Peter Geoghegan <pg@heroku.com> writes:
> FWIW, I think that that bug tells us a lot about hash index usage in
> the field. It took many months for someone to complain about what
> ought to have been a really obvious bug. Clearly, hardly anybody is
> using hash indexes. I broke hash index tuplesort builds in a similar
> way at one point, too. The slightest bit of regression test coverage
> would have caught either bug, I believe.
We *do* have regression test coverage, but that code is set up to not
kick in at any index scale that would be sane to test in the regression
tests. See
https://www.postgresql.org/message-id/12194.1466724741@sss.pgh.pa.us
regards, tom lane