List of problems with EventEcho on windows

[ShrykeWindgrace]

EventEcho demo program in vty-crossplatform, built on win10+windows terminal, ghc-9.4.5 (lts-21.5)

• no mouse events (clicks or moves, but we can ignore moves for now); should be fixable by an additional flag in https://github.com/chhackett/vty-windows/blob/084d57abf7c2c2c9033150b630e3f3bfbfe00d2a/src/Graphics/Vty/Platform/Windows/Input.hs#L169
• presses on control keys (Ctrl/Alt/Shift) are all reported as Ctrl+@ (maybe that's by design, I'll check the behavior of that demo on linux tomorrow); it seems off anyway
• F-keys are not parsed; should be just a matter of plugging proper parsers that were implemented in the previous iteration

• Enter is not seen; should be fixable by a special entry in specialSupportKeys I strongly suspect that these issues are purely on the vty-windows side.

  @chhackett, what do you prefer? One big pull-request or multiple small pull-requests?

[chhackett]

Thanks for jumping in! I guess I'd prefer smaller pull requests. I plan on adding tests to vty-windows in the next week or two.

[ShrykeWindgrace]

On linux: same dev tools and repo versions, debian11+st+tmux (or debian11+konsole):

• no mouse events reported at all; I guess that comes from vty-unix, can not attest to that
• control keys are not reported at all; my intuition says that this is the correct behavior
• F-keys reporting is inconsistent across environments; e.g. Ctrl+F4 is not parsed at all under my konsole:

  -- konsole (xterm-256color)
  EvKey (KFun 29) []       -- that's Ctrl+F5
  EvKey (KChar 'S') []     -- and that's Ctrl+F4
  EvKey (KChar '2') []     
  EvKey (KChar 'O') [MMeta]

  -- st+tmux (tmux-256color)
  EvKey (KFun 29) []       -- that's Ctrl+F5
  EvKey (KFun 28) []     -- and that's Ctrl+F4

  @jtdaugherty is that the intended behavior for the unix version? I'll test vty/master some time later.

[jtdaugherty]

Frankly I have no idea, and terminal input is a mess so I wouldn't be surprised if the behavior you're seeing is in some sense "right." Modifier key support is especially spotty (not because of vty, but because of how keys are encoded in the terminal protocol). If you're willing, I'd be curious to see what those particular test cases above do on the master implementation of the event echo program. I suspect they'll be the same (i.e. I don't think the vty-unix split has changed behavior.)

[ShrykeWindgrace]

@jtdaugherty I agree, pretty sure that the behavior would be the same. I'll test next week, I am on windows till at least Monday.

[chhackett]

I'd like to note that I've checked in some changes to vty-windows. I believe the issues @ShrykeWindgrace identified here have been resolved, along with improving VT input parsing/handling overall. If Shryke agrees, I think we can close this issue.

[ShrykeWindgrace]

Yep, @chhackett fixed these issues, thanks!