Re: gnucash-docs 1.9.0-1 MIGRATED to testing

[Thomas Bushnell BSG <tb@becket.net>]

Steve Langasek <vorlon@debian.org> writes:

> I don't see anything broken in britney's handling of this package. 
> gnucash-docs doesn't depend on gnucash, so this version of gnucash-docs is
> perfectly installable in testing right now -- it's just not co-installable
> with gnucash itself, which is not a requirement.

Hrm, I guess I see the logic here, though it seems like a better
solution could be found than:

> If the package shouldn't have been allowed to progress into testing without
> gnucash, the way to ensure this is by either opening a dummy serious bug on
> gnucash-docs or by making gnucash-docs depend on the corresponding version
> of gnucash instead of merely suggesting it.

A dependency would certainly be wrong, since it doesn't actually
depend.  The point is that the documentation is inconsistent with the
old version of the program.  Of course, a conflict is, strictly
speaking, also wrong.

Maybe what we need is a negative version of "Recommends", that is,
just as Conflicts is the negative version of Depends, perhaps we need
a way to say "it is recommended that you do *not* install X along with
this."

Thomas