RFR: JDK-8246712 doclint incorrectly reports some HTML elements as empty

[Pavel Rappo]

Jon,

Having in mind my extremely limited experience with jdk.compiler, the code change looks okay. However, the test could be even better: it would be cleaner if the test were built around actual tags rather than tag trees. Otherwise, the proposed test is of "grey-box" kind as it uses the fact that {@code} and {@literal}, {@link} and {@linkplain} are implemented using the same constructs.

Now the question is: where to get an exhaustive, always-up-to-date list of inline tags? (Yes, I read and understood the below comment

* Since there is not yet any mapping between DocTree.Kind and
* the corresponding subtype DocTree (see javac Tree.Kind, Tree)
* the list of all current inline tag classes is determined by
* scanning DocTreeVisitor.)

-Pavel

> On 8 Jun 2020, at 21:29, Jonathan Gibbons <jonathan.gibbons at oracle.com> wrote:
> 
> Please review a simple change to ensure that inline tags are taken into account when checking if an HTML element is empty.
> 
> As well as testing all existing inline tags, the test also tries to avoid future instances of this problem, by testing all subtypes of InlineTagTree, and ensuring that each have a corresponding test case.
> 
> -- Jon
> 
> JBS: https://bugs.openjdk.java.net/browse/JDK-8246712
> Webrev: http://cr.openjdk.java.net/~jjg/8246712/webrev.00
>