CPT - look back at 2025

We did five releases of Copr that included these new features:

We did eight releases of Mock with these features:

  • Buildroot lockfile generator to enable hermetic builds.
  • New plugin unbreq to detect unused BuildRequires.
  • We passed `–suppress-sync=yes` to `systemd-nspawn` but then we disabled it as it caused problems to some packages. It is available as opt-in.
  • Automatic mapping to the appropriate user-static QEMU binary when building for new architectures (like RISC-V).
  • Numerous new configs have been added.

We wrote two articles for Fedora Magazine and six other blogposts.
Pavel developed norpm, which allows safe parsing of SPEC files.
We handled several issues in labs where Copr builders are: insufficient resources for s390x, problems with networking of ppc64le, moving x86_64 builders to a different datacenter, allocated high-performing builders for ppc64le…
Jakub added descriptions to resource pools so we can thank our donors.
Jakub wrote a nice page describing why you should use Mock.
Pavel upstreamed Tekton Pipelines for Konflux.
You built 1.4 million packages in Copr in 2025. And Copr serves about 44TB of DNF repositories.

Released SW

Besides Copr, we maintain several projects and this year we did these notable releases:

Log Detective

We successfully launched Log Detective as a prototype in January. Put it fully in production mode in May. And we have been improving it since then.
Jiri set up metrics to measure whether changes in LLM models are improving responses.

SPDX

  • Miroslav continued working on SPDX Change:
    • With the help of Richard Fontana, Miroslav added 75 new entries to fedora-license-data.
    • Miroslav added (with the help of Jilayne Lovejoy) 24 new licenses to SPDX license-list’s upstream repository.

Talks

What can you expect from us in 2026?

We finished the development phase of Pulp integration. The first month of 2026 will be focused on operation - migrating users’ data to the Pulp backend.

We want to put rpmeta in production and dynamically allocate high-performance builders when your build will take too much time on a normal builder.

If you want to affect what else we do in 2026, you can check our backlog and cast your vote (use an emoticon for comment #0) or add a comment on how much time the feature will save if implemented. And we will choose the card that will have the most significant impact.