Closed eliericha closed 1 year ago
hi, what is your proposed fix?
I suggest that after ordering by region start, we order markers with the same region start by their message. That way we should get a deterministic outcome.
so the end/stop you wont consider?
Ah that's a good point. Indeed we could sort by end, and then by message. From the looks of the code it's easy to chain comparators.
can you provide a pr
Yes. I also have a minor fix that avoid a NPE elsewhere. Should I submit in the same PR or open a different issue?
I opted to include the minor NPE fix, but let me know if you'd rather proceed otherwise.
thx for the fix @eliericha
Markers are currently sorted by region start only. So when multiple markers have the same region start, they are ordered undeterministically among each other. That makes it difficult to use in some test scenarios as we can't compare to a fixed reference.
I'll try to make a submission to fix this.