Location and Time
Co-located with the Yocto Project Summit held on May 25-26, 2021.
The Developers Meeting is scheduled for May 25th between 15:30 and 20:00 UTC. The exact times for each individual topic are found below in the schedule.
As always, we will collect topics on the wiki at https://www.openembedded.org/OEDVM_2021.
For the actual developer meeting, there will be pre-assigned timeslots for each topic. The moderator(s) have the option of opening with a short introduction/presentation to introduce the topic.
- BSPs: best practice exemplars, cross-project issue tracker, linters, incentive loop design
- Moderator(s): Rich Persaud, Philip Balister
- Insight into the life of a Maintainer
- Moderator(s): Armin Kuster
- The good, bad and ugly of Maintaining Poky, meta-openembedded, meta-security and a BSP.
- Moderator(s): Armin Kuster
- X11 is dead; long live X11! what's to become of core-image-sato?
- Moderator(s): Trevor Woerner, Alexander Kanavin, Joshua Watt, Ross Burton
- the Yocto Project provides a sample distribution (poky) and images (core-image-minimal, core-image-base, core-image-full-cmdline…) to give users examples to follow and provide a basis for testing purposes
- core-image-sato was created to fill the GUI niche as an example and for testing
- core-image-sato is based on gtk+ 3.x and x11
- both gtk+ 3 and x11 are EOL/unmaintained
- do we need a GUI image going forward (as an example, for testing purposes)?
- how much testing does core-image-sato receive?
- how many teams have based their work on core-image-sato?
- if a GUI image is still needed, upon which toolkit and compositor should it be based?
- what's to become of core-image-sato?
- what's to become of x11 support in oecore?
- SBOM (Software Bill of Materials)
- Moderator(s): Trevor Woerner, Armin Kuster, Mikko Murto (meta-doubleopen)
- the requirement to provide a software bill of materials when delivering software to a customer/user is becoming more and more common
- e.g. a recent Executive Order in the United States requires an SBOM for security reasons
- as a project that creates images from sources, YP/OE is perfectly positioned to generate SBOMs for its artifacts
- we already generate similar information for software licence compliance via SPDX
- meta-doubleopen seems to be moving in this direction
- what information is required for an SBOM, what are the requirements to create a legally compliant SBOM?
- SPDX seems to be the best format for us to use, any objections?
- in our builds when/where do we generate the SBOM? do_package/do_packagedata? archiver/do_populate_lic?
- we already generate various manifests (e.g. buildhistory) should we replace this information with proper SBOMs?
- LTS (Long Term Support)
- Moderator(s): Trevor Woerner, Armin Kuster, Khem Raj
- for the first time ever, the Yocto Project experimented with having an LTS as well as its regular releases
- did anyone notice?
- did anyone use it?
- what did people like about it?
- what could be changed?
- should we do it again?
- what repercussions are there for the larger YP/OE community (layer maintainers)?
- is 2 years too much? not enough? just right?
- Improving Layer quality: Layerindex combined with a layerchecker
- Moderator(s): Jan-Simon Möller (email@example.com)
- Layers need to interop well. There is work to do. Let's chop some wood!
- Overview on the current state
- Your own pain points ?
- The good and the bad examples ?!
- How can we improve collectively ?
- How can we support that process ?
- Project Documentation
- Moderator(s): Michael Opdenacker, Nicolas Dechesne
- the Yocto Project takes documentation very seriously and strives to have relevant, up-to-date documentation available for all users
- feedback about the current state of the documentation
- ongoing work
- guidelines for contributing
- what's missing?
- what should be fixed?
- what's obsolete?
- Automation of CVE Verification
- Moderators: David Reyna, Shachar Menashe
- Propose Sharing CVE information via Layer Index to assist automation
- CVE checking by package version does not capture OE patches, so false positives
- CVE management is expensive, automation makes it feasible, data must be programmatically available
- Proposal to contribute to the Layer Index to share CVE information
- Proposal to validate this for internal tools (CVE Checker, SRTool)
- Proposal to validate this for external tools (VDoo, ...)
- Open discussion in ways to help CVE management
All times are in UTC.
- 1530 - 1555: CVE
- 1555 - 1615: BSOM
- 1615 - 1640: Documentation
- 1650 - 1720: X11/sato
- 1730 - 1800: Layer Quality
- 1810 - 1840: BSP
- 1850 - 1920: Maintainer's Life
- 1930 - 2000: LTS