Lots of cleanup finished, INN 2.4 status

[Russ Allbery]

I think I've done all the major bits of cleanup that I recall planning on
doing for INN 2.4.  macros.h is no more, lots of macros have been ripped
out, and then for good measure I did a lot of mass conversion to strlcpy
and strlcat.

I'm now pretty happy with our ability to audit the code for string
overflows, although there are still some problem areas.  I have some ideas
in mind for fixing the problem areas in INN 2.5 or later, by adding a few
more functions to the buffer API and using buffers in more places instead
of static char arrays.  There are also a few places where internal APIs
should be changed to pass lengths around with variables.

I don't see any real problems, except possibly in a few places in
innfeed/imap_connection.c, just a few places where it requires a lot of
work to convince oneself that the code is okay.

I've also fixed a bunch of places where xmalloc and friends weren't being
used, and changed various calls to xmalloc plus other functions to
xcalloc, xstrdup, xstrndup, concat, or other more appropriate functions.

I'll put the current code on a test machine to flush out any cases where I
missed something.  I think I'm pretty much ready for INN 2.4 from my end.
I think the only other changes that we had outstanding were some patches
from Alex for the nnrpd cache and ESTALE in tradindexed and the changes to
deprecate access in readers.conf.

Should the Python changes go in before INN 2.4 as well?

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.