Re: request to reconsider libnettle/libhogweed

[Ted Zlatanov]

On Thu, 2 Mar 2017 09:58:57 -0800 Paul Eggert <address@hidden> wrote: 

PE> On 03/02/2017 06:59 AM, Ted Zlatanov wrote:
>> * we already link to GnuTLS, which means those C functions are available
>> already.

PE> If Emacs's module system were always present and worked well, this could be
PE> supported via a module that consisted entirely of glue, i.e., a module that
PE> merely exposes C functions already present in Emacs. Since the module system
PE> isn't guaranteed, though, it may be better to create new built-in Elisp
PE> functions for this, at least for now.

There won't be anything in these functions that precludes moving them.

PE> Is there some way that we can say "these functions are built-in now, but 
may be
PE> moved to a module later"? If not, perhaps there should be.

I'll be happy to annotate them as needed, but I don't know who will care
about such annotations? Perhaps it's better to mark the functions as
dependent on a particular Emacs feature, sort of like
`gnutls-available-p' but as a per-function tag.

On Thu, 02 Mar 2017 17:55:07 +0200 Eli Zaretskii <address@hidden> wrote: 

EZ> Please show the patch, perhaps after updating it to match the current
EZ> master, and let's discuss then.

That's the technical side, and I'll gladly do it if John agrees it's
acceptable in principle. I'd rather not spend hours on it otherwise.

Ted