feat: make the trace deterministic

[letypequividelespoubelles]

No description provided.

[DavePearce]

Notes:

• ExpOperation needs a ExpOperationComparator as Exp implements OperationSetModule<ExpOperation>.

• StpOperation needs a StpOperationComparator as Stp implements OperationSetModule<StpOperation>.

• Does ExpOperation need the annotation @EqualsAndHashCode(onlyExplicitlyIncluded = true, callSuper = false) ?

• There is some inconsistency, in that e.g. BlockHash does things differently. But perhaps it has to?

[DavePearce]

It seems odd that you've implemented all the Comparators separately from the class itself. You can have e.g. done class AddOperation extends ModuleOperation implements Comparable<AddOperation> then added a compareTo() method as required for the interface. Just saves lines I guess --- not a big deal though.

[DavePearce]

I suppose the reason for doing it this way is to avoid implemented a Comparator for all Operation classes? I think you still could do it though by modifying the signature of OperationSetModule.sortOperations(). I'd have to check it worked though --- so not worth bothering!

[DavePearce]

Still, one reason for having AddOperation implement Comparable<AddOperation> is that you can then modify ModuleOperationStackedSet to require that its elements must be sortable like so:

public class ModuleOperationStackedSet<E extends ModuleOperation & Comparable<E>>
    extends StackedSet<E> {

That just helps reduce the chance of a problem.

[letypequividelespoubelles]

There is some inconsistency, in that e.g. BlockHash does things differently. But perhaps it has to?

made more consistent. For BlockHash we have to sort before commit time, as we have to trigger WCP.

Does ExpOperation need the annotation @EqualsAndHashCode(onlyExplicitlyIncluded = true, callSuper = false)

added

[DavePearce]

Otherwise, LGTM

[DavePearce]

This doesn't seem ideal. I would have thought it makes more sense to add compareTo to the ExpCall interface and, hence, its implementations (ExplogExpCall and ModexpLogExpCall). Otherwise, this is a bit fragile is someone adds a new implementation of ExpCall ... they might forget to update this bit.

[letypequividelespoubelles]

Do agree.

[DavePearce]

I think there is still a problem with ext module.

I think there is still a problem with ext module. TBC

[DavePearce]

Minor typo in ext -- found via testing :)

[DavePearce]

Yeah, with that final typo fixed ... I think it looks good.

[DavePearce]

Nice job. LGTM.