Sam Steingold <[EMAIL PROTECTED]> writes:
> I think we should keep all phone numbers as strings, and have a user
> variable bbdb-phone-canonicalize which should be nil for "as is" or a
> function which is applied to the phone and should return the
> canonicalized phone string.
>
> Same for zip codes (bbdb-zip-canonicalize).
>
> Unless someone objects, I will do something like this next week.
Hm, if we have bbdb-zip-canonicalize, then the old problem re-appears:
whenever a new format of zip codes is "discovered", somebody has to
update the code. The current solution allows people to add output
functions and print functions which take an address and do some output
(either to the BBDB buffer or to the bbdb.tex file). I like that
approach better. If you really want to canonicalize US zip codes, you
can write your own little function to frob all US addresses.
The patch I've just mailed to the list ("zip codes as plain strings --
patch and file format change?") transforms all zip codes into strings
and tries to get rid of all the zip code formatting stuff: The zip
code itself is printed as the user entered it. The downside of all of
this is of course, that this required a new file format.
The transformation from strings into zip code datastructures and back
is still present, of course. It has moved into bbdb-migration.el --
never to be updated again. This solves my major problem: I had to
modify the code for dutch zip codes, for russion zip codes, for
brazilian zip codes... and I'm sick of it.
Alex.
--
http://www.geocities.com/kensanata/emacs.html
_______________________________________________
bbdb-info mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/mailman/listinfo/bbdb-info
