1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
|
# Copyright 1999-2022 Gentoo Authors
# Distributed under the terms of the GNU General Public License v2
EAPI=8
TOOLCHAIN_PATCH_DEV="sam"
PATCH_VER="1"
PATCH_GCC_VER="12.2.0"
MUSL_VER="1"
MUSL_GCC_VER="12.2.0"
inherit toolchain
# Don't keyword live ebuilds
if ! tc_is_live && [[ -z ${TOOLCHAIN_USE_GIT_PATCHES} ]] ; then
# fails to compile on Solaris and macOS, need to check why
: KEYWORDS="~amd64-linux ~x86-linux ~arm64-macos ~ppc-macos ~x64-macos ~x64-solaris"
KEYWORDS="~arm64-macos ~x64-macos ~x64-solaris"
fi
# use alternate source for Apple M1 (also works for x86_64)
IANSGCCVER="gcc-12.2-darwin-r0"
SRC_URI+=" elibc_Darwin? (
https://github.com/iains/gcc-12-branch/archive/refs/tags/${IANSGCCVER}.tar.gz )"
# Technically only if USE=hardened *too* right now, but no point in complicating it further.
# If GCC is enabling CET by default, we need glibc to be built with support for it.
# bug #830454
RDEPEND="!prefix-guest? ( elibc_glibc? ( sys-libs/glibc[cet(-)?] ) )"
DEPEND="${RDEPEND}"
BDEPEND="
kernel_linux? ( >=${CATEGORY}/binutils-2.30[cet(-)?] )
kernel_Darwin? (
|| ( ${CATEGORY}/binutils-apple ${CATEGORY}/native-cctools )
)"
src_unpack() {
if use elibc_Darwin ; then
# just use Ian's source, not the main one
S="${WORKDIR}/gcc-12-branch-${IANSGCCVER}"
fi
default
}
src_prepare() {
if [[ ${CHOST} == *-darwin* ]] ; then
# https://bugs.gentoo.org/898610#c17
# kill no_pie patch, it breaks things here
rm "${WORKDIR}"/patch/09_all_nopie-all-flags.patch || die
fi
toolchain_src_prepare
eapply_user
# make sure 64-bits native targets don't screw up the linker paths
eapply "${FILESDIR}"/gcc-12-no-libs-for-startfile.patch
if use prefix; then
eapply "${FILESDIR}"/gcc-12-prefix-search-dirs.patch
# try /usr/lib32 in 32bit profile on x86_64-linux (needs
# --enable-multilib), but this does make sense in prefix only
eapply -p0 "${FILESDIR}"/${PN}-4.8.3-linux-x86-on-amd64.patch
fi
# make it have correct install_names on Darwin
eapply -p1 "${FILESDIR}"/4.3.3/darwin-libgcc_s-installname.patch
if [[ ${CHOST} == powerpc*-darwin* ]] ; then
# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=44107
sed -i -e 's|^ifeq (/usr/lib,|ifneq (/usr/lib,|' \
libgcc/config/t-slibgcc-darwin || die
fi
if [[ ${CHOST} == *-solaris* ]] ; then
# madvise is not available in the compatibility mode GCC uses,
# posix_madvise however, is
sed -i -e 's/madvise/posix_madvise/' gcc/cp/module.cc || die
fi
if [[ ${CHOST} == *-darwin* ]] ; then
# GCC' Darwin fork enables support for "-stdlib=libc++"
# unconditionally, and its default include path is invalid,
# causing package build failures due to missing header.
# But more importantly, it breaks the assumption of many build
# scripts and changes their CFLAGS and linking behaviors. The
# situation is tricky and needs careful considerations.
# For now, just disable support for "-stdlib=libc++".
eapply "${FILESDIR}"/gcc-12.2.0-disable-stdlib-option-on-darwin.patch
fi
}
src_configure() {
local myconf=()
case ${CTARGET}:" ${USE} " in
powerpc*-darwin*)
# bug #381179
filter-flags "-mcpu=*" "-mtune=*"
# bug #657522
# A bug in configure checks whether -no-pie works, but the
# compiler doesn't pass -no-pie onto the linker if -fno-PIE
# isn't passed, so the configure check always finds -no-pie
# is accepted. (Likewise, when -fno-PIE is passed, the
# compiler passes -no_pie onto the linker.)
# Since our linker doesn't grok this, avoid above checks to
# be run
# NOTE: later ld64 does grok -no_pie, not -no-pie (as checked)
export gcc_cv_c_no_fpie=no
export gcc_cv_no_pie=no
;;
*-darwin20)
# use sysroot with the linker, #756160
export gcc_cv_ld_sysroot=yes
;;
*-solaris*)
# todo: some magic for native vs. GNU linking?
myconf+=( --with-gnu-ld --with-gnu-as --enable-largefile )
# Solaris 11 defines this in its headers, but that causes a
# mismatch whilst compiling, bug #657514
export ac_cv_func_aligned_alloc=no
export ac_cv_func_memalign=no
export ac_cv_func_posix_memalign=no
;;
i[34567]86-*-linux*:*" prefix "*)
# to allow the linux-x86-on-amd64.patch become useful, we need
# to enable multilib, even if there is just one multilib option.
myconf+=( --enable-multilib )
if [[ ${CBUILD:-${CHOST}} == "${CHOST}" ]]; then
# we might be on x86_64-linux, but don't do cross-compile, so
# tell the host-compiler to really create 32bits (for stage1)
# (real x86-linux-gcc also accept -m32).
append-flags -m32
fi
;;
esac
# Since GCC 4.1.2 some non-posix (?) /bin/sh compatible code is used, at
# least on Solaris, and AIX /bin/sh is way too slow,
# so force it to use $BASH (that portage uses) - it can't be EPREFIX
# in case that doesn't exist yet
export CONFIG_SHELL="${CONFIG_SHELL:-${BASH}}"
toolchain_src_configure "${myconf[@]}"
}
|
GitWeb