Being able to build a "stage1" stripped down version of libraries
shipped by src:util-linux breaks the cyclic build-dependency loop during
bootstrapping. Previously this had to be done manually but can now be
automated thanks to support for Build-Profiles.

Closes: #757147
Thanks: Helmut Grohne

The original issue has vanished, but might not be properly resolved.
This added test should help us automatically detect when we end up with
lots of padded zeroes, if that ever happens again in the future.

Closes: #785075
Thanks: RjY for spotting the original issue.

Everyone upgraded to Jessie should now be free of any
symlinked doc directories. No need to keep removing them.

Seems like util-linux once in a time shipped nfs-related utilities
which was since moved to nfs-common. This is probably pre-squeeze
era.

Thanks lintian.

This is the recommended by upstream, and less buggy, build option.

Closes: #581631

In preparations for switching from slang to ncurses, we need
to do something about the cfdisk-udeb.
Either add a libncursesw5-udeb or remove cfdisk-udeb, as
has already been discussed in:
https://lists.debian.org/debian-boot/2011/02/msg00990.html

- 0002-lib-fileutils-Add-new-dup_fd_cloexec-function.patch
- 0003-lib-sysfs-Use-dup_fd_cloexec-instead-of-direct-call-.patch

This is a temporary workaround until glibc gets fixed (see #785796).

Thanks: Guillem Jover for patches and Steven Chamberlain for testing.
Closes: #787042
Git-Dch: Full

gbp:warning: Old style config section [git-buildpackage] found please rename to [buildpackage]
gbp:warning: Old style config section [gbp-pq] found please rename to [pq]

This isn't technically true, but ensures a robust upgrade also from jessie.

Closes: #786469

When util-linux breaks initscripts and initscripts breaks util-linux,
this happens:

E: This installation run will require temporarily removing the
   essential package initscripts:amd64 due to a
   Conflicts/Pre-Depends loop. This is often bad, but if you really
   want to do it, activate the APT::Force-LoopBreak option.
E: Internal Error, Could not early remove initscripts:amd64 (2)

Replaces should have a counterpart in either a versioned Breaks,
Conflicts or Depends. Bump the existing dependency and drop the Breaks
to avoid the cyclic Breaks to give apt the chance for a smooth upgrade.

- This suits non-linux archs better, since mount is linux-any.

Git-Dch: Full

Fixes previous commit.

Besides not being sure if it's good that fdisk can mean many different
things based or which architecture you're on...
... we should not reinvent the update-alternatives system badly.

The mac-fdisk package already handles not being able to claim
/sbin/fdisk gracefully, so no need for a lockstep upgrade...

The timer was only shipped during a very short time when Debian Jessie
was "testing" and the purging got added before Jessie was released,
so no longer needed. All systems should have been cleaned up by now.

The mistake edit snuck in during commit bb680296
"Add libfdisk1, libfdisk-dev and libfdisk-udeb packages"

This commit reverts the first hunk so libblkid1 get the
old short description back.

Closes: #784822

the new util-linux 2.26.2 FTBFSes on non-Linux architectures, because
of Linux-only symbols in libmount.

Attached there is the easy update to debian/libmount1.symbols.

Closes: #784728

These completions where previously shipped by "bash-completion" pkg.
(See Bug#755986)

Bump the versioned breaks/replaces used by util-linux, since dmesg
is shipped here. Add equivalent to bsdutils, since renice is shipped here.

These where previously shipped by sysvinit-utils (src:sysvinit).

Closes: #784566

This utility was previously shipped by the initscripts (src:sysvinit)
package.

Closes: #753779

- cloud-utils 0.27-1 should now be fixed (see Bug#783826)

Git-Dch: Full

This reverts commit f7edc8ec.

- No longer needed since the test now sets TERM itself.

Git-Dch: Full

Martin Pitt wrote:
"... configure with --enable-libmount-force-mountinfo, to avoid bugs
like
http://lists.freedesktop.org/archives/systemd-devel/2015-February/028180.html
if people still have a broken /etc/mtab?
See
http://lists.freedesktop.org/archives/systemd-devel/2015-February/028719.html
"

Thanks: Martin Pitt for the suggestion and pointers to background info.

libmount gained new symbols since last version...

(This should have been done at the same time libmount1.symbols
where updated.)

Needed to generate correct dependencies for udebs.

- this should be made a versioned breaks once Bug#783826 gets fixed.

Git-Dch: Full

"the library is distributed as a shared library with a stable API and a
standard header file"