On Mon, Jan 27, 2014 at 10:11:14AM +0100, Ludovic Court??s wrote: > Ok. Andraes' and Ludo's explanations convince me. However I'm skeptical that > the Octave devs would be quite so convinced. And removing the propagates-inputs > will mean patching to the Octave source and I don't know how difficult this will be. The patch that would be great upstream is: AC_PATH_PROG([MAKEINFO], [makeinfo]) AC_SUBST([MAKEINFO]) and then use ???@MAKEINFO@??? wherever ???makeinfo??? is expected in the source (similarly for ???less???, etc.)
Ludo???. Having thought about this some more, looked to see what is currently in the octave source and "discussed" the issue on #octave I think now the best solution is to simply remove all the propagated-inputs from the package (and leave inputs and native-inputs as they are). Rationale: * Octave "works" without all these programs (albeit in a rather featureless fashion). If a user wants to add the feature, then she just needs to guix package -i <foo>. * It seems to have been a deliberate decision by the octave developers to rely on $PATH to select the appropriate version of these external programs. * Changing this behaviour would involve alterations to the octave source touching many files, and I think upstream would be unlikely to cooperate with us. The disadvantage of this approach is, that a guix user who installs octave, but not the other packages, gets only a barely functional version. Perhaps we need a (recommended-inputs `(...)) like in debian. Comments? J' -- PGP Public key ID: 1024D/2DE827B3 fingerprint = 8797 A26D 0854 2EAB 0285 A290 8A67 719C 2DE8 27B3 See http://sks-keyservers.net or any PGP keyserver for public key.