meaning of number of lines in the message (%l in index_format)

[Moritz Barsnick]

On Tue, Jun 25, 2019 at 21:01:44 +0200, Vincent Lefevre wrote:
> Note that the %c value also depends on the encoding, but this may
> be less surprising.

Quite obviously. It also includes PGP and S/MIME boilerplate and so on.
Misleading regarding the size of the actual text, but quite clear about
the message size.

> > hard to see the difference between 1.3k and 1.3M at a quick glimpse
> > across the index.
>
> Yes, perhaps the reason I do not use %c. And the fact that the
> scaling facteur is on the right instead of the left makes this
> even harder (that's a bit like the American dates MMDDYYYY).

That's mutt_pretty_size(). If we wanted to overengineer this, this
function could be expanded or cloned to be fixed to a certain
granularity (such as k or M). (I'm not making this a feature request.
OTOH, why not craft my own patch? ;-))

Cheers,
Moritz