configure | 8 ++----- docs/meson.build | 14 +++++------ meson | 2 +- meson.build | 54 ++++++++++++++++++++----------------------- plugins/meson.build | 4 ++-- scripts/mtest2make.py | 7 ++---- target/hexagon/meson.build | 3 ++- tests/qapi-schema/meson.build | 4 ++-- tests/qtest/meson.build | 2 +- tests/unit/meson.build | 2 +- trace/meson.build | 4 ++-- 11 files changed, 47 insertions(+), 57 deletions(-)
The following changes since commit 831aaf24967a49d7750090b9dcfd6bf356f16529: Merge remote-tracking branch 'remotes/marcandre/tags/misc-pull-request' into staging (2021-09-14 18:14:56 +0100) are available in the Git repository at: https://gitlab.com/bonzini/qemu.git tags/for-upstream for you to fetch changes up to ccc3f971c37bd2202a21abc9f0be093f10426364: hexagon: use env keyword argument to pass PYTHONPATH (2021-09-15 09:12:55 +0200) ---------------------------------------------------------------- * Update Meson to 0.58.2 ---------------------------------------------------------------- Paolo Bonzini (3): meson: bump submodule to 0.58.2 meson: switch minimum meson version to 0.58.2 hexagon: use env keyword argument to pass PYTHONPATH configure | 8 ++----- docs/meson.build | 14 +++++------ meson | 2 +- meson.build | 54 ++++++++++++++++++++----------------------- plugins/meson.build | 4 ++-- scripts/mtest2make.py | 7 ++---- target/hexagon/meson.build | 3 ++- tests/qapi-schema/meson.build | 4 ++-- tests/qtest/meson.build | 2 +- tests/unit/meson.build | 2 +- trace/meson.build | 4 ++-- 11 files changed, 47 insertions(+), 57 deletions(-) -- 2.31.1
On Sat, 18 Sept 2021 at 10:34, Paolo Bonzini <pbonzini@redhat.com> wrote: > > The following changes since commit 831aaf24967a49d7750090b9dcfd6bf356f16529: > > Merge remote-tracking branch 'remotes/marcandre/tags/misc-pull-request' into staging (2021-09-14 18:14:56 +0100) > > are available in the Git repository at: > > https://gitlab.com/bonzini/qemu.git tags/for-upstream > > for you to fetch changes up to ccc3f971c37bd2202a21abc9f0be093f10426364: > > hexagon: use env keyword argument to pass PYTHONPATH (2021-09-15 09:12:55 +0200) > > ---------------------------------------------------------------- > * Update Meson to 0.58.2 > > ---------------------------------------------------------------- > Paolo Bonzini (3): > meson: bump submodule to 0.58.2 > meson: switch minimum meson version to 0.58.2 > hexagon: use env keyword argument to pass PYTHONPATH So, I tried merging this, ran into some "is this an issue with this pullreq or is it just an intermittent or infrastructure" issues, and decided to postpone the merge for a bit and retry it later. It has made an absolute mess of my incremental build setups. They now all fail with errors like this, even after having blown away the build directory and re-created it: Darwin manooth.archaic.org.uk 19.6.0 x86_64 make: Entering directory '/Users/pm215/src/qemu-for-merges/build/all' /usr/local/bin/ninja build.ninja && touch build.ninja.stamp GIT ui/keycodemapdb meson tests/fp/berkeley-testfloat-3 tests/fp/berkeley-softfloat-3 dtc capstone slirp [0/1] Regenerating build files. Traceback (most recent call last): File "/Users/pm215/src/qemu-for-merges/meson/mesonbuild/mesonmain.py", line 140, in run return options.run_func(options) File "/Users/pm215/src/qemu-for-merges/meson/mesonbuild/msetup.py", line 245, in run app.generate() File "/Users/pm215/src/qemu-for-merges/meson/mesonbuild/msetup.py", line 154, in generate env = environment.Environment(self.source_dir, self.build_dir, self.options) File "/Users/pm215/src/qemu-for-merges/meson/mesonbuild/environment.py", line 523, in __init__ self.coredata = coredata.load(self.get_build_dir()) File "/Users/pm215/src/qemu-for-merges/meson/mesonbuild/coredata.py", line 1016, in load obj = pickle.load(f) ModuleNotFoundError: No module named 'mesonbuild.mesonlib.universal'; 'mesonbuild.mesonlib' is not a package FAILED: build.ninja Is there anything that can be done to make meson version bumps not a horrific pain to back out ? This seems to go wrong pretty much every time. thanks -- PMM
On Tue, Sep 21, 2021 at 01:40:36PM +0100, Peter Maydell wrote: > On Sat, 18 Sept 2021 at 10:34, Paolo Bonzini <pbonzini@redhat.com> wrote: > > > > The following changes since commit 831aaf24967a49d7750090b9dcfd6bf356f16529: > > > > Merge remote-tracking branch 'remotes/marcandre/tags/misc-pull-request' into staging (2021-09-14 18:14:56 +0100) > > > > are available in the Git repository at: > > > > https://gitlab.com/bonzini/qemu.git tags/for-upstream > > > > for you to fetch changes up to ccc3f971c37bd2202a21abc9f0be093f10426364: > > > > hexagon: use env keyword argument to pass PYTHONPATH (2021-09-15 09:12:55 +0200) > > > > ---------------------------------------------------------------- > > * Update Meson to 0.58.2 > > > > ---------------------------------------------------------------- > > Paolo Bonzini (3): > > meson: bump submodule to 0.58.2 > > meson: switch minimum meson version to 0.58.2 > > hexagon: use env keyword argument to pass PYTHONPATH > > So, I tried merging this, ran into some "is this an issue with this > pullreq or is it just an intermittent or infrastructure" issues, > and decided to postpone the merge for a bit and retry it later. > > It has made an absolute mess of my incremental build setups. > They now all fail with errors like this, even after having blown > away the build directory and re-created it: Might want to 'git submodule deinit --all --force' to purge all currently checked out submodules, so it'll then re-init them in a clean (older) state. In theory this should be handled by the git-submodule.sh script we have in QEMU, but 'deinit' is a big hammer to try if wierd things appear to be happening despite cleaning the build dir. > Is there anything that can be done to make meson version bumps > not a horrific pain to back out ? This seems to go wrong pretty > much every time. Save your previous 'build' directory contents somewhere safe and record the submodule checkout hashes, so they can be reset to that ? Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
On Tue, 21 Sept 2021 at 13:48, Daniel P. Berrangé <berrange@redhat.com> wrote: > > On Tue, Sep 21, 2021 at 01:40:36PM +0100, Peter Maydell wrote: > > Is there anything that can be done to make meson version bumps > > not a horrific pain to back out ? This seems to go wrong pretty > > much every time. > > Save your previous 'build' directory contents somewhere safe > and record the submodule checkout hashes, so they can be > reset to that ? I want this fixed on the Meson end, because if I run into this kind of thing while testing and then backing out pull requests then it seems quite likely that it will also happen to people as they try to bisect across the merge once it eventually gets into git. -- PMM
On 21/09/21 14:40, Peter Maydell wrote: > Is there anything that can be done to make meson version bumps > not a horrific pain to back out ? This seems to go wrong pretty > much every time. Downgrading works fine in 0.58 and newer releases, but the fix is not in 0.55.x. The reason why it goes wrong every time is that we've never updated past a version that has the fix. What were the issues? Paolo
On 21/09/21 14:40, Peter Maydell wrote: > even after having blown > away the build directory and re-created it: Doesn't look like you fully blew away the build directory though. Paolo
On Tue, 21 Sept 2021 at 13:40, Peter Maydell <peter.maydell@linaro.org> wrote: > > On Sat, 18 Sept 2021 at 10:34, Paolo Bonzini <pbonzini@redhat.com> wrote: > > > > The following changes since commit 831aaf24967a49d7750090b9dcfd6bf356f16529: > > > > Merge remote-tracking branch 'remotes/marcandre/tags/misc-pull-request' into staging (2021-09-14 18:14:56 +0100) > > > > are available in the Git repository at: > > > > https://gitlab.com/bonzini/qemu.git tags/for-upstream > > > > for you to fetch changes up to ccc3f971c37bd2202a21abc9f0be093f10426364: > > > > hexagon: use env keyword argument to pass PYTHONPATH (2021-09-15 09:12:55 +0200) > > > > ---------------------------------------------------------------- > > * Update Meson to 0.58.2 > > > > ---------------------------------------------------------------- > > Paolo Bonzini (3): > > meson: bump submodule to 0.58.2 > > meson: switch minimum meson version to 0.58.2 > > hexagon: use env keyword argument to pass PYTHONPATH > > So, I tried merging this, ran into some "is this an issue with this > pullreq or is it just an intermittent or infrastructure" issues, > and decided to postpone the merge for a bit and retry it later. I had another go at this. As far as I can tell it causes all the vm-build-{openbsd,netbsd,freebsd} build-and-tests to hang. At any rate the VMs are sat there eating host CPU. Does 'make vm-build-freebsd' etc work for you ? -- PMM
On 22/09/21 20:22, Peter Maydell wrote: > I had another go at this. As far as I can tell it causes all > the vm-build-{openbsd,netbsd,freebsd} build-and-tests to hang. > At any rate the VMs are sat there eating host CPU. > Does 'make vm-build-freebsd' etc work for you ? It has been hanging for some time but I hadn't debugged it until now. It's caused by a missing dependency that causes bios-tables-test to fail (badly). It's plausible that the meson upgrade triggered a different build ordering for make/ninja, and made it hang for you as well. There was also a related bug that caused the test to hang if bzip2 was not available. Paolo
© 2016 - 2024 Red Hat, Inc.