debathena-reactivate (2.0.41) unstable; urgency=low

  * Bump the chroot's ramdisk to 50% of RAM or 1G, whichever is higher
    (Trac: #1207)
  * Cleanup cached user dmrc files and AccountService entries (Trac:
    #1259)

 -- Jonathan Reed <jdreed@mit.edu>  Wed, 26 Sep 2012 17:08:56 -0400

debathena-reactivate (2.0.40) unstable; urgency=low

  * Cleanup from r24057, where we had to special-case quickstations via
    su-error and sudo-error, and cleanup those conffiles in a Debian-esque
    way
  * Add a trailing NUL to the end of sudo-warning to workaround a bug
    in sudo 1.8.3p1 (fixed upstream, but not in Precise)
  * Stop scribbling over sudoers and use sudoers.d (Trac: #1219)
  * Stop building for lucid.

 -- Jonathan Reed <jdreed@mit.edu>  Fri, 07 Sep 2012 14:39:54 -0400

debathena-reactivate (2.0.39) unstable; urgency=low

  * Yes, the pam_env hack is still needed.

 -- Jonathan Reed <jdreed@mit.edu>  Wed, 08 Aug 2012 14:33:36 -0400

debathena-reactivate (2.0.38) unstable; urgency=low

  * Ship a session wrapper that runs initgroups(3) and stop abusing sudo
    (Trac: #928)
  * Stop depending on kexec-tools, which hasn't been needed since 2.0.25
  * Wrap Depends line for easier reading
  * Drop the schroot pam_env hack, it's not needed anymore (and stop
    diverting /etc/pam.d/schroot)

 -- Jonathan Reed <jdreed@mit.edu>  Wed, 01 Aug 2012 12:52:32 -0400

debathena-reactivate (2.0.37) unstable; urgency=low

  * Set the time in reactivate, because NTP apparently can't (Trac: #749)
  * Implement a blacklist for services that should never start inside the
    chroot

 -- Jonathan Reed <jdreed@mit.edu>  Mon, 09 Jul 2012 14:16:53 -0400

debathena-reactivate (2.0.36) unstable; urgency=low

  * Sigh.  Order matters in the mounts file because with the /run
    transition, /run/shm needs to get bind-mounted after /run, otherwise
    it ends up with the wrong mode bits.

 -- Jonathan Reed <jdreed@mit.edu>  Thu, 05 Jul 2012 13:24:33 -0400

debathena-reactivate (2.0.35) unstable; urgency=low

  * Update debian/copyright for 3-clause BSD and one GPL'd file
  * Ship 16killprocs-no-really to end sessions faster (Trac: #775)
  * Use new lightdm-config hooks to actually run session setup and cleanup
  * Recover from a 'set -e' failure in snapshot-run and don't leave
    chroots lying around if possible
  * Avoid race conditions by inhibiting dbus-daemon-launch-helper during a
    chroot shutdown (Trac: #1052, probably)

 -- Jonathan Reed <jdreed@mit.edu>  Tue, 03 Jul 2012 11:45:03 -0400

debathena-reactivate (2.0.34) unstable; urgency=low

  * dbus-daemon-launch-helper moved in precise
  * Bump Standards-Version

 -- Jonathan Reed <jdreed@mit.edu>  Tue, 05 Jun 2012 17:03:44 -0400

debathena-reactivate (2.0.33) unstable; urgency=low

  * Actually install the diverted lightdm-session.debathena file

 -- Jonathan Reed <jdreed@mit.edu>  Tue, 05 Jun 2012 13:52:59 -0400

debathena-reactivate (2.0.32) unstable; urgency=low

  * Remove unnecessary invocation of cp(1)

 -- Benjamin Kaduk <kaduk@mit.edu>  Fri, 01 Jun 2012 00:37:29 -0400

debathena-reactivate (2.0.31) unstable; urgency=low

  * Stop creating a symlink from /var/run/athena-nologin to /etc/nologin,
    because other thinks like to clobber /etc/nologin.  Our greeter (which
    you have to use on cluster), will check for our own flag file and deny
    logins
  * Use new Debathena-Build-For field in control file to control releases
    this gets built on.
  * Set NSS_NONLOCAL_IGNORE when running getent, because we care about
    local groups, not Hesiod ones
  * Add the user to "sudo" as well, since "admin" is gone in Precise (LP:
    893842)
  * Use SESSION_ID, not CHROOT_NAME in schroot setup script.  SESSION_ID
    was always right, and CHROOT_NAME simply happened to match it due to a
    bug (Debian: 623828)
  * Also divert lightdm-session (lightdm's session wrapper) and wrap it in
    snapshot-run
  * Transform the schroot fstab in the schroot config script to deal with
    the /run transition

 -- Jonathan Reed <jdreed@mit.edu>  Wed, 30 May 2012 17:44:07 -0400

debathena-reactivate (2.0.30) unstable; urgency=low

  * No-change rebuild to pick up changes from r25453 (making the various
    scripts executable)

 -- Jonathan Reed <jdreed@mit.edu>  Thu, 03 May 2012 13:15:13 -0400

debathena-reactivate (2.0.29) unstable; urgency=low

  * Save lots of debugging information if we are rebooting in reactivate.
    It will go to athena-reactivate.log for later forensics.

 -- Benjamin Kaduk <kaduk@mit.edu>  Tue, 08 Nov 2011 21:19:39 -0500

debathena-reactivate (2.0.28) unstable; urgency=low

  * No change rebuild to pick up config-package-dev 4.13 changes

 -- Jonathan Reed <jdreed@mit.edu>  Wed, 26 Oct 2011 18:32:01 -0400

debathena-reactivate (2.0.27) unstable; urgency=low

  * Leave a reason for why we're rebooting in the athena-reactivate log.

 -- Benjamin Kaduk <kaduk@mit.edu>  Sun, 16 Oct 2011 23:10:32 -0400

debathena-reactivate (2.0.26) unstable; urgency=low

  * Run D-Bus-spawned services inside the schroot session, if any.

 -- Geoffrey Thomas <geofft@mit.edu>  Wed, 14 Sep 2011 22:02:58 -0400

debathena-reactivate (2.0.25) unstable; urgency=low

  * Stop kexec'ing, because modern distributions are incapable of doing so
    without sucking

 -- Jonathan Reed <jdreed@mit.edu>  Tue, 23 Aug 2011 10:01:49 -0400

debathena-reactivate (2.0.24) unstable; urgency=low

  * Fix the patch for Trac #656 to use the correct regexp

 -- Jonathan Reed <jdreed@mit.edu>  Tue, 26 Jul 2011 11:28:19 -0400

debathena-reactivate (2.0.23) unstable; urgency=low

  * sudo now checks its supplementary group list, not NSS, for group
    membership, and so it doesn't consider you in the admin group until
    after you run initgroups(3). Since we use sudo itself to run
    initgroups, we need to work around this somehow... the simplest fix
    is to explicitly add $USER to /etc/sudoers inside the chroot.

    At some point we may want to write our own setuid initgroups(3)
    helper and avoid sudo entirely.
  * If chattr fails due to the file not existing, that shouldn't abort
    the rest of reactivate.

 -- Geoffrey Thomas <geofft@mit.edu>  Tue, 19 Jul 2011 17:07:54 -0400

debathena-reactivate (2.0.22) unstable; urgency=low

  * Workaround for schroot/libpam-afs-session stupidity (Trac: #928)

 -- Jonathan Reed <jdreed@mit.edu>  Mon, 11 Jul 2011 16:03:09 -0400

debathena-reactivate (2.0.21) unstable; urgency=low

  * Programs that run df no longer get confused about /tmp. (Trac #656)
    (Thanks to Jon Reed for this patch.)

 -- Liz A. Denys <lizdenys@mit.edu>  Wed, 29 Jun 2011 16:07:21 -0400

debathena-reactivate (2.0.20) unstable; urgency=low

  * policy-rc.d should only return the cups daemons that exist (Trac
    #701).  Fix plagiarized from debathena-cupsys-config's restart_cups.sh

 -- Jonathan Reed <jdreed@mit.edu>  Thu, 10 Mar 2011 23:38:25 -0500

debathena-reactivate (2.0.19) unstable; urgency=low

  [ Jonathan Reed ]
  * Fix package name in maintainer scripts

  [ Evan Broder ]
  * Replace an awful Perl idiom with a much less awful one in the prerm.

 -- Evan Broder <broder@mit.edu>  Thu, 27 May 2010 13:29:16 -0400

debathena-reactivate (2.0.18) unstable; urgency=low

  * Add "lp" to the list of groups we add the user to, to support all
    those cluster machines with parallel port printers attached (Trac #92)

 -- Jonathan Reed <jdreed@mit.edu>  Thu, 15 Apr 2010 14:12:08 -0400

debathena-reactivate (2.0.17) unstable; urgency=low

  * Remove the autostart .desktop files for nm-applet so that users don't
    users don't see configurations they cannot change. (Trac: #400)

 -- Liz A. Denys <lizdenys@mit.edu>  Tue, 13 Apr 2010 22:06:35 -0400

debathena-reactivate (2.0.16) unstable; urgency=low

  * Don't kexec back into recovery mode - regardless of whether or not
    recovery mode was the *very* last thing on the command line.

 -- Evan Broder <broder@mit.edu>  Wed, 07 Apr 2010 00:23:32 -0400

debathena-reactivate (2.0.15) unstable; urgency=low

  [ Greg Brockman ]
  * Change transform_su to error if it fails to apply. (Trac: #416)

  [ Evan Broder ]
  * Don't kexec back into recovery mode.

 -- Evan Broder <broder@mit.edu>  Fri, 02 Apr 2010 19:06:45 -0400

debathena-reactivate (2.0.14) unstable; urgency=low

  * Fix a typo in 2.0.13.

 -- Evan Broder <broder@mit.edu>  Fri, 02 Apr 2010 14:42:17 -0400

debathena-reactivate (2.0.13) unstable; urgency=low

  * Manually enable kexec on logout-triggered reboots. (Trac: #536)

 -- Evan Broder <broder@mit.edu>  Thu, 01 Apr 2010 12:40:22 -0400

debathena-reactivate (2.0.12) unstable; urgency=low

  * Transform /etc/default/schroot to end sessions, not recover them
  * Add build-depends on schroot

 -- Jonathan Reed <jdreed@mit.edu>  Thu, 25 Mar 2010 12:46:37 -0400

debathena-reactivate (2.0.11) unstable; urgency=low

  * Fix a typo in 2.0.9.
  * Merge reactivate's CUPS-restarting logic, which waits for BrowsePolling to
    complete, with debathena-cupsys-config's, which checks to make sure
    that cups is running first, and use a common file for both.

 -- Geoffrey Thomas <geofft@mit.edu>  Mon, 08 Mar 2010 17:21:25 -0500

debathena-reactivate (2.0.10) unstable; urgency=low

  * Be more robust in the preinst when fixing schroot.conf.
  * Bump the debhelper compatibility level as high as we can.

 -- Evan Broder <broder@mit.edu>  Sat, 06 Mar 2010 02:59:55 -0500

debathena-reactivate (2.0.9) unstable; urgency=low

  * Reboot on update if we haven't rebooted for the bind-mount fix from
    2.0.8 already.
  * Bump standards version (no changes required).

 -- Evan Broder <broder@mit.edu>  Wed, 03 Mar 2010 12:05:07 -0500

debathena-reactivate (2.0.8) unstable; urgency=low

  * Fix the check to determine whether we have already bind-mounted
    /media. (Trac: #493)
  * Clean up the per-user files cached locally by the new gdm.
    (Trac: #527)
  
 -- Robert Basch <rbasch@mit.edu>  Mon, 01 Mar 2010 09:52:26 -0500

debathena-reactivate (2.0.7) unstable; urgency=low

  * /lib/init/mount-functions.sh is gone as a side effect of Karmic
    increasingly _actually_ using Upstart, so stop using it and just
    implement the (very small) logic we wanted from it ourselves.
  * Add a two-minute timeout to postinst code that could be an infinite
    loop in appropriately unfortunate circumstances.

 -- Geoffrey Thomas <geofft@mit.edu>  Sat, 13 Feb 2010 00:47:52 -0500

debathena-reactivate (2.0.6) unstable; urgency=low

  * Permit (possibly-backported) upstream schroot at a version that
    includes our patches as an alternative to the debathena-schroot
    dependency.

 -- Geoffrey Thomas <geofft@mit.edu>  Fri, 12 Feb 2010 00:44:06 -0500

debathena-reactivate (2.0.5) unstable; urgency=low

  * Bind-mount /boot (read-only) and /var/log (read-write), but not
    /home. (Trac: #314)
  * Deter people from thinking that /home will persist by getting rid of
    it in the chroot.

 -- Evan Broder <broder@mit.edu>  Mon, 18 Jan 2010 08:29:07 -0500

debathena-reactivate (2.0.4) unstable; urgency=low

  * Add some convenience functions to snapshot-run for running commands in
    the chroot.
  * Make sure that a group exists before we add users to it.
  * Ignore removal of the CUPS remote queue cache.

 -- Evan Broder <broder@mit.edu>  Mon, 21 Dec 2009 12:34:02 -0500

debathena-reactivate (2.0.3) unstable; urgency=low

  * Now that policy-rc.d isn't blindly allowing stops and starts for
    anything in the chroot, we should be sure that the chroot actually
    knows it's a chroot before we try to start CUPS in it.

 -- Evan Broder <broder@mit.edu>  Mon, 23 Nov 2009 16:14:12 -0500

debathena-reactivate (2.0.2) unstable; urgency=low

  * Clean up the CUPS remote queue cache on logout and package upgrades.
  * Fix the policy-rc.d to actually allow for restarting cups and cupsys
    inside the chroot. Apparently pipes can delineate patterns, but don't
    mean anything within a single pattern, such as the one one pattern
    that results from variable expansion.

 -- Evan Broder <broder@mit.edu>  Mon, 23 Nov 2009 15:51:42 -0500

debathena-reactivate (2.0.1) unstable; urgency=low

  * Various bugfixes:
    - Preserve the KRBTKFILE and KRB5CCNAME variables when running
      schroot. (Trac: #441)
    - Run root commands in /, not $HOME.

 -- Evan Broder <broder@mit.edu>  Tue, 17 Nov 2009 18:12:45 -0500

debathena-reactivate (2.0) unstable; urgency=low

  * Rewrite to use schroot 1.3 and aufs union mounts instead of LVM.
    This makes logins on a Dell 760 take about 8 seconds to reach the
    desktop.

 -- Geoffrey Thomas <geofft@mit.edu>  Sun, 08 Nov 2009 16:36:48 -0500

debathena-reactivate (1.25.6) unstable; urgency=low

  * Fix init script metadata (overlooked in 1.21).

 -- Geoffrey Thomas <geofft@mit.edu>  Fri, 16 Oct 2009 23:57:22 -0400

debathena-reactivate (1.25.5) unstable; urgency=low

  * geofft's transformation of /etc/pam.d/su was running outside the
    chroot, and editing the file in place, thus clobbering the symlink. We
    need to fix all of this in the postinst.

 -- Evan Broder <broder@mit.edu>  Wed, 14 Oct 2009 14:47:45 -0400

debathena-reactivate (1.25.4) unstable; urgency=low

  * Always lecture users trying to sudo on quickstations.
  * Stop editing /etc/pam.d/sudo in the snapshot setup, since we're no
    longer diverting it.

 -- Evan Broder <broder@mit.edu>  Wed, 14 Oct 2009 13:52:32 -0400

debathena-reactivate (1.25.3) unstable; urgency=low

  * Specify both of the lecture files in /etc/sudoers: error for
    everybody, and warning for the admin group. The latter overrides the
    former.

 -- Evan Broder <broder@mit.edu>  Wed, 14 Oct 2009 12:39:11 -0400

debathena-reactivate (1.25.2) unstable; urgency=low

  [ Evan Broder ]
  * Use PAM instead of the sudo lecture_file to lecture users.

  [ Geoffrey Thomas ]
  * Do that only outside the cluster login session, and preserve
    lecture_file within the chroot.
  * Revert 1.25.1, since it doesn't actually change anything.

 -- Geoffrey Thomas <geofft@mit.edu>  Wed, 14 Oct 2009 08:58:58 -0400

debathena-reactivate (1.25.1) unstable; urgency=low

  * Display the sudo lecture file for all users, not just the admin
    group. This ensures it will show on quickstations, not just clusters.

 -- Evan Broder <broder@mit.edu>  Sun, 11 Oct 2009 12:20:11 -0400

debathena-reactivate (1.25) unstable; urgency=low

  * Don't put the user in a login snapshot on a quickstation.

 -- Evan Broder <broder@mit.edu>  Wed, 07 Oct 2009 11:45:09 -0400

debathena-reactivate (1.24) unstable; urgency=low

  * Change back to mounting the login snapshot as ext3, since that was
    substantially hurting performance.

 -- Evan Broder <broder@mit.edu>  Mon, 05 Oct 2009 10:12:40 -0400

debathena-reactivate (1.23) unstable; urgency=low

  * Mount the login snapshot file system as ext2, with noatime.

 -- Robert Basch <rbasch@mit.edu>  Wed, 30 Sep 2009 17:49:49 -0400

debathena-reactivate (1.22) unstable; urgency=low

  * Use the new /usr/share/debathena-gdm-config/{Pre,Post}Session.d
    directories in gdm-config.

 -- Evan Broder <broder@mit.edu>  Mon, 28 Sep 2009 14:44:49 -0400

debathena-reactivate (1.21) unstable; urgency=low

  * Avoid sound during first login (Trac: 292).

 -- Michael Mekonnen <mikemeko@mit.edu>  Sun, 20 Sep 2009 15:31:13 -0400

debathena-reactivate (1.20.1) unstable; urgency=low

  * Use the right debian/rules variable for removing files.

 -- Evan Broder <broder@mit.edu>  Wed, 17 Jun 2009 23:21:08 -0400

debathena-reactivate (1.20) unstable; urgency=low

  * Remove the autostart .desktop files for jockey-gtk and update-notifier
    so that users don't get notifications about things they can't change.

 -- Evan Broder <broder@mit.edu>  Wed, 17 Jun 2009 23:07:00 -0400

debathena-reactivate (1.19.1) unstable; urgency=low

  * Make the warning for su appear above the Password: prompt, instead of
    below it.

 -- Geoffrey Thomas <geofft@mit.edu>  Mon, 18 May 2009 19:45:39 -0400

debathena-reactivate (1.19) unstable; urgency=low

  * Display a warning about the login snapshots, once per session
    for sudo and every time you su.

 -- Geoffrey Thomas <geofft@mit.edu>  Fri, 15 May 2009 02:35:03 -0400

debathena-reactivate (1.18) unstable; urgency=low

  * Send SIGTERM before SIGKILL when cleaning up user processes
    using the bind mounts, to give them a chance to exit cleanly.

 -- Robert Basch <rbasch@mit.edu>  Tue, 24 Mar 2009 17:24:07 -0400

debathena-reactivate (1.17) unstable; urgency=low

  * Even more verbosity! Use fuser -kmv to get more output about the
    running processes.

 -- Evan Broder <broder@mit.edu>  Wed, 18 Mar 2009 19:25:53 -0400

debathena-reactivate (1.16) unstable; urgency=low

  * Don't redirect output to /dev/null when we're trying to log it

 -- Evan Broder <broder@mit.edu>  Mon, 16 Mar 2009 20:18:52 -0400

debathena-reactivate (1.15) unstable; urgency=low

  * Include timing information in the logging.

 -- Evan Broder <broder@mit.edu>  Mon, 16 Mar 2009 08:28:51 -0400

debathena-reactivate (1.14) unstable; urgency=low

  * Add instrumentation to track why reactivations are failing.

 -- Evan Broder <broder@mit.edu>  Thu, 12 Mar 2009 20:06:10 -0400

debathena-reactivate (1.13) unstable; urgency=low

  * Instead of generating policy-rc.d at login time, have a general
    policy-rc.d that works both inside and outside of the chroot. This
    fixes a potential issue if there's a reboot during a login session.

 -- Evan Broder <broder@mit.edu>  Sat, 07 Mar 2009 17:20:01 -0500

debathena-reactivate (1.12) unstable; urgency=low

  * Restart a provided list of daemons inside the chroot at login.

 -- Evan Broder <broder@mit.edu>  Thu, 05 Mar 2009 20:46:25 -0500

debathena-reactivate (1.11) unstable; urgency=low

  * Work around an schroot bug that causes you to get dropped from all of
    your groups.

 -- Evan Broder <broder@mit.edu>  Wed, 04 Mar 2009 22:01:39 -0500

debathena-reactivate (1.10) unstable; urgency=low

  * Ignore failure to umount /media.

 -- Robert Basch <rbasch@mit.edu>  Mon, 23 Feb 2009 12:41:53 -0500

debathena-reactivate (1.9) unstable; urgency=low

  * Added /dev/pts and /dev/shm to the bind-mount list such that matlab and
    mathematica can run.
  * Made /media a shared mount (which also required pre-bind-mounting it
    to itself) so that USB automounts are visible inside the chroot.
  * Reversed the order of unmounting such that it's more likely to succeed.

 -- andrew m. boardman <amb@mit.edu>  Thu, 19 Feb 2009 16:51:30 -0500
	
debathena-reactivate (1.8) unstable; urgency=low

  * Try to kill remaining user processes using any of the login
    chroot's bind mounts at session end.
  * Don't die upon failure to umount any of the bind mounts at
    session end; this should force a reboot after the umount of
    the chroot mountpoint fails.

 -- Robert Basch <rbasch@mit.edu>  Sat, 07 Feb 2009 18:38:45 -0500

debathena-reactivate (1.7) unstable; urgency=low

  * Configure locate's cron job to not walk /login.

 -- Tim Abbott <tabbott@mit.edu>  Sat, 07 Feb 2009 17:23:01 -0500

debathena-reactivate (1.6) unstable; urgency=low

  * Fix typo in schroot configuration (environment_filter should be
    environment-filter).

 -- Tim Abbott <tabbott@mit.edu>  Sat, 31 Jan 2009 15:00:34 -0500

debathena-reactivate (1.5) unstable; urgency=low

  * Remove DEB_AUTO_UPDATE_DEBIAN_CONTROL.
  * Change "Debian-Athena Project" to "Debathena Project".
  * Remove debian/control from version control.
  * Clean up debian/copyright.

 -- Tim Abbott <tabbott@mit.edu>  Sat, 24 Jan 2009 18:59:19 -0500

debathena-reactivate (1.4) unstable; urgency=low

  * Add the user to more groups when they login.

 -- Evan Broder <broder@mit.edu>  Sat, 24 Jan 2009 13:36:03 -0500

debathena-reactivate (1.3) unstable; urgency=low

  * Ensure the sound card volume levels are set to 0 at session end.

 -- Robert Basch <rbasch@mit.edu>  Tue, 16 Dec 2008 19:30:25 -0500

debathena-reactivate (1.2) unstable; urgency=low

  * Rewrite to use login chroots.

 -- Greg Hudson <ghudson@mit.edu>  Fri, 22 Aug 2008 15:30:57 -0400

debathena-reactivate (1.1) unstable; urgency=low

  * Initial release.

 -- Greg Hudson <ghudson@mit.edu>  Thu, 14 Aug 2008 16:27:15 -0400
