Could you elaborate on when they wouldn't treat regressions as blockers? I'm not very familiar with LLVM and couldn't gather that from the "How To Release LLVM To The Public" docs [1].
Well, e.g. they say: "Only critical bugs found during this testing phase will be fixed." - so e.g. not all regressions are necessarily release blockers.
Besides, whatever the "process" says, isn't it ultimate up to the release manager?
Well, this says that zig team reported those bugs as "release blockers".
Doesn't mean the LLVM release team also considered them release blockers (or that it had too).