Single argumentless annotation exception from rule 4.8.5 not honored

[JakeWharton]

https://google.github.io/styleguide/javaguide.html#s4.8.5-annotations states:

Exception: A single parameterless annotation may instead appear together with the first line of the signature, for example:

@Override public int hashCode() { ... }

Given code like

@Test public void stuff() {
  // ...
}

and

@Override public void run() {
  // ...
}

GJF formats them like

+@Test
+public void stuff() {
-@Test public void stuff() {
   // ...
 }

and

+@Override
+public void run() {
-@Override public void run() {
   // ...
 }

in contradiction to this exception.

While I understand why you might not want to format the own-line style into the single-line style, changing the single-line style to the own-line style when it's already in compliance with the rules seems erroneous.

I'm observing this with version 1.7.

$ cat Test.before | java -jar google-java-format-1.7-all-deps.jar - > Test.after && diff -u Test.before Test.after
--- Test.before	2019-02-06 13:19:29.000000000 -0500
+++ Test.after	2019-02-06 13:21:33.000000000 -0500
@@ -3,7 +3,8 @@
 import org.junit.Test;

 class Test {
-  @Test public void stuff() {
+  @Test
+  public void stuff() {
     System.out.println("hi");
   }
 }

[JakeWharton]

Also the latest SNAPSHOT build.

$ cat Test.before | java -jar google-java-format-1.8-20190109.235157-1-all-deps.jar - > Test.after && diff -u Test.before Test.after
--- Test.before	2019-02-06 13:19:29.000000000 -0500
+++ Test.after	2019-02-06 13:23:30.000000000 -0500
@@ -3,7 +3,8 @@
 import org.junit.Test;

 class Test {
-  @Test public void stuff() {
+  @Test
+  public void stuff() {
     System.out.println("hi");
   }
 }

[cushon]

While I understand why you might not want to format the own-line style into the single-line style, changing the single-line style to the own-line style when it's already in compliance with the rules seems erroneous.

This comment is related: #318 (comment). The formatter generally doesn't factor the existing formatting in to its decisions.

I think the key word in 4.8.5 is may. Both approaches are allowed by the style guide, and the formatter has to make a decision, and in general it prioritizes making a single consistent decision over preserving existing formatting choices.

[JakeWharton]

Yeah that's fair. And I guess detection of this case just isn't worth it?

For me, the pervasiveness of @Test and @Override as the sole annotation on a methods makes me want this optimization because of the vertical space it otherwise steals.

[cushon]

(I just cc'd you on a related internal bug.)

I think it's still up for debate whether we'd prefer to always leave annotations like @Test or @Override on the same line.