Hi Ludo’ ,

first thanks a lot for your review,

I see two issues here:

  1. This is stepping on the toes of upstream: why are we providing a
     non-trivial program like this downstream?

In my understanding this is the fundamental issue, which could be a shipstopper. Please correct me if I'm wrong. We kind of are obviously even if restic-guix is already in the master branch ( ) . In my opinion the way forward should be: a. In this scenario we merge the current 72803 (after addressing your other comments) and we take this risk

b. In this scenario we remove completely the current incomplete restic-guix command implementation from master, as it makes not much sense to have it incomplete as it is right now.

I view scenario a and scenario b as mutually exclusive but I may be missing some implication, what is your opinion on this?

  2. There are stylistic issues: use of ‘first’ & co. (info "(guix) Data
     Types and Pattern Matching"), use of ‘error’ (it is too generic and
     user-unfriendly), custom argument parsing procedure.

I will address these comments only if we decide to go forward with scenario a.

Thank you so much for your work,

cheers

giacomo