3 - redesign to avoid use of setjmp/longjmp.
7 - clean up -T errors in pptp-command,
8 - implement GRE reordering from PoPToP to make client more resilient.
13 Autoconf/automake this beastie.
14 Fix do_gre_copy semantics to properly handle multiple calls.
15 Like the call manager, there should only be one gre-copy process
16 per connection. This process needs to funnel packets to different
17 output tty's, though, depending on the call-id embedded in the
18 GRE encapsulation. pptp.c must then be modified to use an
19 already-existing gre-copy process in the same way it uses an
20 already-existing call manager.
21 Support server operation.
22 Obviously, this is more process trickery. Upon receiving a call
23 request, the call manager must automagically launch new gre/pppd
24 processes to handle it. Also, some portions of the server control
25 protocol in pptp_ctrl.c have not been completely implemented.
26 More flexible configuration system.
27 pptp needs to read in config options from /etc/pptp or somesuch.
28 This is very necessary for server operation, since presumably we
29 will want to limit the hosts we give PPTP access to. Maybe not,
30 though. We could allow everyone and then allow pppd to do the