This is not the first time multibuild was suggested for kernel. https://jira.suse.com/browse/PM-1015 It has been rejected because OBS is not ready for multibuild. It was not ready then and is not ready now. (In reply to Jiri Srain from comment #5) > You have known about this request since months. What's the kernel team's ETA? OBS people have known about the request to control repository flags for multibuild flavors individually for years. https://github.com/openSUSE/open-build-service/issues/3574 Sure, it's possible to implement workarounds in a lot of packages but maybe fixing the root cause would be more productive? There are is an additional problem with OBS multibuild support: package actions are broken for multibuild in the web UI. https://github.com/openSUSE/open-build-service/issues/12913 https://github.com/openSUSE/open-build-service/issues/7866 https://github.com/openSUSE/open-build-service/issues/7156 Apparently it's confusing repeatedly even to people doing packaging daily, reported repeatedly but not fixed. This is much worse for kernel developers who are often not packagers and interact with OBS only occasionally. For them sane interface is even more important, and figuring out the workaround less likely. Coupled with the previous issue that makes manual intervention more likely required it's pretty damning. (In reply to Richard Brown from comment #18) . > > I'm now very confused. > > The premise of the original bug was that 'source links' need to be removed > from the kernel package in Tumbleweed. > > The premise of this bug many months later is that 'source links' need to be > removed from the kernel package in Tumbleweed and ALP Dolomite. > > kernel-obs-build, kernel-obs-qa, packages in devel:gcc, all seem to me to be > chaff thrown up to confuse. > > Unless I'm missing something they seem to be wildly out of scope to the > matter at hand - the kernel in our future products needs to be built using > methods like multibuild. > > I don't think anyone minds what happens in devel projects, but our > future/now present products are using 'git based source management' and the > long communicated changes needed to facilitate that really need to be > implimented now. Your hyperbolic comments do not contribute to solving the technical problem, and have potential to adding people problem in addition. I want to believe this is hyperbolic and not serious. In any case your contribution towards solving this problem has been less then none.