Allow keybindings with modifiers

[y5nw]

• Goal of the PR
  This PR allows users to create keybindings with modifier keys, such as Control-F10 or Shift-5.
• How does the PR work?
  This PR extends the KeyPress class to include information on the state of modifier keys.
• Does it resolve any reported issue?
  No.
• Does this relate to a goal in the roadmap?
  No
• If not a bug fix, why is this PR needed? What usecases does it solve?
  This provides a partial workaround for SDL: Some keybinds broken due to missing character lookup #14545 by allowing users to use e.g. Shift-7 instead of / for keybindings. Some people may want to continue using e.g. Shift-7, which will no longer be possible with SDL: Use scancodes for keybindings #14964.

Notes

• This PR now treats Left Shift and Right Shift as the same key.
• This PR removes the incompatible feature where Shift+key binds to the "character" rather than the keycode. However, my testing shows that this removed feature did not fully work anyway. (Update: This feature is already dropped by SDL: Use scancodes for keybindings #14964).

To do

This PR is a Draft.

• Adapt input handler to handle such keybindings.
• "Resolve" keybindings to perform action(s) with the closest keybinding.
• Update GUI (depends on Move keybinding settings to (Lua-based) setting menu #15791)
• Fix bugs/regressions
  • Cannot bind an action to the Shift or Control key.
  • Cannot stop sneaking after pressing Shift once.
  • Cannot quit keybinding form with Esc after binding a key to Shift or Ctrl
  • Cannot perform action bound to a normal key when Shift-(key) or Ctrl-(key) is pressed.
  • Chat key does not work when bound to Modifier+Autoforward.
• Add tests
  • KeyPress
  • KeyList
  • MyEventReceiver (src/client/inputhandler.h)

How to test

• The testcases are complemented to reflect the changes made in this PR.
• The above section includes a list of bugs/regressions that occurred during the development of the PR. These should not be reproducible.

[y5nw]

Closing for now. #15934 has higher priority IMO, and its (potential) followups are more impactful.