On Mon, Aug 7, 2023 at 7:19 AM Arsen Arsenović via Gcc-patches <gcc-patches@gcc.gnu.org> wrote: > > Hello, > > This patch set, combined with a sibling patch set sent on the binutils > and GDB MLs, bring up the shared infrastructure between the two projects > in sync again. > > It largely consists of cherry-picks from various people which have been > reviewed and accepted on the binutils-gdb side, as well as a couple of > patches for differences that seemed to get lost during the > pick-and-regenerate process, though I've had to reformat quite a few > commits so that they match what the changelog verifier expects. > > Some of these commits were previously applied and then reverted, namely > "Check if AR works with --plugin and rc", but the reversion reason[1] > seems to be no longer valid as of requiring Binutils 2.35 for LTO > anyway. > > In addition, I'm not entirely certain that the handling of the removed > targets I re-added is correct, nor what their status in binutils and gdb > is, so feedback is welcome there. > > During this process, it appears that I overlooked passing -x to > cherry-pick, so the paper trail got lost. If needed, I can hack > together some code to associate commits based on subjects with their > pair in the other repository, and amend the commits with these > references added. > > These patches ignore the intl/ directory, as I plan to follow up this > patchset with one that can be cleanly applied to both trees which gets > rid of the intl/ directory in favor of out-of-tree gettext (ISL, GMP, et > al. style). > > Bootstrapped and reg-tested on x86_64-pc-linux-gnu. > > OK for master? > > Thanks in advance, have a lovely day. > > [1]: > https://inbox.sourceware.org/gcc-patches/CAMe9rOqFSOSM5Sw5SBMiy20rdgRJkftUXdw=ykexhvdmuty...@mail.gmail.com/ > > Alan Modra (4): > PR29961, plugin-api.h: "Could not detect architecture endianess" > gcc-4.5 build fixes > egrep in binutils > PR27116, Spelling errors found by Debian style checker > > Alexander von Gluck IV (1): > Add support for the haiku operating system > > Arsen Arsenović (3): > toplevel: Substitute GDCFLAGS instead of using CFLAGS > toplevel: Recover tilegx/tilepro targets > configure: reinstate 32b PA-RISC HP-UX target in toplevel > > Christophe Lyon (1): > configure: require libzstd >= 1.4.0 > > Fangrui Song (1): > binutils, gdb: support zstd compressed debug sections > > H.J. Lu (4): > Sync with binutils: GCC: Pass --plugin to AR and RANLIB > GCC: Check if AR works with --plugin and rc > PKG_CHECK_MODULES: Check if $pkg_cv_[]$1[]_LIBS works > PKG_CHECK_MODULES: Properly check if $pkg_cv_[]$1[]_LIBS works > > Indu Bhagat (2): > bfd: linker: merge .sframe sections > toplevel: Makefile.def: add install-strip dependency on libsframe > > John Ericson (1): > Deprecate a.out support for NetBSD targets. > > Luis Machado (1): > Disable year 2038 support on 32-bit hosts by default > > Martin Liska (1): > add --enable-default-compressed-debug-sections-algorithm configure > option > > Nick Alcock (3): > libtool.m4: fix nm BSD flag detection > libtool.m4: fix the NM="/nm/over/here -B/option/with/path" case > libtool.m4: augment symcode for Solaris 11 > > Simon Marchi (1): > Pass PKG_CONFIG_PATH down from top-level Makefile > > Vladimir Mezentsev (1): > gprofng: a new GNU profiler > > Makefile.def | 18 ++ > Makefile.in | 517 ++++++++++++++++++++++++++++++++++++- > Makefile.tpl | 10 +- > config/gcc-plugin.m4 | 40 +++ > config/lib-ld.m4 | 8 +- > config/override.m4 | 2 +- > config/picflag.m4 | 4 +- > config/pkg.m4 | 8 + > config/zstd.m4 | 23 ++ > configure | 224 +++++++++++++++- > configure.ac | 74 +++++- > fixincludes/configure | 3 +- > gcc/configure | 140 ++++++---- > include/collectorAPI.h | 73 ++++++ > include/libcollector.h | 89 +++++++ > include/libfcollector.h | 42 +++ > include/plugin-api.h | 49 ++-- > include/xtensa-dynconfig.h | 2 - > intl/configure | 4 +- > libada/configure | 4 +- > libatomic/configure | 130 ++++++---- > libbacktrace/configure | 130 ++++++---- > libcc1/configure | 132 ++++++---- > libcpp/configure | 4 +- > libffi/configure | 132 ++++++---- > libgcc/configure | 6 +- > libgfortran/configure | 132 ++++++---- > libgm2/configure | 132 ++++++---- > libgomp/configure | 132 ++++++---- > libiberty/Makefile.in | 5 +- > libiberty/aclocal.m4 | 1 + > libiberty/configure | 144 ++++++++++- > libiberty/configure.ac | 12 + > libitm/configure | 132 ++++++---- > libobjc/configure | 130 ++++++---- > libphobos/configure | 130 ++++++---- > libquadmath/configure | 130 ++++++---- > libsanitizer/configure | 132 ++++++---- > libssp/configure | 130 ++++++---- > libstdc++-v3/configure | 148 +++++++---- > libtool.m4 | 130 ++++++---- > libvtv/configure | 132 ++++++---- > lto-plugin/configure | 130 ++++++---- > zlib/configure | 130 ++++++---- > 44 files changed, 2924 insertions(+), 956 deletions(-) > create mode 100644 config/zstd.m4 > create mode 100644 include/collectorAPI.h > create mode 100644 include/libcollector.h > create mode 100644 include/libfcollector.h > > -- > 2.41.0 >
Hi, with the updates to libtool.m4 and pkg.m4, those files originally come from upstream libtool and pkg-config, correct? Won't patching GCC's local copies make re-syncing them with upstream libtool/pkg-config more difficult, or have these patches already been sent there, too? Also, when updating .m4 files, aren't you supposed to increment the serial number that they have near the top, too?
