Re: bug#8050: Gnus does not connect to my IMAP server any more

[Ted Zlatanov <tzz@lifelogs.com>]

On Mon, 07 Mar 2011 12:23:59 -0600 Ted Zlatanov <tzz@lifelogs.com> wrote: 

TZ> On Sat, 05 Mar 2011 15:13:53 -0500 Stefan Monnier <monnier@iro.umontreal.ca> wrote: 
SM> - But when I'm asked for a user name there is no default, whereas earlier
SM> my local user name was used as default (which happens to work for me on
SM> most of the machines to which I can connect).  Please use the local user
SM> name as default in that prompt.

TZ> I'll add that.  I thought I did but must have forgotten.

This is done.

SM> - I'm asked whether to save the password regardless of whether the
SM> password and user names are correct or not.  That's really bad.

TZ> auth-source.el doesn't know at that time whether the authentication will
TZ> be successful.  I think you're saying we need an `auth-source-save'
TZ> function to be called after the fact instead of a ":create t" parameter.

I added this as part of the auth result.  So you can say (example from
nnimap.el, where the 3rd element of the credentials is the
:save-function property of the search result):

                    (when (functionp (nth 2 credentials))
                      (funcall (nth 2 credentials)))

after a successful login.

This means that everyone using the ":create t" parameter will have to
adjust for the :save-function.  Currently that's just nnimap.el and
sieve-manage.el (which I'll modify when you and Lars agree this is the
right approach).

SM> - After refusing to save the password in authinfo.gpg, I get a message
SM> along the lines of "auth-source-search: CREATED 1 results ...".
SM> I don't care whether "results" is replaced by "result" when there's
SM> only 1, but this looks like a debug message which should disappear.

OK, it's gone.  But we show a shorter one with `message'
unconditionally, which I think is the right thing to do.

Ted