I did hit this when creating a patch for www-client/opera. The below patch would make it possible to set GCONF_DEBUG="no" in ebuilds to avoid this behaviour. Intentionally using double negation here, so that only affected ebuilds can be changed, instead of requiring to set GCONF_DEBUG="yes" in all ebuilds that actually have debug feature. Comments? --- gnome2.eclass.orig 2006-04-23 17:36:06.000000000 +0200 +++ gnome2.eclass 2006-05-31 13:10:34.000000000 +0200 @@ -26,7 +26,9 @@ # Path to gconftool-2 GCONFTOOL_BIN=${GCONFTOOL_BIN:="${ROOT}/usr/bin/gconftool-2"} -IUSE="debug" +if [[ ${GCONF_DEBUG} != "no" ]]; then + IUSE="debug" +fi DEPEND=">=sys-apps/sed-4" @@ -40,8 +42,10 @@ gnome2_src_configure() { # Update the GNOME configuration options - use debug && G2CONF="${G2CONF} --enable-debug=yes" - G2CONF="${G2CONF} $(use_enable doc gtk-doc)" + if [[ ${GCONF_DEBUG} != 'no' ]] || use debug ; then + G2CONF="${G2CONF} --enable-debug=yes" + G2CONF="${G2CONF} $(use_enable doc gtk-doc)" + fi # Run libtoolize elibtoolize ${ELTCONF}
And there's of course an error, thanks Mr. Murphy... :P Obviously, I meant: <snip> if [[ ${GCONF_DEBUG} != 'no' ]] || use debug ; then G2CONF="${G2CONF} --enable-debug=yes" fi G2CONF="${G2CONF} $(use_enable doc gtk-doc)" </snip> Sorry.
i don't have any problem with the idea, probably best to do IUSE="${IUSE} debug" just to be safe.
thanks, in cvs.