Junio C Hamano <gitster@xxxxxxxxx> writes: > Derrick Stolee <derrickstolee@xxxxxxxxxx> writes: > >> This is mostly because most consumers in the Git codebase will want >> these methods, and only the most fundamental hashing code will care >> about hashing separate from the_hash_algo. So, the most-natural name >> for "the header file that lets me hash stuff" would be "hash.h" and >> "hash-basics.h" is for "I'm doing very simple, low-level hashing". > > I do agree with the reasoning for "most people want 'hash.h'"; what > is called the "basics" sounds more like for advanced users of the > low-level machinery, but somehow I expect "basic" would the one to > be used in sample programs in tutorials or something. "ll-hash.h" > for low-level hash features? I dunno. > > But whatever the name for the header file meant to be used by > "advanced callers to low-level machinery" would be, I like your > choice of the latter variant between the ones Elijah presented. FWIW, I also think this is a good idea. I suspect that a lot of the libification effort will need a similar distinction between "a subsystem supporting the needs of the git binary" (aka high-level) and "basic, reusable functionality that can be its own library" (aka low-level).
