Improve test case exercising the sorting path for hash index build.
On second thought, we should probably do at least a minimal check that
the constructed index is valid, since the big problem with the most
recent breakage was not whether the sorting was correct but that the
index had incorrect hash codes placed in it.
Branch
------
master
Details
-------
http://git.postgresql.org/pg/commitdiff/606ccc5e7e97914073f991b077712645e125d531
Modified Files
--------------
src/test/regress/expected/create_index.out | 17 +++++++++++++++++
src/test/regress/sql/create_index.sql | 3 +++
2 files changed, 20 insertions(+)