Re: Specifiers

[David Kastrup]

Stefan Monnier <address@hidden> writes:

> Yes, pretty much, except we don't have to worry about combinations of
> buffer-local with other forms of foo-local at the same time (well,
> supposedly we have to worry about it for buffer-local + frame-local,
> but we already know we don't handle that correctly).

That's supposed to be an advantage?

>> and won't get any worse with the proposed new feature.
>
> Try it.
>
>> ISTR that there is already code in specbind to try to DTRT
>> for some of these.
>
> Yes, there's already a fair bit of code trying to get things to work
> right for buffer-local variables (in specbind and a bunch of other
> places).  It's already pretty complex just like that.

I don't quite see what handling independent heaps of complexity for the
three kinds of locality we currently support (and window-locality is
about to come, too, even though for faces, _another_ complexity source
handled independently at the moment) is supposed to be buying us.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum