All ports depending on postgresql shall use the USE_PGSQL=yes knob
defined in Mk/bsd.ports.mk. Bumping portrevisions where needed.
PR: 75344
Approved by: portmgr@ (kris), ade & sean (mentors)
loosely based on the PR mentioned (the original patch in the PR does not
handle either library dependencies nor changing the default credentials).
PORTREVISION *not* bumped, this can wait until I implement WITH_PGSQL_PASS
and WITH_PGSQL_SERVER in the near future.
PR: 61998
Submitted by: Sascha Luck <sascha@c4inet.net>
the original Makefile.in got duplicated in the 'patched' one, leading
to a malformed sh(1) construct that would inevitably break the installation.
Thanks to Ronald Sebastian and Alex Dupre for pointing this out.
Noticed by: Ronald Sebastian <mailsagan@yahoo.com> and
Alex Dupre <sysadmin@alexdupre.com>, among others
While I'm here, remove the autoconf part of the Makefile.in patch, which
actually duplicates what I've been doing in the port's Makefile for
some time now :)
Revamp a bit the creation of the MySQL credentials file.
PR: 59605
Submitted by: Alex Dupre <sysadmin@alexdupre.com> (mostly)
Approved by: portmgr (will)
and other assorted fixes, mostly from Alex's PR.
PR: 56462
Submitted by: Alex Dupre <sysadmin@alexdupre.com>
Ob-woohoo: run-time configuration of default-domain and MySQL info;
at last it is possible to build a single package for
deploying on multiple servers!
Add the WITH_VPOPMAIL_SQWEBMAIL_CREATION knob to make vadduser
(well, make_user_dir(), actually) create an sqwebmail-config file
specifying the outgoing messages' sender address.
- WITH_MYSQL_LIMITS works now, move it where users can actually see it
and remove the BROKEN mark;
- remove the vmysql.h patch that defined the 'limits' table creation -
it was fixed in the meantime;
- remove the annoying configure script check for id == 0.
(username, password, database name, etc) - when I did the changeover
from ${PERL} to ${REINPLACE_CMD}, I plain forgot that sed does not
really like Perl-style $1 subexpression references :(
Reported by: Ventsislav Velkov <veno@evrocom.net>,
"Pavel Novikov" <pavel@ext.by>
The mail/vpopmail and mail/vpopmail-stable ports retain their Perl
dependency, since the pkg-install script is written in Perl. I think
this might need to be changed when I get the time to look into it.
Submitted by: Ulrich Spoerlein <q@uni.de>
Submitted by: gerald <gerald@hannibal.symicon.com> - the original
notification about the vpopmail distfile checksum change
Fix a possible buffer underflow bug in the changed code.
Add a new ENABLE_MYSQL_LIMITS knob for .qmailadmin-limits processing, and
DISABLE the new MySQL limits code - it does not even compile as shipped :(
Use the correct configure script option for specifying the path to
the tcprules program, so vpopmail works even with LOCALBASE != /usr/local.
Submitted by: Mike Meyer <mwm@mired.org> - the tcprules patch
It seems that a couple of new features were added, but I have to take
a closer look at them to make sure nothing is wrong. This will probably
happen as soon as I wake up in a couple of hours.
Submitted by: gerald <gerald@hannibal.symicon.com>
Pointy hat to: Inter 7 for adding new features to an already "released"
tarball instead of releasing a new version :(
autoconf or automake over our local patches can ruin your whole day.
This rarely, if ever, affected the vpopmail ports, causing mysterious
failures at install time, but inevitably killed the stunnel build if
automake-1.5 happened to be installed.
Reported by: Dan Langille <dan@langille.org> for stunnel,
Sitchin Zech <mailsagan@yahoo.com> for vpopmail
There are two new configuration knobs:
- WITHOUT_USERS_BIG_DIR disables the use of 'big' directories for users,
thereby saving a little on disk space and inode usage; IMO, the
savings are negligible;
- WITHOUT_SEEKABLE disables the newly introduced ability of vdelivermail
to examine its input stream for seekability, and read it all into a
file to make it seekable. This works quite fine in the tests that I
have done, so there seems to be no need to enable it on FreeBSD;
however, if things should go wrong (in particular, if delivering mail
to your virtual domains seems to fail), try rebuilding vpopmail with
this knob.
Prompted by: David Phillips <david@acz.org>
support (spelled WITH_CLEARTEXT_PASSWD now), along with a couple of
other tweakable knobs. There is UNTESTED support for Sybase and Oracle
database backends; please report any successes or failures to me.
APOP support discontinuation reported by: "A. Postle" <apostle@kolbas.it>
Introduce two new build knobs:
- WITH_CLEAR_PASSWD actually enable the clear passwords management
that was turned on by default in 5.2 (hello, POLA!)
- WITH_MD5_PASSWORDS for the encrypted passwords, use MD5 hashes.
Prompted by: Claus Rasmussen <claus@webclaus.com>
ports to work once again.
NB: The vpopmail documentation says that the special support for
sqwebmail authorization is no longer needed for sqwebmail 3.0 and above.
If your sqwebmail installation does not work after this vpopmail update,
make sure to try vpopmail built both with and without the WITH_SQWEBMAIL
knob enabled.
Requested and patiently awaited by: much too many to list..
Tested by: NERvOus <nervous@nervous.it>,
Valentin Zahariev <curly@e-card.bg>
the inc_deps and lib_deps files. This should fix qmailadmin,
courier-imap and possibly others, after I broke them by my configure
changes to not create and populate ${PREFIX}/vpopmail/etc/ at
configure time.
Add a couple of configuration knobs.
Bump PORTREVISION for the inc_deps/lib_deps fix.
Reported by: a multitude of innocents :(
Pointy hat: roam (myself)
binaries and libraries of vpopmail. This prevents arbitrary users
from obtaining the username and password for access to the MySQL database,
which are present in plaintext in the vpopmail object files and binaries.
Obtained from: Bugtraq
Let the user override the MySQL username, password and db at build time.
Let the user override the vchkpw gid and the vpopmail uid.
Bump PORTREVISION.
Approved by: nbm (maintainer)
that have been reported to me.
These changes don't affect the package, just if people use some of the
WITH_* or WITHOUT_* options
Submitted by: Yuri L Khachaturian <yukh@ns.otradnoe.net>
