Bug#1026877: opari2: please make the build reproducible
James Addison, le dim. 31 mars 2024 23:44:20 +0100, a ecrit: > Source: opari2 > Followup-For: Bug #1026877 > Control: close -1 2.0.7-2 Thanks!
Bug#1026877: opari2: please make the build reproducible
Source: opari2 Followup-For: Bug #1026877 Control: forwarded -1 https://salsa.debian.org/debian/opari2/-/commit/1cc9b96544cdf1e8cbc733584b3ff1a4109b89e6 https://salsa.debian.org/debian/opari2/-/commit/15155c97944b7c17bc481ff91261d8191f81ae6f
Bug#1026877: opari2: please make the build reproducible
Source: opari2 Followup-For: Bug #1026877 Control: close -1 2.0.7-2
Bug#1026877: Salsa reprotest CI not picky enough? (Was: Bug#1026877: opari2: please make the build reproducible)
Hello, Vagrant Cascadian, le ven. 23 déc. 2022 16:01:48 -0800, a ecrit: > The example for the salsa-ci suggests that build paths variations are > enabled by default under the "Adding extra arguments to reprotest": > > > https://salsa.debian.org/salsa-ci-team/pipeline/-/tree/master#adding-extra-arguments-to-reprotest > > variables: > SALSA_CI_REPROTEST_ARGS: --variations=-build_path > > You could try inverting it, and passing --variations=+build_path That seems to be working indeed: https://salsa.debian.org/debian/opari2/-/jobs/3702025 I let the reproducibility team discuss with CI about really having --variations=+all enabled by default :) Samuel
Bug#1026877: Salsa reprotest CI not picky enough? (Was: Bug#1026877: opari2: please make the build reproducible)
On 2022-12-24, Samuel Thibault wrote: > Chris Lamb, le ven. 23 déc. 2022 05:32:47 +, a ecrit: >> Whilst working on the Reproducible Builds effort [0] we noticed that >> opari2 could not be built reproducibly. >> >> Patch attached that exports CFLAGS from dpkg-buildflags(1), ensuring >> that -fdebug-prefix-map (and similar) to the underlying build system. > > I'm wondering how it is that the salsa reprotest CI didn't catch it: > > https://salsa.debian.org/debian/opari2/-/jobs/3675809 > > Is there some salsa-ci option that I should enable to make it more > picky? https://salsa.debian.org/debian/opari2/-/jobs/3696823#L1036 INFO:reprotest:build "experiment-1": FIX environment, FIX build_path, ... Something is telling reprotest to not vary build paths. I also see weird things suggesting the standard build path reprotest uses is duplicated: https://salsa.debian.org/debian/opari2/-/jobs/3696823#L473 make[1]: Entering directory '/tmp/reprotest.57Bwiu/const_build_path/const_build_path' Normally, there would just be a single "const_build_path" subdir. I noticed this issue on one of the packages I maintain, but haven't poked at it yet. This suggests to me the salsa-ci reprotest jobs are somehow manually adjusting the build paths in some way above-and-beyond how reprotest normally varies the build path? We probably should follow-up with the salsa-ci folks... The example for the salsa-ci suggests that build paths variations are enabled by default under the "Adding extra arguments to reprotest": https://salsa.debian.org/salsa-ci-team/pipeline/-/tree/master#adding-extra-arguments-to-reprotest variables: SALSA_CI_REPROTEST_ARGS: --variations=-build_path You could try inverting it, and passing --variations=+build_path live well, vagrant signature.asc Description: PGP signature
Bug#1026877: Salsa reprotest CI not picky enough? (Was: Bug#1026877: opari2: please make the build reproducible)
Hello, Chris Lamb, le ven. 23 déc. 2022 05:32:47 +, a ecrit: > Whilst working on the Reproducible Builds effort [0] we noticed that > opari2 could not be built reproducibly. > > Patch attached that exports CFLAGS from dpkg-buildflags(1), ensuring > that -fdebug-prefix-map (and similar) to the underlying build system. I'm wondering how it is that the salsa reprotest CI didn't catch it: https://salsa.debian.org/debian/opari2/-/jobs/3675809 Is there some salsa-ci option that I should enable to make it more picky? Samuel
Bug#1026877: opari2: please make the build reproducible
Hi Samuel, > I'm wondering, though, how it is that the salsa reprotest CI didn't > catch it: > > https://salsa.debian.org/debian/opari2/-/jobs/3675809 That's a good question and I don't actually know the answer. :( Indeed, I'm not actually that familiar with reprotest myself, so I'm likely not the best person to ask. Could you quickly try the reproducible-bui...@lists.alioth.debian.org mailing list? Regards, -- ,''`. : :' : Chris Lamb `. `'` la...@debian.org chris-lamb.co.uk `-
Bug#1026877: opari2: please make the build reproducible
Control: tags -1 + pending Hello, Chris Lamb, le ven. 23 déc. 2022 05:32:47 +, a ecrit: > Whilst working on the Reproducible Builds effort [0] we noticed that > opari2 could not be built reproducibly. > > Patch attached that exports CFLAGS from dpkg-buildflags(1), ensuring > that -fdebug-prefix-map (and similar) to the underlying build system. Thanks for the patch! I'm wondering, though, how it is that the salsa reprotest CI didn't catch it: https://salsa.debian.org/debian/opari2/-/jobs/3675809 Is there some salsa-ci option that I should enable to make it more picky? Samuel
Bug#1026877: opari2: please make the build reproducible
Source: opari2 Version: 2.0.7-1 Severity: wishlist Tags: patch User: reproducible-bui...@lists.alioth.debian.org Usertags: buildpath X-Debbugs-Cc: reproducible-b...@lists.alioth.debian.org Hi, Whilst working on the Reproducible Builds effort [0] we noticed that opari2 could not be built reproducibly. Patch attached that exports CFLAGS from dpkg-buildflags(1), ensuring that -fdebug-prefix-map (and similar) to the underlying build system. Patch attached. [0] https://reproducible-builds.org/ Regards, -- ,''`. : :' : Chris Lamb `. `'` la...@debian.org / chris-lamb.co.uk `- --- a/debian/rules 2022-12-23 05:22:27.671938667 + --- b/debian/rules 2022-12-23 05:30:34.856752502 + @@ -17,7 +17,7 @@ dh_autoreconf autoreconf -- -f -i . build-frontend override_dh_auto_configure: - ac_scorep_platform=linux dh_auto_configure -- --enable-shared + ac_scorep_platform=linux dh_auto_configure -- --enable-shared $(shell dpkg-buildflags --export=configure) override_dh_installdocs: dh_installdocs
