Re: Lisp primitives and their calling of the change hooks

[Stefan Monnier]

> Yes, that's what I've done.  Each zlib-decompress-region now has two
> separate pairs of b/a-c-f, one for insert, then the other for delete.
> This is the way that other "coding" primitives, like
> base64-encode-region work.

LGTM.

>> In practice, I'd also expect both b-c-f and a-c-f to be nil (or
>> equivalent) when we call zlib-decompress-region.
> Yes.  But it's nice to have things consistent.  :-)

Agreed.  There's no justification for inconsistency here.

> Anyhow, here's the patch, which appears to work as described above.  Any
> comments?  Otherwise I'll commit it.

The subsequent patch which tries to reduce the number of calls to
bcf/acf seems worse because it's more complex (and the marginal benefit
doesn't seem to justify the complexity).


        Stefan