bug#28402: 25.2; shr.el uses shr-tag-img despite set shr-external-rendering-functions

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: v.schneidermann@gmail.com,  28402@debbugs.gnu.org
> Date: Thu, 05 Oct 2017 15:22:46 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > What if we make shr-indirect-call a defsubst?  Would that take care of
> > this issue?
> 
> Isn't it still (at least) an extra stack frame with a bunch of variable
> bindings?

Yes, but is that really that significant?  You sound like saying that
any non-trivial change in shr-descend should be rejected for these
reasons.  Is that really so?  Do we have measurements that would back
up such extreme care?

> This function is the central bit in shr, and should be as fast as
> possible.

As fast as possible, but not faster, I presume ;-)

Anyway, I see your point.  Vasilij, any thoughts or suggestions to
avoid reverting that part, while still keeping Lars happy?