Systematic Debugging

Overview

Random fixes waste time and create new bugs. Quick patches mask underlying problems.

Core principle: ALWAYS find the root cause before fixing. Symptom fixes are failures.

Violating the letter of this process violates the spirit of debugging.

Iron Rule

NO FIX WITHOUT ROOT CAUSE INVESTIGATION

If you have not completed Phase 1, you may not propose a fix.

When to Use

For EVERY technical problem:

• Test failures
• Production errors
• Unexpected behavior
• Performance issues
• Build failures
• Integration problems

ESPECIALLY when:

• Under time pressure (urgency makes guessing tempting)
• A "quick fix" seems obvious
• Multiple fixes have been tried
• A previous fix did not work
• You do not fully understand the problem

Do not skip because:

• The problem looks simple (simple problems have root causes too)
• You are in a hurry (rushing guarantees rework)
• A manager wants it fixed NOW (systematic is faster than thrashing)

Four Phases

Do NOT proceed to the next phase without completing the current one.

Phase 1: Root Cause Investigation

BEFORE attempting ANY fix:

1. Read Error Messages Carefully
   • Do not skip errors or warnings
   • They usually contain the exact solution
   • Read stack traces completely
   • Note line numbers, file paths, error codes
2. Reproduce Consistently
   • Can you trigger it reliably?
   • What are the exact steps?
   • Does it happen every time?
   • If not reproducible, gather more data -- do NOT guess
3. Check Recent Changes
   • What changed that could cause this?
   • Git diff, recent commits
   • New dependencies, config changes
   • Environment differences

4. Gather Evidence Across Components

   If the system has multiple components (API, service, database, etc.):

   FOR EACH component boundary:
     - Log the data entering the component
     - Log the data leaving the component
     - Verify environment/config propagation
     - Check state at each layer
   
   Run once to gather evidence --> show WHERE it breaks
   THEN analyze evidence to identify the failing component
   THEN investigate that specific component

5. Trace the Data Flow

   If the error is deep in the call stack:

   • Where was the bad value created?
   • Who passed this bad value?
   • Trace backward until you find the source
   • Fix at the source, not at the symptom

Phase 2: Pattern Analysis

Before fixing, find the pattern:

1. Find Working Examples
   • Find similar working code in the same codebase
   • What works that is similar to what is broken?
2. Compare with Reference
   • If you are implementing a pattern, read the reference implementation COMPLETELY
   • Do not skim -- read every line
   • Understand the pattern fully before applying
3. Identify Differences
   • What is different between working and broken?
   • List every difference, no matter how small
   • Do not assume "that is not important"
4. Understand Dependencies
   • What other components does this need?
   • What settings, config, environment?
   • What assumptions are being made?

Phase 3: Hypothesis and Test

Scientific method:

1. Form a Single Hypothesis
   • State clearly: "I believe the root cause is X because Y"
   • Write it down
   • Be specific, not vague
2. Test with Minimum Change
   • Make the SMALLEST POSSIBLE change to test the hypothesis
   • One variable at a time
   • Do not fix multiple things at once
3. Verify Before Proceeding
   • Did it work? Yes --> Phase 4
   • Did it not work? Form a NEW hypothesis
   • Do NOT pile more fixes on top
4. If You Do Not Know
   • Say "I do not understand X"
   • Do not pretend you do
   • Ask for help
   • Research further

Phase 4: Implementation

Fix the root cause, not the symptom:

1. Create a Failing Test
   • Simplest possible reproduction
   • Automated test if possible
   • One-off test script if no framework exists
   • Must exist BEFORE the fix
   • REQUIRED SKILL: tdd-workflow (when writing the fix)
2. Apply a Single Fix
   • Address the identified root cause
   • ONE change at a time
   • No "while I am here" improvements
   • No bundled refactoring
3. Verify the Fix
   • Does the test pass now?
   • Did any other tests break?
   • Is the problem actually resolved?
4. If the Fix Does Not Work
   • STOP
   • Count: How many fixes have you attempted?
   • Fewer than 3: Return to Phase 1, re-analyze with new information
   • 3 or more: STOP and question the architecture (see step 5 below)
   • Do NOT attempt a 4th fix without the architecture discussion

5. 3+ Fixes Failed: Question the Architecture

   Indicators of an architectural problem:

   • Each fix exposes new shared state, coupling, or issues in a different place
   • Fixes require "massive refactoring" to apply
   • Each fix creates new symptoms elsewhere

   STOP and question the fundamentals:

   • Is this pattern fundamentally sound?
   • Are we continuing just because of inertia?
   • Should we refactor the architecture instead of patching symptoms?

   Discuss with the user before attempting more fixes.

   This is not a failed hypothesis -- it is a wrong architecture.

Red Flags -- STOP and Follow the Process

If you catch yourself thinking:

• "Quick fix for now, I will investigate later"
• "Let me just change X and see if it works"
• "Make multiple changes, run the tests"
• "Skip the test, I will verify manually"
• "Probably X, let me fix that"
• "I do not fully understand but this might work"
• "The pattern says X but I will adapt it differently"
• "The main issues are: [list of fixes without investigation]"
• Proposing a solution without tracing the data flow
• "One more fix attempt" (after 2+ failures)
• Each fix exposing new problems in different places

ALL OF THESE MEAN: STOP. Return to Phase 1.

If 3+ fixes have failed: Question the architecture (see Phase 4.5).

Common Rationalizations

Quick Reference

Supporting Techniques

• Binary search: Narrow down the problem by bisecting the code
• Git bisect: Find the commit that introduced the bug (git bisect start)
• Log analysis: Examine system logs, search for patterns
• Minimal reproduction: Reproduce the problem with the smallest possible example

Related skills:

• tdd-workflow -- Creating a failing test (Phase 4, Step 1)
• verification-gate -- Do NOT claim success without verifying the fix works