feat(core): allow disabling identity map and change set tracking

[B4nan]

Sometimes we might want to disable identity map and change set tracking for some query.
This is possible via disableIdentityMap option. Behind the scenes, it will create new
context, load the entities inside that, and clear it afterwards, so the main identity map
will stay clean.

const users = await orm.em.find(User, { email: '[email protected]' }, {
  disableIdentityMap: true,
  populate: { cars: { brand: true } },
});
users[0].name = 'changed';
await orm.em.flush(); // calling flush have no effect, as the entity is not managed

Keep in mind that this can also have
negative effect on the performance.

Closes #1267

[dbettini]

@B4nan hello, I just found this issue while searching for a way to populate only specific relations on a model, without the possibility of loading more data as a consequence of using identity map.
What is the reason for using this.fork(false) instead of this.fork(true)? Looking at the docs, it seems like we don't get a clear identity map this way, but clone the existing one. In other words, the current implementation uses the existing identity map, it just doesn't update it, which will result in sometimes getting more data than required. I.e.:

const one = repo.findOne(id, ['firstRelation'])
const two = repo.findOne(id, { populate: ['secondRelation'], disableIdentityMap: true });

In the above example, two will also have firstRelation loaded, which would not be expected by the naming of disableIdentityMap.
Would you consider changing this behavior? Or, if not, is there any other way to accomplish this, other than manual forking?

[B4nan]

I though it was required for the population to work, but looks like the tests are passing without clearning too. I am fine with the change in v5, for v4 it seems semi breaking.

[dbettini]

Cool, I'll stick to manual forking until v5 then. Thanks for the quick reply!