[m-rev.] diff: fix the behaviour of math.round/1 in the Java grade

[Julien Fischer]

Fix the behaviour of math.round/1 in the Java grade.

library/math.m:
     Implement the documented behaviour for infinities for
     math.round/1 in the Java grade.

Julien.

diff --git a/library/math.m b/library/math.m
index 74cf915..db583da 100644
--- a/library/math.m
+++ b/library/math.m
@@ -455,7 +455,14 @@ e = 2.7182818284590452353602874713526625.
      round(Num::in) = (Rounded::out),
      [will_not_call_mercury, promise_pure, thread_safe],
  "
-    Rounded = java.lang.Math.round(Num);
+    // For +/- infinity Java's round method will return Long.{MAX,MIN}_VALUE.
+    // This does not match the documented Mercury behaviour for round with
+    // infinities.
+    if (java.lang.Double.isInfinite(Num)) {
+        Rounded = Num;
+    } else {
+        Rounded = java.lang.Math.round(Num);
+    }
  ").
  :- pragma foreign_proc("Erlang",
      round(Num::in) = (Rounded::out),