123456789101112131415161718192021222324252627282930313233343536 |
- Features/Changes which will break the API and so must go into v7
- ----------------------------------------------------------------
- - Do not pack the od_control structure in memory. This is plain silly.
- - Do not use the BOOL type in od_control or in any of the public functions.
- Using it causes problems with stuff that defines a BOOL as a different
- size. For example, xpdev uses int as a BOOL whereas OpenDoors uses a char.
- - Support telnet IAC escaping, add an item to the od_control struct to
- keep track of the current status of IAC escapes.
- - Fix up the local display for Win32. It's horribly slow currently.
- Possibly just dust of the DOS stuff and use ciolib?
- - Make output functions take an optional position and attribute to cut down
- on small send()s.
- - Audit for small send()s in general.
- - Possibly improve the output thread and make *nix multi-threaded for block
- buffering ala Synchronet.
- - Change to 24-line default. This means fixing the personalities, and the
- screen stuff.
- - Autodetect if passed handle is a socket (Like I did for MyCroft doors)
- - We need some non-socket way of dealing with I/O for Win32. Something that
- operates along the lines of stdio on *nix. When BBSs commonly start doing
- SSH, is the door supposed to do SSH also? A local socket would work, but it
- seems like the Wrong Thing. The main reason stdio is out is that there's no
- standard way of passing hangup info etc.
- - Check the timestamp on the dropfile... do not use an old one! Add and
- od_control variable to prevent this behaviour.
- - Add a bunch of status-line specific functions to allow new personalities:
- od_status_set_cursor()
- od_status_set_attrib()
- od_status_puttext()
- od_status_gettext()
- od_status_printf()
- od_status_disp_str()
- od_status_putch()
- od_status_addkey()
- od_status_delkey()
- - OD_DISP_EMU() has problems with ANSI music (of course)
|