Re: so_rerror

[Roy Marples <roy%marples.name@localhost>]

On 06/11/2018 00:31, Christos Zoulas wrote:
> I agree as I wrote above, the issue is to do it in a non-intrusive way or
> make sure that nothing breaks...

I'm not sure how much more non intrusive it can be - everything I've 
done is standards compliant, documented by POSIX and implemented 
previously in at least one other OS.

bind even had code for the error path (although it made the wrong choice).
Many other applications such as nsd and unbound also have prior existing 
code for it but thankfully made the right choice about what to do.

> And we discover more and more things
> breaking.

As it stands, the *only* application that's been reported broken since 
8.0 without any existing error path for it which I'm aware of is syslogd 
because logging the error as you pointed out only makes the problem 
worse. We've now fixed that, and you've said you're happy with it.

So, I'll ask. What are these more and more things that are breaking ? 
Bbecause they are not reported in gnats, irc or to an email list to 
which I belong.

Roy