Question 2 (parenthesis): hints not matching the real problem of user's code about tie HOT 8 CLOSED

Thanks for filing, @skywalker-lili! This looks like an important bug, but there isn't enough detail to reproduce it.

Would you mind providing clear reproduction instructions, expected/observed behavior, and a screenshot, similar to #83?

from tie.

Sure. Here is how to reproduce it:

1. Open "app.html"

2. Navigate to "Question 2: Balanced Parentheses"

3. Modify the isBalanced(s) function to return False. Click "run" button to get a hint.
   

4. Then modify the isBalanced(s) function to return False or False. Click "run" button to get a hint.
   

5. Then modify the isBalanced(s) function to return False or False or False. Click "run" button to get a new hint.
   

6. The hints from step 3-5 are different even though the code made the same mistakes by just returning False. The last hint is especially misleading by assuming user's code is counting number of parenthesis.

Expected behavior: hints should be stable when user's code made same mistakes or when making the same wrongs in test cases. Also, hints shouldn't mislead the user what kind of mistake its code has.

from tie.

/cc @rujunm

from tie.

@skywalker-lili, thanks for adding the detail!

from tie.

Just wanted to say that this behavior seems intended: if you provide multiple hints for the same buggyOutputTest, then it'll just iterate through all of them as long as the code is edited in some way but returns the same result. So this feels like two separate issues:

1. The criteria for when to give the next hint should be stricter than just "the code isn't exactly the same". Maybe some sort of edit distance or timestamp check?

2. The "countNumberOfParentheses" function isn't doing the intended thing; I think the "return numLeft == numRight" line should be outside the for loop.

from tie.

I can fix 2 (removing the tab does indeed fix it), but should 1 be moved to a separate issue? @seanlip

from tie.

Yes, let's move 1 to a separate issue. Thanks @andersonw!

from tie.

Second part fixed in #212; first part filed as #211. Thanks @andersonw!

from tie.