Post History

The answer is "it depends". It depends on the language, how much buffering you're willing to allow, how much you're willing to accept some approximate/incorrect syntax highlighting, and whether you'd accept control codes that correct the syntax highlighting after the fact.

If your bar for syntax highlighting was simply that language keywords, operators, and other "punctuation" are highlighted, then almost all programming languages allow this with only a small amount of buffering needed. Some amount of buffering is needed because I can't tell if "i" is part of the keyword "if" until I see the next two characters.

On the other hand, if you wanted to highlight unmatched brackets, then you may not be able to correctly highlight an opening bracket at the beginning of the file until the very end of the file.

This is an issue in most languages for many non-trivial uses of syntax highlighting. Specifically, this is an issue with markdown if you'd like to, for example, use bold text for text wrapped in **s.

As the highlighting of this paragraph indicates, whether the text following ** is made bold depends on whether there is a following ** among other rules. I didn't do any escaping in this case.

Ultimately, the only (non-trivial) solution in general is for the streaming syntax highlighting to make a guess and sometimes be wrong. If allowed, it might be possible to use control codes to edit the previously displayed output when the syntax highlighter has enough context to realize it made a mistake.

[`bat`](https://github.com/sharkdp/bat), for example, just assumes `**` starts a bold span and makes the following text bold even when this is incorrect.