Bug#642757: apt-listbugs

[Andreas Beckmann]

On 2011-10-17 21:37, Francesco Poli wrote:
> I hope I expressed myself clearly enough.
> Maybe I should add the above explanation
> to /usr/share/doc/apt-listbugs/README.Debian.gz
> Do you think that it would help in understanding why apt-listbugs
> ignores the "affects" field?

I think it could help.

It's some timing issue with the affects and what gets upgraded when
In the following example oldpackage=nvidia newpackage=xorg

- old versions of oldpackage and newpackage are installed and work well
- newpackage gets updated
- newpackage works fine unless oldpackage is installed
- newpackage breaks oldpackage, bugs are files to newpackage and
  reassigned to oldpackage with affects newpackage
- no upgrade for oldpackage is available, eventually oldpackage is
  removed from testing
- people continue to upgrade newpackage, apt-list-bugs does not see the
  problem due to affects


In the this case its a "complex relationship" that causes the bug and
can't be modeled currently:

   nvidia-graphics-drivers
   AND xserver-xorg-core (>= 2:1.10.99)
   AND PCIID(nvidia gpu AND NOT recent gtx 2xx or fermi based gpu)

Also adding a Breaks: xserver-xorg-core (>= 2:1.10.99)
to nvidia-graphics-drivers does not work because
* the driver works well with recent cards
* it will just prevent upgrades from buggy configuration without
  declared conflict to buggy configuration with declared conflict
  which does not help at all.

Doing a "dummy upload" of nvidia-graphics-drivers won't help either as
not upgrading to that dummy won't prevent the problem, it's already there.


Andreas