On Fri, 05 Jun 2015 14:57:40 -0400 Daniel Kahn Gillmor <d...@fifthhorseman.net> wrote: > On Fri 2015-06-05 13:25:42 -0400, Michael Biebl wrote: > > Well, assuming that seahorse does work properly with gnupg2. > > That's basically my question. > > The seahorse source code seems to actually behave completely differently > depending on whether it is built with modern versions of any branch > (meaning: gpg >= 1.4.10, or gpg2 >= 2.0.12) versus older versions. (see > pgp/seahorse-gpgme-key-op.h). :(
Right, I just had glanced quickly at configure.ac which indicated that this is a compile time behaviour and not something which seahorse determines during runtime. > This is not great engineering practice, because the version built > against isn't guaranteed to match the version that's running. I agree. > That said, even oldoldstable builds and runs "modern versions" by this > metric. Right, but since we build against gnupg by default, a hard depends atm seems still warranted, and we can't drop it just yet. > I just tested seahorse on a minimal-ish unstable gnome install, where i > did "dpkg --force-depend --purge gnupg". > > Unfortunately, it looks like seahorse embeds the string "gpg" in it, so > it's looking for /usr/bin/gpg. > > Running seahorse in this configuration produces lots of errors of this > form: > > operation-Message: couldn't initialize gnupg properly: Invalid crypto > engine > > The attached patch should resolve things for future versions of > seahorse, though, both on build-time detection and on runtime > flexibility. Since carrying patches downstream comes with a maintenance cost, we try to avoid that as much as possible. It thus would be great if you can forward this upstream, since this doesn't look like a Debian specific issue. > (the attached patch touches both ./configure.ac and ./configure -- since > the package appears to be doing autoreconf, maybe the modifications to > ./configure are unnecessary) > > The only failures i'm now running into with seahorse like this are > failures due to gcr_importer hard-coding paths to gpg as well, so those > are bugs i'll file separately.. Same here, would be great if you can file them also upstream. Thanks a lot Daniel, your efforts are really appreciated. Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
signature.asc
Description: OpenPGP digital signature