[opensuse-factory] reproducible builds status 2023-02
Hi, Did some patches this month. Nothing spectacular. Here are the autogenerated bits: last month's status: https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/R... Last months' reproducible builds project updates (including my work): https://reproducible-builds.org/reports/2023-01/ I uploaded https://rb.zq1.de/compare.factory-20230301/ today https://rb.zq1.de/spec/glossar.txt explains the meaning of below values: total-packages: 14802 (+61) build-tried: 14744 (+63) build-failed: 30 (+3) build-n-a: 255 (+29) build-succeeded: 14459 (+30) build-official-failed+na: 173 (+67) build-compare-failed: 468 (-6) build-compare-succeeded: 13991 (+36) verify-failed: 182 (+74) verified-semi-reproducible: 10913 (+246) bit-by-bit-identical: 13849 (+154) not-bit-by-bit-identical: 615 (-119) not-bit-by-bit-identicalcheck: 610 (-124) https://rb.zq1.de/compare.factory-20230301/graph.png shows the change over time https://rb.zq1.de/compare.factory-20230301/unreproduciblerings.txt lists very unreproducible core packages (bootstrap+DVD) Of the badly unreproducible packages, 5 were in ring0 53 were in ring1 That makes it 58/3316 => 1.75 % which is below the overall average of 468/14459 => 3.24 % 615/14459 => 4.25 % of packages are not perfectly reproducible package notes: ==> ./boost/.rb.notes <== = https://github.com/boostorg/boost/issues/741 2038-date influences order +++ new /usr/lib64/libboost_log.so.1.81.0 (objdump) ==> ./buildkit/.rb.notes <== go1.19 toolchain parallelism ==> ./dnscrypt-proxy/.rb.notes <== .src.rpm captures randomness from %vendored_licenses_packager_prep -vendored_licenses_packager install "/tmp/tmp_vendored_dir_6bSEoMWEIw" /home/abuild/rpmbuild/BUILDROOT/dnscrypt-proxy-2.1.3-0.0.x86_64""/usr/share/licenses/dnscrypt-proxy/vendored"" ==> ./docker-compose/.rb.notes <== go1.19 toolchain build-id parallelism ==> ./dpdk/.rb.notes <== parallelism in /usr/share/doc/packages/dpdk/dpdk/html/searchindex.js => SR 1067125 dpdk clean Sphinx leftover files ==> ./ebumeter/.rb.notes <== => SR 1066747 CPU: -march=native from source/Makefile:CXXFLAGS += -march=native ==> ./elementary-xfce-icon-theme/.rb.notes <== = https://gitlab.gnome.org/GNOME/pango/-/issues/729 toolchain pango ASLR+readdir order --- old /usr/share/icons/elementary-xfce/apps/24/accessories-dictionary.png (hex) ==> ./findbugs/.rb.notes <== date+time gets injected into docs and .jar: ==> ./firecracker/.rb.notes <== => https://github.com/firecracker-microvm/firecracker/pull/3445 = https://github.com/firecracker-microvm/firecracker/issues/3439 hashmap order ==> ./galera-4/.rb.notes <== FTBFS-2037 found range good=2179559341 bad=2179561266 (2039-01-25T09:41:06) ==> ./gnutls/.rb.notes <== date in man - probably because patch updates mtime # filterdiff zcat R*/usr/share/man/man1/gnutls-cli.1.gz ==> ./jhead/.rb.notes <== = https://bugzilla.opensuse.org/show_bug.cgi?id=1208386 gcc12 toolchain random tmpdir from LTO /usr/bin/jhead differs in ELF sections ==> ./ldas-tools-framecpp/.rb.notes <== doxygen doc toolchain, filesys ==> ./libhugetlbfs/.rb.notes <== => SR 1068331 libhugetlbfs drop unreproducible unused file = https://bugzilla.opensuse.org/show_bug.cgi?id=1159558 report unreproducible .ldscript file ==> ./nut/.rb.notes <== date in /usr/share/doc/packages/nut/cables/Makefile ==> ./ongres-stringprep/.rb.notes <== xmvn toolchain random UUID, javadoc ==> ./openEMS/.rb.notes <== octave debuginfo, toolchain issue? ==> ./osgi-service-log/.rb.notes <== java date+time +++ new//usr/share/java/osgi-service-log/org.osgi.service.log.jar/META-INF/maven/org.osgi/org.osgi.service.log/pom.properties 2023-01-29 07:14:54.625684100 +0000 ==> ./pentaho-libxml/.rb.notes <== javadoc toolchain ==> ./perl-DBD-MariaDB/.rb.notes <== = https://bugzilla.opensuse.org/show_bug.cgi?id=1203310 mariadb toolchain FTBFS-2038 strings /usr/sbin/mariadbd|grep dates.later ==> ./postfish/.rb.notes <== postfish-wisdomrc random order from ==> ./prosody/.rb.notes <== => SR 1066043 /etc/prosody/certs/example.com.crt and other random key+certs - still in version 0.12.2 ==> ./python-aiohttp/.rb.notes <== => https://github.com/aio-libs/aiohttp/pull/7191 FTBFS-2039 found range good=2177495253 bad=2177497178 (2039-01-01T12:19:38) ==> ./python-azure-agent/.rb.notes <== python3.10 toolchain .pyc ASLR ==> ./python-azure-ai-ml/.rb.notes <== python3.10 toolchain .pyc ASLR ==> ./python-dtaidistance/.rb.notes <== SR 1068462 drop unreproducible unnecessary file ==> ./python-future/.rb.notes <== python3 toolchain ASLR ==> ./python-pyinstaller-hooks-contrib/.rb.notes <== python3.10 toolchain .pyc ASLR ==> ./python-quantities/.rb.notes <== date+time in /usr/lib/python3.9/site-packages/quantities/constants/_codata.py ==> ./python-sqlalchemy-migrate/.rb.notes <== => SR 1067129 python-sqlalchemy-migrate clean Sphinx leftover files ==> ./python-sunpy/.rb.notes <== = https://github.com/sunpy/sunpy/issues/6030 FTBFS-2023 found range good=1687994158 bad=1687997370 (2023-06-29T00:09:30) ==> ./qca/.rb.notes <== FTBFS-2031 found range good=1924935388 bad=1924937313 (2030-12-31T08:48:33) ==> ./qt6-declarative/.rb.notes <== /usr/lib64/libQt6QuickDialogs2QuickImpl.so.6.4.2 differs in ELF sections order variation in qtdeclarative-everywhere-src-6.4.2/build/build.ninja , ColorDialog_qml.cpp from qmlcachegen ==> ./qt6-quick3d/.rb.notes <== TODO /usr/lib64/libQt6Quick3DHelpers.so.6.4.2 differs in ELF sections ==> ./qt6-webengine/.rb.notes <== Qt toolchain issue ; hard to debug because of slow compile time ==> ./smuxi/.rb.notes <== = https://github.com/mono/mono/issues/20172 mono mdb ==> ./spotifyd/.rb.notes <== = https://github.com/Spotifyd/spotifyd/issues/1167 => https://github.com/librespot-org/librespot/commit/9de1f38e926de827cf2f05b4d0... random ==> ./tigervnc/.rb.notes <== = https://bugzilla.opensuse.org/show_bug.cgi?id=1208478 RSA key ==> ./transfig/.rb.notes <== = https://bugs.ghostscript.com/show_bug.cgi?id=696765 date+time from ghostscript toolchain /usr/share/doc/packages/transfig/fig2mpdf/sample-presentation.pdf differs at offset '13201' (PDF document, version 1.5) ==> ./velociraptor/.rb.notes <== #= https://bugzilla.opensuse.org/show_bug.cgi?id=1207369 report date filterdiff strings RPMS*/usr/bin/velociraptor ==> ./xemacs-packages/.rb.notes <== date+time in .elc date+time from psgml ==> ./zellij/.rb.notes <== reproducible again for unknown reason. was: /usr/bin/zellij differs in ELF sections from LTO? random?
On Mittwoch, 1. März 2023 15:46:59 CET Bernhard M. Wiedemann wrote:
==> ./transfig/.rb.notes <== = https://bugs.ghostscript.com/show_bug.cgi?id=696765 date+time from ghostscript toolchain /usr/share/doc/packages/transfig/fig2mpdf/sample-presentation.pdf differs at offset '13201' (PDF document, version 1.5)
PDFs created from transfig/fig2dev should be somewhat more reproducible now. "Somewhat", as the PDF itself is not fully reproducible (time based IDs), but all the parts which are copied into the final document by pdflatex are. Patch has just landed in Factory, and was also accepted by upstream. (There is a second patch which also fixes it for the fig2mpdf feature. As fig2mpdf is not part of upstream, this resides in the openSUSE package only.) This fixes R-B for transfig itself, and also for many packages which use fig2dev as part of the build process. Kind regards, Stefan
On 21/03/2023 19.38, Stefan Brüns wrote:
Patch has just landed in Factory, and was also accepted by upstream.
Thanks for getting that fix in. I linked it in the draft at https://reproducible-builds.org/reports/2023-03/ Ciao Bernhard M.
participants (2)
-
Bernhard M. Wiedemann
-
Stefan Brüns