Olaf Hering wrote:
> On Fri, Jun 12, Kevin J. McCarthy wrote:
> 
> > Would you mind applying this patch and running mutt with -d 5
> > to get debugging output, just so I have a clearer picture where
> > imap_open_connection is bailing.  I have added a NONULL around the
> > capstr, so hopefully it won't segfault.
> 
> Thanks for the patch. Below is the output, slightly tweaked.
> If the password is correct in the first attempt it goes straight to
> =INBOX. If first attempt is wrong but second is correct no mailbox is
> active. But its possible to go manually to =INBOX.

Ah!  I misunderstood the cause of the crash.  If your first login fails,
then the *second* time you try to login, you were getting the crash.  Is
that right?

I see why this is happening: it's freeing the capstr, even if
authentication fails.

So the question is whether the capstr should be kept, or whether, the
code should do something like:
  if (idata->state == IMAP_DISCONNECTED)
    imap_open_connection (idata);
  else
    imap_check_capabilities (idata);
inside imap_conn_find().

Brendan, do you have any feedback?

-Kevin

Attachment: signature.asc
Description: PGP signature

Reply via email to