On 03/23/12 15:50, Jan Stary wrote:
On Mar 23 13:59:42, RD Thrush wrote:
On 03/23/12 09:01, Jan Stary wrote:
On Mar 23 07:11:45, RD Thrush wrote:
My PORTSDIR is on an nfs server. Mounting the particular nfs
directory on /usr/ports (and setting PORTSDIR accordingly) fails the
new test in bsd.port.mk.
Can you please show how exactly you are mounting it,
and how exactly it fails?
I think my reply to Stuart has those details.
If you 'mount server:/some/dir /usr/ports',
then you shouldn't need to set PORTSDIR at all,
(because it's the default /usr/ports, right?).
I tried that originally but had the problem that triggered the patch.
Is possibly /usr/ports a symlink itself on your machine
(the NFS client)?
Yes, it is. But PORTSDIR is a real directory (nfs mounted).
In the details you provided, PORTSDIR is something under /x2/...
about which we don't know anything.
Thanks. My bad. I was trying to provide details on the i386 environment but
cut the PORTSDIR line from the amd64 /etc/mk.conf which uses the x2 box The
correct cut is:
PORTSDIR=/a8v/pub/OpenBSD/current/ports
and more detail on a8v which is also -current:
v1:logs/i386 81>mount | grep a8v
a8v:/pub on /a8v/pub type nfs (nodev, nosuid, read-only, v3, udp, timeo=100,
retrans=101)
Apparently test -h considers an nfs mount the same as a symlink...
No it doesn't; 'test -h foo' only evaluates as true for symlinks.
I looked at the test manpage as well and had a similar conclusion.
However, I thought I'd try to specifically test for not a directory
rather than for a symlink and had success. I don't yet understand
why.
Because symlinks are followed in test(1).
When the current builds finish, I'll add code to bsd.port.mk to learn more.
