[opensuse-factory] openSUSE reproducible builds status 2020-06
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Hi, last month's status: https://lists.opensuse.org/opensuse-factory/2020-06/msg00003.html Last months' reproducible builds project updates (including my work): https://reproducible-builds.org/reports/2020-05/ This month, several C++ packages became unreproducible from a gcc10 issue https://bugzilla.opensuse.org/show_bug.cgi?id=1172846 I uploaded https://rb.zq1.de/compare.factory-20200630/ today and rbstats are: total-packages: 13000 (+154) build-tried: 12992 (+155) build-failed: 26 (+13) build-n-a: 120 (-28) build-succeeded: 12846 (+170) build-official-failed+na: 285 (-355) build-compare-failed: 423 (+0) build-compare-succeeded: 12423 (+170) verify-failed: 531 (-4) verified-semi-reproducible: 12096 (+727) verified-bit-identical: 0 (+0) bit-by-bit-identical: 12219 (+133) not-bit-by-bit-identical: 628 (+37) not-bit-by-bit-identicalcheck: 627 (+37) https://rb.zq1.de/compare.factory-20200630/graph.png shows the change over time https://rb.zq1.de/compare.factory-20200630/unreproduciblerings.txt lists very unreproducible core packages (bootstrap+DVD) Of the badly unreproducible packages, 1 were in ring0 40 were in ring1 That makes it 41/2901 => 1.41 % which is below the overall average of 423/12846 => 3.29 % 628/12846 => 4.89 % of packages are not perfectly reproducible In Ring0, grep can become more reproducible with https://build.opensuse.org/request/show/813322 In Ring1, we have a newly unreproducible package apache-pdfbox from diffs in the javadoc documentation index http://rb.zq1.de/compare.factory-20200630/apache-pdfbox-compare.out perf package is still unreproducible from random filesystem readdir order via nftw calls in https://github.com/torvalds/linux/blob/54ecb8f7028c5eb3d740bb82b0f1d90f2 df63c5c/tools/perf/pmu-events/jevents.c#L1131 Ciao Bernhard M. -----BEGIN PGP SIGNATURE----- iHUEARYIAB0WIQTykslvYmKwlIQesLNdovN53d8CLgUCXvr7kAAKCRBdovN53d8C Lo53AP9Zm0oJVsd1g+CC7zqWQWeofLVhYF6kZuQ7rfEjDp51twD9FNse83YWkUCl E3xTMLDYpqGyisvjsYWo35ojhTQYvgY= =EwU9 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi, last month's status: https://lists.opensuse.org/opensuse-factory/2020-06/msg00348.html Last months' reproducible builds project updates (including my work): https://reproducible-builds.org/reports/2020-06/ I uploaded https://rb.zq1.de/compare.factory-20200730/ today and rbstats are: total-packages: 13133 (+133) build-tried: 13126 (+134) build-failed: 38 (+12) build-n-a: 126 (+6) build-succeeded: 12962 (+116) build-official-failed+na: 251 (-34) build-compare-failed: 434 (+11) build-compare-succeeded: 12528 (+105) verify-failed: 491 (-40) verified-semi-reproducible: 12104 (+8) verified-bit-identical: 0 (+0) bit-by-bit-identical: 12382 (+163) not-bit-by-bit-identical: 581 (-47) https://rb.zq1.de/compare.factory-20200730/graph.png shows the change over time https://rb.zq1.de/compare.factory-20200730/unreproduciblerings.txt lists very unreproducible core packages (bootstrap+DVD) Of the badly unreproducible packages, 0 were in ring0 37 were in ring1 That makes it 37/2901 => 1.28 % which is below the overall average of 434/12962 => 3.35 % 581/12962 => 4.48 % of packages are not perfectly reproducible notable unreproducible core packages: accounts-qml-module ordering issues in /usr/share/accounts-qml-module/doc/html/onlineaccounts-qml-api.index from /usr/lib64/qt5/bin/qdoc fpc contains mtime of a patched file. Will become reproducible again when the patch reaches the tarball. gcc10 from PGO (Profile Guided Optimizations) go1.13 go1.14 unreproducible debuginfo and .note.go.buildid golang-github-prometheus-prometheus user host date jsch-agent-proxy package-search-index.zip files are randomly symlinks from fdupes -s kopete https://bugreports.qt.io/browse/QTBUG-83186 ASLR in uic needs more love by devs libhugetlbfs https://bugzilla.opensuse.org/show_bug.cgi?id=1159558 python38 same problem as python3 before with nondeterministic .pyc files Ciao Bernhard M.
participants (1)
-
Bernhard M. Wiedemann