So here are my questions:
1. Why are we building some components of Che for multiple arches? Is there a compelling reason to have a dashboard and server/factory, without the ability to launch a workspace running che-code?
2. Can you even install the Che operator without a compatible DevWorkspace operator for your arch? (I would be surprised if this is possible.)
3. Could we simplify the Che build infrastructure to use `podman build` or `docker build` instead of `docker buildx`, and focus on the single end to end deployment of Che on amd64?
For those worried that this means they can't use Windows to Mac to connect to their cluster, don't worry - I'm not suggesting we drop support of darwin or win32 arches.
I'm suggesting we drop container builds for arm64, s390x, and ppc64le, since there is no way to use those containers for a COMPLETE Che 7.60+ deployment with a vscode-based workspace.
What do you think, community members?
Thanks,
Nick
--
Nick Boldt (he/him/his)
Principal Software Engineer, RHCSA
Productization Lead :: OpenShift Dev Spaces (formerly CodeReady Workspaces)
IM: @nickboldt / @nboldt / https://divbyzero.neocities.org
“The Only Thing That Is Constant Is Change” - Heraclitus