How a pytest Run Can Show All Tests Passing While Still Returning Exit Code 1

A developer encountered a paradox during a regression check on 2026-09-27 where all 1,269 collected test cases eventually showed passing results, yet the pytest command consistently returned exit code 1. The discrepancy arose because process exit codes and individual test outcomes answer different questions, and retries can cause these two measures to diverge. Initial runs revealed failures and errors tied to a missing browser-runtime setting and a localhost binding issue, not a fundamental code problem. After carefully isolating the 69 affected test nodes by their exact IDs and authorizing only the necessary browser launch and local bind operations, a targeted retry resolved the outstanding failures. The key takeaway is that a completion report should preserve both the command result and the test roster result, rather than collapsing them into a single green or red verdict.