Re: Attachment counting feature at risk.
On 2006-04-28 16:51:32 -0500, David Champion wrote:
> I'm certainly willing to review this and try to make
> corrections, if you're willing to keep the feature if fixed.
> It's clear that there is a bug in the code.
I think I quashed the coding bug that's there. (Or, maybe, not.)
The point, however, is that attachment counting requires rather
thorough inspection of message bodies, and that this is an
expensive exercise. On the other hand, just inspecting message
headers is quite efficient, but it's not enough for attachment
counting.
> But do remember that it doesn't come up at all unless you're
> using %X/~X, so there's no harm to performance in the
> immediate term. The only damage (if you're not asking for
> it) is in code purity.
Well, my concern is that we're having a feature here that will
make mutt unexpectedly sluggish without users having any reason
to expect this to happen.
--
Thomas Roessler · Personal soap box at <http://log.does-not-exist.org/>.