On Wed, 17 Jan 2024 00:24:01 GMT, Jonathan Gibbons <j...@openjdk.org> wrote:
>> src/jdk.compiler/share/classes/com/sun/source/doctree/RawTextTree.java line >> 40: >> >>> 38: * @apiNote >>> 39: * This class may be used to represent tree nodes containing >>> 40: * {@linkplain DocTree.Kind#MARKDOWN Markdown} text. >> >> This means that there is one-to-many relationship between `RawTextTree` and >> `DocTree.KIND`. This in turn perpetuates the pattern of checking the kind >> followed by casting as opposed to more modern `instanceof` pattern matching. >> There's nothing wrong with it per se, however I wonder what the rationale is >> for leaving this part of the API open-ended. Is it to support other types of >> raw text in the future? > > Yes, the thinking was to allow it to be open-ended. > > In this case, while I think we should specify the ability to use > `instanceof`-style checks with this API, I do not think it would be necessary > ... there is no need for a subtype for `MarkdownTree`, nor would I expect > there to be one for any other kind of raw text. Related: filed [JDK-8324877](https://bugs.openjdk.org/browse/JDK-8324877) to better specify the 1-1 relationship between subtypes of `DocTree` and members of `DocTree.Kind` ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/16388#discussion_r1470351155