Hello Pierre and Ludovic, Pierre Neidhardt writes: > I didn't know Oleg was going to submit a patch, maybe there was some > confusion in the process :p I've mentioned it in our mailing list discussion, see https://lists.gnu.org/archive/html/help-guix/2018-04/msg00130.html Nevertheless thank you for a 31178! :-) > I've quickly reviewed our two patches. Overall I'd opt for mine :p > > - It's the latest version. > - It has more inputs. I thought that we stuck with an ‘inxi’ Perl script version (31178) and there was no patch until I pushed one. It's better to get 31178 version than 31176 ofcourse because of ‘inxi’ version. But I don't like all those ‘propagated-inputs’ except ‘perl-*’ and I guess we could do better. Maybe we could use PATH for non ‘perl-*’ inputs the same way as in 31176. WDYT? > I think we should use propagated-inputs and not native-inputs. What do > you think? I'm not sure about 31178 currently. Usually we prefer to avoid ‘propagated-inputs’ if possible in Guix, because all ‘propagated-inputs’ will be in a Guix profile and could lead to collisions. If you are talking about 31176, then no, because I've used a wrapper to find all ‘inxi-minimal’ or ‘inxi’ requisites. It was easy to do, because ‘inxi’ in 31176 is a Shell script, not a Perl script. Thanks to Chris Marusich for an idea. […] [1] https://debbugs.gnu.org/cgi/bugreport.cgi?bug=31176 Oleg.