Message7923
Applied Stuart's patch to the 2.4.0 code. A few changes:
1. when checking to see if the db exists, add a conn.close() for
the schema branch.
2. add some more logging when a command errors. Include the
command and arguments being run.
3. add logging to schema path to parallel logging to non-schema path.
Also to set up a role/user (rounduptest) to be able to create a schema
in the database rounduptest run:
grant create on database rounduptest to rounduptest;
To add/remove createdb to a role use:
alter role rounduptest CREATEDB;
alter role rounduptest NOCREATEDB;
Also get a better name for db_schema_split(). Maybe get_db_schema_names().
TBD testing:
1) set up another rounduptest_schema db with rounduptest_schema user
2) verify that trying to initialize the db with
RDBMS_NAME=rounduptest_schema
user fails with Runtime Error with permission error.
3) initialize with the RDBMS_NAME=rounduptest_schema.rounduptest
4) run postgresql tests against using rounduptest_schema.rounduptest.
This will take quite a while. Maybe consider a subset of the full
postgresql test suite for use by the schema branch? How to do this?
Maybe use pytest mark pg_schema registered using pytest_configure
to limit the run somehow? |
|
Date |
User |
Action |
Args |
2023-12-27 03:45:02 | rouilj | set | messageid: <1703648702.39.0.817721430909.issue2550852@roundup.psfhosted.org> |
2023-12-27 03:45:02 | rouilj | set | recipients:
+ rouilj, ber, smcgraw |
2023-12-27 03:45:02 | rouilj | link | issue2550852 messages |
2023-12-27 03:45:01 | rouilj | create | |
|