Peter Eisentraut <[EMAIL PROTECTED]> writes: > The developers explained that: > > 1. You need to set a valid, supported terminal to use pinentry. > Terminal type management is a pain, but this is the fault of the > operating system. Making sure the terminal type is right is not a > prohibite requirement in my mind.
It's not that I have problems "managing the terminal type", it's that sometimes I don't have a full-screen terminal available. [I will admit though, I find the full-screen pinentry annoying in that it completely erases the screen and obscures whatever text proceeded it; if there was a non-full-screen mode, I'd probably use it even when I had a proper terminal available.] > 2. Since pinentry is intended to react "asynchronously" to passphrase > requests, it cannot support a mode that writes onto the terminal > inline. I'm not sure what "writes onto the terminal inline" means, but AFAICS, the use of curses to make output pretty is orthogonal to other issues (e.g., if it's necessary to use character-by-character input for some reason, instead of simply using cooked mode with echo turned off, you can do that without curses). > It was pointed out that not using the agent would be better if > you wanted to control gpg's screen output more closely. This last point doesn't make any sense to me. I don't know why they think people use the agent, but I expect for many people it's not because it's a pretty interface, it's because it means you don't have to enter your passphrase 5,000 times in a row if you're encrypting/signing a lot of files. Certainly that's why I use it. So it's not that I want to "control gpg's screen output more closely", it's that I want to use the agent's functionality, but I don't always have the luxury of a full-screen terminal when doing so. > I might add that one could write a no-curses pinentry variant oneself > with relative ease. Er, sure, I guess I'll have to do that. [See my next point though.] > Above all, note that the package is called "pinentry-curses". If you > want "pinentry-plaintext", you might have a valid request, but you have > come to the wrong place. I dunno, it seems natural to me to think of it as really being "pinentry for terminals" -- that it uses curses seems an implementation detail -- and for it to fall-back to some sort of non-full-screen mode when TERM is undefined or specifies a non-capable terminal type. That would certainly be the most straight-forward interface from a user's point of view. So maybe I'll give a shot at implementing a "pinentry-dumb"; it would be nice if I could fold back such changes into an existing tool though -- surely others will encounter the same problem. Thanks, -Miles -- We are all lying in the gutter, but some of us are looking at the stars. -Oscar Wilde -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
