RFR: 8229186: Improve error messages for TestStringIntrinsics failures [v5]

[Evgeny Nikitin]

On Wed, 30 Sep 2020 16:34:10 GMT, Igor Ignatyev <iignatyev at openjdk.org> wrote:

> I think we do, as people tend to first look at exception's messages and only then look thru other logs, so having
> relative information in the exception is always a good thing. should 10K symbols become a problem (this though would
> also mean that you can't print the compared values w/ `Format.asLiteral`), we can revisit this.

Ok, I got it. I was blind and stupid, sorry. Fixed, please check the e6fb6d04cad

> hm, I somehow missed that usage, but you don't need to repeat to the same switch over a component type in
> `ArrayDiff::of`, do you?

<sigh>... I can see no better solution here. The `ArrayDiff::of` checks that both component types are the same in
addition, but I probably could have avoided duplicating the switch (we can ignore double guessing the type). The issue
is that <E> in ArrayDiff is forwarded to first and second, I would need to declare them as generic
(ArrayCodec<Object>). The type system seem to not allow me to fix that nicely.

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

PR: https://git.openjdk.java.net/jdk/pull/112