[systemd-devel] [ANNOUNCE] systemd 19

[Lennart Poettering]

Heya,

Mostly bugfixes:

http://www.freedesktop.org/software/systemd/systemd-19.tar.bz2

Have fun,

Lennart

Alexey Shabalin (1):
      build-sys: update ALTLinux support

Harald Hoyer (2):
      Spelling Corrections
      ratelimit: removed n_printed

Kay Sievers (1):
      systemctl: shutdown(8) - call kexec if kexec kernel is loaded

Lennart Poettering (50):
      tmpfiles: kill double slashes in unix socket names, just in case
      util: beef up logic to find ctty name
      util: retry opening a TTY on EIO
      man: don't do more reloads than necessary in spec files
      readahead: remove misleading error messages
      units: order fsck at .service before basic.target instead of local-fs.target to relax things a little
      main: refuse --test as root
      tmpfiles: never clean up block devices
      unit: don't timeout fsck
      kmsg-syslogd: increase terminate timeout to 5min to generte less debug spew
      kmsg-syslogd: fix build
      mount: make sure network mounts are ordered after network.target
      dbus: don't rely that timer/path units have an initialized unit field
      virtualization: beef virtualization code
      unit: introduce ConditionVirtualization=
      condition: drop pointless double negation
      README: document that we need ipv6
      plymouth: ensure we start not a single getty before plymouth is gone because we never know which ttys plymouth still controls
      util: make string static
      virtualization: check cpuid first, since it is cheaper than DMI
      update TODO
      plymouth: don't wait forever for plymouth in case it is stuck
      main: print warning if /usr is on a seperate partition
      ask-password: supported plymouth cached passwords
      ask-password: fix handling of timeouts when waiting for password
      rescue: terminate plymouth entirely when going into rescue mode
      cryptsetup: try to show the mount point for a crypto disk if we can
      job: start job timer when we begin running the job, not already when we add it to the queue of jobs
      update TODO
      units: synchronize gettys to plymouth even if plymouth is killed by gdm
      general: unify error code we generate on timeout
      dbus: pass along information why a job failed when it failed (dbus api change!)
      logger: leave the logger longer around and allow it do handle more connections
      unit: increase default timeout to 3min
      job: start job timeout on enqueuing not when we start to process a job
      job: when cancelling jobs, make sure to propagate this properly to depending jobs
      job: don't access j->type when j is already freed
      job: also trigger on-failure dependencies when jobs faile due to dependencies, timeout
      systemctl: properly handle job results
      label: udev might be making changes in /dev while we iterate through it
      service: never clean up a service that still has a process in it
      systemctl: shutdown agent explicitly so that it can reset the tty properly
      agent: don't print warnings if a password was removed or timed out
      cryptsetup: add a terse help
      udev: mark hvc devices for exposure in systemd
      udev: expose ttyUSB devices too
      udev: don't ignore non-encrypted block devices with no superblock
      ask-password: make a confused gcc shut up
      update TODO
      build-sys: prepare v19

Mike Kazantsev (1):
      man: fixed typo in SyslogIdentifier=

Zbigniew Jędrzejewski-Szmek (1):
      systemctl: always null-terminate the password

Lennart

-- 
Lennart Poettering - Red Hat, Inc.