Jérémy Frey

Hello, I think this is due to the timeout, too low for an actual network, plus once the resolve fails there is no new attempts. We had similar issue and...

Note: posted too quickly, actually I did not investigate your own workaround it might be better than to keep calling the ContinuousResolver after launch, what we did ^^

I see your point. "effective" sampling rate, but not a constant one. To prevent this false assumption, downstream functions could maybe check that the nominal sampling rate is 0 (i.e....

Seems perfectly fine to me as well, here is a proposal.

Hello, New batches of experiment will start soon in my lab, that would help if these commits are integrated upstream. Is there something I can do to ease the merge?

Hello, I have been using git-subrepo for several personal and professional projects for a couple years. This is a *very* handy tool :+1: I just came across a situation where...