Finding a bug step by step

Your manager opens the weekly support report. The average wait says "error". Last week it worked.

You could stare at the code and guess. Or you could check one thing at a time until the bug has nowhere to hide. The second way works every time.

When Python stops, it prints a traceback. It looks long, but read it from the bottom.

The last line says what went wrong. The line above it says where. Everything higher up shows the calls that led there.

Most of the time, those two lines are all you need.

def average_wait(waits):
    # Minutes each support ticket waited for a first reply.
    total = 0
    for wait in waits:
        total = total + wait
    return total / len(waits)


print(average_wait([12, 30, 18]))
print(average_wait([]))

# Press Run. Read the last line of the error first.
# Then change the function so an empty list returns 0.

Some bugs give no error at all. The code runs and gives a wrong number.

Then add a print inside the code to show the values as they change. Compare each printed value with what you expected. The first line that differs points at the bug.

Take the extra prints out once it is fixed.

def hours_worked(shifts):
    # Each shift is [start hour, end hour] on a 24-hour clock.
    total = 0
    for shift in shifts:
        start = shift[0]
        end = shift[1]
        total = start - end
    return total


week = [[9, 17], [9, 13], [12, 20]]
print("Hours this week:", hours_worked(week))

# The answer should be 20. Add this line inside the loop:
#     print("start", start, "end", end, "total", total)
# Run it, read each line, and fix the two mistakes.

In a long piece of code, check the value in the middle. If it is already wrong, the bug is in the first half. If it is right, the bug is in the second half.

Then check the middle of that half. Each check rules out half of what is left.