WL#1894: Add row id's when sorting after ordered index scan

Affects: Server-5.0   —   Status: Assigned   —   Priority: Medium

When returning rows from ordered scans duplicate rows are
possible. In MySQL merge join algorithm duplicate rows sometimes
have to be filtered out. To do this efficiently a unique row id
is set for each row returned. In the handler this could be done
in two ways:
1) Create row id's in the sorting
2) Fetch the tuple id's