fix: break long assignments after equals

[jtkiesel]

What changed with this PR:

Long assignments (i.e. those that previously violated printWidth) of many types now break after equals the way Prettier JavaScript now does.

Example

Input

class Example {

  void example() {
    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes = new Object()
      .someMethod();

    Object[] aParticularlyLongAndObnoxiousNameForIllustrativePurposes2 = new Object[] {
      new Object(),
      new Object()
    };

    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes3 = SomeClass.someStaticMethod();

    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes = someMethod()
      .anotherMethod();

    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes = anotherValue !=
      null
      ? anotherValue
      : new Object();
  }
}

Output

class Example {

  void example() {
    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes =
      new Object().someMethod();

    Object[] aParticularlyLongAndObnoxiousNameForIllustrativePurposes2 =
      new Object[] { new Object(), new Object() };

    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes3 =
      SomeClass.someStaticMethod();

    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes =
      someMethod().anotherMethod();

    Object aParticularlyLongAndObnoxiousNameForIllustrativePurposes =
      anotherValue != null ? anotherValue : new Object();
  }
}

Relative issues or prs:

Fix #527

[CLAassistant]

All committers have signed the CLA.

[jtkiesel]

@clementdessoude I don't mean to sound impatient, so I apologize in advanced, I just wanted to make sure that I have done everything I needed to do for this pull request. I'm not sure if there's anything I'm missing, or if anyone else handles pull request reviews/merges. Any chance this could get a (hopefully quick) review?

[clementdessoude]

I think that with this implementation, some (oddly placed) comment might disapear. For instance, I think that in one of these two statements, the comment will disapear, as it is attached to the ctx.Equals[0] symbol.

private Map<Integer, String> genericVariable4 /* comment */= new HashMap<Integer, String>(); or private Map<Integer, String> genericVariable4 =/* comment */ new HashMap<Integer, String>();

Can you try to add this test case and check if it is the case ? If the comment disappears, you can try to replace " =" by " ", ctx.Equals[0], here and in the three declarations afterwards

[jtkiesel]

Thanks for catching this! I thought there might be something I was missing. I tested the two scenarios you presented, and the first case (comment before equals) caused the comment to disappear, while the second care (comment after equals) didn't. (I assume in the 2nd case, the comment becomes "attached" to the value, rather than the equals.) I just pushed a fix for it.

Comparing the way that comments are printed to Prettier JavaScript, Prettier Java seems to be a bit different (Prettier JavaScript always separates comments from other elements with a single space, while Prettier Java omits the space), but I confirmed that the comment printing behavior isn't changed by this PR (anymore).

[clementdessoude]

Once again, I'm so sorry for the very long delay in the response. Thanks for pinging me back again !

I have a small comment, let me know if you can tackle it. Otherwise, I should be able to test it in the evening or tomorrow.

[jtkiesel]

No worries at all, I'm sorry if I came off as impatient or rude. Thank you for taking the time to review!

[clementdessoude]

Well done ! I had just one doubt I wanted to clear up before merging, but everything looks fine ! Thanks a lot @jtkiesel !