On Thu, 2020-02-20 at 12:56 +0100, Martin Hundebøll wrote: > On 19/02/2020 16.23, Richard Purdie wrote: > > Add an option to avoid builds depending on the presence of setcap > > from the host system. > > > > Signed-off-by: Richard Purdie <[email protected]> > > --- > > .../gstreamer/gstreamer1.0/capfix.patch | 37 +++++++++++++++++++ > > .../gstreamer/gstreamer1.0_1.16.1.bb | 2 + > > 2 files changed, 39 insertions(+) > > create mode 100644 > > meta/recipes-multimedia/gstreamer/gstreamer1.0/capfix.patch > > > > diff --git a/meta/recipes-multimedia/gstreamer/gstreamer1.0/capfix.patch > > b/meta/recipes-multimedia/gstreamer/gstreamer1.0/capfix.patch > > new file mode 100644 > > index 00000000000..7ca3d5ad4a6 > > --- /dev/null > > +++ b/meta/recipes-multimedia/gstreamer/gstreamer1.0/capfix.patch > > @@ -0,0 +1,37 @@ > > +Currently gstreamer configuration depends on whether setcap is found on > > the host > > +system. Turn this into a configure option to make builds deterinistic. > > + > > +RP 2020/2/19 > > +Signed-off-by: Richard Purdie <[email protected]> > > +Upstream-Status: Pending > > + > > +Index: gstreamer-1.16.1/libs/gst/helpers/meson.build > > +=================================================================== > > +--- gstreamer-1.16.1.orig/libs/gst/helpers/meson.build > > ++++ gstreamer-1.16.1/libs/gst/helpers/meson.build > > +@@ -73,7 +73,12 @@ if have_ptp > > + endif > > + endif > > + > > +- setcap = find_program('setcap', '/usr/sbin/setcap', '/sbin/setcap', > > required : false) > > ++ setcap_feature = get_option('setcap') > > ++ if setcap_feature.disabled() > > ++ setcap = find_program('dontexist', required : false) > > ++ else > > ++ setcap = find_program('setcap', '/usr/sbin/setcap', '/sbin/setcap', > > required : false) > > ++ endif > > I think this can be simplified by using get_option() directly for the > "required" argument: > > setcap = find_program('setcap', '/usr/sbin/setcap', '/sbin/setcap', > required : get_option('setcap'))
Yes, that looks better. I'm no meson expert, this is the first time I've touched it. I'd love if it someone was able to get this resolved with upstream. I've been trying to work on getting various things fixed upstream (e.g. some of the perl changes) but I'm having to take a practical view in some areas. Getting the autobuilder stable/working sanely is currently more of a priority than getting involved in a discussion with upstream on something I know little about :/. I've merged this patch but if anyone wants to improve it, I'm fine with that. Cheers, Richard -- _______________________________________________ Openembedded-core mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-core
