Allows to quit on WM_QUIT message for Windows

[koutoftimer]

Pull requests consist of the following changes:

• constant's type changed from u32 to i32 to stay consistant with winapi constant's type
• listen method modified to interupt infinite loop when WM_QUIT message received
• added post_quit_message helper function that simply calls PostQuitMessage that sends WM_QUIT
• added a bit of docs for listen and post_quit_message

From PostQuitMessage docs:

Indicates to the system that a thread has made a request to terminate (quit).

I have verified that posting WM_QUIT by calling PostQuitMessage message from different process have no effect on the listen loop.

[jamesbirtles]

Looks good, my only concern is that by making the key codes of type i32, it technically makes the windows and linux versions slightly incompatible I guess. I'm not sure if this is really a concern or not though?

also would this not also need to change to i32?
https://github.com/UnwrittenFun/hotkey-rs/blob/1610e8e7726b316fc48a6f132162a42a3af58b15/src/windows.rs#L53

[koutoftimer]

I've got errors about type mismatch by preparing this pull request, that is why I've changed it to i32 and my primary focus was on correct interupting infinite loop, but now I see what happened with that slight change.

I can't see any negative constants, maybe it is "better" to cast i32 to u32.

it technically makes the windows and linux versions slightly incompatible

I'm not a developer of winapi crate and have no idea why do they choose to use i32 for VK_* constants and isize for modifiers. Maybe it is better to ask them whether or not it is safe to cast constants to u32.

I understand that compatibility is an important question, but it is not what I'm concerned with right now. I do not like that "callbacks" are called in the same thread, blocking infinite loop and dispatching events which corresponds to event miss and a ton of additional code for library user.

Recalling your conserns: I do not know right answer.