Heikki Linnakangas <hlinnaka@iki.fi> writes:
> Hmm, using the first value of REGRESS is surprising. How about we change
> the autoconf logic to match what meson does? As attached.
Unless you want to back-patch that as far as 9.2, it's going to make
matters worse not better for predictability of the DB names in the
cross-version upgrade tests. I think it might be better to make the
meson logic match the makefiles.
Having said that, I think what we should be doing is adding whichever
DB names we decide are problematic to the existing logic at about line
89 in AdjustUpgrade.pm:
# remove dbs of modules known to cause pg_upgrade to fail
# anything not builtin and incompatible should clean up its own db
foreach my $bad_module ('adminpack', 'test_ddl_deparse', 'tsearch2')
{
if ($dbnames{"contrib_regression_$bad_module"})
{
_add_st($result, 'postgres',
"drop database contrib_regression_$bad_module");
delete($dbnames{"contrib_regression_$bad_module"});
}
if ($dbnames{"regression_$bad_module"})
{
_add_st($result, 'postgres',
"drop database regression_$bad_module");
delete($dbnames{"regression_$bad_module"});
}
}
Perhaps another alternative is to add "DROP EXTENSION
injection_points" at the end of the test scripts in the
problematic modules.
regards, tom lane
