Debugging Strategies Competent¶

🧠 Soft Skills · Level 2

⏱️ ~2 days

When you'd use this

Systematic debugging, scientific method, rubber duck and common bug patterns.

Debug systematically — reproduce, isolate, hypothesize, verify — instead of changing things at random.

The scientific method for debugging¶

The scientific method for debugging — a key concept in Debugging Strategies.

1. OBSERVE   — What exactly is the symptom?
2. HYPOTHESIZE — What could cause this?
3. PREDICT   — If my hypothesis is right, what should happen when I...?
4. EXPERIMENT — Test the prediction
5. CONCLUDE  — Was I right? If not, new hypothesis.

Strategy ladder (try in order)¶

Strategy ladder (try in order) in Debugging Strategies — what it is and when to use it.

Step	Method	When
1	Read the error message	Always start here — Python errors are clear
2	Reproduce minimally	Strip away everything unrelated
3	Add print/logging	Quick check of values at key points
4	Use breakpoint()	Step through execution
5	Binary search	Comment out half the code — which half breaks?
6	Rubber duck	Explain the problem aloud (or to a duck)
7	Git bisect	Find which commit introduced the bug
8	Sleep on it	Fresh eyes see things tired eyes miss

Common Python bug patterns¶

Common Python bug patterns in Debugging Strategies — what it is and when to use it.

# 1. Mutable default argument (classic trap)
def append_to(element, target=[]):   # BAD — shared list!
    target.append(element)
    return target

# Fix:
def append_to(element, target=None):
    if target is None:
        target = []
    target.append(element)
    return target

# 2. Late binding in closures
functions = [lambda x: x + i for i in range(5)]
print([f(0) for f in functions])   # [4, 4, 4, 4, 4] — all use i=4!

# Fix: capture i as default argument
functions = [lambda x, i=i: x + i for i in range(5)]
print([f(0) for f in functions])   # [0, 1, 2, 3, 4]

# 3. Modifying list while iterating
items = [1, 2, 3, 4, 5]
for item in items:
    if item % 2 == 0:
        items.remove(item)   # SKIPS elements!

# Fix: iterate over a copy or use comprehension
items = [x for x in items if x % 2 != 0]

# 4. == vs is
a = 1000
b = 1000
print(a == b)    # True (same value)
print(a is b)    # False! (different objects — outside cache range)

# 5. Forgetting to await
async def get_data():
    return await fetch("...")   # without await: returns coroutine, not data!

Git bisect — find the breaking commit¶

Git bisect — find the breaking commit, part of Debugging Strategies.

git bisect start
git bisect bad                  # current commit is broken
git bisect good abc1234         # this older commit was working
# Git checks out middle commit
# You test: does the bug exist?
git bisect good   # or git bisect bad
# Repeat until: "abc5678 is the first bad commit"
git bisect reset

Practice Exercises¶

Debug a provided broken function using only print statements (no debugger).
Use breakpoint() to step through a recursive function and understand state at each level.
Find the bug in 5 tricky code snippets (mutable defaults, closures, off-by-one, etc.).
Use git bisect to find which commit broke a test.
Write a bug report for an open-source project — include reproduction steps, expected/actual behavior.

💬 Discussion

Have a question about this topic? Found an error? Share your thoughts below.