RFR: made maven resource plugin and (un)setDetailedOutputOptions expclicit [v4]

[Jiří Vaněk]

On Tue, 9 Jan 2024 13:11:03 GMT, Jiří Vaněk <jvanek at openjdk.org> wrote:

>> That is still sufferign the https://issues.apache.org/jira/browse/MRESOURCES-237 so the filtering was done. Fitlering is excluding proeprties (as they claim they are ascii) and is adding .class sufix for test classes.  Another solution is to use mrp 2.6 which do not suffer the symlinks issue
>> 
>> In both cases, the tests still fail in maven runs, although rummor is, manually it passes (todo, verify)
>
> Jiří Vaněk has updated the pull request incrementally with one additional commit since the last revision:
> 
>   by properly handling static-wide (un)setDetailedOutputOptions fixing case7903558.tests

So the first test-failure was passing in ide and not from cmdline, becasue it was order sensitive.

The Options.(un)setDetailedOutputOptions and its underlying  static private EnumSet<PR> printOptions = JASM; are unlucily static fields, co currently shared over environment, so in world of tests very dangerous.

To move them out of static scope shoud be separate from this PR (if ever)


While doing this, I had removed the reflexive wrapper. it was literally duplicated code  with already existing methods which were based on simple if in two values. The usage of refelction without proper apis made no sense to me, and was making adaptation to (un) setDetailedOutputOptions harder (as it was just Object). i had chosen to remove it instead of casting the output.

If you insists on proper reflexion-based solution, I would be happy to do so.  It will howevere requrire some more interfaces like Compilable, Decompilable and/or Mainable (weird, but serve its purpose). 

There is still one failing test. Am on it. Sorry for delay, and happy new year to whole jasmtools team.

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

PR Comment: https://git.openjdk.org/asmtools/pull/68#issuecomment-1883038199