This release, @worldofpeace asked me to set up a Go / No-Go meeting for the release of NixOS 20.03, Markhor. This meeting is to include stakeholders from various critical subsystems of NixOS. We hope to include people representing server deployments, macOS, ARM, maintainers of various Desktop Environments and GUI ecosystems, and other key participants.
Where / When
We are planning to hold a 1 hour long Jitsi Meet call some time in the next few days. This is video-optional, and participation by IRC is also okay ā but real-time audio participation at a minimum is preferred. Once we get some availability information, I will send out a meeting invitation.
We will try to make the time ideal for everybody, but we will likely not be able to do so. Please fill out this form with your availability so we can do our best: NixOS 20.03 Markhor Go / No-Go - When2meet
Come Prepared
During this call we will review bugs in the 20.03 milestone, and this is also your chance to propose other release blocking bugs or pull requests for review. Please attach these proposals to this thread so they can be collected ahead of time for the call. The outcome of this meeting is to decide if weāre ready to release, or if it needs to be delayed within reason. Ultimately this final decision will be up to the discretion of the release managers (@worldofpeace and @disassembler).
What is a blocker?
A blocker issue is a bug that must be fixed and blocks the milestone releases.
We cannot proceed without them being fixed or having known workarounds. Things that arenāt right for this meeting are bugs that would ābe nice to haveā or are non-essential. Blocker bugs will affect critical path things such as everything already in the 20.03 jobsetās tested constituents and anything else we consider to be critical path for NixOS.
Agenda
A brief explanation of the intentions of the meeting
Roll call: participants will identify what subsystems and users they are representing.
Blocker evaluation: review each listed bug and related issues and decide on priority. Make sure your bug is in this thread before the meeting.
Release Notes Review
Go / No-Go Decision: We will likely release even with known issues. The goal here is to decide which ones are blocking and not.
Setting the Goal Date: what day should we have these completed by, or the release is going on without them?
We can usually toggle the āstable releaseā flag and get a release out within a few hours, so it is pretty quick. Though it isnāt uncommon for us to release on the 35th to 40th of the month.
Iāll hopefully be there for NixOS Containers, some services with a non-trivial upgrade paths (such as Nextcloud, matrix-synapse etc.) and (if needed) for issues with bigger rebuilds (after helping out during the gcc9 and glibc-2.31 upgrade).
Please attach these proposals to this thread so they can be collected ahead of time for the call
I think #82295 should be merged. It is just an addition to release notes, informing people that they wonāt get IPv4 addresses on bridges via DHCP on nixos-20.03 anymore.
This almost locked me out of my machine and might break other peoples setup when the release happens.
worldofpeace and disasm decided NO-GO at this time, and to have second review meeting at the same time on Wednesday: 2020-04-08T19:00:00Z ā 2020-04-08T20:00:00Z
Iād also like to mention, since we lack a āNixOS Release Criteriaā, if your subsystem wasnāt represented in this meeting and thereās something youād consider a blocker, now is the time to mention this.
Samuel Leathers (@disassembler) and I decided NO-GO today in the meeting.
Weāre going to have another review meeting on Friday: 2020-04-10T19:00:00Z ā 2020-04-08T20:00:00Z.
I wasnāt prepared to take notes, the video will help with who was representing what at the beginning.
It seems also that some people assumed it would be next week Friday, so we had some missing people.
Reviewed
Anything not reviewed this time has been fixed
I brought it up tangentially, I donāt recall the status of this because I havenāt been able to track kernel development very well. It would be helpful if someone in the know with kernel can followup on it (@vcunat maybe since you did the bump). Last time I checked this was even a problem on one of my laptopās.