On Thu, 2025-02-20 at 18:34 +0000, Richard Hughes via devel wrote: > Hi all, > > My CI has been failing for the last 24h with: > > [ 4/48] Installing fwupd-0:2.0.7-0.1alpha.fc41.x86_64 100% | 39.9 > MiB > terminate called after throwing an instance of 'std::length_error' > what(): basic_string::_M_replace_aux > contrib/ci/fedora-test.sh: line 3: 7 Aborted > (core dumped) dnf install -y dist/*.rpm > > Has anyone else seen this? Frustratingly I can't reproduce locally. We have been seeing it sporadically in the Anaconda CI in the last few weeks, in just a single specific test workflow (that validates Anaconda RPMs have been built correctly) - I failed run looks like this: https://github.com/rhinstaller/anaconda/actions/runs/13498853227/job/37712147166?pr=6202 [ 1/323] Verify package files 0% | 0.0 B [ 1/323] Verify package files 21% | 69.0 B [ 1/323] Verify package files 34% | 112.0 B [ 1/323] Verify package files 42% | 137.0 B [ 1/323] Verify package files 57% | 183.0 B [ 1/323] Verify package files 73% | 237.0 B [ 1/323] Verify package files 94% | 303.0 B [ 1/323] Verify package files 100% | 321.0 B [ 1/323] Verify package files 100% | 321.0 B [ 2/323] Prepare transaction 63% | 203.0 B [ 2/323] Prepare transaction 100% | 321.0 B [ 2/323] Prepare transaction 100% | 321.0 B [ 3/323] Installing libblockdev-utils-0:3.3.0- 99.20250213132035617502.master.0.gf5ed6a16.fc42.x86_64 0% | 0.0 B [ 3/323] Installing libblockdev-utils-0:3.3.0- 99.20250213132035617502.master.0.gf5ed6a16.fc42.x86_64 100% | 44.5 KiB [ 3/323] Installing libblockdev-utils-0:3.3.0- 99.20250213132035617502.master.0.gf5ed6a16.fc42.x86_64 100% | 44.5 KiB [ 4/323] Installing python3-gobject-base-0:3.50.0-3.fc42.x86_64 100% | 1.5 MiB [ 4/323] Installing python3-gobject-base-0:3.50.0-3.fc42.x86_64 100% | 1.5 MiB [ 5/323] Installing libnl3-0:3.11.0-3.fc42.x86_64 100% | 1.0 MiB [ 5/323] Installing libnl3-0:3.11.0-3.fc42.x86_64 100% | 1.0 MiB [ 6/323] Installing libmnl-0:1.0.5-7.fc42.x86_64 100% | 56.4 KiB [ 6/323] Installing libmnl-0:1.0.5-7.fc42.x86_64 100% | 56.4 KiB [ 7/323] Installing satyr-0:0.43-5.fc42.x86_64 100% | 345.6 KiBterminate called after throwing an instance of 'std::length_error' what(): basic_string::_M_replace_aux make: *** [Makefile.am:329: container-rpm-test] Error 134 Its definitely a race condition, as it sometimes happens, and sometimes not. We have also made a few other observations: * we have never seen this happen outside of the GitHub CI infra -> eq. doing 100-200 locally running RPM test runs did not trigger this * is most often happens with the satyr package, but rarely it crashed with a different package -> likely satyr is just the first package that does something that has a chance to trigger this * the chance of this occurring seems to fluctuate over time - some days its not reproducible at all, some days it happens all the time -> maybe its related to load in the GitHub infra or some other side effects ? Unfortunately the environment this is running in is rather barebones and under many layers of infrastructure, so we were not able to dig up some more debugging data or a crash dump so far from when it happens. :P > > I'm *guessing* it's a regression in dnf5-5.2.10.0-2.fc41 -- i.e. > https://bodhi.fedoraproject.org/updates/FEDORA-2025-665751b0e4 > > Ideas welcome, thanks, > > Richard. > -- _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
