Hi Joe,

On 4/08/2020 2:24 am, Joe Darcy wrote:
Hi David,

To provide a bit more context, the API was consciously designed to be future-proofed against the possibility of Java supporting multiple bounds here (JDK-4891872). As of yet, that change has not come to pass,

Okay so we want to keep the API. :)

and as you note may never come to pass. So besides in addition to the links into the JLS, I think it is reasonable to include an API note acknowledging the situation. How about:

    This method supports the existence of multiple bounds, but to date, a wildcard may have at most one upper/lower bound.

"to date" is just another form of "in the current version".

Sorry I find it hard to suggest suitable text when I don't think it necessary to say anything.

Best to let others endorse this one.

Cheers,
David
-----

-Joe


On 8/2/2020 7:22 PM, David Holmes wrote:
Hi Joe,

I was under the impression that using phrases like "the current version of" were superfluous as all such documentation implicitly refers to the current version, unless explicitly stated otherwise.

Personally I'm not sure this clarification really makes a difference to anything as the main source of confusion with this API is why it states "bounds" rather than "bound" and thus returns an array. By adding this clarification you will require a future update if the language allows multiple bounds, forcing the docs to carry around historical data "prior to Java N only a single bound were allowed but since then multiple bounds ...". This seems unnecessary clutter to me. Given the ship has sailed on the shape of this API the developer must simply embrace it and always process the return value in a loop without assuming any knowledge of how many bounds may actually exist.

I think the reference to the JLS section suffices to instruct the developer on where to find out how many bounds to expect, whatever that may be in their current version of the JLS.

If we think we will never support multiple bounds then these methods should be deprecated (for removal) and replaced by a singular counterpart that returns a single Type rather than a Type[].

Cheers,
David
-----

On 1/08/2020 8:12 am, Joe Darcy wrote:
Hello,

Please review the doc clarification below to address

     8250660: Clarify that WildcardType and AnnotatedWildcardType bounds methods return one

Thanks,

-Joe

diff -r b0553ba43ba1 src/java.base/share/classes/java/lang/reflect/AnnotatedWildcardType.java --- a/src/java.base/share/classes/java/lang/reflect/AnnotatedWildcardType.java Fri Jul 31 11:35:25 2020 -0700 +++ b/src/java.base/share/classes/java/lang/reflect/AnnotatedWildcardType.java Fri Jul 31 15:12:14 2020 -0700
@@ -1,5 +1,5 @@
  /*
- * Copyright (c) 2012, 2015, Oracle and/or its affiliates. All rights reserved. + * Copyright (c) 2012, 2020, Oracle and/or its affiliates. All rights reserved.
   * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   *
   * This code is free software; you can redistribute it and/or modify it
@@ -30,6 +30,7 @@
   * wildcard type argument, whose upper or lower bounds may themselves represent
   * annotated uses of types.
   *
+ * @jls 4.5.1 Type Arguments of Parameterized Types
   * @since 1.8
   */
  public interface AnnotatedWildcardType extends AnnotatedType {
@@ -39,6 +40,9 @@
       * If no lower bound is explicitly declared, the lower bound is the
       * type of null. In this case, a zero length array is returned.
       *
+     * @apiNote In the current version of the Java Programming
+     * Language, a wildcard may have at most one lower bound.
+     *
       * @return the potentially annotated lower bounds of this wildcard type or
       * an empty array if no lower bound is explicitly declared.
       * @see WildcardType#getLowerBounds()
@@ -50,6 +54,9 @@
       * If no upper bound is explicitly declared, the upper bound is
       * unannotated {@code Object}
       *
+     * @apiNote In the current version of the Java Programming
+     * Language, a wildcard may have at most one upper bound.
+     *
       * @return the potentially annotated upper bounds of this wildcard type
       * @see WildcardType#getUpperBounds()
       */
diff -r b0553ba43ba1 src/java.base/share/classes/java/lang/reflect/WildcardType.java --- a/src/java.base/share/classes/java/lang/reflect/WildcardType.java Fri Jul 31 11:35:25 2020 -0700 +++ b/src/java.base/share/classes/java/lang/reflect/WildcardType.java Fri Jul 31 15:12:14 2020 -0700
@@ -1,5 +1,5 @@
  /*
- * Copyright (c) 2003, 2015, Oracle and/or its affiliates. All rights reserved. + * Copyright (c) 2003, 2020, Oracle and/or its affiliates. All rights reserved.
   * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   *
   * This code is free software; you can redistribute it and/or modify it
@@ -29,6 +29,7 @@
   * WildcardType represents a wildcard type expression, such as
   * {@code ?}, {@code ? extends Number}, or {@code ? super Integer}.
   *
+ * @jls 4.5.1 Type Arguments of Parameterized Types
   * @since 1.5
   */
  public interface WildcardType extends Type {
@@ -45,6 +46,9 @@
       *  <li>Otherwise, B is resolved.
       * </ul>
       *
+     * @apiNote In the current version of the Java Programming
+     * Language, a wildcard may have at most one upper bound.
+     *
       * @return an array of Types representing the upper bound(s) of this
       *     type variable
       * @throws TypeNotPresentException if any of the
@@ -69,6 +73,9 @@
       *   <li>Otherwise, B is resolved.
       * </ul>
       *
+     * @apiNote In the current version of the Java Programming
+     * Language, a wildcard may have at most one lower bound.
+     *
       * @return an array of Types representing the lower bound(s) of this
       *     type variable
       * @throws TypeNotPresentException if any of the
@@ -78,6 +85,4 @@
       *     for any reason
       */
      Type[] getLowerBounds();
-    // one or many? Up to language spec; currently only one, but this API
-    // allows for generalization.
  }

Reply via email to