1 Sun Dec 13 13:35:12 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
2 * context_loop.c: Moved "remove session" into its own function.
3 Check for dead webcit child process before each transaction.
4 * Ignore SIGPIPE; this seems to magically make the program bulletproof
5 against broken HTTP connections.
6 * Added a housekeeping thread to periodically remove dead sessions
7 and kill idle sessions.
8 * messages.c: added "delete message" functionality
9 * messages.c: added "move message" functionality
10 * userlist.c: added. This adds "userlist" and "show user" functions.
11 * roomops.c: fixed buggy implementation of ungoto
12 * Set up an option in output_headers() to optionally print the most
13 often used opening HTML as well as the HTTP headers.
14 * Display express messages in cute little JavaScript popup window.
16 Fri Dec 11 21:14:36 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
17 * Brought over message reading and entry functions from old WebCit
19 1998-12-10 Nathan Bryant <bryant@cs.usm.maine.edu>
20 * context_loop.c: really fix the SO_LINGER stuff
21 * webserver.c: remove commented-out references to SO_LINGER
23 Wed Dec 9 18:50:46 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
24 * context_loop.c: After fork(), child process closes the HTTP socket
25 so when the parent process closes it, it actually _does_ close.
26 Otherwise, aIEeeee and possibly other browsers will hang.
27 * Added support for graphics images retrieved from the Citadel server.
28 * Set up login banner graphic and background image graphic
30 1998-12-09 Nathan Bryant <bryant@cs.usm.maine.edu>
31 * context_loop.c, webserver.c, webserver.h: SO_LINGER and locking fix
33 Wed Dec 9 11:13:48 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
34 * Smoothed out some of the login/logout code. Failed authentication
35 now works the way it should.
36 * Added support for new user login (still need to handle registration).
38 Tue Dec 8 15:45:42 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
39 * Discovered that the HTTP socket problem is related to a socket that
40 isn't properly lingering on close. Temporarily hacked a sleep(10)
41 into context_loop.c until it is figured out.
43 Mon Dec 7 17:03:18 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
44 * Did a little debugging, still trying to figure out why logging in
45 displays "Connection reset by peer" error. It used to work...
47 1998-12-07 Nathan Bryant <bryant@cs.usm.maine.edu>
48 * Makefile.in: updated dependencies
50 Sun Dec 6 19:50:55 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
51 * Fixed broken interpretation of variables specified in the URL
52 * added new "action" variable to make main loop simpler
54 1998-12-06 Nathan Bryant <bryant@cs.usm.maine.edu>
55 * webserver.c: prototype vsnprintf
56 * context_loop.c: warning fix
58 1998-12-05 Nathan Bryant <bryant@cs.usm.maine.edu>
59 * tcp_sockets.c: removed some dead code
60 * context_loop.c: oops, I made a memory leak. fix fix fix. also close
61 unused ends of pipes, fix a file descriptor leak and a race condition
62 * portability enhancements
64 1998-12-04 Nathan Bryant <bryant@cs.usm.maine.edu>
65 * webserver.c, context_loop.c, webcit.c: add commandline args for host
67 * converted to autoconf and began port to Digital UNIX
69 * context_loop.c: moved the monster 64k buffer array off the stack
70 * tcp_sockets.c: getservbyname() doesn't set errno.
72 1998-12-03 Nathan Bryant <bryant@cs.usm.maine.edu>
73 * webserver.c: warning fix