Under current design, out-of-order timestamped updates result in fixing up of existing newer updates with older timestamps (hint __hs_fixup_out_of_order_from_pos()). We intend to keep the transaction ids on the updates intact so that older readers can still read based on transaction based visibility rules.

In WT-7200, we observed that a bug in this logic was introduced but none of the existing WiredTiger tests was able to catch it. The aim of this ticket is to write a more rigorous test for out-of-order timestamp fix-up logic.