On Tue, Mar 24, 2020 at 11:25:09PM +0100, Johannes Schindelin wrote: > > Er, isn't that what my patch did? I'm fine if you have another approach > > to present, but I'm worried we might be duplicating effort. > > I missed that your second patch made `GPG` lazy, too. > > My version is slightly different from yours, though: I do not insist on > setting the environment variable `GNUPGHOME` only after the `mkdir` > succeeds, as the `gpg --sign` later on might fail anyway, which means that > we _already_ could end up with `GNUPGHOME` set and the prereq `GPG` _not_ > set. Yes. That mkdir could also be pushed down into the GPG prereq for the same reason. It's pretty unlikely the mkdir would fail, and I thought it would be less confusing (in the off chance that anybody even looks at it when GPG isn't set) if we had a GNUPGHOME that pointed to an _actual_ directory, instead of something that didn't exist. But that's kind of an arbitrary cutoff. The GPG prereq is also importing keys and other stuff, so the state of that directory when GPG isn't set is undefined (but again, nobody's supposed to be looking at it, so...). -Peff
