Re: Default MTA and dependencies
On Sat, 3 Nov 2001, Eray Ozkural (exa) wrote:
> On Saturday 03 November 2001 11:44, Bdale Garbee wrote:
> > That's sufficient and correct. Depending on just a virtual package should
> > be fine as along as at least one real package providing that virtual
> > package is available.
> What's the purpose of virtual, otherwise? If it is used in the sense of
> object-orientation, then it is used to abstract over a set of objects (or
> classes if you prefer) whose names/implementations are not known to the
> programmers who specify or use the virtual objects (classes).
> I'd be much more comfortable with just virtual dependencies where possible
> unless there is a preference.
Why do you thing it's more comfortable if a dependency on a virtual
package does provide an additional "If you have no clue which of the
packages that provide this virtual package you want to install I'd duggest
you install package abc."? This doesn't weaken the dependency and I don't
see any problems for a user who wants to use another package that provides
this virtual package instead - it's simply a service that makes it more
comfortable for users that are happy when the package they want to install
works but don't want to think about the different packages that provide
this virtual package.
Get my GPG key: finger email@example.com | gpg --import
Fingerprint: B29C E71E FE19 6755 5C8A 84D4 99FC EA98 4F12 B400