Re: i386 autopkgtest [Re: i386 in focal: an update]
Hello, On Tue, Dec 10, 2019 at 4:04 PM Steve Langasek wrote: > > team, and 68% still need investigation. Details, including a list of > packages to investigate, produced by the attached script: > > success:74 > blacklisted:26 > failed, needs investigation:211 > krb5 I grabbed krb5 for a look, and I think it boils down to missing Multi-Arch in krb5-config (src: kerberos-configs) and krb5-doc (src: krb5). -- ubuntu-devel mailing list ubuntu-devel@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
Re: i386 autopkgtest [Re: i386 in focal: an update]
A further update - after some judicious pruning of the build-dependencies of gst-plugins-bad1.0, the total number of i386 packages for autopkgtesting is reduced from 311 to 299 (attached). Here is the new list of packages with autopkgtests that warrant investigation. We now have 29% of the i386 autopkgtests passing, 24% confirmed not testable, and 47% still needing investigation. I have also now added 'badtest' hints for the current version of each of these remaining packages, so as of now they should no longer be causing proposed-migration delays for any other packages. As you upload new versions of these packages, please make sure to do the analysis of the autopkgtest failures and work with the release team to either fix the tests or have them permanently blacklisted, as appropriate. success:86 blacklisted:73 failed, needs investigation:140 acl alsa-lib ann bind9 binutils boost1.67 brotli bubblewrap bzip2 capnproto cluster-glue cmake colord confuse cunit cups cython datefudge dbus devscripts dietlibc dlm docbook-to-man double-conversion doxygen fakeroot faketime fenix fenix-plugins fftw3 freeglut fribidi game-data-packager gcc-8 gcc-9 gcc-snapshot gdbm gem2deb glibc gnupg2 gpgme1.0 graphite2 grep gtk+3.0 gvfs gzip ibus icu inetutils init-system-helpers iproute2 jemalloc krb5 kronosnet lapack lbfgsb libarchive libassuan libcaca libcap2 libdaemon libdatrie libgd2 libgpg-error libgphoto2 libiptcdata libjsoncpp libkate libmicrohttpd libnfs libofa libomxil-bellagio libpipeline libpsl libqb libraw libsass libsdl2 libsdl-sge libseccomp libsoxr libthai libtommath libusb-1.0 libutempter libuv1 libvisual libvncserver libxcb libxml2 livecd-rootfs llvm-toolchain-9 matplotlib mir multipath-tools ncurses network-manager nghttp2 nss-wrapper numpy openjdk-lts openssh openssl pacemaker perl pillow pinentry postgresql-12 protobuf pupnp-1.8 python2.7 python3.7 python3.8 python-apt python-bcrypt python-cffi python-coverage python-evdev python-numpy python-scipy pyyaml re2 ruby2.5 ruby-json ruby-nokogiri samba sane-backends sdlgfx stunnel4 systemd systemtap talloc tdb ubuntu-drivers-common uchardet udisks2 valgrind vim xterm zsh On Tue, Dec 10, 2019 at 11:03:30AM -0800, Steve Langasek wrote: > Hi all, > > After a few days of bugfixing tests to be compatible with the new > autopkgtest cross-arch infrastructure, I thought it would be useful to give > folks some stats, and point you to a list of packages requiring further > investigation that you can help drive down. > > Of 311 autopkgtests that we care about on i386, 24% are currently passing, > 8% have been confirmed as not-cross-testable and blacklisted by the release > team, and 68% still need investigation. Details, including a list of > packages to investigate, produced by the attached script: > > success: 74 > blacklisted: 26 > failed, needs investigation: 211 > acl > alglib > alsa-lib > ann > bind9 > binutils > boost1.67 > brotli > bubblewrap > bzip2 > capnproto > cfitsio > cluster-glue > cmake > colord > confuse > coreutils > cunit > cups > cython > datefudge > dbus > devscripts > dietlibc > distro-info > dlm > docbook-to-man > double-conversion > doxygen > faketime > fenix > fenix-plugins > fftw3 > fpc > freeglut > fribidi > fyba > game-data-packager > gcc-8 > gcc-9 > gcc-snapshot > gdbm > gem2deb > gl2ps > glibc > gnupg2 > gpgme1.0 > graphite2 > grep > gtk+2.0 > gtk+3.0 > gvfs > gzip > harfbuzz > ibus > icu > inetutils > init-system-helpers > iproute2 > jemalloc > json-glib > krb5 > kronosnet > lapack > lbfgsb > libanyevent-perl > libarchive > libassuan > libasync-interrupt-perl > libb-hooks-op-check-perl > libbsd-resource-perl > libcaca > libcap2 > libclone-perl > libcommon-sense-perl > libdaemon > libdatrie > libdevel-callchecker-perl > libevent-perl > libev-perl > libfile-libmagic-perl > libfilter-perl > libgd2 > libgd-perl > libgpg-error > libgphoto2 > libhtml-parser-perl > libinput > libio-pty-perl > libiptcdata > libjsoncpp > libkate > liblist-moreutils-perl > liblocale-gettext-perl > libmicrohttpd > libnetaddr-ip-perl > libnet-ssleay-perl > libnfs > libofa > libomxil-bellagio > libopenmpt > libparams-classify-perl > libparams-util-perl > libparams-validate-perl > libperlio-gzip-perl > libpipeline > libproxy > libpsl > libqb > libraw > libsass > libsdl2 > libsdl-sge > libseccomp > libsecret > libsocket6-perl > libsort-key-perl > libsoup2.4 > libsoxr > libssh > libsub-identify-perl > libsub-name-perl > libtaint-runtime-perl > libtest-leaktrace-perl > libtest-taint-perl > libtext-charwidth-perl > libtext-iconv-perl > libthai > libtommath > libunicode-utf8-perl > libusb-1.0 > libutempter > libuuid-perl > libuv1 > libvariable-magic-perl > libvisual > libvncserver > libvorbis > libwacom > libwant-perl > libxcb > libxml2 > libxml-libxml-perl > libxml-parser-perl > libyaml-libyaml-perl > linux > livecd-rootfs > llvm-toolchain-9 > matplotlib > mir > mpi4py >
i386 autopkgtest [Re: i386 in focal: an update]
Hi all, After a few days of bugfixing tests to be compatible with the new autopkgtest cross-arch infrastructure, I thought it would be useful to give folks some stats, and point you to a list of packages requiring further investigation that you can help drive down. Of 311 autopkgtests that we care about on i386, 24% are currently passing, 8% have been confirmed as not-cross-testable and blacklisted by the release team, and 68% still need investigation. Details, including a list of packages to investigate, produced by the attached script: success:74 blacklisted:26 failed, needs investigation:211 acl alglib alsa-lib ann bind9 binutils boost1.67 brotli bubblewrap bzip2 capnproto cfitsio cluster-glue cmake colord confuse coreutils cunit cups cython datefudge dbus devscripts dietlibc distro-info dlm docbook-to-man double-conversion doxygen faketime fenix fenix-plugins fftw3 fpc freeglut fribidi fyba game-data-packager gcc-8 gcc-9 gcc-snapshot gdbm gem2deb gl2ps glibc gnupg2 gpgme1.0 graphite2 grep gtk+2.0 gtk+3.0 gvfs gzip harfbuzz ibus icu inetutils init-system-helpers iproute2 jemalloc json-glib krb5 kronosnet lapack lbfgsb libanyevent-perl libarchive libassuan libasync-interrupt-perl libb-hooks-op-check-perl libbsd-resource-perl libcaca libcap2 libclone-perl libcommon-sense-perl libdaemon libdatrie libdevel-callchecker-perl libevent-perl libev-perl libfile-libmagic-perl libfilter-perl libgd2 libgd-perl libgpg-error libgphoto2 libhtml-parser-perl libinput libio-pty-perl libiptcdata libjsoncpp libkate liblist-moreutils-perl liblocale-gettext-perl libmicrohttpd libnetaddr-ip-perl libnet-ssleay-perl libnfs libofa libomxil-bellagio libopenmpt libparams-classify-perl libparams-util-perl libparams-validate-perl libperlio-gzip-perl libpipeline libproxy libpsl libqb libraw libsass libsdl2 libsdl-sge libseccomp libsecret libsocket6-perl libsort-key-perl libsoup2.4 libsoxr libssh libsub-identify-perl libsub-name-perl libtaint-runtime-perl libtest-leaktrace-perl libtest-taint-perl libtext-charwidth-perl libtext-iconv-perl libthai libtommath libunicode-utf8-perl libusb-1.0 libutempter libuuid-perl libuv1 libvariable-magic-perl libvisual libvncserver libvorbis libwacom libwant-perl libxcb libxml2 libxml-libxml-perl libxml-parser-perl libyaml-libyaml-perl linux livecd-rootfs llvm-toolchain-9 matplotlib mir mpi4py multipath-tools ncurses network-manager nghttp2 nss-wrapper numpy openjdk-lts openssh openssl pacemaker perl pillow pinentry poppler postgresql-12 protobuf pulseaudio pupnp-1.8 python2.7 python3.7 python3.8 python-apt python-bcrypt python-cffi python-coverage python-cryptography python-evdev python-lz4 python-nacl python-numpy python-scipy py-ubjson pyyaml raqm re2 ruby2.5 ruby-json ruby-nokogiri samba sane-backends sdlgfx sed stunnel4 superlu systemd systemtap talloc tdb twisted ubuntu-drivers-common uchardet udisks2 upower util-linux valgrind vtk6 vtk7 xterm zope.interface zsh Cheers, -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. Ubuntu Developer https://www.debian.org/ slanga...@ubuntu.com vor...@debian.org On Fri, Dec 06, 2019 at 03:57:21PM -0800, Steve Langasek wrote: > Hi folks, > > The changes to i386 have proceeded as described in my earlier email. The > autopkgtest infrastructure has been converted to run tests of i386 binaries > on an amd64 VM, and i386 binaries not in the whitelist are now in the > process of being removed. As of this moment, there are 1,679 source > packages in the whitelist for i386, and 8,921 source packages with published > i386 binary packages in focal, vs. 14,344 source packages with published > amd64 binaries. At the current rate, I expect the binary removals to > complete some time tomorrow. > > On the autopkgtest side, there are now a total of 311 source packages > producing i386 binaries which also have autopkgtests[1]. Any packages > reporting autopkgtest regressions on i386 which are not on this list should > have their test failures ignored; if you see i386 autopkgtest failures for > other packages blocking migrations out of focal-proposed, please ask a > member of the release team to override. > > It's good that the number of remaining autopkgtests for i386 is as small as > it is, because while the /infrastructure/ is now capable of cross-testing > i386 packages on amd64, the vast majority of these tests don't run correctly > in such an environment without modification. I have triggered runs of all > of these tests on the amd64 testbed, so current status of the tests is > visible on autopkgtest.ubuntu.com for each package. > > Since these tests have also de facto regressed in the release pocket due to > changes to the infrastructure, it is appropriate for those test failures to > be ignored for proposed-migration purposes as well. I
