Know issues
## Document / Syntax
We would like to introduce an incremental parsing phase (see `tree-sitter`).
Because, when you have tokens (which may be as simple as words) or an AST,
completion / suggestion / highlighting / validation become easy.
So we need to send events to a lexer/parser, update `Document` accordingly.
And fix `Completer` / `Hinter` / `Highlighter` API such as they have access to `Document`.
See [lex_document](https://python-prompt-toolkit.readthedocs.io/en/master/pages/advanced_topics/rendering_flow.html#the-rendering-flow).
## Repaint / Refresh
Currently, performance is poor because, most of the time, we refresh the whole line (and prompt).
We would like to transform events on prompt/line/hint into partial repaint.
See `termwiz` design (`Surface`).
## Action / Command
We would like to support user defined actions that interact nicely with undo manager and kill-ring.
To do so, we need to refactor current key event dispatch.
See `replxx` design (`ACTION_RESULT`, `action_trait_t`).
## Line wrapping (should be fixed with verions >= 6.1.2)
On Unix platform, we assume that `auto_right_margin` (`am`) is enabled.
And on Windows, we activate `ENABLE_WRAP_AT_EOL_OUTPUT`.
But on Windows 10, `ENABLE_WRAP_AT_EOL_OUTPUT` and `ENABLE_VIRTUAL_TERMINAL_PROCESSING` seems to be incompatible.
## Colors
We assume that ANSI colors are supported.
Which is not the case on Windows (except on Windows 10)!