Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Bug#1032708: please drop transitional package libtinfo-dev from src:ncurses
12 views
Skip to first unread message
Holger Levsen
Mar 11, 2023, 3:50:05 AM3/11/23
to
Package: libtinfo-dev
Version: 6.4-2
Severity: normal
user: qa.deb...@packages.debian.org
usertags: transitional
Please drop the transitional package libtinfo-dev (from the source package
ncurses) after the release of bookworm, it has been released with buster and
bullseye already...
Description: transitional package for libncurses-dev
Package: libtinfo-dev
Version: 6.1+20181013-2+deb10u2
Version: 6.2+20201114-2
Version: 6.4-2
Thanks for maintaining ncurses!
--
cheers,
Holger
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ holger@(debian|reproducible-builds|layer-acht).org
⢿⡄⠘⠷⠚⠋⠀ OpenPGP: B8BF54137B09D35CF026FE9D 091AB856069AAA1C
⠈⠳⣄
In Germany we don‘t say „Happy Valentine‘s Day, I love you“, we say „ich werde
diesen vom Markt kreierten, konsumorientierten Trend des Kapitalismus nicht
unterstützen,“ and I think that’s beautiful. (Hazel Brugger)
Version: 6.4-2
Severity: normal
user: qa.deb...@packages.debian.org
usertags: transitional
Please drop the transitional package libtinfo-dev (from the source package
ncurses) after the release of bookworm, it has been released with buster and
bullseye already...
Description: transitional package for libncurses-dev
Package: libtinfo-dev
Version: 6.1+20181013-2+deb10u2
Version: 6.2+20201114-2
Version: 6.4-2
Thanks for maintaining ncurses!
--
cheers,
Holger
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ holger@(debian|reproducible-builds|layer-acht).org
⢿⡄⠘⠷⠚⠋⠀ OpenPGP: B8BF54137B09D35CF026FE9D 091AB856069AAA1C
⠈⠳⣄
In Germany we don‘t say „Happy Valentine‘s Day, I love you“, we say „ich werde
diesen vom Markt kreierten, konsumorientierten Trend des Kapitalismus nicht
unterstützen,“ and I think that’s beautiful. (Hazel Brugger)
Sven Joachim
Mar 11, 2023, 5:20:04 AM3/11/23
to
Control: tags -1 + moreinfo
On 2023-03-11 09:40 +0100, Holger Levsen wrote:
> Package: libtinfo-dev
> Version: 6.4-2
> Severity: normal
> user: qa.deb...@packages.debian.org
> usertags: transitional
>
> Please drop the transitional package libtinfo-dev (from the source package
> ncurses) after the release of bookworm, it has been released with buster and
> bullseye already...
I know, but some packages are still build-depending on it. If I am
informed correctly, sbuild prefers real packages over virtual ones when
resolving build dependencies, and if libtinfo-dev were to be dropped,
the old version remains as cruft in the archive and is uninstallable.
Note that libtinfo-dev has a strict versioned dependency on libtinfo6
because of the /usr/share/doc symlink.
If that analysis is correct, packages build-depending on libtinfo-dev
will enter status BD-Uninstallable, which is bad. Would you like to
file bugs on those packages and add them as blockers for this bug?
Thanks for taking care of crufty transitional packages.
Cheers,
Sven
On 2023-03-11 09:40 +0100, Holger Levsen wrote:
> Package: libtinfo-dev
> Version: 6.4-2
> Severity: normal
> user: qa.deb...@packages.debian.org
> usertags: transitional
>
> Please drop the transitional package libtinfo-dev (from the source package
> ncurses) after the release of bookworm, it has been released with buster and
> bullseye already...
informed correctly, sbuild prefers real packages over virtual ones when
resolving build dependencies, and if libtinfo-dev were to be dropped,
the old version remains as cruft in the archive and is uninstallable.
Note that libtinfo-dev has a strict versioned dependency on libtinfo6
because of the /usr/share/doc symlink.
If that analysis is correct, packages build-depending on libtinfo-dev
will enter status BD-Uninstallable, which is bad. Would you like to
file bugs on those packages and add them as blockers for this bug?
Thanks for taking care of crufty transitional packages.
Cheers,
Sven
Sven Joachim
Mar 12, 2023, 10:40:03 AM3/12/23
to
On 2023-03-11 11:11 +0100, Sven Joachim wrote:
> Control: tags -1 + moreinfo
>
> On 2023-03-11 09:40 +0100, Holger Levsen wrote:
>
>> Package: libtinfo-dev
>> Version: 6.4-2
>> Severity: normal
>> user: qa.deb...@packages.debian.org
>> usertags: transitional
>>
>> Please drop the transitional package libtinfo-dev (from the source package
>> ncurses) after the release of bookworm, it has been released with buster and
>> bullseye already...
>
> I know, but some packages are still build-depending on it. If I am
> informed correctly, sbuild prefers real packages over virtual ones when
> resolving build dependencies, and if libtinfo-dev were to be dropped,
> the old version remains as cruft in the archive and is uninstallable.
> Note that libtinfo-dev has a strict versioned dependency on libtinfo6
> because of the /usr/share/doc symlink.
>
> If that analysis is correct, packages build-depending on libtinfo-dev
> will enter status BD-Uninstallable, which is bad.
After taking a closer look, this seems to be all red herring. Sbuild
> Control: tags -1 + moreinfo
>
> On 2023-03-11 09:40 +0100, Holger Levsen wrote:
>
>> Package: libtinfo-dev
>> Version: 6.4-2
>> Severity: normal
>> user: qa.deb...@packages.debian.org
>> usertags: transitional
>>
>> Please drop the transitional package libtinfo-dev (from the source package
>> ncurses) after the release of bookworm, it has been released with buster and
>> bullseye already...
>
> I know, but some packages are still build-depending on it. If I am
> informed correctly, sbuild prefers real packages over virtual ones when
> resolving build dependencies, and if libtinfo-dev were to be dropped,
> the old version remains as cruft in the archive and is uninstallable.
> Note that libtinfo-dev has a strict versioned dependency on libtinfo6
> because of the /usr/share/doc symlink.
>
> If that analysis is correct, packages build-depending on libtinfo-dev
> will enter status BD-Uninstallable, which is bad.
uses apt for resolving the build dependencies, and while apt prefers
virtual packages over real ones, it has no problem to use the virtual
one (libncurses-dev in this case) in case the real one is uninstallable
or insufficient.
I guess I'll drop the transitional packages in an upload to experimental
and see what happens with the pseudo-excuses[1].
Filing bugs against reverse (build-)dependencies would be doable for
libtinfo-dev, but for libncursesw5-dev (#1032740) and libncurses5-dev
(for which you did not file a bug, for whatever reason) this is a
non-starter due to the large number of them.
Cheers,
Sven
1. https://release.debian.org/britney/pseudo-excuses-experimental.html
Holger Levsen
Mar 12, 2023, 10:52:20 AM3/12/23
to
On Sun, Mar 12, 2023 at 03:36:33PM +0100, Sven Joachim wrote:
> After taking a closer look, this seems to be all red herring. Sbuild
> uses apt for resolving the build dependencies, and while apt prefers
> virtual packages over real ones, it has no problem to use the virtual
> one (libncurses-dev in this case) in case the real one is uninstallable
> or insufficient.
>
> I guess I'll drop the transitional packages in an upload to experimental
> and see what happens with the pseudo-excuses[1].
cool!
> After taking a closer look, this seems to be all red herring. Sbuild
> uses apt for resolving the build dependencies, and while apt prefers
> virtual packages over real ones, it has no problem to use the virtual
> one (libncurses-dev in this case) in case the real one is uninstallable
> or insufficient.
>
> I guess I'll drop the transitional packages in an upload to experimental
> and see what happens with the pseudo-excuses[1].
> Filing bugs against reverse (build-)dependencies would be doable for
> libtinfo-dev, but for libncursesw5-dev (#1032740) and libncurses5-dev
> (for which you did not file a bug, for whatever reason) this is a
> non-starter due to the large number of them.
--
cheers,
Holger
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ holger@(debian|reproducible-builds|layer-acht).org
⢿⡄⠘⠷⠚⠋⠀ OpenPGP: B8BF54137B09D35CF026FE9D 091AB856069AAA1C
⠈⠳⣄
Sven Joachim
Jul 13, 2023, 12:50:06 PM7/13/23
to
Control: tags -1 - moreinfo
On 2023-03-12 14:44 +0000, Holger Levsen wrote:
testing. :-)
>> Filing bugs against reverse (build-)dependencies would be doable for
>> libtinfo-dev, but for libncursesw5-dev (#1032740) and libncurses5-dev
>> (for which you did not file a bug, for whatever reason) this is a
>
> I did: #1032741
Err, of course. Somehow I managed to overlook that bug.
>> non-starter due to the large number of them.
>
> bug filing can be scripted ;)
Surely, but filing 300+ bugs to get rid of three empty packages is still
a very bad ratio and causes a lot of busy work which is not really
necessary. Rather, I filed _one_ bug against ftp.debian.org[1] to remove
the cruft packages from unstable where they currently remain.
FWIW, what I did looks like a good strategy for other transitional
packages with lots of reverse (build-)dependencies, maybe you would like
to inform their maintainers of it.
Cheers,
Sven
1. https://bugs.debian.org/1040983
On 2023-03-12 14:44 +0000, Holger Levsen wrote:
> On Sun, Mar 12, 2023 at 03:36:33PM +0100, Sven Joachim wrote:
>> After taking a closer look, this seems to be all red herring. Sbuild
>> uses apt for resolving the build dependencies, and while apt prefers
>> virtual packages over real ones, it has no problem to use the virtual
>> one (libncurses-dev in this case) in case the real one is uninstallable
>> or insufficient.
>>
>> I guess I'll drop the transitional packages in an upload to experimental
>> and see what happens with the pseudo-excuses[1].
>
> cool!
Everything went well, and the transitional packages are no longer in
>> After taking a closer look, this seems to be all red herring. Sbuild
>> uses apt for resolving the build dependencies, and while apt prefers
>> virtual packages over real ones, it has no problem to use the virtual
>> one (libncurses-dev in this case) in case the real one is uninstallable
>> or insufficient.
>>
>> I guess I'll drop the transitional packages in an upload to experimental
>> and see what happens with the pseudo-excuses[1].
>
> cool!
testing. :-)
>> Filing bugs against reverse (build-)dependencies would be doable for
>> libtinfo-dev, but for libncursesw5-dev (#1032740) and libncurses5-dev
>> (for which you did not file a bug, for whatever reason) this is a
>
> I did: #1032741
>> non-starter due to the large number of them.
>
> bug filing can be scripted ;)
a very bad ratio and causes a lot of busy work which is not really
necessary. Rather, I filed _one_ bug against ftp.debian.org[1] to remove
the cruft packages from unstable where they currently remain.
FWIW, what I did looks like a good strategy for other transitional
packages with lots of reverse (build-)dependencies, maybe you would like
to inform their maintainers of it.
Cheers,
Sven
1. https://bugs.debian.org/1040983
Holger Levsen
Jul 17, 2023, 10:41:24 AM7/17/23
to
On Thu, Jul 13, 2023 at 06:44:21PM +0200, Sven Joachim wrote:
> >> I guess I'll drop the transitional packages in an upload to experimental
> >> and see what happens with the pseudo-excuses[1].
> >> I guess I'll drop the transitional packages in an upload to experimental
> >> and see what happens with the pseudo-excuses[1].
> Everything went well, and the transitional packages are no longer in
> testing. :-)
very nice!
> testing. :-)
> >> non-starter due to the large number of them.
> > bug filing can be scripted ;)
> Surely, but filing 300+ bugs to get rid of three empty packages is still
> a very bad ratio and causes a lot of busy work which is not really
> necessary. Rather, I filed _one_ bug against ftp.debian.org[1] to remove
> the cruft packages from unstable where they currently remain.
> FWIW, what I did looks like a good strategy for other transitional
> packages with lots of reverse (build-)dependencies, maybe you would like
> to inform their maintainers of it.
But for the next transitional bug filings after the trixie release,
maybe together we can draft a paragraph to include in those bugs?
How about:
=====
Please note that you can stage dropping the transitional package in experimental
so that pseudo-excuses will show you which and how many reverse (build)-depends
need to be changed.
https://release.debian.org/britney/pseudo-excuses-experimental.html
=====
?
--
cheers,
Holger
⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ holger@(debian|reproducible-builds|layer-acht).org
⢿⡄⠘⠷⠚⠋⠀ OpenPGP: B8BF54137B09D35CF026FE9D 091AB856069AAA1C
⠈⠳⣄
Man diskutiert mit ihnen nicht, hat die Geschichte gezeigt“...
0 new messages