Implement for par_iter and par_iter_mut on Query

[TheRawMeatball]

No description provided.

[Frizi]

Shouldn't that variant be called par_map?

[TheRawMeatball]

Alright, I rewrote the whole thing, and now it exposes par_iter methods on queries which let you use the parallel iterator interface in bevy_tasks.

[CGMossa]

I find myself in need of this PR. I have many places where I am filtering, and collecting based on query-data, and I wish to make these parallel executed.
If I can help with this PR, then that would be great.

[ManevilleF]

What is the status of this MR ?

[TheRawMeatball]

I'll fix it up eventually, though probably not that soon, given it's likely to break again.

[alice-i-cecile]

@TheRawMeatball, are you interested in cleaning up the PR description and rebasing? Or should this wait on future work.

Totally fine if it's not a priority, but I'll add a Abandoned label so other folks know to pick up the work. This would be valuable, so I don't want to leave it sitting forever.

[ManevilleF]

I think this PR could be very valuable, it can avoid using Mutex or RwLock when extracting parallelized block data

[mockersf]

Query::par_iter and Query::par_iter_mut now exists