Bug#1063664: gcc-13-cross: file conflicts between gnat-13- and gnat-{9,10}-

[Andreas Beckmann]

Source: gcc-13-cross
Version: 14
Severity: serious

Hi,

there are undeclared file conflicts between gnat-13- and
gnat-{9,10}- in sid. (but not between -9- and -10-).
Maybe it would be sufficient to rebuild the package against gcc-13
13.2.0-13 which had some gnat conflict fixes.

The conflicting files e.g. for gnat-13-aarch64-linux-gnu on amd64 are

usr/aarch64-linux-gnu/lib/libgnarl.so
usr/aarch64-linux-gnu/lib/libgnat.so
usr/bin/aarch64-linux-gnu-gnat
usr/bin/aarch64-linux-gnu-gnatbind
usr/bin/aarch64-linux-gnu-gnatchop
usr/bin/aarch64-linux-gnu-gnatclean
usr/bin/aarch64-linux-gnu-gnatgcc
usr/bin/aarch64-linux-gnu-gnathtml
usr/bin/aarch64-linux-gnu-gnatkr
usr/bin/aarch64-linux-gnu-gnatlink
usr/bin/aarch64-linux-gnu-gnatls
usr/bin/aarch64-linux-gnu-gnatmake
usr/bin/aarch64-linux-gnu-gnatname
usr/bin/aarch64-linux-gnu-gnatprep
usr/share/man/man1/aarch64-linux-gnu-gnat.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatbind.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatchop.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatclean.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnathtml.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatkr.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatlink.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatls.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatmake.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatname.1.gz
usr/share/man/man1/aarch64-linux-gnu-gnatprep.1.gz


Andreas

Bug#1061648: gnat-13-x86-64-linux-gnu: needs to inherit the Conflicts from gnat-13

[Andreas Beckmann]

Package: gnat-13-x86-64-linux-gnu
Version: 13.2.0-12
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package failed to install
because it tries to overwrite other packages files.

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package gnat-13-x86-64-linux-gnu.
  Preparing to unpack .../5-gnat-13-x86-64-linux-gnu_13.2.0-12_amd64.deb ...
  Unpacking gnat-13-x86-64-linux-gnu (13.2.0-12) ...
  dpkg: error processing archive 
/tmp/apt-dpkg-install-N8eYXp/5-gnat-13-x86-64-linux-gnu_13.2.0-12_amd64.deb 
(--unpack):
   trying to overwrite '/usr/bin/x86_64-linux-gnu-gnatgcc', which is also in 
package gnat-12 12.3.0-14
  Errors were encountered while processing:
   /tmp/apt-dpkg-install-N8eYXp/5-gnat-13-x86-64-linux-gnu_13.2.0-12_amd64.deb

There are currently

Package: gnat-13
Version: 13.2.0-12
Depends: gnat-13-x86-64-linux-gnu (= 13.2.0-12), gcc-13-base (= 13.2.0-12), 
gcc-13 (>= 13)
Suggests: gnat-13-doc, ada-reference-manual-2012, gnat-13-sjlj
Conflicts: gnat-10, gnat-11, gnat-12, gnat-4.9, gnat-5, gnat-6, gnat-7, gnat-8, 
gnat-9

but these Conflicts are no longer effective (and can probably be
removed) due to the introduction of gnat-13-

This bug probably affects all gnat-13- packages for all
architectures.


cheers,

Andreas

Bug#1061647: libgccjit-14-doc: missing Breaks+Replaces: libgccjit-13-doc (and maybe more old versions)

[Andreas Beckmann]

Package: libgccjit-14-doc
Version: 14-20240127-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'testing'.
It installed fine in 'testing', then the upgrade to 'sid' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#overwriting-files-and-replacing-packages-replaces

>From the attached log (scroll to the bottom...):

  Preparing to unpack .../libgccjit-14-doc_14-20240127-1_all.deb ...
  Unpacking libgccjit-14-doc (14-20240127-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgccjit-14-doc_14-20240127-1_all.deb (--unpack):
   trying to overwrite '/usr/share/info/libgccjit-figures/factorial.png', which 
is also in package libgccjit-13-doc 13.2.0-10
  Errors were encountered while processing:
   /var/cache/apt/archives/libgccjit-14-doc_14-20240127-1_all.deb

The files in conflict are

  usr/share/info/libgccjit-figures/factorial.png
  usr/share/info/libgccjit-figures/factorial1.png
  usr/share/info/libgccjit-figures/sum-of-squares.png
  usr/share/info/libgccjit-figures/sum-of-squares1.png
  usr/share/info/libgccjit.info.gz


cheers,

Andreas


libgccjit-13-doc=13.2.0-10_libgccjit-14-doc=14-20240127-1.log.gz
Description: application/gzip

Bug#1036641: closed by Debian FTP Masters (reply to Matthias Klose ) (Bug#1036641: fixed in gcc-12 12.3.0-3)

[Andreas Beckmann]

Please try to get this fixed in bookworm, too.

Thanks

Andreas

Bug#999414: postrm: bash -> sh

[Andreas Beckmann]

Followup-For: Bug #999414
Control: found -1 0.189-2
Control: found -1 0.189-3
Control: severity -1 serious
Control: tag -1 sid trixie

the postrm script contains a bashism in purge_debuginfod_ubuntu:
  local -r
and another bug: calling purge_debuginfod_ubuntu even for Debian
resulting in this purge failure:

  (Reading database ... 4780 files and directories currently installed.)
  Purging configuration files for libdebuginfod-common (0.189-3) ...
  /var/lib/dpkg/info/libdebuginfod-common.postrm: 22: local: -r: bad variable 
name
  dpkg: error processing package libdebuginfod-common (--purge):
   installed libdebuginfod-common package post-removal script subprocess 
returned error exit status 2
  Errors were encountered while processing:
   libdebuginfod-common


Andreas


libdebuginfod-common_0.189-3.log.gz
Description: application/gzip

Bug#1036641: gcc-12-base: please bump the Breaks: gnat (<< 12) for smoother upgrades from bullseye

[Andreas Beckmann]

Followup-For: Bug #1036641
Control: tag -1 patch

I've been running bullseye->bookworm upgrade tests with the following
change applied to src:gcc-12, this had the best results w.r.t. upgrading
gnat related packages:

--- gcc-12-12.2.0/debian/rules.conf 2023-01-02 14:06:37.0 +0100
+++ gcc-12-12.2.0/debian/rules.conf 2023-05-24 09:42:40.0 +0200
@@ -615,7 +615,7 @@
   arch_gnutype_map = $(DEB_TARGET_ARCH)=$(TARGET_ALIAS)
 endif # cross compiler

-BASE_BREAKS := gnat (<< 7)
+BASE_BREAKS := gnat (<< 12), gnat-10 (<< 10.3)
 # these would need proper updates, and are only needed for upgrades
 ifneq (,$(filter $(distrelease),stretch jessie trusty xenial bionic))
   BASE_BREAKS :=

The Breaks against gnat-10 is versioned in a way that gnat-10/sid stays
installable in sid.


Andreas

Bug#1036641: gcc-12-base: please bump the Breaks: gnat (<< 12) for smoother upgrades from bullseye

[Andreas Beckmann]

Package: gcc-12-base
Version: 12.2.0-14
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

As usual, it is helpful to bump the Breaks against gnat (which is not
co-installable) for smoother upgrades from bullseye to ensure the
obsolete gnat-10 stack gets removed.

Andreas

Bug#1026245: gm2-12-doc,gm2-13-doc: ships unversioned /usr/share/info/gm2.info.gz

[Andreas Beckmann]

Package: gm2-12-doc,gm2-13-doc
Severity: serious
Control: found -1 12.2.0-10
Control: found -1 13-20221214-1

Hi,

there is a file conflict between gm2-12-doc and gm2-13-doc:

  Preparing to unpack .../gm2-13-doc_13-20221214-1_all.deb ...
  Unpacking gm2-13-doc (13-20221214-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/gm2-13-doc_13-20221214-1_all.deb (--unpack):
   trying to overwrite '/usr/share/info/gm2.info.gz', which is also in package 
gm2-12-doc 12.2.0-10
  Errors were encountered while processing:
   /var/cache/apt/archives/gm2-13-doc_13-20221214-1_all.deb

Please rename that to gm2-XX.info.gz and provide a symlink
in gm2-doc (does not yet exist) like for e.g. gcc-doc.
Do not forget versioned Breaks+Replaces in gm2-doc for taking over
gm2.info.gz from these two packages.


Andreas


gm2-12-doc=12.2.0-10_gm2-13-doc=13-20221214-1.log.gz
Description: application/gzip

Bug#1023324: gm2-12-doc: missing Breaks+Replaces: gm2-12 (<< 12.2.0-8)

[Andreas Beckmann]

Package: gm2-12-doc
Version: 12.2.0-8
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.
This error may also be triggered by having a predecessor package from
'sid 'installed while installing the package from 'experimental'.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#overwriting-files-and-replacing-packages-replaces

>From the attached log (scroll to the bottom...):

  Preparing to unpack .../gm2-12-doc_12.2.0-8_all.deb ...
  Unpacking gm2-12-doc (12.2.0-8) ...
  dpkg: error processing archive 
/var/cache/apt/archives/gm2-12-doc_12.2.0-8_all.deb (--unpack):
   trying to overwrite '/usr/share/doc/gcc-12-base/m2/gm2.html', which is also 
in package gm2-12 12.2.0-7
  dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
  Errors were encountered while processing:
   /var/cache/apt/archives/gm2-12-doc_12.2.0-8_all.deb


cheers,

Andreas


gm2-12=12.2.0-7_gm2-12-doc=12.2.0-8.log.gz
Description: application/gzip

Bug#1016936: dwz: fails while building assaultcube

[Andreas Beckmann]

Package: dwz
Version: 0.14-1
Severity: serious
Control: block 1016935 with -1

Hi,

assaultcube recently started to FTBFS with a dwz error. Since both
src:assaultcube and src:dwz haven't been changed for a long time and
successfully built previously, I suspect this is related to GCC 12
becoming the default compiler recently.

   dh_dwz -O--sourcedirectory=source/src
dwz: debian/assaultcube/usr/lib/games/assaultcube/ac_client: Unknown debugging 
section .debug_addr
dwz: debian/assaultcube/usr/lib/games/assaultcube/ac_server: Unknown debugging 
section .debug_addr
dwz: Too few files for multifile optimization
dh_dwz: error: dwz 
-mdebian/assaultcube/usr/lib/debug/.dwz/x86_64-linux-gnu/assaultcube.debug 
-M/usr/lib/debug/.dwz/x86_64-linux-gnu/assaultcube.debug -- 
debian/assaultcube/usr/lib/games/assaultcube/ac_client 
debian/assaultcube/usr/lib/games/assaultcube/ac_server returned exit code 1
dh_dwz: error: Aborting due to earlier error
make: *** [debian/rules:9: binary] Error 25

(the full buildlog can be found in #1016935)

(my most recent successful assaultcube build in sid was on July 11)

Andreas

Bug#999764: closed by Debian FTP Masters (reply to Matthias Klose ) (Bug#999764: fixed in gcc-12 12-20211117-1)

[Andreas Beckmann]

Control: found -1 12-2027-1

All fixed but gcc-11-locales=11.2.0-10 vs. gcc-12-locales=12-2027-1:

usr/share/locale/be/LC_MESSAGES/cpplib-11.mo
usr/share/locale/ca/LC_MESSAGES/cpplib-11.mo
usr/share/locale/da/LC_MESSAGES/cpplib-11.mo
usr/share/locale/de/LC_MESSAGES/cpplib-11.mo
usr/share/locale/el/LC_MESSAGES/cpplib-11.mo
usr/share/locale/eo/LC_MESSAGES/cpplib-11.mo
usr/share/locale/es/LC_MESSAGES/cpplib-11.mo
usr/share/locale/fi/LC_MESSAGES/cpplib-11.mo
usr/share/locale/fr/LC_MESSAGES/cpplib-11.mo
usr/share/locale/id/LC_MESSAGES/cpplib-11.mo
usr/share/locale/ja/LC_MESSAGES/cpplib-11.mo
usr/share/locale/nl/LC_MESSAGES/cpplib-11.mo
usr/share/locale/pt_BR/LC_MESSAGES/cpplib-11.mo
usr/share/locale/ru/LC_MESSAGES/cpplib-11.mo
usr/share/locale/sr/LC_MESSAGES/cpplib-11.mo
usr/share/locale/sv/LC_MESSAGES/cpplib-11.mo
usr/share/locale/tr/LC_MESSAGES/cpplib-11.mo
usr/share/locale/uk/LC_MESSAGES/cpplib-11.mo
usr/share/locale/vi/LC_MESSAGES/cpplib-11.mo
usr/share/locale/zh_CN/LC_MESSAGES/cpplib-11.mo
usr/share/locale/zh_TW/LC_MESSAGES/cpplib-11.mo


Andreas

Bug#999764: gcc-12: file conflicts with gcc-11 due to missing Breaks+Replaces or renaming

[Andreas Beckmann]

Source: gcc-12
Version: 12-2023-1
Severity: serious

To avoid repeating this bug on every new major upstrema release,
I'd suggest to add
  Conflicts+Replaces+Provides: some-new-virtual-package
to all intentionally not co-installable packages in addition to the
Breaks+Replaces that is needed now.

File conflicts with src:gcc-11:

gcc-12-locales:

usr/share/locale/be/LC_MESSAGES/cpplib-11.mo
usr/share/locale/be/LC_MESSAGES/gcc-11.mo
usr/share/locale/ca/LC_MESSAGES/cpplib-11.mo
usr/share/locale/da/LC_MESSAGES/cpplib-11.mo
usr/share/locale/da/LC_MESSAGES/gcc-11.mo
usr/share/locale/de/LC_MESSAGES/cpplib-11.mo
usr/share/locale/de/LC_MESSAGES/gcc-11.mo
usr/share/locale/el/LC_MESSAGES/cpplib-11.mo
usr/share/locale/el/LC_MESSAGES/gcc-11.mo
usr/share/locale/eo/LC_MESSAGES/cpplib-11.mo
usr/share/locale/es/LC_MESSAGES/cpplib-11.mo
usr/share/locale/es/LC_MESSAGES/gcc-11.mo
usr/share/locale/fi/LC_MESSAGES/cpplib-11.mo
usr/share/locale/fi/LC_MESSAGES/gcc-11.mo
usr/share/locale/fr/LC_MESSAGES/cpplib-11.mo
usr/share/locale/fr/LC_MESSAGES/gcc-11.mo
usr/share/locale/hr/LC_MESSAGES/gcc-11.mo
usr/share/locale/id/LC_MESSAGES/cpplib-11.mo
usr/share/locale/id/LC_MESSAGES/gcc-11.mo
usr/share/locale/ja/LC_MESSAGES/cpplib-11.mo
usr/share/locale/ja/LC_MESSAGES/gcc-11.mo
usr/share/locale/nl/LC_MESSAGES/cpplib-11.mo
usr/share/locale/nl/LC_MESSAGES/gcc-11.mo
usr/share/locale/pt_BR/LC_MESSAGES/cpplib-11.mo
usr/share/locale/ru/LC_MESSAGES/cpplib-11.mo
usr/share/locale/ru/LC_MESSAGES/gcc-11.mo
usr/share/locale/sr/LC_MESSAGES/cpplib-11.mo
usr/share/locale/sr/LC_MESSAGES/gcc-11.mo
usr/share/locale/sv/LC_MESSAGES/cpplib-11.mo
usr/share/locale/sv/LC_MESSAGES/gcc-11.mo
usr/share/locale/tr/LC_MESSAGES/cpplib-11.mo
usr/share/locale/tr/LC_MESSAGES/gcc-11.mo
usr/share/locale/uk/LC_MESSAGES/cpplib-11.mo
usr/share/locale/uk/LC_MESSAGES/gcc-11.mo
usr/share/locale/vi/LC_MESSAGES/cpplib-11.mo
usr/share/locale/vi/LC_MESSAGES/gcc-11.mo
usr/share/locale/zh_CN/LC_MESSAGES/cpplib-11.mo
usr/share/locale/zh_CN/LC_MESSAGES/gcc-11.mo
usr/share/locale/zh_TW/LC_MESSAGES/cpplib-11.mo
usr/share/locale/zh_TW/LC_MESSAGES/gcc-11.mo

gnat-12:

usr/bin/gnat
usr/bin/gnatbind
usr/bin/gnatchop
usr/bin/gnatclean
usr/bin/gnatgcc
usr/bin/gnathtml
usr/bin/gnatkr
usr/bin/gnatlink
usr/bin/gnatls
usr/bin/gnatmake
usr/bin/gnatname
usr/bin/gnatprep
usr/bin/x86_64-linux-gnu-gnat
usr/bin/x86_64-linux-gnu-gnatbind
usr/bin/x86_64-linux-gnu-gnatchop
usr/bin/x86_64-linux-gnu-gnatclean
usr/bin/x86_64-linux-gnu-gnatgcc
usr/bin/x86_64-linux-gnu-gnathtml
usr/bin/x86_64-linux-gnu-gnatkr
usr/bin/x86_64-linux-gnu-gnatlink
usr/bin/x86_64-linux-gnu-gnatls
usr/bin/x86_64-linux-gnu-gnatmake
usr/bin/x86_64-linux-gnu-gnatname
usr/bin/x86_64-linux-gnu-gnatprep
usr/lib/x86_64-linux-gnu/libgnarl.so
usr/lib/x86_64-linux-gnu/libgnat.so
usr/share/man/man1/gnat.1.gz
usr/share/man/man1/gnatbind.1.gz
usr/share/man/man1/gnatchop.1.gz
usr/share/man/man1/gnatclean.1.gz
usr/share/man/man1/gnathtml.1.gz
usr/share/man/man1/gnatkr.1.gz
usr/share/man/man1/gnatlink.1.gz
usr/share/man/man1/gnatls.1.gz
usr/share/man/man1/gnatmake.1.gz
usr/share/man/man1/gnatname.1.gz
usr/share/man/man1/gnatprep.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnat.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatbind.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatchop.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatclean.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnathtml.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatkr.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatlink.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatls.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatmake.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatname.1.gz
usr/share/man/man1/x86_64-linux-gnu-gnatprep.1.gz

libstdc++6-12-dbg:

usr/lib/x86_64-linux-gnu/debug/libstdc++.a
usr/lib/x86_64-linux-gnu/debug/libstdc++.so
usr/lib/x86_64-linux-gnu/debug/libstdc++.so.6
usr/lib/x86_64-linux-gnu/debug/libstdc++fs.a

lib32stdc++6-12-dbg:

usr/lib32/debug/libstdc++.a
usr/lib32/debug/libstdc++.so
usr/lib32/debug/libstdc++.so.6
usr/lib32/debug/libstdc++fs.a

libx32stdc++6-12-dbg:

usr/libx32/debug/libstdc++.a
usr/libx32/debug/libstdc++.so
usr/libx32/debug/libstdc++.so.6
usr/libx32/debug/libstdc++fs.a

libgccjit-12-doc:

usr/share/info/factorial.png
usr/share/info/factorial1.png
usr/share/info/libgccjit.info.gz
usr/share/info/sum-of-squares.png
usr/share/info/sum-of-squares1.png

libstdc++-12-doc:

usr/share/man/man3/C++Intro.3cxx.gz
usr/share/man/man3/SGIextensions.3cxx.gz
usr/share/man/man3/__gnu_cxx.3cxx.gz
usr/share/man/man3/__gnu_cxx::_Caster.3cxx.gz
usr/share/man/man3/__gnu_cxx::_Char_types.3cxx.gz
usr/share/man/man3/__gnu_cxx::_ExtPtr_allocator.3cxx.gz
usr/share/man/man3/__gnu_cxx::_Invalid_type.3cxx.gz
...

libtsan2:

usr/lib/x86_64-linux-gnu/libtsan_preinit.o


Andreas

Bug#998233: gcc-doc-defaults: FTBFS with gcc-11 being the default

[Andreas Beckmann]

Source: gcc-doc-defaults
Version: 5:22
Severity: serious
Tags: ftbfs
Justification: fails to build from source (but built successfully in the past)

   debian/rules override_dh_gencontrol-arch
make[1]: Entering directory '/build/gcc-doc-defaults-22'
# call dh_gencontrol one at a time for *-doc packages
dh_gencontrol -p cpp-doc -- -v5:;  dh_gencontrol -p gcc-doc -- -v5:;  
dh_gencontrol -p gfortran-doc -- -v5:;  dh_gencontrol -p gnat-doc -- 
-v5:10.1.0-2;  dh_gencontrol -p gccgo-doc -- -v5:;
...
   dh_builddeb
dpkg-deb: error: parsing file 'debian/gcc-doc/DEBIAN/control' near line 3 
package 'gcc-doc':
 'Version' field value '5:': nothing after colon in version number
dpkg-deb: error: parsing file 'debian/gfortran-doc/DEBIAN/control' near line 3 
package 'gfortran-doc':
 'Version' field value '5:': nothing after colon in version number
dpkg-deb: error: parsing file 'debian/gccgo-doc/DEBIAN/control' near line 3 
package 'gccgo-doc':
 'Version' field value '5:': nothing after colon in version number
dh_builddeb: error: dpkg-deb --build debian/gfortran-doc .. returned exit code 2
dh_builddeb: error: dpkg-deb --build debian/gcc-doc .. returned exit code 2
dh_builddeb: error: dpkg-deb --build debian/gccgo-doc .. returned exit code 2
dh_builddeb: error: Aborting due to earlier error
make: *** [debian/rules:113: binary] Error 25

There are some versions missing ...


Andreas

Bug#985044: libdebuginfod-common: modifies conffiles: /etc/profile.d/debuginfod.{c,}sh

[Andreas Beckmann]

Package: libdebuginfod-common
Version: 0.183-4
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package modifies conffiles.
This is forbidden by the policy, see
https://www.debian.org/doc/debian-policy/ch-files.html#configuration-files

10.7.3: "[...] The easy way to achieve this behavior is to make the
configuration file a conffile. [...] This implies that the default
version will be part of the package distribution, and must not be
modified by the maintainer scripts during installation (or at any
other time)."

Note that once a package ships a modified version of that conffile,
dpkg will prompt the user for an action how to handle the upgrade of
this modified conffile (that was not modified by the user).

Further in 10.7.3: "[...] must not ask unnecessary questions
(particularly during upgrades) [...]"

If a configuration file is customized by a maintainer script after
having asked some debconf questions, it may not be marked as a
conffile. Instead a template could be installed in /usr/share and used
by the postinst script to fill in the custom values and create (or
update) the configuration file (preserving any user modifications!).
This file must be removed during postrm purge.
ucf(1) may help with these tasks.
See also https://wiki.debian.org/DpkgConffileHandling

In https://lists.debian.org/debian-devel/2012/09/msg00412.html and
followups it has been agreed that these bugs are to be filed with
severity serious.

debsums reports modification of the following files,
from the attached log (scroll to the bottom...):

0m12.6s ERROR: FAIL: debsums reports modifications inside the chroot:
  /etc/profile.d/debuginfod.sh
  /etc/profile.d/debuginfod.csh


Reading your changelog, this sounds like you need to manage these files
with ucf instead of shipping them as conffiles.


cheers,

Andreas


libdebuginfod-common_0.183-4.log.gz
Description: application/gzip

Bug#972936: libgcc-s1 buster -> bullseye upgrade issues

[Andreas Beckmann]

On 15/02/2021 10.07, Matthias Klose wrote:

On 2/14/21 5:58 PM, Simon McVittie wrote:

On Sat, 13 Feb 2021 at 19:52:10 +0100, Paul Gevers wrote:

[The release team are] pretty concerned about a couple of known RC bugs
which need the proper attention of people familiar with upgrade paths
as there's potential to leave upgrading systems unbootable and/or
without a working apt.

...

https://bugs.debian.org/972936 / libgcc-s1



He reports that at least the issue he originally experienced, and perhaps
others, can be fixed by adding transitional binary packages like these:
https://salsa.debian.org/rpavlik/gcc-10-compat



can somebody confirm that as a working solution?


I finally tried to look into this issue and whether I can reproduce it 
in piuparts. But reading the mentioned bug logs seems to indicate that 
all known instances are fixed in current bullseye, or did I miss 
something? So far I didn't stumble upon something related in piuparts...
And with the Breaks: gcc-8-base (<< 8.4) I suggested in #984574 we 
should be even safer against fallout from gcc-8.

That said, having real transitional packages is probably a good idea anyway.


Andreas

Bug#984574: gcc-10-base: please add Breaks: gcc-8-base (<< 8.4)

[Andreas Beckmann]

Package: gcc-10-base
Version: 10.2.1-6
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

while investigating some buster->bullseye upgrade failures in piuparts I
came across a case where apt tries to keep libstdc++-8-dev installed to
satisfy the libstdc++-dev dependency of some boost package.
Adding a versioned
  Breaks: gcc-8-base (<< 8.4)
should ensure the full gcc-8 stack from buster gets removed during the
upgrade (while systems running testing/unstable that still have a newer
version of gcc-8 installed would be able to keep that installed).
We had similar fixes for previous releases as well.

Andreas

Bug#980429: g++-10: spurious c++17 mode segmentation fault in append_to_statement_list_1 (tree-iterator.c:65)

[Andreas Beckmann]

Package: g++-10
Version: 10.2.1-6
Severity: serious

Hi,

while rebuilding src:nheko I noticed an ICE with SEGFAULT.
I've somewhat minimized the testcase with delta (from 3.3 MB to 15 KB).

Build the attached code with

x86_64-linux-gnu-g++-10 -std=c++17 -Wno-return-type -c gcc-10-segfault.C

* it needs the -std=c++17 flag, without it no segfault occurs
* -Wno-return-type suppresses warnings caused by minimizing return statements 
away
* it is a regression from GCC-9, which always accepts the code
* it does not segfault always, but very likely needs less than 5 repetitions
  to result in segfault (out of 1000 test compiles 526 died with a segfault)
* there are two similar backtraces that may happen:

$ x86_64-linux-gnu-g++-10 -std=c++17 -Wno-return-type -c gcc-10-segfault.C
In file included from :
/usr/include/stdc-predef.h: In substitution of 'template std::function::function(_Functor) [with 
_Functor = ;  = ; 
 = ]':
gcc-10-segfault.C:283:12:   required from 'static void 
tweeny::detail::easingresolve::impl(FunctionTuple&, 
tweeny::easing::linearEasing, Fs ...) [with int I = 0; TypeTuple = 
std::array; FunctionTuple = std::tuple >; Fs = {}]'
gcc-10-segfault.C:289:64:   required from 'void 
tweeny::detail::easingfill(EasingCollectionT&, EasingT, 
tweeny::detail::int2type<0>) [with TypeTupleT = std::array; 
EasingCollectionT = std::tuple >; 
EasingT = tweeny::easing::linearEasing]'
gcc-10-segfault.C:297:41:   required from 'void 
tweeny::detail::tweenpoint::via(F) [with F = tweeny::easing::linearEasing; 
Ts = {float}]'
gcc-10-segfault.C:294:5:   required from 
'tweeny::detail::tweenpoint::tweenpoint(Ts ...) [with Ts = {float}]'
gcc-10-segfault.C:141:2:   required from 'static constexpr 
std::_Require::__construct_helper<_Tp, _Args>::type>, 
std::is_constructible<_Tp, _Args ...> > > 
std::allocator_traits<_Alloc>::_S_construct(_Alloc&, _Tp*, _Args&& ...) [with 
_Tp = tweeny::detail::tweenpoint; _Args = {float&}; _Alloc = 
std::allocator >; 
std::_Require::__construct_helper<_Tp, _Args>::type>, 
std::is_constructible<_Tp, _Args ...> > > = void; typename 
std::allocator_traits<_Alloc>::__construct_helper<_Tp, _Args>::type = 
std::integral_constant]'
gcc-10-segfault.C:144:14:   required from 'static decltype 
(std::allocator_traits<_Alloc>::_S_construct(__a, __p, 
(forward<_Args>)(std::allocator_traits<_Alloc>::construct::__args)...)) 
std::allocator_traits<_Alloc>::construct(_Alloc&, _Tp*, _Args&& ...) [with _Tp 
= tweeny::detail::tweenpoint; _Args = {float&}; _Alloc = 
std::allocator >; decltype 
(std::allocator_traits<_Alloc>::_S_construct(__a, __p, 
(forward<_Args>)(std::allocator_traits<_Alloc>::construct::__args)...)) = void]'
gcc-10-segfault.C:224:26:   required from 'std::vector<_Tp, _Alloc>::reference 
std::vector<_Tp, _Alloc>::emplace_back(_Args&& ...) [with _Args = {float&}; _Tp 
= tweeny::detail::tweenpoint; _Alloc = 
std::allocator >; std::vector<_Tp, 
_Alloc>::reference = tweeny::detail::tweenpoint&]'
gcc-10-segfault.C:310:21:   required from 'tweeny::tween& 
tweeny::tween::to(T) [with T = float]'
gcc-10-segfault.C:322:47:   required from here
/usr/include/stdc-predef.h:32:111: internal compiler error: Segmentation fault
   32 |whether the overall intent is to support these features; otherwise,
  | 
  ^
   33 |presume an older compiler has intent to support these features and
  |~
   
0xa65400 crash_signal
../../src/gcc/toplev.c:328
0x7f575b1ddd5f ???
./signal/../sysdeps/unix/sysv/linux/x86_64/sigaction.c:0
0xf09c3b append_to_statement_list_1
../../src/gcc/tree-iterator.c:65
0xf09c3b append_to_statement_list_force(tree_node*, tree_node**)
../../src/gcc/tree-iterator.c:105
0xf09c3b add_stmt(tree_node*)
../../src/gcc/cp/semantics.c:393
0x13a55c2 synthesize_method(tree_node*)
../../src/gcc/cp/method.c:1585
0x64cece maybe_instantiate_noexcept(tree_node*, int)
../../src/gcc/cp/pt.c:25338
0x768644 resolve_overloaded_unification
../../src/gcc/cp/pt.c:22255
0x768644 unify_one_argument
../../src/gcc/cp/pt.c:21801
0x111d442 type_unification_real
../../src/gcc/cp/pt.c:21945
0x111c93f fn_type_unification(tree_node*, tree_node*, tree_node*, tree_node* 
const*, unsigned int, tree_node*, unification_kind_t, int, conversion**, bool, 
bool)
../../src/gcc/cp/pt.c:21325
0x111c02a add_template_candidate_real(z_candidate**, tree_node*, tree_node*, 
tree_node*, tree_node*, vec const*, tree_node*, 
tree_node*, tree_node*, int, tree_node*, unification_kind_t, int) [clone 
.isra.0]
../../src/gcc/cp/call.c:3417
0xf538cd add_template_candidate
../../src/gcc/cp/call.c:3502
0xf538cd add_candidates
../../src/gcc/cp/call.c:5855
0x11289ec build_user_type_conversion_1

Bug#977461: gcc-1[01]-offload-amdgcn,gcc-1[01]-hppa64-linux-gnu: shipping unversioned binaries causes file conflicts

[Andreas Beckmann]

Package: 
gcc-10-offload-amdgcn,gcc-11-offload-amdgcn,gcc-10-hppa64-linux-gnu,gcc-11-hppa64-linux-gnu
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts
Control: found -1 10.2.1-1
Control: found -1 11-20201208-1

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files.

>From the attached log (scroll to the bottom...):

  Preparing to unpack .../gcc-11-offload-amdgcn_11-20201208-1_amd64.deb ...
  Unpacking gcc-11-offload-amdgcn (11-20201208-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/gcc-11-offload-amdgcn_11-20201208-1_amd64.deb 
(--unpack):
   trying to overwrite '/usr/amdgcn-amdhsa/bin/ar', which is also in package 
gcc-10-offload-amdgcn 10.2.1-1
  Errors were encountered while processing:
   /var/cache/apt/archives/gcc-11-offload-amdgcn_11-20201208-1_amd64.deb

These unversioned files are in conflict between the gcc versions:

usr/amdgcn-amdhsa/bin/ar
usr/amdgcn-amdhsa/bin/as
usr/amdgcn-amdhsa/bin/ld
usr/amdgcn-amdhsa/bin/nm
usr/amdgcn-amdhsa/bin/ranlib

usr/bin/hppa64-linux-gnu-lto-dump

Maybe the unversioned binaries should be managed by unversioned packages
from gcc-defaults instead.

cheers,

Andreas


gcc-10-offload-amdgcn=10.2.1-1_gcc-11-offload-amdgcn=11-20201208-1.log.gz
Description: application/gzip

Bug#977408: libstdc++-11-doc: missing Conflicts: libstdc++-10-doc

[Andreas Beckmann]

Followup-For: Bug #977408

Add libgccjit-11-doc which misses a Conflicts: libgccjit-10-doc to the list.

There are file conflicts on .png files with very generic names:

usr/share/info/factorial.png
usr/share/info/factorial1.png
usr/share/info/libgccjit.info.gz
usr/share/info/sum-of-squares.png
usr/share/info/sum-of-squares1.png

which could easily lead to clashes with other packages in the future.


Andreas

Bug#977409: libgm2-17: contains only libraries with SOVERSION 15

[Andreas Beckmann]

Package: libgm2-17
Version: 11-20201208-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships 

-rw-r--r-- root/root 31064 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2cor.so.15.0.0
-rw-r--r-- root/root189592 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2iso.so.15.0.0
-rw-r--r-- root/root 56168 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2log.so.15.0.0
-rw-r--r-- root/root 13968 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2min.so.15.0.0
-rw-r--r-- root/root172720 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2pim.so.15.0.0
lrwxrwxrwx root/root 0 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2cor.so.15 -> libm2cor.so.15.0.0
lrwxrwxrwx root/root 0 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2iso.so.15 -> libm2iso.so.15.0.0
lrwxrwxrwx root/root 0 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2log.so.15 -> libm2log.so.15.0.0
lrwxrwxrwx root/root 0 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2min.so.15 -> libm2min.so.15.0.0
lrwxrwxrwx root/root 0 2020-12-08 14:08 
./usr/lib/x86_64-linux-gnu/libm2pim.so.15 -> libm2pim.so.15.0.0

which is inconsistent with being renamed to libgm2-17.
It also causes file conflicts with libgm2-15 in sid.


Andreas

Bug#977408: libstdc++-11-doc: missing Conflicts: libstdc++-10-doc

[Andreas Beckmann]

Package: libstdc++-11-doc
Version: 11-20201208-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#overwriting-files-and-replacing-packages-replaces

>From the attached log (scroll to the bottom...):

  Preparing to unpack .../libstdc++-11-doc_11-20201208-1_all.deb ...
  Unpacking libstdc++-11-doc (11-20201208-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libstdc++-11-doc_11-20201208-1_all.deb (--unpack):
   trying to overwrite '/usr/share/man/man3/C++Intro.3cxx.gz', which is also in 
package libstdc++-10-doc 10.2.1-1
  dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
  Errors were encountered while processing:
   /var/cache/apt/archives/libstdc++-11-doc_11-20201208-1_all.deb

You should consider adding Conflicts+Replaces+Provides on some (internally
used) virtual package to avoid running into this problem again in the
future.


cheers,

Andreas


libstdc++-10-doc=10.2.1-1_libstdc++-11-doc=11-20201208-1.log.gz
Description: application/gzip

Bug#961159: lib32{,x}gcc-s1: please provide lib{,x}32gcc1 (= 1:${binary:Version})

[Andreas Beckmann]

I have a similar usecase like Bill: I have lots of gcc-*-multilib
packages installed (from ancient to to experimental) for doing
compilation tests for $software with many different compiler versions.

nearly all lib(x)32XYZ packages built by gcc-N (N < 9? 8?) depend on
lib(x)32gcc1

In the end I worked around it with equivs:


### Commented entries have reasonable defaults.
### Uncomment to edit them.
# Source: 
Section: misc
Priority: optional
# Homepage: 
Standards-Version: 3.9.2

Package: lib32gcc1
Architecture: amd64
Version: 1:10.1.0-1
Depends: lib32gcc-s1 (>= 10.1.0-1)
Description: GCC support library (dependency package, 32bit)
 This is a dependency package, and can be safely removed after upgrade.
Homepage: http://gcc.gnu.org/


### Commented entries have reasonable defaults.
### Uncomment to edit them.
# Source: 
Section: misc
Priority: optional
# Homepage: 
Standards-Version: 3.9.2

Package: libx32gcc1
Architecture: amd64
Version: 1:10.1.0-1
Depends: libx32gcc-s1 (>= 10.1.0-1)
Description: GCC support library (x32)
 This is a dependency package, and can be safely removed after upgrade.
Homepage: http://gcc.gnu.org/


Andreas

Bug#966126: gcc-doc-defaults: need to be updated for GCC 10 being the default compiler

[Andreas Beckmann]

Package: gcc-doc-defaults
Version: 5:20
Severity: serious
Tags: ftbfs
Justification: fails to build from source (but built successfully in the past)

Hi,

gcc-doc-defaults needs to be updated after GCC 10 was made the default
compiler. It does currently FTBFS.


Andreas

Bug#961159: lib32{,x}gcc-s1: please provide lib{,x}32gcc1 (= 1:${binary:Version})

[Andreas Beckmann]

On 21/05/2020 14.05, Matthias Klose wrote:
> On 5/20/20 10:32 PM, Andreas Beckmann wrote:
>> With the transitional packages gone in 10.1.0-2, please add versioned
>> (epoched!) provides on the old names (as already done in libgcc-s1)
>> in order to keep old packages installable along the latest gcc.
> 
> I'd like to avoid that.  Please build the nvidia packages using the new 
> package
> names.

This has nothing to do with nvidia. This breaks keeping old compilers
around, which so far worked fine for a long time.

Andreas

Bug#961159: lib32{,x}gcc-s1: please provide lib{,x}32gcc1 (= 1:${binary:Version})

[Andreas Beckmann]

Package: lib32gcc-s1,libx32gcc-s1
Version: 10.1.0-2
Severity: important

With the transitional packages gone in 10.1.0-2, please add versioned
(epoched!) provides on the old names (as already done in libgcc-s1)
in order to keep old packages installable along the latest gcc.

Thanks

Andreas

Bug#954826: libgcc-8-dev: depends on unavailable version of libgcc-s1

[Andreas Beckmann]

On 02/04/2020 11.13, Matthias Klose wrote:
> On 4/2/20 1:30 AM, Andreas Beckmann wrote:
>> On Thu, 26 Mar 2020 21:02:59 +0100 Matthias Klose  wrote:
>>> yes, but we cannot rebuild the package, because it build-depends on gnat-8.
>>> Filed a removal bug for gcc-8 instead.
>>
>> An easy fix would probably be to temporarily add a versioned
>>   Provides: libgcc-s1 (= 1:${binary:Version})
or rather libgcc1-s1 (= ${gcc:EpochVersion})
>> to libgcc-s1 in src:gcc-10 which should make libgcc-8-dev installable
>> again, s.t. gcc-8 can be rebuilt fixed.
> 
> that's a good idea.

I had gcc-10 rebuilding last night with that change (see attached patch)
and just tried it in a pbuilder sid chroot.
* libgcc-8-dev was installable again
* downgrading gcc-10 to the sid version (i.e. switching again to a
version without that provides) worked as well (after removing
libgcc-8-dev first, of course).

Andreas
diff -Nru gcc-10-10-20200324/debian/changelog gcc-10-10-20200324/debian/changelog
--- gcc-10-10-20200324/debian/changelog	2020-03-24 13:38:16.0 +0100
+++ gcc-10-10-20200324/debian/changelog	2020-04-02 01:59:01.0 +0200
@@ -1,3 +1,10 @@
+gcc-10 (10-20200324-1.1) UNRELEASED; urgency=medium
+
+  * Non-maintainer upload.
+  * work around broken gcc-8 dependencies (Closes: #954826)
+
+ -- Andreas Beckmann   Thu, 02 Apr 2020 01:59:01 +0200
+
 gcc-10 (10-20200324-1) unstable; urgency=medium
 
   * GCC snapshot, taken from the trunk (20200324, 906b3eb9df6).
diff -Nru gcc-10-10-20200324/debian/control gcc-10-10-20200324/debian/control
--- gcc-10-10-20200324/debian/control	2020-03-24 13:19:44.0 +0100
+++ gcc-10-10-20200324/debian/control	2020-04-02 01:59:01.0 +0200
@@ -44,7 +44,7 @@
 Section: libs
 Priority: optional
 Depends: gcc-10-base (= ${gcc:Version}), ${shlibs:Depends}, ${misc:Depends}
-Provides: libgcc1 (= ${gcc:EpochVersion}), libgcc-s1-armel [armel], libgcc-s1-armhf [armhf]
+Provides: libgcc1-s1 (= ${gcc:EpochVersion}), libgcc1 (= ${gcc:EpochVersion}), libgcc-s1-armel [armel], libgcc-s1-armhf [armhf]
 Multi-Arch: same
 Pre-Depends: ${misc:Pre-Depends}
 Breaks: ${libgcc:Breaks}
diff -Nru gcc-10-10-20200324/debian/control.m4 gcc-10-10-20200324/debian/control.m4
--- gcc-10-10-20200324/debian/control.m4	2020-03-20 20:41:20.0 +0100
+++ gcc-10-10-20200324/debian/control.m4	2020-04-02 01:59:01.0 +0200
@@ -206,7 +206,7 @@
 Section: ifdef(`TARGET',`devel',`libs')
 Priority: optional
 Depends: BASELDEP, ${shlibs:Depends}, ${misc:Depends}
-Provides: libgcc1`'LS (= ${gcc:EpochVersion}), ifdef(`TARGET',`libgcc-s1-TARGET-dcv1',`libgcc-s1-armel [armel], libgcc-s1-armhf [armhf]')
+Provides: libgcc-s1 (= ${gcc:EpochVersion}), libgcc1`'LS (= ${gcc:EpochVersion}), ifdef(`TARGET',`libgcc-s1-TARGET-dcv1',`libgcc-s1-armel [armel], libgcc-s1-armhf [armhf]')
 ifdef(`MULTIARCH', `Multi-Arch: same
 Pre-Depends: ${misc:Pre-Depends}
 ')`'dnl

Bug#954826: libgcc-8-dev: depends on unavailable version of libgcc-s1

[Andreas Beckmann]

On Thu, 26 Mar 2020 21:02:59 +0100 Matthias Klose  wrote:
> yes, but we cannot rebuild the package, because it build-depends on gnat-8.
> Filed a removal bug for gcc-8 instead.

An easy fix would probably be to temporarily add a versioned
  Provides: libgcc-s1 (= 1:${binary:Version})
to libgcc-s1 in src:gcc-10 which should make libgcc-8-dev installable
again, s.t. gcc-8 can be rebuilt fixed.

Andreas

Bug#946487: libgomp-plugin-nvptx1: please suggest libcuda1 | libcuda1-any

[Andreas Beckmann]

Package: libgomp-plugin-nvptx1
Version: 9.2.1-21
Severity: normal

Hi,

libgomp-plugin-nvptx1 currently comes with 
  Suggests: libcuda1
Please update that to
  Suggests: libcuda1 | libcuda1-any
to include the virtual package provided by all libcuda1 variants in
non-free.
Or if you prefer to have an existing real package as first
alternative on ppc64el use
  Suggests: libcuda1 [amd64] | libnvidia-tesla-cuda1 [amd64 ppc64el] | 
libcuda1-any
Let this change propagate to gcc-10 in experimental, too.

Thanks

Andreas

PS: in sid/amd64
Package libcuda1-any is a virtual package provided by:
  libnvidia-tesla-cuda1 418.87.01-2
  libnvidia-legacy-390xx-cuda1 390.132-1
  libnvidia-legacy-340xx-cuda1 340.107-7
  libnvidia-legacy-304xx-cuda1 304.137-7
  libcuda1 430.64-1

Bug#929552: g++-x86-64-linux-gnu: broken symlink: /usr/share/doc/g++-x86_64-linux-gnu -> cpp-x86_64-linux-gnu

[Andreas Beckmann]

Package: g++-x86-64-linux-gnu
Version: 4:8.3.0-1
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts
Control: found -1 g++-x86-64-linux-gnu/4:8.3.0-1

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m30.7s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/g++-x86_64-linux-gnu -> cpp-x86_64-linux-gnu 
(g++-x86-64-linux-gnu)


Note that the incorrect link uses '_' instead of '-' in source and target.

There were similar bugs in other packages in the past,
e.g. #915678 in gcc-x86-64-linux-gnux32


cheers,

Andreas


g++-x86-64-linux-gnu_4:8.3.0-1.log.gz
Description: application/gzip

Bug#924911: closed by Matthias Klose (Bug#924911: fixed in gcc-8 8.3.0-4)

[Andreas Beckmann]

On 2019-03-26 13:21, Debian Bug Tracking System wrote:
> #924911: gcc-8-base: please add Breaks: gnat-6 (<< 6.4)

That worked as expected :-)
Thanks!

Andreas

Bug#924911: gcc-8-base: please add Breaks: gnat-6 (<< 6.4)

[Andreas Beckmann]

Package: gcc-8-base
Version: 8.3.0-3
Severity: important

Hi,

this should fix some upgrade paths from stretch to buster as encountered
by piuparts. In some cases apt tries to keep gnat-6 installed instead of
installing gnat-8. The existing Breaks: gnat (<< 7) is not sufficient to
"solve" this properly. While I haven't tried to rebuild gcc-8 with this
change and rerun the test with that version available, I'm pretty sure
from my experience in the last years that this will work as intended.
The version (<< 6.4) is chosen to cover upgrades from stretch (which
currently has 6.3.0-18+deb9u1), but not touch setups where people still
have the newer 6.4 or 6.5 (that was previously in sid/testing) installed
(in case such setups are actually possible).

Andreas

PS: I have a machine that still has every g++-x(.y) installed that was
in sid since g++-4.3 (and the corresponding for clang) :-)


libaws3.3.2-dev_None.log.gz
Description: application/gzip

Bug#915678: gcc-x86-64-linux-gnux32: broken symlink: /usr/share/doc/gcc-x86_64-linux-gnux32 -> cpp-x86_64-linux-gnux32 (gcc-x86-64-linux-gnux32)

[Andreas Beckmann]

Followup-For: Bug #915678
Control: found -1 4:8.3.0-1
Control: found -1 g++-x86-64-linux-gnux32/4:8.3.0-1
Control: affects -1 + g++-x86-64-linux-gnux32

Hi,

this invalid symlink is still present in some packages:

0m28.0s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/g++-x86_64-linux-gnux32 -> cpp-x86_64-linux-gnux32 
(g++-x86-64-linux-gnux32)


Andreas

Bug#918903: dwz: fails during pocl i386 build

[Andreas Beckmann]

Package: dwz
Version: 0.12-3
Severity: important

Hi,

I tried updating pocl to debhelper-compat (= 12), but it failed on i386
(while amd64 built fine):

[...]
   dh_install -O--buildsystem=cmake
   dh_installdocs -O--buildsystem=cmake
dh_installdocs: Cannot auto-detect main package for pocl-doc.  If the default 
is wrong, please use --doc-main-package
   dh_sphinxdoc -O--buildsystem=cmake
   dh_installchangelogs -O--buildsystem=cmake
   dh_installinit -O--buildsystem=cmake
   dh_installsystemduser -O--buildsystem=cmake
   dh_lintian -O--buildsystem=cmake
   dh_perl -O--buildsystem=cmake
   dh_link -O--buildsystem=cmake
   dh_strip_nondeterminism -O--buildsystem=cmake
   dh_compress -O--buildsystem=cmake
   dh_fixperms -O--buildsystem=cmake
   debian/rules override_dh_missing
make[1]: Entering directory '/build/pocl-1.2'
dh_missing --list-missing
make[1]: Leaving directory '/build/pocl-1.2'
   dh_dwz -O--buildsystem=cmake
dwz: debian/libpocl2/usr/lib/i386-linux-gnu/libpocl.so.2.2.0: Unknown debugging 
section .debug_addr
dwz: debian/libpocl2/usr/lib/i386-linux-gnu/pocl/libllvmopencl.so: Unknown 
debugging section .debug_addr
dwz: Too few files for multifile optimization
objcopy: 'debian/libpocl2/usr/lib/debug/.dwz/i386-linux-gnu/libpocl2.debug': No 
such file
dh_dwz: objcopy --compress-debug-sections 
debian/libpocl2/usr/lib/debug/.dwz/i386-linux-gnu/libpocl2.debug returned exit 
code 1
make: *** [debian/rules:57: binary] Error 2
dpkg-buildpackage: error: debian/rules binary subprocess returned exit status 2


Andreas

Bug#918445: libgccjit-9-doc: missing Conflicts: libgccjit-8-doc

[Andreas Beckmann]

Package: libgccjit-9-doc
Version: 9-20190103-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

libgccjit-9-doc needs to conflict with libgccjit-8-doc like it does with
the older ones.


Andreas


libgccjit-8-doc=8.2.0-13_libgccjit-9-doc=9-20190103-1.log.gz
Description: application/gzip

Bug#915870: gccgo-doc: broken symlinks: /usr/share/man/man1/gccgo.1.gz, /usr/share/info/gccgo.info.gz

[Andreas Beckmann]

Package: gccgo-doc
Version: 5:7.2.0-2
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m46.5s ERROR: FAIL: Broken symlinks:
  /usr/share/man/man1/gccgo.1.gz -> gccgo-7.1.gz (gccgo-doc)
  /usr/share/info/gccgo.info.gz -> gccgo-7.info.gz (gccgo-doc)


cheers,

Andreas


gccgo-doc_5:7.2.0-2.log.gz
Description: application/gzip

Bug#915678: gcc-x86-64-linux-gnux32: broken symlink: /usr/share/doc/gcc-x86_64-linux-gnux32 -> cpp-x86_64-linux-gnux32 (gcc-x86-64-linux-gnux32)

[Andreas Beckmann]

Package: gcc-x86-64-linux-gnux32
Version: 4:8.2.0-2
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m26.3s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/gcc-x86_64-linux-gnux32 -> cpp-x86_64-linux-gnux32 
(gcc-x86-64-linux-gnux32)
^  ^

The package also ships the correct
  /usr/share/doc/gcc-x86-64-linux-gnux32 -> cpp-x86-64-linux-gnux32
^  ^
(Notice the normalization of '_' to '-').


cheers,

Andreas


gcc-x86-64-linux-gnux32_4:8.2.0-2.log.gz
Description: application/gzip

Bug#913016: cpp-6-x86-64-linux-gnu:amd64: ships /usr/bin/x86_64-linux-gnu-cpp-6, already in cpp-6:amd64

[Andreas Beckmann]

Package: cpp-6-x86-64-linux-gnu
Version: 6.5.0-2cross1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package failed to install
because it tries to overwrite other packages files.

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package cpp-6-x86-64-linux-gnu.
  Preparing to unpack .../cpp-6-x86-64-linux-gnu_6.5.0-2cross1_amd64.deb ...
  Unpacking cpp-6-x86-64-linux-gnu (6.5.0-2cross1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/cpp-6-x86-64-linux-gnu_6.5.0-2cross1_amd64.deb 
(--unpack):
   trying to overwrite '/usr/bin/x86_64-linux-gnu-cpp-6', which is also in 
package cpp-6 6.5.0-2
  dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
  Errors were encountered while processing:
   /var/cache/apt/archives/cpp-6-x86-64-linux-gnu_6.5.0-2cross1_amd64.deb


This seems to affect several other -cross packages as well.


cheers,

Andreas


cpp-6=6.5.0-2_cpp-6-x86-64-linux-gnu=6.5.0-2cross1.log.gz
Description: application/gzip

Bug#911633: gcc-8-base: please add Breaks: gnat (<< 7)

[Andreas Beckmann]

Package: gcc-8-base
Version: 8.2.0-8
Severity: important

Hi doko,

the
  Breaks: gnat (<< 7)
is present in gcc-7-base, but missing in gcc-8-base. Adding this should
fix some upgrade issues from stretch to buster where old gnat-6 built
packages are kept installed instead of being removed. gcc-7-base no
longer has a sufficiently high score to take care of this, but
gcc-8-base has. You way want to bump this to (<< 8) once the gnat-8
transition has completed.


Andreas

Bug#911599: Bug#911090: libapt-pkg5.0: incompatible with apt/stretch

[Andreas Beckmann]

On 2018-10-22 15:13, Matthias Klose wrote:
> The compat symbol was in GCC 7, and then removed in GCC 8. Sure, that doesn't
> help with the distro upgrade.  There is nothing in GCC to change, so if you 
> want
> to fix that, you need a rename of the affected library packages.

That sounds like someone forgot to initiate a proper c++-v8 transition.

How could someone find all libraries in stretch that provide any of the
affected "old-style" mangled symbols?


Andreas

Re: Bug#911090: libapt-pkg5.0: incompatible with apt/stretch

[Andreas Beckmann]

Control: clone -1 -2
Control: reassign -2 g++-8
Control: retitle -2 g++-8: removal of old abi mangling breaks compatibility 
with stretch

On 2018-10-15 18:18, Julian Andres Klode wrote:
> So, I think this affects more than just apt. gcc 7 broke the ABI
> by adding a new mangling
> 
>   URI::operator std::__cxx11::basic_string, 
> std::allocator >()
> 
> and only linking to that. It seems that some new gcc version then
> got rid of the old one
> 
>   URI::operator std::__cxx11::basic_string, 
> std::allocator >[abi:cxx11]()
> 
> and now it's crashing.
> 
> And that's a problem for _every_ library with operator string(), not
> just apt.

Andreas

Bug#909614: cpp-x86-64-linux-gnu: broken symlink: /usr/share/doc/cpp-x86_64-linux-gnu/README.Bugs -> ../gcc-8-x86_64-linux-gnu-base/README.Bugs

[Andreas Beckmann]

Package: cpp-x86-64-linux-gnu
Version: 4:8.2.0-1
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts
Control: affects -1 + gcc-x86-64-linux-gnu

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m35.6s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/cpp-x86_64-linux-gnu/README.Bugs -> 
../gcc-8-x86_64-linux-gnu-base/README.Bugs

Both source and target are missing a s/x86_64/x86-64/
to be valid package names.


cheers,

Andreas


gcc-x86-64-linux-gnu_4:8.2.0-1.log.gz
Description: application/gzip

Bug#905022: gcc-doc-defaults: FTBFS, needs update for gcc-8 being default

[Andreas Beckmann]

Source: gcc-doc-defaults
Version: 5:17.1
Severity: serious
Justification: fails to build from source (but built successfully in the past)

Hi,

gcc-doc-defaults FTBFS in sid. It probably needs some updates now that
gcc-8 has become the default gcc.


[...]
   debian/rules override_dh_gencontrol-arch
make[1]: Entering directory '/build/gcc-doc-defaults-17.1'
# call dh_gencontrol one at a time for *-doc packages
dh_gencontrol -p cpp-doc -- -v5:;  dh_gencontrol -p gcc-doc -- -v5:;  
dh_gencontrol -p gfortran-doc -- -v5:;  dh_gencontrol -p gnat-doc -- 
-v5:7.2.0-2;  dh_gencontrol -p gcj-doc -- -v5:;  dh_gencontrol -p gccgo-doc -- 
-v5:;
dh_gencontrol: Requested unknown package gcj-doc via -p/--package, expected one 
of: gcc-doc cpp-doc gfortran-doc gnat-doc gccgo-doc gcc-doc-base
dh_gencontrol: unknown option or error during option parsing; aborting
dh_gencontrol -a --remaining-packages
dh_gencontrol: No packages to build. Architecture mismatch: amd64, want: all 
alpha amd64 arm64 armel armhf i386 ia64 mips mips64 mips64el mipsel powerpc 
powerpcspe ppc64 ppc64el s390 s390x sparc sparc64 x32 any
make[1]: Leaving directory '/build/gcc-doc-defaults-17.1'
   debian/rules override_dh_gencontrol-indep
make[1]: Entering directory '/build/gcc-doc-defaults-17.1'
dh_gencontrol -pgcc-doc-base -- -v7.2.0-2
dh_gencontrol -i --remaining-packages
dh_gencontrol: No packages to build. Architecture mismatch: amd64, want: all 
alpha amd64 arm64 armel armhf i386 ia64 mips mips64 mips64el mipsel powerpc 
powerpcspe ppc64 ppc64el s390 s390x sparc sparc64 x32 any
make[1]: Leaving directory '/build/gcc-doc-defaults-17.1'
   dh_md5sums
   dh_builddeb
dpkg-deb: error: parsing file 'debian/gcc-doc/DEBIAN/control' near line 3 
package 'gcc-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dh_builddeb: dpkg-deb --build debian/gcc-doc .. returned exit code 2
dpkg-deb: error: parsing file 'debian/gccgo-doc/DEBIAN/control' near line 3 
package 'gccgo-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dpkg-deb: error: parsing file 'debian/gfortran-doc/DEBIAN/control' near line 3 
package 'gfortran-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dh_builddeb: dpkg-deb --build debian/gfortran-doc .. returned exit code 2
dh_builddeb: dpkg-deb --build debian/gccgo-doc .. returned exit code 2
dh_builddeb: Aborting due to earlier error
make: *** [debian/rules:111: binary] Error 2
dpkg-buildpackage: error: fakeroot debian/rules binary subprocess returned exit 
status 2


Andreas


gcc-doc-defaults_5%17.1.log.gz
Description: application/gzip

Bug#904509: gcc-x86-64-linux-gnux32: broken symlink: /usr/share/doc/cpp-x86_64-linux-gnux32/README.Bugs -> ../gcc-8-x86_64-linux-gnux32-base/README.Bugs

[Andreas Beckmann]

Package: gcc-x86-64-linux-gnux32
Version: 4:8.1.0-1
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m33.0s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/cpp-x86_64-linux-gnux32/README.Bugs -> 
../gcc-8-x86_64-linux-gnux32-base/README.Bugs


cheers,

Andreas


gcc-x86-64-linux-gnux32_4:8.1.0-1.log.gz
Description: application/gzip

Bug#901048: gnat-*-x86-64-linux-gnu: missing Conflicts: gnat-5, gnat-6

[Andreas Beckmann]

On 2018-06-08 17:51, Matthias Klose wrote:
> you should not run piuparts against outdated binaries.

These tests take everything available in the Packages files ...

>  I filed #899045
> to remove these packages, but that didn't get yet processed.  Closing
> this issue now.

Interestingly these packages still don't show up in the ftp-master cruft
report ... (otherwise I might have spotted their cruftness).


Andreas

Bug#901048: gnat-*-x86-64-linux-gnu: missing Conflicts: gnat-5, gnat-6

[Andreas Beckmann]

Package: gnat-7-x86-64-linux-gnu,gnat-8-x86-64-linux-gnu
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'stretch'.
It installed fine in 'stretch', then the upgrade to 'buster' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/#overwriting-files-and-replacing-packages-replaces

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package gnat-7-x86-64-linux-gnu.
  Preparing to unpack .../34-gnat-7-x86-64-linux-gnu_7.3.0-18cross1_amd64.deb 
...
  Unpacking gnat-7-x86-64-linux-gnu (7.3.0-18cross1) ...
  dpkg: error processing archive 
/tmp/apt-dpkg-install-PxZEOj/34-gnat-7-x86-64-linux-gnu_7.3.0-18cross1_amd64.deb
 (--unpack):
   trying to overwrite '/usr/bin/x86_64-linux-gnu-gnat', which is also in 
package gnat-6 6.3.0-18+deb9u1
  Errors were encountered while processing:
   
/tmp/apt-dpkg-install-PxZEOj/34-gnat-7-x86-64-linux-gnu_7.3.0-18cross1_amd64.deb

  Selecting previously unselected package gnat-8-x86-64-linux-gnu.
  Preparing to unpack .../31-gnat-8-x86-64-linux-gnu_8.1.0-1cross1_amd64.deb ...
  Unpacking gnat-8-x86-64-linux-gnu (8.1.0-1cross1) ...
  dpkg: error processing archive 
/tmp/apt-dpkg-install-L5HI3j/31-gnat-8-x86-64-linux-gnu_8.1.0-1cross1_amd64.deb 
(--unpack):
   trying to overwrite '/usr/bin/x86_64-linux-gnu-gnat', which is also in 
package gnat-6 6.3.0-18+deb9u1
  Errors were encountered while processing:
   
/tmp/apt-dpkg-install-L5HI3j/31-gnat-8-x86-64-linux-gnu_8.1.0-1cross1_amd64.deb

  Selecting previously unselected package gnat-7-x86-64-linux-gnu.
  Preparing to unpack .../25-gnat-7-x86-64-linux-gnu_7.3.0-18cross1_amd64.deb 
...
  Unpacking gnat-7-x86-64-linux-gnu (7.3.0-18cross1) ...
  dpkg: error processing archive 
/tmp/apt-dpkg-install-Le5E3m/25-gnat-7-x86-64-linux-gnu_7.3.0-18cross1_amd64.deb
 (--unpack):
   trying to overwrite '/usr/bin/x86_64-linux-gnu-gnat', which is also in 
package gnat-5 5.5.0-12
  Errors were encountered while processing:
   
/tmp/apt-dpkg-install-Le5E3m/25-gnat-7-x86-64-linux-gnu_7.3.0-18cross1_amd64.deb


cheers,

Andreas


gnat-6=6.3.0-18+deb9u1_gnat-7-x86-64-linux-gnu=7.3.0-18cross1.log.gz
Description: application/gzip

Bug#898374: gcc-8-cross: file conflicts with {gcc,g++,...}-8 on /usr/bin/-*-8

[Andreas Beckmann]

Source: gcc-8-cross
Version: 12
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package failed to install
because it tries to overwrite other packages files.

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package gcc-8-x86-64-linux-gnu.
  Preparing to unpack .../16-gcc-8-x86-64-linux-gnu_8.1.0-1cross1_amd64.deb ...
  Unpacking gcc-8-x86-64-linux-gnu (8.1.0-1cross1) ...
  dpkg: error processing archive 
/tmp/apt-dpkg-install-GsQJpd/16-gcc-8-x86-64-linux-gnu_8.1.0-1cross1_amd64.deb 
(--unpack):
   trying to overwrite '/usr/bin/x86_64-linux-gnu-gcc-8', which is also in 
package gcc-8 8.1.0-1
  dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
  Errors were encountered while processing:
   
/tmp/apt-dpkg-install-GsQJpd/01-cpp-8-x86-64-linux-gnu_8.1.0-1cross1_amd64.deb
   
/tmp/apt-dpkg-install-GsQJpd/16-gcc-8-x86-64-linux-gnu_8.1.0-1cross1_amd64.deb

Similar problems with cpp-8,g++-8,... as well as for gcc-7-cross and
{gcc,g++,...}-7.


cheers,

Andreas


gcc-8=8.1.0-1_gcc-8-x86-64-linux-gnu=8.1.0-1cross1.log.gz
Description: application/gzip

Bug#881775: gcc-7-base: please add Breaks: gnat (<< 7)

[Andreas Beckmann]

Package: gcc-7-base
Version: 7.2.0-14
Severity: normal

Hi,

please add a 
  Breaks: gnat (<< 7)
to gcc-7-base (and merge this to gcc-8-base ...) to fix some stretch ->
buster upgrade paths involving old ada libraries where apt may prefer to
keep gnat-6 installed instead of switching to gnat-7 ...

(This would probably solve itself in case src:gcc-6 leaves buster.)
(I assume there is a good reason for gnat-6 and gnat-7 not being
co-installable.)


Thanks

Andreas

Bug#881272: gcc-5-: broken symlink /usr/lib/gcc-cross//5/libcc1.so -> ../../../..//lib/libcc1.so.0

[Andreas Beckmann]

On 11/09/2017 04:43 PM, Matthias Klose wrote:
> is this with the gcc-5-cross* packages which are currently built but not 
> uploaded?

I waited a bit, but no new binaries appeared. The tested packages are
versioned 5.5.0-3cross1.

There is another broken link in some gcj-5- packages, that does
remind me of #855640 in gcc-6:

0m40.1s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc-cross/alpha-linux-gnu/5/ecj1 -> ../../ecj1


Andreas

Bug#881272: gcc-5-: broken symlink /usr/lib/gcc-cross//5/libcc1.so -> ../../../..//lib/libcc1.so.0

[Andreas Beckmann]

Source: gcc-5
Version: 5.5.0-3
Severity: important
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m25.1s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc-cross/mips-linux-gnu/5/libcc1.so -> 
../../../../mips-linux-gnu/lib/libcc1.so.0


That seems to be the last broken link in the gcc-5-cross and
gcc-5-cross-ports packages.
(gcc-6-* and gcc-7-* look clean now :-)


cheers,

Andreas


gcc-5-mips-linux-gnu_5.5.0-3cross1.log.gz
Description: application/gzip

Bug#879883: libx32stdc++6-8-dbg: missing Conflicts: libx32stdc++6-7-dbg

[Andreas Beckmann]

Package: libx32stdc++6-8-dbg
Version: 8-20171023-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi doko,

the last round of adding Conflicts seems to have missed this one:

  Selecting previously unselected package libx32stdc++6-8-dbg.
  Preparing to unpack .../15-libx32stdc++6-8-dbg_8-20171023-1_amd64.deb ...
  Unpacking libx32stdc++6-8-dbg (8-20171023-1) ...
  dpkg: error processing archive 
/tmp/apt-dpkg-install-ntuDRD/15-libx32stdc++6-8-dbg_8-20171023-1_amd64.deb 
(--unpack):
   trying to overwrite '/usr/libx32/debug/libstdc++.a', which is also in 
package libx32stdc++6-7-dbg 7.2.0-12
  dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
  Errors were encountered while processing:
   /tmp/apt-dpkg-install-ntuDRD/15-libx32stdc++6-8-dbg_8-20171023-1_amd64.deb


cheers,

Andreas

Bug#872233: lib{, x}32objc-5-dev: broken symlink: /usr/lib/gcc/x86_64-linux-gnu/5/{, x}32/libobjc_gc.so -> ../../../../../lib32/libobjc_gc.so.4

[Andreas Beckmann]

Followup-For: Bug #872233
Control: affects -1 + lib64objc-5-dev

There is also a similar broken symlink in lib64objc-5-dev

0m40.2s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/i686-linux-gnu/5/64/libobjc_gc.so -> 
../../../../../lib64/libobjc_gc.so.4


Andreas

Bug#877441: gcc-8: missing Breaks+Replaces against several gcc-7 packages

[Andreas Beckmann]

Source: gcc-8
Version: 8-20170923-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts
Control: affects -1 + lib32stdc++6-8-dbg libgccjit-8-doc libstdc++-8-doc 
libstdc++6-8-dbg libx32stdc++6-8-dbg

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

Not attaching the logs, just listing the packages with their conflicting
files:

lib32stdc++6-7-dbg=7.2.0-7 lib32stdc++6-8-dbg=8-20170923-1
usr/lib32/debug/libstdc++.a
usr/lib32/debug/libstdc++.so
usr/lib32/debug/libstdc++.so.6
usr/lib32/debug/libstdc++fs.a

libgccjit-7-doc=7.2.0-7 libgccjit-8-doc=8-20170923-1
usr/share/info/libgccjit.info.gz

libstdc++-7-doc=7.2.0-7 libstdc++-8-doc=8-20170923-1
usr/share/man/man3/*.3cxx.gz

libstdc++6-7-dbg=7.2.0-7 libstdc++6-8-dbg=8-20170923-1
usr/lib/x86_64-linux-gnu/debug/libstdc++.a
usr/lib/x86_64-linux-gnu/debug/libstdc++.so
usr/lib/x86_64-linux-gnu/debug/libstdc++.so.6
usr/lib/x86_64-linux-gnu/debug/libstdc++fs.a

libx32stdc++6-7-dbg=7.2.0-7 libx32stdc++6-8-dbg=8-20170923-1
usr/libx32/debug/libstdc++.a
usr/libx32/debug/libstdc++.so
usr/libx32/debug/libstdc++.so.6
usr/libx32/debug/libstdc++fs.a


cheers,

Andreas

Bug#872233: lib{, x}32objc-5-dev: broken symlink: /usr/lib/gcc/x86_64-linux-gnu/5/{, x}32/libobjc_gc.so -> ../../../../../lib32/libobjc_gc.so.4

[Andreas Beckmann]

Followup-For: Bug #872233
Control: found -1 5.4.1-12

Hi,

the broken symlinks are still here:

0m46.5s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/x86_64-linux-gnu/5/32/libobjc_gc.so -> 
../../../../../lib32/libobjc_gc.so.4

0m57.3s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/x86_64-linux-gnu/5/x32/libobjc_gc.so -> 
../../../../../libx32/libobjc_gc.so.4


Andreas

Bug#872301: gcc-6: broken symlink: /usr/lib/gcc-cross//6/libobjc_gc.so -> ../../../..//lib/libobjc_gc.so.4

[Andreas Beckmann]

Source: gcc-6
Version: 6.4.0-3
Severity: important
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m42.6s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc-cross/mips64-linux-gnuabi64/6/libobjc_gc.so -> 
../../../../mips64-linux-gnuabi64/lib/libobjc_gc.so.4


cheers,

Andreas


libobjc-6-dev-mips64-cross_6.4.0-3cross1.log.gz
Description: application/gzip

Bug#872233: lib{, x}32objc-5-dev: broken symlink: /usr/lib/gcc/x86_64-linux-gnu/5/{, x}32/libobjc_gc.so -> ../../../../../lib32/libobjc_gc.so.4

[Andreas Beckmann]

Package: lib32objc-5-dev,libx32objc-5-dev
Version: 5.4.1-11
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m26.7s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/x86_64-linux-gnu/5/32/libobjc_gc.so -> 
../../../../../lib32/libobjc_gc.so.4

0m26.6s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/x86_64-linux-gnu/5/x32/libobjc_gc.so -> 
../../../../../libx32/libobjc_gc.so.4


cheers,

Andreas


lib32objc-5-dev_5.4.1-11.log.gz
Description: application/gzip

Bug#871220: gcc-doc-defaults: FTBFS with GCC-7

[Andreas Beckmann]

Package: gcc-doc-defaults
Version: 5:16
Severity: serious
Tags: sid buster
User: debian-gcc@lists.debian.org
Usertags: ftbfs-gcc-7

gcc-doc-defaults FTBFS since GCC-7 was made the default compiler:

[...]
   debian/rules override_dh_gencontrol-arch
make[1]: Entering directory '/build/gcc-doc-defaults-16'
# call dh_gencontrol one at a time for *-doc packages
dh_gencontrol -p cpp-doc -- -v5:;  dh_gencontrol -p gcc-doc -- -v5:;  
dh_gencontrol -p gfortran-doc -- -v5:;  dh_gencontrol -p gnat-doc -- 
-v5:6.3.0-1;  dh_gencontrol -p gcj-doc -- -v5:6.3.0-1;  dh_gencontrol -p 
gccgo-doc -- -v5:;
dh_gencontrol -a --remaining-packages
dh_gencontrol: No packages to build. Architecture mismatch: amd64, want: all 
alpha amd64 arm64 armel armhf i386 ia64 m68k mips mips64 mips64el mipsel 
powerpc powerpcspe ppc64 ppc64el s390 s390x sparc sparc64 x32 any
make[1]: Leaving directory '/build/gcc-doc-defaults-16'
   debian/rules override_dh_gencontrol-indep
make[1]: Entering directory '/build/gcc-doc-defaults-16'
dh_gencontrol -pgcc-doc-base -- -v6.3.0-1
dh_gencontrol -i --remaining-packages
dh_gencontrol: No packages to build. Architecture mismatch: amd64, want: all 
alpha amd64 arm64 armel armhf i386 ia64 m68k mips mips64 mips64el mipsel 
powerpc powerpcspe ppc64 ppc64el s390 s390x sparc sparc64 x32 any
make[1]: Leaving directory '/build/gcc-doc-defaults-16'
   dh_md5sums
   dh_builddeb
dpkg-deb: error: parsing file 'debian/cpp-doc/DEBIAN/control' near line 3 
package 'cpp-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dh_builddeb: dpkg-deb --build debian/cpp-doc .. returned exit code 2
dpkg-deb: error: parsing file 'debian/gcc-doc/DEBIAN/control' near line 3 
package 'gcc-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dh_builddeb: dpkg-deb --build debian/gcc-doc .. returned exit code 2
dpkg-deb: building package 'gnat-doc' in '../gnat-doc_6.3.0-1_amd64.deb'.
dpkg-deb: error: parsing file 'debian/gfortran-doc/DEBIAN/control' near line 3 
package 'gfortran-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dh_builddeb: dpkg-deb --build debian/gfortran-doc .. returned exit code 2
dpkg-deb: building package 'gcc-doc-base' in '../gcc-doc-base_6.3.0-1_all.deb'.
dpkg-deb: building package 'gcj-doc' in '../gcj-doc_6.3.0-1_amd64.deb'.
dpkg-deb: error: parsing file 'debian/gccgo-doc/DEBIAN/control' near line 3 
package 'gccgo-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dh_builddeb: dpkg-deb --build debian/gccgo-doc .. returned exit code 2
dh_builddeb: Aborting due to earlier error
debian/rules:105: recipe for target 'binary' failed
make: *** [binary] Error 2


The 'dh_gencontrol -- -v5:' looks like some required variable is now unset


Andreas


gcc-doc-defaults_5%16.log.gz
Description: application/gzip

Bug#870895: gcc-7: libobjc-7-dev--cross: broken symlink: /usr/lib/gcc-cross//7/libobjc_gc.so -> ../../../..//lib/libobjc_gc.so.4

[Andreas Beckmann]

Source: gcc-7
Version: 7.1.0-11
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

new version, new broken symlinks ;-)

During a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m22.6s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc-cross/mips64-linux-gnuabi64/7/libobjc_gc.so -> 
../../../../mips64-linux-gnuabi64/lib/libobjc_gc.so.4

That seems to be the only issue in gcc-7 so far :-)


cheers,

Andreas


libobjc-7-dev-mips64-cross_7.1.0-11cross1.log.gz
Description: application/gzip

Bug#866150: lib64objc-6-dev: broken symlink: /usr/lib/gcc/i686-linux-gnu/6/64/libobjc_gc.so -> ../../../../../lib64/libobjc_gc.so.4

[Andreas Beckmann]

Followup-For: Bug #866150
Control: affects -1 + libx32objc-6-dev lib64objc-7-dev libx32objc-7-dev

Similar problems in libx32objc-6-dev, lib64objc-7-dev, libx32objc-7-dev:

0m26.6s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/i686-linux-gnu/6/x32/libobjc_gc.so -> 
../../../../../libx32/libobjc_gc.so.4

1m15.7s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/i686-linux-gnu/7/64/libobjc_gc.so -> 
../../../../../lib64/libobjc_gc.so.4

1m13.2s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/i686-linux-gnu/7/x32/libobjc_gc.so -> 
../../../../../libx32/libobjc_gc.so.4


There seems to be only one regular libobjc_gc.so.4:

libobjc4: /usr/lib/i386-linux-gnu/libobjc_gc.so.4
libobjc4: /usr/lib/i386-linux-gnu/libobjc_gc.so.4.0.0

(and of course many more from the *-cross packages in /usr//...)


Andreas

Bug#866150: lib64objc-6-dev: broken symlink: /usr/lib/gcc/i686-linux-gnu/6/64/libobjc_gc.so -> ../../../../../lib64/libobjc_gc.so.4

[Andreas Beckmann]

Package: lib64objc-6-dev
Version: 6.3.0-19
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m27.1s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc/i686-linux-gnu/6/64/libobjc_gc.so -> 
../../../../../lib64/libobjc_gc.so.4


cheers,

Andreas


lib64objc-6-dev_6.3.0-19.log.gz
Description: application/gzip

Bug#863845: libstdc++6: adding more Breaks to smoothen upgrades from jessie to stretch

[Andreas Beckmann]

On 2017-05-31 22:55, Andreas Beckmann wrote:
>   libktoblzcheck1c2a  (replaced by libktoblzcheck1v5)
>   libterralib (replaced by libterralib3)
>   libmagickcore-6.q16-2   (replaced by 
> libmagickcore-6.q16-3)
>   liblhapdf0  (replaced by liblhapdf0v5)
>   libspice-client-glib-2.0-8 (<< 0.33)

> I'm now going to try whether that works as expected ...

After some iterations of testing I can confirm that adding these 4
Breaks improves some upgrade paths:

libktoblzcheck1c2a,
libaqbanking34-plugins,
liblhapdf0,
libpythia8,

There is a different problem with libterralib that needs to be fixed in
jessie, adding a Breaks in libstdc++ won't help.
libmagickcore-6.q16-2 and libspice-client-glib-2.0-8 need more
investigation.


Andreas

Bug#863845: libstdc++6: adding more Breaks to smoothen upgrades from jessie to stretch

[Andreas Beckmann]

Package: libstdc++6
Version: 6.3.0-18
Severity: important

Hi,

during my piuparts upgrade tests I identified a few upgrade paths that
could likely be smoothened by adding a few Breaks to libstdc++6:

  libktoblzcheck1c2a(replaced by libktoblzcheck1v5)
  libterralib   (replaced by libterralib3)
  libmagickcore-6.q16-2 (replaced by 
libmagickcore-6.q16-3)
  liblhapdf0(replaced by liblhapdf0v5)
  libspice-client-glib-2.0-8 (<< 0.33)

In these upgrade paths the problematic packages are not removed
or upgraded, but the jessie versions of these (and some other) packages
are kept installed.

I'm now going to try whether that works as expected ...


Andreas

Bug#863820: gcc-6-base: adding Breaks: tzdata-java to smoothen the openjdk 7 -> 8 upgrade path from jessie to stretch

[Andreas Beckmann]

On 2017-05-31 16:01, Andreas Beckmann wrote:
> I'll now build a gcc-6 with that Breaks added ...

... and can confirm that it fixes the upgrade path of
libreoffice-subsequentcheckbase (where I previously observed a kept back
tzdata (and openjdk-7) from jessie).

Andreas

Bug#863820: gcc-6-base: adding Breaks: tzdata-java to smoothen the openjdk 7 -> 8 upgrade path from jessie to stretch

[Andreas Beckmann]

Package: gcc-6-base
Version: 6.3.0-18
Severity: important

Hi Doko,

would it be possible to use gcc-6-base as the bigger hammer to smoothen
the openjdk 7 -> 8 upgrade from jessie to stretch?
The Breaks: tzdata-java added in openjdk-8-jre-headless (#857992)
improved the situation a lot, but I still see some corner cases where
switching to openjdk-8 is not considered by apt, and instead some jessie
packages are kept installed (the openjdk-7 stack, and especially tzdata 
is kept at the jessie version (that still has tzdata-java)).

A versioned Breaks against some openjdk-7 package would not work due to
new upstream releases being uploaded to (old)stable frequently.
An unversioned Breaks against some openjdk-7 package would render the
packages from experimental uninstallable in sid.

gcc-6-base should have a high enough score in apt to stop any attempts
to keep tzdata-java or openjdk-7 (from jessie) installed.

I'll now build a gcc-6 with that Breaks added ...


Andreas

Bug#861060: libstdc++6: please add Breaks: libsigc++-2.0-0c2a (<= 2.4.1-1+b1)

[Andreas Beckmann]

Package: libstdc++6
Version: 6.3.0-14
Severity: important
User: debian...@lists.debian.org
Usertags: piuparts
Control: affects -1 + gir1.2-ganv-1.0

Hi,

I just found an upgrade path on i386 that doesn't work as expected:

  Starting 2 pkgProblemResolver with broken count: 2
  Investigating (0) libsigc++-2.0-0v5 [ i386 ] < none -> 2.10.0-1 > ( libs )
  Broken libsigc++-2.0-0v5:i386 Conflicts on libsigc++-2.0-0c2a [ i386 ] < 
2.4.0-1 > ( libs )
Considering libsigc++-2.0-0c2a:i386 4 as a solution to 
libsigc++-2.0-0v5:i386 0
Holding Back libsigc++-2.0-0v5:i386 rather than change 
libsigc++-2.0-0c2a:i386
  Investigating (0) libganv-1-1v5 [ i386 ] < none -> 1.4.2~dfsg0-2 > ( libs )
  Broken libganv-1-1v5:i386 Depends on libsigc++-2.0-0v5 [ i386 ] < none -> 
2.10.0-1 > ( libs ) (>= 2.8.0)
Considering libsigc++-2.0-0v5:i386 0 as a solution to libganv-1-1v5:i386 0
Holding Back libganv-1-1v5:i386 rather than change libsigc++-2.0-0v5:i386
  Investigating (1) gir1.2-ganv-1.0 [ i386 ] < 1.4.2~dfsg0-1 -> 1.4.2~dfsg0-2 > 
( introspection )
  Broken gir1.2-ganv-1.0:i386 Depends on libganv-1-1v5 [ i386 ] < none -> 
1.4.2~dfsg0-2 > ( libs )
Considering libganv-1-1v5:i386 0 as a solution to gir1.2-ganv-1.0:i386 0
Holding Back gir1.2-ganv-1.0:i386 rather than change libganv-1-1v5:i386
   Try to Re-Instate (2) gir1.2-ganv-1.0:i386
  Done

So let us use the bigger hammer and add another Breaks to libstdc++6.
2.4.1-1+b1 was the last version before libsigc++-2.0-0v5 according to
snapshot.d.o.


Thanks

Andreas


ganv.log.gz
Description: application/gzip

Bug#858112: gcc-defaults: broken symlinks: gcc-: /usr/share/doc/cpp-/README.Bugs -> ../gcc-6/README.Bugs

[Andreas Beckmann]

Control: severity -1 important

On Wed, 5 Apr 2017 21:21:06 +0200 Andreas Beckmann <a...@debian.org> wrote:
> Control: severity -1 serious

> and resulting in packages without copyright file after upgrade
> e.g. gdc-mips64-linux-gnuabi64 4:6.3.0-1 -> 4:6.3.0-2

This is tracked separately as #860117


Andreas

Bug#858112: gcc-defaults: broken symlinks: gcc-: /usr/share/doc/cpp-/README.Bugs -> ../gcc-6/README.Bugs

[Andreas Beckmann]

Followup-For: Bug #858112
Control: reopen -1
Control: found -1 1.168

Hi,

the broken README.Bugs symlink is still there ...

  /usr/share/doc/cpp-mips-linux-gnu/README.Bugs -> ../gcc-6-base/README.Bugs
  /usr/share/doc/cpp-s390x-linux-gnu/README.Bugs -> ../gcc-6-base/README.Bugs
  /usr/share/doc/cpp-alpha-linux-gnu/README.Bugs -> ../gcc-6-base/README.Bugs
...


Andreas

Bug#860120: gcc-{hppa, sh4}-linux-gnu: broken symlink: /usr/share/doc/cpp-/README.Bugs -> ../gcc-6-base/README.Bugs

[Andreas Beckmann]

Package: gcc-hppa-linux-gnu,gcc-sh4-linux-gnu
Version: 4:6.3.0-4
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m39.9s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/cpp-sh4-linux-gnu/README.Bugs -> ../gcc-6-base/README.Bugs

It's a bit confusing since gcc-* ships a symlink in /usr/share/doc/cpp-* ...
Also this affects only these two packages built from gcc-defaults-ports,
the other ones seem to be OK.


cheers,

Andreas


gcc-hppa-linux-gnu_4:6.3.0-4.log.gz
Description: application/gzip

Bug#860117: gcc-defaults: gdc: unhandled directory to symlink conversion: /usr/share/doc/PACKAGE

[Andreas Beckmann]

Source: gcc-defaults
Version: 1.168
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

a test with piuparts revealed that your package misses the copyright
file after an upgrade, which is a violation of Policy 12.5:
https://www.debian.org/doc/debian-policy/ch-docs.html#s-copyrightfile

After the upgrade /usr/share/doc/$PACKAGE/ is just an empty directory.

This was observed on the following upgrade paths:

  stretch -> sid

>From the attached log (scroll to the bottom...):

1m1.8s DEBUG: Starting command: ['adequate', '--root', 
'/tmp/piupartss/tmp1lLvlN', 'gdc-mips-linux-gnu']
1m2.0s DUMP: 
  gdc-mips-linux-gnu: missing-copyright-file 
/usr/share/doc/gdc-mips-linux-gnu/copyright
1m2.0s DEBUG: Command ok: ['adequate', '--root', '/tmp/piupartss/tmp1lLvlN', 
'gdc-mips-linux-gnu']

  MISSING COPYRIGHT FILE: /usr/share/doc/gdc-mips-linux-gnu/copyright
  # ls -lad /usr/share/doc/gdc-mips-linux-gnu
  drwxr-xr-x 2 root root 40 Apr  9 05:26 /usr/share/doc/gdc-mips-linux-gnu
  # ls -la /usr/share/doc/gdc-mips-linux-gnu/
  total 0
  drwxr-xr-x  2 root root   40 Apr  9 05:26 .
  drwxr-xr-x 92 root root 2460 Apr  9 05:26 ..


Additional info may be available here:
https://wiki.debian.org/MissingCopyrightFile

Note that dpkg intentionally does not replace directories with symlinks
and vice versa, you need the maintainer scripts to do this.
See in particular the end of point 4 in
https://www.debian.org/doc/debian-policy/ch-maintainerscripts.html#s-unpackphase

It is recommended to use the dpkg-maintscript-helper commands
'dir_to_symlink' and 'symlink_to_dir' (available since dpkg 1.17.14)
to perform the conversion, ideally using d/$PACKAGE.maintscript.
Do not forget to add 'Pre-Depends: ${misc:Pre-Depends}' in d/control.
See dpkg-maintscript-helper(1) and dh_installdeb(1) for details.


The previous versions (up to 6.3.0-2) shipped regular files
(and the messed up symlink):

drwxr-xr-x root/root 0 2017-03-16 15:08 
./usr/share/doc/gdc-mips-linux-gnu/
-rw-r--r-- root/root  6957 2016-08-03 13:58 
./usr/share/doc/gdc-mips-linux-gnu/README.Debian
-rw-r--r-- root/root 12987 2017-03-16 15:08 
./usr/share/doc/gdc-mips-linux-gnu/changelog.Debian.gz
-rw-r--r-- root/root  2207 2013-06-12 23:03 
./usr/share/doc/gdc-mips-linux-gnu/copyright
lrwxrwxrwx root/root 0 2017-03-16 15:08 
./usr/share/doc/gdc-mips-linux-gnu/cpp-mips-linux-gnu -> cpp-mips-linux-gnu


cheers,

Andreas


gdc-mips-linux-gnu_4:6.3.0-4.log.gz
Description: application/gzip

Bug#859914: libstdc++6: please add Breaks: libopencv-core2.4

[Andreas Beckmann]

Package: libstdc++6
Version: 6.3.0-11
Severity: important
User: debian...@lists.debian.org
Usertags: piuparts
Control: affects -1 + libsikuli-script-jni

Hi doko,

please add another Breaks to libstdc++: against libopencv-core2.4

In piuparts I found that the jessie->stretch upgrade of
libsikuli-script-jni only works partially, apt problem resolver output:

  Investigating (0) libopencv-core2.4v5 [ amd64 ] < none -> 2.4.9.1+dfsg1-1 > ( 
libs )
  Broken libopencv-core2.4v5:amd64 Conflicts on libopencv-core2.4 [ amd64 ] < 
2.4.9.1+dfsg-1+deb8u1 > ( libs )
Considering libopencv-core2.4:amd64 51 as a solution to 
libopencv-core2.4v5:amd64 3
Holding Back libopencv-core2.4v5:amd64 rather than change 
libopencv-core2.4:amd64
  Investigating (0) libopencv-imgproc2.4v5 [ amd64 ] < none -> 2.4.9.1+dfsg1-1 
> ( libs )
  Broken libopencv-imgproc2.4v5:amd64 Depends on libopencv-core2.4v5 [ amd64 ] 
< none -> 2.4.9.1+dfsg1-1 > ( libs ) (= 2.4.9.1+dfsg1-1)
Considering libopencv-core2.4v5:amd64 3 as a solution to 
libopencv-imgproc2.4v5:amd64 1
Holding Back libopencv-imgproc2.4v5:amd64 rather than change 
libopencv-core2.4v5:amd64
  Investigating (0) libopencv-highgui2.4-deb0 [ amd64 ] < none -> 
2.4.9.1+dfsg1-1 > ( libs )
  Broken libopencv-highgui2.4-deb0:amd64 Depends on libopencv-core2.4v5 [ amd64 
] < none -> 2.4.9.1+dfsg1-1 > ( libs )
Considering libopencv-core2.4v5:amd64 3 as a solution to 
libopencv-highgui2.4-deb0:amd64 0
Holding Back libopencv-highgui2.4-deb0:amd64 rather than change 
libopencv-core2.4v5:amd64
  Investigating (0) libsikulixapi-jni [ amd64 ] < none -> 1.1.0-2 > ( java )
  Broken libsikulixapi-jni:amd64 Depends on libopencv-core2.4v5 [ amd64 ] < 
none -> 2.4.9.1+dfsg1-1 > ( libs )
Considering libopencv-core2.4v5:amd64 3 as a solution to 
libsikulixapi-jni:amd64 -1
Holding Back libsikulixapi-jni:amd64 rather than change 
libopencv-core2.4v5:amd64
  Investigating (0) libsikuli-script-jni [ amd64 ] < 
1.0~x~rc3.tesseract3-dfsg1-8 -> 1.1.0-2 > ( java )
  Broken libsikuli-script-jni:amd64 Depends on libsikulixapi-jni [ amd64 ] < 
none -> 1.1.0-2 > ( java )
Considering libsikulixapi-jni:amd64 -1 as a solution to 
libsikuli-script-jni:amd64 -1
Holding Back libsikuli-script-jni:amd64 rather than change 
libsikulixapi-jni:amd64
   Try to Re-Instate (1) libsikuli-script-jni:amd64
  Done

That very much looks like we need the big hammer to get rid of
libopencv-core2.4 in favor of libopencv-core2.4v5 ...

Thanks


Andreas

PS: the Breaks against libopenmpi1.6 worked nicely, I'm now
investigating the 10 remaining upgrade paths :-)

Bug#858112: gcc-defaults: broken symlinks: gcc-: /usr/share/doc/cpp-/README.Bugs -> ../gcc-6/README.Bugs

[Andreas Beckmann]

Control: severity -1 serious

On Sat, 18 Mar 2017 15:07:06 +0100 Andreas Beckmann <a...@debian.org> wrote:
> and another one in gdc-
> 
>   /usr/share/doc/gdc-/cpp- -> 
> cpp-

and resulting in packages without copyright file after upgrade
e.g. gdc-mips64-linux-gnuabi64 4:6.3.0-1 -> 4:6.3.0-2

  MISSING COPYRIGHT FILE: /usr/share/doc/gdc-mips64-linux-gnuabi64/copyright
  # ls -lad /usr/share/doc/gdc-mips64-linux-gnuabi64
  drwxr-xr-x 2 root root 40 Mar 20 17:35 
/usr/share/doc/gdc-mips64-linux-gnuabi64
  # ls -la /usr/share/doc/gdc-mips64-linux-gnuabi64/
  total 0
  drwxr-xr-x  2 root root   40 Mar 20 17:35 .
  drwxr-xr-x 94 root root 2500 Mar 20 17:35 ..

Bug#858112: gcc-defaults: broken symlinks: gcc-: /usr/share/doc/cpp-/README.Bugs -> ../gcc-6/README.Bugs

[Andreas Beckmann]

Followup-For: Bug #858112

and another one in gdc-

  /usr/share/doc/gdc-/cpp- -> cpp-


Andreas

Bug#858112: gcc-defaults: broken symlinks: gcc-: /usr/share/doc/cpp-/README.Bugs -> ../gcc-6/README.Bugs

[Andreas Beckmann]

Source: gcc-defaults
Version: 1.165
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m35.0s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/cpp-aarch64-linux-gnu/README.Bugs -> ../gcc-6/README.Bugs

That happens in gcc-defaults-ports, too:

0m34.2s ERROR: FAIL: Broken symlinks:
  /usr/share/doc/cpp-alpha-linux-gnu/README.Bugs -> ../gcc-6/README.Bugs

A possible candidate would be
  /usr/share/doc/gcc-6-base/README.Bugs
or rather
  /usr/share/doc/gcc-6--base/README.Bugs

Greetings from the MG BSP :-)


cheers,

Andreas

Bug#857384: gnat-6: broken symlinks: /usr/share/man/man1/{, x86_64-linux-gnu-}gnatgcc.1.gz -> x86_64-linux-gnu-gcc-6.1.gz

[Andreas Beckmann]

Followup-For: Bug #857384

Hi,

there is a similar problem in the -cross packages:

0m28.6s ERROR: FAIL: Broken symlinks:
  /usr/share/man/man1/alpha-linux-gnu-gnatgcc.1.gz -> alpha-linux-gnu-gcc-6.1.gz

as well as in gnat-5:

0m52.1s ERROR: FAIL: Broken symlinks:
  /usr/share/man/man1/gnatgcc.1.gz -> gcc-5.1.gz


Andreas

Bug#857583: Acknowledgement (gcc-6-hppa64-linux-gnu: broken symlinks: /usr/share/man/man1/hppa64-linux-gnu-(cpp|gcc-.*)-6.1.gz -> (cpp|gcc-.*)-6.1.gz)

[Andreas Beckmann]

same problem in gcc-5-hppa64-linux-gnu:

0m30.3s ERROR: FAIL: Broken symlinks:
  /usr/share/man/man1/hppa64-linux-gnu-gcc-ranlib-5.1.gz -> gcc-ranlib-5.1.gz
  /usr/share/man/man1/hppa64-linux-gnu-gcc-nm-5.1.gz -> gcc-nm-5.1.gz
  /usr/share/man/man1/hppa64-linux-gnu-gcc-ar-5.1.gz -> gcc-ar-5.1.gz
  /usr/share/man/man1/hppa64-linux-gnu-cpp-5.1.gz -> cpp-5.1.gz

Bug#855640: gcc-6-cross: gcj-6- ship broken symlink /usr/lib/gcc-cross//6/ecj1 -> ../../ecj1

[Andreas Beckmann]

Followup-For: Bug #855640

Hi,

this problem also exists in the gcc-5-cross packages, should I clone
this bug?

Andreas

Bug#857583: gcc-6-hppa64-linux-gnu: broken symlinks: /usr/share/man/man1/hppa64-linux-gnu-(cpp|gcc-.*)-6.1.gz -> (cpp|gcc-.*)-6.1.gz

[Andreas Beckmann]

Package: gcc-6-hppa64-linux-gnu
Version: 6.3.0-8
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m39.4s ERROR: FAIL: Broken symlinks:
  /usr/share/man/man1/hppa64-linux-gnu-gcc-ranlib-6.1.gz -> gcc-ranlib-6.1.gz
  /usr/share/man/man1/hppa64-linux-gnu-gcc-nm-6.1.gz -> gcc-nm-6.1.gz
  /usr/share/man/man1/hppa64-linux-gnu-gcc-ar-6.1.gz -> gcc-ar-6.1.gz
  /usr/share/man/man1/hppa64-linux-gnu-cpp-6.1.gz -> cpp-6.1.gz

Not sure if this is the only cross compiler package with these broken
links, but the again, it's the only one built from src:gcc-6 instead of
src:gcc-6-cross or src:gcc-6-cross-ports.
I think I already reported a similar bug in another package, but can't
find it right now. Ah, there is is: #851698, and it wasn't me :-)


cheers,

Andreas


gcc-6-hppa64-linux-gnu_6.3.0-8.log.gz
Description: application/gzip

Bug#857384: gnat-6: broken symlinks: /usr/share/man/man1/{, x86_64-linux-gnu-}gnatgcc.1.gz -> x86_64-linux-gnu-gcc-6.1.gz

[Andreas Beckmann]

Package: gnat-6
Version: 6.3.0-8
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package ships (or creates)
a broken symlink.

>From the attached log (scroll to the bottom...):

0m40.5s ERROR: FAIL: Broken symlinks:
  /usr/share/man/man1/x86_64-linux-gnu-gnatgcc.1.gz -> 
x86_64-linux-gnu-gcc-6.1.gz
  /usr/share/man/man1/gnatgcc.1.gz -> x86_64-linux-gnu-gcc-6.1.gz


cheers,

Andreas


gnat-6_6.3.0-8.log.gz
Description: application/gzip

Bug#856875: gcc-6-cross: gcc-6- ship broken symlink: /usr/lib/gcc-cross//6/libcc1.so -> ../../../..//lib/libcc1.so.0

[Andreas Beckmann]

Source: gcc-6-cross
Version: 18
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed the gcc-6- cross compilers
ship a broken symlink.

>From the attached log (scroll to the bottom...):

0m32.5s ERROR: FAIL: Broken symlinks:
  /usr/lib/gcc-cross/mipsel-linux-gnu/6/libcc1.so -> 
../../../../mipsel-linux-gnu/lib/libcc1.so.0

There seems to be no foreign-arch variant of that library available
- do I have to install some foreign dependencies first?
But even installing libcc1-0:mipsel wouldn't satisfy the target path.

These related files/symlinks are shipped (on amd64):

gcc-6: /usr/lib/gcc/x86_64-linux-gnu/6/libcc1.so
gcc-6-aarch64-linux-gnu: /usr/lib/gcc-cross/aarch64-linux-gnu/6/libcc1.so
gcc-6-alpha-linux-gnu: /usr/lib/gcc-cross/alpha-linux-gnu/6/libcc1.so
gcc-6-arm-linux-gnueabi: /usr/lib/gcc-cross/arm-linux-gnueabi/6/libcc1.so
gcc-6-arm-linux-gnueabihf: /usr/lib/gcc-cross/arm-linux-gnueabihf/6/libcc1.so
gcc-6-hppa-linux-gnu: /usr/lib/gcc-cross/hppa-linux-gnu/6/libcc1.so
gcc-6-m68k-linux-gnu: /usr/lib/gcc-cross/m68k-linux-gnu/6/libcc1.so
gcc-6-mips-linux-gnu: /usr/lib/gcc-cross/mips-linux-gnu/6/libcc1.so
gcc-6-mips64-linux-gnuabi64: 
/usr/lib/gcc-cross/mips64-linux-gnuabi64/6/libcc1.so
gcc-6-mips64el-linux-gnuabi64: 
/usr/lib/gcc-cross/mips64el-linux-gnuabi64/6/libcc1.so
gcc-6-mipsel-linux-gnu: /usr/lib/gcc-cross/mipsel-linux-gnu/6/libcc1.so
gcc-6-powerpc-linux-gnu: /usr/lib/gcc-cross/powerpc-linux-gnu/6/libcc1.so
gcc-6-powerpc-linux-gnuspe: /usr/lib/gcc-cross/powerpc-linux-gnuspe/6/libcc1.so
gcc-6-powerpc64-linux-gnu: /usr/lib/gcc-cross/powerpc64-linux-gnu/6/libcc1.so
gcc-6-powerpc64le-linux-gnu: 
/usr/lib/gcc-cross/powerpc64le-linux-gnu/6/libcc1.so
gcc-6-s390x-linux-gnu: /usr/lib/gcc-cross/s390x-linux-gnu/6/libcc1.so
gcc-6-sh4-linux-gnu: /usr/lib/gcc-cross/sh4-linux-gnu/6/libcc1.so
gcc-6-sparc64-linux-gnu: /usr/lib/gcc-cross/sparc64-linux-gnu/6/libcc1.so
libcc1-0: /usr/lib/x86_64-linux-gnu/libcc1.so.0
libcc1-0: /usr/lib/x86_64-linux-gnu/libcc1.so.0.0.0

gcc-5-cross and gcc-7-cross have similar problems.


cheers,

Andreas


gcc-6-mipsel-linux-gnu_6.3.0-8cross1.log.gz
Description: application/gzip

Bug#855640: gcc-6-cross: gcj-6- ship broken symlink /usr/lib/gcc-cross//6/ecj1 -> ../../ecj1

[Andreas Beckmann]

Source: gcc-6-cross
Version: 16
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed that the gcj-6- packages
built from src:gcc-6-cross ship a broken symlink:

  /usr/lib/gcc-cross//6/ecj1 -> ../../ecj1

I couldn't find the target /usr/lib/gcc-cross/ecj1 in any package,
but there is /usr/lib/gcc/ecj1 (in package ecj1).


Andreas

Bug#851886: gcc: broken manpage symlinks

[Andreas Beckmann]

Control: reassign -1 gcc-6 6.3.0-3

On 2017-01-19 17:14, Matthias Klose wrote:
> That was fixed in today's gcc-6 6.3.0-3 upload. The man pages should be there.

looks better, but not perfect, since there is now a new issue:

0m28.5s ERROR: FAIL: Broken symlinks:
  /usr/share/man/man1/gcov-6.1 -> x86_64-linux-gnu-gcc-6.1
  /usr/share/man/man1/gcc-6.1 -> x86_64-linux-gnu-gcc-6.1

* link source and target are missing the .gz suffix
* there is no x86_64-linux-gnu-gcc-6.1.gz in any package

# ls -ls /usr/share/man/man*/*gcc* /usr/share/man/man*/*gcov*
  4 -rw-r--r-- 1 root root   1365 Jun 12  2013 /usr/share/man/man1/c89-gcc.1.gz
  4 -rw-r--r-- 1 root root   1420 Jan 18 16:11 /usr/share/man/man1/c99-gcc.1.gz
  0 lrwxrwxrwx 1 root root 24 Jan 18 19:02 /usr/share/man/man1/gcc-6.1 -> 
x86_64-linux-gnu-gcc-6.1
284 -rw-r--r-- 1 root root 287108 May  7  2016 /usr/share/man/man1/gcc-6.1.gz
  0 lrwxrwxrwx 1 root root 13 Jan 18 16:11 /usr/share/man/man1/gcc-ar.1.gz 
-> gcc-ar-6.1.gz
  0 lrwxrwxrwx 1 root root 13 Jan 18 16:11 /usr/share/man/man1/gcc-nm.1.gz 
-> gcc-nm-6.1.gz
  0 lrwxrwxrwx 1 root root 17 Jan 18 16:11 
/usr/share/man/man1/gcc-ranlib.1.gz -> gcc-ranlib-6.1.gz
  0 lrwxrwxrwx 1 root root 10 Sep  7 20:48 /usr/share/man/man1/gcc.1.gz -> 
gcc-6.1.gz
  0 lrwxrwxrwx 1 root root 24 Jan 18 19:02 /usr/share/man/man1/gcov-6.1 -> 
x86_64-linux-gnu-gcc-6.1
 12 -rw-r--r-- 1 root root   9132 May  7  2016 /usr/share/man/man1/gcov-6.1.gz
  0 lrwxrwxrwx 1 root root 11 Sep  7 20:48 /usr/share/man/man1/gcov.1.gz -> 
gcov-6.1.gz
  4 -rw-r--r-- 1 root root269 Jan 18 19:02 
/usr/share/man/man1/x86_64-linux-gnu-gcc-ar-6.1.gz
  0 lrwxrwxrwx 1 root root 13 Jan 18 16:11 
/usr/share/man/man1/x86_64-linux-gnu-gcc-ar.1.gz -> gcc-ar-6.1.gz
  4 -rw-r--r-- 1 root root269 Jan 18 19:02 
/usr/share/man/man1/x86_64-linux-gnu-gcc-nm-6.1.gz
  0 lrwxrwxrwx 1 root root 13 Jan 18 16:11 
/usr/share/man/man1/x86_64-linux-gnu-gcc-nm.1.gz -> gcc-nm-6.1.gz
  4 -rw-r--r-- 1 root root274 Jan 18 19:02 
/usr/share/man/man1/x86_64-linux-gnu-gcc-ranlib-6.1.gz
  0 lrwxrwxrwx 1 root root 17 Jan 18 16:11 
/usr/share/man/man1/x86_64-linux-gnu-gcc-ranlib.1.gz -> gcc-ranlib-6.1.gz

# dpkg -S /usr/share/man/man*/*gcc* /usr/share/man/man*/*gcov*
gcc: /usr/share/man/man1/c89-gcc.1.gz
gcc: /usr/share/man/man1/c99-gcc.1.gz
gcc-6: /usr/share/man/man1/gcc-6.1
gcc-6-doc: /usr/share/man/man1/gcc-6.1.gz
gcc: /usr/share/man/man1/gcc-ar.1.gz
gcc: /usr/share/man/man1/gcc-nm.1.gz
gcc: /usr/share/man/man1/gcc-ranlib.1.gz
gcc-doc: /usr/share/man/man1/gcc.1.gz
gcc-6: /usr/share/man/man1/x86_64-linux-gnu-gcc-ar-6.1.gz
gcc: /usr/share/man/man1/x86_64-linux-gnu-gcc-ar.1.gz
gcc-6: /usr/share/man/man1/x86_64-linux-gnu-gcc-nm-6.1.gz
gcc: /usr/share/man/man1/x86_64-linux-gnu-gcc-nm.1.gz
gcc-6: /usr/share/man/man1/x86_64-linux-gnu-gcc-ranlib-6.1.gz
gcc: /usr/share/man/man1/x86_64-linux-gnu-gcc-ranlib.1.gz
gcc-6: /usr/share/man/man1/gcov-6.1
gcc-6-doc: /usr/share/man/man1/gcov-6.1.gz
gcc-doc: /usr/share/man/man1/gcov.1.gz


Andreas

Bug#851886: gcc: broken manpage symlinks

[Andreas Beckmann]

Package: gcc
Version: 4:6.3.0-1
Severity: normal
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

the gcc metapackage installs a few broken manpage symlinks:

  /usr/share/man/man1/x86_64-linux-gnu-gcc-ranlib.1.gz -> gcc-ranlib-6.1.gz
  /usr/share/man/man1/x86_64-linux-gnu-gcc-nm.1.gz -> gcc-nm-6.1.gz
  /usr/share/man/man1/x86_64-linux-gnu-gcc-ar.1.gz -> gcc-ar-6.1.gz
  /usr/share/man/man1/gcc-ranlib.1.gz -> gcc-ranlib-6.1.gz
  /usr/share/man/man1/gcc-nm.1.gz -> gcc-nm-6.1.gz
  /usr/share/man/man1/gcc-ar.1.gz -> gcc-ar-6.1.gz

I cannot find /usr/share/man/man1/gcc-ranlib-6.1.gz in the archive,
(or in the chroot after installing the suggested gcc-doc package)
but there is

$ apt-file search  gcc-ranlib-6.1.gz
gcc-6: /usr/share/man/man1/x86_64-linux-gnu-gcc-ranlib-6.1.gz
gcc-6-aarch64-linux-gnu: /usr/share/man/man1/aarch64-linux-gnu-gcc-ranlib-6.1.gz
gcc-6-alpha-linux-gnu: /usr/share/man/man1/alpha-linux-gnu-gcc-ranlib-6.1.gz
gcc-6-arm-linux-gnueabi: /usr/share/man/man1/arm-linux-gnueabi-gcc-ranlib-6.1.gz
gcc-6-arm-linux-gnueabihf: 
/usr/share/man/man1/arm-linux-gnueabihf-gcc-ranlib-6.1.gz
...


Andreas

Bug#833829: libgcc-6-dev: Missing crtfastmath.o on kfreebsd-*

[Andreas Beckmann]

Control: retitle -1 libgcc-6-dev: Missing crtfastmath.o on kfreebsd-*
Control: affects -1 + src:vlc

On Tue, 9 Aug 2016 11:39:54 +0200 Matthias Klose  wrote:
> > I'm unable to build a package because crtfastmath.o is missing from this
> > package. Architecture is kfreebsd-i386. I don't see this bug on the
> > kfreebsd-amd64
> > 
> > ,
> > | cc -Wall -DPIC   -O2 -pipe -fno-tree-dominator-opts -fno-tree-pre 
> > -ffast-math -DUSE_MMX -DUSE_SSE -DUSE_SSE2 -g -D_FILE_OFFSET_BITS=64 
> > -D_LARGEFILE_SOURCE -fPIC -pthreadluma.c   -o luma
> > | /usr/bin/ld: cannot find crtfastmath.o: No such file or directory
> > `
> 
> https://buildd.debian.org/status/fetch.php?pkg=gcc-6=kfreebsd-i386=6.1.1-11=1470331624
> 
> looks like this is never built, and the install step succeeds despite it
> complains about not finding this file.

Seems to be specific to gcc-6, the file is available in libgcc-4.9-dev,
libgcc-5-dev, and gcc-snapshot (7.0.0).

This bug is the cause for the FTBFS of vlc on both kfreebsd-amd64 and
kfreebsd-i386:

https://buildd.debian.org/status/fetch.php?pkg=vlc=kfreebsd-amd64=2.2.4-3%2Bb4=1472902929
https://buildd.debian.org/status/fetch.php?pkg=vlc=kfreebsd-i386=2.2.4-3%2Bb4=1472905628

Andreas

Bug#826583: gdc-5: fails to upgrade from 'testing' - trying to overwrite /usr/lib/gcc/x86_64-linux-gnu/5/include/d/__entrypoint.di

[Andreas Beckmann]

Package: gdc-5
Version: 5.4.0-3
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'testing'.
It installed fine in 'testing', then the upgrade to 'sid' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package gdc-5.
  Preparing to unpack .../gdc-5_5.4.0-3_amd64.deb ...
  Unpacking gdc-5 (5.4.0-3) ...
  dpkg: error processing archive 
/var/cache/apt/archives/gdc-5_5.4.0-3_amd64.deb (--unpack):
   trying to overwrite 
'/usr/lib/gcc/x86_64-linux-gnu/5/include/d/__entrypoint.di', which is also in 
package libphobos-5-dev 5.3.1-21
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
  Processing triggers for libc-bin (2.22-9) ...
  Errors were encountered while processing:
   /var/cache/apt/archives/gdc-5_5.4.0-3_amd64.deb


cheers,

Andreas


libphobos-5-dev=5.3.1-21_gdc-5=5.4.0-3.log.gz
Description: application/gzip

Bug#825361: libgcj-bc: shlibs file generates unsatisfiable dependency

[Andreas Beckmann]

Package: libgcj-bc
Version: 5.3.1-2
Severity: serious

Hi,

the libgcj-bc.shlibs file contains

libgcj_bc 1 libgcj-bc (>= 5.3.1-17~)

which is unsatisfiable and causes e.g. uninstallable ant packages:
https://packages.qa.debian.org/a/ant.html

Should this have been

libgcj_bc 1 libgcj16 (>= 5.3.1-17~)

instead?


Andreas

Bug#825320: gcc-doc-defaults: FTBFS in sid: error in 'Version' field string '5:': nothing after colon in version number

[Andreas Beckmann]

Source: gcc-doc-defaults
Version: 5:14
Severity: serious
Justification: fails to build from source (but built successfully in the past)

Hi,

gcc-doc-defaults FTBFS in curent sid:

   dh_builddeb
dpkg-deb: building package 'gcc-doc' in '../gcc-doc_5.1.0-1_amd64.deb'.
dpkg-deb: building package 'gccgo-doc' in '../gccgo-doc_5.1.0-1_amd64.deb'.
dpkg-deb: building package 'gcj-doc' in '../gcj-doc_5.1.0-1_amd64.deb'.
dpkg-deb: error: parsing file 'debian/gnat-doc/DEBIAN/control' near line 3 
package 'gnat-doc':
 error in 'Version' field string '5:': nothing after colon in version number
dpkg-deb: building package 'gfortran-doc' in 
'../gfortran-doc_5.1.0-1_amd64.deb'.
dh_builddeb: dpkg-deb --build debian/gnat-doc .. returned exit code 2
dpkg-deb: building package 'gcc-doc-base' in '../gcc-doc-base_4.9.2-1_all.deb'.
dpkg-deb: building package 'cpp-doc' in '../cpp-doc_5.1.0-1_amd64.deb'.
debian/rules:100: recipe for target 'binary' failed
make: *** [binary] Error 1
dpkg-buildpackage: error: fakeroot debian/rules binary gave error exit status 2


Andreas


gcc-doc-defaults_5%14.log.gz
Description: application/gzip

Bug#818873: gcc-5-cross: FTBFS when built with dpkg-buildpackage -A (No such file or directory)

[Andreas Beckmann]

Control: tag -1 - help + pending

On 2016-05-09 11:54, Matthias Klose wrote:
> On 09.05.2016 10:17, Andreas Beckmann wrote:
>> Followup-For: Bug #818873
> 
> that's already fixed in the VCS.

Good, so let's fix the tags :-)

Andreas

Bug#818873: gcc-5-cross: FTBFS when built with dpkg-buildpackage -A (No such file or directory)

[Andreas Beckmann]

Followup-For: Bug #818873

Hi,

in 
https://buildd.debian.org/status/fetch.php?pkg=gcc-5-cross=all=21=1462410260
 the failing command is 

DEB_HOST_ARCH=s390x ARCH=s390x MAKEFLAGS="CC=something" dh_shlibdeps 
-plibgnat-5-s390x-cross \

-l/lib/s390x-linux-gnu:/usr/lib/s390x-linux-gnu:/s390x-linux-gnu/lib:/usr/s390x-linux-gnu/lib:/«PKGBUILDDIR»/gcc/debian/tmp/usr/s390x-linux-gnu/lib:/«PKGBUILDDIR»/gcc/debian/tmp/usr/s390x-linux-gnu/lib
dpkg-shlibdeps: error: no dependency information found for 
/«PKGBUILDDIR»/gcc/debian/tmp/usr/s390x-linux-gnu/lib/libgcc_s.so.1 (used by 
debian/libgnat-5-s390x-cross/usr/s390x-linux-gnu/lib/libgnarl-5.so.1)
Hint: check if the library actually comes from a package.
dh_shlibdeps: dpkg-shlibdeps -Tdebian/libgnat-5-s390x-cross.substvars 
-l/lib/s390x-linux-gnu -l/usr/lib/s390x-linux-gnu -l/s390x-linux-gnu/lib 
-l/usr/s390x-linux-gnu/lib 
-l/«PKGBUILDDIR»/gcc/debian/tmp/usr/s390x-linux-gnu/lib 
-l/«PKGBUILDDIR»/gcc/debian/tmp/usr/s390x-linux-gnu/lib 
debian/libgnat-5-s390x-cross/usr/s390x-linux-gnu/lib/libgnat-5.so.1 
debian/libgnat-5-s390x-cross/usr/s390x-linux-gnu/lib/libgnarl-5.so.1 returned 
exit code 2

That's actually in the gcc source code being built by gcc-5-cross.
This call looks correct and there is also a
/usr/s390x-linux-gnu/lib/libgcc_s.so.1 available.
But if you dig into dpkg-shlibdeps source code, you see that the -l option
works rather counterintuitive. The -l paths get prepended to the list of
(default) search paths one-by-one, so that the last -l option gets actually
searched first.

I'm not sure who should be blamed here: dpkg, debhelper or gcc.
Neither dpkg-shlibdeps nor dh_shlibdeps document anything about the order
in their manpages.
Why is /«PKGBUILDDIR»/gcc/debian/tmp/usr/s390x-linux-gnu/lib included in
the search path anyway? AIUI if some library can (only) be satisfied from
there, dpkg-shlibdeps will always fail.

I could reproduce this in a sid/amd64 for an arch-indep build. (didn't verify
what is different in a full arch+indep build since this seems to take extremely
long). Changing the search path order lets this command succeed.


Andreas

Bug#810809: gnat-6-, libgnat{prj, vsn}6-dev--cross: fails to upgrade from 'sid' - trying to overwrite /usr/bin/-gnatxref, /usr//lib/libgnat{vsn, prj}.a

[Andreas Beckmann]

Followup-For: Bug #810809
Control: reassign -1 src:gcc-6 6-20160205-1
Control: reopen -1
Control: affects -1 + src:gcc-6-cross src:gcc-6-cross-ports

Hi,

the *gnat*-cross package relationships seem not to work correctly, yet:

  Selecting previously unselected package gnat-6-aarch64-linux-gnu.
  Preparing to unpack .../gnat-6-aarch64-linux-gnu_6-20160205-1cross1_amd64.deb 
...
  Unpacking gnat-6-aarch64-linux-gnu (6-20160205-1cross1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/gnat-6-aarch64-linux-gnu_6-20160205-1cross1_amd64.deb 
(--unpack):
   trying to overwrite '/usr/bin/aarch64-linux-gnu-gnatxref', which is also in 
package gnat-5-aarch64-linux-gnu 5.3.1-8cross1
  Processing triggers for libc-bin (2.21-7) ...
  Errors were encountered while processing:
   /var/cache/apt/archives/gnat-6-aarch64-linux-gnu_6-20160205-1cross1_amd64.deb

and so on ...


Andreas

Bug#814527: libgccjit-6-doc: should Conflicts: libgccjit-5-doc, not itself

[Andreas Beckmann]

Package: libgccjit-6-doc
Version: 6-20160205-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package libgccjit-6-doc.
  Preparing to unpack .../libgccjit-6-doc_6-20160205-1_all.deb ...
  Unpacking libgccjit-6-doc (6-20160205-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgccjit-6-doc_6-20160205-1_all.deb (--unpack):
   trying to overwrite '/usr/share/info/libgccjit.info.gz', which is also in 
package libgccjit-5-doc 5.3.1-8
  Errors were encountered while processing:
   /var/cache/apt/archives/libgccjit-6-doc_6-20160205-1_all.deb

Package: libgccjit-6-doc
Source: gcc-6
Version: 6-20160205-1
Conflicts: libgccjit-6-doc

That should probably be
  Conflicts: libgccjit-5-doc


cheers,

Andreas

Bug#813081: lib{, x}32stdc++6-6-dbg, libgccjit-6-doc: fails to upgrade from 'sid' - trying to overwrite /usr/lib{, x}32/debug/libstdc++fs.a, /usr/share/info/libgccjit.info.gz

[Andreas Beckmann]

Package: libx32stdc++6-6-dbg,lib32stdc++6-6-dbg,libgccjit-6-doc
Version: 6-20160122-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package libx32stdc++6-6-dbg.
  Preparing to unpack .../libx32stdc++6-6-dbg_6-20160122-1_amd64.deb ...
  Unpacking libx32stdc++6-6-dbg (6-20160122-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libx32stdc++6-6-dbg_6-20160122-1_amd64.deb (--unpack):
   trying to overwrite '/usr/libx32/debug/libstdc++fs.a', which is also in 
package libx32stdc++6-5-dbg 5.3.1-7
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
  Processing triggers for libc-bin (2.21-7) ...
  Errors were encountered while processing:
   /var/cache/apt/archives/libx32stdc++6-6-dbg_6-20160122-1_amd64.deb


  Selecting previously unselected package lib32stdc++6-6-dbg.
  Preparing to unpack .../lib32stdc++6-6-dbg_6-20160122-1_amd64.deb ...
  Unpacking lib32stdc++6-6-dbg (6-20160122-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/lib32stdc++6-6-dbg_6-20160122-1_amd64.deb (--unpack):
   trying to overwrite '/usr/lib32/debug/libstdc++fs.a', which is also in 
package lib32stdc++6-5-dbg 5.3.1-7
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
  Processing triggers for libc-bin (2.21-7) ...
  Errors were encountered while processing:
   /var/cache/apt/archives/lib32stdc++6-6-dbg_6-20160122-1_amd64.deb


  Selecting previously unselected package libgccjit-6-doc.
  Preparing to unpack .../libgccjit-6-doc_6-20160122-1_all.deb ...
  Unpacking libgccjit-6-doc (6-20160122-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgccjit-6-doc_6-20160122-1_all.deb (--unpack):
   trying to overwrite '/usr/share/info/libgccjit.info.gz', which is also in 
package libgccjit-5-doc 5.3.1-7
  Errors were encountered while processing:
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
   /var/cache/apt/archives/libgccjit-6-doc_6-20160122-1_all.deb


cheers,

Andreas


libx32stdc++6-5-dbg=5.3.1-7_libx32stdc++6-6-dbg=6-20160122-1.log.gz
Description: application/gzip

Bug#810514: closed by Matthias Klose <d...@debian.org> (Re: Bug#810514: libgcj17{,-awt}: missing Breaks+Replaces on libgcj16{,-awt})

[Andreas Beckmann]

Control: found -1 6-20160109-1

On 2016-01-09 20:51, Debian Bug Tracking System wrote:
> Version: 6-20160109-1

I can still reproduce these overwrite conflicts with the 20160109 snapshot.


Andreas

Bug#810809: gnat-6-, libgnat{prj, vsn}6-dev--cross: fails to upgrade from 'sid' - trying to overwrite /usr/bin/-gnatxref, /usr//lib/libgnat{vsn, prj}.a

[Andreas Beckmann]

Source: gcc-6-cross
Version: 6-20160109-1cross1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

During a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

>From the attached log (scroll to the bottom...):
(log from arm64-cross, but the file conflict happens on all cross architectures)

  Selecting previously unselected package gnat-6-aarch64-linux-gnu.
  Preparing to unpack .../gnat-6-aarch64-linux-gnu_6-20160109-1cross1_amd64.deb 
...
  Unpacking gnat-6-aarch64-linux-gnu (6-20160109-1cross1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/gnat-6-aarch64-linux-gnu_6-20160109-1cross1_amd64.deb 
(--unpack):
   trying to overwrite '/usr/bin/aarch64-linux-gnu-gnatxref', which is also in 
package gnat-5-aarch64-linux-gnu 5.3.1-5cross1
  Selecting previously unselected package libgnatvsn6-arm64-cross.
  Preparing to unpack .../libgnatvsn6-arm64-cross_6-20160109-1cross1_all.deb ...
  Unpacking libgnatvsn6-arm64-cross (6-20160109-1cross1) ...
  Selecting previously unselected package libgnatprj6-arm64-cross.
  Preparing to unpack .../libgnatprj6-arm64-cross_6-20160109-1cross1_all.deb ...
  Unpacking libgnatprj6-arm64-cross (6-20160109-1cross1) ...
  Selecting previously unselected package libgnatvsn6-dev-arm64-cross.
  Preparing to unpack 
.../libgnatvsn6-dev-arm64-cross_6-20160109-1cross1_all.deb ...
  Unpacking libgnatvsn6-dev-arm64-cross (6-20160109-1cross1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgnatvsn6-dev-arm64-cross_6-20160109-1cross1_all.deb 
(--unpack):
   trying to overwrite '/usr/aarch64-linux-gnu/lib/libgnatvsn.a', which is also 
in package libgnatvsn5-dev-arm64-cross 5.3.1-5cross1
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
  Selecting previously unselected package libgnatprj6-dev-arm64-cross.
  Preparing to unpack 
.../libgnatprj6-dev-arm64-cross_6-20160109-1cross1_all.deb ...
  Unpacking libgnatprj6-dev-arm64-cross (6-20160109-1cross1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgnatprj6-dev-arm64-cross_6-20160109-1cross1_all.deb 
(--unpack):
   trying to overwrite '/usr/aarch64-linux-gnu/lib/libgnatprj.a', which is also 
in package libgnatprj5-dev-arm64-cross 5.3.1-5cross1
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
  Processing triggers for libc-bin (2.21-6) ...
  Errors were encountered while processing:
   /var/cache/apt/archives/gnat-6-aarch64-linux-gnu_6-20160109-1cross1_amd64.deb
   
/var/cache/apt/archives/libgnatvsn6-dev-arm64-cross_6-20160109-1cross1_all.deb
   
/var/cache/apt/archives/libgnatprj6-dev-arm64-cross_6-20160109-1cross1_all.deb


cheers,

Andreas


libgnatprj5-dev-arm64-cross=5.3.1-5cross1_libgnatprj6-dev-arm64-cross=6-20160109-1cross1.log.gz
Description: application/gzip

Bug#810717: libstdc++-6-doc: package corrupted?

[Andreas Beckmann]

Package: libstdc++-6-doc
Version: 6-20160109-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package failed to install. As
per definition of the release team this makes the package too buggy for
a release, thus the severity.

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package libstdc++-6-doc.
  Preparing to unpack .../libstdc++-6-doc_6-20160109-1_all.deb ...
  Unpacking libstdc++-6-doc (6-20160109-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libstdc++-6-doc_6-20160109-1_all.deb (--unpack):
   unable to clean up mess surrounding 
'./usr/share/man/man3/__gnu_pbds_detail_tree_traits_
  Key,   Mapped,   Cmp_Fn,  
 Node_Update,   ov_tree_tag,
' before installing another version: File name too long
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)
  Errors were encountered while processing:
   /var/cache/apt/archives/libstdc++-6-doc_6-20160109-1_all.deb

md5sum:
d99dce5587f0aa05758639e3f598275d  
/var/cache/apt/archives/libstdc++-6-doc_6-20160109-1_all.deb

cheers,

Andreas

Bug#810514: libgcj17{,-awt}: missing Breaks+Replaces on libgcj16{,-awt}

[Andreas Beckmann]

Package: libgcj17,libgcj17-awt
Version: 6-20160103-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package failed to install
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package libgcj17:amd64.
  Preparing to unpack .../libgcj17_6-20160103-1_amd64.deb ...
  Unpacking libgcj17:amd64 (6-20160103-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgcj17_6-20160103-1_amd64.deb (--unpack):
   trying to overwrite 
'/usr/lib/jvm/java-1.5.0-gcj-6-amd64/jre/lib/security/java.security', which is 
also in package libgcj16:amd64 6-20160101-1


  Selecting previously unselected package libgcj16-awt:amd64.
  Preparing to unpack .../libgcj16-awt_6-20160101-1_amd64.deb ...
  Unpacking libgcj16-awt:amd64 (6-20160101-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgcj16-awt_6-20160101-1_amd64.deb (--unpack):
   trying to overwrite '/usr/lib/jvm/java-1.5.0-gcj-6-amd64/lib/libjawt.so', 
which is also in package libgcj17-awt:amd64 6-20160103-1


cheers,

Andreas


libgcj16=6-20160101-1_libgcj17=6-20160103-1.log.gz
Description: application/gzip

Bug#810514: lib32stdc++6-6-dbg, libx32stdc++6-6-dbg, libgccjit-6-doc: missing Breaks+Replaces on their *-5-* counterparts

[Andreas Beckmann]

Severity: serious
Followup-For: Bug #810514
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package lib32stdc++6-6-dbg.
  Preparing to unpack .../lib32stdc++6-6-dbg_6-20160103-1_amd64.deb ...
  Unpacking lib32stdc++6-6-dbg (6-20160103-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/lib32stdc++6-6-dbg_6-20160103-1_amd64.deb (--unpack):
   trying to overwrite '/usr/lib32/debug/libstdc++.a', which is also in package 
lib32stdc++6-5-dbg 5.3.1-5

  Selecting previously unselected package libgccjit-6-doc.
  Preparing to unpack .../libgccjit-6-doc_6-20160103-1_all.deb ...
  Unpacking libgccjit-6-doc (6-20160103-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libgccjit-6-doc_6-20160103-1_all.deb (--unpack):
   trying to overwrite '/usr/share/info/libgccjit.info.gz', which is also in 
package libgccjit-5-doc 5.3.1-5

  Selecting previously unselected package libx32stdc++6-6-dbg.
  Preparing to unpack .../libx32stdc++6-6-dbg_6-20160103-1_amd64.deb ...
  Unpacking libx32stdc++6-6-dbg (6-20160103-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libx32stdc++6-6-dbg_6-20160103-1_amd64.deb (--unpack):
   trying to overwrite '/usr/libx32/debug/libstdc++.a', which is also in 
package libx32stdc++6-5-dbg 5.3.1-5


cheers,

Andreas

Bug#807885: libstdc++6-6-dbg: fails to upgrade from 'sid' - trying to overwrite /usr/lib/x86_64-linux-gnu/debug/libstdc++.a

[Andreas Beckmann]

Package: libstdc++6-6-dbg
Version: 6-20151211-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package fails to upgrade from
'sid' to 'experimental'.
It installed fine in 'sid', then the upgrade to 'experimental' fails
because it tries to overwrite other packages files without declaring a
Breaks+Replaces relation.

See policy 7.6 at
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces

>From the attached log (scroll to the bottom...):

  Selecting previously unselected package libstdc++6-6-dbg:amd64.
  (Reading database ... ^M(Reading database ... 5%^M(Reading database ... 
10%^M(Reading database ... 15%^M(Reading database ... 20%^M(Reading database 
... 25%^M(Reading database ... 30%^M(Reading database ... 35%^M^M(Reading 
database ... 40%^M(Reading database ... 45%^M(Reading database ... 
50%^M(Reading database ... 55%^M(Reading database ... 60%^M(Reading database 
... 65%^M(Reading database ... 70%^M(Reading database ... 75%^M(Reading 
database ... 80%^M(Reading database ... 85%^M(Reading database ... 
90%^M(Reading database ... 95%^M(Reading database ... 100%^M(Reading database 
... 9570 files and directories currently installed.)
  Preparing to unpack .../libstdc++6-6-dbg_6-20151211-1_amd64.deb ...
  Unpacking libstdc++6-6-dbg:amd64 (6-20151211-1) ...
  dpkg: error processing archive 
/var/cache/apt/archives/libstdc++6-6-dbg_6-20151211-1_amd64.deb (--unpack):
   trying to overwrite '/usr/lib/x86_64-linux-gnu/debug/libstdc++.a', which is 
also in package libstdc++6-5-dbg:amd64 5.3.1-3
  Errors were encountered while processing:
   /var/cache/apt/archives/libstdc++6-6-dbg_6-20151211-1_amd64.deb
  dpkg-deb: error: subprocess paste was killed by signal (Broken pipe)

There are similar problems with lib32stdc++6-6-dbg,
libx32stdc++6-6-dbg and libgccjit-6-doc.


cheers,

Andreas


libstdc++6-5-dbg=5.3.1-3_libstdc++6-6-dbg=6-20151211-1.log.gz
Description: application/gzip

Bug#805616: gcc-4.9: GCC++ 4.9 compiler had ICE (internal compiler error)

[Andreas Beckmann]

verified in g++-4.9 built from svn:

$ g++-4.9.x -v -std=gnu++11 -fno-strict-aliasing -O2 -c 805616..ii
Using built-in specs.
COLLECT_GCC=/opt/software/x86_64/gcc-4.9.x/bin/g++-4.9.x
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-4_9-branch/configure 
--prefix=/opt/software/x86_64/gcc-4.9.x --program-suffix=-4.9.x 
--enable-languages=c,c++,fortran --enable-checking
Thread model: posix
gcc version 4.9.4 20151121 (prerelease) (GCC) 
COLLECT_GCC_OPTIONS='-v' '-std=gnu++11' '-fno-strict-aliasing' '-O2' '-c' 
'-shared-libgcc' '-mtune=generic' '-march=x86-64'
 
/opt/software/x86_64/gcc-4.9.x/libexec/gcc/x86_64-unknown-linux-gnu/4.9.4/cc1plus
 -fpreprocessed 805616..ii -quiet -dumpbase 805616..ii -mtune=generic 
-march=x86-64 -auxbase 805616. -O2 -std=gnu++11 -version -fno-strict-aliasing 
-o /tmp/ccmf892Y.s
GNU C++ (GCC) version 4.9.4 20151121 (prerelease) (x86_64-unknown-linux-gnu)
compiled by GNU C version 4.9.4 20151121 (prerelease), GMP version 
6.0.0, MPFR version 3.1.3, MPC version 1.0.3
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
GNU C++ (GCC) version 4.9.4 20151121 (prerelease) (x86_64-unknown-linux-gnu)
compiled by GNU C version 4.9.4 20151121 (prerelease), GMP version 
6.0.0, MPFR version 3.1.3, MPC version 1.0.3
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
Compiler executable checksum: c8476ec2ebe2d7bd0e72dc80313e75a7
In file included from 
/new-hd1/extra/ishikawa/TB-3HG/NEW-COMMSRC/mozilla/gfx/skia/skia/include/effects/SkLayerRasterizer.h:13:0,
 from 
/new-hd1/extra/ishikawa/TB-3HG/NEW-COMMSRC/mozilla/gfx/2d/DrawTargetSkia.cpp:24,
 from 
/home/ishikawa/objdir-tb3/gfx/2d/Unified_cpp_gfx_2d0.cpp:83:
/new-hd1/extra/ishikawa/TB-3HG/NEW-COMMSRC/mozilla/gfx/skia/skia/include/core/SkRasterizer.h:
 In member function ‘virtual void mozilla::gfx::DrawTargetTiled::PopClip()’:
/new-hd1/extra/ishikawa/TB-3HG/NEW-COMMSRC/mozilla/gfx/skia/skia/include/core/SkRasterizer.h:21:7:
 internal compiler error: Segmentation fault
0xb78fff crash_signal
../../gcc-4_9-branch/gcc/toplev.c:337
0x9a1669 is_gimple_reg_type
../../gcc-4_9-branch/gcc/gimple-expr.h:75
0x9a1669 is_gimple_val(tree_node*)
../../gcc-4_9-branch/gcc/gimple-expr.c:814
0xb9e35d verify_types_in_gimple_min_lval
../../gcc-4_9-branch/gcc/tree-cfg.c:3007
0xb9e35d verify_types_in_gimple_reference
../../gcc-4_9-branch/gcc/tree-cfg.c:3149
0xba12d0 verify_gimple_assign_single
../../gcc-4_9-branch/gcc/tree-cfg.c:4004
0xba12d0 verify_gimple_assign
../../gcc-4_9-branch/gcc/tree-cfg.c:4183
0xba12d0 verify_gimple_stmt
../../gcc-4_9-branch/gcc/tree-cfg.c:4417
0xba7cad verify_gimple_in_cfg(function*)
../../gcc-4_9-branch/gcc/tree-cfg.c:4882
0xace5d2 execute_function_todo
../../gcc-4_9-branch/gcc/passes.c:1853
0xacef33 execute_todo
../../gcc-4_9-branch/gcc/passes.c:1887
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See  for instructions.

There are several google hits for
  "segmentation fault is_gimple_reg_type"
  "segmentation fault is_gimple_reg_type is_gimple_val"
but all with different backtraces
  "segmentation fault is_gimple_reg_type is_gimple_val 
verify_types_in_gimple_min_lval"
gets no results at all, so this seems to be a new PR :-)


Minimizing the example to something managable ... and understood by different 
compiler versions as well

4.9.4 (SVN): segfault
4.9.3-5 (Debian): segfault
4.8.5-1 (Debian): pass
4.7.4-3 (Debian): compile error
5.2.1-23 (Debian): segfault, different backtrace


$ g++-4.9.x -v -std=gnu++11 -O2 -c 805616_min.hM.min.ii
Using built-in specs.
COLLECT_GCC=/opt/software/x86_64/gcc-4.9.x/bin/g++-4.9.x
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc-4_9-branch/configure 
--prefix=/opt/software/x86_64/gcc-4.9.x --program-suffix=-4.9.x 
--enable-languages=c,c++,fortran --enable-checking
Thread model: posix
gcc version 4.9.4 20151121 (prerelease) (GCC) 
COLLECT_GCC_OPTIONS='-v' '-std=gnu++11' '-O2' '-c' '-shared-libgcc' 
'-mtune=generic' '-march=x86-64'
 
/opt/software/x86_64/gcc-4.9.x/libexec/gcc/x86_64-unknown-linux-gnu/4.9.4/cc1plus
 -fpreprocessed 805616_min.hM.min.ii -quiet -dumpbase 805616_min.hM.min.ii 
-mtune=generic -march=x86-64 -auxbase 805616_min.hM.min -O2 -std=gnu++11 
-version -o /tmp/ccdwpTpg.s
GNU C++ (GCC) version 4.9.4 20151121 (prerelease) (x86_64-unknown-linux-gnu)
compiled by GNU C version 4.9.4 20151121 (prerelease), GMP version 
6.0.0, MPFR version 3.1.3, MPC version 1.0.3
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
GNU C++ (GCC) version 4.9.4 20151121 (prerelease) (x86_64-unknown-linux-gnu)
compiled by GNU C version 4.9.4 20151121 (prerelease), GMP version 
6.0.0, MPFR version 3.1.3, MPC version 1.0.3
GGC heuristics: --param ggc-min-expand=30 --param 

Bug#799953: gcc-4.9: incorrect double to integer conversion on i386

[Andreas Beckmann]

Control: forwarded -1 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=323

On Thu, 24 Sep 2015 18:42:14 +0200 Miroslav Urbanek  
wrote:
> I believe I've found a bug in GCC that affects plymouth and maybe
> other packages on i386. The following minimal code produces an
> incorrect result.

At least it's an unexpected result ... seems like you are running in some 
excess precision issue, probably PR323


I can also reproduce it on amd64 when explicitly using the 387 FPU:

$ gcc-5 -O1 -mfpmath=387 -o test test.c && ./test 100
f = 999.00
d = 99872.00
g = 99872.00
fractional part of d is 28.00

and I cannot reproduce it on armhf (in qemu).

gcc-5 -O2 -o test test.c && ./test 100
f = 999.00
d = 99872.00
g = 99872.00
fractional part of d is 0.00


amd64 disassembly with -mfpmath=387:
-O0:

00400546 :
#  400546:   55  push   %rbp
#  400547:   48 89 e5mov%rsp,%rbp
#  40054a:   48 83 ec 40 sub$0x40,%rsp
#  40054e:   89 7d dcmov%edi,-0x24(%rbp)
#  400551:   48 89 75 d0 mov%rsi,-0x30(%rbp)
  400555:   d9 05 6d 01 00 00   flds   0x16d(%rip)# 4006c8 
<_IO_stdin_used+0x38>
  40055b:   d9 5d fcfstps  -0x4(%rbp)   = f 
(float)
#  40055e:   48 8b 45 d0 mov-0x30(%rbp),%rax
#  400562:   48 83 c0 08 add$0x8,%rax
#  400566:   48 8b 00mov(%rax),%rax
#  400569:   48 89 c7mov%rax,%rdi
#  40056c:   e8 cf fe ff ff  callq  400440 
  400571:   89 45 c8mov%eax,-0x38(%rbp) = 
argv[1] (int)
  400574:   db 45 c8fildl  -0x38(%rbp)  
int_to_extended(argv(1))
  400577:   d8 4d fcfmuls  -0x4(%rbp)
  40057a:   dd 5d f0fstpl  -0x10(%rbp)  = d 
(double)
  40057d:   dd 45 f0fldl   -0x10(%rbp)
  400580:   d9 5d ecfstps  -0x14(%rbp)  = g 
(float)
  400583:   f2 0f 10 45 f0  movsd  -0x10(%rbp),%xmm0
  400588:   f2 0f 2c c0 cvttsd2si %xmm0,%eax(int)d 
(double2int with truncation)
  40058c:   89 45 e8mov%eax,-0x18(%rbp) = i 
(int)
#  40058f:   d9 45 fcflds   -0x4(%rbp)  print f
#  400592:   dd 5d c8fstpl  -0x38(%rbp)
#  400595:   f2 0f 10 45 c8  movsd  -0x38(%rbp),%xmm0
#  40059a:   bf 94 06 40 00  mov$0x400694,%edi
#  40059f:   b8 01 00 00 00  mov$0x1,%eax
#  4005a4:   e8 67 fe ff ff  callq  400410 
#  4005a9:   48 8b 45 f0 mov-0x10(%rbp),%raxprint d
#  4005ad:   48 89 45 c8 mov%rax,-0x38(%rbp)
#  4005b1:   f2 0f 10 45 c8  movsd  -0x38(%rbp),%xmm0
#  4005b6:   bf 9c 06 40 00  mov$0x40069c,%edi
#  4005bb:   b8 01 00 00 00  mov$0x1,%eax
#  4005c0:   e8 4b fe ff ff  callq  400410 
#  4005c5:   d9 45 ecflds   -0x14(%rbp) print g
#  4005c8:   dd 5d c8fstpl  -0x38(%rbp)
#  4005cb:   f2 0f 10 45 c8  movsd  -0x38(%rbp),%xmm0
#  4005d0:   bf a4 06 40 00  mov$0x4006a4,%edi
#  4005d5:   b8 01 00 00 00  mov$0x1,%eax
#  4005da:   e8 31 fe ff ff  callq  400410 
  4005df:   db 45 e8fildl  -0x18(%rbp)  
int_to_extended(i)
  4005e2:   dd 45 f0fldl   -0x10(%rbp)  
double_ro_extended(d)
  4005e5:   de e1   fsubp  %st,%st(1)
  4005e7:   dd 5d c8fstpl  -0x38(%rbp)  = d-i
#  4005ea:   f2 0f 10 45 c8  movsd  -0x38(%rbp),%xmm0   print 
d-i
#  4005ef:   bf ac 06 40 00  mov$0x4006ac,%edi
#  4005f4:   b8 01 00 00 00  mov$0x1,%eax
#  4005f9:   e8 12 fe ff ff  callq  400410 
#  4005fe:   90  nop
#  4005ff:   c9  leaveq 
#  400600:   c3  retq   

-O1:

00400546 :
#  400546:   48 83 ec 28 sub$0x28,%rsp
#  40054a:   48 8b 7e 08 mov0x8(%rsi),%rdi
#  40054e:   ba 0a 00 00 00  mov$0xa,%edx
#  400553:   be 00 00 00 00  mov$0x0,%esi
#  400558:   e8 e3 fe ff ff  callq  400440 
  40055d:   48 89 44 24 08  mov%rax,0x8(%rsp)   = 
argv[1] (long)
  400562:   db 44 24 08 fildl  0x8(%rsp)
int_to_extended(argv(1))fpu stack: argv(1)
  400566:   d9 05 4c 01 00 00   flds   0x14c(%rip)# 4006b8 
<_IO_stdin_used+0x38>fpu 

Bug#794329: gdc-multilib: uninstallable in sid: Depends: gdc (= 4:5.2.1-2), but sid has gdc (5.2.1-2)

[Andreas Beckmann]

Package: gdc-multilib
Version: 4:5.2.1-2
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts

Hi,

during a test with piuparts I noticed your package is not
installable in sid:

gdc-multilib/amd64 unsatisfiable Depends: gdc (= 4:5.2.1-2)
gdc-multilib/i386 unsatisfiable Depends: gdc (= 4:5.2.1-2)

since

 gdc | 5.1.1-1 | stretch | amd64, arm64, armel, armhf, i386, mips, 
mipsel, powerpc, ppc64el, s390x
 gdc | 5.1.1-1 | sid | arm64, armel
 gdc | 5.2.1-2 | sid | amd64, armhf, hurd-i386, i386, 
kfreebsd-amd64, kfreebsd-i386, mips, mipsel, powerpc, ppc64el, s390x


Cheers,

Andreas


-- 
To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: 
https://lists.debian.org/20150801130152.31864.70841.report...@zam581.zam.kfa-juelich.de

Bug#787486: ppl: FTBFS with swi-prolog 7.x

[Andreas Beckmann]

Control: retitle -1 ppl: FTBFS with swi-prolog 7.x

Since it built in 'stretch/testing' successfully I looked at the
difference in installed packages after satisfying the build dependencies
... the only interesting package with a newer upstream version in sid
seemed to be swi-prolog ... and then rebuilt it again in 'sid/unstable'
but with swi-prolog 6.x from 'testing', which worked well.


Andreas


-- 
To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/55acb480@debian.org