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
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
|
# Copyright 1999-2023 Gentoo Authors
# Distributed under the terms of the GNU General Public License v2
EAPI=8
inherit gnuconfig libtool flag-o-matic multilib-minimal toolchain-funcs
MY_PV=${PV/_p*}
MY_PV=${MY_PV/_/-}
MANUAL_PV=${MY_PV}
MANUAL_PV=6.2.1
MY_P=${PN}-${MY_PV}
PLEVEL=${PV/*p}
DESCRIPTION="Library for arbitrary-precision arithmetic on different type of numbers"
HOMEPAGE="https://gmplib.org/"
SRC_URI="
https://gmplib.org/download/gmp/${MY_P}.tar.xz
mirror://gnu/${PN}/${MY_P}.tar.xz
doc? ( https://gmplib.org/${PN}-man-${MANUAL_PV}.pdf )
"
S="${WORKDIR}"/${MY_P%a}
LICENSE="|| ( LGPL-3+ GPL-2+ )"
# The subslot reflects the C & C++ SONAMEs.
SLOT="0/10.4"
KEYWORDS="~alpha ~amd64 ~arm ~arm64 ~hppa ~ia64 ~loong ~m68k ~mips ~ppc ~ppc64 ~riscv ~s390 ~sparc ~x86 ~amd64-linux ~x86-linux ~arm64-macos ~ppc-macos ~x64-macos ~x64-solaris"
IUSE="+asm doc +cpudetection +cxx pic static-libs"
REQUIRED_USE="cpudetection? ( asm )"
RESTRICT="!cpudetection? ( bindist )"
BDEPEND="
app-arch/xz-utils
sys-devel/m4
"
DOCS=( AUTHORS ChangeLog NEWS README doc/configuration doc/isa_abi_headache )
HTML_DOCS=( doc )
MULTILIB_WRAPPED_HEADERS=( /usr/include/gmp.h )
PATCHES=(
"${FILESDIR}"/${PN}-6.1.0-noexecstack-detect.patch
"${FILESDIR}"/${PN}-6.2.1-no-zarch.patch
)
pkg_pretend() {
if use cpudetection && ! use amd64 && ! use x86 ; then
elog "Using generic C implementation on non-amd64/x86 with USE=cpudetection"
elog "--enable-fat is a no-op on alternative arches."
elog "To obtain an optimized build, set USE=-cpudetection, but binpkgs should not then be made."
fi
}
src_prepare() {
default
# We cannot run autotools here as gcc depends on this package
elibtoolize
# bug #536894
if [[ ${CHOST} == *-darwin* ]] ; then
eapply "${FILESDIR}"/${PN}-6.1.2-gcc-apple-4.0.1.patch
fi
# GMP uses the "ABI" env var during configure as does Gentoo (econf).
# So, to avoid patching the source constantly, wrap things up.
mv configure configure.wrapped || die
cat <<-\EOF > configure
#!/usr/bin/env sh
exec env ABI="${GMPABI}" "$0.wrapped" "$@"
EOF
# Patches to original configure might have lost the +x bit.
chmod a+rx configure{,.wrapped} || die
# Save the upstream files named config.{guess,sub} which are
# wrappers around the gnuconfig versions.
mkdir "${T}"/gmp-gnuconfig || die
mv config.guess "${T}"/gmp-gnuconfig/config.guess || die
mv config.sub "${T}"/gmp-gnuconfig/config.sub || die
# Grab fresh copies from gnuconfig.
touch config.guess config.sub || die
gnuconfig_update
# Rename the fresh copies to the filenames the wrappers from GMP
# expect.
mv config.guess configfsf.guess || die
mv config.sub configfsf.sub || die
}
multilib_src_configure() {
# Generally a very fragile package
strip-flags
# Miscompiled with LTO at least on arm64, bug #889948
# 6.3.0 says it now supports LTO, but needs retesting on a variety
# of platforms. Fix was maybe https://gmplib.org/repo/gmp-6.3/rev/9c324044f4b5.
filter-lto
# Because of our 32-bit userland, 1.0 is the only HPPA ABI that works
# https://gmplib.org/manual/ABI-and-ISA.html#ABI-and-ISA (bug #344613)
if [[ ${CHOST} == hppa2.0-* ]] ; then
GMPABI="1.0"
fi
# ABI mappings (needs all architectures supported)
case ${ABI} in
32|x86) GMPABI=32;;
64|amd64|n64) GMPABI=64;;
[onx]32) GMPABI=${ABI};;
esac
export GMPABI
tc-export CC
# https://gmplib.org/manual/Notes-for-Package-Builds
local myeconfargs=(
CC_FOR_BUILD="$(tc-getBUILD_CC)"
--localstatedir="${EPREFIX}"/var/state/gmp
--enable-shared
$(use_enable asm assembly)
# fat is needed to avoid gmp installing either purely generic
# or specific-to-used-CPU (which our config.guess refresh prevents at the moment).
# Both Fedora and opensuse use this option to tackle the issue, bug #883201.
#
# This only works for amd64/x86, so to get accelerated performance
# (i.e. not using the generic C), one needs USE=-cpudetection if
# on non-amd64/x86.
#
# (We do not mask USE=cpudetection on !amd64/x86 because we want
# the flag to be useful on other arches to allow opting out of the
# config.guess logic below.)
$(use_enable cpudetection fat)
$(use_enable cxx)
$(use_enable static-libs static)
# --with-pic forces static libraries to be built as PIC
# and without TEXTRELs. musl does not support TEXTRELs: bug #707332
$(use pic && echo --with-pic)
)
# Move the wrappers from GMP back into place (may have been destroyed by previous econf run)
cp "${T}"/gmp-gnuconfig/config.guess "${S}"/config.guess || die
cp "${T}"/gmp-gnuconfig/config.sub "${S}"/config.sub || die
# See bug #883201 again.
if ! use cpudetection && ! tc-is-cross-compiler ; then
local gmp_host=$("${S}"/config.guess || die "failed to run config.guess")
if [[ -z ${gmp_host} ]] ; then
die "Empty result from GMP's custom config.guess!"
fi
einfo "GMP guessed processor type: ${gmp_host}"
ewarn "This build will only work on this machine. Enable USE=cpudetection for binary packages!"
export ac_cv_build="${gmp_host}"
export ac_cv_host="${gmp_host}"
fi
ECONF_SOURCE="${S}" econf "${myeconfargs[@]}"
}
multilib_src_install() {
emake DESTDIR="${D}" install
# Should be a standalone lib
rm -f "${ED}"/usr/$(get_libdir)/libgmp.la
# This requires libgmp
local la="${ED}/usr/$(get_libdir)/libgmpxx.la"
if ! use static-libs ; then
rm -f "${la}" || die
fi
}
multilib_src_install_all() {
einstalldocs
use doc && cp "${DISTDIR}"/gmp-man-${MANUAL_PV}.pdf "${ED}"/usr/share/doc/${PF}/
}
|
GitWeb