Following on from "[gentoo-dev] Reliance upon || ( use? ( ) ) behaviour": the following packages look like they're misusing || ( use? ( ) ) constructs. * If you must use || ( use? ( ) ), which you really shouldn't, you need to combine all use calls with a has_version. But that's horrid, so don't. * The form || ( foo? ( ... ) bar/baz ) is unnecessary if foo is not used to modify build parameters. If your package needs either X or Y, but doesn't care at compile time, use || ( X Y ) with no use flag. app-emulation/fuse-0.7.0 Conditional on 'gtk' inside || ( ) block in RDEPEND Conditional on 'sdl' inside || ( ) block in RDEPEND Conditional on 'svga' inside || ( ) block in RDEPEND Conditional on 'fbcon' inside || ( ) block in RDEPEND Conditional on 'svga' inside || ( ) block in RDEPEND Conditional on 'fbcon' inside || ( ) block in RDEPEND Conditional on 'sdl' inside || ( ) block in RDEPEND herds: no-herd maintainer: <maintainer-needed@gentoo.org> app-text/estraier-1.2.25 Conditional on 'chasen' inside || ( ) block in DEPEND Conditional on 'mecab' inside || ( ) block in DEPEND Conditional on 'kakasi' inside || ( ) block in DEPEND Conditional on 'chasen' inside || ( ) block in RDEPEND Conditional on 'mecab' inside || ( ) block in RDEPEND Conditional on 'kakasi' inside || ( ) block in RDEPEND app-text/estraier-1.2.29-r1 (the same) app-text/estraier-1.2.29 (the same) herds: cjk maintainer: <hattya@gentoo.org> games-arcade/koules-1.4-r2 Conditional on 'svga' inside || ( ) block in DEPEND Conditional on 'tk' inside || ( ) block in DEPEND Conditional on 'svga' inside || ( ) block in RDEPEND Conditional on 'tk' inside || ( ) block in RDEPEND herds: games games-emulation/epsxe-1.5.2-r1 Conditional on 'opengl' inside || ( ) block in RDEPEND games-emulation/epsxe-1.6.0-r3 Conditional on 'opengl' inside || ( ) block in RDEPEND herds: games games-emulation/pcsx-1.5-r1 Conditional on 'opengl' inside || ( ) block in RDEPEND Conditional on 'x86' inside || ( ) block in RDEPEND herds: games games-fps/imaze-1.4 Conditional on 'Xaw3d' inside || ( ) block in DEPEND Conditional on 'Xaw3d' inside || ( ) block in RDEPEND herds: games games-sports/foobillard-3.0a Conditional on 'sdl' inside || ( ) block in DEPEND Conditional on 'sdl' inside || ( ) block in RDEPEND herds: games gnustep-base/gnustep-base-1.10.3-r1 Conditional on 'ffcall' inside || ( ) block in DEPEND Conditional on 'gcc-libffi' inside || ( ) block in DEPEND Conditional on 'ffcall' inside || ( ) block in RDEPEND Conditional on 'gcc-libffi' inside || ( ) block in RDEPEND gnustep-base/gnustep-base-1.10.3 Conditional on 'gcc-libffi' inside || ( ) block in DEPEND Conditional on 'gcc-libffi' inside || ( ) block in RDEPEND gnustep-base/gnustep-base-1.12.0 (the same) gnustep-base/gnustep-base-1.13.0-r1 (the same) herds: gnustep mail-filter/sqlgrey-1.6.7 Conditional on 'postgres' inside || ( ) block in DEPEND Conditional on 'sqlite' inside || ( ) block in DEPEND Conditional on 'mysql' inside || ( ) block in DEPEND Conditional on 'postgres' inside || ( ) block in RDEPEND Conditional on 'sqlite' inside || ( ) block in RDEPEND Conditional on 'mysql' inside || ( ) block in RDEPEND mail-filter/sqlgrey-1.7.4 (the same) herds: net-mail maintainer: <vivo@gentoo.org> net-analyzer/ipac-ng-1.31-r2 Conditional on 'sqlite' inside || ( ) block in DEPEND Conditional on 'postgres' inside || ( ) block in DEPEND Conditional on 'sqlite' inside || ( ) block in RDEPEND Conditional on 'postgres' inside || ( ) block in RDEPEND herds: netmon net-dns/mydns-1.1.0 Conditional on 'mysql' inside || ( ) block in DEPEND Conditional on 'postgres' inside || ( ) block in DEPEND Conditional on 'postgres' inside || ( ) block in DEPEND Conditional on 'mysql' inside || ( ) block in RDEPEND Conditional on 'postgres' inside || ( ) block in RDEPEND Conditional on 'postgres' inside || ( ) block in RDEPEND maintainer: MATSUU Takuto <matsuu@gentoo.org> net-print/hplip-0.9.7-r3 Conditional on 'X' inside || ( ) block in RDEPEND net-print/hplip-1.6.12 Conditional on 'X' inside || ( ) block in DEPEND Conditional on 'X' inside || ( ) block in RDEPEND net-print/hplip-1.7.1 (the same) herds: printing net-print/hpoj-0.91-r3 Conditional on 'X' inside || ( ) block in DEPEND Conditional on 'X' inside || ( ) block in RDEPEND herds: printing sci-libs/gdal-1.2.6-r4 Conditional on 'postgres' inside || ( ) block in DEPEND Conditional on 'mysql' inside || ( ) block in DEPEND Conditional on 'netcdf' inside || ( ) block in DEPEND Conditional on 'hdf' inside || ( ) block in DEPEND Conditional on 'postgres' inside || ( ) block in RDEPEND Conditional on 'mysql' inside || ( ) block in RDEPEND Conditional on 'netcdf' inside || ( ) block in RDEPEND Conditional on 'hdf' inside || ( ) block in RDEPEND sci-libs/gdal-1.3.0-r1 (the same) sci-libs/gdal-1.3.1 Conditional on 'netcdf' inside || ( ) block in DEPEND Conditional on 'hdf' inside || ( ) block in DEPEND Conditional on 'netcdf' inside || ( ) block in RDEPEND Conditional on 'hdf' inside || ( ) block in RDEPEND sci-libs/gdal-1.3.2 (the same) herds: sci-geosciences maintainer: Steve Arnold <nerdboy@gentoo.org>
games done
netmon done.
mydns done.
hplip fixed, hpoj is asked for removal
estraier is p.masked for removal.
For as far as I understand this bug, gnustep-base uses the || ( use? ( ) ) construct legitimately, because it builds against one chosen implementation of a foreign function call library. It is not just a runtime dependency. Please re-add gnustep if I didn't get the point.
No, gnustep gets it wrong, since it uses the use flag to control behaviour: pkg_setup() { »···if use gcc-libffi; then It is possible that gcc-libffi is enabled but the || ( ) block has been satisfied by dev-libs/ffcall. For that matter, abusing gcc-config like that is pretty horrible...
(In reply to comment #7) > It is possible that gcc-libffi is enabled but the || ( ) block has been > satisfied by dev-libs/ffcall. Ok, do I get close when I think that the DEPEND should not have the || but only the use? That way it is forced to have the right library implementation for the chosen method. ... ffcall? ( dev-libs/ffcall ) gcc-libffi? ( >=sys-devel/gcc-3.3.5 ) !ffcall? ( !gcc-libffi? ( >=dev-libs/libffi-3.3.5 ) ) ...
That's an acceptable solution, yes.
Ok, thanks for the insight. Should be fixed now.
Still unfixed: app-emulation/fuse-0.7.0 maintainer: <maintainer-needed@gentoo.org> mail-filter/sqlgrey-1.6.7 mail-filter/sqlgrey-1.7.4 herds: net-mail sci-libs/gdal-1.3.1 sci-libs/gdal-1.3.2 sci-libs/gdal-1.4.2 herds: sci-geosciences maintainer: Steve Arnold <nerdboy@gentoo.org> Ping!
adding darkside@g.o for fuse
(In reply to comment #12) > adding darkside@g.o for fuse > Adding the real maintainer for app-emulation/fuse. ;)
Gdal fixed from version 1.7.
net-mail done.
All packages done, closing.