vfs_zfsacl.c: use of undeclared identifier 'ace'

[Jeremy Allison]

On Mon, Nov 13, 2017 at 07:02:32PM +0000, Youzhong Yang wrote:
> I don't disagree. But FreeBSD does ship zfsacl.so: 
> 
> /usr/local/lib/shared-modules/vfs/zfsacl.so
> 
> Last month or so I've been playing with Linux (Centos, Ubuntu) and FreeBSD, in terms of tracing and debugging, Solaris/illumos is still unbeatable. If FreeBSD can have 'mdb' integrated, it would be very attractive. I know someone is working on it ...
> 
> Anyway, it's not a place to debate which OS is good or bad. For Solaris/illumos, can I volunteer to be the tester? We can try to have some retired hardware dedicated for this purpose. I'd like to try my best to make this happen, although I cannot guarantee the resource and my time. Please let me know.

No, I'm not passing judgement, other than saying that
the fact we can't get RichACLs into the Linux kernel
due to one intransigent developer is a f%&king disgrace :-(.

I'm happy for you to keep riding us hard on FreeBSD/Illumos
breakage (although less on the Illumos - now Oracle killed
Solaris the pale, hungry children are doomed to die a slow,
ignominious death :-).

But without auto-regression tests on FreeBSD triggered on
every code submission, we are going to get occasional
compile breakage, and I don't see a way around that.