> Am Donnerstag, den 02.12.2021, 20:16 +0000 schrieb Jaft:
> > > Am Donnerstag, den 02.12.2021, 02:10 +0000 schrieb Jaft:
> > > > I had noticed that the core-updates-frozen branch had been merged
> > > > so
> > > > I upgraded but found things pretty much the same as before.
> > > Please come back, you're within the wrong timeline.
> >
> > Oh, I don't mean that I used another branch; I saw it got merged to
> > master (I believe) so I just did a guix pull and then guix upgrade.
> > I'm still using stable. Sorry about the confusion!
> I am jokingly referring to the fact that core-updates-frozen is not yet
> merged to master. If you do live two years in the future, please tell
> me the lotto numbers. I need them before I die.
Ohhh; haha. Now I get it. Welp; seems I must've misread something, somewhere.
> > > > I saw an old patch (
> > > > http://git.savannah.gnu.org/cgit/guix.git/commit/?id=e311ef4f87f7ad8db2114e5f89961eea0240893b
> > > > ) and, while I'd checked rofi for gdk-pixbuf+svg – before –, –
> > > > somehow – it made me think to check librsvg, this time, and found
> > > > that it was using gdk-pixbuf, rather than gdk-pixbuf+svg. I then
> > > > made a package inheriting librsvg but using gdk-pixbuf+svg,
> > > > instead, and made a package which inherited rofi but used my
> > > > librsvg package and, with that installed, rofi worked with .svgs,
> > > > then.
> > > >
> > > > Am I right in assuming librsvg ought to be using the latter, as
> > > > the library deals directly with handling SVGs? If so, I can put
> > > > together a patch to submit.
> > > Have you checked using gdk-pixbuf+svg as input to rofi directly? I
> > > don't see why we would have to go in circles for librsvg, the
> > > component you're trying to use is gdk-pixbuf.
> >
> > I just checked and it does; I was going off of the package formation
> > in Guix but, checking the listed dependencies on the rofi GitHub page
> > (
> > https://github.com/davatorium/rofi/blob/next/INSTALL.md#external-libraries
> > ), it does list gdk-pixbuf as one so, perhaps, it makes more sense to
> > build with that instead of librsvg.
> >
> > I had assumed the package inputs for rofi were already accurate and,
> > if gdk-pixbuf doesn't have SVG support while gdk-pixbuf+svg does, it
> > seemed plausible that gdk-pixbuf+svg would be the preferred package
> > for librsvg as librsvg is dealing with SVGs, perhaps part of the
> > reason for SVG icons not getting rendered in applications like
> > Thunar, XFCE, etc. (that being said, I'm unfamiliar with the librsvg
> > code so, perhaps, this assumption of how the gdk-pixbuf dependency is
> > being used is incorrect, on my part).
> Writing a short letter takes time. So to summarize, librsvg is not
> actually a dependency of rofi, gdk-pixbuf (with SVG support) is.
> Anything missing?
I believe that's accurate but I don't have much of any experience with work such as this so I was including my reasoning, in case I was off or misguided at all. That being said, I'd hazard that yours is an accurate summary, in totality.
> > In any case, librsvg is not listed as a dependency for rofi while
> > gdk-pixbuf is and swapping librsvg for gdk-pixbuf+svg in the rofi
> > package still seemed to build it alright (and render SVGs) so, at
> > least directly for rofi, directly using dgk-pixbuf+svg would still
> > solve the SVG issue for it.
>
> Now that that's cleared up, you might want to synthesize a patch from
> it. Is there anything else that was swept under the rug and that we'd
> need to actually resolve before closing this bug after fixing rofi?
Taking another look at some of the other programs I'd mentioned, I'd noticed that file-roller and viewnior are also using gdk-pixbuf; switching those inputs to gdk-pixbuf+svg made them render the icons from Papirus so was thinking to make patches for those, as well?
Including gdk-pixbuf+svg as an input for thunar resulted in it being able to fully render icons appropriately, finally, but I couldn't figure out where gdk-pixbuf had been used (neither for thunar nor any dependencies), if at all. I'm assuming that simply adding it as an input, rather than trying to trace if gdk-pixbuf is used elsewhere in thunar's dependency graph, is considered bad practice, right?