WARNING: 'use' function invoked by sys-devel/gcc-4.4.7 during metadata regeneration. WARNING: 'use' function invoked by sys-devel/gcc-4.5.4 during metadata regeneration. WARNING: Please ensure upstream is aware of it and working to remove it.
*** This bug has been marked as a duplicate of bug 174407 ***
*** Bug 566562 has been marked as a duplicate of this bug. ***
Let's move to this bug, if you really insist. Nevertheless it's not a duplicate but a potentially dependant bug, if at all. And please keep it open as long as the problem is there; otherwise you can't easily use trackers. Furthermore, an alternative working solution of simple ricer variable was proposed already and ignored, so please don't state that no alternative has been given.
The Council has decided today that the remaining uses of use* in global scope have to be fixed before the next meeting (including existing ebuilds). Please choose one of the solutions proposed so far and implement it before the deadline. Otherwise, QA will proceed as it sees fit. I'm removing the cross-bug dependency as dynamic SLOTs will most likely never be implemented, and certainly not in the requested timeframe.
easy enough to just enable multislot all the time for all the packages
(In reply to SpanKY from comment #6) > easy enough to just enable multislot all the time for all the packages can we not just drop multislot? or just move it to the toolchain overlay?
(In reply to Anthony Basile from comment #7) that doesn't make much sense. the overlay contains old versions & dev snapshots.
(In reply to SpanKY from comment #8) > (In reply to Anthony Basile from comment #7) > > that doesn't make much sense. the overlay contains old versions & dev > snapshots. i was just thinking where to put the current eclass if we wanted to keep USE=multislot eclass *somewhere*. would the following be bad? and similar for toolchain-binutils.eclass. --- toolchain.eclass.orig 2016-02-15 11:50:08.521015223 -0500 +++ toolchain.eclass 2016-02-15 11:51:26.417016533 -0500 @@ -131,7 +131,7 @@ LICENSE="GPL-2+ LGPL-2.1+ FDL-1.1+" fi -IUSE="multislot regression-test vanilla" +IUSE="regression-test vanilla" IUSE_DEF=( nls nptl ) if [[ ${PN} != "kgcc64" && ${PN} != gcc-* ]] ; then @@ -160,7 +160,7 @@ IUSE+=" ${IUSE_DEF[*]/#/+}" # Support upgrade paths here or people get pissed -if ! tc_version_is_at_least 4.8 || is_crosscompile || use multislot || [[ ${GCC_PV} == *_alpha* ]] ; then +if ! tc_version_is_at_least 4.8 || is_crosscompile || [[ ${GCC_PV} == *_alpha* ]] ; then SLOT="${GCC_CONFIG_VER}" elif ! tc_version_is_at_least 5.0 ; then SLOT="${GCC_BRANCH_VER}" @@ -2091,10 +2091,6 @@ local curr_branch_ver=$(get_version_component_range 1-2 ${curr_config_ver}) - # If we're using multislot, just run gcc-config if we're installing - # to the same profile as the current one. - use multislot && return $([[ ${curr_config_ver} == ${GCC_CONFIG_VER} ]]) - if [[ ${curr_branch_ver} == ${GCC_BRANCH_VER} ]] ; then return 0 else
(In reply to Anthony Basile from comment #9) there's no need to keep a copy anymore once the eclass does: SLOT="${GCC_CONFIG_VER}"
gcc has multislot always enabled now *** This bug has been marked as a duplicate of bug 174407 ***
