On Thu, Aug 8, 2024 at 3:08 AM Joe Orton <jor...@redhat.com> wrote:
>
> On Wed, Aug 07, 2024 at 09:18:29AM -0400, Eric Covener wrote:
> > On Wed, Aug 7, 2024 at 8:40 AM Joe Orton <jor...@redhat.com> wrote:
> > > In CI this seems to be hanging on MacOS - possibly in this new test,
> > > though the job was not even starting for a while so I'm not 100% sure.
> > >
> > > https://github.com/apache/apr/actions/runs/10282155892
> > >
> > > Can anybody with MacOS try the new test case and see what happens?
> > >
> > > ./testall -v testshm
> >
> > For me it failed, both the fchown and fchmod if I change the order.
> >
> > estshm             : |Line 315: Could not change permissions of shm
> > segment (22): Invalid argument
>
> Thanks. I wonder if this true on BSDs more generally as well, at least
> the MacOS build is using shm_open() and POSIX is clear that fchmod
> should work on a shared memory object:
>
> https://pubs.opengroup.org/onlinepubs/007904975/functions/fchmod.html
>
> I disabled the test for DARWIN in r1919728.

Should we do APR_ENOTIMPL in the runtime?

> Anybody with a "real" BSD who can also test this?

None here, but from the other side of the family tree AIX and Solaris are OK.

-- 
Eric Covener
cove...@gmail.com

Reply via email to