After having manually reviewed a bunch of patches generated by aflow, I noticed that the more complex they tend to be - the less likely they are to be accepted by upstream.
It's not a 100% match but there's at least a strong correlation. I attribute this to occam's razor. Maybe this could be an interesting output metric for the patching flow ?
After having manually reviewed a bunch of patches generated by aflow, I noticed that the more complex they tend to be - the less likely they are to be accepted by upstream.
It's not a 100% match but there's at least a strong correlation. I attribute this to occam's razor. Maybe this could be an interesting output metric for the patching flow ?