Re: [exim] Segfault in 4.86 (mail queue handleing)

> On 6/10/2015, at 2:52 am, Heiko Schlittermann <hs@???> wrote:
>
> The +fixes branch was invented to provide fast access to the most
> important subset of changes without the need for new release.
>
> But you're right, if the urgency for (some of) the patches rises, we
> should release a 4.86.1 or whatever.

For what it’s worth, I would also agree that bug 1671 would be a valid reason to have a 4.86.1.

We upgraded to 4.86 this week (I am working on moving over to +fixes, but haven’t managed that yet, and missed the significance of this bug in the changes included there), and hit this on a lot of servers.

It is a fairly minor impact (a number, sometimes very large, of extra frozen messages in the queue until they are forced or otherwise unfrozen), but seeing a lot of segfaults (in the Exim log and, e.g. syslog) is quite concerning, and it’s hard to be completely sure that the segfault occurred after delivery was successful, so the bug can look at lot worse than it actually is. Since it’s tricky to reproduce (unless you know what’s happening) that makes it even more worrisome.

There are other changes in +fixes that I’d really like to see - but in my opinion only 1671 would warrant a 4.86.1.

As always, thanks heaps to everyone for all the work they do on Exim; it’s greatly appreciated!

Regards,
Tony