Remove parking_lot dependency

[madsmtm]

Fixes #650.

std's Mutex has recently improved a lot, so there's not really a performance argument for keeping i any more. Also means we won't have to use the same windows-sys version as parking_lot

Dependency tree (Windows):

// Before
winit v0.27.1
├── bitflags v1.3.2
├── instant v0.1.12
│   └── cfg-if v1.0.0
├── log v0.4.17
│   └── cfg-if v1.0.0
├── once_cell v1.13.0
├── parking_lot v0.12.1
│   ├── lock_api v0.4.7
│   │   └── scopeguard v1.1.0
│   │   [build-dependencies]
│   │   └── autocfg v1.1.0
│   └── parking_lot_core v0.9.3
│       ├── cfg-if v1.0.0
│       ├── smallvec v1.9.0
│       └── windows-sys v0.36.1
│           └── windows_x86_64_gnu v0.36.1
├── raw-window-handle v0.5.0
│   └── cty v0.2.2
└── windows-sys v0.36.1 (*)

// After
winit v0.27.1
├── bitflags v1.3.2
├── instant v0.1.12
│   └── cfg-if v1.0.0
├── log v0.4.17
│   └── cfg-if v1.0.0
├── once_cell v1.13.0
├── raw-window-handle v0.5.0
│   └── cty v0.2.2
└── windows-sys v0.36.1
    └── windows_x86_64_gnu v0.36.1

Dependency tree (X11):

// Before
winit v0.27.1
├── bitflags v1.3.2
├── instant v0.1.12
│   └── cfg-if v1.0.0
├── libc v0.2.129
├── log v0.4.17
│   └── cfg-if v1.0.0
├── mio v0.8.4
│   ├── libc v0.2.129
│   └── log v0.4.17 (*)
├── once_cell v1.13.0
├── parking_lot v0.12.1
│   ├── lock_api v0.4.7
│   │   └── scopeguard v1.1.0
│   │   [build-dependencies]
│   │   └── autocfg v1.1.0
│   └── parking_lot_core v0.9.3
│       ├── cfg-if v1.0.0
│       ├── libc v0.2.129
│       └── smallvec v1.9.0
├── percent-encoding v2.1.0
├── raw-window-handle v0.5.0
│   └── cty v0.2.2
└── x11-dl v2.20.0
    ├── lazy_static v1.4.0
    └── libc v0.2.129
    [build-dependencies]
    └── pkg-config v0.3.25

// After
winit v0.27.1
├── bitflags v1.3.2
├── instant v0.1.12
│   └── cfg-if v1.0.0
├── libc v0.2.129
├── log v0.4.17
│   └── cfg-if v1.0.0
├── mio v0.8.4
│   ├── libc v0.2.129
│   └── log v0.4.17 (*)
├── once_cell v1.13.0
├── percent-encoding v2.1.0
├── raw-window-handle v0.5.0
│   └── cty v0.2.2
└── x11-dl v2.20.0
    ├── lazy_static v1.4.0
    └── libc v0.2.129
    [build-dependencies]
    └── pkg-config v0.3.25

• Tested on all platforms changed
  • Nope, please test this!
• Added an entry to CHANGELOG.md if knowledge of this change could be valuable to users
• Updated documentation to reflect any user-facing changes, including notes of platform-specific behavior
• Created or updated an example program if it would help users understand this functionality
• Updated feature matrix, if new features were added or implemented

[notgull]

This looks good to me; I'll always support removing a dependency that's been obsoleted by the standard library.

[notgull]

While this is entirely fine, I'd suggest having a macro like this somewhere in the crate:

macro_rules! lock {
    ($mutex: expr) => {{
        match ($mutex).lock() {
            Ok(guard) => guard,
            Err(err) => {
                log::warn!("Mutex is poisoned: {:?}", &err);
                err.into_inner()
            }
        }
    }}
}

And instead calling it like this:

lock!(XLIB_ERROR_HOOKS).push(hook);

This ensures two things:

1. A poisoned lock won't blow up the entire system.
2. If we decide to switch to another mutex scheme, for some reason, we can easily adjust lock! to fit our use case.

[madsmtm]

I went with unwrap because it's what's used in the other backends, and I'm not at all sure how to handle mutex poisoning

[notgull]

Hmm, I guess I should add it in a separate PR then...

[madsmtm]

That would be nice, please do!

[notgull]

Great! I'll wait until after this PR is merged, so I can make a macro that works across the entire crate.

[madsmtm]

Hey @notgull, this PR has been merged, so if you want to make this change, now you have the opportunity!