Hi,
While working on [1], I found the internal error "tuple concurrently updated"
is raised by simple_heap_update and other similar functions, and the comments
on them says "Any failure is reported via ereport()". However, I could not understand
the intension of this comments because I suppose the unexpected errors are usual
reported via elog() not ereport and in fact elog() is used in these functions.
I wonder this statement should be fixed as the attached patch or could be removed for
less confusion. Maybe, I am just missing something, though....
[1] https://www.postgresql.org/message-id/flat/20250331200057.00a62760966a821d484ea904%40sraoss.co.jp
--
Yugo Nagata <nagata@sraoss.co.jp>