Re: New Emacs facilities from Gnus

[Daiki Ueno]

2007/11/1, Simon Josefsson <address@hidden>:

> > * password.el:
>
> I don't care strongly about it, but it was created as a consequence of
> finding duplicated code in various places that did something similar.  I
> think those code duplications are still present.

In that case, I'd suggest that the low-level libraries which need
passphrase input should provide a callback mechanism for reading
passphrases.  For instance, though epg.el does not cache passphrase by
itself, application programmers can add their own cache mechanisms in
the caller side as they desire:

(let ((c (epg-make-context)))
  (epg-context-set-passphrase-callback
   c
   (lambda (context handback)  ...passphrase caching code...))
  (epg-encrypt-string c "test" nil))

If all the libraries support this, application programmers can put
these code in the single place, and library programmers will never be
troubled with the default behavior of passphrase caching ;-)

Regards,
-- 
Daiki Ueno