Created attachment 299089 [details] gtk-vnc-0.5.0 build log gtk-vnc's configure script appears to be capable of detecting libpulse and appropriately disabling its support if not present (see attached build log), a USE flag to remove the dependency would be appreciated.
Created attachment 299091 [details, diff] Patch for gtk-vnc-0.5.0.ebuild
(In reply to comment #0) > gtk-vnc's configure script appears to be capable of detecting libpulse and > appropriately disabling its support if not present Yes, and that's a problem, see http://www.gentoo.org/proj/en/qa/automagic.xml To make the dependency controllable by a USE flag, it must be made non-automagic, which in this case requires patching configure.ac and adding an eautoreconf call.
Created attachment 307823 [details] Automagick patch for gtk-vnc
I wrote simple patch for build system, working ebuild can be found in my overlay (layman/maksbotan). If you will not commit it in reasonable time, i will do it myself and report upstream.
(In reply to comment #4) > I wrote simple patch for build system, working ebuild can be found in my > overlay (layman/maksbotan). If you will not commit it in reasonable time, i > will do it myself and report upstream. Report this to upstream just now please -> bugzilla.gnome.org And post the link here
As talked with tetromino, we usually want to provide also an "auto" option that behaves as automagic as is the default option, that way upstream will probably get the patch. Then, would be nice if you could update your patch for providing that option also and send it to upstream to see the bug fixed by them also as soon as possible ;) Thanks
(In reply to comment #6) > As talked with tetromino, we usually want to provide also an "auto" option > that behaves as automagic as is the default option, that way upstream will > probably get the patch. Then, would be nice if you could update your patch > for providing that option also and send it to upstream to see the bug fixed > by them also as soon as possible ;) > > Thanks If you look closely at my patch, you will see that: +AC_ARG_WITH(pulseaudio, +[ --with-pulseaudio enable audio support via pulseaudio], +[case "${withval}" in + yes|no) ;; + *) AC_MSG_ERROR([bad value ${withval} for pulseaudio option]) ;; + esac],[withval=yes]) Last argument ([withval=yes]) is "action-if-not-given" and provides automagick behavior that you talk about. I will send it upstream today.
Here is upstream bug: https://bugzilla.gnome.org/show_bug.cgi?id=673570
(In reply to comment #7) > (In reply to comment #6) > > As talked with tetromino, we usually want to provide also an "auto" option > > that behaves as automagic as is the default option, that way upstream will > > probably get the patch. Then, would be nice if you could update your patch > > for providing that option also and send it to upstream to see the bug fixed > > by them also as soon as possible ;) > > > > Thanks > > If you look closely at my patch, you will see that: > > +AC_ARG_WITH(pulseaudio, > +[ --with-pulseaudio enable audio support via pulseaudio], > +[case "${withval}" in > + yes|no) ;; > + *) AC_MSG_ERROR([bad value ${withval} for pulseaudio option]) ;; > + esac],[withval=yes]) > > Last argument ([withval=yes]) is "action-if-not-given" and provides > automagick behavior that you talk about. I will send it upstream today. But yes should mean that pulseaudio support is forced and, if not found, it should die. For automagic, "auto" should be used instead
(In reply to comment #9) > (In reply to comment #7) > > (In reply to comment #6) > > > As talked with tetromino, we usually want to provide also an "auto" option > > > that behaves as automagic as is the default option, that way upstream will > > > probably get the patch. Then, would be nice if you could update your patch > > > for providing that option also and send it to upstream to see the bug fixed > > > by them also as soon as possible ;) > > > > > > Thanks > > > > If you look closely at my patch, you will see that: > > > > +AC_ARG_WITH(pulseaudio, > > +[ --with-pulseaudio enable audio support via pulseaudio], > > +[case "${withval}" in > > + yes|no) ;; > > + *) AC_MSG_ERROR([bad value ${withval} for pulseaudio option]) ;; > > + esac],[withval=yes]) > > > > Last argument ([withval=yes]) is "action-if-not-given" and provides > > automagick behavior that you talk about. I will send it upstream today. > > But yes should mean that pulseaudio support is forced and, if not found, it > should die. For automagic, "auto" should be used instead Where do you see it dying? Here AC_MSG_NOTICE, not ERROR is used: + PKG_CHECK_MODULES(PULSEAUDIO, libpulse-simple, WANT_PULSEAUDIO="yes", AC_MSG_NOTICE([Not building against libpulse]))
Fixed in 0.5.0-r1.
