On Thu, Dec 19, 2019 at 03:06:27PM +0100, Leo Prikler wrote:
> Am Donnerstag, den 19.12.2019, 15:38 +0200 schrieb Efraim Flashner:
> > On Thu, Dec 19, 2019 at 01:44:47PM +0100, Leo Prikler wrote:
> > > Am Donnerstag, den 19.12.2019, 12:00 +0100 schrieb Leo Prikler:
> > > > Am Donnerstag, den 19.12.2019, 11:20 +0200 schrieb Efraim
> > > > Flashner:
> > > > > What does this need glib and glib:bin for? Is it just for
> > > > > building
> > > > > the
> > > > > schemas or does it actually need it at runtime?
> > > > To be honest, I'm not quite sure.  I've copied this part from my
> > > > dash-
> > > > to-dock extension, wherein I copied it from the gnome-shell-
> > > > extensions
> > > > package.
> > > > 
> > > > As far as I'm aware both packages do build schemas, but I'm not
> > > > sure
> > > > how extensions handle them at runtime.  Perhaps this is already
> > > > wrong
> > > > in the package I originally copied the snippet from.  I'll try to
> > > > see
> > > > how far I can get with depropagation.
> > > > 
> > > > Regards,
> > > > Leo
> > > Upon closer inspection, it appears depropagation is indeed
> > > possible. 
> > > See the attached patch.
> > 
> > It makes sense to me that glib:bin should be a native-input but I
> > assume
> > glib, if it's needed at runtime, would probably need to be propagated
> > since the extension doesn't refer to it. Likely it's getting glib
> > from
> > another package in the environment.
> I'm not really sure, what the correct thing would be here.  I did have
> very weird behaviour with hide-app-icon, where having glib as input vs.
> not having it made the difference of having to reload the extension
> after changing settings vs. not having to.  However, I'm really not
> sure whether that was due to the extension or a weirdness in GNOME
> Shell.
> 

In any case, I've pushed the patch with glib:bin in native-inputs and
glib in propagated-inputs.

> > Also, the depropagation patch should really be a separate patch for
> > each
> > package if we go that route.
> Perhaps.  I just wanted to "prove" that you can at least depropagate
> glib:bin and probably glib as well.
> 

-- 
Efraim Flashner   <efraim@flashner.co.il>   אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted