I'm running current CVS bbdb.
I have a few records where a user has two net records with the same
prefix. E.g.:
[EMAIL PROTECTED], [EMAIL PROTECTED]
If I type "bob" and then issue bbdb-complete-name, I want both
addresses to be a candidate for completion (I have
bbdb-completion-type set to nil).
What happens now is that one of the two is chosen randomly, based on
whichever one is first in bbdb-hashtable, I guess.
This comment from bbdb-complete-name seems to shed some light.
;; If we have two completions which expand to the same record, only
;; treat one as a completion. For example, if the user asked for
;; completion on "foo" and there was a record of "Foo Bar <foo@baz>",
;; pretend the first completion ("Foo Bar") is valid and the second
;; ("foo@baz") is not, since they're actually the *same* completion
;; even though they're textually different.
;;
;; realistically, because of the hash, you're not guaranteed
;; which will actually be selected. this hurts, of course.
Except that it is not always a choice between just "Foo Bar" and
"foo@baz", there could also be a "foo@yahoo" associated with the same
record. I think current implementation chooses randomly among the
three when completing "foo"
--
matt
_______________________________________________
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/bbdb-info
BBDB Home Page: http://bbdb.sourceforge.net/
