How to have accurate expected-actual test after Level-7?

[iamgenechua]

After level 7, we will be able to persist data on our harddrive.
Now, everytime we start up our program, we continue to have the tasks from before.
These tasks are ever changing, and the 'list' command will show different tasks depending on what i have done.
My expected output will behave as if it is the first time we run the program.
How do i do my testing accurately then?

[erisjacey]

If by testing you mean continuing to use .bat or .sh files to script testing, the way I did it was pretty rudimentary:

I created a feature to support the "delete all" user input command. This command (as its name suggests) deletes all existing tasks (if any) in the list of tasks.

All I did afterwards was make sure this command was the first line of my input.txt. This allowed the task list to "reset" every time the script is run.

A cool side effect to adding this command is that users of my Duke chatbot will be able to use it in normal circumstances as well (because I think the default implementation of delete 2, etc. can be quite clunky especially for substantially big lists).

I'm sure there are more clever ways to go about doing this though.

[dianneloh9]

For me, for the .sh (or .bat) file, I added a line that deleted the data file before running the test if the file exists initially. So that the test can run from a clean slate.