Re: Is Elisp really that slow?

[Stefan Monnier]

> As C-Mode has no associated terminal nor it composes messages, the
> request would amount to "unbind C-c C-c",

That's right.

In my book, prog-mode should bind C-c C-c to `compile` and C-mode has no
need to have its own binding for it (because I can't think of a better
way to "compile/run" C code than via `compile`).  And it definitely
doesn't need a specific binding for comment-region since M-; already
covers this need.

> something no current or future C/C++ user would benefit from.

They would because the bindings would be more uniform across modes.

> Pretending that every mode conforms to the same rules is
> counterproductive, because there are vast differences among them.

Pretending that every detail of every major mode is unique is similarly
counter-productive.

> We shall strive for pragmatism, not for bureaucracy.

This is not about bureaucracy.  It's actually about streamlining the
overall system (by removing an unneeded chunk of code from CC-mode in
this case).

> comment-dwin != comment-region

I don't think that justifies having both bind to such short
key-sequences, nor does it justify binding comment-region to
a key-sequence usually bound to something completely unrelated.

> In principle, fixing those would be beneficial. OTOH, forcing existing
> users to adapt just for the cause of coherence with modes they don't use
> is not desirable either.

Short term pain for longer term gain.
There's no question that it's a tradeoff.

This particular C-c C-c binding in CC-mode has been on the "short term
gain for long term pain" for more than 10 years.


        Stefan