debug

Start by reproducing the failure with the exact command and reading the full error output and stack trace. Then form a hypothesis about the failing component, isolate it with targeted logging or bisection, fix the root cause, and confirm the test passes.

Use git bisect to binary-search through commit history between a known-good and known-bad state. For uncommitted code, comment out half the changes and narrow down until the offending line is found.

Symptom patches like swallowing exceptions or adding null checks without understanding the cause hide the real defect and let the same class of bug resurface elsewhere. Fixing the root cause prevents the entire category of failure.

Avoid catching and ignoring exceptions, adding null guards without understanding why the value is null, retrying flaky tests instead of fixing them, and disabling tests or linters to bypass failures.

Add a regression test whenever feasible after fixing a bug, especially for defects that reached production or escaped existing test coverage. The test should reproduce the original failure condition so the bug cannot silently return.