Re: Fwd: 16/35: gnu: r-ggplot2: Update to 2.2.0.

[Ben Woodcroft]

On 23/01/17 18:57, Ricardo Wurmus wrote:
> Ben Woodcroft <address@hidden> writes:
>
> > > -------- Forwarded Message --------
> > > Subject:        16/35: gnu: r-ggplot2: Update to 2.2.0.
> > > Date:   Thu, 29 Dec 2016 20:49:25 +0000 (UTC)
> > > From:   Ricardo Wurmus <address@hidden>
> > > Reply-To:       Ricardo Wurmus <address@hidden>
> > > To:     address@hidden
> > >
> > >
> > >
> > > rekado pushed a commit to branch master
> > > in repository guix.
> > >
> > > commit ab01f1cdb11520f0d56249719bcf627cad802bc2
> > > Author: Ricardo Wurmus <address@hidden>
> > > Date:   Thu Dec 29 20:51:26 2016 +0100
> > >
> > >       gnu: r-ggplot2: Update to 2.2.0.
> > >
> > >       * gnu/packages/statistics.scm (r-ggplot2): Update to 2.2.0.
> > >       [propagated-inputs]: Remove r-proto and r-svglite; add r-lazyeval and
> > >       r-tibble.
> > Is there are particular reason why svglite was removed from propagated-inputs? In 
> > response I'm now unable to save SVG files with "ggsave", unless I install 
> > r-svglite separately.
> >
> > Perhaps this is a consequence of the updater suggesting some inputs can be 
> > removed?
> It is.  My apologies!
>
> “r-svglite” is optional and was reported by my modified updater, which
> compares the current inputs with the inputs a fresh import would
> suggest.  If optional inputs are added it would be good to comment on
> why they were added.  I probably wouldn’t have removed them in that
> case.

OK, np, pushed as e4b6c565beef27a1ce2203a9bc119e84a126e5e0.

I wonder if the importer could import the previous version again and 
compare to those, rather than just the current definition to avoid other 
situations like this.

ben