Re: license packages

["Michal Suchanek" <hramrach@centrum.cz>]

2009/1/15 Daniel Baumann <daniel@debian.org>:
> schoappied wrote:
>> Now I'm in doubt whether aptitude or apt-get is preferable for a mixed
>> testing/unstable system..
>
> i *personally* use apt-get only, i don't like aptitude.
>
> apart from that i've the impression that aptitude being slower, it has a
> very strange and allegedly not determinstic dependency and recommends
> resolver (in the meaning that it depends on whatever other packages have
> recommends for a  particular package, because aptitude then tries to be
> clever and installs the package that has the 'most recommendet level'
> across the set of installed packages, and not complying to the order of
> alternatives that were given by the maintainer). or in otherwords: it
> tries to be too smart and therefore fails in some cases, which i don't
> like. i prefere a tool that behaves always the same way (even if it is
> not optimal in some cases, then i'll make sure that it behaves in these
> cases as i want it to)...
>

FWIW if the first alternative was always used what use for
alternatives is there?

I have run into a problem with X packages where naming the vmmouse
driver makes aptitude not install any other input drivers.

The X server depends on input-all | input-2, and vmmouse provides
input-2 so there is no need for a keyboard driver, eh?
Further, the vmmouse driver packaging has a few bugs that prevent it
from doing anything either, leaving the user with inputless X. Very
good security of the box after starting X ;-)

If the input-all was always chosen it would make the alternative
dependency useless, and the package would always work. The X people
say not to use aptitude so would apt always choose input-all, even if
explicitly named vmmouse already provides input-2?

If so it makes alternatives pretty much useless for real options, they
would be useful only for packages that are obsoleted, replaced, and
removed from the repository.

Thanks

Michal