AgentSkillsCN

pythonista-debugging

适用于遇到错误、Bug 或问题时使用。可通过“Bug”、“错误”、“修复”、“调试”、“回溯”、“堆栈跟踪”、“异常”、“崩溃”、“损坏”、“无法运行”、“失败”、“问题”、“权宜之计”、“封装”、“黑客手段”、“重现”、“二分查找”等短语触发，或当您心存侥幸，试图通过增加复杂性来回避真正问题的解决时使用。

SKILL.md
--- frontmatter
name: pythonista-debugging
description: Use when encountering errors, bugs, or problems. Triggers on "bug", "error", "fix", "debug", "traceback", "stack trace", "exception", "crash", "broken", "not working", "failing", "issue", "workaround", "wrapper", "hack", "reproduce", "bisect", or when tempted to add complexity to avoid fixing the real problem.

Debugging and Root Cause Fixing

Core Philosophy

Find and fix the root cause. NEVER work around problems with wrappers or complexity.

Debugging Workflow

1. Read the Stack Trace

python
# Stack traces read bottom-to-top. The last frame is where error occurred.
Traceback (most recent call last):
  File "main.py", line 10, in <module>      # Entry point
    result = process_data(data)              # Calling function
  File "processor.py", line 25, in process_data
    return transform(item)                   # Where it failed
TypeError: 'NoneType' has no attribute 'items'  # What went wrong

Key info: File, line number, function name, actual error message.

2. Reproduce the Bug

bash
# Create minimal reproduction
pytest tests/test_module.py::test_specific -v

# Or write a quick script
python -c "from module import func; func(problematic_input)"

3. Use Debugger or Print Statements

python
# Quick debugging with print (remove after!)
print(f"DEBUG: {variable=}, {type(variable)=}")

# Better: Use debugger
import pdb; pdb.set_trace()  # Interactive debugger

# Or breakpoint() in Python 3.7+
breakpoint()

4. Bisect if Needed

bash
# Find which commit introduced the bug
git bisect start
git bisect bad HEAD
git bisect good v1.0.0
# Git will guide you through testing commits

The Anti-Pattern: Working Around Instead of Fixing

When you encounter an error, your instinct may be to:

  1. •Add a wrapper class
  2. •Use __getattr__ for dynamic delegation
  3. •Use # type: ignore to suppress errors
  4. •Create helper classes that "adapt" interfaces

This is WRONG. These are signs you're working around instead of fixing.

Common Workarounds to Avoid

Wrapper Classes with __getattr__

python
# WRONG - Magic delegation
class TestWrapper:
    def __init__(self, obj):
        self.obj = obj
    def __getattr__(self, name):
        return getattr(self.obj, name)

# CORRECT - Simple helper function
def test_helper(obj):
    pass

Adapter Classes

python
# WRONG - Adapter to hide interface mismatch
class TestAdapter:
    def __init__(self, production_obj):
        self.obj = production_obj
    def adapted_method(self):
        return self.obj.method().to_dict()

# CORRECT - Transform inline, fix the interface
def test_something(production_obj):
    result = production_obj.method()
    assert result.to_dict()["field"] == expected

Type Ignores to Hide Problems

python
# WRONG - Hiding the problem
result = maybe_none.items()  # type: ignore

# CORRECT - Handle the None case
if maybe_none is not None:
    result = maybe_none.items()

Red Flags

You're working around instead of fixing if you're:

  1. •Creating a wrapper class "just for tests"
  2. •Using __getattr__ or other magic methods
  3. •Adding complexity to avoid changing existing code
  4. •Thinking "I'll just adapt this interface..."
  5. •Using # type: ignore without investigating why

The Right Approach

  1. •Stop when you realize you're working around
  2. •Identify the root cause (read stack trace, reproduce)
  3. •Fix the root cause with simple, explicit code
  4. •Delete any workarounds you created

Questions to Ask

  • •Am I adding complexity to avoid fixing the real problem?
  • •Is there a simpler, more direct way to do this?
  • •Would someone reading this code understand what's happening?
  • •Can I reproduce this bug with a minimal test case?

Related Skills