Run the client and CLI tests with the testbed server in postgre mode

[vxgmichel]

Those tests typically never run with postgresql, which could help find bugs.

Here is the procedure:

# [Shell 1] Run a postgresql cluster 
$ cd server
$ PG_CLUSTER_PORT=38135 pytest --run-postgresql-cluster
Creating PostgreSQL cluster...
PostgreSQL url:  postgresql://postgres@localhost:38135/postgres
usage: PG_URL=postgresql://postgres@localhost:38135/postgres py.test --postgresql tests
Press enter when you're done with...

# [Shell 2] Run a testbed server with the cluster
$ parsec testbed --with-postgresql postgresql://postgres@localhost:38135/postgres
2025-01-24T15:53:55.832483Z [info     ] Parsec version                 version=3.2.5-a.0+dev
All set !
Don't forget to export `TESTBED_SERVER` environ variable:
export TESTBED_SERVER='parsec3://127.0.0.1:6770?no_ssl=true'
2025-01-24T15:53:55.855393Z [info     ] Started server process [120107]
2025-01-24T15:53:55.855655Z [info     ] Uvicorn running on http://127.0.0.1:6770 (Press CTRL+C to quit)

# [Shell 3] Run the client tests in sequential mode
$ export TESTBED_SERVER='parsec3://127.0.0.1:6770?no_ssl=true'
$ cargo nextest run -p libparsec_client -j 1

# Then run the client tests in sequential mode
$ export TESTBED_SERVER='parsec3://127.0.0.1:6770?no_ssl=true'
$ cargo build -p parsec-cli
$ cargo nextest run -p parsec-cli -j 1 

Those tests could be integrated in a daily job in the CI.