[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#368411: tex-common keeps asking the question about group on each upgrade



Norbert Preining <preining@logic.at> wrote:

> On Mon, 22 Mai 2006, Frank Küster wrote:
>> > on each upgrade of the tex-common package, i'm being asked the same
>> > question over and over again: "Group that should own the TeX font cache"?
>> > It seems that the answer is not properly remembered by the configuration.
>> 
>> I've noticed that, too.  I thought I had properly catered for the
>> renaming:
>> 
>> # The groupname question has been split up, but it need not be shown again
>> db_fget tex-common/groupname seen || true
>> if [ "$RET" = true ]; then
>>  db_fset tex-common/groupname_single seen true || true
>>  db_fset tex-common/groupname_multi seen true || true
>> fi
>> db_unregister tex-common/groupname || true
>
> Although we should check here for groupname_singe and _multi, too, 

Why should we?  There's no point in setting the seen flag of a message
to its current setting, is there?  We don't set any seen flag to false,
nowhere, do we?

> We should check before whether this question has been seen or not...
>
> So it should be
>
> while true; do
>       db_fget tex-common/groupname_$groupname_variant seen || true
>       if [ "$RET" = false ] ; then
>         db_input $cache_debconf_priority tex-common/groupname_$groupname_variant || true
>         db_go || true
>       fi

I guess I'm just blind (frustrated by my measurements going wrong
today...), but:  How should that have effect?  If the question is
already seen, it won't be shown again anyway.  

However, I have some other questions to think about:

- will there ever be a wrong  entry in groupname, i.e. a non-existent
  group, from earlier invocations of the script?  I guess no, but for
  upgraders we cannot be sure.  Does that need special treatment? 

- If the answer given the first time is a non-existent group, the loop
  is executed once more.  However, I expected that the question would
  not be shown again, except when dpkg-reconfigure is used. 

  However, this is not the case:  The question is shown over and over
  again, during a fresh installation, until I enter a new value.

Anyway, if it turns out that Ralf's idea can be implemented without side
effects, dropping the whole stuff completely would be much nicer.

Regards, Frank
-- 
Frank Küster
Single Molecule Spectroscopy, Protein Folding @ Inst. f. Biochemie, Univ. Zürich
Debian Developer (teTeX)



Reply to: