Stabilize test_decoding touching with sequences
Some of the test_decoding regression tests are unstable due to modifying
a sequence. The first increment of a sequence after a checkpoint is
always logged (and thus decoded), which makes the output unpredictable.
The runs are usually much shorter than a checkpoint internal, so these
failures are rare, but we've seen a couple of them on animals that are
either slow or are running with valgrind/clobber cache/...
Fixed by skipping sequence decoding in most tests, with the exception of
the test aimed at testing decoding of sequences.
Reported-by: Amita Kapila
Discussion: https://postgr.es/m/d045f3c2-6cfb-06d3-5540-e63c320df8bc@enterprisedb.com
Branch
------
master
Details
-------
https://git.postgresql.org/pg/commitdiff/a180c2b34de0989269fdb819bff241a249bf5380
Modified Files
--------------
contrib/test_decoding/expected/ddl.out | 26 +++++++++++-----------
.../test_decoding/expected/decoding_in_xact.out | 2 +-
contrib/test_decoding/expected/replorigin.out | 8 +++----
contrib/test_decoding/expected/rewrite.out | 2 +-
contrib/test_decoding/expected/slot.out | 6 ++---
contrib/test_decoding/sql/ddl.sql | 26 +++++++++++-----------
contrib/test_decoding/sql/decoding_in_xact.sql | 2 +-
contrib/test_decoding/sql/replorigin.sql | 8 +++----
contrib/test_decoding/sql/rewrite.sql | 2 +-
contrib/test_decoding/sql/slot.sql | 6 ++---
10 files changed, 44 insertions(+), 44 deletions(-)