diff options
| author | Tom Lane <tgl@sss.pgh.pa.us> | 2023-01-17 16:00:39 -0500 |
|---|---|---|
| committer | Tom Lane <tgl@sss.pgh.pa.us> | 2023-01-17 16:01:06 -0500 |
| commit | 58393ea56ef8e75bcd60af1f88d0a81b1d8e21eb (patch) | |
| tree | 303fcb509df5978e30e625e125e34f42be61a069 /src | |
| parent | 0e4d38100acb90d93e68694674050f4cebdd816a (diff) | |
| download | postgresql-58393ea56ef8e75bcd60af1f88d0a81b1d8e21eb.tar.gz postgresql-58393ea56ef8e75bcd60af1f88d0a81b1d8e21eb.zip | |
AdjustUpgrade.pm should zap test_ext_cine, too.
test_extensions' test_ext_cine extension has the same upgrade hazard
as test_ext7: the regression test leaves it in an updated state
from which no downgrade path to default is provided. This causes
the update_extensions.sql script helpfully provided by pg_upgrade
to fail. So drop it in cross-version-upgrade testing.
Not entirely sure how come I didn't hit this in testing yesterday;
possibly I'd built the upgrade reference databases with
testmodules-install-check disabled.
Backpatch to v10 where this module was introduced.
Diffstat (limited to 'src')
| -rw-r--r-- | src/test/perl/PostgreSQL/Test/AdjustUpgrade.pm | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/test/perl/PostgreSQL/Test/AdjustUpgrade.pm b/src/test/perl/PostgreSQL/Test/AdjustUpgrade.pm index dbd755cedf6..47c2352e586 100644 --- a/src/test/perl/PostgreSQL/Test/AdjustUpgrade.pm +++ b/src/test/perl/PostgreSQL/Test/AdjustUpgrade.pm @@ -88,12 +88,13 @@ sub adjust_database_contents } } - # avoid version number issues with test_ext7 + # avoid no-path-to-downgrade-extension-version issues if ($dbnames{contrib_regression_test_extensions}) { _add_st( $result, 'contrib_regression_test_extensions', + 'drop extension if exists test_ext_cine', 'drop extension if exists test_ext7'); } |