Quotes affect REPL highlighting

Currently it treats the entire racket-repl-mode buffer as something to font-lock using mostly the same rules as a racket-mode edit buffer.

Is this good or bad?

A REPL buffer is kind of a hopeless mix of:

prompts
input expressions, possibly invalid with unmatched delimiters
output values from the print in read-eval-print-loop
other output from prints and writes in your end user program
arbitrary output from your program, e.g. (display "I am not a Racket expression").

Strictly speaking, only some of this stuff should be font-locked as Racket expressions or values.

Given this, you could argue that it's a mistake to font-lock the buffer, at all. Eventually it will produce a surprising or bad result. It can also be slow when the program prints many, large values.

You could argue it should be disabled by default, and advise users they can use racket-repl-mode-hook to enable font-lock-mode if they want it.

On the other hand, in my own years of using it I've found the font-lock to be useful enough -- it helps me visually parse things -- to like it despite the occasional glitches.

So the status quo is it's enabled by default, and you can use racket-repl-mode-hook disable font-lock-mode.

So my personal opinion is that it's not really a bug, except maybe a documentation bug? But I'm open to discussion.

I think there's a potential feature, which you hinted at: Have the back end somehow distinguish which parts of current-output-port come from print vs. display so the front end can font-lock those or not, as appropriate. And maybe add more "region distinctions" in the front end as necessary.

This feels like a lot of work (plus the risk of breaking things) for a not-huge benefit. At least that's my initial feeling.

It might make more sense to undertake this as part of the work to support a lang lexer mode (and I'll tag this issue with that, for now).

greghendershott / racket-mode

Quotes affect REPL highlighting #633