Fix stats_ext test in 32 bit machines
Because tuple packing is different (because of the MAXALIGN difference),
the expected costs of a seqscan is different.
The commonly used trick of eliding costs in EXPLAIN output (COSTS OFF)
would make the tests completely pointless. Instead, add an alternative
expected file.
Branch
------
master
Details
-------
http://git.postgresql.org/pg/commitdiff/6e9aaf6e471b446e5ccad6e9e174c3490978ae13
Modified Files
--------------
src/test/regress/expected/stats_ext_1.out | 155 ++++++++++++++++++++++++++++++
1 file changed, 155 insertions(+)