. Current behavior:

someVeryLongIdentifier.getThings().filter(Boolean).map(x =>
                                                       anotherVeryLongExpressionThatWillNotFitOnASingleLineNicely());

                                                       nextLine();

Expected behavior:

someVeryLongIdentifier.getThings().filter(Boolean).map(x =>
  anotherVeryLongExpressionThatWillNotFitOnASingleLineNicely());

nextLine();

[wchargin]

The following patch works for me. Can someone familiar with the repo make sure this seems reasonable? If it does, I'll pull-request it.

diff --git a/indent/javascript.vim b/indent/javascript.vim
index a935a87..f1329cb 100644
--- a/indent/javascript.vim
+++ b/indent/javascript.vim
@@ -55,7 +55,7 @@ let s:continuation_regex = '\%([\\*+/.:]\|\%(<%\)\@<![=-]\|\W[|&?]\|||\|&&\|[^=]
 " TODO: this needs to deal with if ...: and so on
 let s:msl_regex = s:continuation_regex.'|'.s:expr_case

-let s:one_line_scope_regex = '\<\%(if\|else\|for\|while\)\>[^{;]*' . s:line_term
+let s:one_line_scope_regex = '\%(\<\%(if\|else\|for\|while\)\>\|=>\)[^{;]*' . s:line_term

 " Regex that defines blocks.
 let s:block_regex = '\%([{[]\)\s*\%(|\%([*@]\=\h\w*,\=\s*\)\%(,\s*[*@]\=\h\w*\)*|\)\=' . s:line_term

I just added it to the one-liners regex. I had to put it in another non-capturing group so that the => was separate from the \>, because > doesn't count as a word boundary.

[davidchambers]

This looks reasonable to me. Also, none of the current maintainers is familiar with the indentation-related code, so you may be the closest thing we have to an expert. ;)

[wchargin]

Expert? That's a scary thought! :)

I opened #339.

[davidchambers]

Resolved by #339