Compare saved revisions and restore the right wording

1. Choose a known article and record its baseline

Use the accepted Three things to check before seasonal garden work article from lesson 42. Sign in as Cedar Editor, open it in Posts, and record its post ID, Published status and complete wording in your private exercise record. Open the visitor page separately and check the same content. Use the recovery-tool worksheet to distinguish a revision from Undo, autosave, Trash, an XML export and a complete backup. This exercise uses a disposable course article. For important content, first create and verify a recovery point suitable for the change. A revision restores supported Post or Page content; it does not restore plugins, uploads, settings or a whole site, so those failures need a separate full-site backup and restore plan.

2. Save two intentional revisions

At the end of the article, add the heading Check the work area and the sentence Describe the work area without publishing private access details. Click Save, leave the editor and reopen the article. Record this complete saved state as Revision A. Change the new heading to Check access and the work area. Change its sentence to Check the route and work area, then keep private access details in the agreed enquiry channel. Save and reopen again. Record this as Revision B. The two full editor screenshots below were captured after saving and reloading each version. Keep the original introduction, three sections and Services link unchanged. Deliberately saved versions are the objects you will compare.

3. Observe a real autosave without saving the marker publicly

Add a temporary paragraph: Autosave test — do not keep. Do not click Save. Leave the editor open and let its normal autosave run. In this recorded WordPress 7.1 session the configured interval was 60 seconds; use the actual behavior on your installation rather than assuming a fixed wait always succeeds. The temporary marker below was captured after the editor performed its own timed autosave. A separate signed-out page still shows saved Revision B and contains no marker. Leave and reopen the editor: WordPress displays There is an autosave of this post that is more recent than the version below, with View the autosave. This notice is recovery evidence. It does not mean the temporary paragraph became the published article or a whole-site backup. If the notice does not appear, inspect your actual autosave state before continuing; do not label a manually saved marker as an autosave.

4. Open the current history, then compare A with B

In the Post sidebar, find Revisions and click the number beside it. WordPress 7.1 opens its current Revisions screen. The timeline, author entries and Show changes control help inspect each saved state; the autosave is identified separately. Select the saved B entry and inspect its changes. For the explicit two-version text comparison in this exercise, click Open classic revisions screen. The screenshot below shows the actual resulting classic page. Enable Compare any two revisions. Use the left and right timeline handles to select A and B. Check the exact heading and sentence: the left side removes Check the work area and its first sentence; the right side adds Check access and the work area and the improved sentence. Read the surrounding unchanged content too. Use the wording, author and recorded revision identifiers to identify the version. If date displays differ between interfaces, investigate the timezone and keep the exact-content comparison as your deciding evidence. Do not restore an entry just because it looks like the right time.

5. Select and restore Revision A

In the classic screen, clear Compare any two revisions to return to one selected revision. Move the timeline to the earlier saved A entry. Read the right-hand content and confirm Check the work area and its first sentence. The B wording and temporary autosave marker must not be the selected result. Click Restore This Revision only when A is the intended selection. WordPress returns to the editor. Read the entire article and compare it with your recorded A version. The screenshot below shows the actual restored editor; its saved content matched A exactly. This fixture restored the published post directly, so an extra Save was not needed to make A public. Your interface may require an explicit Save or Update if there are further unsaved changes; inspect its current state rather than applying a button sequence blindly.

6. Verify what the visitor receives

Open the article in a separate signed-out window. Confirm Check the work area and its original sentence appear. Confirm Check access and the work area and Autosave test — do not keep do not appear. Read the unchanged introduction, original three sections and Services link as well. The full page below proves the earlier version reached the visitor without losing unrelated article content. A successful editor return alone would not prove that.

7. Return the exercise to its accepted final article

Use your baseline record to remove the temporary revision-exercise section and restore the accepted wording from lesson 42. Click Save, leave the editor, reopen it and compare the full saved content with the baseline. Open the visitor article again at desktop and phone widths. Confirm that the accepted Check access before the visit section and Services link remain, and all exercise/autosave text is absent. Record the revision IDs, restoration result and final Published status in the worksheet. Historical revisions may remain in the private history. The cleanup check is that the current accepted article is correct and no temporary test text is public; it does not pretend that WordPress erased every historical record.

Practise

Create saved A and B versions of one disposable article, observe a real autosave, compare the exact differences, restore A through the native control, check the signed-out result and finish by restoring the accepted baseline.

Check your result

A and B were deliberately saved; a separate timed autosave was observed and stayed private. The native restore matched A exactly, the visitor received it with unrelated content intact, and the final published article returned to its accepted baseline.

Optional notes