Raise warning when using SQLite3 in-memory test database with --reuse-db

[schmitch]

Currently I've running several unittests with py.test that needs a database. The tests working fine.
But I#ve running with --reuse-dband the following database settings:

DATABASES = {
    'default': {
        'ENGINE': 'django.db.backends.sqlite3',
        'NAME': 'db.sqlite3'
    }
}

The Tests will work fine, but the databse db.sqlite3 won't getting created / the file is missing while the tests run / after the tests.
I use django unittests and run py.test with the following: py.test -v -s --ds=envisia.test.settings --reuse-db also I'm running on Python 3.4.1, pytest 2.6.1 and Django 1.7rc2.
On PostgreSQL the database will be created and will exists..

Edit: Also PostgreSQL tests run faster even when I run sqlite3 on :memory: which is somehow strange?! But I think that's cause of PostgreSQL --reuse-db works.

[blueyed]

The Tests will work fine, but the databse db.sqlite3 won't getting created / the file is missing while the tests run / after the tests.

For SQLite, the default test database will be :memory:.

See tests/settings_sqlite_file.py for how to force using a file based database:

DATABASES = {
    'default': {
        'ENGINE': 'django.db.backends.sqlite3',
        'NAME': db_name,
        # # Django > (1, 7)
        'TEST': {'NAME': test_db_name},
        # # Django < (1, 7)
        'TEST_NAME': test_db_name,
    },
}

Also PostgreSQL tests run faster even when I run sqlite3 on :memory: which is somehow strange?! But I think that's cause of PostgreSQL --reuse-db works.

I am not sure I understand what you mean. Do you have tests that are explicitly using PostgreSQL, even when you configure to use SQLite?

[pelme]

Maybe there should be an explicit error when using --reuse-db with in memory sqlite to avoid confusion? Or possibly a warning?

[schmitch]

@blueyed oh, damit I always though that 'NAME' will work, too..
And to PostgreSQL if my run my test suite with sqlite3 it is slower than postgresql. who cares I will close this since I've got it working.

[blueyed]

@pelme

Maybe there should be an explicit error when using --reuse-db with in memory sqlite to avoid confusion? Or possibly a warning?

A warning would be good.

@c-schmitt

And to PostgreSQL if my run my test suite with sqlite3 it is slower than postgresql.

You could run py.test with --durations=10 to get a report of the 10 slowest test parts (once with SQLite and then PostgreSQL), which might indicate something.
I could imagine that PostgreSQL with --reuse-db outperforms SQLite, but it seems a bit odd.

(Re-opening the issue for now, until we have decided how to proceed)