hello -
i've been using keychain for some time now, and for a while i was confused by the result of trying to start a fresh ssh-agent via `keychain --timeout 600.` i'd been expecting that ssh-agent itself would be passed the timeout via its "-t life" option, ensuring that any keys added to it share that timeout. instead it looks like the timeout is only passed to the ssh-add invocations for keys to be immediately added. since the syntax for the ssh-add and ssh-agent timeouts is the same, it looks like it would be a relatively minor change.
are there aspects of this suggested change that wouldn't make sense or that might have unexpected ramifications? it would be a change in behavior, which may be a strike against it... failing all else, perhaps the man page could be changed from describing the timeout as "conveyed to ssh-agent" to "conveyed to ssh-add?" (i'd guess that was the source of my initial confusion)
thanks for the great utility!
andy