I have a situation where my assume() call rejects more than 90% of example inputs. It's hard to randomly generate inputs that pass assume() in my case, so I'd like Hypothesis to just keep trying. This code stops iterating after max_examples * 10 tries, however, with the message "settings.max_examples={s.max_examples}, but < 10% of examples satisfied assumptions". The logic is not configurable. I propose to make it configurable, if that's something you're interested in? I don't yet have a proposal for what the option will be called.

(Background: I'm generating inputs to test a program which uses a SAT solver. It's hard to predict which combinations will lead to a solution or no-solution. I don't want to figure that out, especially since whatever logic I use to make solvable inputs might be overly restrictive and not fully test the program. I'd rather keep making random inputs, assume() that the SAT solver finds a solution, then assert things about the solution.