RFR: 8291643: Consider omitting type annotations from type error diagnostics

[Vicente Romero]

On Thu, 9 Nov 2023 00:47:10 GMT, Liam Miller-Cushon <cushon at openjdk.org> wrote:

> Hi,
> 
> Please consider this fix for [JDK-8291643](https://bugs.openjdk.org/browse/JDK-8291643), which causes javac to remove type-use annotations when formatting types in diagnostics.
> 
> For the example like the one in the bug, this change causes javac to emit the diagnostic like `incompatible types: List<String> cannot be converted to List<Number>` rather than `incompatible types: List<String> cannot be converted to List<@Nullable Number>`. Including the type annotations can be confusing, because they do not impact the type checking done by the compiler.
> 
> An alternative I considered was to remove type annotations from types for specific diagnostics, instead of doing it unconditionally in diagnostic formatting. I can revisit that if the current approach seems too broad.
> 
> The test update to `test/langtools/tools/javac/lambda/LambdaConv25.out` is because a `ForAll` is being formatted, and `stripMetadata()` uses a `StructuralTypeMapping` which rewrites away the `ForAll`.

looks good to me

I guess your patch have uncovered a bug in stripMetadata

-------------

Marked as reviewed by vromero (Reviewer).

PR Review: https://git.openjdk.org/jdk/pull/16578#pullrequestreview-1723748188
PR Comment: https://git.openjdk.org/jdk/pull/16578#issuecomment-1804901144