Your message dated Tue, 24 Sep 2024 20:27:15 +0200
with message-id
<cagkjw9+oofnnusfwcboigub6rkburmytc2cuan8dr64lqm+...@mail.gmail.com>
and subject line Re: Bug#1082558: dpkg: confusing warning message
has caused the Debian Bug report #1082558,
regarding dpkg: confusing warning message
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)
--
1082558: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1082558
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Source: dpkg
Version:1.22.11
Severity: wishlist
Dear dpkg maintainers,
On a biarch system (amd64,i386) and using --force options (note that I am
aware of the taken risk ;-)), dpkg shows:
$ dpkg --force-all -i libllvm19_1%3a19.1.0-3_i386.deb
(Reading database ... 722987 files and directories currently installed.)
Preparing to unpack libllvm19_1%3a19.1.0-3_i386.deb ...
Unpacking libllvm19:i386 (1:19.1.0-3) over (1:19.1.0-3) ...
dpkg: warning: overriding problem because --force enabled:
dpkg: warning: trying to overwrite shared
'/usr/lib/llvm-19/lib/libLLVM.so.19.1', which is different from other
instances of package libllvm19:i386
I think that the given message is confusing while including the arch,
here (:i386) which is the one of this package itself and not the
possible other instances.
Shouldn't it be better to have: 'which is different from other
instances of package libllvm19'?
Regards,
Patrice
--- End Message ---
--- Begin Message ---
Sorry then to create such a confusion.
My mind is probably too formatted by another classic message to me:
"trying to overwrite '%.250s', "
"which is also in package %.250s %.250s"
(src/main/archive.c:927-928)
and where the current installing package is not part of
but the one installed is.
This confuses me when I read the message for this other case of "shared"
file conflict where the cited package is the one currently to be installed and
not the one(s) installed in conflict.
And I agree that listing these other instances may not be a good idea if the
current system has a lot of them (a multi multi arch system! :-)).
I am curious to use dpkg-query -S to check the result and how such a full list
can be obtained and also how to get all the shared files currently on my system.
I am closing this report to not disturb any more on that.
Many many thanks for all.
Best,
Paptrice
Le mar. 24 sept. 2024 à 14:55, Guillem Jover <[email protected]> a écrit :
>
> Hi!
>
> On Tue, 2024-09-24 at 09:47:55 +0200, Patrice Duroux wrote:
> > Also, the installed package providing the current file that is
> > conflicting might not be related
> > to that package (multiarch or not). I mean that it could be another
> > unrelated package.
>
> No, this message is always only concerned about multi-arch instances,
> of the same package. There is no support for ref-counting for anything
> that is not a Multi-Arch:same package.
>
> > I wasn't sure if it was a hint message from dpkg that determined the
> > current proprietary
> > package of the file (and so it should be another arch) or if it was a
> > generic message for
> > all cases.
>
> The message is only about a conflict that occurs when installing a
> multi-arch instance when other instances of the same package are
> already unpacked, and the new one does not match the digest for the
> other ones.
>
> > So maybe if dpkg can't determine at that time which already installed
> > package is conflicting
> > for the file to provide better guidance, the message should be very
> > simple and avoid any
> > additional sentences. No?
>
> See above, the current instance conflicts with all existing ones (at
> the same version). It could also print instead all the existing ones,
> but that would make the message longer, and the information for the
> current instance is readily available, so I'm not sure what would be
> the gain with a reversal of the affected packages?
>
> (I'm now wondering whether I really understand the source of the
> confusion for the error message. :)
>
> Thanks,
> Guillem
--- End Message ---
