[Pkg-libvirt-commits] [Git][libvirt-team/libvirt][debian/sid] 28 commits: New upstream version 6.6.0

Guido Günther gitlab at salsa.debian.org
Sun Aug 23 07:46:43 BST 2020 


Guido Günther pushed to branch debian/sid at Libvirt Packaging Team / libvirt


Commits:
ecdcc72c by Andrea Bolognani at 2020-08-04T00:38:17+02:00
New upstream version 6.6.0
- - - - -
eb5d02e2 by Andrea Bolognani at 2020-08-10T22:47:04+02:00
gbp: Switch debian-branch to debian/experimental

Gbp-Dch: Ignore

- - - - -
751e1467 by Andrea Bolognani at 2020-08-10T22:47:04+02:00
upstream: Add key for Jiří Denemark

The keyring now contains the following keys:

  pub   dsa1024 2000-05-31 [SC]
        C74415BA7C9C7F78F02E1DC34606B8A5DE95BC1F
  uid           [ unknown] Daniel Veillard (Red Hat work email) <veillard at redhat.com>
  uid           [ unknown] Daniel Veillard <Daniel.Veillard at w3.org>
  sub   elg1024 2000-05-31 [E]
  sub   rsa2048 2016-12-01 [S]
  sub   rsa2048 2016-12-01 [E]

  pub   rsa4096 2020-07-20 [SC] [expires: 2025-07-19]
        453B65310595562855471199CA68BE8010084C9C
  uid           [ unknown] Jiří Denemark <jdenemar at redhat.com>
  sub   rsa4096 2020-07-20 [E] [expires: 2025-07-19]

The former was used to sign releases up to 6.5.0, while the
latter will be used from 6.6.0 onwards.

- - - - -
d0f2e4d0 by Andrea Bolognani at 2020-08-10T22:47:18+02:00
Update upstream source from tag 'upstream/6.6.0'

Update to upstream version '6.6.0'
with Debian dir 7ffd2d401cb6e8bf2db110e469b93836f31b48b9

- - - - -
fccbf6fe by Andrea Bolognani at 2020-08-10T22:51:00+02:00
changelog: Create entry for 6.6.0-1

Gbp-Dch: Ignore

- - - - -
803cec46 by Andrea Bolognani at 2020-08-10T22:51:03+02:00
patches: Refresh patches

Gbp-Dch: Ignore

- - - - -
c14f66b4 by Andrea Bolognani at 2020-08-10T22:51:03+02:00
symbols: Bump symbol versions

libvirt 6.6.0 doesn't introduce any new public symbols.

Gbp-Dch: Ignore

- - - - -
ab2a1b4a by Andrea Bolognani at 2020-08-10T22:51:03+02:00
control: Add Build-Depends on libtirpc-dev

libvirt now specifically looks for libtirpc and is not happy
with the XDR headers included in libc6-dev.

Same as upstream commit d7147b379738

- - - - -
8714f7d9 by Andrea Bolognani at 2020-08-10T22:51:03+02:00
control: Drop Build-Depends on libncurses5-dev

It's not actually supposed to be there, since libvirt only uses
ncurses indirectly through readline.

Same as upstream commit 588d2834d799

- - - - -
f2a885c7 by Andrea Bolognani at 2020-08-10T23:17:48+02:00
copyright: Remove obsolete entries

This addresses a bunch of

  unused-file-paragraph-in-dep5-copyright
  wildcard-matches-nothing-in-dep5-copyright

Lintian tags that started showing up as a consequence of
upstream dropping a number of files.

Gbp-Dch: Ignore

- - - - -
a7dbff15 by Andrea Bolognani at 2020-08-10T23:18:14+02:00
copyright: Remove duplicate entries

This addresses a bunch of

  duplicate-globbing-patterns

Lintian tags that went unnoticed until now.

Gbp-Dch: Ignore

- - - - -
ceab4030 by Christian Ehrhardt at 2020-08-14T07:30:45+02:00
d/control, d/rules: feature architecture parity

Enable systemtap, numa and numad on architecture having the build-deps
nowadays.

Ubuntu enabled these features on ppc64el and s390x that work fine there and
have no reason to be non-enabled there based on these architectures.

On review it was pointed out that the build-deps are linux-any these
days, therefore enable it less selective on linux-any.

Signed-off-by: Dimitri John Ledkov <xnox at ubuntu.com>
Signed-off-by: Stefan Bader <stefan.bader at canonical.com>
Signed-off-by: Christian Ehrhardt <christian.ehrhardt at canonical.com>

- - - - -
dd2d1a9e by Christian Ehrhardt at 2020-08-14T07:30:49+02:00
Drop d/p/apparmor-Allow-virt-aa-helper-to-access-the-name-service-.patch

The abstraction brings too much permissions, see discussion when
upstreaming the change:
https://www.redhat.com/archives/libvir-list/2020-August/msg00099.html

(To me) with libvirt >=6.0 (actually even with the old versions, maybe
kernel dependent) this isn't reproducible anymore. We should drop the
rule and once anyone can reproduce it again we can try if we either
want to go with:

  # virt-aa-helper dependent libraries read (and if successful, other
  # files) this but virt-aa-helper itself doesn't require the access,
  # so silence the denial.
  deny /etc/nsswitch.conf r,

Or allowing a very reduced set (Ubuntu had that for a short while):

  # virt-aa-helper dependent libraries might read nss info, but do not
  # need full nameservice access.
  /etc/gai.conf r,
  /etc/hosts r,
  /etc/host.conf r,
  /etc/nsswitch.conf r,

Signed-off-by: Christian Ehrhardt <christian.ehrhardt at canonical.com>

- - - - -
36eaf769 by Andrea Bolognani at 2020-08-18T20:17:55+02:00
patches: Drop debian/Don-t-enable-default-network-on-boot.patch

We're going to achieve the same result without needing to patch
the upstream build system.

Gbp-Dch: Ignore

- - - - -
7beca8e1 by Andrea Bolognani at 2020-08-18T20:19:30+02:00
rules: Don't autostart the default network

Upstream enables this by default, but we don't want it for
Debian. To disable autostart, all we need to do is remove the
symlink before dh_install has a chance to pick it up.

Gbp-Dch: Ignore

- - - - -
cb9de24b by Andrea Bolognani at 2020-08-19T07:05:56+00:00
rules: Enable ZFS support correctly

Starting with

  commit 3ce2f8690933b72f65c42ebaa2ccbe8821e1eecd
  Author: Guido Günther <agx at sigxcpu.org>
  Date:   Sun Aug 14 15:28:28 2016 +0200

    Support zfs on Linux too

ZFS support is supposed to be enabled both for Linux and for
non-Linux, which makes it unnecessary to use the WITH_STORAGE_ZFS
variable.

Moreover, while

  commit ee7cd36509d10535628ca402990675644578671b
  Author: Guido Günther <agx at sigxcpu.org>
  Date:   Tue Aug 9 06:58:44 2016 +0200

    Enable zfs pool support

    Closes: #827245

was intended to initially enable ZFS support on non-Linux only,
in fact the logic that was introduced with that commit was along
the lines of

  if linux:
      zfs = False
  else:
      if linux:
          zfs = True

which means that ZFS only ended up being enabled on non-Linux
because the newly-introduced Build-Depends on zfsutils triggered
libvirt's feature autodetection logic.

Gbp-Dch: Ignore

- - - - -
6193c399 by Andrea Bolognani at 2020-08-19T07:05:57+00:00
rules: Enable sanlock support correctly

When sanlock support was introduced with

  commit e4714b7d895a154ccc74a471d65b120122e35c49
  Author: Guido Günther <agx at sigxcpu.org>
  Date:   Fri Jun 8 16:13:24 2012 +0200

    Add sanlock support

the non-Linux branch was not filled in, and the contents of the
WITH_SANLOCK variable were never passed to configure, which means
that sanlock support only ended up being enabled on Linux because
the newly-introduced Build-Depends on libsanlock-dev triggered
libvirt's feature autodetection logic.

Gbp-Dch: Ignore

- - - - -
6cbe83eb by Andrea Bolognani at 2020-08-19T07:05:57+00:00
rules: Enable OpenVZ support correctly

When OpenVZ was made Linux-only with

  commit 17570fc900f26deaf45c2072dd81c5f0d3fb727b
  Author: Guido Günther <agx at sigxcpu.org>
  Date:   Sat Jun 11 12:31:45 2011 +0200

    Enable OpenVZ on Linux only

the non-Linux branch was not updated accordingly.

Gbp-Dch: Ignore

- - - - -
84a04e69 by Andrea Bolognani at 2020-08-19T07:05:58+00:00
rules: Enable GlusterFS support correctly

When GlusterFS support was enabled with

  commit f0d27edfb47bd8338b55e47f5d3f35021c6e877d
  Author: Guido Günther <agx at sigxcpu.org>
  Date:   Fri Sep 8 11:37:33 2017 +0200

    Enable Gluster support

only the Linux branch was modified accordingly.

Gbp-Dch: Ignore

- - - - -
d0d54bdb by Andrea Bolognani at 2020-08-19T07:05:59+00:00
rules: Don't use --with-systemd-daemon

The option was introduced in

  commit 448ab8c5ba51ebe91d8c5062ef4027b1c7de7045
  Author: Guido Günther <agx at sigxcpu.org>
  Date:   Fri Mar 7 10:54:46 2014 +0100

    Build-depend on libsystemd-daemon-dev on linux

but it was never passed to configure: if one tries that today,
the result is

  configure: WARNING: unrecognized options: --with-systemd-daemon

The libsystemd-daemon dependency itself no longer exists, having
been replaced with a libsystemd dependency in 5409dc10c9e7 and
ultimately being abandoned in favor of libdbus in 903cbca07140.

Gbp-Dch: Ignore

- - - - -
d31eba58 by Christian Ehrhardt at 2020-08-20T07:19:40+02:00
fix device mapper issues

As reported on:
https://www.redhat.com/archives/libvir-list/2020-August/msg00236.html
https://www.redhat.com/archives/libvir-list/2020-August/msg00592.html
- virdevmapper-Don-t-cache-device-mapper-major.patch
- virdevmapper-Ignore-all-errors-when-opening-dev-mapper-co.patch
- virdevmapper-Handle-kernel-without-device-mapper-support.patch

- - - - -
3145e312 by Christian Ehrhardt at 2020-08-21T10:08:52+02:00
tools: fix libvirt-guests.sh text assignments

- - - - -
d273e0d3 by Andrea Bolognani at 2020-08-22T10:10:12+00:00
patches: Drop debian/Use-upstreams-polkit-rule.patch

We're going to achieve the same result without needing to patch
the upstream build system.

Gbp-Dch: Ignore

- - - - -
2a2db7f3 by Andrea Bolognani at 2020-08-22T10:10:13+00:00
rules: Rename upstream polkit rules

While upstream has included this file for a long time, they use
a slightly different name than the one we want to use in Debian
for historical reasons as well as to match the other polkit file
we install. Change its name before dh_install is called to
achieve the desired results.

Gbp-Dch: Ignore

- - - - -
676d8ff8 by Andrea Bolognani at 2020-08-22T17:03:04+02:00
patches: Backport libvirt-guests fix correctly

When the patch was originally backported, it was cherry-picked
to the top of the patch queue, and more specifically on top of

  debian/Debianize-libvirt-guests.patch

which it conflicts with. The result is no longer a straighforward
cherry-pick, which is undesirable, and also merely delays the
need to tweak the Debian-specific patch until we import 6.7.0.

Move the backport to the bottom of the patch queue, before any
Debian-specific changes.

Gbp-Dch: Ignore

- - - - -
1137e334 by Andrea Bolognani at 2020-08-22T17:03:30+02:00
patches: Assign topic to all patches

Divide them into three groups:

  * debian - patches that are Debian-specific;

  * backport - patches that have been cherry-picked from a
    newer libvirt version and will naturally be dropped as soon
    as that version is imported into Debian;

  * forward - patches that are not taken from upstream but are
    not inherently Debian-specific either. These should be sent
    upstream so that we can eventually drop them from the Debian
    package.

- - - - -
51e52ab3 by Andrea Bolognani at 2020-08-22T17:03:30+02:00
patches: Reorder patches

In order to make rebases more convenient and avoid unnecessary
churn, backports should be placed at the bottom of the patch
queue and Debian-specific patches at the very top.

- - - - -
1589f54f by Andrea Bolognani at 2020-08-22T19:24:37+02:00
gbp: Switch debian-branch to debian/sid

Gbp-Dch: Ignore

- - - - -


30 changed files:

- AUTHORS
- − INSTALL
- Makefile.in
- NEWS.rst
- aclocal.m4
- build-aux/config.guess
- build-aux/config.sub
- build-aux/ltmain.sh
- build-aux/syntax-check.mk
- config.h.in
- configure
- configure.ac
- debian/changelog
- debian/control
- debian/copyright
- debian/libvirt0.symbols
- − debian/patches/apparmor-Allow-virt-aa-helper-to-access-the-name-service-.patch
- + debian/patches/backport/tools-fix-libvirt-guests.sh-text-assignments.patch
- + debian/patches/backport/virdevmapper-Don-t-cache-device-mapper-major.patch
- + debian/patches/backport/virdevmapper-Handle-kernel-without-device-mapper-support.patch
- + debian/patches/backport/virdevmapper-Ignore-all-errors-when-opening-dev-mapper-co.patch
- debian/patches/debian/Debianize-libvirt-guests.patch
- − debian/patches/debian/Don-t-enable-default-network-on-boot.patch
- debian/patches/debian/Prefer-sbin-over-usr-sbin.patch
- debian/patches/Set-defaults-for-zfs-tools.patch → debian/patches/debian/Set-defaults-for-zfs-tools.patch
- − debian/patches/debian/Use-upstreams-polkit-rule.patch
- debian/patches/debian/apparmor_profiles_local_include.patch
- debian/patches/Include-etc-pki-qemu-in-apparmor.patch → debian/patches/forward/Include-etc-pki-qemu-in-apparmor.patch
- debian/patches/Pass-GPG_TTY-env-var-to-the-ssh-binary.patch → debian/patches/forward/Pass-GPG_TTY-env-var-to-the-ssh-binary.patch
- debian/patches/Reduce-udevadm-settle-timeout-to-10-seconds.patch → debian/patches/forward/Reduce-udevadm-settle-timeout-to-10-seconds.patch


The diff was not included because it is too large.


View it on GitLab: https://salsa.debian.org/libvirt-team/libvirt/-/compare/793ad15c1dc3c6008ef49d71522e5655eb264730...1589f54fc09f34f69e688ecab9a6873fdedad028

-- 
View it on GitLab: https://salsa.debian.org/libvirt-team/libvirt/-/compare/793ad15c1dc3c6008ef49d71522e5655eb264730...1589f54fc09f34f69e688ecab9a6873fdedad028
You're receiving this email because of your account on salsa.debian.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/pkg-libvirt-commits/attachments/20200823/f1d33e0a/attachment-0001.html>

More information about the Pkg-libvirt-commits mailing list