On Tue, Jun 19, 2012 at 08:07:03AM +0200, Apelete Seketeli wrote:
> Le 19-Jun-12, Martin Jansa a écrit:
> > On Tue, Jun 19, 2012 at 01:42:10AM +0200, Apelete Seketeli wrote:
> > > This fix regenerates the gtk+-2.20.1/no-demos.patch file because the
> > > targeted Makefile.am file changed since gtk+-2.10.1.
> > > The gtk+_2.20.1.bb recipe is also updated to fix an issue where
> > > no-demos.patch was not taken into account during the do_patch() stage.
> > > 
> > > Signed-off-by: Apelete Seketeli <apelete@seketeli.net>
> > > ---
> > >  recipes/gtk+/gtk+-2.20.1/no-demos.patch |    7 ++++---
> > >  recipes/gtk+/gtk+_2.20.1.bb             |    3 +--
> > >  2 files changed, 5 insertions(+), 5 deletions(-)
> > > 
> > > diff --git a/recipes/gtk+/gtk+-2.20.1/no-demos.patch b/recipes/gtk+/gtk+-2.20.1/no-demos.patch
> > > index 0fc4c48..ac0730d 100644
> > > --- a/recipes/gtk+/gtk+-2.20.1/no-demos.patch
> > > +++ b/recipes/gtk+/gtk+-2.20.1/no-demos.patch
> > > @@ -1,7 +1,8 @@
> > > ---- gtk+-2.10.1/Makefile.am.orig	2006-08-08 12:37:30.000000000 +0100
> > > -+++ gtk+-2.10.1/Makefile.am	2006-08-08 12:37:48.000000000 +0100
> > > -@@ -1,6 +1,6 @@
> > > +--- gtk+-2.20.1/Makefile.am.orig	2012-06-18 22:40:19.348040648 +0200
> > > ++++ gtk+-2.20.1/Makefile.am	2012-06-18 22:40:41.272149318 +0200
> > > +@@ -1,7 +1,7 @@
> > >   ## Makefile.am for GTK+
> > > + include $(top_srcdir)/Makefile.decl
> > >   
> > >  -SRC_SUBDIRS = gdk-pixbuf gdk gtk modules demos tests perf contrib
> > >  +SRC_SUBDIRS = gdk-pixbuf gdk gtk modules tests perf contrib
> > > diff --git a/recipes/gtk+/gtk+_2.20.1.bb b/recipes/gtk+/gtk+_2.20.1.bb
> > > index 9cffee1..6e3c291 100644
> > > --- a/recipes/gtk+/gtk+_2.20.1.bb
> > > +++ b/recipes/gtk+/gtk+_2.20.1.bb
> > > @@ -1,10 +1,9 @@
> > >  require gtk+.inc
> > >  
> > > -SRC_URI_append_virtclass-native = " file://no-demos.patch \
> > > -"
> > >  SRC_URI_append = "file://gtk-dnd-grab-deadlock-fix.patch \
> > >  		  file://cross-nm.patch \
> > >  		  file://xkb-ifdef.patch \
> > > +		  file://no-demos.patch \
> > 
> > I guess this was because demos were working for target case and were
> > intentionally disabled only for native..
> 
> In my oe setup I did not build gtk+-native. I added gtk+-native to the
> ASSUME_PROVIDED list in local.conf instead, and installed
> libgtk2.0-dev on my host machine.
> As a consequence, I faced this gtk+ issue while baking a target
> package, not a native one.
> 
> Do you mean I shouldn't have fixed the issue the way I did ? What
> would be the best course of action then ?

As you can see in
http://git.openembedded.org/openembedded/commit/recipes/gtk+?id=a3b0920d0cc2cecac9eea4993ad18a64b207d7af
and
http://git.openembedded.org/openembedded/commit/recipes/gtk+?id=a3b0920d0cc2cecac9eea4993ad18a64b207d7af

no-demos.patch is intentionally applied only for gtk+-native and even
correctly packaged for target. So if you have issues building
gtk+_2.20.1.bb then rebasing gtk+-2.20.1/no-demos.patch is right, but
moving that patch from SRC_URI_append_virtclass-native to SRC_URI_append
is wrong.

BTW: I'm not using oe-classic at all, and I'm still reading oe-devel ML, please 
don't put me to To: directly, Thanks.

Cheers,

> 
> Regards.
> -- 
>         Apelete

-- 
Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com