Use LockMode::NONE as default lock mode

[BenMorel]

As discussed in doctrine/dbal#4391, the $lockMode parameter should not be nullable anywhere; what should be used as a default is LockMode::NONE.

There was a long standing confusion in the DBAL that can be summarized to:

• in all platforms but SQL Server / SQL Anywhere, null was equivalent to LockMode::NONE;
• in SQL Server / SQL Anywhere, LockMode::NONE resulted in using WITH (NOLOCK), which changed the effective isolation level to READ UNCOMMITTED for the table; this is not the contract of LockMode::NONE, and was qualified as a bug.

This has been fixed in doctrine/dbal#4400, and now null is not a valid value anymore for lock modes.

This PR is the sequel to the above:

• $lockMode defaults to LockMode::NONE everywhere, null is not allowed anymore
• no transaction is required anymore when using LockMode::NONE, which didn't make any sense

Notes:

• the transaction requirement for LockMode::NONE was highly inconsistent anyway: the Query required one, but EntityManager::find() didn't
• there were many inconsistencies in the documented types for $lockMode as well, which did document in many places that only an int was acceptable, but did accept or even expect null at some point: 1, 2, 3

[BenMorel]

Looks like composer lock file is broken; CI can't run.

[BenMorel]

@beberlei Could you please take a look at this PR? This should be an easy pick.

[SenseException]

You need to re-target your PR to a different branch, but I'm not sure if this is supposed to be a bugfix or an improvement for a newer branch. composer.lock isn't in the versioning in 2.8.x and 2.9.x anymore and shouldn't be a problem.

[BenMorel]

@SenseException This is not a bugfix, rather an improvement. Should this go into 2.9.x? What's master currently, is it 3.x?

[SenseException]

If master is the correct target branch, then it needs a new upmerge. It might be a current blocker.

[BenMorel]

That's the point, I'm not sure if master is the correct branch; is it 3.x?

[SenseException]

@doctrine/team-orm Is there a suggestion about the target branch for this PR?