- Érico Nogueira Rolim
- Gustavo de Souza dos Reis
- Henrique Simões
- Gustavo Montenari Pechta
- Jakub Wlodek
- Mark Rivers
- Tejas Guruswamy
- Freddie Akeroyd
- Famous Alele
- Retrospective: What are your thoughts on how the meetings have been going?
- Format, usefulness, efficiency
- Mark: have a presentation from someone every meeting (new detectors, developments)
- How can we improve engagement of developers and maintainers? How can we encourage folks to attend meetings?
- Meeting agenda hasn't been posted with 1 week advance, that should be fixed
- What's missing for more people to interact in general issues?
- We should review the maintainer list. What to do for those who still haven't responded?
- Mark: add recent contributors as maintainers for some of them, such as Kaz do ADMythen
- Xiaoqiang: remove from ADMythen
- Do we have long-term roadmaps? Maybe revisit milestones?
- Mark: there are wishlist items (e.g. refactors to use smart pointers), but they will take time, skill, and require thought towards updating drivers/plugins as well
- Jure: does the single NDArrayPool cause issues with attribute list reuse? Would be interesting to benchmark effect of allocating for every frame
- Jure (February Codeathon): destructible areaDetector drivers
- Implement support in ADCore
- Implement support in ADAravis
- Mark: there are wishlist items (e.g. refactors to use smart pointers), but they will take time, skill, and require thought towards updating drivers/plugins as well
- Items below discussed this meeting
- Documentation gaps (issue tracking effort):
- Driver documentation
- Design decision documentation
- Mark: this kind of documentation should go in the Architecture section in User Guide
- Development and code standards:
- It would allow us to have a document to point out to contributors
- Mark: document further driver guidelines, improve base class in order to simplify it
- Is there interest in this?
- Suggestions:
- Using commit messages to document development choices
- Avoid inclusion of commented out code
- Prefer modern C++ practices (?)
- Freddie: what C++ standard can be used?
- Mark and Jakub: older machines with limitations for C++ standard, standard compiler support
- Mark: some things are already C++11
- Mark: Windows support is important and care should be taken (e.g. PVXS); dbd files in O.Common can cause build issues when using the same tree
- It would allow us to have a document to point out to contributors
- Rules/procedures for taking care of the org:
- Technical disagreements, how to solve?
- Each maintainer has the last word for their drivers; or
- External consensus can override that, and certain standards/designs have to be followed
- Situations where merging without maintainer approval is fine
- Mark: things have been going well, might not be necessary to spend time on such rules. Maybe clean up current permissions and have a procedure for new projects?
- Technical disagreements, how to solve?
- areaDetector:
- Issues:
- Building on macOS 26.2: bundled zlib fails to compile
- Check external zlib build support (naming issue on MacOS? libz vs libzlib)
- Might be okay to support MacOS if it's simple
- Supported in CI https://docs.github.com/pt/actions/reference/runners/github-hosted-runners#standard-github-hosted-runners-for-public-repositories
- Building on macOS 26.2: bundled zlib fails to compile
- Issues:
- ADCore:
- Issues:
- Strange INP field to record named "test"
- Useful to have example, could be comment, or try and find innocuous internal PV
- Strange INP field to record named "test"
- Issues:
- ADAndor3:
- Issues:
- ADEiger:
- Issues:
- Support for reading and using timestamps from Stream2
- Pending further testing, Tejas and Jakub
- Eiger2 MEDM Counting Mode is flipped
- Waiting for feedback from firmware update
- The lz4 data in Stream2 is different from Stream
- Jakub tested with newer LZ4 version (RHEL9) and the issue still happens
- Support for reading and using timestamps from Stream2
- Issues:
- ADMythen:
- Issues:
- PRs:
- ADPICam:
- ADRIXSCam:
- PRs:
- Fix grab_setup being called with every acquisition
- Gustavo didn't have permission to merge it, though it's ready
- Fix grab_setup being called with every acquisition
- PRs:
- PRs:
- Stale Items:
- Stale Issues:
- Stale PRs:
- github actions build failing on windows-x64
- Fixing incompatible pointer assignment that causes build on RHEL 10 to fail
- Camera on Windows, ImageJ PVA plugin on same machine, gets really slow (1 FPS), no issue with CA or different machines
- Stale Items:
- Find a way to stop misbehaving plugins from slowing down driver
- 2025-05-08: close after documenting expected plugin behavior
- documentation
- 2025-08-07: Jakub is setting one up, can look at docs
- No time to take a look at it yet
- Find a way to stop misbehaving plugins from slowing down driver