RFC: PoC C1 Control Sequences

[lwi]

Issues I discovered:

• we have "char" in the codebase everywhere and sometimes casts to "unsigned int" which could be very wrong. we should migrate all places to "unsigned char". Then we can avoid casts at places where it counts.
• we do not handle UTF-8 (self) synchronization and probably by looking at the source handle such cases very badly.
• considering the set of C1 control sequences, the handling of ESC and control sequences should be merged as they are can be refered to from C1 at the same time. we should also get rid of the naming concept of "escape sequences" since with C1 we can start a control sequence without escape.

This pull request just shows a proof of concept embedding C1 in escape_simple and _osc and handling it like an "escape sequence". Also see https://github.com/TragicWarrior/libvterm/issues/147

[lwi]

I added support for UTF-8 encoded C1 control characters.

[TragicWarrior]

@lwi , the code looks good to me. are you content with?

[lwi]

@TragicWarrior the code is okayish as it is, however the original issues remain (in the code there is only a concept of "escape sequences" implying they start with a ESC which is not accurated anymore with C1 controls). Also the code supports raw C1 and UTF-8 encoded. xterm only supports them raw, VTE only UTF-8 encoded. UTF-8 encoded C1 are very rare and Fedora would be one of the first users. A very strict UTF-8 terminal emulator (like tmux) would probably discard raw C1 because it can be mistaken for broken UTF-8 (an out of context UTF-8 continuation byte to be exact). So its debatable whether and to what extend we should support C1 controls. That being said, if C1 control support in raw and UTF-8 encoded form is fine with you, this pull request should be fine.