From lamby at debian.org Tue Sep 24 19:35:34 2024 From: lamby at debian.org (Chris Lamb) Date: Tue, 24 Sep 2024 11:35:34 -0700 Subject: Bug#1082702: magic-wormhole-transit-relay: please make the build reproducible Message-ID: <172720284515.3589.12933783724374213386@a588bbe472a7> Source: magic-wormhole-transit-relay Version: 0.3.1-0.1 Severity: wishlist Tags: patch User: reproducible-builds at lists.alioth.debian.org Usertags: randomness X-Debbugs-Cc: reproducible-bugs at lists.alioth.debian.org Hi, Whilst working on the Reproducible Builds effort [0], we noticed that magic-wormhole-transit-relay could not be built reproducibly. This is because the testsuite leaves around a bunch of nondeterministic files which then get installed into the binary package: /usr/lib/python3/dist-packages/wormhole_transit_relay.test.test/Create/test_create/sia6i7v6/temp/created.db A patch is attached that removes this "wormhole_transit_relay.test.test" directory after the tests have completed. [0] https://reproducible-builds.org/ Regards, -- ,''`. : :' : Chris Lamb `. `'` lamby at debian.org / chris-lamb.co.uk `- -------------- next part -------------- --- a/debian/rules 2024-09-24 11:28:51.007696346 -0700 --- b/debian/rules 2024-09-24 11:32:31.120093198 -0700 @@ -2,6 +2,7 @@ export PYBUILD_NAME=magic-wormhole export CRYPTOGRAPHY_OPENSSL_NO_LEGACY=1 +export PYBUILD_AFTER_TEST=rm -rfv {build_dir}/wormhole_transit_relay.test.test %: dh $@ --buildsystem=pybuild From lamby at debian.org Tue Sep 24 19:57:33 2024 From: lamby at debian.org (Chris Lamb) Date: Tue, 24 Sep 2024 11:57:33 -0700 Subject: Bug#1082705: python-inline-snapshot: documentation includes inline error messages instead of output Message-ID: <172720403497.4436.12800163664794374127@0c730f6c9a62> Source: python-inline-snapshot Version: 0.13.1-1 Severity: wishlist User: reproducible-builds at lists.alioth.debian.org Usertags: buildpath X-Debbugs-Cc: reproducible-bugs at lists.alioth.debian.org Hi, Whilst working on the Reproducible Builds effort [0], we noticed that python-inline-snapshot could not be built reproducibly. This is because the documentation embeds a number of error messages that in turn include nondeterminstic contents: > pytest test_something.py --inline-snapshot=create,report ERROR: usage: pytest [options] [file_or_dir] [file_or_dir] [...] pytest: error: unrecognized arguments: --inline-snapshot=create,report inifile: None rootdir: /tmp/tmp.0ENzKZTyyS  There are also other errors that don't include nondeterminstic content, hence this more general bug: > pytest test_something.py --inline-snapshot=short-report environment: line 10: python: command not found It appears that the documentation cannot find the pytest plugin within its own build directory, which may be because the setuptools entry point isn't available [1] as it isn't installed... but I don't know how to fix that. [0] https://reproducible-builds.org/ [1] https://docs.pytest.org/en/7.1.x/how-to/writing_plugins.html#setuptools-entry-points Regards, -- ,''`. : :' : Chris Lamb `. `'` lamby at debian.org / chris-lamb.co.uk `- From lamby at debian.org Tue Sep 24 19:59:00 2024 From: lamby at debian.org (Chris Lamb) Date: Tue, 24 Sep 2024 11:59:00 -0700 Subject: Bug#1082706: python-sphobjinv: please make the build reproducible Message-ID: <172720427625.4007.4487842241498286269@03027f5d9cd5> Source: python-sphobjinv Version: 2.3.1.1-2 Severity: wishlist Tags: patch User: reproducible-builds at lists.alioth.debian.org Usertags: randomness X-Debbugs-Cc: reproducible-bugs at lists.alioth.debian.org Hi, Whilst working on the Reproducible Builds effort [0], we noticed that python-sphobjinv could not be built reproducibly. This is because every documentation page embedded an arbitrary memory reference: ? ? ? ??? ./usr/share/doc/python-sphobjinv-doc/html/api/data.html ? ? ? ? @@ -497,15 +497,15 @@ ? ? ? ? ? ? ? ? ? ? ? ? Built with Sphinx using a ? ? ? ? theme ? ? ? ? provided by Read the Docs. ? ? ? ? ? ? ? ? ? ? ? ? - ? ? ? ? + ? ? ? ? ? ? ? ?

... which was in turn caused by a faulty call to super(). Patch attached. [0] https://reproducible-builds.org/ Regards, -- ,''`. : :' : Chris Lamb `. `'` lamby at debian.org / chris-lamb.co.uk `- -------------- next part -------------- --- a/debian/patches/docs-Call-super-correctly.patch 1969-12-31 16:00:00.000000000 -0800 --- b/debian/patches/docs-Call-super-correctly.patch 2024-09-24 11:48:45.007670165 -0700 @@ -0,0 +1,15 @@ +Description: Make the build reproducible +Author: Chris Lamb +Last-Update: 2024-09-24 + +--- python-sphobjinv-2.3.1.1.orig/doc/source/_templates/footer.html ++++ python-sphobjinv-2.3.1.1/doc/source/_templates/footer.html +@@ -2,7 +2,7 @@ + + {%-block extrafooter %} + +-{{ super }} ++{{ super() }} + +

+ --- a/debian/patches/series 2024-09-24 11:29:12.923722801 -0700 --- b/debian/patches/series 2024-09-24 11:48:43.971665640 -0700 @@ -2,3 +2,4 @@ tests-Replace-python-with-python3-in-test_cli.py.patch docs-Use-packaged-intersphinx-resources.patch docs-Use-a-local-referenced-PNG-file.patch +docs-Call-super-correctly.patch From lamby at debian.org Tue Sep 24 20:13:45 2024 From: lamby at debian.org (Chris Lamb) Date: Tue, 24 Sep 2024 12:13:45 -0700 Subject: Bug#1082707: lomiri-content-hub: please make the build reproducible Message-ID: <172720516158.414158.2792906170627625829@copycat> Source: lomiri-content-hub Version: 2.0.0-2 Severity: wishlist Tags: patch User: reproducible-builds at lists.alioth.debian.org Usertags: randomness X-Debbugs-Cc: reproducible-bugs at lists.alioth.debian.org Hi, Whilst working on the Reproducible Builds effort [0], we noticed that lomiri-content-hub could not be built reproducibly. This is because it ships a results.txt report generated by rdfind: ? ? ??? ./usr/share/doc/lomiri-content-hub/cpp/html/results.txt ? ? ? ? @@ -1,25 +1,25 @@ ? ? ? ? # Automatically generated ? ? ? ? # duptype id depth size device inode priority name ? ? ? ? -DUPTYPE_FIRST_OCCURRENCE 188 1 179 65057 5641205 3 ./search/all_12.js ? ? ? ? -DUPTYPE_WITHIN_SAME_TREE -188 1 179 65057 5640133 3 ./search/functions_12.js ? ? ? ? -DUPTYPE_FIRST_OCCURRENCE 196 1 327 65057 5640161 3 ./search/all_8.js ? ? ? ? -DUPTYPE_WITHIN_SAME_TREE -196 1 327 65057 5640927 3 ./search/functions_8.js ? ? ? ? -DUPTYPE_FIRST_OCCURRENCE 202 1 538 65057 5640154 3 ./search/all_e.js ? ? ? ? -DUPTYPE_WITHIN_SAME_TREE -202 1 538 65057 5641199 3 ./search/functions_e.js ? ? ? ? -DUPTYPE_FIRST_OCCURRENCE 189 1 1236 65057 5640139 3 ./search/all_13.js A patch is attached that suppresses the creation of these reports. [0] https://reproducible-builds.org/ Regards, -- ,''`. : :' : Chris Lamb `. `'` lamby at debian.org / chris-lamb.co.uk `- -------------- next part -------------- diff --git a/debian/rules b/debian/rules index 56ce3cf..ece0504 100755 --- a/debian/rules +++ b/debian/rules @@ -35,9 +35,9 @@ endif override_dh_install: rmdir debian/tmp/usr/share/doc/lomiri-content-hub/qml/html/images/ - cd debian/tmp/usr/share/doc/lomiri-content-hub/cpp/html/ && rdfind -makesymlinks true . + cd debian/tmp/usr/share/doc/lomiri-content-hub/cpp/html/ && rdfind -makesymlinks true -makeresultsfile false . cd debian/tmp/usr/share/doc/lomiri-content-hub/cpp/html/ && symlinks -rc . - cd debian/tmp/usr/share/doc/lomiri-content-hub/qml/html/ && rdfind -makesymlinks true . + cd debian/tmp/usr/share/doc/lomiri-content-hub/qml/html/ && rdfind -makesymlinks true -makeresultsfile false . cd debian/tmp/usr/share/doc/lomiri-content-hub/qml/html/ && symlinks -rc . dh_install From lamby at debian.org Thu Sep 26 14:59:28 2024 From: lamby at debian.org (Chris Lamb) Date: Thu, 26 Sep 2024 06:59:28 -0700 Subject: Bug#1082796: python-mt-940: please make the build reproducible Message-ID: <172735899961.3754.11000768625620136421@3a28f4ebbd78> Source: python-mt-940 Version: 4.30.0-2 Severity: wishlist Tags: patch User: reproducible-builds at lists.alioth.debian.org Usertags: timestamps X-Debbugs-Cc: reproducible-bugs at lists.alioth.debian.org Hi, Whilst working on the Reproducible Builds effort [0], we noticed that python-mt-940 could not be built reproducibly. This is because it ships test coverage in HTML format in the binary package that contains timestamps. A patch is attached that simply removes this coverage entirely after running the tests. However, the coverage was being installed into the incorrect directory /usr/lib/python3/dist-packages/htmlcov, so if you would prefer to keep the test coverage, you would of course need to install it into a more appropriate location also. [0] https://reproducible-builds.org/ Regards, -- ,''`. : :' : Chris Lamb `. `'` lamby at debian.org / chris-lamb.co.uk `- -------------- next part -------------- --- a/debian/rules 2024-09-26 06:35:01.043556873 -0700 --- b/debian/rules 2024-09-26 06:56:18.013446724 -0700 @@ -2,6 +2,7 @@ export PYBUILD_NAME=mt-940 export PYBUILD_TEST_ARGS={dir}/mt940_tests/ +export PYBUILD_AFTER_TEST=rm -rfv {build_dir}/htmlcov # try to prevent unsanctioned downloads export http_proxy=http://127.0.0.1:9/ From lamby at debian.org Thu Sep 26 17:15:58 2024 From: lamby at debian.org (Chris Lamb) Date: Thu, 26 Sep 2024 09:15:58 -0700 Subject: Bug#1082806: tree-puzzle: please make the build reproducible Message-ID: <172736726847.14372.5468282786227902716@4293a02465dc> Source: tree-puzzle Version: 5.3~rc16+dfsg-10 Severity: wishlist Tags: patch User: reproducible-builds at lists.alioth.debian.org Usertags: randomness X-Debbugs-Cc: reproducible-bugs at lists.alioth.debian.org Hi, Whilst working on the Reproducible Builds effort [0], we noticed that tree-puzzle could not be built reproducibly (again). This is something of a followup from #1005901. Here, however, it looks like the .puzzle files at fault. Here is the diff on amd64: ? ? ? ??? ./usr/share/doc/tree-puzzle/examples/tests/qp-hky-rhet-clock-nucl.nucl.puzzle ? ? ? ? @@ -320,18 +320,18 @@ ? ? ? ? : ? ? ? ? :-----1 Thylacinus ? ? ? ? ? ? ? ? ? ? ? ? branch length S.E. branch length S.E. ? ? ? ? Thylacinus 1 0.13266 0.07077 9 0.14146 0.08889 ? ? ? ? Sarcophilu 2 0.16398 0.05399 10 0.15536 0.07510 ? ? ? ? -Dasyurus 3 0.00001 0.00399 11 0.07879 0.07178 ? ? ? ? +Dasyurus 3 0.00001 0.00141 11 0.07879 0.07178 ? ? ? ? Echymipera 4 0.20360 0.08378 ? ? ? ? Trichosuru 5 0.15345 0.07805 ? ? ? ? -Phalanger 6 0.18858 0.08737 ? ? ? ? +Phalanger 6 0.18858 0.08736 A patch is attached that cleans the tests directory entirely by using tree-puzzle's own clean target. Seems much more sensible than playing whack-a-mole. [0] https://reproducible-builds.org/ Regards, -- ,''`. : :' : Chris Lamb `. `'` lamby at debian.org / chris-lamb.co.uk `- -------------- next part -------------- --- a/debian/rules 2024-09-26 09:08:18.502165005 -0700 --- b/debian/rules 2024-09-26 09:14:00.523766494 -0700 @@ -31,13 +31,9 @@ mv `pwd`/debian/tmp/usr/bin/puzzle `pwd`/debian/$(pkg)/usr/bin/$(pkg) mv `pwd`/debian/tmp/usr/bin/ppuzzle `pwd`/debian/$(ppkg)/usr/bin/$(ppkg) -override_dh_installexamples: - rm -f tests/*.log tests/*.trs -ifneq (,$(filter $(DEB_BUILD_ARCH),i386)) - # Does not generate deterministic output on i386. - rm -f tests/qp-tn-nucl.nucl* -endif - dh_installexamples +execute_before_dh_installexamples: + # Don't install output generated by tests, some of which is nodeterminstic + $(MAKE) -C tests clean override_dh_fixperms: dh_fixperms