public inbox for gentoo-commits@lists.gentoo.org
 help / color / mirror / Atom feed
From: "Sam James" <sam@gentoo.org>
To: gentoo-commits@lists.gentoo.org
Subject: [gentoo-commits] repo/gentoo:master commit in: dev-libs/glib/, dev-libs/glib/files/
Date: Tue, 27 Sep 2022 18:17:00 +0000 (UTC)	[thread overview]
Message-ID: <1664302614.0a9a261cfe28fd20a512dfeb705b9fb4a5b3dbc2.sam@gentoo> (raw)

commit:     0a9a261cfe28fd20a512dfeb705b9fb4a5b3dbc2
Author:     Sam James <sam <AT> gentoo <DOT> org>
AuthorDate: Tue Sep 27 18:16:09 2022 +0000
Commit:     Sam James <sam <AT> gentoo <DOT> org>
CommitDate: Tue Sep 27 18:16:54 2022 +0000
URL:        https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=0a9a261c

dev-libs/glib: backport crash fix

As requested by e.g. GIMP maintainers
at https://twitter.com/zemarmot/status/1574362564015841281.

Signed-off-by: Sam James <sam <AT> gentoo.org>

 .../glib/files/glib-2.74.0-crash-gparamspec.patch  |  27 ++
 dev-libs/glib/glib-2.74.0-r1.ebuild                | 289 +++++++++++++++++++++
 2 files changed, 316 insertions(+)

diff --git a/dev-libs/glib/files/glib-2.74.0-crash-gparamspec.patch b/dev-libs/glib/files/glib-2.74.0-crash-gparamspec.patch
new file mode 100644
index 000000000000..84edcde4189d
--- /dev/null
+++ b/dev-libs/glib/files/glib-2.74.0-crash-gparamspec.patch
@@ -0,0 +1,27 @@
+https://gitlab.gnome.org/GNOME/glib/-/commit/ea3f17d598d550345e94e4571130e429443e91cb
+https://gitlab.gnome.org/GNOME/glib/-/issues/2770
+
+From: Emmanuele Bassi <ebassi@gnome.org>
+Date: Sun, 25 Sep 2022 14:20:24 +0100
+Subject: [PATCH] Empty values are not valid GParamSpec
+
+The validate() vfunc for GParamSpecParam returns FALSE for empty GValue,
+which means the is_valid() vfunc should do the same.
+
+This avoids a segfault when calling g_param_value_is_valid() on a
+GParamSpecParam.
+
+Fixes: #2770
+--- a/gobject/gparamspecs.c
++++ b/gobject/gparamspecs.c
+@@ -894,6 +894,9 @@ param_param_is_valid (GParamSpec   *pspec,
+ {
+   GParamSpec *param = value->data[0].v_pointer;
+ 
++  if (param == NULL)
++    return FALSE;
++
+   return g_value_type_compatible (G_PARAM_SPEC_TYPE (param), G_PARAM_SPEC_VALUE_TYPE (pspec));
+ }
+ 
+GitLab

diff --git a/dev-libs/glib/glib-2.74.0-r1.ebuild b/dev-libs/glib/glib-2.74.0-r1.ebuild
new file mode 100644
index 000000000000..9c91688a8c0a
--- /dev/null
+++ b/dev-libs/glib/glib-2.74.0-r1.ebuild
@@ -0,0 +1,289 @@
+# Copyright 1999-2022 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+PYTHON_REQ_USE="xml(+)"
+PYTHON_COMPAT=( python3_{8..11} )
+
+inherit flag-o-matic gnome.org gnome2-utils linux-info meson-multilib multilib python-any-r1 toolchain-funcs xdg
+
+DESCRIPTION="The GLib library of C routines"
+HOMEPAGE="https://www.gtk.org/"
+
+LICENSE="LGPL-2.1+"
+SLOT="2"
+IUSE="dbus debug +elf gtk-doc +mime selinux static-libs sysprof systemtap test utils xattr"
+RESTRICT="!test? ( test )"
+REQUIRED_USE="gtk-doc? ( test )" # Bug #777636
+
+KEYWORDS="~alpha ~amd64 ~arm ~arm64 ~hppa ~ia64 ~loong ~m68k ~mips ~ppc ~ppc64 ~riscv ~s390 ~sparc ~x86 ~amd64-linux ~x86-linux"
+
+# * elfutils (via libelf) does not build on Windows. gresources are not embedded
+# within ELF binaries on that platform anyway and inspecting ELF binaries from
+# other platforms is not that useful so exclude the dependency in this case.
+# * Technically static-libs is needed on zlib, util-linux and perhaps more, but
+# these are used by GIO, which glib[static-libs] consumers don't really seem
+# to need at all, thus not imposing the deps for now and once some consumers
+# are actually found to static link libgio-2.0.a, we can revisit and either add
+# them or just put the (build) deps in that rare consumer instead of recursive
+# RDEPEND here (due to lack of recursive DEPEND).
+RDEPEND="
+	!<dev-util/gdbus-codegen-${PV}
+	>=virtual/libiconv-0-r1[${MULTILIB_USEDEP}]
+	>=dev-libs/libpcre2-10.32:0=[${MULTILIB_USEDEP},static-libs?]
+	>=dev-libs/libffi-3.0.13-r1:=[${MULTILIB_USEDEP}]
+	>=sys-libs/zlib-1.2.8-r1[${MULTILIB_USEDEP}]
+	>=virtual/libintl-0-r2[${MULTILIB_USEDEP}]
+	kernel_linux? ( >=sys-apps/util-linux-2.23[${MULTILIB_USEDEP}] )
+	selinux? ( >=sys-libs/libselinux-2.2.2-r5[${MULTILIB_USEDEP}] )
+	xattr? ( !elibc_glibc? ( >=sys-apps/attr-2.4.47-r1[${MULTILIB_USEDEP}] ) )
+	elf? ( virtual/libelf:0= )
+	sysprof? ( >=dev-util/sysprof-capture-3.40.1:4[${MULTILIB_USEDEP}] )
+"
+DEPEND="${RDEPEND}"
+# libxml2 used for optional tests that get automatically skipped
+BDEPEND="
+	app-text/docbook-xsl-stylesheets
+	dev-libs/libxslt
+	>=sys-devel/gettext-0.19.8
+	gtk-doc? ( >=dev-util/gtk-doc-1.33
+		app-text/docbook-xml-dtd:4.2
+		app-text/docbook-xml-dtd:4.5 )
+	systemtap? ( >=dev-util/systemtap-1.3 )
+	${PYTHON_DEPS}
+	test? ( >=sys-apps/dbus-1.2.14 )
+	virtual/pkgconfig
+"
+# TODO: >=dev-util/gdbus-codegen-${PV} test dep once we modify gio/tests/meson.build to use external gdbus-codegen
+
+PDEPEND="
+	dbus? ( gnome-base/dconf )
+	mime? ( x11-misc/shared-mime-info )
+"
+# shared-mime-info needed for gio/xdgmime, bug #409481
+# dconf is needed to be able to save settings, bug #498436
+
+MULTILIB_CHOST_TOOLS=(
+	/usr/bin/gio-querymodules$(get_exeext)
+)
+
+PATCHES=(
+	"${FILESDIR}"/${PN}-2.64.1-mark-gdbus-server-auth-test-flaky.patch
+	"${FILESDIR}"/${P}-clang-build.patch
+	"${FILESDIR}"/${P}-crash-gparamspec.patch
+)
+
+pkg_setup() {
+	if use kernel_linux ; then
+		CONFIG_CHECK="~INOTIFY_USER"
+		if use test ; then
+			CONFIG_CHECK="~IPV6"
+			WARNING_IPV6="Your kernel needs IPV6 support for running some tests, skipping them."
+		fi
+		linux-info_pkg_setup
+	fi
+	python-any-r1_pkg_setup
+}
+
+src_prepare() {
+	if use test; then
+		# TODO: Review the test exclusions, especially now with meson
+		# Disable tests requiring dev-util/desktop-file-utils when not installed, bug #286629, upstream bug #629163
+		if ! has_version dev-util/desktop-file-utils ; then
+			ewarn "Some tests will be skipped due dev-util/desktop-file-utils not being present on your system,"
+			ewarn "think on installing it to get these tests run."
+			sed -i -e "/appinfo\/associations/d" gio/tests/appinfo.c || die
+			sed -i -e "/g_test_add_func/d" gio/tests/desktop-app-info.c || die
+		fi
+
+		# gdesktopappinfo requires existing terminal (gnome-terminal or any
+		# other), falling back to xterm if one doesn't exist
+		#if ! has_version x11-terms/xterm && ! has_version x11-terms/gnome-terminal ; then
+		#	ewarn "Some tests will be skipped due to missing terminal program"
+		# These tests seem to sometimes fail even with a terminal; skip for now and reevulate with meson
+		# Also try https://gitlab.gnome.org/GNOME/glib/issues/1601 once ready for backport (or in a bump) and file new issue if still fails
+		sed -i -e "/appinfo\/launch/d" gio/tests/appinfo.c || die
+		# desktop-app-info/launch* might fail similarly
+		sed -i -e "/desktop-app-info\/launch-as-manager/d" gio/tests/desktop-app-info.c || die
+		#fi
+
+		# https://bugzilla.gnome.org/show_bug.cgi?id=722604
+		sed -i -e "/timer\/stop/d" glib/tests/timer.c || die
+		sed -i -e "/timer\/basic/d" glib/tests/timer.c || die
+
+		ewarn "Tests for search-utils have been skipped"
+		sed -i -e "/search-utils/d" glib/tests/meson.build || die
+
+		# Play nice with network-sandbox, but this approach would defeat the purpose of the test
+		#sed -i -e "s/localhost/127.0.0.1/g" gio/tests/gsocketclient-slow.c || die
+	else
+		# Don't build tests, also prevents extra deps, bug #512022
+		sed -i -e '/subdir.*tests/d' {.,gio,glib}/meson.build || die
+	fi
+
+	# Don't build fuzzing binaries - not used
+	sed -i -e '/subdir.*fuzzing/d' meson.build || die
+
+	# gdbus-codegen is a separate package
+	sed -i -e '/install_dir/d' gio/gdbus-2.0/codegen/meson.build || die
+
+	# Same kind of meson-0.50 issue with some installed-tests files; will likely be fixed upstream soon
+	sed -i -e '/install_dir/d' gio/tests/meson.build || die
+
+	cat > "${T}/glib-test-ld-wrapper" <<-EOF
+		#!/usr/bin/env sh
+		exec \${LD:-ld} "\$@"
+	EOF
+	chmod a+x "${T}/glib-test-ld-wrapper" || die
+	sed -i -e "s|'ld'|'${T}/glib-test-ld-wrapper'|g" gio/tests/meson.build || die
+
+	default
+	gnome2_environment_reset
+	# TODO: python_name sedding for correct python shebang? Might be relevant mainly for glib-utils only
+}
+
+multilib_src_configure() {
+	if use debug; then
+		append-cflags -DG_ENABLE_DEBUG
+	else
+		append-cflags -DG_DISABLE_CAST_CHECKS # https://gitlab.gnome.org/GNOME/glib/issues/1833
+	fi
+
+	# TODO: figure a way to pass appropriate values for all cross properties that glib uses (search for get_cross_property)
+	#if tc-is-cross-compiler ; then
+		# https://bugzilla.gnome.org/show_bug.cgi?id=756473
+		# TODO-meson: This should be in meson cross file as 'growing_stack' property; and more, look at get_cross_property
+		#case ${CHOST} in
+		#hppa*|metag*) export glib_cv_stack_grows=yes ;;
+		#*)            export glib_cv_stack_grows=no ;;
+		#esac
+	#fi
+
+	local emesonargs=(
+		-Ddefault_library=$(usex static-libs both shared)
+		$(meson_feature selinux)
+		$(meson_use xattr)
+		-Dlibmount=enabled # only used if host_system == 'linux'
+		-Dman=true
+		$(meson_use systemtap dtrace)
+		$(meson_use systemtap)
+		$(meson_feature sysprof)
+		$(meson_native_use_bool gtk-doc gtk_doc)
+		$(meson_use test tests)
+		-Dinstalled_tests=false
+		-Dnls=enabled
+		-Doss_fuzz=disabled
+		$(meson_native_use_feature elf libelf)
+		-Dmultiarch=false
+	)
+	meson_src_configure
+}
+
+multilib_src_test() {
+	export XDG_CONFIG_DIRS=/etc/xdg
+	export XDG_DATA_DIRS=/usr/local/share:/usr/share
+	export G_DBUS_COOKIE_SHA1_KEYRING_DIR="${T}/temp"
+	export LC_TIME=C # bug #411967
+	export TZ=UTC
+	unset GSETTINGS_BACKEND # bug #596380
+	python_setup
+
+	# https://bugs.gentoo.org/839807
+	local -x SANDBOX_PREDICT=${SANDBOX_PREDICT}
+	addpredict /usr/b
+
+	# Related test is a bit nitpicking
+	mkdir "$G_DBUS_COOKIE_SHA1_KEYRING_DIR"
+	chmod 0700 "$G_DBUS_COOKIE_SHA1_KEYRING_DIR"
+
+	meson_src_test --timeout-multiplier 2 --no-suite flaky
+}
+
+multilib_src_install() {
+	meson_src_install
+	keepdir /usr/$(get_libdir)/gio/modules
+}
+
+multilib_src_install_all() {
+	# These are installed by dev-util/glib-utils
+	# TODO: With patching we might be able to get rid of the python-any deps and removals, and test depend on glib-utils instead; revisit now with meson
+	rm "${ED}/usr/bin/glib-genmarshal" || die
+	rm "${ED}/usr/share/man/man1/glib-genmarshal.1" || die
+	rm "${ED}/usr/bin/glib-mkenums" || die
+	rm "${ED}/usr/share/man/man1/glib-mkenums.1" || die
+	rm "${ED}/usr/bin/gtester-report" || die
+	rm "${ED}/usr/share/man/man1/gtester-report.1" || die
+	# gdbus-codegen manpage installed by dev-util/gdbus-codegen
+	rm "${ED}/usr/share/man/man1/gdbus-codegen.1" || die
+}
+
+pkg_preinst() {
+	xdg_pkg_preinst
+
+	# Make gschemas.compiled belong to glib alone
+	local cache="/usr/share/glib-2.0/schemas/gschemas.compiled"
+
+	if [[ -e ${EROOT}${cache} ]]; then
+		cp "${EROOT}"${cache} "${ED}"/${cache} || die
+	else
+		touch "${ED}"${cache} || die
+	fi
+
+	multilib_pkg_preinst() {
+		# Make giomodule.cache belong to glib alone
+		local cache="/usr/$(get_libdir)/gio/modules/giomodule.cache"
+
+		if [[ -e ${EROOT}${cache} ]]; then
+			cp "${EROOT}"${cache} "${ED}"${cache} || die
+		else
+			touch "${ED}"${cache} || die
+		fi
+	}
+
+	# Don't run the cache ownership when cross-compiling, as it would end up with an empty cache
+	# file due to inability to create it and GIO might not look at any of the modules there
+	if ! tc-is-cross-compiler ; then
+		multilib_foreach_abi multilib_pkg_preinst
+	fi
+}
+
+pkg_postinst() {
+	xdg_pkg_postinst
+	# glib installs no schemas itself, but we force update for fresh install in case
+	# something has dropped in a schemas file without direct glib dep; and for upgrades
+	# in case the compiled schema format could have changed
+	gnome2_schemas_update
+
+	multilib_pkg_postinst() {
+		gnome2_giomodule_cache_update \
+			|| die "Update GIO modules cache failed (for ${ABI})"
+	}
+	if ! tc-is-cross-compiler ; then
+		multilib_foreach_abi multilib_pkg_postinst
+	else
+		ewarn "Updating of GIO modules cache skipped due to cross-compilation."
+		ewarn "You might want to run gio-querymodules manually on the target for"
+		ewarn "your final image for performance reasons and re-run it when packages"
+		ewarn "installing GIO modules get upgraded or added to the image."
+	fi
+
+	for v in ${REPLACING_VERSIONS}; do
+		if ver_test "$v" "-lt" "2.63.6"; then
+			ewarn "glib no longer installs the gio-launch-desktop binary. You may need"
+			ewarn "to restart your session for \"Open With\" dialogs to work."
+		fi
+	done
+}
+
+pkg_postrm() {
+	xdg_pkg_postrm
+	gnome2_schemas_update
+
+	if [[ -z ${REPLACED_BY_VERSION} ]]; then
+		multilib_pkg_postrm() {
+			rm -f "${EROOT}"/usr/$(get_libdir)/gio/modules/giomodule.cache
+		}
+		multilib_foreach_abi multilib_pkg_postrm
+		rm -f "${EROOT}"/usr/share/glib-2.0/schemas/gschemas.compiled
+	fi
+}


             reply	other threads:[~2022-09-27 18:17 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-27 18:17 Sam James [this message]
  -- strict thread matches above, loose matches on Subject: below --
2025-02-21 21:55 [gentoo-commits] repo/gentoo:master commit in: dev-libs/glib/, dev-libs/glib/files/ Sam James
2024-12-21 22:18 Sam James
2024-03-25 10:53 Sam James
2021-12-11 13:34 Mart Raudsepp
2021-08-05  1:56 Sam James
2020-06-07 21:28 Mart Raudsepp
2019-07-23 13:03 Mart Raudsepp
2018-06-28 18:33 Mart Raudsepp
2018-06-10  8:25 Mart Raudsepp
2017-03-19 14:43 Mart Raudsepp
2016-11-11 12:47 Gilles Dartiguelongue
2016-03-26 21:25 Alexandre Rostovtsev

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1664302614.0a9a261cfe28fd20a512dfeb705b9fb4a5b3dbc2.sam@gentoo \
    --to=sam@gentoo.org \
    --cc=gentoo-commits@lists.gentoo.org \
    --cc=gentoo-dev@lists.gentoo.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox