Re: Where to put *.pth files for custom Python site packages?

[Florent Rougon]

Hi,

Le 17/09/2024, Christian Gelinek  a écrit:

> The Python documentation[0] states that the search path can be extended via
> .pth files which should be added to specific directories.

(...)

> Given that these sub-folders are managed by Debian/`apt`, where should I place
> our custom.pth file pointing to other (network drive) folders we would like to
> include in Python's search paths?

You can put .pth files in directories such as

  /usr/local/lib/python3.12/dist-packages

(substitute the correct Python version). If you use a venv based at
, the following directory would also work *for the venv*:

  /lib/python3.12/site-packages

Alternatively, of course, you can rely on the PYTHONPATH environment
variable.

Regards

-- 
Florent

Re: Where to put *.pth files for custom Python site packages?

[tomas]

On Tue, Sep 17, 2024 at 11:01:00AM +0930, Christian Gelinek wrote:
> We're using a library of our own Python modules & packages.
> 
> The Python documentation[0] states that the search path can be extended via
> .pth files which should be added to specific directories. While the
> documentation specifically mentions "lib/pythonX.Y/site-packages (on Unix
> and macOS)", I found that Debian seems to do this differently:
> 
> $ python -c 'import site ; print(site.getsitepackages())'
> ['/usr/local/lib/python3.11/dist-packages',
> '/usr/lib/python3/dist-packages', '/usr/lib/python3.11/dist-packages']
> 
> I.e. instead of `site-packages`, it uses `dist-packages` sub-folders.
> 
> Given that these sub-folders are managed by Debian/`apt`, where should I
> place our custom.pth file pointing to other (network drive) folders we would
> like to include in Python's search paths?

I think /usr/local is the one for you: the Debian package manager shouldn't
touch it. It's at the front to allow you to override the distro's choices.

This is actually a standard pattern in GNU flavoured Unix-ish environments.

Cheers
-- 
t


signature.asc
Description: PGP signature

Where to put *.pth files for custom Python site packages?

[Christian Gelinek]

We're using a library of our own Python modules & packages.

The Python documentation[0] states that the search path can be extended 
via .pth files which should be added to specific directories. 
While the documentation specifically mentions 
"lib/pythonX.Y/site-packages (on Unix and macOS)", I found that Debian 
seems to do this differently:


$ python -c 'import site ; print(site.getsitepackages())'
['/usr/local/lib/python3.11/dist-packages', 
'/usr/lib/python3/dist-packages', '/usr/lib/python3.11/dist-packages']


I.e. instead of `site-packages`, it uses `dist-packages` sub-folders.

Given that these sub-folders are managed by Debian/`apt`, where should I 
place our custom.pth file pointing to other (network drive) folders we 
would like to include in Python's search paths?


Thanks for your time.

[0]: https://docs.python.org/3/library/site.html

Re: php packages for debian 10 and debian 11

[Andrew M.A. Cater]

On Wed, Aug 07, 2024 at 01:28:45PM +, Naik, Anil Natha wrote:
> 
> Hi Team,
> 

Hi Anik,

> As per the Debian repository I could see below php versions
> 
> Debian 10
> php<https://packages.debian.org/buster/php/php> (2:7.3+69)
> 
> Debian 11
> php (2:7.4+76)
> 

Both PHP 7.*

> We have installed php 8.1.1 version for Debian 11.
> 
> We wanted to know, what is the stable version of php available for Debian 10 
> and Debian 11 in repository.

As above. Debian stable doesn't change after release. Others have suggested
Ondrej Sury's PHP packages - that's definitely the best idea.

> Is it possible to upgrade from 7.3 to 8.3 on Debian 10 or 11?
> 

It is possible, potentially BUT ...

If you have old machines that are Internet connected, you might want to look
at moving them forward a major version or two of Debian itself. Debian
security support is nominally for one year after the release of the 
next version before it gets passed to LTS. On a two year release cycle,
a verion of Debian will be supported for three years before LTS.
https://wiki.debian.org/LTS

Debian 10 is out of security support from Debian.
It is also out of support from Debian LTS but in support from Debian ELTS.
That's a service that is provided from a separate group.
https://wiki.debian.org/LTS/Extended

Debian 11 is about to pass from Debian main security support to  Debian LTS.
There's a final release planned for August 31st. See, for example,
https://endoflife.date for Debian.

Hope this helps. With every good wish, as ever,

Andy Cater
(amaca...@debian.org)

> Regards,
> Anil Naik
> 
> 
> The information contained in this message may be confidential and legally 
> protected under applicable law. The message is intended solely for the 
> addressee(s). If you are not the intended recipient, you are hereby notified 
> that any use, forwarding, dissemination, or reproduction of this message is 
> strictly prohibited and may be unlawful. If you are not the intended 
> recipient, please contact the sender by return e-mail and destroy all copies 
> of the original message.

Re: php packages for debian 10 and debian 11

[Eduardo M KALINOWSKI]

On 07/08/2024 10:28, Naik, Anil Natha wrote:

Hi Team,

As per the Debian repository I could see below php versions

Debian 10

/*php* * (2:7.3+69)*/

Debian 11

*/php (2:7.4+76)/*

//

We have installed php 8.1.1 version for Debian 11.

*We wanted to know, what is the stable version of php available for 
Debian 10 and Debian 11 in repository.*


Is it possible to upgrade from 7.3 to 8.3 on Debian 10 or 11?


This repository has newer php versions installable on older Debian 
releases: https://deb.sury.org/


Just be aware that while it's done by a Debian developer, it's not 
officially part of Debian.


--
I think I'm schizophrenic.  One half of me's paranoid and the other half's
out to get him.

Eduardo M KALINOWSKI
edua...@kalinowski.com.br

Re: php packages for debian 10 and debian 11

[Wesley]

i always run php8 on debian 11.
There is no problem for me.

Thanks 



> 
> Hi Team, 
> 
> As per the Debian repository I could see below php versions 
> 
> Debian 10
> 
> *php https://packages.debian.org/buster/php/php ** (2:7.3+69)***
> 
> Debian 11
> 
> **php (2:7.4+76)**
> 
> We have installed php 8.1.1 version for Debian 11. 
> 
> **We wanted to know, what is the stable version of php available for Debian 
> 10 and Debian 11 in repository.**
> 
> Is it possible to upgrade from 7.3 to 8.3 on Debian 10 or 11?
> 
> Regards, 
> 
> Anil Naik
> 
> 
> 
> The information contained in this message may be confidential and legally 
> protected under applicable law. The message is intended solely for the 
> addressee(s). If you are not the intended recipient, you are hereby notified 
> that any use, forwarding, dissemination, or reproduction of this message is 
> strictly prohibited and may be unlawful. If you are not the intended 
> recipient, please contact the sender by return e-mail and destroy all copies 
> of the original message.
>

php packages for debian 10 and debian 11

[Naik, Anil Natha]

Hi Team,

As per the Debian repository I could see below php versions

Debian 10
php (2:7.3+69)

Debian 11
php (2:7.4+76)

We have installed php 8.1.1 version for Debian 11.

We wanted to know, what is the stable version of php available for Debian 10 
and Debian 11 in repository.
Is it possible to upgrade from 7.3 to 8.3 on Debian 10 or 11?


Regards,
Anil Naik


The information contained in this message may be confidential and legally 
protected under applicable law. The message is intended solely for the 
addressee(s). If you are not the intended recipient, you are hereby notified 
that any use, forwarding, dissemination, or reproduction of this message is 
strictly prohibited and may be unlawful. If you are not the intended recipient, 
please contact the sender by return e-mail and destroy all copies of the 
original message.

Listing packages installed on broken system, was systemd errors

[David Wright]

On Tue 16 Jul 2024 at 21:35:39 (+0100), mick.crane wrote:
> I installed on a fresh disk the nightly build of Trixie and it works a
> treat and it configured the monitor to it's highest resolution using
> the nouveau module thing.
> Unfortunately I broke my previous Trixie installation trying to get
> rid of the nvidia module.
> It still works with "startx" but the resolution is low.
> Apt let me install the 6.9.8 kernel whereas wouldn't dist-upgrade,
> probably as it had done that before.
> I saw on the net to "update-initrmfs -u" so I did that.
> So I think that earlier install has a kernel now free of the nvidia
> but there must be some systemd left referencing stuff I've deleted
> because there is some message at boot about can't find or load a
> module and X wont automatically start.
> I'd quite like to get that earlier install working cleanly to remind
> me of what was installed.

You don't need to be able to run a system merely to list what's
installed on it. Just mount the root filesystem somewhere, like
/mnt, and type:

  $ dpkg-query --admindir=/mnt/var/lib/dpkg -W --showformat 
'${Package}_${Version}\n' | less

Note the single quotes, to avoid the shell wiping out the format string.

Cheers,
David.

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[songbird]

Thomas Schmitt wrote:
...
> Thanks for this description of a real world procedure.
> Now i know at least that i am not the only one who cares about the
> post-upgrade steps in the manual. I already began to think that everybody
> lets the surplus packages rot in the dark.

  i have been running the same general routine for
years.  every morning i boot up, run the updates
for apt (but i use apt-get out of habit) and then
do the updates that make sense.  since i am running
testing i may delay some updates for a period of
time until issues look to be resolved.

  if there are packages that are no longer needed
i will usually leave them in place until after i've
made sure i don't need them any more (here or there
i have had to downgrade to fix an issue so i don't
want to have to recreate a configuration file) before
i purge them from my system, but it may only be a
few days where i don't have the purged.  i also used
to run orphaner, but i've not used that in quite a
long time and it doesn't even appear in the archive
any more.

  after upgrading i reboot.  i want to know something
is broken when i have time to fix it instead of 
finding out when i don't have time.  i do keep a 
bootable stable partition though for emergencies and
comparison.

  once in a while i remove any unneeded downloaded
packages from my apt cache so i'm not wasting too
much space on my system, but that's once a month or
so when things seem to be quiet enough that i'm not
at risk of removing a package i might need to
downgrade.


> Have a nice day :)

  :)


> Thomas


  songbird

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Thomas Schmitt]

Hi,

Florent Rougon wrote:
> AFAIK, these are not wildcards; each star appended to a package name
> indicates that the package is going to be purged

At least it is a good way to catch the attention of the apt operator. :))


> tl;dr: aptitude praise

Thanks for this description of a real world procedure.
Now i know at least that i am not the only one who cares about the
post-upgrade steps in the manual. I already began to think that everybody
lets the surplus packages rot in the dark.


Have a nice day :)

Thomas

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Florent Rougon]

Le 01/06/2024, Florent Rougon  a écrit:

> FWIW, removal of “obsolete or local” packages is easily done
> interactively in aptitude: you go the the corresponding section of the
> main screen, hit Enter, etc. The [ key recursively unfolds a section
> (use ] to fold it back). You ask to purge a package by typing _
> (removing with -, as in the venerable dselect).

Forgot to say: one can perform the operation (remove with -, purge with
_) on a whole section at once by pressing the key on the section title,
where by “section” I mean: a foldable group of packages (e.g.: admin,
kernel, libsdevel, libs, etc., plus actual Debian sections: main,
contrib, nonfree, and presumably also the new non-free-firmware).

So, for instance, a whole bunch of obsolete library pakages can be
marked all at once for purge with a single _ keypress (this doesn't
exempt one from _reading_ the list of packages marked with this
keypress—make sure you didn't overlook something!).

Regards

-- 
Florent

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Florent Rougon]

Le 31/05/2024, "Thomas Schmitt"  a écrit:

> Then it offered me a list with slightly frightening wildcards:
>
>   The following packages will be REMOVED:
> fuse* libreoffice-avmedia-backend-gstreamer* linux-image-4.19.0-17-amd64*
> linux-image-4.19.0-20-amd64* linux-image-4.19.0-9-amd64* python*
> python-twisted-core* wicd-daemon* wicd-gtk*

AFAIK, these are not wildcards; each star appended to a package name
indicates that the package is going to be purged (no config files left
afterwards), as opposed to simply removed (leaving configuration files
in place).

> I will probably run "apt autoremove" after verifying that the few
> worthy local packages are not in the list proposed for autoremoval.

tl;dr: aptitude praise
~~

FWIW, removal of “obsolete or local” packages is easily done
interactively in aptitude: you go the the corresponding section of the
main screen, hit Enter, etc. The [ key recursively unfolds a section
(use ] to fold it back). You ask to purge a package by typing _
(removing with -, as in the venerable dselect).

For actually obsolete packages, doing so will occasionally trigger a
“dependency problem” because another package that depends on it hasn't
been marked as “to be removed or purged” yet. But normally, that other
package is also obsolete, so it *will* be marked shortly after when you
get to it. So basically, once you've gone through all the obsolete
packages marking each one as “to be purged”, having only left intact the
local ones you do want to keep, there should be no dependency problem to
resolve. ⇒ Hit g (for “go”), check the preview, hit g again if it looks
fine, otherwise q.

Note that the preview (of what is going to be done) is shown in a new
tab (yes, these are tabs, you can switch between them with F6), and that
tab gets closed if you cancel the operation with q. Also, you can act
directly in the Preview tab to unmark an operation, etc. And you can
undo with Ctrl-u, including outside the Preview tab.

Generally and especially for this kind of use (removing obsolete
packages that are still installed), I find that the following lines are
a must-have in /etc/apt/apt.conf:

  // Similar to dselect
  Aptitude::UI::Advance-On-Action "true";

(I also like “Aptitude::Auto-Upgrade "true";” but it is irrelevant
here.)

apt and aptitude have different algorithms for handling upgrades, so:
for stable-to-next-stable upgrades, I do stick to what the Release Notes
recommend. In most other situations where there isn't a huge number of
packages to upgrade, I find that aptitude does a great job:
  - interactive control over what is going to be done;
  - visualization of packages marked as auto-installed vs. those not
marked as such (and you can flip this bit using m or M);
  - interactive, regexp-based search (with powerful features if you look
up the syntax in the manual);
  - interactive package list limited by a user-defined filter (this is
Limit Display in the Search menu);
  - interactive inspection of the (deps a package + other control
fields) with Enter; of its reverse deps with the r key (and you can
quicky recurse in order to find why you need to have this pkg
installed);
  - works in a terminal;
  - etc.

There is the occasional crash, fortunately I've never seen one happen
while dpkg was installing, removing, etc., so the crashes I've seen were
all rather harmless (restart aptitude and proceed again). The worst I
had was on sid during the time_t transition: at some point, aptitude
couldn't start without crashing; however, after upgrading a few packages
with 'apt', it all went back into order. :-)

Regards

-- 
Florent

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Thomas Schmitt]

Hi,

just for the archive:

I think i found the source code which emits the "[...]" strings of
apt-list:

  https://sources.debian.org/src/apt/2.9.4/apt-private/private-output.cc/#L292

The possible status strings are:

  [installed,upgradable to: ...]
  [installed,local]
  [installed,auto-removable]
  [installed,automatic]
  [installed]
  [upgradable from: ...]"
  [residual-config]

More seem not to exist.

Their meaning has to be guessed from the names of methods, variables, and
constants in the code. (I believe "pkgCache::Flag::Auto" is a constant and
not an overloaded monster car. With C++ one never knows what's behind a
name.)

   if (P->CurrentVer != 0)
   {
  if (P.CurrentVer() == V)
  {
 if (state.Upgradable() && state.CandidateVer != NULL)
strprintf(StatusStr, _("[installed,upgradable to: %s]"),
 else if (V.Downloadable() == false)
StatusStr = _("[installed,local]");
 else if(V.Automatic() == true && state.Garbage == true)
StatusStr = _("[installed,auto-removable]");
 else if ((state.Flags & pkgCache::Flag::Auto) == pkgCache::Flag::Auto)
StatusStr = _("[installed,automatic]");
 else
StatusStr = _("[installed]");
  }
  else if (state.CandidateVer == V && state.Upgradable())
 strprintf(StatusStr, _("[upgradable from: %s]"),
   InstalledVerStr.c_str());
   }
   else if (V.ParentPkg()->CurrentState == pkgCache::State::ConfigFiles)
  StatusStr = _("[residual-config]");

-
About my post-upgrade activities:

I ran the command that is proposed in
  
https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.html#purge-removed-packages
  apt purge '~c'

It flooded me with package names under the headline

  The following packages were automatically installed and are no longer 
required:
  ...
  Use 'apt autoremove' to remove them.

Then it offered me a list with slightly frightening wildcards:

  The following packages will be REMOVED:
fuse* libreoffice-avmedia-backend-gstreamer* linux-image-4.19.0-17-amd64*
linux-image-4.19.0-20-amd64* linux-image-4.19.0-9-amd64* python*
python-twisted-core* wicd-daemon* wicd-gtk*

After my confirmation it purged the 9 configurations which were reported
by "apt list '~c'"
  Purging configuration files for fuse (2.9.9-5) ...
  ...
  Purging configuration files for python (2.7.16-1) ...
and as last line said
  Processing triggers for dbus (1.14.10-1~deb12u1) ...

Somewhat mistrusting about the removal of "python" i asked apt-file from
where my current /usr/bin/python stems. Answer: python-is-python3 .
(As side result i now wonder how the result of /usr/bin/python3-pasteurize
might taste and how long it stays fresh.)


I will probably run "apt autoremove" after verifying that the few
worthy local packages are not in the list proposed for autoremoval.


Have a nice day :)

Thomas

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Curt]

On 2024-05-30, Thomas Schmitt  wrote:
>
> Nevertheless it would be nice to find documentation about this kind of
> info in the output of "apt list".

I found this from an old post about Synaptic (the apt front-end), in the 
latter's
"help page":

 Obsolete or locally installed - Display only packages that are not (for longer)
 (sic) included in one of the specified repositories.

I guess the terminology is intended to cover all conceivable cases of why a
package isn't found in the repositories, in the event the user grew confused by
one or the other term that wasn't relevant to her situation, although for APT
the two cases are indistinguishable.


>
> Have a nice day :)
>
> Thomas
>
>


-- 

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Thomas Schmitt]

Hi,

i wrote:
> > What kind of programming language can have inspired the developers
> > to define such a syntax ?

Max Nikulin:
> https://blog.jak-linux.org/2019/08/15/apt-patterns/

This points to aptitude. The package description of aptitude says
"mutt-like syntax for matching packages". Indeed
  https://www.sendmail.org/~ca/email/mutt/manual-4.html
has in its second half some lines which resemble apt-patterns.


> As to obsolete vs. local packages, my guess is that apt may label some
> version as obsolete if another version of the same package is still
> available from some repository. Otherwise it is local.

To me it seems that apt-patterns simply calls "obsolete" what apt-list
then marks in its output as "[... local]".
Obviously these terms refer to different reasons why a package is not
found in the official repos. But these reasons seem to be
indistinguishable. So in the end both terms depict the same status.


Have a nice day :)

Thomas

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[David Wright]

On Wed 29 May 2024 at 18:20:25 (+0200), Thomas Schmitt wrote:
> i wonder why none of the electricians on this list has an anecdote to
> share about dealing with "obsolete" packages after upgrade.
> No triumphs, defeats, or global catastrophes ?

Nowadays I install new releases from scratch, helped by the fact that
for years I've always had two systems on each machine, the current and
the previous. (/home is shared.) That tends to limit cruft as well.

> I wrote:
> > > https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.h
> tml#purge-removed-packages
> > > What does "[residual-config]" mean ?

It is odd that the terminology used in the output differs from that
used to provoke it, as in:

  $ apt list '?config-files'
  Listing... Done
  mlocate/oldstable,now 0.26-5 amd64 [residual-config]
  $ 

Both man apt-patterns and aptitude's Search Term Reference ought to
include the bracketed items if there's no intention to unify terms.
Perhaps it's related to the tendency to underdocument the output from
programs.

> The predicate "obsolete" is not the same as "automatically installed".
> I understand that obsolete means having no successor package in the
> upgraded Debian release.

"Obsolete" is an unfortunately loaded word. I think aptitude expresses
it a bit more clearly: "This term matches any installed package which
is not available in any version from any archive. These packages
appear as “Obsolete or Locally Installed” in the visual interface."

  $ apt list '?installed ?obsolete'
  Listing... Done
  xtoolwait/now 1.3-6.2 amd64 [installed,local]
  yt-dlp/now 2024.05.26-1 all [installed,local]
  $ 

The first is from squeeze, the second from trixie (hardly "obsolete"),
both installed with apt-get fullpath (previously I'd have used dpkg -i).

> Is there a way to do a dry run which only tells what would happen if i
> were more courageous ?

Both apt* and dpkg have --no-act --dry-run --simulate to prevent
acting. (apt* has additional synonyms -s --just-print --recon.)
With dpkg, it's safest to place the option first, as it only
protects what follows it in the command line.

Typically you can also not be root to help protect yourself,
as in:

  $ apt-get -s remove libc6 
  NOTE: This is only a simulation!
apt-get needs root privileges for real execution.
Keep also in mind that locking is deactivated,
so don't depend on the relevance to the real current situation!
  Reading package lists...
  Building dependency tree...
  Reading state information...
  The following packages were automatically installed and are no longer 
required:
  [ … ]
  Use 'apt autoremove' to remove them.
  The following packages will be REMOVED:
  [ … ]
  WARNING: The following essential packages will be removed.
  This should NOT be done unless you know exactly what you are doing!
  [ … ]
  0 upgraded, 0 newly installed, 1702 to remove and 0 not upgraded.
  [ … ]

Cheers,
David.

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Max Nikulin]

On 30/05/2024 16:22, Thomas Schmitt wrote:

Max Nikulin wrote:

apt-patterns(7)


Wow. What kind of programming language can have inspired the developers
to define such a syntax ?


https://blog.jak-linux.org/2019/08/15/apt-patterns/

"apt list" has some limitations in comparison to "aptitude search". I 
have no idea what was the source of inspiration for aptitude, but this 
query language appeared perhaps a quarter of century ago.


https://www.debian.org/doc/manuals/aptitude/ch02s04s05.en.html

For me it is hard to compose a query beyond trivial ones. My adventures 
with "apt list":
Re: List packages from non-default repositories. Wed, 4 Oct 2023 
17:26:47 +0700

https://lists.debian.org/msgid-search/ufjel7$l9m$1...@ciao.gmane.io

However I did not post last variant of the query. It is for aptitude.

As to obsolete vs. local packages, my guess is that apt may label some 
version as obsolete if another version of the same package is still 
available from some repository. Otherwise it is local.

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Florent Rougon]

Le 30/05/2024, "Thomas Schmitt"  a écrit:

> So "local" would be just another word for "obsolete" ?

My understanding is that “obsolete” and “local” may mean different
things to the person who installed the packages (“obsolete” would
correspond to the first item of the list at the end of my previous post,
“local” to the second one); however, apt and aptitude can't distinguish
between them: both categories are comprised of “packages that are
installed but not available from the sources scanned during the last
'apt update' run (or 'aptitude update').

I believe someone already wrote something along these lines in this
thread (maybe Max). In aptitude, the packages in question are all
grouped in the category named “Obsolete and Locally Created Packages”,
IMHO because there is no good way to programmatically distinguish
between them.

(A private package could very well be made available and installed from
a private repository; or alternatively, installed with 'dpkg -i' without
ever being put in an apt repository; therefore “has been installed in
the past from an apt repository” is not a good criterion to distinguish
between “obsolete packages” and “local” ones.)

Note: I mentioned private packages to simplify wording, but the
  “Obsolete and Locally Created Packages” category would also
  contain packages that users sometimes download from third-party
  websites[1], installing them with 'dpkg -i' or the 'apt' command
  line tool, without adding any repository to their
  sources.list(.d). All these are “local packages” from my POV.

Regards

[1] Printer drivers in .deb form, libdvdcss stuff, etc. (make sure the
source is trustworthy!)

-- 
Florent

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Thomas Schmitt]

Hi,

i wrote:
> > Next documenation riddle is what the word "local" means in output lines
> > like
> >   linux-image-5.10.0-rc2-ts/now 5.10.0-rc2-ts-37 amd64 [installed,local]

Florent Rougon wrote:
> I don't use this but guess it is as in aptitude, where “obsolete/local
> packages” are packages that are installed (from dpkg's POV) but not
> available from any of the repositories scanned in the last 'apt update'
> run.

So "local" would be just another word for "obsolete" ?

Indeed, the output of

  apt list '?installed !?obsolete' | grep local

shows only some packages with "locale" in their name, but none with
"local" in the []-brackets. On the other hand

  apt list '?installed ?obsolete' | grep -v 'local]$'

shows no packages, i.e. all lines of obsolete packages end by "local]".


Nevertheless it would be nice to find documentation about this kind of
info in the output of "apt list".


Have a nice day :)

Thomas

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Florent Rougon]

Hi Thomas,

Le 30/05/2024, "Thomas Schmitt"  a écrit:

> Next documenation riddle is what the word "local" means in output lines
> like
>
>   linux-image-5.10.0-rc2-ts/now 5.10.0-rc2-ts-37 amd64 [installed,local]

I don't use this but guess it is as in aptitude, where “obsolete/local
packages” are packages that are installed (from dpkg's POV) but not
available from any of the repositories scanned in the last 'apt update'
run.

This happens in particular with:

  - packages that used to be in a repo seen by 'apt update' (often, you
installed said packages at that time), but are not included in your
current apt sources (/etc/apt/sources.list,
/etc/apt/sources.list.d/); this usually happens between Debian
    releases for some packages shipped by Debian;

  - packages that are not in any of the repos seen by 'apt update' and
that you installed from .deb with 'dpkg -i' (I believe the apt
command line tool can also do this); for instance, local packages
you prepared yourself but didn't bother to put in an apt repository.

Regards

-- 
Florent

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Thomas Schmitt]

Hi,

i wrote:
> > But i am not sure whether the commercial package which i have to keep
> > will be preserved with "apt autoremove".
> > Is there a way to do a dry run which only tells what would happen if i
> > were more courageous ?

Mike Kupfer wrote:
> When I use "apt autoremove", I am given a list of proposed removals and
> a prompt about whether I want to proceed.

Good to know that there are safeguards when i finally remove some of the
"obsolete" packages.


I wrote:
> > How could i get a list of only the automatically installed obsolete
> > packages ?
> > (I still did not find any documentation about the '~c' or '~o' with
> > "apt list".)

Max Nikulin wrote:
> apt-patterns(7)

Wow. What kind of programming language can have inspired the developers
to define such a syntax ?
But hey, at least there is logic provided. \o/

So i try

  apt list '?installed ?obsolete ?automatic'

This narrows the list from 220 to 192 packages.
Even better, i don't have to diff the lists but can see the 28 other
obsolete packages by

  apt list '?installed ?obsolete !?automatic'

Among them are "hfsprogs", the self-made kernels, and the commercial
package which i need to keep.


Next documenation riddle is what the word "local" means in output lines
like

  linux-image-5.10.0-rc2-ts/now 5.10.0-rc2-ts-37 amd64 [installed,local]

(I may have missed something in the man pages of dpkg and dpkg-query, but
their occurences of the word "local" do not look like related to the info
from "apt list".)


Have a nice day :)

Thomas

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Max Nikulin]

On 29/05/2024 23:20, Thomas Schmitt wrote:

How could i get a list of only the automatically installed obsolete
packages ?
(I still did not find any documentation about the '~c' or '~o' with
"apt list".)


apt-patterns(7) and dpkg(1). Apt can not distinguish packages installed 
by dpkg directly and by apt but not present in any configured repository 
any more. Both are "obsolete".

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Mike Kupfer]

Thomas Schmitt wrote:

> But i am not sure whether the commercial package which i have to keep
> will be preserved with "apt autoremove".
> Is there a way to do a dry run which only tells what would happen if i
> were more courageous ?

When I use "apt autoremove", I am given a list of proposed removals and
a prompt about whether I want to proceed.

cheers,
mike

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Thomas Schmitt]

Hi,

i wonder why none of the electricians on this list has an anecdote to
share about dealing with "obsolete" packages after upgrade.
No triumphs, defeats, or global catastrophes ?


I wrote:
> > https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.h
tml#purge-removed-packages
> > What does "[residual-config]" mean ?

Marco Moock wrote:
> Packages include system-wide configuration files. If packages are
> removed, this configuration will not be deleted. You need to purge such
> packages to remove it.

So the smaller list of packages can be dealt with what the upgrade
instructions propose:
  apt purge '~c'


There remains the list of 220 "obsolete" packages.

> > https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.html#obsolete

> Packages have dependencies. Those will be marked as automatically
> installed. They can be removed if no other package depends on them.

But several of those packages were surely installed manually by me via
dpkg -i after being made made as descibed in
  
https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s-common-building
The predicate "obsolete" is not the same as "automatically installed".
I understand that obsolete means having no successor package in the
upgraded Debian release.

How could i get a list of only the automatically installed obsolete
packages ?
(I still did not find any documentation about the '~c' or '~o' with
"apt list".)


> Be aware: If you install software beyond apt/dpkg that depends on files
> in installed packages, you need to mark them as manually installed to
> avoid being removed by autoremove.

Google leads me to apt-mark for that purpose.
But i am not sure whether the commercial package which i have to keep
will be preserved with "apt autoremove".
Is there a way to do a dry run which only tells what would happen if i
were more courageous ?


Have a nice day :)

Thomas

Re: moving some packages back to bookworm stable

[Max Nikulin]

On 29/05/2024 00:51, Michael Grant wrote:

The culprits that seemed to be causing the massive dependencies were
libsasl2-2 and libsasl2-modules-db.  Though not libsasl2-modules which
i also have installed.


With adjusted priorities these packages are not an issue for "apt upgrade".

More serious concern may be whether data written by newer versions are 
readable by older applications or whether some command in package 
scripts can fail due to downgrade of libc or some other library. That is 
why I consider downgrade as a last resort option. Trying it you should 
be prepared to a severe failure and clean reinstall.


Actually the config I posted is incomplete. There is a pitfall similar 
to APT::Default-Release (and -t option). Besides bookworm (or stable) it 
is necessary to explicitly specify bookworm-security and 
bookworm-updates (stable-security, stable-updates). Anyway it can be 
easily spotted in


apt policy

output that has to be run to check effect of changed config files.

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[The Wanderer]

On 2024-05-28 at 15:02, Marco Moock wrote:

> Am 28.05.2024 um 20:38:46 Uhr schrieb Thomas Schmitt:

>> What does "[residual-config]" mean ?
> 
> Packages include system-wide configuration files. If packages are 
> removed, this configuration will not be deleted. You need to purge
> such packages to remove it.

On brief analysis, it looks like this reflects the same status which is
reported in the Status line of the output of 'dpkg -s PACKAGENAME' with
the string "deinstall ok config-files". YMMV, but I find that intuitive
enough: the package used to be installed, and isn't anymore, but its
config files have been left behind.

(Note that if you have the package installed from multiple
architectures, you will apparently need to specify the arch qualifier
suffix to the package name in order for the command to not error out.)

I don't remember using that dpkg command very often, but I do remember
seeing that string often enough in the past, so there are probably other
commands which will also report it if applicable.

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw



signature.asc
Description: OpenPGP digital signature

Re: After upgrade, what do you do about "removed" and "obsolete" packages ?

[Marco Moock]

Am 28.05.2024 um 20:38:46 Uhr schrieb Thomas Schmitt:

> today i upgraded a Debian 11 system to 12 and am now scratching my
> head over the final steps as described in
>   
> https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.html#purge-removed-packages
>   
> https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.html#obsolete

Packages have dependencies. Those will be marked as automatically
installed. They can be removed if no other package depends on them.

You can do that with the autopurge/autoremove apt command.

Be aware: If you install software beyond apt/dpkg that depends on files
in installed packages, you need to mark them as manually installed to
avoid being removed by autoremove. dpkg doesn't care about stuff
manually installed.

> What does "[residual-config]" mean ?

Packages include system-wide configuration files. If packages are
removed, this configuration will not be deleted. You need to purge such
packages to remove it.


-- 
Gruß
Marco

Send unsolicited bulk mail to 1716921526mu...@cartoonies.org

After upgrade, what do you do about "removed" and "obsolete" packages ?

[Thomas Schmitt]

Hi,

today i upgraded a Debian 11 system to 12 and am now scratching my head
over the final steps as described in
  
https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.html#purge-removed-packages
  
https://www.debian.org/releases/stable/amd64/release-notes/ch-upgrading.en.html#obsolete

The command
  apt list '~c'
shows 9 "removed packages":
  fuse/stable 2.9.9-6+b1 amd64 [residual-config]
  libreoffice-avmedia-backend-gstreamer/now 1:7.0.4-4+deb11u9 amd64 
[residual-config]
  linux-image-4.19.0-17-amd64/now 4.19.194-3 amd64 [residual-config]
  linux-image-4.19.0-20-amd64/now 4.19.235-1 amd64 [residual-config]
  linux-image-4.19.0-9-amd64/now 4.19.118-2+deb10u1 amd64 [residual-config]
  python-twisted-core/now 18.9.0-3 all [residual-config]
  python/now 2.7.16-1 amd64 [residual-config]
  wicd-daemon/now 1.7.4+tb2-6 all [residual-config]
  wicd-gtk/now 1.7.4+tb2-6 all [residual-config]
What does "[residual-config]" mean ?
The man page of apt is quite sparse. Is there something more detailed
available online ?

  apt list '~o'
shows 220 "obsolete packages", of which at least one is a commercial
non-Debian package which is on the machine for commercial reasons.
Others are gcc-{8,9,10}, hfsprogs, linux-image-* from Debian 10 and from
my own kernel experiments, and lots of stuff of which i have no clue.

I wonder how others sift through such a list and decide what to do.


Have a nice day :)

Thomas

Re: moving some packages back to bookworm stable

[Michael Grant]

Max, your list looks very similiar to what I'm seeing.

I seem to have suceeded in removing all of the testing packages from
my backup instance, now, just need to flip the ips around and see if
the ship still floats.

The culprits that seemed to be causing the massive dependencies were
libsasl2-2 and libsasl2-modules-db.  Though not libsasl2-modules which
i also have installed.

Using apt to try and remove the first 2 were causing this:

The following packages will be REMOVED:
  apache2 apache2-bin clamav clamav-daemon clamav-freshclam clamav-milter 
clamav-unofficial-sigs clamdscan colord curl dirmngr git gnupg gnupg2 
gpg-wks-client
  libapache2-mod-php8.2 libapache2-mod-ruid2 libaprutil1-ldap libclamav11 
libcurl3-gnutls libcurl4 libgphoto2-6 libldap-2.5-0 libmailutils9 
libmemcached11 libpq5 libsane1
  libsasl2-2 mailutils mongo-tools opendkim opendkim-tools python-apt 
python3-certbot-apache python3-debianbts python3-pycurl python3-pysimplesoap 
python3-reportbug reportbug
  sane-utils sendmail sendmail-bin sensible-mda

I sucked down those 3 packages and downgraded them via 'dpkg -i' and
was able to uninstall and reinstall sendmail by apt and now, no more
packages from testing.

Whew, I won't do that again.  But it's good to know how these things work!

Thanks all for your help.


signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[Max Nikulin]

On 29/05/2024 00:00, Michael Grant wrote:


4) dpkg -i libc6_whatever.deb libwhomever.deb 

5) Repeat until it works.

Apt is NOT built for downgrading.


Agree.


Ah I see, I did not realise that's what you meant by downgrading it,
thanks.


The thread is becoming excessively long. Have you posted output of "apt 
policy" to figure out what is your current configuration? You must have 
either additional config or pinning besides sources list for testing.


Out of curiosity I have tried to install sendmail and openssh-server 
from testing. I removed testing from sources list, dropped the following 
file


/etc/apt/preferences.d/80-downgrade.pref
Package: *
Pin: release n=bookworm
Pin-Priority: 1001

and after

 apt upgrade

only a couple of packages from testing survived:

libdb5.3t64/now 5.3.28+dfsg2-7 amd64 [installed,local]
libssl3t64/now 3.2.1-3 amd64 [installed,local]

I have no idea if all sendmail state and spool files have compatible 
format. So arbitrary failures may be expected.


Before downgrade it was:

libc-bin/testing,now 2.38-11 amd64 [installed]
libc-l10n/testing,now 2.38-11 all [installed,automatic]
libc6/testing,now 2.38-11 amd64 [installed]
libdb5.3t64/testing,now 5.3.28+dfsg2-7 amd64 [installed,automatic]
libsasl2-2/testing,now 2.1.28+dfsg1-6 amd64 [installed,automatic]
libsasl2-modules-db/testing,now 2.1.28+dfsg1-6 amd64 [installed,automatic]
libsasl2-modules/testing,now 2.1.28+dfsg1-6 amd64 [installed,automatic]
libssl3t64/testing,now 3.2.1-3 amd64 [installed,automatic]
libzstd1/testing,now 1.5.5+dfsg2-2 amd64 [installed]
locales/testing,now 2.38-11 all [installed]
openssh-client/testing,now 1:9.7p1-5 amd64 [installed]
openssh-server/testing,now 1:9.7p1-5 amd64 [installed]
openssh-sftp-server/testing,now 1:9.7p1-5 amd64 [installed,automatic]
openssl/testing,now 3.2.1-3 amd64 [installed,automatic]
sendmail-base/testing,now 8.18.1-3 all [installed,automatic]
sendmail-bin/testing,now 8.18.1-3 amd64 [installed,automatic]
sendmail-cf/testing,now 8.18.1-3 all [installed,automatic]
sendmail/testing,now 8.18.1-3 all [installed]
sensible-mda/testing,now 8.18.1-3 amd64 [installed,automatic]

Re: moving some packages back to bookworm stable

[Greg Wooledge]

On Tue, May 28, 2024 at 01:00:24PM -0400, Michael Grant wrote:
> So once I've done this dpkg -i to install a package, I can do that
> without removing the old one first?

Yes, dpkg will upgrade or downgrade the existing package.

> And, once I've hammered a package into place with dpkg, in the future,
> will apt take it into account as a dependency of things already
> installed even though apt itself didn't install or rather downgrade
> the package itself?

Yes.  Dpkg is the lower level tool.  Apt is the higher level tool.
The set of installed packages is tracked by dpkg (it's in the file
/var/lib/dpkg/status which you can read if you like).  Apt calls dpkg
to do all of the installing, removing, etc.  Apt just calculates the
dependency tree and downloads the .deb files for dpkg to use.

Re: moving some packages back to bookworm stable

[Michael Grant]

> > # apt remove -s libc6
> 
> DO NOT do this.
> 
> Downgrade it.  DO NOT remove it and then hope to reinstall it later.
> Removing libc6 will break everything.
> 
> You seem to be flailing, so let me spell this out as explicitly as
> possible.  When I say "downgrade a library package", I mean:
> 
> 1) Download the .deb file for the bookworm(-security) version of the
>library package.
> 
> 2) Run "dpkg -i libc6_whatever.deb".
> 
> 3) When you inevitably get dependency conflicts, download the additional
>library packages that need to be downgraded at the same time, and add
>them to the list.
> 
> 4) dpkg -i libc6_whatever.deb libwhomever.deb 
> 
> 5) Repeat until it works.
> 
> 6) Helpful post-mess cleanup commands include "dpkg --configure -a" and
>"apt-get -f install".  (Yes, that last one has install with no package
>names.)
> 
> Apt is NOT built for downgrading.  If you happen to get any positive
> results from an apt command that involves downgrading, you can consider
> that a pleasant surprise.  Usually you need to invoke dpkg directly.

Ah I see, I did not realise that's what you meant by downgrading it,
thanks.

So once I've done this dpkg -i to install a package, I can do that
without removing the old one first?

And, once I've hammered a package into place with dpkg, in the future,
will apt take it into account as a dependency of things already
installed even though apt itself didn't install or rather downgrade
the package itself?  The fact that I am dpkg installing it over a
package that apt itself installed, perhaps this keeps apt happy?

Thanks for your help.  I use apt all the time to do upgrades but
rarely do I ever need to get into weeds with it.  It's a bit of a
black box to me.




signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[Greg Wooledge]

On Tue, May 28, 2024 at 09:12:18AM -0400, Michael Grant wrote:
> > You will most likely need to remove the testing versions of these packages
> > (apache2, git and so on) and then install the bookworm versions afterward.
> 
> Those dependent packages (most if not all) are not from testing.
> apache2, perl, they are all installed from bookworm or
> bookworm-security.

I am skeptical of this.

> That db5.3 from testing is uninstalled and reinstalling from stable is
> causing these other packages from stable to be uninstalled.  I find
> that confusing.

Let's say that you're right.  You actually *did* download and downgrade
these packages (and just didn't show us the details for some reason), but
the packaging system is unhappy about what you're telling it to do,
and it gets overly aggressive and wants to remove some things that you
feel could remain in place.

You could just *let it remove them*, and then reinstall them.  If you've
already downloaded and downgraded them to bookworm versions, then you
probably still have the .deb files, so it wouldn't even require another
download.

> But what about libc6?  That one really worries me.

As it should!

> # apt remove -s libc6

DO NOT do this.

Downgrade it.  DO NOT remove it and then hope to reinstall it later.
Removing libc6 will break everything.

You seem to be flailing, so let me spell this out as explicitly as
possible.  When I say "downgrade a library package", I mean:

1) Download the .deb file for the bookworm(-security) version of the
   library package.

2) Run "dpkg -i libc6_whatever.deb".

3) When you inevitably get dependency conflicts, download the additional
   library packages that need to be downgraded at the same time, and add
   them to the list.

4) dpkg -i libc6_whatever.deb libwhomever.deb 

5) Repeat until it works.

6) Helpful post-mess cleanup commands include "dpkg --configure -a" and
   "apt-get -f install".  (Yes, that last one has install with no package
   names.)

Apt is NOT built for downgrading.  If you happen to get any positive
results from an apt command that involves downgrading, you can consider
that a pleasant surprise.  Usually you need to invoke dpkg directly.

Re: moving some packages back to bookworm stable

[Michael Grant]

> So, which part are you confused about?  Did you think there was some
> easy way to FIX a frankendebian?  Are you confused because you keep
> thinking "there must be some single apt command that will do all the
> work for me"?
> 
> There's not.  You get to do all the work by hand.

I am trying to do it by hand.  There's not many packages to deal
with at this point, doing this by hand looks like 10 or so packages.

> You will most likely need to remove the testing versions of these packages
> (apache2, git and so on) and then install the bookworm versions afterward.

Those dependent packages (most if not all) are not from testing.
apache2, perl, they are all installed from bookworm or
bookworm-security.

That db5.3 from testing is uninstalled and reinstalling from stable is
causing these other packages from stable to be uninstalled.  I find
that confusing.

But what about libc6?  That one really worries me.

# apt remove -s libc6
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
a few pages of dependicies...

> The things to watch out for are config files (hence your backup), and
> any crazy dependency situations.  In the ideal case, you'll simply be
> able to remove all the packages that aren't libs, then downgrade the
> libs, then reinstall the packages.  And make sure you have sensible
> config files.  If you get stuck, there's always the big hammer
> (dpkg --force-depends and so on).
> 
> If/when it breaks, you get to reinstall from scratch.

I have a running second week old version of the same vm.  I'm rapidly
moving to abandon this and just swapping the instances around.

> This is why we tell people DO NOT MIX BINARY PACKAGES FROM MULTIPLE
> RELEASES.

Yup.  But this whole experience does make me wonder if there are
situations where it is safe.  For instance, if the thing you're
installing from a different release does not cause an update anything
from the current release to a new release.  It feels like apt might be
able to suss that out and if so, pop an "Are you sure??? (y/N)" in the
terminal.


signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[fxkl47BF]

On Tue, 28 May 2024, Michael Grant wrote:

> On Mon, May 27, 2024 at 12:59:34PM -0500, David Wright wrote:
>> So what did it say after that?
>
> Sorry, here's the entire output of one of the tries:
>
> [bottom /etc/mail #1168] apt install libdb5.3/bookworm db5.3-util/bookworm 
> db-util/bookworm
> Reading package lists... Done
> Building dependency tree... Done
> Reading state information... Done
> Selected version '5.3.28+dfsg2-1' (Debian:12.5/stable [amd64]) for 'libdb5.3'
> Selected version '5.3.28+dfsg2-1' (Debian:12.5/stable [amd64]) for 
> 'db5.3-util'
> Selected version '5.3.2' (Debian:12.5/stable [all]) for 'db-util'
> The following packages were automatically installed and are no longer 
> required:
>  acl apache2-data apache2-utils augeas-lenses avahi-daemon clamav-base 
> colord-data git-man gnupg-l10n gnupg-utils gpg-wks-server guile-3.0-libs 
> ipp-usb libapr1 libaprutil1
>  libaprutil1-dbd-sqlite3 libaugeas0 libavahi-core7 libcolorhug2 libdaemon0 
> libexif12 libgphoto2-l10n libgphoto2-port12 libgudev-1.0-0 libgusb2 
> libhashkit2 libieee1284-3 libldap-common
>  liblua5.3-0 libnspr4 libnss-mdns libnss3 libopendbx1 libopendbx1-sqlite3 
> libopendkim11 libpoppler-glib8 libpoppler126 libpython2-stdlib libpython3.11 
> librbl1 librtmp1 libsane-common
>  libsnmp-base libsnmp40 libssh2-1 libvbr2 mailutils-common python2 
> python2-minimal python3-augeas sane-airscan update-inetd usb.ids
> Use 'apt autoremove' to remove them.
> The following additional packages will be installed:
>  php8.2-fpm
> Suggested packages:
>  php-pear
> The following packages will be REMOVED:
>  apache2 apache2-bin clamav clamav-daemon clamav-freshclam clamav-milter 
> clamav-unofficial-sigs clamdscan colord curl dirmngr git gnupg gnupg2 
> gpg-wks-client libapache2-mod-php8.2
>  libapache2-mod-ruid2 libaprutil1-ldap libclamav11 libcurl3-gnutls libcurl4 
> libdb5.3t64 libgphoto2-6 libldap-2.5-0 libmailutils9 libmemcached11 libpq5 
> libsane1 libsasl2-2
>  libsasl2-modules-db mailutils mongo-tools opendkim opendkim-tools python-apt 
> python3-certbot-apache python3-debianbts python3-pycurl python3-pysimplesoap 
> python3-reportbug reportbug
>  sane-utils sasl2-bin sendmail sendmail-bin sensible-mda
> The following NEW packages will be installed:
>  libdb5.3 php8.2-fpm
> The following packages will be DOWNGRADED:
>  db-util db5.3-util
> 0 upgraded, 2 newly installed, 2 downgraded, 46 to remove and 0 not upgraded.
> Need to get 1,743 kB/2,507 kB of archives.
> After this operation, 234 MB disk space will be freed.
> Do you want to continue? [Y/n] n
> Abort.
>
>>> Is there some way to get apt to reinstall a package such that it does
>>> not think it has to uninstall things which depend on it because it's
>>> being immediatly reinstalled?
>>
>> That is the idea behind reinstall, though downgrading is always
>> a test of its ability to succeed.
>
> What it says it's going to do is actually remove those 46 packages and
> not reinstall them.  I believe it!  Clearly apt is unwinding the
> dependencies.  It seems like it's not taking into account the
> downgraded libdb5.3 is a valid dependency for all the things it's
> about to uninstall so it doesn't need to uninstall those things.  I
> thought it should do that, but for some reason, it's not doing that
> for me.
>
>

i ran into exactly this same situation
apt and apt-get wanted to remove a passel of packages
i tried aptitude and it removed only the package i wanted
the system still works fine
just my experience

Re: moving some packages back to bookworm stable

[Dan Ritter]

to...@tuxteam.de wrote: 
> On Mon, May 27, 2024 at 02:02:47PM -0400, Stefan Monnier wrote:
> 
> ISTR that "apt-get install =" will unconditionally
> install  of , if necessary pulling in dependencies.
> 
> But I've never tried it :-)

That pulls in dependencies but does not install packages that
would otherwise be forbidden by the priority system.

E.g.: if you have foobar 1.5 in stable and foobar 2.1 in
backports, and they each depend on libfoobar of the same version
number, then

apt-get install foobar=2.1

will fail saying that it requires libfoobar 2.1 but version 1.5
is to be installed.

You can then solve that by saying

apt-get install foobar=2.1 libfoobar=2.1

but many interesting packages will have a web of dependencies,
and sometimes following them will get you to a place it is hard
to escape.

The backports repository is generally safe (or safe-ish) because
the packages in it are meant to work in a mostly-stable system.

Other repos are less accomodating.

-dsr-

Re: moving some packages back to bookworm stable

[Greg Wooledge]

On Tue, May 28, 2024 at 07:09:16AM -0400, Michael Grant wrote:
> On Tue, May 28, 2024 at 06:59:50AM -0400, Greg Wooledge wrote:
> > On Tue, May 28, 2024 at 06:10:11AM -0400, Michael Grant wrote:
> > > The following packages will be REMOVED:
> > >   [...] libdb5.3t64 [...]
> > 
> > You've *clearly* still got testing packages installed.
> 
> YES.  As I originally said, I created this mess by installing sendmail
> from testing.  And then, a month or so later, I did an
> apt-get upgrade (to do updates, not a full upgrade) which pulled in
> some more things from testing.  I'm trying to get back to all being
> from stable.

So, which part are you confused about?  Did you think there was some
easy way to FIX a frankendebian?  Are you confused because you keep
thinking "there must be some single apt command that will do all the
work for me"?

There's not.  You get to do all the work by hand.

You will most likely need to remove the testing versions of these packages
(apache2, git and so on) and then install the bookworm versions afterward.

The things to watch out for are config files (hence your backup), and
any crazy dependency situations.  In the ideal case, you'll simply be
able to remove all the packages that aren't libs, then downgrade the
libs, then reinstall the packages.  And make sure you have sensible
config files.  If you get stuck, there's always the big hammer
(dpkg --force-depends and so on).

If/when it breaks, you get to reinstall from scratch.

This is why we tell people DO NOT MIX BINARY PACKAGES FROM MULTIPLE
RELEASES.

Re: moving some packages back to bookworm stable

[Michael Grant]

On Tue, May 28, 2024 at 06:59:50AM -0400, Greg Wooledge wrote:
> On Tue, May 28, 2024 at 06:10:11AM -0400, Michael Grant wrote:
> > The following packages will be REMOVED:
> >   [...] libdb5.3t64 [...]
> 
> You've *clearly* still got testing packages installed.

YES.  As I originally said, I created this mess by installing sendmail
from testing.  And then, a month or so later, I did an
apt-get upgrade (to do updates, not a full upgrade) which pulled in
some more things from testing.  I'm trying to get back to all being
from stable.

Unless this is somehow easily fixable, I am leaning towards reverting
to my backup before the apt-get upgrade which has just sendmail from
testing.  I can more easily remove that and reinstall that before
doing the update.

The more I futz with this live machine, the deeper I seem to dig
myself into a hole.


signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[Greg Wooledge]

On Tue, May 28, 2024 at 06:10:11AM -0400, Michael Grant wrote:
> The following packages will be REMOVED:
>   [...] libdb5.3t64 [...]

You've *clearly* still got testing packages installed.

Re: moving some packages back to bookworm stable

[Michael Grant]

On Mon, May 27, 2024 at 12:59:34PM -0500, David Wright wrote:
> So what did it say after that?

Sorry, here's the entire output of one of the tries:

[bottom /etc/mail #1168] apt install libdb5.3/bookworm db5.3-util/bookworm 
db-util/bookworm
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Selected version '5.3.28+dfsg2-1' (Debian:12.5/stable [amd64]) for 'libdb5.3'
Selected version '5.3.28+dfsg2-1' (Debian:12.5/stable [amd64]) for 'db5.3-util'
Selected version '5.3.2' (Debian:12.5/stable [all]) for 'db-util'
The following packages were automatically installed and are no longer required:
  acl apache2-data apache2-utils augeas-lenses avahi-daemon clamav-base 
colord-data git-man gnupg-l10n gnupg-utils gpg-wks-server guile-3.0-libs 
ipp-usb libapr1 libaprutil1
  libaprutil1-dbd-sqlite3 libaugeas0 libavahi-core7 libcolorhug2 libdaemon0 
libexif12 libgphoto2-l10n libgphoto2-port12 libgudev-1.0-0 libgusb2 libhashkit2 
libieee1284-3 libldap-common
  liblua5.3-0 libnspr4 libnss-mdns libnss3 libopendbx1 libopendbx1-sqlite3 
libopendkim11 libpoppler-glib8 libpoppler126 libpython2-stdlib libpython3.11 
librbl1 librtmp1 libsane-common
  libsnmp-base libsnmp40 libssh2-1 libvbr2 mailutils-common python2 
python2-minimal python3-augeas sane-airscan update-inetd usb.ids
Use 'apt autoremove' to remove them.
The following additional packages will be installed:
  php8.2-fpm
Suggested packages:
  php-pear
The following packages will be REMOVED:
  apache2 apache2-bin clamav clamav-daemon clamav-freshclam clamav-milter 
clamav-unofficial-sigs clamdscan colord curl dirmngr git gnupg gnupg2 
gpg-wks-client libapache2-mod-php8.2
  libapache2-mod-ruid2 libaprutil1-ldap libclamav11 libcurl3-gnutls libcurl4 
libdb5.3t64 libgphoto2-6 libldap-2.5-0 libmailutils9 libmemcached11 libpq5 
libsane1 libsasl2-2
  libsasl2-modules-db mailutils mongo-tools opendkim opendkim-tools python-apt 
python3-certbot-apache python3-debianbts python3-pycurl python3-pysimplesoap 
python3-reportbug reportbug
  sane-utils sasl2-bin sendmail sendmail-bin sensible-mda
The following NEW packages will be installed:
  libdb5.3 php8.2-fpm
The following packages will be DOWNGRADED:
  db-util db5.3-util
0 upgraded, 2 newly installed, 2 downgraded, 46 to remove and 0 not upgraded.
Need to get 1,743 kB/2,507 kB of archives.
After this operation, 234 MB disk space will be freed.
Do you want to continue? [Y/n] n
Abort.

> > Is there some way to get apt to reinstall a package such that it does
> > not think it has to uninstall things which depend on it because it's
> > being immediatly reinstalled?
> 
> That is the idea behind reinstall, though downgrading is always
> a test of its ability to succeed.

What it says it's going to do is actually remove those 46 packages and
not reinstall them.  I believe it!  Clearly apt is unwinding the
dependencies.  It seems like it's not taking into account the
downgraded libdb5.3 is a valid dependency for all the things it's
about to uninstall so it doesn't need to uninstall those things.  I
thought it should do that, but for some reason, it's not doing that
for me.



signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[tomas]

On Mon, May 27, 2024 at 02:02:47PM -0400, Stefan Monnier wrote:
> >> > # apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin
> >> 
> >> I can never remember exactly what `-t` really does, but I suspect you'll
> >> need things like
> >> 
> >> apt install libc-bin/bookworm
> >
> > To install a single backported (or other release) package, 
> > apt-get install packagename/releasename
> >
> > and to install a backported package plus dependencies which
> > are also from that specific release, use 
> > apt-get -t releasename packagename
> 
> But that's not the whole story of what `-t` does since the above does
> not explain why his attempt to use `-t` to downgrade some packages
> resulted in `apt` saying " is already the newest version".

ISTR that "apt-get install =" will unconditionally
install  of , if necessary pulling in dependencies.

But I've never tried it :-)

Cheers
-- 
t


signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[Max Nikulin]

On 28/05/2024 01:02, Stefan Monnier wrote:

But that's not the whole story of what `-t` does since the above does
not explain why his attempt to use `-t` to downgrade some packages
resulted in `apt` saying " is already the newest version".


My guess is that -t increases priority of the specified release. However 
the value is not enough to allow downgrades.


Another issue is the libdb5.3t64 explicit argument. This package does 
not exist in bookworm.

Re: moving some packages back to bookworm stable

[David Wright]

On Mon 27 May 2024 at 21:46:24 (+0200), Detlef Vollmann wrote:
> On 5/27/24 20:02, Stefan Monnier wrote:
> > > > > # apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin
> > > > 
> > > > I can never remember exactly what `-t` really does, but I suspect you'll
> > > > need things like
> > > > 
> > > >  apt install libc-bin/bookworm
> > > 
> > > To install a single backported (or other release) package,
> > > apt-get install packagename/releasename
> > > 
> > > and to install a backported package plus dependencies which
> > > are also from that specific release, use
> > > apt-get -t releasename packagename
> > 
> > But that's not the whole story of what `-t` does since the above does
> > not explain why his attempt to use `-t` to downgrade some packages
> > resulted in `apt` saying " is already the newest version".
> 
> Sometimes '-t' works for me, and does what I expect, and sometimes
> it doesn't.

And, of course, what would interest the list is what it says
when it doesn't work.

> So I generelly use now the explicit version:
> 
> apt install libc-bin=2.36-9+deb12u7

Cheers,
David.

Re: moving some packages back to bookworm stable

[Detlef Vollmann]

On 5/27/24 20:02, Stefan Monnier wrote:

# apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin


I can never remember exactly what `-t` really does, but I suspect you'll
need things like

 apt install libc-bin/bookworm


To install a single backported (or other release) package,
apt-get install packagename/releasename

and to install a backported package plus dependencies which
are also from that specific release, use
apt-get -t releasename packagename


But that's not the whole story of what `-t` does since the above does
not explain why his attempt to use `-t` to downgrade some packages
resulted in `apt` saying " is already the newest version".


Sometimes '-t' works for me, and does what I expect, and sometimes
it doesn't.  So I generelly use now the explicit version:

apt install libc-bin=2.36-9+deb12u7

  Detlef

Re: moving some packages back to bookworm stable

[David Wright]

On Mon 27 May 2024 at 14:02:47 (-0400), Stefan Monnier wrote:
> >> > # apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin
> >> 
> >> I can never remember exactly what `-t` really does, but I suspect you'll
> >> need things like
> >> 
> >> apt install libc-bin/bookworm
> >
> > To install a single backported (or other release) package, 
> > apt-get install packagename/releasename
> >
> > and to install a backported package plus dependencies which
> > are also from that specific release, use 
> > apt-get -t releasename packagename
> 
> But that's not the whole story of what `-t` does since the above does
> not explain why his attempt to use `-t` to downgrade some packages
> resulted in `apt` saying " is already the newest version".

Neither syntax will specify a newer version for plain "install"
to install or upgrade.

Cheers,
David.

Re: moving some packages back to bookworm stable

[Stefan Monnier]

>> > # apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin
>> 
>> I can never remember exactly what `-t` really does, but I suspect you'll
>> need things like
>> 
>> apt install libc-bin/bookworm
>
> To install a single backported (or other release) package, 
> apt-get install packagename/releasename
>
> and to install a backported package plus dependencies which
> are also from that specific release, use 
> apt-get -t releasename packagename

But that's not the whole story of what `-t` does since the above does
not explain why his attempt to use `-t` to downgrade some packages
resulted in `apt` saying " is already the newest version".


Stefan

Re: moving some packages back to bookworm stable

[David Wright]

On Mon 27 May 2024 at 12:23:41 (-0400), Michael Grant wrote:
> [ … ]
> so I thought I'd try the same process with db5.3, but removing db5.3
> wants to remove a slew of packages:
> 
> # apt reinstall -s libdb5.3/bookworm
> ...
> Selected version '5.3.28+dfsg2-1' (Debian:12.5/stable [amd64]) for 'libdb5.3'
> The following packages were automatically installed and are no longer 
> required:
>   acl apache2-data apache2-utils augeas-lenses avahi-daemon clamav-base 
> colord-data git-man gnupg-l10n gnupg-utils gpg-wks-server guile-3.0-libs 
> ipp-usb libapr1 libaprutil1
>   libaprutil1-dbd-sqlite3 libaugeas0 libavahi-core7 libcolorhug2 libdaemon0 
> libexif12 libgphoto2-l10n libgphoto2-port12 libgudev-1.0-0 libgusb2 
> libhashkit2 libieee1284-3 libldap-common
>   liblua5.3-0 libnspr4 libnss-mdns libnss3 libopendbx1 libopendbx1-sqlite3 
> libopendkim11 libpoppler-glib8 libpoppler126 libpython2-stdlib libpython3.11 
> librbl1 librtmp1 libsane-common
>   libsnmp-base libsnmp40 libssh2-1 libvbr2 mailutils-common python2 
> python2-minimal python3-augeas sane-airscan update-inetd usb.ids
> Use 'apt autoremove' to remove them

So what did it say after that?

> Is there some way to get apt to reinstall a package such that it does
> not think it has to uninstall things which depend on it because it's
> being immediatly reinstalled?

That is the idea behind reinstall, though downgrading is always
a test of its ability to succeed.

Cheers,
David.

Re: moving some packages back to bookworm stable

[Dan Ritter]

Stefan Monnier wrote: 
> > # apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin
> 
> I can never remember exactly what `-t` really does, but I suspect you'll
> need things like
> 
> apt install libc-bin/bookworm

To install a single backported (or other release) package, 
apt-get install packagename/releasename

and to install a backported package plus dependencies which
are also from that specific release, use 
apt-get -t releasename packagename

-dsr-

Re: moving some packages back to bookworm stable

[Michael Grant]

Hans, thanks for that but I am a bit confused following your
instructions.  Did you mean to I should remove the lines for 'stable'
from sources.list?  Or remove the lines for 'testing'?  I am trying to
get the packages to go back to stable.

I am more familiar with apt than aptitude.

I managed to do part of what Greg recommended.  I removed sendmail and
sasl2-bin and reinstalled them from stable.  That seemed to work fine,
I have fewer testing pkgs installed now:

$ apt-show-versions | g testing
db-util:all/testing 5.3.3 uptodate
db5.3-util:amd64/testing 5.3.28+dfsg2-7 uptodate
libc-bin:amd64/testing 2.38-11 uptodate
libc-dev-bin:amd64/testing 2.38-11 uptodate
libc-devtools:amd64/testing 2.38-11 uptodate
libc-l10n:all/testing 2.38-11 uptodate
libc6:amd64/testing 2.38-11 uptodate
libc6-dev:amd64/testing 2.38-11 uptodate
libdb5.3t64:amd64/testing 5.3.28+dfsg2-7 uptodate
libmilter1.0.1:amd64/testing 8.18.1-3 uptodate
libsasl2-2:amd64/testing 2.1.28+dfsg1-6 uptodate
libsasl2-modules:amd64/testing 2.1.28+dfsg1-6 uptodate
libsasl2-modules-db:amd64/testing 2.1.28+dfsg1-6 uptodate
libssl3t64:amd64/testing 3.2.1-3 uptodate
libzstd1:amd64/testing 1.5.5+dfsg2-2 uptodate
locales:all/testing 2.38-11 uptodate
openssh-client:amd64/testing 1:9.7p1-5 uptodate
openssh-server:amd64/testing 1:9.7p1-5 uptodate
openssh-sftp-server:amd64/testing 1:9.7p1-5 uptodate
openssl:amd64/testing 3.2.1-3 uptodate
zstd:amd64/testing 1.5.5+dfsg2-2 uptodate

so I thought I'd try the same process with db5.3, but removing db5.3
wants to remove a slew of packages:

# apt reinstall -s libdb5.3/bookworm
...
Selected version '5.3.28+dfsg2-1' (Debian:12.5/stable [amd64]) for 'libdb5.3'
The following packages were automatically installed and are no longer required:
  acl apache2-data apache2-utils augeas-lenses avahi-daemon clamav-base 
colord-data git-man gnupg-l10n gnupg-utils gpg-wks-server guile-3.0-libs 
ipp-usb libapr1 libaprutil1
  libaprutil1-dbd-sqlite3 libaugeas0 libavahi-core7 libcolorhug2 libdaemon0 
libexif12 libgphoto2-l10n libgphoto2-port12 libgudev-1.0-0 libgusb2 libhashkit2 
libieee1284-3 libldap-common
  liblua5.3-0 libnspr4 libnss-mdns libnss3 libopendbx1 libopendbx1-sqlite3 
libopendkim11 libpoppler-glib8 libpoppler126 libpython2-stdlib libpython3.11 
librbl1 librtmp1 libsane-common
  libsnmp-base libsnmp40 libssh2-1 libvbr2 mailutils-common python2 
python2-minimal python3-augeas sane-airscan update-inetd usb.ids
Use 'apt autoremove' to remove them

Is there some way to get apt to reinstall a package such that it does
not think it has to uninstall things which depend on it because it's
being immediatly reinstalled?

And for those of you telling me to have a backup, I do.  I have booted
a snapshot from about a week ago.  However, to make that the live one
and dump this one, it's not so easy but possible.  That snapshot has
only sendmail from testing. Hard to know what is more work, going down
this route or making the other instance live.  I'm starting to think
about abandoning this and reconfiguring the backup instance.


signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[Hans]

Doing "apt-get upgrade" will only upgrade all installed packages, but no new 
ones (even, if they are needed).

Better is to do an "apt-get full-upgrade", which will install the whole system 
from stable to testing. However, this might also uninstall some wanted 
packages, thus often it is calles the "intelligent" upgrade. Intelligent does 
not mean, the upgrade is intelligent, but the one doing this upgrade (mostly 
the person, who is root) should be intelligent.

Downgrading is not an easy way, but managable. But it is a lot of intelligent 
work.

How can you do this? This is, how I am figured out (best way for me!)

First, remove the entry from stable off your sources.list.

Then start aptitude and update the list. 

Next manually search all packages you want to downgrade to the needed 
versions. The last apt-get or aptitude log should help.


Mark all installed versions to "remove" (magenta coluur) and needed versions 
to "install" (green colour).

Now, dive manually into all dependencies (these are marked with the red 
coulour)  and do the same as above (mark the installed version first "remove" 
then the correct as "install").

Important: Check that ALL dependencies are correct and no libs or anything 
else is set with red colour.

This process must be done very, very correct! 

After this press "g" (which is for "install now") and if everything was set 
correct, all packages are now downgraded.

Note: If you have missed something, you have to restart again!

>From my experiences this doing so is still faster, than to setup a new system.

Hope this helps.

Good luck!

Best

Hans 


 

Re: moving some packages back to bookworm stable

[Max Nikulin]

On 27/05/2024 21:28, Michael Grant wrote:

What I want to do is get the system back to just using the packages
from stable rather than testing.


I have never tried the following, so it is better to test it in a 
virtual machine or inside a container. I would try to set priority of 
bookworm release high enough to allow downgrade. See apt_preferences(5) 
for details concerning pinning. However downgrade of libc makes "apt 
upgrade" really risky.


Do not forget to remove configuration with excessive priority afterwards.

Re: moving some packages back to bookworm stable

[Stefan Monnier]

> I needed to install a version of sendmail from testing a while back to
> test it.

Downgrading Debian packages is not well supported, by and large.
So installing `testing` packages into a `stable` install is manageable
(tho it itself can bring trouble) but going back to `stable` afterwards
tends to be a lot more complicated.

Transitions like the t64 transition going on right now in `testing` make
it yet more troublesome.

I recommend the use of snapshots when you want to try such a thing with
the intention of "going back" later.

> # apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin

I can never remember exactly what `-t` really does, but I suspect you'll
need things like

apt install libc-bin/bookworm

to state more explicitly what you want.
Maybe you can do something like

apt install $(apt-show-versions | sed -n 's|/testing.*|/stable|p')


- Stefan

Re: moving some packages back to bookworm stable

[David Wright]

On Mon 27 May 2024 at 09:56:54 (-0400), Michael Grant wrote:
> What's the best way to get back to running just the bookworm stable
> packages?  I tried what I thought was the obvious way to fix this by
> running:
> 
> # apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin 
> libc-devtools libc-l10n libc6 libc6-dev libdb5.3t64 libmilter1.0.1 libsasl2-2 
> libsasl2-modules libsasl2-modules-db libssl3t64 libzstd1 locales 
> openssh-client openssh-server openssh-sftp-server openssl sasl2-bin sendmail 
> sendmail-base sendmail-bin sendmail-cf sensible-mda zstd

As Greg wrote: backups come first.

But in the above, you need reinstall, either as a command, or
as an option --reinstall.

Cheers,
David.

Re: moving some packages back to bookworm stable

[Greg Wooledge]

On Mon, May 27, 2024 at 10:28:37AM -0400, Michael Grant wrote:
> On Mon, May 27, 2024 at 10:19:48AM -0400, Greg Wooledge wrote:
> > On Mon, May 27, 2024 at 09:56:54AM -0400, Michael Grant wrote:
> > > I needed to install a version of sendmail from testing a while back to
> > > test it.
> > 
> > Your subject header says "bookworm stable".  You don't install binary
> > packages from testing on a stable system.  You use backports instead.
> 
> ugh no, wait, I may be using the wrong terminology.  I'm not wanting
> to install special packages and definitely don't need to build my own.
> 
> What I want to do is get the system back to just using the packages
> from stable rather than testing.  Only those few packages before
> things get worse in the next update.  There's not many.

Downgrading essential libraries (libc6 and friends) that were brought
in when you tainted your stable system with testing packages is going
to be risky.

Definitely make a backup before you do ANYTHING.

Once that's done, you can try purging the non-essential testing packages,
and then downgrading the essential ones.  If at any point the system
becomes utterly broken, reinstall stable, and then restore your backup.

Re: moving some packages back to bookworm stable

[Michael Grant]

On Mon, May 27, 2024 at 10:19:48AM -0400, Greg Wooledge wrote:
> On Mon, May 27, 2024 at 09:56:54AM -0400, Michael Grant wrote:
> > I needed to install a version of sendmail from testing a while back to
> > test it.
> 
> Your subject header says "bookworm stable".  You don't install binary
> packages from testing on a stable system.  You use backports instead.

ugh no, wait, I may be using the wrong terminology.  I'm not wanting
to install special packages and definitely don't need to build my own.

What I want to do is get the system back to just using the packages
from stable rather than testing.  Only those few packages before
things get worse in the next update.  There's not many.



signature.asc
Description: PGP signature

Re: moving some packages back to bookworm stable

[Greg Wooledge]

On Mon, May 27, 2024 at 09:56:54AM -0400, Michael Grant wrote:
> I needed to install a version of sendmail from testing a while back to
> test it.

Your subject header says "bookworm stable".  You don't install binary
packages from testing on a stable system.  You use backports instead.

https://backports.debian.org/Instructions/

There is already a bookworm backported version of sendmail available:

https://packages.debian.org/search?keywords=sendmail&searchon=names§ion=all&suite=bookworm-backports

If that one isn't new enough, then you may need to build your own backport
package.  This is usually either trivially easy ("type these commands")
or utterly impossible, depending on which dependencies have changed
since bookworm.  There's no in-between.

moving some packages back to bookworm stable

[Michael Grant]

I needed to install a version of sendmail from testing a while back to
test it.  On friday, I ran 'apt upgrade' which looked like it was
going to uninstall and then reinstall the sendmail packages.  I let it
run, when it was done, only some of the sendmail packages had
re-installed.  Basically, I shot myself in the foot with the
dependencies.  In the end, this was not a good idea.

I really do not want to reinstall the entire box, there's a lot of
config that has gone into this over the years.  What I'd like to do is
just get back to stable bookworm packages.

Before the update, these are the packages that were installed from
testing:

$ apt-show-versions | grep testing
libmilter1.0.1:amd64/testing 8.18.1-3 uptodate
libsasl2-2:amd64/testing 2.1.28+dfsg1-4+b1 upgradeable to 2.1.28+dfsg1-6
libsasl2-modules:amd64/testing 2.1.28+dfsg1-4+b1 upgradeable to 2.1.28+dfsg1-6
libsasl2-modules-db:amd64/testing 2.1.28+dfsg1-4+b1 upgradeable to 
2.1.28+dfsg1-6
sasl2-bin:amd64/testing 2.1.28+dfsg1-4+b1 upgradeable to 2.1.28+dfsg1-6
sendmail:all/testing 8.18.1-3 uptodate
sendmail-base:all/testing 8.18.1-1 upgradeable to 8.18.1-3
sendmail-bin:amd64/testing 8.18.1-1 upgradeable to 8.18.1-3
sendmail-cf:all/testing 8.18.1-1 upgradeable to 8.18.1-3
sensible-mda:amd64/testing 8.18.1-3 uptodate

After a day of frantically trying to get things working again, I found
I had some package that depended on some t64 version of some library.
In the end, I think it was sasl2-bin.  I ended up installing sasl2-bin
from testing, then sendmail started working again.  This is what I
ended up with installed from testing:

$ apt-show-versions | grep testing
db-util:all/testing 5.3.3 uptodate
db5.3-util:amd64/testing 5.3.28+dfsg2-7 uptodate
libc-bin:amd64/testing 2.38-11 uptodate
libc-dev-bin:amd64/testing 2.38-11 uptodate
libc-devtools:amd64/testing 2.38-11 uptodate
libc-l10n:all/testing 2.38-11 uptodate
libc6:amd64/testing 2.38-11 uptodate
libc6-dev:amd64/testing 2.38-11 uptodate
libdb5.3t64:amd64/testing 5.3.28+dfsg2-7 uptodate
libmilter1.0.1:amd64/testing 8.18.1-3 uptodate
libsasl2-2:amd64/testing 2.1.28+dfsg1-6 uptodate
libsasl2-modules:amd64/testing 2.1.28+dfsg1-6 uptodate
libsasl2-modules-db:amd64/testing 2.1.28+dfsg1-6 uptodate
libssl3t64:amd64/testing 3.2.1-3 uptodate
libzstd1:amd64/testing 1.5.5+dfsg2-2 uptodate
locales:all/testing 2.38-11 uptodate
openssh-client:amd64/testing 1:9.7p1-5 uptodate
openssh-server:amd64/testing 1:9.7p1-5 uptodate
openssh-sftp-server:amd64/testing 1:9.7p1-5 uptodate
openssl:amd64/testing 3.2.1-3 uptodate
sasl2-bin:amd64/testing 2.1.28+dfsg1-6 uptodate
sendmail:all/testing 8.18.1-3 uptodate
sendmail-base:all/testing 8.18.1-3 uptodate
sendmail-bin:amd64/testing 8.18.1-3 uptodate
sendmail-cf:all/testing 8.18.1-3 uptodate
sensible-mda:amd64/testing 8.18.1-3 uptodate
zstd:amd64/testing 1.5.5+dfsg2-2 uptodate

I did not manually install most of that.  Only sasl2-bin and sendmail
from testing.

What's the best way to get back to running just the bookworm stable
packages?  I tried what I thought was the obvious way to fix this by
running:

# apt install -t=bookworm db-util db5.3-util libc-bin libc-dev-bin 
libc-devtools libc-l10n libc6 libc6-dev libdb5.3t64 libmilter1.0.1 libsasl2-2 
libsasl2-modules libsasl2-modules-db libssl3t64 libzstd1 locales openssh-client 
openssh-server openssh-sftp-server openssl sasl2-bin sendmail sendmail-base 
sendmail-bin sendmail-cf sensible-mda zstd
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
db-util is already the newest version (5.3.3).
db5.3-util is already the newest version (5.3.28+dfsg2-7).
libc-bin is already the newest version (2.38-11).
libc-dev-bin is already the newest version (2.38-11).
libc-devtools is already the newest version (2.38-11).
libc-l10n is already the newest version (2.38-11).
libc6 is already the newest version (2.38-11).
libc6-dev is already the newest version (2.38-11).
libdb5.3t64 is already the newest version (5.3.28+dfsg2-7).
libmilter1.0.1 is already the newest version (8.18.1-3).
libsasl2-2 is already the newest version (2.1.28+dfsg1-6).
libsasl2-modules is already the newest version (2.1.28+dfsg1-6).
libsasl2-modules-db is already the newest version (2.1.28+dfsg1-6).
libssl3t64 is already the newest version (3.2.1-3).
libzstd1 is already the newest version (1.5.5+dfsg2-2).
locales is already the newest version (2.38-11).
openssh-client is already the newest version (1:9.7p1-5).
openssh-server is already the newest version (1:9.7p1-5).
openssh-sftp-server is already the newest version (1:9.7p1-5).
openssl is already the newest version (3.2.1-3).
sasl2-bin is already the newest version (2.1.28+dfsg1-6).
sendmail is already the newest version (8.18.1-3).
sendmail-base is already the newest version (8.18.1-3).
sendmail-bin is already the newest version (8.18.1-3).
sendmail-cf is already the newest version (8.18.1-3).
sensible-mda is already the newest version (8.18.1-3).
z

Re: RE: Problems installing QEMU packages on Debian 12 (stable)

[Lukas Nagy]

Hi,

thanks for checking, in the end I solved this by switching mirrors from 
default http://deb.debian.org/debian to http://ftp.cz.debian.org/debian 
- after updating I got the correct version of QEMU package.


Maybe something was cached somewhere for several days, strange that I 
had to change the mirror.

Problems installing QEMU packages on Debian 12 (stable)

[Lukas Nagy]

Hi,

I am trying to make KVM/QEMU work on my Debian 12. I follow 
https://wiki.debian.org/KVM but I get stuck already on installation, 
because apt-get reports non-existent packages on debian repos.


I ran

sudo apt install qemu-system libvirt-daemon-system virt-manager

It resolves packages, but when fails on 404 on qemu / xen packages

Full output is here:

sudo apt install qemu-system libvirt-daemon-system virt-manager
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
The following packages were automatically installed and are no longer 
required:

  libcapi20-3 libodbc2 libosmesa6 libz-mingw-w64
Use 'sudo apt autoremove' to remove them.
The following additional packages will be installed:
  gir1.2-ayatanaappindicator3-0.1 gir1.2-gtk-vnc-2.0 
gir1.2-libosinfo-1.0 gir1.2-libvirt-glib-1.0 gir1.2-spiceclientglib-2.0 
gir1.2-spiceclientgtk-3.0 gir1.2-vte-2.91 gnutls-bin ibverbs-providers 
ipxe-qemu
  libcacard0 libcapstone4 libdaxctl1 libexecs0 libfdt1 libfmt9 
libgfapi0 libgfrpc0 libgfxdr0 libglusterfs0 libgtk-vnc-2.0-0 
libgvnc-1.0-0 libibverbs1 libiscsi7 libisoburn1 libndctl6 libnss-mymachines
  libphodav-3.0-0 libphodav-3.0-common libpmem1 librados2 librbd1 
librdmacm1 libspice-client-glib-2.0-8 libspice-client-gtk-3.0-5 
libspice-server1 libtpms0 liburing2 libusbredirhost1 libusbredirparser1
  libvdeplug2 libvirglrenderer1 libvirt-clients libvirt-daemon 
libvirt-daemon-config-network libvirt-daemon-config-nwfilter 
libvirt-daemon-driver-lxc libvirt-daemon-driver-qemu 
libvirt-daemon-driver-vbox
  libvirt-daemon-driver-xen libvirt-daemon-system-systemd 
libvirt-glib-1.0-0 libvirt-glib-1.0-data libvirt-l10n libvirt0 
libxencall1 libxendevicemodel1 libxenevtchn1 libxenforeignmemory1 
libxengnttab1
  libxenhypfs1 libxenmisc4.17 libxenstore4 libxentoolcore1 
libxentoollog1 libxml2-utils mdevctl netcat-openbsd ovmf python3-libvirt 
python3-libxml2 qemu-block-extra qemu-efi-aarch64 qemu-efi-arm
  qemu-system-arm qemu-system-common qemu-system-data qemu-system-gui 
qemu-system-mips qemu-system-misc qemu-system-ppc qemu-system-sparc 
qemu-system-x86 qemu-utils seabios spice-client-glib-usb-acl-helper
  swtpm swtpm-libs swtpm-tools systemd-container virt-viewer virtinst 
xorriso

Suggested packages:
  libvirt-clients-qemu libvirt-login-shell 
libvirt-daemon-driver-storage-gluster 
libvirt-daemon-driver-storage-iscsi-direct 
libvirt-daemon-driver-storage-rbd libvirt-daemon-driver-storage-zfs 
numad auditd
  nfs-common open-iscsi pm-utils systemtap zfsutils samba vde2 trousers 
python3-guestfs ssh-askpass xorriso-tcltk jigit cdck

The following NEW packages will be installed:
  gir1.2-ayatanaappindicator3-0.1 gir1.2-gtk-vnc-2.0 
gir1.2-libosinfo-1.0 gir1.2-libvirt-glib-1.0 gir1.2-spiceclientglib-2.0 
gir1.2-spiceclientgtk-3.0 gir1.2-vte-2.91 gnutls-bin ibverbs-providers 
ipxe-qemu
  libcacard0 libcapstone4 libdaxctl1 libexecs0 libfdt1 libfmt9 
libgfapi0 libgfrpc0 libgfxdr0 libglusterfs0 libgtk-vnc-2.0-0 
libgvnc-1.0-0 libibverbs1 libiscsi7 libisoburn1 libndctl6 libnss-mymachines
  libphodav-3.0-0 libphodav-3.0-common libpmem1 librados2 librbd1 
librdmacm1 libspice-client-glib-2.0-8 libspice-client-gtk-3.0-5 
libspice-server1 libtpms0 liburing2 libusbredirhost1 libusbredirparser1
  libvdeplug2 libvirglrenderer1 libvirt-clients libvirt-daemon 
libvirt-daemon-config-network libvirt-daemon-config-nwfilter 
libvirt-daemon-driver-lxc libvirt-daemon-driver-qemu 
libvirt-daemon-driver-vbox
  libvirt-daemon-driver-xen libvirt-daemon-system 
libvirt-daemon-system-systemd libvirt-glib-1.0-0 libvirt-glib-1.0-data 
libvirt-l10n libvirt0 libxencall1 libxendevicemodel1 libxenevtchn1 
libxenforeignmemory1
  libxengnttab1 libxenhypfs1 libxenmisc4.17 libxenstore4 
libxentoolcore1 libxentoollog1 libxml2-utils mdevctl netcat-openbsd ovmf 
python3-libvirt python3-libxml2 qemu-block-extra qemu-efi-aarch64 
qemu-efi-arm
  qemu-system qemu-system-arm qemu-system-common qemu-system-data 
qemu-system-gui qemu-system-mips qemu-system-misc qemu-system-ppc 
qemu-system-sparc qemu-system-x86 qemu-utils seabios
  spice-client-glib-usb-acl-helper swtpm swtpm-libs swtpm-tools 
systemd-container virt-manager virt-viewer virtinst xorriso

0 upgraded, 96 newly installed, 0 to remove and 0 not upgraded.
Need to get 97,7 MB/143 MB of archives.
After this operation, 974 MB of additional disk space will be used.
Do you want to continue? [Y/n] y
Err:1 http://deb.debian.org/debian bookworm/main amd64 gnutls-bin amd64 
3.7.9-2+deb12u1

  404  Not Found [IP: 2a04:4e42:41::644 80]
Err:2 http://deb.debian.org/debian bookworm/main amd64 systemd-container 
amd64 252.19-1~deb12u1

  404  Not Found [IP: 2a04:4e42:41::644 80]
Err:3 http://deb.debian.org/debian bookworm/main amd64 libnss-mymachines 
amd64 252.19-1~deb12u1

  404  Not Found [IP: 2a04:4e42:41::644 80]
Err:4 http://deb.debian.org/debian bookworm/main amd64 libxentoolcore1 
amd64 4.17.2+76-ge1f9cb16e2-1~deb12u1

  404  Not Found [IP: 2a04:4e

Re: Old control sums for packages.

[Kamil Jońca]

Simon Hollenbach  writes:

> Hello KJ,
>
> there is the snapshot archive at https://snapshot.debian.org/ - You
> can get older Packages files from there.

Thanks. I was not aware of this service.
KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html

Re: Old control sums for packages.

[Simon Hollenbach]

Hello KJ,

there is the snapshot archive at https://snapshot.debian.org/ - You
can get older Packages files from there.

I don't know if you mean the links like `Packages.gz ->
by-hash/SHA256/c039245acc063d9b42cade368a874bf5e0ee3025a7bb2634f3f3bc601f15bb89`
or the actual contents of Packages: These by-hash links are in use
since somewhen in 2017 for sid, but since you were talking about
looking back a month or two, this should not be a problem.

Please also note that there was a time when Packages had no SHA256 fields.

Ciao,
Simon


On Mon, 1 Apr 2024 at 12:37, Kamil Jońca  wrote:
>
> At http://deb.debian.org/debian/dists/sid/main/binary-amd64/ we can find
> files with SHA256 sums of packages. Unfortunately they are only 2 weeks
> old. Is this possible to have little older files? (For example month or
> 2)?
> KJ
>
> --
> http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
>

Old control sums for packages.

[Kamil Jońca]

At http://deb.debian.org/debian/dists/sid/main/binary-amd64/ we can find
files with SHA256 sums of packages. Unfortunately they are only 2 weeks
old. Is this possible to have little older files? (For example month or
2)?
KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html

Re: how to downgrade nvidia-graphics-drivers packages?

[Dan Ritter]

Harald Dunkel wrote: 
> Hi folks,
> 
> Looking at a set of installed binary packages built from the same source
> package, I would like to keep the version numbers consistent. There might
> be exceptions, but in general you won't like to mix unstable and experimental
> binary packages from the nvidia-graphics-drivers, for example.
> 
> Question is, how can I tell apt to avoid mixing version numbers?

If they come from different repositories (i.e. backports,
unstable, experimental) you can set priorities in
/etc/apt/preferences.d/ -- read the man page for
apt_preferences, because it's not intuitive.

Package: *
Pin: release a=bookworm
Pin-Priority: 900

Package: *
Pin: release a=bookworm-backports
Pin-Priority: 50

Or you can set things per-package by name, or a variety of other
mechanisms.

-dsr-

how to downgrade nvidia-graphics-drivers packages?

[Harald Dunkel]

Hi folks,

Looking at a set of installed binary packages built from the same source
package, I would like to keep the version numbers consistent. There might
be exceptions, but in general you won't like to mix unstable and experimental
binary packages from the nvidia-graphics-drivers, for example.

Question is, how can I tell apt to avoid mixing version numbers?

Regards
Harri

Re: packages listed vs. apt-rdepends --follow=Depends ...

[David Wright]

On Wed 06 Dec 2023 at 23:08:41 (+), Albretch Mueller wrote:
> On 12/2/23, Albretch Mueller  wrote:
> > On 12/2/23, Tom Furie  wrote:
> >> 'apt depends ' would list the direct dependencies without
> >> recursion.
> > $ apt depends wget 2>&1 | grep "  Depends: " | awk '{ print $2}'
> 
>  that didn't work,

Tom said it wouldn't, and went on to give a recursive version.

> dpkg would still demand dependencies,

(dependencies of dependencies)

> so I decided
> to change the strategy to:
>  1) using apt-get install ...
>  2) save the install log into a file (apt-get install reports to you
> the order of installation) from which you can then created a dpkg
> based script
>  3) move all packages from /var/cache/ ... to wherever is needed.

I don't remember ever having to worry about the order. I would just
transfer the package cache from the connected PC (at work) to the
unconnected one (at home), and tell dpkg to install the lot, in one
gulp. At worst, the commands   dpkg --configure -a   and
apt-get -f install   would clear any logjams.

> On 12/2/23, Darac Marjal  wrote:
> > There used to be "apt-zip" (no longer in Debian), which was
> > built around the idea of using ZIP disks for transferring files.
> > "apt-zip-list" would use the state of packages on the disconnected
> > system to product a "want list" of files to be downloaded. This "want
> > list" would be a shell script consisting of various wget or curl
> > commands. The script would be taken over to the connected system and
> > run, to pull the required packages onto a high-capacity removable medium
> > (such as a USB drive or ZIP drive). Back at the disconnected system,
> > "apt-zip-inst" would complete the process, installing the files from the
> > removable medium.
> 
>  Hmm! ... and the apt-zip functionality doesn't exist anymore in the
> same way that it rains and thunders when the Gods decide? When a
> package is removed or discontinued, is there a formal explanation as
> to why?

My guess would be that diminishing use with increasing broadband
availability led to no interest in maintaining it, particularly
through times when changes were made to APT.

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=815456

Have you looked at apt-offline?

>  I don't know why and I decided to change my approach, but I tried to use the:
>  https://manpages.ubuntu.com/manpages/noble/en/man8/apt-get.8.html
>  -s, --simulate, --just-print, --dry-run, --recon, --no-act
>  functionality, but it didn't work:
>  "E: Command line option --dry-run is not understood in combination
> with the other options"
>  then I found confusing explanations about users being confused:
>  
> https://serverfault.com/questions/1074702/apt-get-update-dry-run-command-does-not-work-anymore
 ↑↑
As the reference says, they were trying to use -s with update, but
update doesn't involve packages, only the packages index. Well,
you know update is going to get you—just read your sources.list—so
what's the point of -s.

Cheers,
David.

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Albretch Mueller]

On 12/2/23, Albretch Mueller  wrote:
> On 12/2/23, Tom Furie  wrote:
>> 'apt depends ' would list the direct dependencies without
>> recursion.
> $ apt depends wget 2>&1 | grep "  Depends: " | awk '{ print $2}'

 that didn't work, dpkg would still demand dependencies, so I decided
to change the strategy to:
 1) using apt-get install ...
 2) save the install log into a file (apt-get install reports to you
the order of installation) from which you can then created a dpkg
based script
 3) move all packages from /var/cache/ ... to wherever is needed.
~
On 12/2/23, Darac Marjal  wrote:
> There used to be "apt-zip" (no longer in Debian), which was
> built around the idea of using ZIP disks for transferring files.
> "apt-zip-list" would use the state of packages on the disconnected
> system to product a "want list" of files to be downloaded. This "want
> list" would be a shell script consisting of various wget or curl
> commands. The script would be taken over to the connected system and
> run, to pull the required packages onto a high-capacity removable medium
> (such as a USB drive or ZIP drive). Back at the disconnected system,
> "apt-zip-inst" would complete the process, installing the files from the
> removable medium.

 Hmm! ... and the apt-zip functionality doesn't exist anymore in the
same way that it rains and thunders when the Gods decide? When a
package is removed or discontinued, is there a formal explanation as
to why?

 I don't know why and I decided to change my approach, but I tried to use the:
 https://manpages.ubuntu.com/manpages/noble/en/man8/apt-get.8.html
 -s, --simulate, --just-print, --dry-run, --recon, --no-act
 functionality, but it didn't work:
 "E: Command line option --dry-run is not understood in combination
with the other options"
 then I found confusing explanations about users being confused:
 
https://serverfault.com/questions/1074702/apt-get-update-dry-run-command-does-not-work-anymore
~
On 12/3/23, Greg Wooledge  wrote:
> After that, it was revealed that the whole project is based on some
> paranoid fantasy.  The non-networked computer is non-networked only
> because the OP believes that "they" (that's literally the word which
> was used) are using "AI" to watch the OP "24/7".  This makes me less
> inclined to take the project seriously.

 Greg, quite honestly, I'd wish it would just be my "paranoid
fantasy", but, unfortunately, I will have disappoint you, with all the
streams of data "they" are collecting from everyone of us, "they" are
keeping a data Doppelgänger of everyone of us.

 lbrtchx

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Greg Wooledge]

On Sat, Dec 02, 2023 at 10:28:14PM -0600, David Wright wrote:
> On Sat 02 Dec 2023 at 13:48:34 (+), Darac Marjal wrote:
> > On 02/12/2023 04:22, Greg Wooledge wrote:
> > apt-get has the side effect of installing the packages on the
> > connected system.
> 
> Not with the -d option. I think Greg may have suggested apt-get
> because apt autocleans as a side effect, but that can be prevented
> with a configuration option, IIRC it's APT::FtpArchive::Clean.

Well, I initially suggested "apt install ./myfile.deb" because I
thought the goal was to install a .deb file and its dependencies.

Then it was mentioned that the procedure would need to be repeated on
a non-networked computer, so I changed it to "apt-get install ./myfile.deb"
which would leave the dependencies in /v/c/a/a so they could be copied
to the non-networked computer along with the initial .deb file.

After that, it was revealed that the whole project is based on some
paranoid fantasy.  The non-networked computer is non-networked only
because the OP believes that "they" (that's literally the word which
was used) are using "AI" to watch the OP "24/7".  This makes me less
inclined to take the project seriously.

Even with that revelation, however, I still feel the most obvious
way to proceed would be to have a networked computer which "mirrors"
the non-networked one.  When you need to install something on the
non-networked computer, you first do it on the networked one, then copy
everything over to the non-networked one.

It doesn't have to be a whole computer.  It could be a VM, or just a
simple chroot directory.

Re: packages listed vs. apt-rdepends --follow=Depends ...

[David Wright]

On Sat 02 Dec 2023 at 13:48:34 (+), Darac Marjal wrote:
> On 02/12/2023 04:22, Greg Wooledge wrote:
> > On Fri, Dec 01, 2023 at 10:01:54PM -0600, David Wright wrote:
> > > On Fri 01 Dec 2023 at 21:55:42 (-0500), Greg Wooledge wrote:
> > > >  apt install ./myfile.deb
> > > That requires you to be online, aka "exposed mode". The OP only
> > > exposes a live USB to the outside world, not their "real" system.
> > > 
> > > I dimly recollect something called apt-move, but I never needed
> > > to use it. Back in the days of dial-up, when I had a real job,
> > > I would upgrade my desk's tower, copy the (uncleaned) archives/
> > > directory onto a Zip drive, take it home and install the .debs
> > > onto my home desktop, configured identically, with dpkg.
> > In that case, use apt-get instead of apt.  That way the downloaded .deb
> > files will not be removed afterward.  Then you can just sweep 'em up
> > from /var/cache/apt/archives, copy them to a stack of floppies, put
> > the floppies in a box, tie the box to a trained ferret, send the ferret
> > across town
> 
> apt-get has the side effect of installing the packages on the
> connected system.

Not with the -d option. I think Greg may have suggested apt-get
because apt autocleans as a side effect, but that can be prevented
with a configuration option, IIRC it's APT::FtpArchive::Clean.

Cheers,
David.

Re: packages listed vs. apt-rdepends --follow=Depends ...

[David Wright]

On Sat 02 Dec 2023 at 07:06:37 (+0100), to...@tuxteam.de wrote:
> On Sat, Dec 02, 2023 at 02:52:25AM +, Albretch Mueller wrote:
> >  direct dependencies of packages which haven't been downloaded,
> > install. I need to download those packages.
> >  These should be a straightforward way to do that or an easy hack.
> >  lbrtchx
> 
> I /think/ this hack might involve iterations until you hit a fixed
> point.
> 
> See -- package dependencies are listed in the package itself, so
> you know your original package's direct dependencies. Off you go,
> download those, look into those packages, find the dependencies...
> and off you go, download the dependencies's dependencies.
> 
> Until you reach the fixed point.
> 
> There's one package, apt-cache, which can look at a pre-made complete
> dependencies's network. But to update that database you have to be
> online...
> 
> Sounds like quite the fun.

Would it not be more straightforward to download APT's lists, and copy
them to the unconnected machine. # apt-get update   will bitch and
moan about Release files, but still recreate the .bin caches, and
$ apt-get -s -d install foo   will list all the missing dependencies
(± --no-install-recommends to taste) in one fell swoop.

Cheers,
David.

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Greg Wooledge]

On Sat, Dec 02, 2023 at 06:15:17AM +, Albretch Mueller wrote:
> They are even using "AI" to mess with
> people they target and it doesn't matter if they know well (which they
> have actually told me) that you are not a criminally minded dude, a
> threat to society, ... and they are quite literally
> watching/monitoring you 24x7.

Yah, OK.

You do know that those emails you receive claiming you've been "hacked"
or that "I know what you did" are just phishing, right?

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Darac Marjal]

On 02/12/2023 04:22, Greg Wooledge wrote:

On Fri, Dec 01, 2023 at 10:01:54PM -0600, David Wright wrote:

On Fri 01 Dec 2023 at 21:55:42 (-0500), Greg Wooledge wrote:

 apt install ./myfile.deb

That requires you to be online, aka "exposed mode". The OP only
exposes a live USB to the outside world, not their "real" system.

I dimly recollect something called apt-move, but I never needed
to use it. Back in the days of dial-up, when I had a real job,
I would upgrade my desk's tower, copy the (uncleaned) archives/
directory onto a Zip drive, take it home and install the .debs
onto my home desktop, configured identically, with dpkg.

In that case, use apt-get instead of apt.  That way the downloaded .deb
files will not be removed afterward.  Then you can just sweep 'em up
from /var/cache/apt/archives, copy them to a stack of floppies, put
the floppies in a box, tie the box to a trained ferret, send the ferret
across town


apt-get has the side effect of installing the packages on the connected 
system. There used to be "apt-zip" (no longer in Debian), which was 
built around the idea of using ZIP disks for transferring files. 
"apt-zip-list" would use the state of packages on the disconnected 
system to product a "want list" of files to be downloaded. This "want 
list" would be a shell script consisting of various wget or curl 
commands. The script would be taken over to the connected system and 
run, to pull the required packages onto a high-capacity removable medium 
(such as a USB drive or ZIP drive). Back at the disconnected system, 
"apt-zip-inst" would complete the process, installing the files from the 
removable medium.


The nice thing about "apt-zip" was that it took the guesswork out of the 
equation. The files required were the ones that the target system 
required, no more no less. Also, the connected system didn't have to be 
debian; there was an option to write the script in a DOS-compatible 
manner, so you could run it on Windows, for example.


I don't know if there's a direct replacement for apt-zip nowadays.



If the OP doesn't have a same-release, same-architecture connected
system to use for this purpose, then I don't have an answer.  I don't
deal with this stone-age crap any longer, and I am unable to express
how *happy* I am that this is the case.
There are still use cases for fully disconnected systems these days. The 
most common one might be an offline Certificate Authority (best practice 
says that the host holding your certificate authority certificate should 
NEVER have network access, to prevent any possibility of compromise), 
but some security professionals prescribe "air-gap" security for other 
systems (think of the Iranian Uranium Enrichment system that Stuxnet 
compromised). For these sorts of systems, you're stuck with using 
something like apt-zip, or else just downloading the point-release ISOs 
and burning them.


OpenPGP_signature.asc
Description: OpenPGP digital signature

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Albretch Mueller]

On 12/2/23, Tom Furie  wrote:
> 'apt depends ' would list the direct dependencies without
> recursion.

 Thank you, I think I got what I needed (at least for now).

$ apt depends wget
wget
  Depends: libc6 (>= 2.28)
  Depends: libgnutls30 (>= 3.7.0)
  Depends: libidn2-0 (>= 0.6)
  Depends: libnettle8
  Depends: libpcre2-8-0 (>= 10.22)
  Depends: libpsl5 (>= 0.16.0)
  Depends: libuuid1 (>= 2.16)
  Depends: zlib1g (>= 1:1.1.4)
  Conflicts: 
  Recommends: ca-certificates
$

$ apt depends wget 2>&1 | grep "  Depends: " | awk '{ print $2}'
libc6
libgnutls30
libidn2-0
libnettle8
libpcre2-8-0
libpsl5
libuuid1
zlib1g
$

On 12/2/23, Greg Wooledge  wrote:
> On Fri, Dec 01, 2023 at 10:01:54PM -0600, David Wright wrote:
>> That requires you to be online, aka "exposed mode". The OP only
>> exposes a live USB to the outside world, not their "real" system.
>
> If the OP doesn't have a same-release, same-architecture connected
> system to use for this purpose, then I don't have an answer.  I don't
> deal with this stone-age crap any longer, and I am unable to express
> how *happy* I am that this is the case.

 At times *being* and *keeping moving* is entirely fine, you don't
have to be *happy* and I am not sure if I am living in the stone-age
or the brave new world in which we are living these days. One thing is
sure, I can't be naive about any of it.

 I do have the "same-release, same-architecture connected system" I
just want to make as sure (among other things meaning -on a hardware
level-), as I possibly could that, by simply rebooting a live DVD you
can "switch" (as it were) from one to the other.

 As a simple matter of economy (I use almost exclusively amd64 based
hardware), I am trying to streamline a way of:
 a) booting a live Deb Linux from a DVD (which, physically, you can't
write onto)
 b) running a short script with an array of utility packages which are
not part of §a's install base preferably from a pen drive or a
subdirectory of a hard drive you could mount
 then you would:
 1) boot up
 2) (from the USB pen drive) run §b's dpkg-based script
 3) install the extra stuff you need including Selenium-automation
versions of firefox (gecko) and brave (chromium)
~
 4) go into "exposed mode": setup networking, macchanger (I know in my
case it is silly, but I do it anyway), ...
 ...
 after going about the exposed-mode tests you need to and all
etceteras you had to check out online
~
 (n-2)) disconnect yourself from the internet (software + hardware by
removing the wifi USB dongle or cable) ...
 (n-1)) optionally, run script to check which files were changed
during your run and how
 n) shutdown

 Some of you have suggested using apt-clone and apt-move, but I think
this is a simpler way to solve what you see as an XY problem and it
doesn't really matter if you use a "different", "newer" version of Deb
Linux. Most probably all of this is plain nonsense to all of you, but
I have no other way to work. They are even using "AI" to mess with
people they target and it doesn't matter if they know well (which they
have actually told me) that you are not a criminally minded dude, a
threat to society, ... and they are quite literally
watching/monitoring you 24x7.

 lbrtchx

Re: packages listed vs. apt-rdepends --follow=Depends ...

[tomas]

On Sat, Dec 02, 2023 at 02:52:25AM +, Albretch Mueller wrote:
>  direct dependencies of packages which haven't been downloaded,
> install. I need to download those packages.
>  These should be a straightforward way to do that or an easy hack.
>  lbrtchx

I /think/ this hack might involve iterations until you hit a fixed
point.

See -- package dependencies are listed in the package itself, so
you know your original package's direct dependencies. Off you go,
download those, look into those packages, find the dependencies...
and off you go, download the dependencies's dependencies.

Until you reach the fixed point.

There's one package, apt-cache, which can look at a pre-made complete
dependencies's network. But to update that database you have to be
online...

Sounds like quite the fun.

Cheers
-- 
t


signature.asc
Description: PGP signature

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Greg Wooledge]

On Fri, Dec 01, 2023 at 10:01:54PM -0600, David Wright wrote:
> On Fri 01 Dec 2023 at 21:55:42 (-0500), Greg Wooledge wrote:
> > apt install ./myfile.deb
> 
> That requires you to be online, aka "exposed mode". The OP only
> exposes a live USB to the outside world, not their "real" system.
> 
> I dimly recollect something called apt-move, but I never needed
> to use it. Back in the days of dial-up, when I had a real job,
> I would upgrade my desk's tower, copy the (uncleaned) archives/
> directory onto a Zip drive, take it home and install the .debs
> onto my home desktop, configured identically, with dpkg.

In that case, use apt-get instead of apt.  That way the downloaded .deb
files will not be removed afterward.  Then you can just sweep 'em up
from /var/cache/apt/archives, copy them to a stack of floppies, put
the floppies in a box, tie the box to a trained ferret, send the ferret
across town

If the OP doesn't have a same-release, same-architecture connected
system to use for this purpose, then I don't have an answer.  I don't
deal with this stone-age crap any longer, and I am unable to express
how *happy* I am that this is the case.

Re: packages listed vs. apt-rdepends --follow=Depends ...

[David Wright]

On Fri 01 Dec 2023 at 21:55:42 (-0500), Greg Wooledge wrote:
> On Sat, Dec 02, 2023 at 02:52:25AM +, Albretch Mueller wrote:
> >  direct dependencies of packages which haven't been downloaded,
> > install. I need to download those packages.
> >  These should be a straightforward way to do that or an easy hack.
> 
> I'm still struggling to figure out what the X is in this gigantic X-Y
> problem.
> 
> Do you have a .deb file, and you want to install it, along with all of
> its dependencies?  If that's the X, then you do this:
> 
> apt install ./myfile.deb
> 
> That's it.

[In view of the time of day:]

That requires you to be online, aka "exposed mode". The OP only
exposes a live USB to the outside world, not their "real" system.

I dimly recollect something called apt-move, but I never needed
to use it. Back in the days of dial-up, when I had a real job,
I would upgrade my desk's tower, copy the (uncleaned) archives/
directory onto a Zip drive, take it home and install the .debs
onto my home desktop, configured identically, with dpkg.

Cheers,
David.

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Tom Furie]

Albretch Mueller  writes:

>  How can you list just the direct dependencies? and how safe is it
> downloading and installing only those via dpkg?

'apt depends ' would list the direct dependencies without
recursion.

Why do you want to download them individually and install directly with
dpkg when apt can handle it all cleanly? Whatever it is you're trying to
do, this feels like the wrong way to go about it...

Cheers,
Tom

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Greg Wooledge]

On Sat, Dec 02, 2023 at 02:52:25AM +, Albretch Mueller wrote:
>  direct dependencies of packages which haven't been downloaded,
> install. I need to download those packages.
>  These should be a straightforward way to do that or an easy hack.

I'm still struggling to figure out what the X is in this gigantic X-Y
problem.

Do you have a .deb file, and you want to install it, along with all of
its dependencies?  If that's the X, then you do this:

apt install ./myfile.deb

That's it.

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Albretch Mueller]

 direct dependencies of packages which haven't been downloaded,
install. I need to download those packages.
 These should be a straightforward way to do that or an easy hack.
 lbrtchx

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Albretch Mueller]

On 12/2/23, Tom Furie  wrote:
...
> This is a recursive search, also showing dependencies of dependencies,
> etc.

 How can you list just the direct dependencies? and how safe is it
downloading and installing only those via dpkg?

 lbrtchx

Re: packages listed vs. apt-rdepends --follow=Depends ...

[Tom Furie]

Albretch Mueller  writes:

> https://packages.debian.org/bullseye/wget
>
> shows 8 packages as "depends"
>
> dep: libc6 (>= 2.28)
> dep: libgnutls30 (>= 3.7.0)
> dep: libidn2-0 (>= 0.6)
> dep: libnettle8
> dep: libpcre2-8-0 (>= 10.22)
> dep: libpsl5 (>= 0.16.0)
> dep: libuuid1 (>= 2.16)
> dep: zlib1g (>= 1:1.1.4)

These are direct dependencies.

>  vs. 17 using apt-rdepends --follow=Depends:
>
> $ apt-rdepends --follow=Depends "wget" | grep --invert-match ^\
> Reading package lists... Done
> Building dependency tree... Done
> Reading state information... Done
> wget
> libc6
> libcrypt1
> libgcc-s1
> gcc-10-base
> libgnutls30
> libgmp10
> libhogweed6
> libnettle8
> libidn2-0
> libunistring2
> libp11-kit0
> libffi7
> libtasn1-6
> libpcre2-8-0
> libpsl5
> libuuid1
> zlib1g

This is a recursive search, also showing dependencies of dependencies,
etc.

Cheers,
Tom

packages listed vs. apt-rdepends --follow=Depends ...

[Albretch Mueller]

https://packages.debian.org/bullseye/wget

shows 8 packages as "depends"

dep: libc6 (>= 2.28)
dep: libgnutls30 (>= 3.7.0)
dep: libidn2-0 (>= 0.6)
dep: libnettle8
dep: libpcre2-8-0 (>= 10.22)
dep: libpsl5 (>= 0.16.0)
dep: libuuid1 (>= 2.16)
dep: zlib1g (>= 1:1.1.4)
~
 vs. 17 using apt-rdepends --follow=Depends:

$ apt-rdepends --follow=Depends "wget" | grep --invert-match ^\
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
wget
libc6
libcrypt1
libgcc-s1
gcc-10-base
libgnutls30
libgmp10
libhogweed6
libnettle8
libidn2-0
libunistring2
libp11-kit0
libffi7
libtasn1-6
libpcre2-8-0
libpsl5
libuuid1
zlib1g
~
 these packages weren't listed on packages.debian.org:

libcrypt1
libgcc-s1
gcc-10-base
libgmp10
libhogweed6
libunistring2
libp11-kit0
libffi7
libtasn1-6
~
 why? and which list should one trust (more)?

 lbrtchx

Re: used vs. unused packages installed

[Nate Bargmann]

For this sort of thing I prefer the aptitude TUI.  Highlight the package
in question and hit 'r' and the list of reverse dependencies appears.
Installed packages will be in bold (also bright white with my terminal
settings).  One can continue up the chain by highlighting one of the
installed reverse dependencies and pressing 'r' again, and so on.

Another trick I have used is to use 'M' to mark the package as
automatically installed, if nothing depends on it then it will be marked
for removal.  Be careful here as a manually installed top level package
you might want to keep such as vim will be marked for removal and all
packages that were automatically installed with it.  To reverse the
proposed removal action, use 'Ctrl-u'.

While I often use apt at the command line, I've been using aptitude
since early 2001 and often prefer its TUI for doing drastic things!

- Nate

-- 
"The optimist proclaims that we live in the best of all
possible worlds.  The pessimist fears this is true."
Web: https://www.n0nb.us
Projects: https://github.com/N0NB
GPG fingerprint: 82D6 4F6B 0E67 CD41 F689 BBA6 FB2C 5130 D55A 8819



signature.asc
Description: PGP signature

Re: used vs. unused packages installed

[Michel Verdier]

On 2023-11-30, David Wright wrote:

> deborphan -Ps   or orphaner

Perhaps
deborphan -Ps --ignore-suggests

Or even
deborphan -Ps --ignore-suggests --ignore-recommends

Re: used vs. unused packages installed

[Michel Verdier]

On 2023-11-30, Greg Wooledge wrote:

> My first thought is that popularity-contest should be able to tell you
> this, because it's able to tell *Debian* which packages are "old"

I should live on the "old" but mandatory edge :)

20 tk
20 tcl
14 g++

Re: used vs. unused packages installed

[David Wright]

On Thu 30 Nov 2023 at 16:06:06 (-0600), Mike McClain wrote:
> Is there any way to determine which packages are used of the many
> that come with an install?

I don't know of one.

> My Raspberry Pi install of bookworm has some 1800 packages
> installed many of which I know I don't use, many others I suspect I
> don't use but don't know if some program I do use depends on them at
> some point in its life.

My principal bullseye has 2064, and has no DE, but includes
*TeX, LibreOffice, and a generous number of fonts. Without
knowing this sort of information, it's difficult to judge.

An obvious method is --no-install-recommends, but don't be surprised
when some packages lack functionality that you expect to be present.

> $ apropos editor | wc   reports 23 hits
> Six of which are various versions of VI which I don't use but pico,
> nano, mcedit, mousepad and mu-editor are also included. I only use jed
> but don't know what would break if I purged the others and am loathe
> to break a working system.

Some of these are in the same package (bits of vim ± gui) or part
of another package (mc/mcedit), and some are too small to worry about
(nano/pico). Others are too specialised to be thought of as just
editors (editres, gparted, mid3v2), and the lack of some will break
your system (sed).

> There are 259 packages whose name starts with 'python', admittedly I
> could purge one a week and see if anything breaks, that would only take
> 5 years but I'm not quite that patient.

Only 58 here; what am I doing wrong? No, actually I thought you would
approve of the number, as it goes to show how much python has been
fragmented so that you only have to install the parts needed.

> Suggestions?

apt-get --purge autoremove
deborphan -Ps   or orphaner
cruft-ng
debfoster

The last of these is, I think, like --no-install-recommends,
something you set up beforehand.

BTW are you seriously short of space, or just a tidy person?
(You don't have to answer.)

Cheers,
David.

Re: used vs. unused packages installed

[John Hasler]

 Mike McClain writes:
> Is there any way to determine which packages are used of the many that
> come with an install?  My Raspberry Pi install of bookworm has some
> 1800 packages installed many of which I know I don't use, many others
> I suspect I don't use but don't know if some program I do use depends
> on them at some point in its life.

When you remove a package using apt it removes packges that depend on
it.

Run

apt -s remove  | grep Remv

This will list all the packages that depend on  and therefor
would also be removed. "apt -s" simulates the action without actually
changing anything.  It needn't be run as root.
-- 
John Hasler 
j...@sugarbit.com
Elmwood, WI USA

Re: used vs. unused packages installed

[Greg Wooledge]

On Thu, Nov 30, 2023 at 04:06:06PM -0600, Mike McClain wrote:
> Is there any way to determine which packages are used of the many
> that come with an install?

My first thought is that popularity-contest should be able to tell you
this, because it's able to tell *Debian* which packages are "old"
(not recently used).

One of the programs that comes with it is /usr/sbin/popcon-largest-unused
which looks like what you want.

See <https://manpages.debian.org/popcon-largest-unused>.  Looks like
you have to allow popcon to run at least once.

Re: used vs. unused packages installed

[Dan Ritter]

Mike McClain wrote: 
> Is there any way to determine which packages are used of the many
> that come with an install?
> My Raspberry Pi install of bookworm has some 1800 packages
> installed many of which I know I don't use, many others I suspect I
> don't use but don't know if some program I do use depends on them at
> some point in its life.
> $ apropos editor | wc   reports 23 hits
> Six of which are various versions of VI which I don't use but pico,
> nano, mcedit, mousepad and mu-editor are also included. I only use jed
> but don't know what would break if I purged the others and am loathe
> to break a working system.
> There are 259 packages whose name starts with 'python', admittedly I
> could purge one a week and see if anything breaks, that would only take
> 5 years but I'm not quite that patient.
> 
> Suggestions?

Pick a suspect package and run

apt-cache rdepends --installed PACKAGE

That shows you all the packages which are installed and depend
on this PACKAGE.

That saves you from the uninstall-and-see-what-breaks dance.

-dsr-

used vs. unused packages installed

[Mike McClain]

Is there any way to determine which packages are used of the many
that come with an install?
My Raspberry Pi install of bookworm has some 1800 packages
installed many of which I know I don't use, many others I suspect I
don't use but don't know if some program I do use depends on them at
some point in its life.
$ apropos editor | wc   reports 23 hits
Six of which are various versions of VI which I don't use but pico,
nano, mcedit, mousepad and mu-editor are also included. I only use jed
but don't know what would break if I purged the others and am loathe
to break a working system.
    There are 259 packages whose name starts with 'python', admittedly I
could purge one a week and see if anything breaks, that would only take
5 years but I'm not quite that patient.

Suggestions?

Thanks,
Mike McClain
--
Every problem has a gift for you in its hands.
- Richard Bach

[SOLVED] Uninstalling unwanted PHP and mariadb packages [WAS Re: How could the missing MySQL extension required by WordPress be installed?]

[Susmita/Rajib]

For all the debian-users who might fall into the same ditch into which
I had fallen, to alert them ...

I thank Mr. Cater for his advice at
https://lists.debian.org/debian-user/2023/11/msg00443.html. I raed his
advice and understood that while uninstalling I should follow a path
chronologically reverse to the one I had followed while installing all
those packages on the 10th Nov 2023.

So what I did was that I noted down his advice on a plain text editor,
then opted for "uninstall completely ..." the following packages:
php-common
apache2
apache2-bin
apache2-data
mariadb-client-10.5

Following my practice of safety, I kept a note of the packages I had
installed, as recorded in my first post of this thread, at:
https://lists.debian.org/debian-user/2023/11/msg00384.html
as follows:
[quote]
The nearest that are available in Debian Bullseye :
default-mysql-server
default-mysql-server-core
php-mysql
php7.4-mysql
apache2
php-db
along with their dependencies.

I installed them all. ...
[/quote]

Then I went to my list of synaptic installation history and from the
block categories described earlier in:
https://lists.debian.org/debian-user/2023/11/msg00411.html,
individually selected the still unselected packages. All of them,
listed on the History list of 10th November 2023.

With a little anxiety, I proceeded for a complete uninstallation.

The Uninstallation completed without an error message (synaptic
reported in the "Changes Applied" window: "Successfully applied all
changes...") as follows:

[output]
(Reading database ... 270111 files and directories currently installed.)
Removing apache2 (2.4.56-1~deb11u2) ...
Removing apache2-bin (2.4.56-1~deb11u2) ...
dpkg: warning: while removing apache2-bin, directory
'/var/lib/apache2' not empty so not removed
Removing apache2-data (2.4.56-1~deb11u2) ...
Removing apache2-utils (2.4.56-1~deb11u2) ...
Removing default-mysql-server (1.0.7) ...
Removing default-mysql-server-core (1.0.7) ...
Removing mariadb-server-10.5 (1:10.5.21-0+deb11u1) ...
Removing galera-4 (26.4.11-0+deb11u1) ...
Removing libaprutil1-ldap:amd64 (1.6.1-5+deb11u1) ...
Removing libaprutil1-dbd-sqlite3:amd64 (1.6.1-5+deb11u1) ...
Removing libaprutil1:amd64 (1.6.1-5+deb11u1) ...
Removing libapr1:amd64 (1.7.0-6+deb11u2) ...
Removing libcgi-fast-perl (1:2.15-1) ...
Removing libhtml-template-perl (2.97-1.1) ...
Removing libcgi-pm-perl (4.51-1) ...
Removing mariadb-client-10.5 (1:10.5.21-0+deb11u1) ...
Removing libconfig-inifiles-perl (3.03-1) ...
Removing libdbd-mariadb-perl (1.21-3) ...
Removing libfcgi-bin (2.4.2-2) ...
Removing libfcgi-perl:amd64 (0.79+ds-2) ...
Removing libfcgi0ldbl:amd64 (2.4.2-2) ...
Removing liblua5.3-0:amd64 (5.3.3-1.1+deb11u1) ...
Removing mailutils (1:3.10-3+b1) ...
Removing libmailutils7:amd64 (1:3.10-3+b1) ...
Removing mariadb-client-core-10.5 (1:10.5.21-0+deb11u1) ...
Removing libmariadb3:amd64 (1:10.5.18-0+deb11u1) ...
Removing libterm-readkey-perl (2.38-1+b2) ...
Removing mariadb-server-core-10.5 (1:10.5.21-0+deb11u1) ...
Removing mariadb-common (1:10.5.21-0+deb11u1) ...
update-alternatives: using /etc/mysql/my.cnf.fallback to provide
/etc/mysql/my.cnf (my.cnf) in auto mode
Removing php-db (1.10.0-1) ...
Removing php-pear (1:1.10.12+submodules+notgz+20210212-1) ...
Removing php-cli (2:7.4+76) ...
update-alternatives: using /usr/bin/php7.4 to provide /usr/bin/php
(php) in auto mode
update-alternatives: using /usr/bin/phar7.4 to provide /usr/bin/phar
(phar) in auto mode
update-alternatives: using /usr/bin/phar.phar7.4 to provide
/usr/bin/phar.phar (phar.phar) in auto mode
Removing php-xml (2:7.4+76) ...
Removing php7.4-xml (7.4.33-1+deb11u4) ...
Removing php7.4-cli (7.4.33-1+deb11u4) ...
Removing php7.4-readline (7.4.33-1+deb11u4) ...
Removing php-mysql (2:7.4+76) ...
Removing php7.4-opcache (7.4.33-1+deb11u4) ...
Removing php7.4-json (7.4.33-1+deb11u4) ...
Removing php7.4-mysql (7.4.33-1+deb11u4) ...
Removing socat (1.7.4.1-3) ...
Removing php7.4-common (7.4.33-1+deb11u4) ...
Removing php-common (2:76) ...
Warning: Stopping phpsessionclean.service, but it can still be activated by:
  phpsessionclean.timer
Processing triggers for man-db (2.9.4-2) ...
Processing triggers for menu (2.1.48) ...
Processing triggers for libc-bin (2.31-13+deb11u5) ...
(Reading database ... 268625 files and directories currently installed.)
Purging configuration files for php-common (2:76) ...
Purging configuration files for apache2 (2.4.56-1~deb11u2) ...
Purging configuration files for mariadb-common (1:10.5.21-0+deb11u1) ...
Purging configuration files for mariadb-client-10.5 (1:10.5.21-0+deb11u1) ...
Processing triggers for menu (2.1.48) ...
[/output]

So a SUMMARY becomes due:
To note down the bulk packages while installing, follow a path
chronologically reverse to the one followed during installing, then
from the "History of installed, upgraded and removed packages" also
select the associated packages installed along with the bulk packages
a

Uninstalling unwanted PHP and mariadb packages [WAS Re: How could the missing MySQL extension required by WordPress be installed?]

[Susmita/Rajib]

From: "Andrew M.A. Cater" 
Date: Sat, 11 Nov 2023 16:39:01 +
Message-id: <[🔎] zu-ujrqewxkor...@einval.com>
In-reply-to: <[🔎]
CAEG4cZXquvxryjdFLCgHBiteyx8mFm=e+kjcusxnhhj0+zp...@mail.gmail.com>,
https://lists.debian.org/debian-user/2023/11/msg00415.html

On Sun, 12 Nov 2023 at 07:42, Susmita/Rajib  wrote:
>
> On Sat, Nov 11, 2023 at 05:44:29PM +0530, Susmita/Rajib wrote:
> > Apologise for the typos and incorrect sentence constructions in my
[ ... ]
> > Possible?
> >
>
> I'd suggest you remove
>
> php-common
> apache2
> apache2-bin
> apache2-data
> mariadb-client-10.5
>
> and then an apt autoremove.
>
> Then perhaps an apt-purge apache2 .. and so on.
> Then go back and remove any outlying packages that have not already been
> removed until you have no more left.
[ ... ]

Ok. Thank you Mr. Cater, for your guidance.

[ ... ]
> For the future - for any reader of this list, not just Rajib:
>
> If you don't know quite what you're installing and a single package installs
> lots of other packages, stop and consider before you install. Meta-packages
> like desktops pull in lots of dependencies, but so, as you've seen does
> PHP
>
> If you're really not sure: read lots then ask. If you say

I sincerely apologise, Mr. Cater.
I am not posting an excuse, but sometimes, the excitement of the
prospect of a new perspective and more wisdom becomes too overpowering
to restrain myself from undue risk-taking. A millennium of lost
opportunities, trampled hopes and prospects, the lack of a beautiful
and fulfilling community life and a prosperous society enhances the
excitement exponentially. In this regard we share our common pain with
the Israelis, suffering from multiple-centuries of genocide even many
times more than the 2nd World War Holocaust. Our foodgrains were
repeatedly diverted, causing many man-made famines for more than a
century, killing and displacing my ancestors and made us more
susceptible to diabetes and heart diseases. But as the Israelis also
endorse, the only people that didn't persecute them were us.

My conscience cries for the sufferings of a grand continuous
civilisation and her peoples, Mr. Cater.

Prof. C K Raju's articles entice with the prospect of identifying the
real perpetrators of such inhuman atrocities with more evidence. So
the excitement couldn't be contained.

> "I want to do X. I think that means I need to install Y and configure Z.
> >From reading ABC pages, I _think_ that means that I need to do the
> following but I'm not sure because page C says to do this and I don't
> understand. I'm running Debian 12.2 updated this week. "
>
>  - it becomes a lot easier to understand what you want.

Yes, I have followed the general sensible guidelines. But in
excitement I couldn't restrain myself from experimenting. I didn't
consider the problem to be so irritating. So I didn't consider that I
would have to ask for support later. I sincerely apologise, Mr. Cater.

[ ... ]
> If you explain _exactly_ what error messages you get and how you get them
> that helps significantly. If you've done a web search and found a possible
> answer or that other people have had a similar problem, add that in.
>
> "I'm getting error message DEF from running this command. From the net,
> it looks as if other people have been having similar problems on older
> versions of Debian but not recently - am I reading this correctly"
>
> It shows that you have made an attempt to problem solve for yourself and
> people will be more willing to help and explain their suggestions.
[ ... ]

By good coincidence, I follow this sensible recording of experiences.
But thank you for reminding and consolidating the idea behind the
actions in a clearer language.

> If the advice you get solves a problem - post that back to the list
> in the thread and change the message subject slightly to make it
> clear that this is an answer. For example:
>
> [SOLVED] Package foo fails to install in Debian 12.2. [Was: Cannot install
>  foo]
[ ... ]

There is one problem with this approach, as stated earlier. For the
Mailing list, any change of subject from my Gmail webmail email-server
makes such an email detached from the main thread and treats it as a
different subject. Had Mailman also had an email editor for posting
messages, that would have been a better option.

However, I understand the difficulties and the legalities associated
with this added functionality and acknowledge the associated problems,
since FSF has to bear with the Deep State more, bringing an added
peril to itself. We already know the fate of Assange, Snowden and
myriads of unknown others.

In countries where the Law owes its existence to Xtian Morality

Uninstalling unwanted PHP and mariadb packages [WAS Re: How could the missing MySQL extension required by WordPress be installed?]

[Andrew M.A. Cater]

On Sat, Nov 11, 2023 at 05:44:29PM +0530, Susmita/Rajib wrote:
> Apologise for the typos and incorrect sentence constructions in my
> earlier post. But I am sure that Mr.  would be able to bypass the
> errors to extract the information intended for him.
> 
> My synaptic has a Log file for the installed packages yesterday:
> 
> 
> 
> Commit Log for Fri Nov 10 21:04:10 2023
> 
> Upgraded the following packages:
> mariadb-common (1:10.5.18-0+deb11u1) to 1:10.5.21-0+deb11u1
> 
> Installed the following packages:
> default-mysql-server (1.0.7)
> galera-4 (26.4.11-0+deb11u1)
> libcgi-fast-perl (1:2.15-1)
> libcgi-pm-perl (4.51-1)
> libconfig-inifiles-perl (3.03-1)
> libdbd-mariadb-perl (1.21-3)
> libdbi-perl (1.643-3+b1)
> libfcgi-bin (2.4.2-2)
> libfcgi-perl (0.79+ds-2)
> libfcgi0ldbl (2.4.2-2)
> libhtml-template-perl (2.97-1.1)
> libterm-readkey-perl (2.38-1+b2)
> mariadb-client-10.5 (1:10.5.21-0+deb11u1)
> mariadb-client-core-10.5 (1:10.5.21-0+deb11u1)
> mariadb-server-10.5 (1:10.5.21-0+deb11u1)
> mariadb-server-core-10.5 (1:10.5.21-0+deb11u1)
> socat (1.7.4.1-3)
> 
> 
> 
> Commit Log for Fri Nov 10 21:13:50 2023
> 
> Installed the following packages:
> default-mysql-server-core (1.0.7)
> 
> 
> 
> Commit Log for Fri Nov 10 21:25:16 2023
> 
> Installed the following packages:
> php-common (2:76)
> php-mysql (2:7.4+76)
> php7.4-common (7.4.33-1+deb11u4)
> php7.4-mysql (7.4.33-1+deb11u4)
> 
> 
> 
> Commit Log for Fri Nov 10 21:37:54 2023
> 
> Installed the following packages:
> apache2 (2.4.56-1~deb11u2)
> apache2-bin (2.4.56-1~deb11u2)
> apache2-data (2.4.56-1~deb11u2)
> apache2-utils (2.4.56-1~deb11u2)
> libapr1 (1.7.0-6+deb11u2)
> libaprutil1 (1.6.1-5+deb11u1)
> libaprutil1-dbd-sqlite3 (1.6.1-5+deb11u1)
> libaprutil1-ldap (1.6.1-5+deb11u1)
> liblua5.3-0 (5.3.3-1.1+deb11u1)
> 
> 
> 
> Commit Log for Fri Nov 10 21:44:41 2023
> 
> Installed the following packages:
> php-cli (2:7.4+76)
> php-db (1.10.0-1)
> php-pear (1:1.10.12+submodules+notgz+20210212-1)
> php-xml (2:7.4+76)
> php7.4-cli (7.4.33-1+deb11u4)
> php7.4-json (7.4.33-1+deb11u4)
> php7.4-opcache (7.4.33-1+deb11u4)
> php7.4-readline (7.4.33-1+deb11u4)
> php7.4-xml (7.4.33-1+deb11u4)
> 
> 
> So what do I need to do to clean-uninstall them all?
> 
> Possible?
>

I'd suggest you remove 

php-common
apache2 
apache2-bin
apache2-data
mariadb-client-10.5

and then an apt autoremove.

Then perhaps an apt-purge apache2 .. and so on.
Then go back and remove any outlying packages that have not already been
removed until you have no more left.

For the future - for any reader of this list, not just Rajib:

If you don't know quite what you're installing and a single package installs
lots of other packages, stop and consider before you install. Meta-packages
like desktops pull in lots of dependencies, but so, as you've seen does
PHP

If you're really not sure: read lots then ask. If you say 

"I want to do X. I think that means I need to install Y and configure Z.
>From reading ABC pages, I _think_ that means that I need to do the 
following but I'm not sure because page C says to do this and I don't
understand. I'm running Debian 12.2 updated this week. "

 - it becomes a lot easier to understand what you want.

If you explain _exactly_ what error messages you get and how you get them
that helps significantly. If you've done a web search and found a possible
answer or that other people have had a similar problem, add that in. 

"I'm getting error message DEF from running this command. From the net,
it looks as if other people have been having similar problems on older
versions of Debian but not recently - am I reading this correctly" 

It shows that you have made an attempt to problem solve for yourself and
people will be more willing to help and explain their suggestions.
If the advice you get solves a problem - post that back to the list
in the thread and change the message subject slightly to make it
clear that this is an answer. For example:

[SOLVED] Package foo fails to install in Debian 12.2. [Was: Cannot install
 foo]

It is always possible that no-one else has ever seen anything similar. The
advice you get from people here may always be a best guess as to a solution.
The more clearly you can give information, the easier the problem solving
becomes.

If you have to ask lots of questions around the same subject, consider
stopping and writing a summary post to write down what you've learned
and solutions you have tried so far.

SUMMARY - Package bar still failing to install after several tries

"I have been given solution A in a post in this thread on November 1st.
I did this but then got error message B - see post on November 3rd
 so I did C on November 4th. The problem I have now is that the
error message is DE" 

Hope this helps,

Andy

[amaca...@debian.org]

Nix on Debian for installing packages

[Nicolas George]

Hi.

I am considering using Nix to install packages that are not available in
Debian, or not available in the version I need. But I ear NixOS has a
quite different taste than usual Linux distros, and I know Debian, and
all our homemade admin scripts are tailored for Debian-based systems, so
I will keep Debian as the base system.

I was hoping somebody here might already have done something like that
and have notes they could share to kick-start me.

I can start with a simple scenario, a simple program with only library
dependencies and no inter-dependency configuration.

So, starting from a standard stable Debian installation, what do I need
to type in order to have Nix install (for example) OpenVPN somewhere?
After that, I would like to be able to forget it was installed using Nix
until it is time to upgrade it: I just configure and run it, like I
installed it from source.

So, will somebody give me the hints that will save me some times in
trial and error to get my head around the logic?

Thanks in advance.

-- 
  Nicolas George


signature.asc
Description: PGP signature

Re: Which Virtual Manager? Was: EASY way to install packages from trixie/sid to stable?

[The Wanderer]

On 2023-10-28 at 00:25, Max Nikulin wrote:

> On 28/10/2023 02:02, The Wanderer wrote:
> 
>> for the case of hierarchical snapshots
> 
> qemu-img(1) allows to create snapshots of disk images that are stored
> in the same file. In addition the "create" command has the "-b 
> BACKING_FILE" option

Does virt-manager expose this feature via a convenient
create-a-new-snapshot GUI, showing the tree of which snapshots descend
from what? I believe I was under the impression that, for the case of
qemu, no such thing was available.

(I note, again, that it's been a long time since I tried this; I have
spent the past couple of years, at least, attempting with various
degrees of desperation to avoid more stress than I can handle, and after
my last try with virt-manager hit that wall this conversation is the
first time I've been able to handle any kind of try-it-again.)

>> If the option BACKING_FILE is specified, then the image will record
>> only the differences from BACKING_FILE.
> 
> Is it something close to "hierarchical snapshots"?

If one snapshot can descend from another, and you can delete any
snapshot (again, from that GUI) and have any references to it in other
snapshots automagically cleaned up to point to any relevant new parent
(such that the data seen through those other snapshots is still the same
as before the deletion), then probably.

It sounds like it might be worth my giving this another try, with qemu
as a backend, and seeing if I get any better results. Not sure exactly
when I may do that - I'm dealing with major stress from another source
right now, and would probably have difficulty handling it if the attempt
failed - but the idea is at least on my radar.

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw



signature.asc
Description: OpenPGP digital signature

Re: Which Virtual Manager? Was: EASY way to install packages from trixie/sid to stable?

[Max Nikulin]

On 28/10/2023 02:02, The Wanderer wrote:

for the case of hierarchical snapshots


qemu-img(1) allows to create snapshots of disk images that are stored in 
the same file. In addition the "create" command has the "-b 
BACKING_FILE" option



If the option BACKING_FILE is specified, then the image will record only
the differences from BACKING_FILE.


Is it something close to "hierarchical snapshots"?

Re: Which Virtual Manager? Was: EASY way to install packages from trixie/sid to stable?

[Charles Curley]

On Fri, 27 Oct 2023 14:46:52 +
Minecraftchest1  wrote:

> With Virt-Manager, you should have the option to choose an existing
> disk image.

It can also create a disk image for you. On which you will have to make
partitions and file systems.

-- 
Does anybody read signatures any more?

https://charlescurley.com
https://charlescurley.com/blog/

Re: Which Virtual Manager? Was: EASY way to install packages from trixie/sid to stable?

[The Wanderer]

On 2023-10-27 at 10:46, Minecraftchest1 wrote:

> With Virt-Manager, you should have the option to choose an existing 
> disk image.

That only helps if you've already created a disk image, which will not
be the case when creating a new VM from scratch. Having to resort to the
command line (or to other tools) to create the initial disk image - even
if, potentially, just creating an empty file (or a file of specified
size, filled with zeroes) would work - is not as friendly or as
straightforward a workflow as being able to do it from the GUI.

> In that dialog, you can create an image in any of the pools (you can
> also add pools in that dialog), and that will let you change the file
> name and disk size.

Hold up. Where do "pools" (which I'm guessing is short for "storage
pools") come into things? The other virtualization solutions I've seen
and worked with (short of full-system-level things, such as I understand
VMWare ESXi to be) don't require being aware of or handling storage
pools; they work with disk image files (or, for the case of hierarchical
snapshots, cascading stacks thereof) directly, and do not require those
files to be part of any "storage pool" in any way that the user needs to
be aware of.

If you're starting with the assumption that "storage pools" will - much
less need to - be involved somewhere, you're already not matching the
convenience etc. of the workflow I know from those other tools.

Just offhand, I would expect that a "storage pools" paradigm would block
off some of the convenient things that can be done in that other
workflow, such as being able to move or copy a virtual machine by moving
or copying the directory that its files (disk images, configuration
files, et cetera) are stored in - because you'd also have to fiddle with
whatever it is that defines the "storage pool" that those files are part
of, and that definition would presumably be outside of the directory
that defines the virtual machine.

> I am not ay my laptop currently, but I can take and share some 
> screenshots later today.

Regardless of the above, that might be useful.

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw



signature.asc
Description: OpenPGP digital signature

Re: Which Virtual Manager? Was: EASY way to install packages from trixie/sid to stable?

[Minecraftchest1]

With Virt-Manager, you should have the option to choose an existing disk image. 
In that dialog, you can create an image in any of the pools (you can also add 
pools in that dialog), and that will let you change the file name and disk size.

I am not ay my laptop currently, but I can take and share some screenshots 
later today.

On October 27, 2023 9:17:46 AM UTC, The Wanderer  wrote:
>On 2023-10-26 at 15:28, Andrew M.A. Cater wrote:
>
>> Apt-get install virt-manager will pull in all the associated
>> qemu/KVM packages you might need. It should be at least as
>> straightforward to use as Virtualbox.
>
>I've seen people state or suggest multiple times that virt-manager
>should be, as you say, as straightforward to use as VirtualBox, and that
>was what I expected to find when I tried to use it myself - but even
>once I got around the issues that arose from my not using systemd et
>cetera and could actually use the program, it is not what I actually found.
>
>(I very much hope that what I am about to describe is wrong, and that
>people will explain how/why it's wrong, such that I can get out of the
>situation described and into a state where I can actually use
>virt-manager in a way that I could find useful. It is not my intention
>to spread FUD or falsehoods, even if some of the below may look like it
>would fall within those categories.)
>
>
>From what I recall from having used VirtualBox in the past, its workflow
>is fairly similar to what I see in using VMWare Workstation in a
>(work-related) Windows environment. When you create a new VM, it prompts
>you for where the VM's files should be stored, and then for details
>about the VM's configuration (disk sizes, hardware devices, et cetera),
>and optionally lets you specify what will be done to install the OS that
>will run in the VM - and then with that done, the VM is ready, and you
>can boot it up or create a snapshot (using a graphical
>snapshot-management interface) or make further configuration edits or do
>whatever else you will with it.
>
>With virt-manager, from what I recall (it has been a while since I last
>tried), the workflow was quite different. IIRC, I didn't even try using
>qemu as a backend, because AFAIK it doesn't support hierarchical VM
>snapshots and that's a feature I very much expect to rely on; instead I
>think I went with KVM. With that backend, AFAIR I didn't even get
>prompted for where the VM's file should be stored; instead, the location
>where the system stores files appears to be defined in a system-wide
>config file, and to not be modifiable on a per-VM basis (except relative
>to that system-wide root). That's a problem, because when I partitioned
>this system I expected to be able to store VM files in the same massive
>data partition as I allocated for other large data; the default
>system-wide location doesn't have the space to do much with. It also
>doesn't work when the system may have multiple users who may want to
>manage VMs separately from one another (though, fortunately, this is
>more an abstract concern rather than one that affects me in practice).
>
>With VMWare Workstation and what I think I I recall from VirtualBox,
>once a VM is created, the resulting files are owned by the user who ran
>the program. With what I recall from when I tried virt-manager, even if
>I redirected the file storage location to be under the larger data
>partition, the files were owned by another user, related to libvirt.
>That's undesirable when trying to store VM files per-user in a per-user
>location, since the user won't be able to work with them (moving them
>around, editing details, etc.) except through programs running with that
>other user's access.
>
>When I accepted that and tried to proceed anyway, for the sake of
>experimentation, IIRC, I ran into obstacles trying to set up the
>necessary virtual hardware for the VM - in particular, IIRC, a virtual
>CD drive pointed at the ISO that would be used to install the OS. (This
>part I am less certain about than even the above; it's been rather a
>while, and I was stressed enough by the time I hit this point that I may
>have blanked out more of the details in self-defense.) At that point, I
>gave up, at least in part for the sake of not piling more and more
>stress on myself trying to get the ability to do things that would
>hopefully enable me to reduce stress in other areas.
>
>(Writing this mail is already bringing back up all that stress, and I
>hope it will not just wind up making things worse.)
>
>
>So... either I somehow have managed to do things *100% completely
>wrong*, or the workflow with virt-manager is not even remotely as
>straightforward(ly usable) as the

Re: Which Virtual Manager? Was: EASY way to install packages from trixie/sid to stable?

[Andrew M.A. Cater]

On Fri, Oct 27, 2023 at 05:17:46AM -0400, The Wanderer wrote:
> On 2023-10-26 at 15:28, Andrew M.A. Cater wrote:
> 
> > Apt-get install virt-manager will pull in all the associated
> > qemu/KVM packages you might need. It should be at least as
> > straightforward to use as Virtualbox.
> 
> I've seen people state or suggest multiple times that virt-manager
> should be, as you say, as straightforward to use as VirtualBox, and that
> was what I expected to find when I tried to use it myself - but even
> once I got around the issues that arose from my not using systemd et
> cetera and could actually use the program, it is not what I actually found.
> 
> (I very much hope that what I am about to describe is wrong, and that
> people will explain how/why it's wrong, such that I can get out of the
> situation described and into a state where I can actually use
> virt-manager in a way that I could find useful. It is not my intention
> to spread FUD or falsehoods, even if some of the below may look like it
> would fall within those categories.)
> 

People's experience varies: I've used Virtualbox in the past, Hyper-V under
Windows and VMWare Workstation - Virt-manager provides as easy a front end to
 KVM for me.

It does help if you select the "Customise configuration" option before 
beginning install at step 5 of 5


> 
> With virt-manager, from what I recall (it has been a while since I last
> tried), the workflow was quite different. IIRC, I didn't even try using
> qemu as a backend, because AFAIK it doesn't support hierarchical VM
> snapshots and that's a feature I very much expect to rely on; instead I
> think I went with KVM. With that backend, AFAIR I didn't even get
> prompted for where the VM's file should be stored; instead, the location
> where the system stores files appears to be defined in a system-wide
> config file, and to not be modifiable on a per-VM basis (except relative
> to that system-wide root). That's a problem, because when I partitioned
> this system I expected to be able to store VM files in the same massive
> data partition as I allocated for other large data; the default
> system-wide location doesn't have the space to do much with. It also
> doesn't work when the system may have multiple users who may want to
> manage VMs separately from one another (though, fortunately, this is
> more an abstract concern rather than one that affects me in practice).
> 
> With VMWare Workstation and what I think I I recall from VirtualBox,
> once a VM is created, the resulting files are owned by the user who ran
> the program. With what I recall from when I tried virt-manager, even if
> I redirected the file storage location to be under the larger data
> partition, the files were owned by another user, related to libvirt.
> That's undesirable when trying to store VM files per-user in a per-user
> location, since the user won't be able to work with them (moving them
> around, editing details, etc.) except through programs running with that
> other user's access.
> 
> When I accepted that and tried to proceed anyway, for the sake of
> experimentation, IIRC, I ran into obstacles trying to set up the
> necessary virtual hardware for the VM - in particular, IIRC, a virtual
> CD drive pointed at the ISO that would be used to install the OS. (This
> part I am less certain about than even the above; it's been rather a
> while, and I was stressed enough by the time I hit this point that I may
> have blanked out more of the details in self-defense.) At that point, I
> gave up, at least in part for the sake of not piling more and more
> stress on myself trying to get the ability to do things that would
> hopefully enable me to reduce stress in other areas.
> 
See the "configure before install" which opens this up more - you can
also see this by viewing/modifying settings - but you normally have to make
sure that the VM is shut down.

> (Writing this mail is already bringing back up all that stress, and I
> hope it will not just wind up making things worse.)
> 
> 
> So... either I somehow have managed to do things *100% completely
> wrong*, or the workflow with virt-manager is not even remotely as
> straightforward(ly usable) as the one I see with VMWare Workstation and
> think I remember seeing with VirtualBox.
> 
> I would *love* to be wrong about that, because there is a *lot* of stuff
> that I'd like to do that would be *far* easier if I had discardable VM
> snapshots to do it in. However, I also do not have the personal stress
> to spare for experimenting with this blindly and bashing my head against
> walls getting nowhere in those experiments.
> 
> If there *is* a way to 

Re: Which Virtual Manager? Was: EASY way to install packages from trixie/sid to stable?

[The Wanderer]

On 2023-10-26 at 15:28, Andrew M.A. Cater wrote:

> Apt-get install virt-manager will pull in all the associated
> qemu/KVM packages you might need. It should be at least as
> straightforward to use as Virtualbox.

I've seen people state or suggest multiple times that virt-manager
should be, as you say, as straightforward to use as VirtualBox, and that
was what I expected to find when I tried to use it myself - but even
once I got around the issues that arose from my not using systemd et
cetera and could actually use the program, it is not what I actually found.

(I very much hope that what I am about to describe is wrong, and that
people will explain how/why it's wrong, such that I can get out of the
situation described and into a state where I can actually use
virt-manager in a way that I could find useful. It is not my intention
to spread FUD or falsehoods, even if some of the below may look like it
would fall within those categories.)


From what I recall from having used VirtualBox in the past, its workflow
is fairly similar to what I see in using VMWare Workstation in a
(work-related) Windows environment. When you create a new VM, it prompts
you for where the VM's files should be stored, and then for details
about the VM's configuration (disk sizes, hardware devices, et cetera),
and optionally lets you specify what will be done to install the OS that
will run in the VM - and then with that done, the VM is ready, and you
can boot it up or create a snapshot (using a graphical
snapshot-management interface) or make further configuration edits or do
whatever else you will with it.

With virt-manager, from what I recall (it has been a while since I last
tried), the workflow was quite different. IIRC, I didn't even try using
qemu as a backend, because AFAIK it doesn't support hierarchical VM
snapshots and that's a feature I very much expect to rely on; instead I
think I went with KVM. With that backend, AFAIR I didn't even get
prompted for where the VM's file should be stored; instead, the location
where the system stores files appears to be defined in a system-wide
config file, and to not be modifiable on a per-VM basis (except relative
to that system-wide root). That's a problem, because when I partitioned
this system I expected to be able to store VM files in the same massive
data partition as I allocated for other large data; the default
system-wide location doesn't have the space to do much with. It also
doesn't work when the system may have multiple users who may want to
manage VMs separately from one another (though, fortunately, this is
more an abstract concern rather than one that affects me in practice).

With VMWare Workstation and what I think I I recall from VirtualBox,
once a VM is created, the resulting files are owned by the user who ran
the program. With what I recall from when I tried virt-manager, even if
I redirected the file storage location to be under the larger data
partition, the files were owned by another user, related to libvirt.
That's undesirable when trying to store VM files per-user in a per-user
location, since the user won't be able to work with them (moving them
around, editing details, etc.) except through programs running with that
other user's access.

When I accepted that and tried to proceed anyway, for the sake of
experimentation, IIRC, I ran into obstacles trying to set up the
necessary virtual hardware for the VM - in particular, IIRC, a virtual
CD drive pointed at the ISO that would be used to install the OS. (This
part I am less certain about than even the above; it's been rather a
while, and I was stressed enough by the time I hit this point that I may
have blanked out more of the details in self-defense.) At that point, I
gave up, at least in part for the sake of not piling more and more
stress on myself trying to get the ability to do things that would
hopefully enable me to reduce stress in other areas.

(Writing this mail is already bringing back up all that stress, and I
hope it will not just wind up making things worse.)


So... either I somehow have managed to do things *100% completely
wrong*, or the workflow with virt-manager is not even remotely as
straightforward(ly usable) as the one I see with VMWare Workstation and
think I remember seeing with VirtualBox.

I would *love* to be wrong about that, because there is a *lot* of stuff
that I'd like to do that would be *far* easier if I had discardable VM
snapshots to do it in. However, I also do not have the personal stress
to spare for experimenting with this blindly and bashing my head against
walls getting nowhere in those experiments.

If there *is* a way to get virt-manager to support a VMWare( and, I
think, VirtualBox)-like workflow - with support for hierarchical nested
snapshots, and graphical management thereof, among other things - and
have things more-or-less Just Work, I would *love* to learn about