Dnia 2014-08-04, o godz. 11:02:27 Michał Górny napisał(a): > The original checks assume that no-blocker is denoted as 'None'. > However, currently atom.blocker returns just 'False'. So, applying this patch restores the 'dependency.unknown' check that wasn't working for some time. As a result, I think it slows down repoman near GNOME & KDE categories (though no real measurements), and gives 4779 new warnings. Some of those warnings concern only || () dependencies, e.g.: x11-wm/awesome/awesome-3.5.5.ebuild: DEPEND: =x11-libs/libX11-1.3.99.901 ) in the ebuild. However, some look like false positives and need further investigation: app-emulation/emul-linux-x86-baselibs/emul-linux-x86-baselibs-20140406-r4.ebuild: RDEPEND: >=app-arch/bzip2-1.0.6-r4[abi_x86_32(-)], >=app-arch/xz-utils-5.0.5-r1[abi_x86_32(-)], >=app-text/libpaper-1.1.24-r2[abi_x86_32(-)], >=dev-db/sqlite-3.8.3:3[abi_x86_32(-)], >=dev-libs/dbus-glib-0.100.2-r1[abi_x86_32(-)], >=dev-libs/elfutils-0.155-r1[abi_x86_32(-)], >=dev-libs/expat-2.1.0-r3[abi_x86_32(-)], >=dev-libs/glib-2.34.3:2[abi_x86_32(-)], >=dev-libs/gmp-5.1.3-r1[abi_x86_32(-)], >=dev-libs/json-c-0.11-r1[abi_x86_32(-)], >=dev-libs/libelf-0.8.13-r2[abi_x86_32(-)], >=dev-libs/libgcrypt-1.5.3-r100:11[abi_x86_32(-)], >=dev-libs/libgcrypt-1.6.1-r1:0[abi_x86_32(-)], >=dev-libs/libgpg-error-1.12-r1[abi_x86_32(-)], >=dev-libs/libpcre-8.33-r1[abi_x86_32(-)], >=dev-libs/libtasn1-3.4-r1[abi_x86_32(-)], >=dev-libs/libxml2-2.9.1-r4[abi_x86_32(-)], >=dev-libs/libxslt-1.1.28-r2[abi_x86_32(-)], >=dev-libs/lzo-2.06-r1[abi_x86_32(-)], >=dev-libs/nettle-2.7.1-r1[abi_x86_32(-)], >=dev-libs/openssl-1.0.1h-r2[abi_x86_32(-)], >=dev-libs/udis86-1.7-r2[abi_x86_32(-)], >=media-libs/giflib-4.2.3-r1[abi_x86_32(-)], >=media-libs/lcms-2.5-r1:2[abi_x86_32(-)], >=media-libs/libpng-1.2.51:1.2[abi_x86_32(-)], >=media-libs/libpng-1.6.10:0[abi_x86_32(-)], >=media-libs/tiff-3.9.7-r1:3[abi_x86_32(-)], >=media-libs/tiff-4.0.3-r6:0[abi_x86_32(-)], >=sys-apps/acl-2.2.52-r1[abi_x86_32(-)], >=sys-apps/attr-2.4.47-r1[abi_x86_32(-)], >=sys-apps/dbus-1.6.18-r1[abi_x86_32(-)], >=sys-apps/file-5.18-r1[abi_x86_32(-)], >=sys-apps/tcp-wrappers-7.6.22-r1[abi_x86_32(-)], >=sys-apps/util-linux-2.24.1-r3[abi_x86_32(-)], >=sys-devel/gettext-0.18.3.2[abi_x86_32(-)], >=sys-devel/libtool-2.4.2-r1[abi_x86_32(-)], >=sys-devel/llvm-3.3-r3[abi_x86_32(-)], >=sys-libs/e2fsprogs-libs-1.42.9[abi_x86_32(-)], >=sys-libs/gdbm-1.10-r1[abi_x86_32(-)], >=sys-libs/gpm-1.20.7-r2[abi_x86_32(-)], >=sys-libs/libavc1394-0.5.4-r1[abi_x86_32(-)], >=sys-libs/libraw1394-2.1.0-r1[abi_x86_32(-)], >=sys-libs/ncurses-5.9-r3[abi_x86_32(-)], >=sys-libs/readline-6.2_p5-r1:0[abi_x86_32(-)], >=sys-libs/slang-2.2.4-r1[abi_x86_32(-)], >=sys-libs/zlib-1.2.8-r1[abi_x86_32(-)] CC-ing multilib@ for consultation :). I don't think this should prevent us from committing the patch, however. If we believe the potential false positives are important enough, we should just add explanatory 'if False' instead of keeping a code bug ;). -- Best regards, Michał Górny