[ubuntu/precise] postgresql-common 124 (Accepted)
martin.pitt at ubuntu.com
Sat Oct 15 11:33:12 UTC 2011
postgresql-common (124) unstable; urgency=medium
* Keeping urgency medium, as 123 did not make it into testing yet.
* PgCommon.pm, cluster_info(): Use /bin/ps instead of reading /proc/*/comm,
the latter was reported to not work on some systems.
* pg_ctlcluster: Use PgCommon::check_pidfile_running and drop duplicated
* pg_wrapper: The previous multiarch globbing was wrong, as it looked for
the architecture in uname() (which doesn't work on i386, where uname says
i686). Now get the multiarch library path from whereever psql expects
libedit.so to be, so that this even works if you install postgresql-client
for a foreign architecture.
postgresql-common (123) unstable; urgency=medium
* Urgency medium, as bug 644078 slipped into testing, and is quite a nasty
regression for people that it affects.
* t/001_packages.t: Check that the -plpython3 package is installed for
* t/020_create_sql_remove.t: Check that the "plpython3u" language works.
* t/020_create_sql_remove.t: More thoroughly check PL/Perl and PL/PerlU.
* PgCommon.pm, cluster_info(): If we have a PID file and can read it (i. e.
as root), prefer doing this over probing the port. This delivers the
correct result with e. g. "pg_ctlcluster restart" when the port got
changed in the configuration file. (Closes: #643850)
* pg_wrapper: Drop dpkg-architecture call; we don't want to introduce a
dependency to dpkg-dev, and DEB_HOST_MULTIARCH also isn't available in
backports. Just extend the libreadline.so globbing to cover multiarch
paths as well. (Closes: #644078)
* t/070_non_postgres_clusters.t: Test pg_upgradecluster. This reproduces
* pg_upgradecluster: Supply socket argument for the tablespace test. Thanks
to Piotr Szydełko for the patch! (Closes: #644477)
Date: 2011-10-08 21:26:32.689144+00:00
Changed-By: Martin Pitt <martin.pitt at ubuntu.com>
-------------- next part --------------
Sorry, changesfile not available.
More information about the Precise-changes