I've run into this tool, love the idea of migra and tusker complements it nicely, in order to stop committing endless small migrations files, and just commit the whole schema, goddammit.
I was trying to run tusker diff with a schema.sql generated by a pg_dump --schema-only, modified to remove what I think was useless for the business rules/logic. Apparently the final .sql was somehow invalid (missing a schema creation). But tusker/migra was not showing that error, I just had a sqlalchemy error on the bottom of a long SQL output:
(Background on this error at: https://sqlalche.me/e/14/f405)
Looking into it, the output was actually truncated by my terminal scrollback size, and the eror(psycopg2.errors.InvalidSchemaName: schema "***" does not exist) was all at the top.
I guess one way to have better UI for this would be to catch the exception in or around execute_sql_file, print the error (and not the whole SQL), because you never know maybe the user is a doofus is trying to migrate to an invalid database schema :-)
I've run into this tool, love the idea of migra and tusker complements it nicely, in order to stop committing endless small migrations files, and just commit the whole schema, goddammit.
I was trying to run
tusker diff
with a schema.sql generated by apg_dump --schema-only
, modified to remove what I think was useless for the business rules/logic. Apparently the final .sql was somehow invalid (missing a schema creation). But tusker/migra was not showing that error, I just had a sqlalchemy error on the bottom of a long SQL output:Looking into it, the output was actually truncated by my terminal scrollback size, and the eror(
psycopg2.errors.InvalidSchemaName: schema "***" does not exist
) was all at the top.I guess one way to have better UI for this would be to catch the exception in or around
execute_sql_file
, print the error (and not the whole SQL), because you never know maybe the user is a doofus is trying to migrate to an invalid database schema :-)