Gardner Cohen writes:
 > Is the interface to vm-auto-select-folder public?  Can I rely on
 > getting at the message pointer argument in my advice (which would
 > allow me to replace the bbdb/vm-update-record call above with
 > something far simpler and less reliant on the inner workings of vm)?

No guarantees about the internal interfaces.  On the other hand,
I don't change them whimsically either.  vm-auto-select-folder
has had the same interface since it was created.

The problem is, I think, BBDB doesn't yet understand virtual
messages and that the function you're using shoudln't be called
int he context in which you're calling it.  For a message m
(virtual or real), the buffer where the message contents can be
found is
  (vm-buffer-of (vm-real-message-of m))
The message starts at
  (vm-start-of (vm-real-message-of m))
and ends at
  (vm-end-of (vm-real-message-of m))

This is very different from what bbdb/vm-update-record is doing.
I think bbdb/vm-update-record is meant to be run only from
vm-select-message-hook.  Even so, it is peeking at internal VM
variables, but I think it will still work anyway.

Reply via email to