To help detect and isolate memory corruption of a page after it was read from disk and after its checksum was validated, it would be helpful to add a few basic data format checks closer to page read, e.g., checking that the page type is valid.

If the checks fail, we should dump the page image contents to the logs as described in WT-17081 and WT-17082 to help us diagnose the issue.

If page decoding (e.g., __wt_page_inmem happens sufficiently close to page reading for this ticket to be practical, please add the explanation as a comment and close the ticket.