Improve Cvss.from() parsing performance

[anderruiz]

What does it do?

Improve parsing performance by optimizing several parts:

• Use regex groups to do the parsing (that we are doing anyway) to avoid have to use StringTokenizer + split
• Use equals in enums instead of equalsIgnoreCase as it is already impossible based on the regexes
• Do not use all the regexes at the same time (at the initial part of the method) and reorder them based on the probability of success (V3 more likely that V2).

The performance was measured with a JMH test that I did not push, but that I could if it can be interesting

Update

There is an improvement made on top of this PR here, that basically replace the usage of String (whenever possible) by chars which ends in a faster parsing and serialization (getVector). I also included additional test for the parsing part. I would vote for using the char based one, but we can split it in two different PRs if needed

[anderruiz]

@stevespringett any advance with this?

[stevespringett]

LGTM

[stevespringett]

Let me know if you'd like a release or if there's another performance optimization you'd like to include beforehand.

[anderruiz]

Let me know if you'd like a release or if there's another performance optimization you'd like to include beforehand.

I created a new PR, it would be great if we could merge it and release a version. I'm considering including support for CVSS 4.0 if you are fine with it.

[stevespringett]

Thanks for the optimizations @anderruiz Very much appreciated. v1.4.2 has been released to Maven Central and should be publicly available in the next 15 minutes or so. Support for CVSSv4 would be amazing.