Changing a page address preserves its redirect
When a page's address changes, the old one redirects to the new one and leaves the sitemap. A defect made that record fail silently.
Changing the address of a published page has a cost: old links, bookmarks and search results point to the previous address. The system records a permanent redirect so they keep arriving.
The defect
Recording the redirect used a query that, when it did not find the page, ended the request. The save was interrupted midway, with no warning, and the redirect was not created.
The fix
The query now looks the page up directly in the table, by its current identifier and language. If it is not found, the redirect is skipped and the reason goes to the log, without taking down the save.
- Going from A to B, back to A and to B again does not create repeated records.
- In the sitemap, the old address is removed before the new one enters. Without that, both were listed, and the search engine kept visiting an address that already redirects.
The fix applies to regular pages and to publications, which had the same defect.
Source: Core · req-112 · BATCH-112