It is expected that a significant amount of repair examples will be artificially synthesized. Two potential methods have been discussed:
[x] #18
[ ] #19
In either case, a choice needs to be made regarding where and how these synthesized commits will be preserved as part of the dataset (since they cannot be incorporated in the source repositories).
It is expected that a significant amount of repair examples will be artificially synthesized. Two potential methods have been discussed:
In either case, a choice needs to be made regarding where and how these synthesized commits will be preserved as part of the dataset (since they cannot be incorporated in the source repositories).