Grzegorz Śliwiński

Okay, that fixture would have to be separated at least by name from the proc/noproc duo: ```python from pytest_postgresql import factories postgresql_my_proc = factories.postgresql_proc() postgresql_table = factories.postgresql_table("postgresql_my_proc", dbname="subset_of_heavy_tests", optional_proc="postgresql_my_proc", load=[Path("/path/to/heavy.sql")])...

@Modjular the test database isn't created directly by postgresql_proc, but bu postgresql fixture. Instead of following the rollback approach every test, pytest-postgresql utilises database templates approach. Try following these steps:...

@Modjular I did some performance checks when I introduced and the performance of the templating database should be on par with transaction rollback. With additional plus here. For the transaction...

@astoff Being able to modify the template database in a separate fixture would be.... tricky. Could you describe to me your use case? From the process fixture up to the...

@astoff if you want to check the uuid4 functionality, then you should assign a Unique name for the Person, if that's not important, then you should assign a known uuid4...

Introducing #914 might allow you to do what you want, but... I'd not advise it. Having a fixture in between the process fixture and the test itself, that will add...

@GFuhr the issue is solved on master. I'll probably release the new functionality version next week, depending on the time if it'll let me. (I think the minimum would be...

@GFuhr check version 6.0.0. Mind that paths will now have to be pathlib.Path instances

* https://github.com/psycopg/psycopg/issues/391 * https://github.com/psycopg/psycopg/pull/62

@dmitry-shibanov I think I know the exact cause, i'm installing libmysql package that's being built on installation. And I have two runs on given python version on macos each is...