Message ID | 20200906175344.5042-1-pbonzini@redhat.com |
---|---|
Headers | show |
Series | Next round of Meson bugfixes and cleanups | expand |
On Mon, Sep 7, 2020 at 2:35 AM Peter Maydell <peter.maydell@linaro.org> wrote: > On Sun, 6 Sep 2020 at 18:56, Paolo Bonzini <pbonzini@redhat.com> wrote: > > > > The following changes since commit > 227de21ed0759e275a469394af72c999d0134bb5: > > > > Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20200903' into > staging (2020-09-05 15:30:41 +0100) > > > > are available in the Git repository at: > > > > https://gitlab.com/bonzini/qemu.git tags/for-upstream > > > > for you to fetch changes up to 6264b35324d3766d3c2ff369c4e8ecba8bd5b571: > > > > meson: remove linkage of sdl to baum (2020-09-06 19:50:57 +0200) > > > > ---------------------------------------------------------------- > > meson related: > > * convert unit tests > > * bugfixes for mtest2make > > * miscellaneous bugfixes > > * dead code removal and configure cleanups > > * oss-fuzz fixes > > * msys fixes > > Build failure, Windows (this is the second "uint" type usage > I've seen today...): > > ../../tests/test-vmstate.c: In function 'int_cmp': > ../../tests/test-vmstate.c:884:5: error: unknown type name 'uint' > uint ua = GPOINTER_TO_UINT(a); > ^ > ../../tests/test-vmstate.c:885:5: error: unknown type name 'uint' > uint ub = GPOINTER_TO_UINT(b); > ^ > I've already fixes this issue, that patch can be quenued > Makefile.ninja:5443: recipe for target > 'tests/test-vmstate.exe.p/test-vmstate.c.obj' failed > make: *** [tests/test-vmstate.exe.p/test-vmstate.c.obj] Error 1 > make: *** Waiting for unfinished jobs.... > ../../tests/test-util-filemonitor.c: In function > 'test_file_monitor_events': > ../../tests/test-util-filemonitor.c:620:17: error: too many arguments > to function 'mkdir' > if (mkdir(pathsrc, 0700) < 0) { > ^ > In file included from > /usr/lib/mxe/usr/x86_64-w64-mingw32.shared/include/unistd.h:10:0, > from > /home/petmay01/qemu-for-merges/include/qemu/osdep.h:93, > from ../../tests/test-util-filemonitor.c:21: > /usr/lib/mxe/usr/x86_64-w64-mingw32.shared/include/io.h:280:15: note: > declared here > int __cdecl mkdir (const char *) __MINGW_ATTRIB_DEPRECATED_MSVC2005; > ^ > > > > Build failure, OpenBSD: > > In file included from ../src/hw/arm/nseries.c:30: > In file included from > /home/qemu/qemu-test.yhbDti/src/include/hw/arm/omap.h:24: > In file included from > /home/qemu/qemu-test.yhbDti/src/include/hw/input/tsc2xxx.h:14: > /home/qemu/qemu-test.yhbDti/src/include/ui/console.h:11:11: fatal > error: 'epoxy/gl.h' file not found > # include <epoxy/gl.h> > ^~~~~~~~~~~~ > 1 error generated. > gmake: *** [Makefile.ninja:1735: > libqemu-aarch64-softmmu.fa.p/hw_arm_nseries.c.o] Error 1 > > > > Odd warning on most but not all of the builds, though they went on to > complete OK: > > make: Entering directory '/home/peter.maydell/qemu-freebsd/build' > /home/peter.maydell/qemu-freebsd/tests/Makefile.include:144: warning: > overriding recipe for target 'check-block' > Makefile.mtest:1345: warning: ignoring old recipe for target 'check-block' > config-host.mak is out-of-date, running configure > cross containers no > > NOTE: guest cross-compilers enabled: aarch64-linux-gnu-gcc cc > aarch64-linux-gnu-gcc cc > /usr/bin/python3 /home/peter.maydell/qemu-freebsd/meson/meson.py > --internal regenerate /home/peter.maydell/qemu-freebsd > /home/peter.maydell/qemu-freebsd/build --backend ninja > The Meson build system > Version: 0.55.1 > [etc] > > > x86-64 clang build failed at the link stage (this is config we've > talked about before with > '../../configure' '--cc=clang' '--cxx=clang++' '--enable-gtk' > '--extra-cflags=-fsanitize=undefined -fno-sanitize=shift-base > -Werror' > but where the clang++ doesn't work because the right libstdc++ happens > not to be present): > > Linking target qemu-alpha > libcommon.fa.p/cpus-common.c.o: In function `cpu_list_add': > /home/petmay01/linaro/qemu-for-merges/build/clang/../../cpus-common.c:80: > undefined reference to `__ubsan_handle_type_mismatch_v1' > /home/petmay01/linaro/qemu-for-merges/build/clang/../../cpus-common.c:80: > undefined reference to `__ubsan_handle_type_mismatch_v1' > libcommon.fa.p/cpus-common.c.o: In function `cpu_get_free_index': > /home/petmay01/linaro/qemu-for-merges/build/clang/../../cpus-common.c:68: > undefined reference to `__ubsan_handle_type_mismatch_v1' > /home/petmay01/linaro/qemu-for-merges/build/clang/../../cpus-common.c:68: > undefined reference to `__ubsan_handle_type_mismatch_v1' > /home/petmay01/linaro/qemu-for-merges/build/clang/../../cpus-common.c:67: > undefined reference to `__ubsan_handle_type_mismatch_v1' > > libcommon.fa.p/cpus-common.c.o:/home/petmay01/linaro/qemu-for-merges/build/clang/../../cpus-common.c:67: > more undefined references to `__ubsan_handle_type_mismatch_v1' follow > collect2: error: ld returned 1 exit status > > thanks > -- PMM > >
On 11/09/20 13:05, Max Reitz wrote: > On closer inspection it seems like it’s because of the > “build_by_default: false”, which seems like a rather conscious decision. > Was I only lucky that the socket_scm_helper was built by default so > far? Should I have explicitly built it all this time? Yes, you were lucky but that's not a reason not to change it. The problem with touching a whole build system is that you are not going to test everybody's usecase, and yours is definitely reasonable and not "M-x butterfly" territory. Paolo