Expose One_or_two as a Mina_stdlib submodule
This change begins the incremental process of consolidating small libraries into larger, more cohesive modules under Mina_stdlib. As a first step, One_or_two is now accessible as Mina_stdlib.One_or_two while maintaining backward compatibility.
Modified files have been updated to reference Mina_stdlib.One_or_two instead of directly importing One_or_two. This follows a pattern that will be extended to other libraries like String_sign, truth, state_or_error, linked_tree, immutable_array, hex, gadt_lib, bash_colors, and allocation_functors.
!ci-build-me
!ci-bypass-changelog
!ci-build-me
Won't this slow down compilation?
I don't think it would be significant in this case. The modules inside Mina_stdlib are small, fast to compile and independent (therefore can be compiled in parallel). Mina_stdlib will be compiled fast and the other modules, initially depending on One_or_two, will be compiled. In addition to that, we don't modify Mina_stdlib very often or One_or_two, so we don't often need to re-compile them. It might have a low overhead (I would say in us or ms - ?) so negligible for us.
It is a good point to raise though, in particular when modules are huge. Thanks for asking.
!ci-build-me
!ci-build-me