Add non-existent entity behavior to Has doc

[nicopap]

Objective

Has<T> in some niche cases may behave in an unexpected way.

Specifically, when using Query::get on a Has<T> with a despawned entity.

Solution

Add precision about cases wehre Query::get could return an Err.

[Nilirad]

I hope I don't sound too rude, but to me, at least, it is obvious that Query::get is fallible, because it returns a result. If I mindlessly unwrap it I should not expect to not encounter errors.

Also, I don't understand how this is specific to Has, as it seems to involve Query::get in general.

[nicopap]

This PR is a follow up on this discord discussion: https://discord.com/channels/691052431525675048/730525730601041940/1185535439461941298.

So we are at least 2 (me and @NiseVoid) who would overlook it.

The reasoning was "if this matches every existing entities, it's safe to unwrap". Unlike say Query<&Transform> which only match entities with a Transform component, Query<Has<Transform>> matches every entities. But the reasoning is flawed for the reasons mentioned in the new doc paragraph.