bug#33275: 27.0.50; Image cache pruning

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: 33275@debbugs.gnu.org
> Date: Mon, 05 Nov 2018 20:13:27 +0100
> 
> If I understand you correctly, you seem to favour introducing code that
> will break code that's working today (if you insert an image into a
> buffer today in a memory-pressure situation, that won't signal an error
> today) instead of tweaking a non-documented ad-hoc caching strategy (by
> taking the image cache size into consideration), or fixing the caching
> (by using weak hash tables), and I'm not quite sure why.

Because your use case is not important enough to change image caching
strategy that works well for many years.

> `image-size' didn't use to cache images, but was introduced as an
> optimisation.

And that optimization had a reason, right?

Btw, why doesn't xmalloc return NULL in your case before invoking the
OOM killer?  Is that expected behavior on GNU/Linux systems?