This page is running without the lab, so the two live Drupal sites, logins, live accessibility checks and Lighthouse are not available here. The steps below and what to look for still apply. To run this comparison:
Ticks are saved in this browser for this issue and included in the downloaded log. Nothing depends on them.
After the last step, press Run checks. Each check runs inside both frames and says Pass or Fail against what is expected. The rule: a fix check should fail on Before and pass on After; a regression check should pass on both; a precondition must pass on both or nothing else means anything.
| Type | Check | Expected | Before | After | Verdict |
|---|
The two frames stay in step, so you can confirm the change by eye. Do the steps, click the "URL alias" link yourself in each frame (Mirror off), then answer for each side. Your answers are compared with what is expected and saved in the downloaded log. This does not replace testing with a keyboard and a screen reader.
| Did you see this? | Expected Before / After | Before | After | Result |
|---|
Turn it on under Audits. axe-core runs in both frames after every page load and after interactions. Because both frames are driven the same way, a difference means the change affects accessibility. "Violations" are elements axe flags as failing WCAG A/AA rules. Automated checks find only a subset of problems; they do not replace a keyboard and screen-reader test.
Waiting for both frames to finish loading…
| Side | Critical | Serious | Moderate | Minor | Total |
|---|
| Rule | Impact | Before | After | Change |
|---|
| Time | Page (reported by) | Before | After | Difference |
|---|
Turn on background audits under Audits. Lighthouse runs in its own headless Chrome and audits the page as it loads, logged in. It does not see states you reach by clicking (for example an error summary after Save) and takes about 10 seconds per site. Accessibility and best-practice audits are shown by default; use axe for live, interactive checking.
| Category | Before | After | Change |
|---|
| Audit | Before | After | Change |
|---|
Mirror repeats real clicks, typed text, checkboxes and selects in the other frame to save setup time. It does not mirror the keyboard (Tab, Enter, focus), because script-generated key events do not behave like real ones. Do the final keyboard check in each frame yourself.
Viewer build: static. If this differs from the server's start, reload the page (an old tab keeps running old code).
Frames are separate sites on before.localhost and after.localhost, so logins do not interfere. This compares behaviour in a browser. It is not a substitute for testing with a keyboard and a screen reader.