Re: RFM: How to make a buffer-local var in C layer

[Phillip Lord]

Stefan Monnier <address@hidden> writes:

> Both are C variables of type Lisp_Object.
> The "V" one contains the value of the variable
> `fontification-functions', while the Q one contains the symbol
> `fontification-functions'.

Ah, yes, okay.

>> Fsetq(Qfontification_functions,Qt)
>
> BTW, one notable difference (other than performance) between
>
>     Fset (Qfontification_functions, Qt);
> and
>     Vfontification_functions = Qt;
>
> is that the former will obey make-variable-buffer-local while the
> latter won't.  So if you want the var to be buffer-local, either use the
> former, or make sure to call Fmake_local_variable explicitly at
> some point.


Okay. Performance is irrelevant here, to be honest, so I've gone for
Fset.

I've pushed those changes up to my branch now and they appear to be
working. I think that I have all the C support that I need, and I'd
welcome any comments that you have (don't worry about all the commits
and change messages -- I'm planning on squashing before I'd think about
moving to master).

I'll add the emacs-lisp support next, which should be relatively
simple(.el).

Phil