[Blocked] Relax render/compute pass lifetimes

[kvark]

Mostly reverts #168 given that wgpu-core is now properly handling error IDs.
Closes #603
Closes #188

Has a major positive usability impact 🚆 It is no longer required by the type system to keep resources alive for the duration of a render/compute pass encoding. It's still required by the runtime: dropping a pass will produce an error if some of the dependencies are dropped before it.

Looking at the error messages may be difficult until an approach similar to gfx-rs/wgpu#931 (comment) is rolled out in wgpu-rs (cc @scoopr). We'll make it a priority before releasing wgpu-0.7.

Edit: actually, the IDs here are not "error IDs" that wgpu-core handles. It will panic straight upon seeing such a stale ID...

[monocodus]

This is an autogenerated code review.

Checker summary (by rust_clippy):
The tool has found 10 warnings, 1 errors.

The .monocodus config not found in your repo. Default config is used.
Check config documentation here

[scoopr]

I think the "stale IDs" will need to be handled more gracefully anyway. Consider, pushErrorScope, loadScene, popErrorScope, I would imagine should work without paniccing, no matter if some of the IDs are invalid on creation or dropped at the wrong time.

I'm so conflicted about this. I do love when the compiler tells me something is wrong, but in this particular case, the papercut is so such a big wound that I do hope good documentation and error messages will be enough.

[monocodus]

This is an autogenerated code review.

Checker summary (by rust_clippy):
The tool has found 10 warnings, 1 errors.

The .monocodus config not found in your repo. Default config is used.
Check config documentation here

[ghost]

Just curious, what is this blocked by? Deciding if I should wait for this.

[kvark]

It's blocked because of this:

Edit: actually, the IDs here are not "error IDs" that wgpu-core handles. It will panic straight upon seeing such a stale ID...

We don't want the panics for this, since WebGPU API doesn't allow panics there. At the same time, we don't want to add special handling of this in wgpu-core because it would have no way to tell if the ID was wrong to begin with (in which case we should panic, since this can't happen in WebGPU proper), or just got removed after it was used...

So @definitelynotrobot should probably not wait for this. If you are constrained by the lifetimes here, consider typed_arena to keep stuff that needs extended lifetimes.

[pythonesque]

I actually really like the lifetime discipline enforced here, I'd much rather get an error at compile time than runtime for stuff like this.

[kvark]

Project moved to wgpu