Bug#929132: unblock (pre-approval): dbus/1.12.14-1

[Cyril Brulebois]

Niels Thykier  (2019-05-19):
> Ok. I have added an unblock and age-days 8 hint.  Also CC'ing KiBi for
> a d-i ack before adding an unblock-udeb hint.

Apologies for the delay; no objections.


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Bug#929132: unblock (pre-approval): dbus/1.12.14-1

[Samuel Thibault]

Hello,

Simon McVittie, le jeu. 30 mai 2019 19:04:49 +0100, a ecrit:
> The only change that I think could possibly cause regressions
> for d-i's use case (AT-SPI)

d-i doesn't use AT-SPI yet actually.

Samuel

Bug#929132: unblock (pre-approval): dbus/1.12.14-1

[Simon McVittie]

On Sun, 19 May 2019 at 15:35:00 +, Niels Thykier wrote:
> Ok. I have added an unblock and age-days 8 hint.  Also CC'ing KiBi for a
> d-i ack before adding an unblock-udeb hint.

This is now only waiting for a d-i ack, and I haven't had any regression
reports. Any opinions?

Full details: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=929132#5

The only change that I think could possibly cause regressions
for d-i's use case (AT-SPI) is that the tighter validation in
bus/desktop-file.c could break AT-SPI if it ships syntactically invalid
.service files, but /usr/share/dbus-1/services/org.a11y.Bus.service and
/usr/share/dbus-1/accessibility-services/org.a11y.atspi.Registry.service
both seem valid, so that shouldn't be an issue.

smcv

Bug#929132: unblock (pre-approval): dbus/1.12.14-1

[Niels Thykier]

Control: tags -1 d-i confirmed

Simon McVittie:
> Control: tags -1 - moreinfo
> 
> On Fri, 17 May 2019 at 18:59:00 +, Niels Thykier wrote:
>> I am ok with these changes on the premise that you are ready to promptly
>> rollback to the bare minimum changes in case of regressions (regardless
>> of whether we see them before or after the migration).
>>
>> If you agree with this, please go ahead with the upload and remove the
>> moreinfo tag once it is ready to be unblocked.
> 
> Understood. If there are regressions, I can revert individual fixes if
> the cause is obvious, or if necessary do a 1.12.14+really1.12.12 upload.
> 
> I'm busy next weekend and will probably be checking email less frequently,
> so it would be ideal if you're able to set the age-days to cause migration
> on the 27th - that way I'll be more able to respond to any new regression
> reports that come from testing users shortly after migration.
> 
> I've uploaded 1.12.14-1 (no further changes, other than mentioning #928877
> in the changelog) and it has built on all release architectures, except
> 'all' which is still Needs-Build.
> 
> Thanks,
> smcv
> 

Ok. I have added an unblock and age-days 8 hint.  Also CC'ing KiBi for a
d-i ack before adding an unblock-udeb hint.

Thanks,
~Niels

Bug#929132: unblock (pre-approval): dbus/1.12.14-1

[Simon McVittie]

Control: tags -1 - moreinfo

On Fri, 17 May 2019 at 18:59:00 +, Niels Thykier wrote:
> I am ok with these changes on the premise that you are ready to promptly
> rollback to the bare minimum changes in case of regressions (regardless
> of whether we see them before or after the migration).
> 
> If you agree with this, please go ahead with the upload and remove the
> moreinfo tag once it is ready to be unblocked.

Understood. If there are regressions, I can revert individual fixes if
the cause is obvious, or if necessary do a 1.12.14+really1.12.12 upload.

I'm busy next weekend and will probably be checking email less frequently,
so it would be ideal if you're able to set the age-days to cause migration
on the 27th - that way I'll be more able to respond to any new regression
reports that come from testing users shortly after migration.

I've uploaded 1.12.14-1 (no further changes, other than mentioning #928877
in the changelog) and it has built on all release architectures, except
'all' which is still Needs-Build.

Thanks,
smcv

Bug#929132: unblock (pre-approval): dbus/1.12.14-1

[Niels Thykier]

Control: tags -1 moreinfo confirmed

Simon McVittie:
> Package: release.debian.org
> Severity: normal
> User: release.debian@packages.debian.org
> Usertags: unblock
> 
> I would like to follow the dbus upstream 1.12.x stable branch in buster,
> like I did for 1.8.x in jessie and 1.10.x in stretch. I am an upstream
> maintainer and prepared all recent upstream releases.
> 
> If the timing of this release is not suitable to make it into
> buster r0, it might be a good idea to backport the changes related to
> _dbus_rlimit_raise_fd_limit() (Debian bug #928877) as a patch, so that
> *those* can go into r0. As a result I haven't uploaded to unstable yet,
> to keep it possible to upload a backport via unstable if you'd prefer.
> 
> Annotated diffstat for the attached diff, filtered with "filterdiff -p1
> --exclude=Makefile.in --exclude=aclocal.m4 --exclude=build-aux/ltmain.sh
> --exclude=configure --exclude='*/Makefile.in'
> --exclude=aminclude_static.am --exclude=m4/libtool.m4":
> 
> [...]
> 
> Release preparation.
> 
> Thoughts?
> 
> Thanks,
> smcv
> 

Hi Simon,

I am ok with these changes on the premise that you are ready to promptly
rollback to the bare minimum changes in case of regressions (regardless
of whether we see them before or after the migration).

If you agree with this, please go ahead with the upload and remove the
moreinfo tag once it is ready to be unblocked.

Thanks,
~Niels

Bug#929132: unblock (pre-approval): dbus/1.12.14-1

[Simon McVittie]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

I would like to follow the dbus upstream 1.12.x stable branch in buster,
like I did for 1.8.x in jessie and 1.10.x in stretch. I am an upstream
maintainer and prepared all recent upstream releases.

If the timing of this release is not suitable to make it into
buster r0, it might be a good idea to backport the changes related to
_dbus_rlimit_raise_fd_limit() (Debian bug #928877) as a patch, so that
*those* can go into r0. As a result I haven't uploaded to unstable yet,
to keep it possible to upload a backport via unstable if you'd prefer.

Annotated diffstat for the attached diff, filtered with "filterdiff -p1
--exclude=Makefile.in --exclude=aclocal.m4 --exclude=build-aux/ltmain.sh
--exclude=configure --exclude='*/Makefile.in'
--exclude=aminclude_static.am --exclude=m4/libtool.m4":

 bus/bus.c  |8 -
 dbus/dbus-sysdeps-util-unix.c  |   64 +---
 dbus/dbus-sysdeps-util-win.c   |3 
 dbus/dbus-sysdeps.h|3 

Fix #928877, which gives system services the wrong RLIMIT_NOFILE under
recent systemd versions (while logging a warning). See NEWS or #928877
for more details.

 cmake/DBus1Config.pkgconfig.in |   27 +++

Avoid overwriting pkg-config-related global variables when a CMake project
depends on libdbus.

 Makefile.am|2 
 bus/Makefile.am|6 
 dbus/Makefile.am   |2 
 m4/ax_ac_append_to_file.m4 |   32 
 m4/ax_ac_print_to_file.m4  |   32 
 m4/ax_add_am_macro_static.m4   |   28 +++
 m4/ax_am_macros_static.m4  |   38 +
 m4/ax_code_coverage.m4 |  272 +
 m4/ax_file_escapes.m4  |   30 
 test/Makefile.am   |9 -

Adapt to an API change in newer-than-buster autoconf-archive versions.
This is not directly helpful for buster, but we needed it upstream so that
users of more recently-updated autoconf-archive can still build dbus.
Unfortunately there does not seem to be anything we could do that would
be compatible with both older and newer versions, so I've had to bundle
the newer version. The built binaries shouldn't be affected.

 bus/desktop-file.c |   10 -
 dbus/dbus-sysdeps-unix.c   |8 +
 tools/dbus-send.c  |   11 -

Silence compiler warnings that apparently cause FTBFS on some systems
with newer-than-buster toolchains. The change in bus/desktop-file.c
could in principle reject some malformed .service files that would
previously have been wrongly accepted (although I suspect none exist).

 configure.ac   |6 

Avoid FTBFS with some newer-than-buster autoconf-archive versions that
break assumptions we previously made about their use of AX_-prefixed
strings. Again, this is not directly helpful for buster, but we needed it
upstream so that users of more recently-updated autoconf-archive can
still build dbus.

 test/Makefile.am   |9 -

Work around a build failure on MSYS2. The built binaries are unaffected.

 debian/rules   |1 
 debian/tests/gnome-desktop-testing |1 

Disable tests for code paths that are only reached when malloc() returns
NULL, which in practice doesn't happen on Linux anyway. This significantly
speeds up the build, mitigating slowdown from increased test coverage in
recent releases.

 debian/tests/control   |8 -
 debian/tests/system-bus|   54 +++

Add smoke-tests to make sure that both code paths for system bus
activation still work, backported from experimental on the basis that
more coverage is better.

 NEWS   |   42 +
 configure.ac   |6 
 debian/changelog   |   11 +

Release preparation.

Thoughts?

Thanks,
smcv
 Makefile.am|2 
 NEWS   |   42 +
 bus/Makefile.am|6 
 bus/bus.c  |8 -
 bus/desktop-file.c |   10 -
 cmake/DBus1Config.pkgconfig.in |   27 +++
 configure.ac   |6 
 dbus/Makefile.am   |2 
 dbus/dbus-sysdeps-unix.c   |8 +
 dbus/dbus-sysdeps-util-unix.c  |   64 +---
 dbus/dbus-sysdeps-util-win.c   |3 
 dbus/dbus-sysdeps.h|3 
 debian/changelog   |   11 +
 debian/rules   |1 
 debian/tests/control   |8 -
 debian/tests/gnome-desktop-testing |1 
 debian/tests/system-bus|   54 +++
 m4/ax_ac_append_to_file.m4 |   32 
 m4/ax_ac_print_to_file.m4  |   32 
 m4/ax_add_am_macro_static.m4   |   28 +++
 m4/ax_am_macros_static.m4  |   38 +
 m4/ax_code_coverage.m4 |  272