Reverse Sampler Behavior – Identical sample maps in two samplers will both be reversed

[d.healey]

@Christoph-Hart I've been playing around with the shared monolith reference and it works well until I try to use two samples maps with different numbers of round robins.

Here's a project demonstrating the issue:
https://www.dropbox.com/s/77oloqbssef2bt1/Shared Samples Test.zip?dl=0

Any ideas?

[Casey Kolb]

@d-healey Major bump on this. I'm realizing that this reverse issue is broken for expansions.

@Christoph-Hart Do we know if there's a solution? Engine.setAllowDuplicateSamples(false);

[d.healey]

@casey-kolb Should work with expansions, Christoph did some trouble shooting on this with me. When I'm back at my computer I'll check if there is any special requirement for this. But note that it doesn't work if the sample maps have different numbers of RR groups.

[d.healey]

@casey-kolb Ah I think you might need to use Expansion.setAllowDuplicateSamples(false); Let me know if it works for you.

[Christoph Hart]

@d-healey said in Reverse Sampler Behavior – Identical sample maps in two samplers will both be reversed:

Ah I think you might need to use Expansion.setAllowDuplicateSamples(false);

Yup that's the case - I remember we discussed whether this is necessary, and came to the conclusion that it's better if you have to decide this per expansion (since there might be applications where expansions can be used in a very wide context).

[d.healey]

@christoph-hart Any ideas about the shared monolith reference not working with different amount of RR groups?

[Casey Kolb]

@christoph-hart Got it. Will test this right away!

[Casey Kolb]

@d-healey David and Christoph save the day again! Thanks guys. That did the trick.

[d.healey]

I've opened a github issue relating to the reference monolith problem when the sample maps have different amounts of RR groups.