Jeremy Maitin-Shepard <[EMAIL PROTECTED]> writes: [..]
> This patch does work because there is no guarantee that the match value > that is returned to the caller is actually a suitable value for the user > to have typed, since getting a value from the history is equivalent in > terms of the completion system to the user having typed it. In > particular, the match value might not even be a string. Notice that > with this patch, if you use C-x b, pick something, and then open C-x b > again, and look at the history, you'll see that [Object object] is > listed, because with switch-to-buffer, the match value is the actual > buffer itself, not a string. Indeed, I wasn't aware of that problem. Is the buffer history the only one not storing strings? > I agree that what you want is a useful feature in some cases like when > completing a URL, but I think it would require some change to the > completion system in order for it to work. Even then, it is not clear > how to make it work with switch-to-buffer. It wasn't so much in the case of urls that I thought it would be useful since in this case auto completion already does what I want (namely it has the effect that what is stored in the history is the url I ended up visiting). It was actually in the case of the command history since having entries like "cmd_" in the history didn't seem to me to be very useful. Of course, a workaround to this would be to enable auto completion for the command history as well -- but how to do this? It seems like minibuffer_auto_complete_preferences is ignored in this case. Is that right? And why? -- Martin _______________________________________________ Conkeror mailing list [email protected] https://www.mozdev.org/mailman/listinfo/conkeror
