Print vs pdb

print 和 pdb

≈ 20 min · python-debugging · Open in WolfCode

Two semantic-error functions. Print debugging finds both. Brief intro to pdb for the day prints stop being enough.

Bug 1 — forgot to divide

Python · runnable

def average(nums):
    return sum(nums)             # ← BUG: never divides

print(average([10, 20, 30]))     # 60 — should be 20.0

Loading Python runtime (~5 MB, one-time)…

Bug 2 — boundary check

Python · runnable

def in_range(value, low, high):
    return low < value < high    # ← BUG: < excludes boundaries

print(in_range(1, 1, 10))        # False — should be True (inclusive)

Loading Python runtime (~5 MB, one-time)…

The fixes

Python · runnable

def average(nums):
    if not nums:
        return 0.0               # guard against ZeroDivisionError
    return sum(nums) / len(nums)

def in_range(value, low, high):
    return low <= value <= high

print(average([10, 20, 30]))    # 20.0
print(average([]))               # 0.0
print(in_range(1, 1, 10))       # True
print(in_range(10, 1, 10))      # True
print(in_range(0, 1, 10))       # False

Loading Python runtime (~5 MB, one-time)…

When print stops being enough — try pdb

In a real script you'd add breakpoint() and run with python (not in browser)

# In your terminal — Python 3.7+:
#
#     def average(nums):
#         breakpoint()                 # ← drops into (Pdb) prompt here
#         return sum(nums) / len(nums)
#
# At the (Pdb) prompt:
#   p nums           # print the value of nums
#   p sum(nums)      # try any expression
#   n                # next line
#   c                # continue
#   q                # quit
#
# For an entry-level course, print covers 90% of bugs.
# pdb is the tool for when print produces too much noise.

print("(pdb is interactive — try it in your terminal, not the browser)")

Loading Python runtime (~5 MB, one-time)…