I just had an idea about using deep learning to train a bot to judge the quality of a merge. Does it have tests? Does it pass lint? Is the subset of the grammar used in the patch the same as the rest of the source? If a patch is "clean" it should get automatically merged and only require human intervention to reject it.
I just had an idea about using deep learning to train a bot to judge the quality of a merge. Does it have tests? Does it pass lint? Is the subset of the grammar used in the patch the same as the rest of the source? If a patch is "clean" it should get automatically merged and only require human intervention to reject it.