I don’t think merging the break to master and waiting for all of the maintainers of dependant packages to fix it is the best approach here.
Ever tried to bump glibc, gcc or openssl? Getting every single leaf-package to build (and also to work which is a totally different story btw) is just not practical. A better solution would be to fix Hydra’s email notifications and to re-enable it after that.
I agree with @jonringer here. We also have ZHF iterations for a reason IMHO.
Also it seems as Hydra-related impurities are forgotten here: sometimes it happens that a package breaks due to too much load on a builder (I guess) or something fails with some kind of Illegal instruction error depending on the builder’s CPU. Another story are timed out tests or the fact that we don’t have too much aarch64 machines so a lot of those builds get cancelled occasionally.