Re: December 2, 2025 Workstation Working Group Meeting

Hi, here is my Gemini-assisted summary for this meeting.

Summary

The Working Group discussed the decision to remove the existing Fedora 
Flatpak repository and replace it with a new one for Fedora 
Workstation. Jef Spaleta expressed strong concerns that this decision 
would lead to unsustainable fracturing, proposing the previously 
rejected Filter Fedora Flatpaks change proposal as a better 
alternative, a sentiment reinforced by discussions with Neal Gompa 
(Conan Kudo), Tomáš Popela, and Matthias Clasen. The group ultimately 
reached a consensus to abandon the plan for a separate Fedora 
Workstation Flatpak repository and adopt the filter proposal, 
contingent upon its approval by FESCo, while also emphasizing the 
long-term goal of leveraging Flathub and the need for better 
coordination regarding architecture support like RISC-V.

Details

Concern over Flatpak Repository Strategy

Jef Spaleta voiced strong concern over the decision to create a new 
Workstation-specific Flatpak repository, arguing that it would lead to 
further fracturing and create an unsustainable precedent for every 
Fedora spin to have its own remote. Michael Catanzaro acknowledged that 
this was not ideal but stated the current goal was to reduce the number 
of Fedora Flatpaks without having control over the existing Fedora 
Flatpak repository or strategy. The decision was made assuming the 
Filter Fedora Flatpaks change proposal was rejected.

Revisiting the Filter Proposal and History of the Flatpak SIG

Jef Spaleta mentioned they had discussed resubmitting the rejected 
filter change proposal to FESCo, suggesting that the Workstation 
Working Group's current plan was an attempt to circumvent the proper 
process. Neal Gompa (Conan Kudo) and Tomáš Popela explained that the 
Flatpak SIG became a one-person effort (Yaakov) due to initial 
complexity with modularity tools, discouragement from Flathub people, 
and key members leaving or having their assignments change. Jef Spaleta 
emphasized that a SIG should not become a one-person effort and such 
efforts should be adjourned if group collaboration cannot be sustained.

Challenges in Managing Flatpak SIG and Technical Solutions

Jef Spaleta asserted that having one-person SIGs is a structural 
problem that needs to be fixed, although Neal Gompa (Conan Kudo) 
pointed out that the majority of SIGs are currently one-person efforts. 
Michael Catanzaro and Jef Spaleta discussed the lack of developers as a 
problem, with Tomáš Popela offering a more optimistic view. Matthias 
Clasen noted that enforcing a minimum quorum might shut down the 
majority of Fedora SIGs, which may not be the best strategy. Jef 
Spaleta believes the filter proposal is the most reasonable approach to 
rescope the Flatpak effort and is aiming to transition Fedora Flatpaks 
to a model similar to EPEL to involve RPM maintainers for a more 
sustainable approach.

Viability of the Filtered Fedora Flatpak Proposal

Jef Spaleta sought confirmation on whether the filter proposal would 
meet the requirements of the Workstation Working Group. Neal Gompa 
(Conan Kudo) highlighted that while the filter proposal is probably 
viable, the main technical challenge is managing the filter to prevent 
applications on pre-existing systems from stopping receiving updates 
when an app is switched for a new release. Jef Spaleta agreed that 
technical requirements like preventing update issues must be addressed 
in the change proposal scope.

Consensus on Filter Proposal and Long-Term Goals

Michael Catanzaro proposed that if FESCO accepts the Filter Fedora 
Flatpaks change proposal, the Fedora Workstation will adopt it and drop 
the plan to create a new Workstation Flatpak repository. Matthias 
Clasen reiterated the long-term goal of moving away from a 
Fedora-specific Flatpak repository entirely, preferring Flathub when 
suitable. Michael Catanzaro clarified that a Fedora repo would still be 
needed for pre-installed apps.

Flathub Coordination and RISC-V Architecture

Jef Spaleta stressed the importance of improving coordination with 
Flathub, particularly concerning the bring-up of the RISC-V 
architecture. They warned that if Flathub is not ready with builders 
when Fedora Workstation starts producing outputs for RISC-V (expected 
around 2026), there will be immense pressure on Fedora to Flatpak 
everything. Michel Lind suggested that reliable metadata on whether a 
package is built from source and for all supported architectures would 
alleviate most of the concern regarding Flathub.

Final Decision and Scope

The Working Group reached a consensus to adopt the filtered Fedora 
Flatpak change proposal if FESCo approves it, and abandon the plan for 
a separate Fedora Workstation Flatpak repository. Neal Gompa (Conan 
Kudo) specifically requested that the filter's scope be tightly noted 
as specific to Workstation, suggesting that the filter might be more 
appropriate for Labs editions rather than general-purpose spins.

Open Floor and Future Flatpak Development

Tomáš Popela announced that a new team member started working on 
Flatpak infrastructure, who is expected to help with Fedora Flatpak 
infrastructure. Michael Catanzaro expressed a desire for continuous 
Flatpak updates tied to RPM builds. Neal Gompa (Conan Kudo) noted that 
they are close to this goal, needing only an infrastructure service to 
listen to RPM updates and kick off Flatpak builds.

Suggested next steps

 - Jef Spaleta will try to resubmit the filter proposal to FESCO.


-- 
_______________________________________________
desktop mailing list -- [email protected]
To unsubscribe send an email to [email protected]
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/[email protected]
Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue