[opensuse-factory] New Tumbleweed snapshot 20201007 released!
Please note that this mail was generated by a script. The described changes are computed based on the x86_64 DVD. The full online repo contains too many changes to be listed here. Please check the known defects of this snapshot before upgrading: https://openqa.opensuse.org/tests/overview?distri=opensuse&groupid=1&version=Tumbleweed&build=20201007 Please do not reply to this email to report issues, rather file a bug on bugzilla.opensuse.org. For more information on filing bugs please see https://en.opensuse.org/openSUSE:Submitting_bug_reports Packages changed: alsa-plugins alsa-utils bison (3.6.4 -> 3.7.2) checkpolicy (3.0 -> 3.1) gettext-runtime (0.20.2 -> 0.21) glibc (2.31 -> 2.32) ibus (1.5.22 -> 1.5.23) krename (5.0.0 -> 5.0.1) libqt5-qtwebkit libselinux (3.0 -> 3.1) libselinux-bindings (3.0 -> 3.1) libsemanage (3.0 -> 3.1) libsepol (3.0 -> 3.1) lightdm linux-glibc-devel (5.7 -> 5.8) mozilla-nss ncurses (6.2.20200711 -> 6.2.20200912) policycoreutils (3.0 -> 3.1) python-semanage (3.0 -> 3.1) sysvinit (2.96 -> 2.97) tcpd x11-tools xdm xinit xmodmap xorg-x11-server zsh === Details === ==== alsa-plugins ==== Subpackages: alsa-plugins-pulse alsa-plugins-pulse-32bit alsa-plugins-speexrate alsa-plugins-upmix - Placeholder for SLE15-SP3 sync: jsc#SLE-11987 no functional changes at all ==== alsa-utils ==== - Conditionally add buildreq alsa-topology-devel to fix build errors ==== bison ==== Version update (3.6.4 -> 3.7.2) Subpackages: bison-lang - GNU bison 3.7.2: * fix all known Bison CVEs relating to the bison-the-program itself, not the generated code. * concurrent build issues * Push parsers always use YYMALLOC/YYFREE * Fix unlikely crashes found by fuzzing - GNU bison 3.7.1: * YYPRINT macro is deprecated * --graph will generate a *.gv file by default * Add Counterexample Generation feature * Add File prefix mapping feature * Various bug fixes and behavior changes ==== checkpolicy ==== Version update (3.0 -> 3.1) - Update to version 3.1 * checkpolicy treats invalid characters as an error - might break rare use cases (intentionally) * Drop extern_te_assert_t.patch, is upstream ==== gettext-runtime ==== Version update (0.20.2 -> 0.21) Subpackages: gettext-tools libtextstyle0 - Add multiple new features (bsc#1165138) - Add patches: * 0001-msgcat-Add-feature-to-use-the-newest-po-file.patch * 0002-msgcat-Merge-headers-when-use-first.patch - Reintroduce utoreconf call - Update to 0.21: * Programming languages support: - Shell: o xgettext now recognizes and ignores 'env' invocations and environment variable assignments in front of commands. - Java: o xgettext now recognizes format strings in the Formatter syntax. They are marked as 'java-printf-format' in POT and PO files. o xgettext now recognizes text blocks as string literals. - JavaScript: xgettext parses JSX expressions more reliably. - Ruby: o xgettext now supports Ruby. o 'msgfmt -c' now verifies the syntax of translations of Ruby format strings. * Improvements for translators: - When msgfmt writes a MO file, it now does so in such a way that processes that are currently using an older copy of the MO file will not crash. * Libtextstyle: - Added support for emitting hyperlinks. - New API for doing formatted output. - The example programs support the NO_COLOR environment variable. ==== glibc ==== Version update (2.31 -> 2.32) Subpackages: glibc-32bit glibc-devel glibc-extra glibc-lang glibc-locale glibc-locale-base nscd - Keep nsswitch.conf in /etc for SLES15 - syslog-locking.patch: Correct locking and cancellation cleanup in syslog functions (bsc#1172085, BZ #26100) - ifunc-fma4.patch: x86-64: Fix FMA4 detection in ifunc (BZ #26534) - Update to glibc 2.32 * Unicode 13.0.0 Support * New locale added: ckb_IQ * The GNU C Library now loads audit modules listed in the DT_AUDIT and DT_DEPAUDIT dynamic section entries of the main executable * powerpc64le supports IEEE128 long double libm/libc redirects when using the -mabi=ieeelongdouble to compile C code on supported GCC toolchains * To help detect buffer overflows and other out-of-bounds accesses several APIs have been annotated with GCC 'access' attribute * On Linux, functions the pthread_attr_setsigmask_np and pthread_attr_getsigmask_np have been added * The GNU C Library now provides the header file <sys/single_threaded.h> which declares the variable __libc_single_threaded * The functions sigabbrev_np and sigdescr_np have been added * The functions strerrorname_np and strerrordesc_np have been added * AArch64 now supports standard branch protection security hardening in glibc when it is built with a GCC that is configured with - -enable-standard-branch-protection (or if -mbranch-protection=standard flag is passed when building both GCC target libraries and glibc, in either case a custom GCC is needed) * The deprecated <sys/sysctl.h> header and the sysctl function have been removed * The sstk function is no longer available to newly linked binaries * The legacy signal handling functions siginterrupt, sigpause, sighold, sigrelse, sigignore and sigset, and the sigmask macro have been deprecated * ldconfig now defaults to the new format for ld.so.cache * The deprecated arrays sys_siglist, _sys_siglist, and sys_sigabbrev are no longer available to newly linked binaries, and their declarations have been removed from <string.h> * The deprecated symbols sys_errlist, _sys_errlist, sys_nerr, and _sys_nerr are no longer available to newly linked binaries, and their declarations have been removed from from <stdio.h> * Both strerror and strerror_l now share the same internal buffer in the calling thread, meaning that the returned string pointer may be invalided or contents might be overwritten on subsequent calls in the same thread or if the thread is terminated * Using weak references to libpthread functions such as pthread_create or pthread_key_create to detect the singled-threaded nature of a program is an obsolescent feature * The "files" NSS module no longer supports the "key" database (used for secure RPC) * The __morecore and __after_morecore_hook malloc hooks and the default implementation __default_morecore have been deprecated * The hesiod NSS module has been deprecated and will be removed in a future version of glibc * CVE-2016-10228: An infinite loop has been fixed in the iconv program when invoked with the -c option and when processing invalid multi-byte input sequences * CVE-2020-10029: Trigonometric functions on x86 targets suffered from stack corruption when they were passed a pseudo-zero argument * CVE-2020-1752: A use-after-free vulnerability in the glob function when expanding ~user has been fixed. * CVE-2020-6096: A signed comparison vulnerability in the ARMv7 memcpy and memmove functions has been fixed - riscv-syscall-clobber.patch, ldbl-96-rem-pio2l.patch, long-double-alias.patch: Removed ==== ibus ==== Version update (1.5.22 -> 1.5.23) Subpackages: ibus-dict-emoji ibus-gtk ibus-gtk-32bit ibus-gtk3 ibus-lang libibus-1_0-5 libibus-1_0-5-32bit typelib-1_0-IBus-1_0 - Update version to 1.5.23 * Generate simple.xml with denylist 6042974 508527d 37db75b 6879879 59b902a 568d58d 6ed34f3 5959d6f 5d67a28 394d9a8 ed7bc8d e938846 3aa670e 0d90da4 e4dd6d1 * Accept xdigits only for Unicode typing a440942 * Update emoji-parser with CLDR emoji annotation release-31-0-1 9a9f828 * Update ibusunicodegen.h with unicode-ucd 13.0.0 e10fc89 * Delete deprecated ENABLE_APPINDICATOR_ENGINE_ICON check aa3a9f0 * Fix SEGV 02105c4 f591381 * Fix some errors in ibus-desktop-testing-runner 7b0d091 8da0167 * Refactor source files 0b9d936 0ad5e9a * Fix string formats in translatable strings 7caead1 f8c468a ce865f6 * Use WAYLAND_DISPLAY on Wayland sessions to make up IBus socket name (Carlos Garnacho) 8ce2520 * Skip parsing of compose sequence with invalid keysyms (Neil Shepperd) 0da3cec * Tell Pango about the engine language in the candidate panel (Aaron Muir Hamilton) 3f098dc 79a09f1 * Fix for several error spotted by static analyzer (ntfs.hard) 00adea6 * Remove glib_check_version() in gtk immodule (Changwoo Ryu) 5765bfd * Build the Emoji dictionaries in parallel (Changwoo Ryu) 59d0de4 * Update translation - Drop 0001-Replace-the-Qt-check-for-appindicator-engine-icon-wi.patch, ibus-use-wayland-display-for-socket-name.patch, ibus-socket-name-compatibility.patch. Merged by upstream - Update ibus.spec: Drop the is_opensuse macro to eliminate the difference between SLE-15 and openSUSE-Leap (jsc#SLE-11653). - Move xim.d files to /usr/etc when available, which is currently Tumbleweed only (boo#1176431) ==== krename ==== Version update (5.0.0 -> 5.0.1) - Update to 5.0.1: * Escape/unescape filenames with special characters (kde#391291) * Close the app if the progress dialog is closed (kde#395084) * Fix building against exiv2-0.27 * Fix build with taglib and ECM >= 5.72.0 * Make icons look sharp with HiDPI (kde#418112) - Drop patches merged upstream: * 0001-Use-cmakedefine01-instead-of-macro_bool_to_01.patch * 0002-Fix-building-against-exiv2-0.27.patch * 0003-Fix-the-previous-commits-and-the-build-with-exiv2-0..patch * use-local-cmake-modules-first.patch * Close-the-app-if-the-progress-dialog-is-closed.patch - Drop lower-minimum-cmake-version.patch, Leap 42.3 is no longer supported - Refresh servicemenus-files.patch ==== libqt5-qtwebkit ==== Subpackages: libQt5WebKit5 libQt5WebKit5-imports libQt5WebKitWidgets5 - Add upstream patch qtwebkit-5.212.0_pre20200309-bison-3.7.patch in order to fix build with a recent bison. ==== libselinux ==== Version update (3.0 -> 3.1) Subpackages: libselinux1 libselinux1-32bit selinux-tools - Update to version 3.1: * selinux/flask.h, selinux/av_permissions.h and sepol/policydb/flask.h were removed. All userspace object managers should have been updated to use the dynamic class/perm mapping support. Use string_to_security_class(3) and string_to_av_perm(3) to map the class and permission names to their policy values, or selinux_set_mapping(3) to create a mapping from class and permission index values used by the application to the policy values. * Removed restrictions in libsepol and checkpolicy that required all declared initial SIDs to be assigned a context. * Support for new policy capability genfs_seclabel_symlinks * selinuxfs is mounted with noexec and nosuid * `security_compute_user()` was deprecated ==== libselinux-bindings ==== Version update (3.0 -> 3.1) - Update to version 3.1: * selinux/flask.h, selinux/av_permissions.h and sepol/policydb/flask.h were removed. All userspace object managers should have been updated to use the dynamic class/perm mapping support. Use string_to_security_class(3) and string_to_av_perm(3) to map the class and permission names to their policy values, or selinux_set_mapping(3) to create a mapping from class and permission index values used by the application to the policy values. * Removed restrictions in libsepol and checkpolicy that required all declared initial SIDs to be assigned a context. * Support for new policy capability genfs_seclabel_symlinks * selinuxfs is mounted with noexec and nosuid * `security_compute_user()` was deprecated * Refreshed python3.8-compat.patch ==== libsemanage ==== Version update (3.0 -> 3.1) Subpackages: libsemanage-migrate-store libsemanage1 - Add /var/lib/selinux - Remove libsemanage-update-map-file.patch to prevent checkers from declining the submission. Keeping the snippet in the spec file in case we try to enable LTO again - Update to version 3.1 * Improved manpage * fsync final files before rename - Disabled LTO again. This breaks e.g. shadow and also other packages in security:SELinux - Fix build with LTO: [bsc#1133102] * Enable LTO (Link Time Optimization) and build with -ffat-lto-objects * Update map file to include new symbols and remove wildcards - Add libsemanage-update-map-file.patch ==== libsepol ==== Version update (3.0 -> 3.1) - Update to version 3.1 * Add support for new polcap genfs_seclabel_symlinks * Initialize the multiple_decls field of the cil db * Return error when identifier declared as both type and attribute * Write CIL default MLS rules on separate lines * Sort portcon rules consistently * Remove leftovers of cil_mem_error_handler * Drop remove_cil_mem_error_handler.patch, is included ==== lightdm ==== Subpackages: liblightdm-gobject-1-0 lightdm-lang - Fixup pam stack a bit more (boo#1176338): + Do not create hardlinks between lightdm pam files and xdm's pam files, but install explicit lightdm pam stack that @includes xdm. This has the advantage that, if an admin creates /etc/pam.d/xdm, we will follow this right away. With the hardlinks, the /usr/etc/pam.d/xdm would still be used, which might not be the expected outcome. - Follow XDM's change of moving the default pam config file to /usr/etc (boo#1176338). - Add pre/posttrans scripts to ensure user modified /etc/pam.d/lightdm* survives our move to /usr/etc (and user modification in /etc/pam.d wins over /usr/etc/pam.d). ==== linux-glibc-devel ==== Version update (5.7 -> 5.8) - Update to kernel headers 5.8 ==== mozilla-nss ==== Subpackages: libfreebl3 libfreebl3-hmac libsoftokn3 libsoftokn3-hmac mozilla-nss-certs mozilla-nss-tools - Add patch to fix build on aarch64 - boo#1176934: * nss-freebl-fix-aarch64.patch ==== ncurses ==== Version update (6.2.20200711 -> 6.2.20200912) Subpackages: libncurses6 ncurses-utils terminfo terminfo-base terminfo-iterm terminfo-screen - Add ncurses patch 20200912 + add configure-check for systre/tre with mingw configuration, to get the library-dependencies as seen in msys2 configuration for mingw64. + build-fixes for the win32-driver configuration. + use more defensive binary mode setting for Win32 (Juergen Pfeifer). - Add ncurses patch 20200907 + fix regression in setupterm validating non-empty $TERM (report by Soren Tempel). - Add ncurses patch 20200906 + merge/adapt in-progress work by Juergen Pfeifer for new version of win32-driver. + correct description of vt330/vt340 (Ross Combs). - Add ncurses patch 20200831 + build-fix for awk-scripts modified for win32-driver (report by Werner Fink). - Drop workaround patch awk-scripts.patch as now upstream fixed - Add patch awk-scripts.patch as workaround for patch 20200829 - Add ncurses patch 20200829 + remove a redundant NCURSES_EXPORT as a build-fix for "Maarten Anonymous". + merge/adapt in-progress work by Juergen Pfeifer for new version of win32-driver. + modify configure script, moving gcc -Werror options to EXTRA_CFLAGS to avoid breaking configure-checks (adapted from ongoing work on mawk and lynx). > errate for terminfo.src (report by Florian Weimer): + correct icl6404 csr + correct ti916 cup + improve ndr9500 - Refresh patch ncurses-6.2.dif - Add ncurses patch 20200822 + improve version-number extraction in MKlib_gen.sh + make the test-package for manpages installable by adjusting the man_db.renames file. + correct an off-by-one loop-limit in convert_strings function (report by Yue Tai). + add CF_SHARED_OPTS cases for HPE NonStop systems (Randall S Becker). + modify CF_SHARED_OPTS case for NetBSD to use the same "-shared" option for the non-rpath case as for the rpath case, to allow gcc to provide suitable runtime initialization (report by Rajeev V Pillai). - Disable wgetch-events as it is deprecated and breaks build of other packages - Add ncurses patch 20200817 + reduce build-warnings by excluding ncurses-internals from deprecation warnings. + mark wgetch-events feature as deprecated. + add definition for $(LIBS) to ncurses/Makefile.in, to simplify builds using the string-hacks option. + prevent KEY_EVENT from appearing in curses.h unless the configure option --enable-wgetch-events is used (report by Werner Fink). - Add ncurses patch 20200816 + amend tic/infocmp check to allow for the respective tool's absence (report by Steve Wills, cf: 20200808). + improved some of the build-scripts with shellcheck + filter out -MT/-MD/-MTd/-MDd options in script for Visual Studio C++ (discussion with "Maarten Anonymous"). - Add ncurses patch 20200808 + improve discussion of the system's tic utility when used as part of cross-compiling (discussion with Keith Marshall). + modify configuration checks for build-time tic/infocmp to use AC_CHECK_TOOL. That can still be overridden by --with-tic-path and - -with-infocmp-path when fallbacks are used, but even if not using fallbacks, the improved check may help with cross-compiling (discussion with Keith Marshall). + other build-fixes for Ada95 with MinGW. + modify Ada95 source-generation utility to write to a file given as parameter rather than to the standard output, allowing builds with MinGW. - Add ncurses patch 20200801 + remove remaining parts of checks for ISC Unix (cf: 20121006). + add user32.lib to LDFLAGS for Visual Studio C++ configuration (discussion with "Maarten Anonymous"). + modify MKkey_defs.sh to hide ncurses' definition of KEY_EVENTS to reduce Visual Studio C++ redefinition warnings. + improve/update checks for external functions in test/configure - Add ncurses patch 20200725 + set LINK_TESTS in CF_SHARED_OPTS for msvc (patch by "Maarten Anonymous") + improved workaround for redefinition-warnings for KEY_EVENT. + improve man/term.5 section on legacy storage format (report by Florian Weimer). - Add ncurses patch 20200718 + reduce redefinition-warnings for KEY_EVENT when building with Visual Studio C++. + define NCURSES_STATIC when compiling programs to link with static libraries, to work with MinGW vs Visual Studio C++. > additional changes for building with Visual Studio C++ and msys2 (reports/patches by "Maarten Anonymous") + modify c++/Makefile.in to set the current directory while compiling the main program, so the linker can find related objects. + several changes to allow the c++/demo program to compile/link. + change an ifdef in test-directory, to use VC++ wide-character funcs. ==== policycoreutils ==== Version update (3.0 -> 3.1) Subpackages: policycoreutils-lang python3-policycoreutils - Add get_os_version.patch get_os_version is implemented in a very RH/Fedora specific way. Ensure that it returns a valid string for SUSE by changing the default. Also remove the RH specific logic when generating HTML versions of the SELinux documentation - Align more with Fedora spec file to get rid of python dependencies in the core system - create new python-utils sub-package - move some tools to devel sub-package - Cleanup dependencies - Proper default permissions for newrole (4755) - Update to version 3.1 * New `setfiles -E` option - treat conflicting specifications as errors, such as where two hardlinks for the same inode have different contexts. * `setsebool -V` reports errors from commit phase * matchpathcon related interfaces are deprecated * New `restorecon -x` option which prevents it from crossing file system * boundaries. * `sepolgen-ifgen` parses a gen_tunable statement as bool * Removed Requires for python3-ipy as the ipaddress module is used. No requires for python-ipaddress as it's assumed this is used only on recent systems * Drop chcat_join.patch, is upstream ==== python-semanage ==== Version update (3.0 -> 3.1) - Update to version 3.1 * Improved manpage * fsync final files before rename - Disabled LTO again. This breaks e.g. shadow and also other packages in security:SELinux - Fix build with LTO: [bsc#1133102] * Enable LTO (Link Time Optimization) * Update map file to include new symbols and remove wildcards - Add libsemanage-update-map-file.patch ==== sysvinit ==== Version update (2.96 -> 2.97) - Drop /bin/pidof and /sbin/pidof, including corresponding man page: let's switch to pidof as provided by procps-ng. - Update to sysvinit 2.97: * Check $(ROOT) filesystem for libcrypt instead of a hardcoded path to /usr. * Code clean-up and making sure we avoid freeing unused memory. * Added shell script which converts systemd unit files into init.d style scripts. * Allow init to load configuration data from files stored in /etc/inittab.d/ * Allow shutdown time to be specified in the format +hh:mm. This is in addition to the existing formats such as hh:mm, +m, and "now". * Fixed typos in manual pages. - Update startpar to 0.65: + Make sure startpar testsuite can find insserv executable in /usr/sbin or /sbin. + Added PREFIX variable to Makefile and testsuite to make location of startpar and insserv more flexible. - Rebase sysvinit-2.90.dif. - Drop SCVER defines: not used in any place. - Drop startpar-sysmacros.patch: fixed upstream. ==== tcpd ==== - tcp_wrappers_7.6-shared-lib.diff: Linux has STRERROR not SYS_ERRLIST [bsc#1175272] ==== x11-tools ==== - xim script: also consider files below /usr/etc/X11/xim.d (boo#1176431) - Upgrade old RPM constructs. - i18n.template: moved example for user specific LANG setting from .profile (bash package) to .i18n skeleton file (boo#1158724) - moved xim files to /usr/etc; xinit is already prepared for this (boo#1176431) ==== xdm ==== Subpackages: xdm-xsession - removed /etc/X11/xdm/Keyboard.map since it meanwhile has been replaced by /usr/share/systemd/kbd-model-map (used by YaST) - Don't create the compat symlinks in /etc on the fly, track them properly as %config - /etc/X11/xdm/scripts/11-ssh-agent * ssh-askpass now moved to /usr/libexec/ssh (boo#1175991) - moved xdm/xdm-np PAM files to /usr/etc/pam.d, since lightdm has been adjusted meanwhile - Switch to /usr/bin/pidof dependency, provided by procps-ng. - fixes in xdm-tarball.patch * /usr/lib/X11/display-manager: fixes sddm, which didn't use xdm_reload_files in xdm_start_proc() of /usr/lib/X11/displaymanagers/sddm (boo#1173049) - fixes in xdm-tarball.patch * symlink also $XDMDIR/xinitrc.common in /etc/X11/xinit; it is still often sourced in old ~/.xinitrc user files ... - fixes in xdm-tarball.patch * fixed include path for generic Xresources in xdm's Xresources file * fixed $XDMDIR in Xsession, so sys.xsession can be found * sys.xsession: fixed syntax error in code, which reads additional xdm scripts * display-manager: also symlink Xstartup and Xreset in /etc/X11/xdm (needed by lighdm) - specfile: * for now use again /etc/pam.d for xdm/xdm-np PAM files instead of /usr/etc/pam.d (using the latter resulted in "authentificataion failure" with lightdm) - reenabled move to /usr/etc/X11/xdm; updated xdm-tarball.patch (includes changes from xinit-UsrEtcMove.patch) [boo#1176212] - xinit-UsrEtcMove.patch * changes needed with xinitrc moving to %{_libexecdir}/xinit and xinitrc.common moving to /usr/etc/X11/xinit/ - reverted move to /usr/etc/X11/xdm for now due to boo#1176212 - updated xdm-tarball.patch * adjusted for changes in xmodmap and xinit (boo#1173049) - xdm-tarball.patch * /usr/etc changes still needed for xdm.tar.bz2; will need more changes once xmodmap and xinit files are also moved to /usr/etc (boo#1173049) - move /etc/X11/xdm --> /usr/etc/X11/xdm and /etc/pam.d -> /usr/etc/pam.d; still TODO: adjust scripts in xdm.tar.bz2 (boo#1173049) - more cleanup * no longer support sle12; adjusted xdm.tar.bz2 and applied xdm-with-update-alternative.patch to this tarball - cleanup * get rid of xdm-fallbacks.tar.bz2, which wasn't used any longer since SUSE 12.1 * no longer support systmed based OSes like SUSE < 12.1 and sle11; removed therefore also xdm-consolekit.diff * no longer support sle11 ==== xinit ==== - Properly track the compatibility symlink - simplified UsrEtcMove enable/disable logic in specfile - xinit-tarball.patch/xinit.spec * reenabled move to /usr/etc/X11/xinit * fixed remaining issues mentioned in boo#1173052, comment#6 and boo#1176212, comment#7 - reverted move to /usr/etc/X11/xinit for now (boo#1173052, comment#6) - moved xinit files to /usr/etc/X11/xinit and removed xinitrc skeleton (boo#1173052) - xinit-tarball.patch * adjust tarball contenct to /usr/etc move (boo#1173052) ==== xmodmap ==== - reenabled move to /usr/etc/X11 for Tumbleweed (boo#1173053) - reverted move to /usr/etc/X11 for now (boo#1173053, comment#3) - moved Xmodmap files to /usr/etc/X11 (boo#1173053) ==== xorg-x11-server ==== Subpackages: xorg-x11-server-Xvfb xorg-x11-server-extra xorg-x11-server-sdk xorg-x11-server-wayland - n_xorg-wrapper-anybody.patch * replace default config /etc/X11/Xwrapper, which allows anybody to use the wrapper, by a patch for the code, i.e. [#] rootonly, console, anybody allowed_users=anybody [#] yes, no, auto needs_root_rights=auto is now the default without any Xwrapper config (needs_root_rights=auto was already the default before) - u_xorg-wrapper-Xserver-Options-Whitelist-Filter.patch * replaced by improved version written by Matthias Gerstner of our security team + simplified the option parsing code a bit + changed the "ignore forbidden argument" logic into an "abort on forbidden argument" logic. This is safer and avoids surprises on the user's end that could occur if the desired command line arguments aren't effective but the Xorg server is still started. + tried to adjust to the coding style present in the file (mostly the function name) + added some logic to apply the option filtering only to non-root users when Xorg is actually started as root. This should allow for full flexibility if root calls the wrapper or if the Xorg server only runs with user privileges. - U_Fix-segfault-on-probing-a-non-PCI-platform-device-on.patch, U_Revert-linux-Fix-platform-device-PCI-detection-for-c.patch, U_Revert-linux-Fix-platform-device-probe-for-DT-based-.patch, U_Revert-linux-Make-platform-device-probe-less-fragile.patch * fix Xserver startup on Raspberry Pi 3 (boo#1176203) - n_xorg-wrapper-rename-Xorg.patch * moved Xorg to Xorg.bin and Xorg.sh to Xorg (boo#1175867) - change default for needs_root_rights to auto in Xwrapper.config (boo#1175867) - reenabled SUID wrapper for TW (boo#1175867) - u_xorg-wrapper-Xserver-Options-Whitelist-Filter.patch * Xserver option whitelist filter (boo#1175867) ==== zsh ==== - Add ncurses-fix.patch in order to fix ncurses failure. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Donnerstag, 8. Oktober 2020, 15:01:13 CEST schrieb Dominique Leuenberger:
glibc (2.31 -> 2.32)
Good bye, OBS, see you in a healthy state tomorrow afternoon again (hopefully.., fingers crossed ;-) Cheers, Pete -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
After this update I could not start Xorg anymore due to the xorg/xinit package changes Now my setup is unusual, as I do not use a DM and just use startx launched in ~/.profile which itself run at console login. I'm documenting the issue I had, for reference. With this update /usr/bin/Xorg is now a shell script running /usr/bin/Xorg.wrap if present or /usr/bin/Xorg.bin otherwise. previously, /usr/bin/Xorg was /usr/bin/Xorg.bin. I had /usr/bin/Xorg setuid root via /etc/permission.local and chkstat, but this obviously did not work anymore since now /usr/bin/Xorg.bin needed to be setuid root instead. Did that and X started successfully but exited immediately with code 0. Investigating further, I found that /etc/X11/xinit/xinitrc had been moved to /usr/libexec/xinit/xinitrc, and since I had my ~/.xinitrc call /etc/X11/xinit/xinitrc, it choked on that (file not found) and updating the path fixed it. Next, I thought it might be a good idea to switch to Xorg.wrap found in package xorg-x11-server-wrapper. So I installed it, removed setuid on /usr/bin/Xorg.bin and startx worked when logged on the console and starting it from there. But it failed when startx was launched in ~/.profile, during console login. Looked at 'man Xrapper.config', created /etc/X11/Xwrapper.conf with allowed_users=anybody But then, no matter what, startx always failed with: /usr/bin/Xorg.wrap: Xserver option ":0" invalid or not in whitelist. Aborting. Thus had to revert to the initial setuid solution. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Michael Pujos composed on 2020-10-08 16:53 (UTC+0200):
After this update I could not start Xorg anymore due to the xorg/xinit package changes
Now my setup is unusual, as I do not use a DM and just use startx launched in ~/.profile which itself run at console login. I'm documenting the issue I had, for reference.
With this update /usr/bin/Xorg is now a shell script running /usr/bin/Xorg.wrap if present or /usr/bin/Xorg.bin otherwise. previously, /usr/bin/Xorg was /usr/bin/Xorg.bin.
I had /usr/bin/Xorg setuid root via /etc/permission.local and chkstat, but this obviously did not work anymore since now /usr/bin/Xorg.bin needed to be setuid root instead.
Did that and X started successfully but exited immediately with code 0. Investigating further, I found that /etc/X11/xinit/xinitrc had been moved to /usr/libexec/xinit/xinitrc, and since I had my ~/.xinitrc call /etc/X11/xinit/xinitrc, it choked on that (file not found) and updating the path fixed it.
Next, I thought it might be a good idea to switch to Xorg.wrap found in package xorg-x11-server-wrapper. So I installed it, removed setuid on /usr/bin/Xorg.bin and startx worked when logged on the console and starting it from there. But it failed when startx was launched in ~/.profile, during console login. Looked at 'man Xrapper.config', created /etc/X11/Xwrapper.conf with
allowed_users=anybody
But then, no matter what, startx always failed with:
/usr/bin/Xorg.wrap: Xserver option ":0" invalid or not in whitelist. Aborting.
Thus had to revert to the initial setuid solution.
Looks like good reason for a regression bug report. Note though that Stefan Dirsch, who usually is responsible for X bug fixes, will usually denigrate startx usage and state that it is "unsupported", even though all other distros I'm familiar with seem not to be of the same persuasion. -- Evolution as taught in public schools, like religion, is based on faith, not on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Thu, Oct 08, 2020 at 02:52:12PM -0400, Felix Miata wrote:
Looks like good reason for a regression bug report. Note though that Stefan Dirsch, who usually is responsible for X bug fixes, will usually denigrate startx usage and state that it is "unsupported", even though all other distros I'm familiar with seem not to be of the same persuasion.
Haha. :-) Yes, that's usually true. But in that case I tried to be helpful. ;-) Already answered to his feeedback ... Thanks, Stefan Public Key available ------------------------------------------------------ Stefan Dirsch (Res. & Dev.) SUSE Software Solutions Germany GmbH Tel: 0911-740 53 0 Maxfeldstraße 5 FAX: 0911-740 53 479 D-90409 Nürnberg http://www.suse.de Germany ---------------------------------------------------------------- (HRB 36809, AG Nürnberg) Geschäftsführer: Felix Imendörffer ---------------------------------------------------------------- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Thu, Oct 08, 2020 at 04:53:49PM +0200, Michael Pujos wrote:
After this update I could not start Xorg anymore due to the xorg/xinit package changes
Now my setup is unusual, as I do not use a DM and just use startx launched in ~/.profile which itself run at console login. I'm documenting the issue I had, for reference.
With this update /usr/bin/Xorg is now a shell script running /usr/bin/Xorg.wrap if present or /usr/bin/Xorg.bin otherwise. previously, /usr/bin/Xorg was /usr/bin/Xorg.bin.
I had /usr/bin/Xorg setuid root via /etc/permission.local and chkstat, but this obviously did not work anymore since now /usr/bin/Xorg.bin needed to be setuid root instead.
Did that and X started successfully but exited immediately with code 0. Investigating further, I found that /etc/X11/xinit/xinitrc had been moved to /usr/libexec/xinit/xinitrc, and since I had my ~/.xinitrc call /etc/X11/xinit/xinitrc, it choked on that (file not found) and updating the path fixed it.
Next, I thought it might be a good idea to switch to Xorg.wrap found in package xorg-x11-server-wrapper. So I installed it, removed setuid on /usr/bin/Xorg.bin and startx worked when logged on the console and starting it from there. But it failed when startx was launched in ~/.profile, during console login. Looked at 'man Xrapper.config', created /etc/X11/Xwrapper.conf with
allowed_users=anybody
But then, no matter what, startx always failed with:
/usr/bin/Xorg.wrap: Xserver option ":0" invalid or not in whitelist. Aborting.
Thus had to revert to the initial setuid solution.
I suggest with your setup to uninstall xorg-x11-server-wrapper, i.e. not to use Xwrapper at all. Background: We introduced it for gdm, so Xserver started by gdm no longer needs to be run as user root. Until now we used a special patch against gdm to start it still as root, which our gdm developers wanted to get rid of. RH/Debian are using Xwrapper since some time for gdm. It sounds weird you're calling xinitrc thru your ~/.xinitrc. The sample .xinitrc in /etc/skel just sources /etc/X11/xinit/xinitrc.common, but I added a compat link for this. When using xinit/startx, /usr/libexec/xinit/xinitrc should be found. Try these first without any ~/.xinitrc. Although we no longer really support starting a Xsession via xinit/startx, at least for simple WMs it should be somewhat possible with a reduced functionality (sound may not work, etc.). So if this is no longer working, I need to fix this.
allowed_users=anybody
This is already the default for our Xwrapper. This doesn't need to be set explicitely. We've added a Xserver option whitelist to Xwrapper for security reasons. Therefore ":0" is not allowed as option (gdm doesn't use/need it). Hope this helps. Thanks for your feedback! CU, Stefan Public Key available ------------------------------------------------------ Stefan Dirsch (Res. & Dev.) SUSE Software Solutions Germany GmbH Tel: 0911-740 53 0 Maxfeldstraße 5 FAX: 0911-740 53 479 D-90409 Nürnberg http://www.suse.de Germany ---------------------------------------------------------------- (HRB 36809, AG Nürnberg) Geschäftsführer: Felix Imendörffer ---------------------------------------------------------------- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Thu, Oct 08, 2020 at 09:39:19PM +0200, Stefan Dirsch wrote:
When using xinit/startx, /usr/libexec/xinit/xinitrc should be found. Try these first without any ~/.xinitrc. Although we no longer really support starting a Xsession via xinit/startx, at least for simple WMs it should be somewhat possible with a reduced functionality (sound may not work, etc.). So if this is no longer working, I need to fix this.
Just had a try with latest TW. xinit and startx works for me. The former starts a xterm, the latter a KDE session. When using the .xinitrc, that we previously shipped in /etc/skel, xinit also starts a KDE session. So I would say: Works for me. ;-) Thanks, Stefan Public Key available ------------------------------------------------------ Stefan Dirsch (Res. & Dev.) SUSE Software Solutions Germany GmbH Tel: 0911-740 53 0 Maxfeldstraße 5 FAX: 0911-740 53 479 D-90409 Nürnberg http://www.suse.de Germany ---------------------------------------------------------------- (HRB 36809, AG Nürnberg) Geschäftsführer: Felix Imendörffer ---------------------------------------------------------------- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 10/8/20 9:39 PM, Stefan Dirsch wrote:
On Thu, Oct 08, 2020 at 04:53:49PM +0200, Michael Pujos wrote:
I suggest with your setup to uninstall xorg-x11-server-wrapper, i.e. not to use Xwrapper at all.
Yup, that's what I did. Xwrapper works fine started from the command-line (in runlevel 3), just not from startx invoked from ~/.profile, itself executed on console login: Xorg fails to start with the same error than when it is not setuid root. This is super minor as I expect the number to call startx in their ~/.profile with xorg-x11-server-wrapper installed to be in the 0.000001% range, but just mentioning it for completeness.
Background: We introduced it for gdm, so Xserver started by gdm no longer needs to be run as user root. Until now we used a special patch against gdm to start it still as root, which our gdm developers wanted to get rid of. RH/Debian are using Xwrapper since some time for gdm. Good to know.
It sounds weird you're calling xinitrc thru your ~/.xinitrc. The sample .xinitrc in /etc/skel just sources /etc/X11/xinit/xinitrc.common, but I added a compat link for this.
When using xinit/startx, /usr/libexec/xinit/xinitrc should be found. Try these first without any ~/.xinitrc. Although we no longer really support starting a Xsession via xinit/startx, at least for simple WMs it should be somewhat possible with a reduced functionality (sound may not work, etc.). So if this is no longer working, I need to fix this.
Again, my setup is a bit unusual with login on tty1 starting Xorg with i3, tty2 starting Plasma, tty3 starting Plasma+i3 (replacing kwin) and tty4 for Gnome. Although these days, I simply use i3. My ~/.xinitrc sets WINDOWMANAGER depending on tty and do a few other things before ultimately calling /usr/libexec/xinit/xinitrc. Full details on my dotfiles repo: https://github.com/bubbleguuum/dotfiles I suppose I could just use SDDM these days (had problem to set it up in the past with multiple monitors and high-dpi displays, but not true anymore), but I like this setup. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Fri, Oct 09, 2020 at 12:23:15PM +0200, Michael Pujos wrote:
On 10/8/20 9:39 PM, Stefan Dirsch wrote:
On Thu, Oct 08, 2020 at 04:53:49PM +0200, Michael Pujos wrote:
I suggest with your setup to uninstall xorg-x11-server-wrapper, i.e. not to use Xwrapper at all.
Yup, that's what I did. Xwrapper works fine started from the command-line (in runlevel 3), just not from startx invoked from ~/.profile, itself executed on console login: Xorg fails to start with the same error than when it is not setuid root. This is super minor as I expect the number to call startx in their ~/.profile with xorg-x11-server-wrapper installed to be in the 0.000001% range, but just mentioning it for completeness.
Ok. This is really a corner case.
It sounds weird you're calling xinitrc thru your ~/.xinitrc. The sample .xinitrc in /etc/skel just sources /etc/X11/xinit/xinitrc.common, but I added a compat link for this.
When using xinit/startx, /usr/libexec/xinit/xinitrc should be found. Try these first without any ~/.xinitrc. Although we no longer really support starting a Xsession via xinit/startx, at least for simple WMs it should be somewhat possible with a reduced functionality (sound may not work, etc.). So if this is no longer working, I need to fix this.
Again, my setup is a bit unusual with login on tty1 starting Xorg with i3, tty2 starting Plasma, tty3 starting Plasma+i3 (replacing kwin) and tty4 for Gnome. Although these days, I simply use i3. My ~/.xinitrc sets WINDOWMANAGER depending on tty and do a few other things before ultimately calling /usr/libexec/xinit/xinitrc. Full details on my dotfiles repo: https://github.com/bubbleguuum/dotfiles
Ah. @bubbleguum. One of our main contributors to suse-prime! :-) This is really a special setup, since ~/.xinitrc is meant to be a *replacement* for /usr/libexec/xinit/xinitrc.
I suppose I could just use SDDM these days (had problem to set it up in the past with multiple monitors and high-dpi displays, but not true anymore), but I like this setup.
I see. The good thing about this is, that people with such a special setup know how to adjust it, if there are any system changes. ;-) I'll see if there are more users, who rely on /etc/X11/xinit/xinitrc. Then I can still add a compat sysmlink. Thanks, Stefan Public Key available ------------------------------------------------------ Stefan Dirsch (Res. & Dev.) SUSE Software Solutions Germany GmbH Tel: 0911-740 53 0 Maxfeldstraße 5 FAX: 0911-740 53 479 D-90409 Nürnberg http://www.suse.de Germany ---------------------------------------------------------------- (HRB 36809, AG Nürnberg) Geschäftsführer: Felix Imendörffer ---------------------------------------------------------------- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On my system snapshot 20201007 breaks the keyboard setup. After starting X11 I have allways an american keyboard until I start yast2 keyboard and configure a german keyboard. Anyone other noticed that , too ? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Op vrijdag 9 oktober 2020 09:46:31 CEST schreef Markus Koßmann:
On my system snapshot 20201007 breaks the keyboard setup. After starting X11 I have allways an american keyboard until I start yast2 keyboard and configure a german keyboard. Anyone other noticed that , too ?
I noticed this under Wayland, but it might have been present under Plasma also, because I used "YaST language" to set my keyboard to "US int" under Wayland and went back to Plasma X11. -- fr.gr. member openSUSE Freek de Kruijf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 09/10/2020 08:46, Markus Koßmann wrote:
On my system snapshot 20201007 breaks the keyboard setup. After starting X11 I have allways an american keyboard until I start yast2 keyboard and configure a german keyboard. Anyone other noticed that , too ?
It happens here on occasions but hasn't happened recently. Regards Sid. -- Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Senior Staff Specialist, Cricket Coach Microsoft Windows Free Zone - Linux used for all Computing Tasks -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
In data venerdì 9 ottobre 2020 13:42:06 CEST, Sid Boyce ha scritto:
On 09/10/2020 08:46, Markus Koßmann wrote:
On my system snapshot 20201007 breaks the keyboard setup. After starting X11 I have allways an american keyboard until I start yast2 keyboard and configure a german keyboard. Anyone other noticed that , too ?
It happens here on occasions but hasn't happened recently.
Regards
Sid. Well, not on the American outlay, but as I am using fcitx I have on every reboot the Chinese one. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi Markus,
On my system snapshot 20201007 breaks the keyboard setup. After starting X11 I have allways an american keyboard until I start yast2 keyboard and configure a german keyboard. Anyone other noticed that , too ?
yep, same here. One has to switch forth and back language, then it gets the correct keyboard. Bye. Michael. -- Michael Hirmke -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
I had the same problem. The worst is, keyboard is changed in the KDE login password dialog. This login dialog has no "eye" button to change from '*' to character display. Is that possible to configure? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi,
I had the same problem.
The worst is, keyboard is changed in the KDE login password dialog. This login dialog has no "eye" button to change from '*' to character display. Is that possible to configure?
this does not happen here. The password dialog uses the correct keyboard setting. The problem starts after having logged in. Bye. Michael. -- Michael Hirmke -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi, I had the same problem, until I added my language (German) to the IBus Panel, i. e. the IBus keyboard selector. Best regards, Daniel -- Daniel Schlieper email: daniel.schlieper@tuxomania.net On Sat, 10 Oct 2020, Michael Hirmke wrote:
Hi,
I had the same problem.
The worst is, keyboard is changed in the KDE login password dialog. This login dialog has no "eye" button to change from '*' to character display. Is that possible to configure?
this does not happen here. The password dialog uses the correct keyboard setting. The problem starts after having logged in.
Bye. Michael. -- Michael Hirmke -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 14.10.20 um 16:28 schrieb Daniel Schlieper:
Hi,
I had the same problem, until I added my language (German) to the IBus Panel, i. e. the IBus keyboard selector.
Best regards, Daniel
-- Daniel Schlieper email: daniel.schlieper@tuxomania.net
On Sat, 10 Oct 2020, Michael Hirmke wrote:
Hi,
I had the same problem.
The worst is, keyboard is changed in the KDE login password dialog. This login dialog has no "eye" button to change from '*' to character display. Is that possible to configure?
this does not happen here. The password dialog uses the correct keyboard setting. The problem starts after having logged in.
Bye. Michael.
FYI: https://bugzilla.suse.com/show_bug.cgi?id=1177545#c22 Regards, Frank -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Freitag, 9. Oktober 2020, 09:46:31 CEST schrieb Markus Koßmann: Opened <https://bugzilla.opensuse.org/show_bug.cgi?id=1177545> -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
El sáb., 10 oct. 2020 a las 6:49, Markus Koßmann (<mkossmann_ml1@gmx.de>) escribió:
Am Freitag, 9. Oktober 2020, 09:46:31 CEST schrieb Markus Koßmann: Opened <https://bugzilla.opensuse.org/show_bug.cgi?id=1177545>
I have trouble after the last update with the sound. In the Yast sound module the sound card appear ok, but in KDE don't appear the sound devices. After reconfiguring the sound with Yast, crashes kded5: [KCrash Handler] #4 0x0000000000000000 in () #5 0x00007f9901371779 in pa_srbchannel_free () at /usr/lib64/pulseaudio/libpulsecommon-13.0.so #6 0x00007f990137051a in pa_pstream_set_srbchannel () at /usr/lib64/pulseaudio/libpulsecommon-13.0.so #7 0x00007f990137063a in pa_pstream_unlink () at /usr/lib64/pulseaudio/libpulsecommon-13.0.so #8 0x00007f99013c0cc5 in () at /usr/lib64/libpulse.so.0 #9 0x00007f99013c0e12 in () at /usr/lib64/libpulse.so.0 #10 0x00007f9901531227 in () at /usr/lib64/libkmixcore.so.5 #11 0x00007f9901531269 in () at /usr/lib64/libkmixcore.so.5 #12 0x00007f9901512486 in Mixer::~Mixer() () at /usr/lib64/libkmixcore.so.5 #13 0x00007f9901517057 in MixerToolBox::deinitMixer() () at /usr/lib64/libkmixcore.so.5 #14 0x00007f99018fcd8c in () at /usr/lib64/qt5/plugins/kf5/kded/kmixd.so #15 0x00005635fd980a84 in () #16 0x00005635fd980c19 in () #17 0x00005635fd97f41f in () #18 0x00007f99308b7e0a in __libc_start_main () at /lib64/libc.so.6 #19 0x00005635fd97fb9a in () [Inferior 1 (process 2120) detached] -- USA LINUX OPENSUSE QUE ES SOFTWARE LIBRE, NO NECESITAS PIRATEAR NADA Y NI TE VAS A PREOCUPAR MAS POR LOS VIRUS Y SPYWARES: http://www.opensuse.org/es/ Puedes visitar mi blog en: http://jerbes.blogspot.com.ar/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (14)
-
CisBug
-
Daniel Schlieper
-
Dominique Leuenberger
-
Felix Miata
-
Frank Krüger
-
Freek de Kruijf
-
Hans-Peter Jansen
-
Juan Erbes
-
Markus Koßmann
-
mh@mike.franken.de
-
Michael Pujos
-
Sid Boyce
-
stakanov.s
-
Stefan Dirsch