Hi Simon, zimoun writes: > On Fri, 17 Sept 2021 at 01:40, Mark H Weaver wrote: >> Liliana Marie Prikler wrote: [...] >> > If done this way, there'd be the benefit that modules with packages >> > using this thing would have to explicitly request the presence of the >> > symbol through their use-modules clauses. >> >> Actually, for better or worse, Guile's '@@' form does not require the >> named module to be imported using 'use-modules', so I don't think this >> benefit strictly exists as stated above. However, I agree that it's >> good practice to list all imported modules in the '#:use-module' clauses >> at the top of the file wherever possible [*], and that there may be some >> benefit in declaring the use of 'computed-origins' at the top of each >> file. > > I am not deeply familiar with Guile module. > > I chose to put this in (guix packages) instead of its own module > because the module would contain only one function and nothing > exported. The aim for now, as discussed, is to not make this 'method' > part of the public API. > > Then if the function is not exported by the module, the '#:use-module' > does not have an effect, right? It's true that it would have no effect on the set of available bindings, and that's an excellent point. Perhaps Liliana could clarify what she had in mind, or better yet, propose a patch. Please don't let me be a blocker on this thread. I contributed a few thoughts, but I don't have time right now to shepherd this issue, sorry. Regards, Mark -- Disinformation flourishes because many people care deeply about injustice but very few check the facts. Ask me about .