building with full optimizations and lots of 3rd party libraries (esp if they are c++ or use lots of swift-generics w/type inference) can indeed take that long

building with full optimizations and lots of 3rd party libraries (esp if they are c++ or use lots of swift-generics w/type inference) can indeed take that long caused by lots of inlining, de-virtualization, static specialization, whole module optimization and static linking (or just linking in general)... a comparable clean build with debug flags (no optimization) might only take a few minutes with incremental builds taking a few seconds or so, i guess thats why its tolerated... the other thing is ci build pipelines are reaaallly slow so if you made the same build locally on your m1 pro its probably gonna take 1/4 the time as the ci (ime anyways)...many ci's are slow af....