Why Import error?

[ShangwenWang]

I have already installed unidiff:

Then I try to run the command python3 run.py Chart 1 Patch1
It is wrong:

Why is it?

[xiongyingfei]

This error shows that unidiff is not correctly installed. Pls check your installation.

[ShangwenWang]

@xiongyingfei Thanks Prof. Xiong.
I have already solved this problem and met another one.
In the README, authors introduce that this experiment was conducted under Defects4j v-1.2. However, in this version, there is no replaceall-current.jar in the folder test_generation which is contradict with Project.pm in the mod-d4j folder in this repo. As a result, I cannot get correct result. Could you please remember why this happens? Maybe you performed the study under another version of Defects4j like v1.3?

[Ultimanecat]

The version we used in the experiment may be a bit older than d4j1.2. It can be an older minor release,. Also, I guess d4j-v1.1 might work. I have no idea whether this script will success on d4j1.3 as we did this experiment before its release. I currently suggest trying d4j-v1.1 if v1.2 fails. We will try to recall the specific version we used in the experiment. Hope this could help you.

[ShangwenWang]

@Ultimanecat Thanks for your reply!
However, I guess it must be a version later than v1.2, since in the versions older than v1.2 (like v1.1), the .init.sh files never install replaceall-current.jar file.
Hope you can recall that cause that's really a problem when trying to apply this tool.

[xinyuan-liu]

You are right @ShangwenWang . This repo was updated in Aug 2018 with the latest commit of d4j at that time (later than v1.2 but before v1.3 release). I've updated the d4j modification files so it should be competable with d4j v1.3 now.

[Ringbo]

@xinyuan-liu Sorry about that, I also meet 2 problems:

1. In the last LIST pass_distance, is it possible the elements be NaN? In the result of PatchRepair5 ,I got the Nan result in pass_distance list.
   such as:

fali_distance: [7.059298104074019E-4]；
pass_distance: [NaN,NaN, 0.0, NaN, NaN, NaN, NaN, NaN, NaN, NaN].

2. A uninitialized warning happened in /APR/defects4j/framework/core/Utils.pm line 143. But when I change the tool from randoop to evosuite, it didn't happened again. And the tmp file has been generated nomarlly both in evosuite and randoop.