Map users to core ports on first button press

[nijm]

Description

The first user to press a button is mapped to core port 0 (player 1), the second is mapped to core port 1 (player 2), and so on.

Adds a 'None' option to the 'Mapped Port' setting for each user under Quick Menu->Controls->Port N Controls.

Configured by the input_remap_on_button_press option. When enabled, all users are unmapped by default (assigned to 'None'/MAX_USERS).

Hotkeys are still only usable by user 0. I have another PR to address that: Hotkeys follow player 1 #18353

My Use Case

I have 6 wireless controllers:

• 2x 8BitDo M30 6 button MegaDrive style controllers
• 2x SNES style controllers
• 2x PSX style controllers

I prefer to play games with the type of controller they were designed for, but I like having the option to use any of them. Each pair of controllers is identical, so it's hard to know if I've picked up SNES controller 1 or 2 - with this feature I can use either.

With the existing device type reservation feature I ensure each controller type is always mapped to the same 2 users:

• user 1: 8BitDo MD controller
• user 2: 8BitDo MD controller
• user 3: SNES Controller
• user 4: SNES Controller
• user 5: PSX Controller
• user 6: PSX Controller

With core input remapping files I can map the core controls in a sensible way for each controller type. This depends on each controller type always being mapped to the same users. I leave the mapped port setting unset and disable saving the remapping file on exit.

Screenshots

Related Issues

Controller Priority Option (request)
#6707

Related Pull Requests

Hotkeys follow player 1 #18353

Works well with this merged PR:
Controller device reservation / preference.
#16647

[hizzlekizzle]

This is a major (optional) change for most platforms, but it's similar to how Android has always worked--though I guess only at a core level rather than the frontend level.

A significant number of users want to use specific original-style controllers, and this seems like a nice way of streamlining that process.

@sonninnos

[nijm]

Rebased on master as there were some conflicting changes in input_driver_poll. Also made it compatible with the input overlay, which I hadn't considered before.