On Mon, Apr 18, 2016 at 4:57 AM, James Hogarth <james.hogarth(a)gmail.com> wrote:
On 18 April 2016 at 03:07, Chris Murphy <lists(a)colorremedies.com> wrote:
>
> On Sun, Apr 17, 2016 at 5:04 PM, Jon LaBadie <jonfu(a)jgcomp.com> wrote:
> > On F22, the manpage for dnf says the update and update-to
> > commands are deprecated aliases for upgrade and upgrade-to.
> >
> > That being the case, shouldn't the update-info and
> > check-update commands also be deprecated and have
> > corresponding upgrade versions?
>
> That would be consistent. File a bug?
>
> I think the vernacular change is questionable, though. For a long time
> update means minor, and upgrade means major. And it makes it more
> confusing where upgrade vs system-upgrade. So system is what makes it
> major and not system is what makes it minor. Uhh?
>
> But I'm also too lazy to go look at the change to see what the logic
> is about. I guess from dnf's standpoint this isn't a meaningful
> distinction.
>
> For now I continue to use update.
>
Once upon a time obsoleting was disabled by default and update would only do
non-obsoleting packages and upgrade would include obsolete in the dependency
resolution.
The obsolete was enabled by default rendering the difference between update
and upgrade meaningless in default situations - hence the deprecation.
Ahh, that makes complete sense. Different histories of update vs
upgrade than what I'm used to.
Honestly I'm surprised both were carried forward to dnf rather
than the
deprecated command being dropped at the yum -> dnf boundary.
Yes, dnf would have been a good time for this. I can only guess early
on something important switching over to dnf blew up spectacularly
without it, and it was easier to put in a backward compatible
deprecated subcommand than fight reality.
--
Chris Murphy