1 |
Le samedi 22 avril 2017 à 23:03 +0200, Michał Górny a écrit : |
2 |
> On sob, 2017-04-22 at 13:43 -0400, Mike Gilbert wrote: |
3 |
> > |
4 |
> > On Sat, Apr 22, 2017 at 4:20 AM, Michał Górny <mgorny@g.o> |
5 |
> > wrote: |
6 |
> > > |
7 |
> > > Respect EROOT when running gtk-query-immodules-* tools, alike |
8 |
> > > other |
9 |
> > > updaters in the eclass. This ensures that x11-libs/gtk+ installs |
10 |
> > > correctly when installing to a ROOT. |
11 |
> > |
12 |
> > I'm not an expert on this eclass, but it seems like this may be |
13 |
> > intentional; calling compiled binaries from $ROOT will break for |
14 |
> > cross-compiles. |
15 |
> > |
16 |
> > Is there some package that installs ${CHOST}-gtk-query-immodules- |
17 |
> > X.0 |
18 |
> > for ROOT == /? I don't have any such binary on my system. |
19 |
> > |
20 |
> |
21 |
> I was thinking about this possibility too but it doesn't look like |
22 |
> any |
23 |
> of the remaining code was done like this, nor the GTK+ ebuild which |
24 |
> calls this function. But I'll wait for the GNOME team to confirm what |
25 |
> they want. |
26 |
> |
27 |
Looking at the ebuilds, it looks like your patch is correct. I honestly |
28 |
cannot remember why we used EPREFIX and not EROOT. |
29 |
|
30 |
-- |
31 |
Gilles Dartiguelongue <eva@g.o> |