TLDR: Focused on OpenSSL 4.0 migration, learning more and investigated an interesting regression.
What a week it has been doing my first +1 maintenance shift! The week has mostly been smooth sailing going through various OpenSSL4.0 migration tasks with some deviations. As I started my shift on Monday, I first messaged others to get some advice for my +1 maintenance shift and looked through the previous reports to get inspiration.
The first half of the week was spent on opening merge-proposals for various FTBFS bugs and the second investigating an interesting regression in pybind11. More information on this regression below (fix already released).
Lastly, I want to take the chance to thank those who helped me throughout the week. @ruancomelli and @vmpyr for all their helpful links and information for a first-time +1 maintenance shift. @ravi-sharma for his guidance and advice throughout my OpenSSL transitions. @ahasenack and @ginggs for all their help with sponsorships and investigating the pybind11 regression.
Work-Needed Items
- The work items recommended by people in previous +1 maintenance shifts.
- There are still a few packages to be fixed for the OpenSSL 4.0 migration. Read the https://lists.ubuntu.com/archives/ubuntu-devel/2026-June/043622.html for more info.
- More research into whether the
pybind11package for Resolute should be updated with the regression fix. So far no packages have been found to be affected but any new builds against the version in Resolute may introduce the regression.
Sponsorship Needed
- Fix for OpenSSL 4.0 compatibility for the “amanda” package
- Rebuild request for the “ecflow” package to pass testing
Full Logs
kristall
This package was failing to build from source when built against OpenSSL 4.0 but is now fixed and released:
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/kristall/+bug/2154981
Merge Proposal: https://code.launchpad.net/~rroessler/ubuntu/+source/kristall/+git/kristall/+merge/509583
Debian Bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1138433
Salsa MR: https://salsa.debian.org/debian/kristall/-/merge_requests/2
Upstream Bug: https://github.com/ikskuh/kristall/issues/308
Upstream PR: https://github.com/ikskuh/kristall/pull/309
xmrig
This package was failing to build from source when built against OpenSSL 4.0 but is now fixed and released:
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/xmrig/+bug/2154929
Merge Proposal: https://code.launchpad.net/~rroessler/ubuntu/+source/xmrig/+git/xmrig/+merge/509641
Debian Bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1138381
Salsa MR: https://salsa.debian.org/cryptocoin-team/xmrig/-/merge_requests/1
Upstream Bug: https://github.com/xmrig/xmrig/issues/3833
Upstream PR: https://github.com/xmrig/xmrig/pull/3808
monero
This package was failing to build from source when built against OpenSSL 4.0 but is now fixed and released:
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/monero/+bug/2154924
Merge Proposal: https://code.launchpad.net/~rroessler/ubuntu/+source/monero/+git/monero/+merge/509642
Debian Bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1138376
Salsa MR: Merge requests are currently blocked from being requested
Upstream Bug: https://github.com/monero-project/monero/issues/10904
Upstream PR: https://github.com/monero-project/monero/pull/10908
feather-wallet
This package was failing to build from source when built against OpenSSL 4.0 but is now fixed and released:
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/feather-wallet/+bug/2154886
Merge Proposal: https://code.launchpad.net/~rroessler/ubuntu/+source/feather-wallet/+git/feather-wallet/+merge/509643
Debian Bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1138339
Salsa MR: https://salsa.debian.org/cryptocoin-team/feather-wallet/-/merge_requests/1
Upstream Bug: https://github.com/feather-wallet/feather/issues/300
Upstream PR: Issues in submitting as seen from the above bug-report
amanda
This package was failing to build from source when built against OpenSSL 4.0 but is now fixed and waiting sponsorship:
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/amanda/+bug/2163689
Merge Proposal: https://code.launchpad.net/~rroessler/ubuntu/+source/amanda/+git/amanda/+merge/509813
Debian Bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1144897
Salsa MR: https://salsa.debian.org/debian/amanda/-/merge_requests/4
Upstream Bug: Unnecessary as noted in PR below
Upstream PR: https://github.com/zmanda/amanda/pull/292
ecflow
This package was failing to build from source when built against OpenSSL 4.0 but is now fixed and committed:
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/ecflow/+bug/2154891
Merge Proposal: https://code.launchpad.net/~rroessler/ubuntu/+source/ecflow/+git/ecflow/+merge/509587
Debian Bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1138344
Salsa MR: https://salsa.debian.org/science-team/ecflow/-/merge_requests/1
Upstream Bug: https://github.com/ecmwf/ecflow/issues/411
Upstream PR: https://github.com/ecmwf/ecflow/pull/412
After the above merge-proposal was accepted, there was a regression in the autopkgtest results. Whenever the ecflow module was imported, the process would exit with a free() invalid pointer crash. After following the backtrace output from gdb, I found that there was a regression in pybind11-3.0.1-3 where on deallocation of specifically constructed objects, a free() would be called on string literals causing the crash. Since the fix to pybind11 has been committed, I have since added an additional request to rebuild ecflow against the patched version.
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/ecflow/+bug/2163679
Merge Proposal: code.launchpad.net/~rroessler/ubuntu/+source/ecflow/+git/ecflow/+merge/509910
pybind11
As per the regression first noticed in ecflow testing, I followed the trail to a recent upstream PR in which the regression was introduced (see: https://github.com/pybind/pybind11/pull/5486). This PR introduced a latent bug where a free() is called on string-literals under certain conditions (showing in the original build for ecflow). Upstream has since patched this and I also prepared a fix for Stonking which has been committed. I also spent the later part of the week confirming that no other packages have been affected (of which there are none in Stonking and Resolute). It should be noted that users that are compiling against the version in Resolute will still have the regression.
Launchpad Bug: https://bugs.launchpad.net/ubuntu/+source/pybind11/+bug/2164501
Merge Proposal: https://code.launchpad.net/~rroessler/ubuntu/+source/pybind11/+git/pybind11/+merge/509742
Debian Bug: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1144807
Salsa MR: Waiting for upstream release (since next will be unaffected)
Upstream Bugs:
Upstream PRs: