[2.7.6] Big regression with indentation because of the fix for ternaries

[boris-petrov]

Congrats on the new release @jtkiesel! However, there's a nasty regression in certain cases.

Prettier-Java 2.7.6

Playground link 1
Playground link 2

--print-width 140

Input:

public abstract class Foo implements MyInterface {
  @Override public String foo() {
    return true ? "" :
		foo.method2(r ->
		  Context.a().m(1, a -> {
        code;
      })
		);
  }
}

Output:

public abstract class Foo implements MyInterface {

	@Override
	public String foo() {
		return true
			? ""
			: foo.method2(r ->
					Context.a().m(1, a -> {
						code;
					})
				);
	}
}

Expected behavior:

public abstract class Foo implements MyInterface {

	@Override
	public String foo() {
		return true
			? ""
			: foo.method2(r ->
				Context.a().m(1, a -> {
					code;
				})
			);
	}
}

The two playground links show regressions for these two previous cases (but this time in a ternary):

#764
#760

I'm not sure there aren't more cases that regressed.

[jtkiesel]

This isn't exactly a regression, but an alignment with the way Prettier formats ternaries in JavaScript/TypeScript (example).

Prettier attempts to align the closing parenthesis with the start of the expression, which is indented by 2 characters due to the : and space preceding it. When spaces are used for indentation, it can do so successfully no matter the tab width, however if tabs are used it simply indents by a single tab, as you noticed.

[boris-petrov]

Oh... OK. I didn't know that. I don't think it makes sense for them to do it that way but I guess that's that. :) Thanks for this and sorry for the fuss!

[boris-petrov]

But... that's extremely bad I think. Look at what prettier does - link. You are consistent which is great! But... how does it make sense for the output to be like that?! HALF of the comment is indented. Do you think that's a bug in Prettier?