Ready for 2.3.10?

[Ken Murchison]

Wesley Craig wrote:
> This is a show stopper, i.e., it results in data loss:
> 
>     https://bugzilla.andrew.cmu.edu/show_bug.cgi?id=2919
> 
> The two fixes are pretty trivial.  I have other comments on running 
> ctl_mboxlist -m from cyrus.conf START, but those are probably not as 
> critical as these fixes.

Applied, with associated manpage update.


> This is broken:
> 
>     https://bugzilla.andrew.cmu.edu/show_bug.cgi?id=2971
> 
> Not really a "show stopper" but it consistently generates useless noise 
> on the support lists.  I don't believe:

I know.  Still looking for the "right" solution.  This doesn't nag me 
enough to float to the top of my TODO list.


>     https://bugzilla.andrew.cmu.edu/show_bug.cgi?id=2805
> 
> is the correct solution, either (tho that has some non-et issues that 
> ought to be teased out).

The other fixes were applied back before 2.3.8.

I spent a lot of time cleaning things up so that the number of warnings 
from 'gcc -W -Wall' are minimal.  The one that has me baffled is this:

warning: type-punning to incomplete type might break strict-aliasing rules

Anyone know what exactly this means and how to fix it?  I can give a 
specific file:lineno is required.


Other than that, there are still some signed/unsigned warnings in 
lib/cyrusdb_skiplist.c and imap/index.c, but I left those along for now, 
because I was afraid to break anything.  I'll probably take another 
crack at them after 2.3.10.


-- 
Kenneth Murchison
Systems Programmer
Project Cyrus Developer/Maintainer
Carnegie Mellon University