I have not had any patch assigned for review for sometime, presumably because I have reviewed a patch which is stuck in the 'in-progress' state.
proposed ways to enhance the matching code:
- soften the count == mincount constraint, by selecting the users having less than mincount+1 patches in-progress or pending-review
- weight the in-progress patches differently than the pending-review patches, e.g. in-progress could have a weight of .5 and pending-review a weight of 1.
- replace the random.choice at the end with the selection of the user having passed a review transition the least recently of all candidate. This will have the effect of making committers less selectable for 1st pass review if they act as active 2nd pass reviewers. It will also increase fairness towards active 1st pass reviewers which do they work by not dispatching reviews to them only (together with point 1).