In net/samba-libsmclient, Makefile includes ../samba3/Makefile, and ../samba3/Makefile includes samba-libsmbclient/Makefile.inc. When WRKDIRPREFIX is defined, samba3/Makefile trys to include ${WRKDIRPREFIX}/usr/ports/net/samba-libsmbclient/Makefile.inc and fails. How-To-Repeat: cd /usr/ports/net/samba-libsmbclient; make WRKDIRPREFIX=/var/tmp install
State Changed From-To: open->feedback Awaiting maintainers feedback
Hi all! My fault, I've copied this lines from another port, without deep thinking about their meaning... So, proposed fix is correct and should be applied to the current samba3 port(dunno, is it possible due freeze right now). Thanks, Niimi, for spotting it so quickly. Kris, that's the fix to the problem you've reported, as well as to the ports/93688. So, can one of the commiters apply this fix ASAP? (Tom has connectivity problems right now, so can't do it by himself...). With best regards, Timur.
State Changed From-To: feedback->closed Tested and committed, thanks.
Responsible Changed From-To: freebsd-ports-bugs->ade Tested and committed, thanks.
On Wed, Feb 22, 2006 at 05:09:58AM +0100, Timur I. Bakeyev wrote: > Hi all! > > My fault, I've copied this lines from another port, without deep > thinking about their meaning... > > So, proposed fix is correct and should be applied to the current samba3 > port(dunno, is it possible due freeze right now). Thanks, Niimi, for > spotting it so quickly. > Kris, that's the fix to the problem you've reported, as well as to the > ports/93688. So, can one of the commiters apply this fix ASAP? (Tom has > connectivity problems right now, so can't do it by himself...). The port is still broken: http://pointyhat.freebsd.org/errorlogs/i386-errorlogs/e.5-exp.2006020203/samba-libsmbclient-3.0.21b.log This port must be fixed immediately since it's in the dependency chain for KDE and GNOME, and it's blocking the testing work that we need to finish before the start of the freeze. If someone can fix it in the next hour we can proceed, otherwise I'll have to revert the broken updates. Kris
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Feb 22, 2006, at 12:36 , Kris Kennaway wrote: > The port is still broken: > > http://pointyhat.freebsd.org/errorlogs/i386-errorlogs/e.5-exp. > 2006020203/samba-libsmbclient-3.0.21b.log Hrm. Packaged fine here. http://tinderbox.lovett.com/ui/index.php?action=describe_port&id=153 Is this with rev 1.148 of net/samba3/Makefile? The problem was there, not within the net/samba-libsmbclient port itself. - -aDe -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) iD8DBQFD/M0mpXS8U0IvffwRAlaGAJ0Wosks/Obv5sVRj3SD1oKRiHqHGgCfWYec IQKtiQfgLfEsJfQCvyhaxLc= =gqdh -----END PGP SIGNATURE-----
On Wed, Feb 22, 2006 at 12:44:22PM -0800, Ade Lovett wrote: > > On Feb 22, 2006, at 12:36 , Kris Kennaway wrote: > > The port is still broken: > > > > http://pointyhat.freebsd.org/errorlogs/i386-errorlogs/e.5-exp. > > 2006020203/samba-libsmbclient-3.0.21b.log > > Hrm. Packaged fine here. > > http://tinderbox.lovett.com/ui/index.php?action=describe_port&id=153 > > Is this with rev 1.148 of net/samba3/Makefile? The problem was > there, not within the net/samba-libsmbclient port itself. Ah, it was indeed missing that update. I'll retry. Kris
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Having looked in more detail at net/samba3 and net/samba- libsmbclient, I'm now of the opinion that the latter should not really be a slave port, but a port in its own right. The amount of hackery that goes on to shoehorn samba-libsmbclient as a slave to samba3 is, imo, simply not worth the pain, given the alternative of simply maintaining two independent ports, and ensuring that distinfo et al. is updated for both on an update. - -aDe -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) iD8DBQFD/YbXpXS8U0IvffwRAtPwAJkB/x7wMlK8DHxY2wN7swdx/S5hpQCeIRSh IRZVyOw//XebEQSLxN1QWTM= =hD9X -----END PGP SIGNATURE-----
Hi all! On Thu, Feb 23, 2006 at 01:56:39AM -0800, Ade Lovett wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Having looked in more detail at net/samba3 and net/samba-libsmbclient, I'm now > of the opinion that the latter should not really be a slave port, but a port in > its own right. I don't agree, that it is necessary. Before the last update slave/master relation was very clear and easy to grasp - compare with net/py-samba, which is of the same level of complexity. Recent hack with Makefile.inc was added mostly for prettiness of OPTIONS knob and is inspired by it - we still can't make it work without inclusion of <bsd.pre.ports.mk>... > The amount of hackery that goes on to shoehorn samba-libsmbclient as a slave to > samba3 is, imo, simply not worth the pain, given the alternative of simply > maintaining two independent ports, and ensuring that distinfo et al. is updated > for both on an update. I don't see much hackery besides that Makefile.inc. All the rest, including tricky Kerberos 5 detection is done for the samba3 port and, in case of split, should be duplicated in the libsmbclient port. BTW, there is jet another port that depends from net/samba3 - ja/samba3. In general, my opinion is that current state of samba3 ports is stable enough and they should remain as is. With best regards, Timur.