Set multipleOf for numbers with @Digits, stop generating pattern

[MikeEdgar]

Remove incorrect mapping of number @Digits to pattern schema constraint and replace with multipleOf when not already set in the schema. Uses multipleOf: 1 for integers (i.e. Digits#fraction() <= 0) and multipleOf based on fraction value when > 0.

Fixes #2384

[MikeEdgar]

@Azquelt please take a look if you have a chance and let me know if anything here appears to you that it may conflict with the spirit of microprofile/microprofile-open-api#717.

[MikeEdgar]

@Postremus FYI. Please take a look and give feedback if this looks amiss.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
100.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[Azquelt]

Looks good, I don't see any conflicts with microprofile/microprofile-open-api#717

I've left a comment but my only other thoughts were additional things we could do:

• We could infer minimum and maximum values based on Digits.integer if it's not otherwise constrained.
  • I'm a little uncertain on this since it's entirely possible the real constraint is tighter than @Digits suggests
  • although strictly speaking @Digits(integer = 6) implies "maxExclusive": 100000, "minExclusive": -100000, having that constraint might be worse than having no constraint documented if the actual constraint were e.g. 0 <= x < 65536
• We could infer "multipleOf": 1 for integer types

[Postremus]

@MikeEdgar looks good to me.

[MikeEdgar]

We could infer minimum and maximum values based on Digits.integer if it's not otherwise constrained.

I considered this one too, but didn't do it here mainly for the reason you mentioned about being inaccurate. I suppose if @Digits is not scanned until after all of the possible min/max constraints, it should be fine.

We could infer "multipleOf": 1 for integer types

You mean when no @Digits is present at all? I think for the integer Java types we should be generating type: [ integer ] so it may be redundant.

[Azquelt]

I think for the integer Java types we should be generating type: [ integer ] so it may be redundant.

Ah that's true, it would then be redundant.

[MikeEdgar]

I've opened #2423 for the additional handling to set min/max constraints so we can think about that in the future and not block this fix/change for now.