On Tue, Jan 03, 2012 at 10:19:42AM -0800, Junio C Hamano wrote: > For people who have been used to interact with git-svn from the terminal > but has *_ASKPASS for reasons other than their use of git-svn in their > environment, the change to the username codepath is technically a > regression, as they used to be able to see and correct typo while giving > their username but with the patch *_ASKPASS will kick in and they have to > type in bline, but I am not particularly worried about it. It is something > you type very often and committed to your muscle memory anyway. There is one difference between how git and ssh use the ASKPASS variable. In git, we try it _first_, and fall back to asking on the terminal. For ssh, they first try the terminal, and fall back to askpass only when the terminal cannot be opened. If we tried the terminal first, then it wouldn't be a big deal to use *_ASKPASS more frequently, since it's a fallback. Of course, that in itself might be a regression for some people. I wonder if we should make the order: 1. GIT_ASKPASS 2. terminal 3. SSH_ASKPASS to help make our use SSH_ASKPASS better match that of ssh. I dunno. I am not an askpass user these days, so I don't know what people expect or want. -Peff -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html