- Cherry-picked from upstream commit 032228c9
  "script: may be hangs"

Git-Dch: Full
Closes: #764547

- Cherry-picked from upstream commit e71b0aad
  "build-sys: check for libtoolize rather than libtool"

Git-Dch: Full
Closes: #764387
Thanks: Helmut Grohne for fixing this upstream (and more).

With the previous fix we want to make sure that the initramfs
gets updated to include it on upgrades from previous version.
Lets just always update it on upgrades as if we only do it
on occation we'll probably forget to do it once it's actually
needed.

- Cherry-picked from upstream commit c4604c38
  "flock: zero timeout is valid"

Git-Dch: Full
Closes: #764660

Ben Hutchings wrote:

If you use NTP and tell the kernel to adjust the system time, that
causes the kernel to periodically write the system time to the RTC too,
because the system time will be more accurate.

For hysterical raisins, the first call to settimeofday() that specifies
a 'time zone' (local time offset) implicitly sets whether the RTC is
supposed to hold local time or UTC: if the time value is unspecified and
the time offset is non-zero then the RTC holds local time, otherwise it
holds UTC.  There is no way to change this later!

'hwclock --hctosys' specifies both time value and time offset, and
therefore implicitly configures the RTC to hold UTC (but only when NTP
is used and it is written periodically by the kernel).

So we absolutely have to run 'hwclock --systz' first.  We could *also*
run 'hwclock --hctosys' straight after that.

Closes: #764552
Thanks: Ben Hutchings for debugging this.

This reverts commit 90a49d45.

A suggested fix for this fix has been proposed, lets try that
first.

- Cherry-picked from upstream commit 473c5fb8
  "libmount: fix mnt_is_readonly() #ifdef"
- Fixes Hurd build failure.

Git-Dch: Full
Closes: #764418
Thanks: Pino Toscano for fixing this upstream!

util-linux currently FTCBFS. It fails to compile sys-utils/ipcutils.c,
because that file used C99 features and the cross build does not pass
-std=gnu99 to gcc, because CC is explicitly set from debian/rules. The
configure run did select "$(DEB_HOST_GNU_TYPE)-gcc -std=gnu99", but this
is overridden by passing CC. So when setting CC, we should include the
-std flag. This is what my patch does. Possibly not passing CC works as
well.

Closes: #764392

We only do this to make sure that uuidd isn't running which will
make usermod fail later. People might not even have uuidd
(uuid-runtime) installed, so lets not worry people for nothing...

Closes: #764159

This reverts commit 5dcd0dd1.

Apparently this breaks when people use localtime in their hardware
clock (on top of the new initramfs-tools 0.118 scheme which sets
up time in the initramfs), as reported by Michael Biebl.

We don't need to be so picky about only changing the shell
if it's set to what we've once shipped as an incorrect default.
Just always set the shell when we rename the user makes things
simpler and hopefully more robust.

Using chsh can cause problems with certain pam configurations.

Closes: #764244

"build-sys: use -lutil for BSD only"
debian/patches/build-sys-use-lutil-for-BSD-only.patch

Git-Dch: Full

The ‘uuid-runtime’ package has the primary purpose of installing
user-facing utility programs. It is not a runtime library package, so
the “libs” section is incorrect for this package.

Closes: #764039
Thanks: Ben Finney for the suggestion

In some cases (klibc) the initramfs might not have touch.
Use shell redirection to create the file instead.

Closes: #763823

initramfs-tools [0.117] has started mounting and fscking both / and /usr in
the initramfs, and that means we need to set the system clock properly
as well. (#763777)

Currently /lib/udev/hwclock-set will happily adjust the system time
twice if the RTC uses local time and it is run in both the
initramfs and the main init system.

Closes: #763823

The different small systemd utility libraries has been merged into one.
Update build-dependencies accordingly instead of using the transitional
packages.

Closes: #763808

If there is a running uuidd while executing the usermod command it will
error out. Attempt to stop it and hope that it doesn't get (socket)
activated before we get a chance to run the usermod command.
(Also hope that there are no other running processes under this
user for some funky reason.)
This should atleast catch most cases. Please note that uuidd is
not getting restarted.... Hopefully it'll get socket activated
if really needed. Also libuuid still should work even if uuidd is
not running. The user rename should happen so rarely that this should
not be an issue. Patches to improve this welcome.

LP: #1374648

This was initially introduced in:

commit 4e0adb16
"Use correct configure options on non-linux"

.. as a workaround for an upstream build system bug,
which was fixed in:

commit 6749ce1e
"build-sys: fix switch_root/pivot_root cut-n-paste error in configure.ac"

... which was part of v2.25.1, so it's no longer needed.

libmount-dev doesn't need to depend on uuid-dev since it
doesn't reference it anywhere directly, but ofcourse
this doesn't really make any practical difference as the
libblkid-dev dependency pulls in uuid-dev anyway.

The mount.pc file references blkid.pc and the package
thus needs to depend on that package to work correctly.

LP: #1096581

Lintian complained that it was named BSD (without anything
declaring which version / how many clauses), but we're
not referencing it anywhere and already has one 3-clause
version included so just drop it.

Upstream version 2.25.1

# gpg: Signature made mån  8 sep 2014 01:48:30 CEST using DSA key ID BDE5F1EE
# gpg: Good signature from "Andreas Henriksson <andreas@fatal.se>"
# gpg:                 aka "Andreas Henriksson <ah@debian.org>"

Just in case the user doesn't have uuid-runtime installed
duplicate the postinst part converting the user to
make sure everyone gets converted over.

This should be dropped once it has shipped in a stable release.

Signed-off-by: Karel Zak <kzak@redhat.com>

Signed-off-by: Karel Zak <kzak@redhat.com>

Signed-off-by: Karel Zak <kzak@redhat.com>