Well, here is my latest problem...
I have done what Jeremy suggested with bpl8's socket_close function to
fix the fail-the-passwd-attempt-to-throw-off-players bug (by copying
close_socket, renaming it close_badpwd, removing the offending code at the end,
and setting up a CON_BADPWD state.) Now I get another bug...
I repeat the attack (I login in, pwd, in CON_PLAYING state), then login
in again and fail the passwd attempt. Instead of closing the connection, the
connection is still ESTABLISHED (as per netstat), and the first bit of input I
give the hung socket throws me off. Now the REAL bug. If I just sit on the
hung socket, and when my legit player goes to quit, BAMN. Instant socket
closure for both sessions. The "legit" session does not get the 0-5 menu as
normal, just the cold "Connection closed by foreign host" thing.
Now, is the code at the end of close_socket hiding a subtle bug, or did
I introduce one in my close_badpwd code (which is an *EXACT* copy of eclose_socket, execept for the code after the comment at the end)? This one bug
has REALLY got me stumped....
TIA,
Mark Crichton
macricht@indyunix.iupui.edu
(Wow....just got an idea for the eq. globbing problem. Got to check if it will
work now....)
This archive was generated by hypermail 2b30 : 12/18/00 PST