On Tue, Nov 12, 2019, at 7:50 PM, Adam Williamson wrote:
On Mon, 2019-11-11 at 12:07 +0100, Kevin Kofler wrote:
> Samuel Sieb wrote:
>
> > On 11/10/19 1:52 PM, Neal Becker wrote:
> > > Well it does work if you do
> > >
> > > dnf --disablerepo=rpmfusion-free,rpmfusion-nonfree download --
> > > enablerepo=rawhide --source mercurial
> > >
> > > But that is not exactly obvious, so I think it's something dnf could
> > > handle better (I hesitate to call it a bug).
> >
> > Wow, that is very non-obvious.
>
> How is it non-obvious? --source enables the source repositories
> corresponding to the binary repositories you have enabled, so if you don't
> want it to enable rpmfusion-free-source, you have to disable the
> rpmfusion-free (not rpmfusion-free-source) repository. Sounds quite obvious
> to me.
But the user's intent is clear and the code almost certainly *could*
interpret it, it just doesn't right now. All it has to do is (re-
)consider the repo(s) that were disabled via user config *after* the
"infer the source repos to enable from the enabled binary repos" flow
has happened. I'd call it a bug, or at least a reasonable feature
request.
I agree it's s feature request. It's a case of "I recognize the correct
answer when it's shown to me"... after seeing this thread, it's obvious
that/why/how it works, but I wouldn't have necessarily thought of it myself.
V/r,
James Cassell