RFR: 8254965: Remove unused Matcher::_ruleName and make ruleName non-product

[Claes Redestad]

On Mon, 19 Oct 2020 17:07:46 GMT, Vladimir Kozlov <kvn at openjdk.org> wrote:

> > Does that print really need to print the name of the rule - or would just printing `mach->rule()` suffice? For
> > sustainability purposes I suppose having the name spelled out is a nice to have rather than a strict necessity.
> 
> mach->rule() value is only available in files generated from .ad file during build. If you don't have those files
> available it will be hard to map the value to mach instruction. You have to rebuild VM again and for that you need to
> have exact sources as one failed. It is doable but hassle.

A hassle surely, but sustaining engineers routinely must set up exact source to debug an issue. So I think it's a
judgement call, and if you prefer keeping it in I'm ok with that: reducing static footprint is nice, but a lump of
things like this should never be paged in anyhow.

A side-note: Matcher has several of these fields that only take a pointer to a .ad-defined array, e.g., _ruleName,
_must_clone, _swallowed, ... - does this do any good, or could the code refer to the static globals directly? If only
used for debugging or printouts like _ruleName it can surely go, but how about the others?

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

PR: https://git.openjdk.java.net/jdk/pull/722