Hello, On Mon, Jun 13, 2016 at 05:07:08PM -0400, Leo Famulari wrote: > On Mon, Jun 13, 2016 at 08:24:23PM +0000, ng0 wrote: > > What if we patched gpg-1 to not colide with gpg-2? > > > > For example, move gpg and gpgv and man pages of them > > for gpg-1 to something which has -1 in its name. > > On the other hand this would have to be consistent and be followed > > straigth to not colide again. I'm not sure if this > > approach will work out in the long run. > > Unless this is a configuration option supported by upstream GnuPG, I > don't want to do that. For the same reason, my patch does not affect > gnupg-2.0.
I agree with this argument - one of our principles is to mess as little as possible with upstream packages. In this case, since there is a special configuration option for gnupg-2, we may use it, but then we would have to pay the price that both gnupg-1 (or gnupg-2.0) and gnupg-2.1 could not be installed together any more. Maybe it would then be consistent to drop gnupg-2.0 from the distribution. Finally I managed to make gnupg-2.1 work together with mutt, but it was not completely straightforward... So we should document the change if we make it. By the way, should we maybe make pinentry a propagated input of gnupg-2.1? If I understand correctly, gnupg-2.1 will not work without it (and mixing pinentry from Debian with gnupg-2.1 from Guix was one of the reasons for gnupg not working at first). Andreas
