I'm reading the CP/M Assembly development books "CP/M Assembly Language Programming" by Ken Barbier and "Mastering CP/M" by Alan Miller, which are interesting also for their their different approaches to reusable code building blocks. Barbier builds a library of subroutines, Miller focuses on macros..

Back in the day, what building blocks were more common or recommended, macros or subroutines? What were the best practices?

Some reasons that I would shy away from macros were that the DRI assemblers did not cleanly show the macros in the listing file (I think "$*MACRO" got close, but was still messy), Macros inline the code so can lead to unexpectedly large binaries, and the DRI macro facility was never quite as robust as I wanted. But, I did use macros plenty, and certainly leveraged the Z80.LIB macros extensively. It's my opinion that macros have a place, as do subroutines, and so it's just a matter of picking the right approach for the situation. There are even hybrid approaches, where macros "soften" the function calls but subroutines avoid code bloat (a macro handles/hides the setup for the subroutine call, but there is still a subroutine library required) - more akin to a high-level-language.