On Sun Nov 30 17:12, Matthew Johnson wrote: > I'd be interested in knowing what caused that line to be added to > gpg.conf, since I've only ever seen gpg-agent used with environment > variables. > > If a package automatically updates gpg.conf and then leaves it in a > non-functioning state, that's a potentially RC issue. In which case this > bug should be cloned and the new one left at RC and this one > downgraded. >
OK, it looks like seahorse (in Etch) munges gpg-agent-info setting in gpg.conf. I _think_ it's meant to remove it on exit, but obviously if it's been killed (or the machine crashed) and then not restarted before upgrade, this will be left in the user's gpg.conf. The version of seahorse in Lenny doesn't seem to do this, but I'm really not sure what to do about these left over values. Seahorse in Lenny could be made to remove them when started as that user, but this doesn't solve the problem of people changing to other gpg agents. I've CC'd the seahorse maintainers for input. I'd reassign it to seahorse, but I'm not sure on the severity or even which version should have the bug assigned to it. Matt -- Matthew Johnson
signature.asc
Description: Digital signature