See also: ideas.txt.

FujiChat features, we're almost at parity!

- Configurable ctcp version response. Does it matter?
- Keyboard buffer, if possible. Right now we miss the occasional
  keystroke when typing fast & furious during network I/O. Not
  sure how to accomplish this with FujiNet (the old FujiChat
  key buffer actually hooked into the SLIP code). Use a VBI I
  guess. If that doesn't do it, it would require a custom SIO
  replacement... well, I have the OS source, so maybe.

Other stuff:

- Implemented, but not sure it was worth doing:
  If an VBI-driven keyboard buffer turns out to be impossible or
  too much of a PITA to contemplate, another idea to avoid dropping
  keystrokes: after every keypress, wait something like 1/4 or 1/3
  second for another keypress, before checking for incoming net
  traffic. As long as the user's typing 4 (or 3) keys per second,
  he won't lose keystrokes. Probably have to put a limit on it,
  a pathological case would be holding down a key long enough to
  fill the edit box with 240 of the same character. That would
  cause a 60 sec delay in net traffic, which might actually cause
  us to time out... need some data on just how fast a fast typist
  really types, and on average, how often they pause (to think or
  read back what they just wrote, etc). Also look into detecting
  key repeats (SRTIMER).
  Remove this when/if I do a proper typeahead buffer.
- Write a cgetc() replacement that doesn't call the OS K: "get one byte"
  routine. I was avoiding it because it will need a 192-byte table
  (keycode -> atascii lookup)... but I'm spending more than 192 bytes
  of code on keycode-filtering, and it's not even complete.
- Filter out the rest of the keystrokes that causes cgetc() to block.
  Includes ctrl-/, ctrl-8, ctrl-9, maybe others (The_Doctor__).
- Load/save config files to N:SD///.FujiNetChat or such. Since we *have*
  to have a FujiNet anyway, might as well make better use of it.
- Rewrite the incoming message parser! It needs to work more like
  the command parser in cmd.c: know how many arguments to expect, and
  not blindly assume they're present. Also, replace strtok() and
  strstr() with nextarg().
- Make scr_activate() *not* set scr_status[scr_active]. Also make
  printing to a screen aware of the proper SCR_* status to set.
  This will matter even more once there's another status (4th statusbar
  color).
- Status bar needs another color (COLOR2). It will mean "new non-chat
  text" (e.g. join/quit/part, numerics). The existing SCR_ACTIVE (COLOR1)
  will mean *only* new text (that doesn't hilite you), and COLOR3 (red)
  will stay the same as it now.
- Filtered-out keystrokes (ctrl-3 and friends) shouldn't even bring up
  the editbox, if it's not already showing.
- Auto-pinging the server seems to work, but needs more testing.
- *Thoroughly* test the nick and channel tab completion for the [private]
  and [server] screens.
- Fix the nick completion in channel screens. It works, but needs some
  polishing up.
- More scrollback. Of course it needs more memory... see doc/scrollback.txt
  for my ideas on this.
- ^U works, but is slow (calls backspace() in a loop, which does a
  memmove() each time).
- Server /ping command is iffy (see do_server_pong() in irc.c).
- "User has kicked <yourname> from #channel", the name should be
  replaced by "you".
- Gracefully handle nicks/channels whose lengths are stupid long.
  At least we shouldn't overflow any buffers.
- At least one keyboard macro (for ChanServ auth). More would
  be nice, if we can afford the RAM.
- Error numerics should go to the current screen (?).
- Add an optional key parameter to /join (key). spec calls for it,
  I've never seen it used.
- 'Dead' screens (channels we've parted, or failed to join) should
  show some kind of indicator, and not accept input. Although maybe
  this isn't worth doing (you already get "cannot send to channel").
- *Possibly* save the config from within the client... though not
  all options can be changed (really only /nick and /beep). To
  make this useful would probably bloat the code too much. Some
  way to change the IRC server/port within the app would be very
  useful though.
- Input box history. Up-arrow to recall previous command. Unfortunately
  this needs a lot of RAM. Maybe just one previous command, so you
  can up-arrow to re-ping someone, etc.

Config file and initial config:

- The config tool should be rewritten in asm, for size. It doesn't
  need to be fast, but it needs to *load* fast.
- The config UI should be nicer. Dialog/curses style, arrows to move, and
  hotkeys like [S]erver, [N]ick. Categories (as tabs? tree?). I'm not
  going to get this right without some user feedback, since I'm mainly
  a CLI and text editor guy.
- There should be a preset list of servers to choose from, or the
  user can enter his own.
- There should be a file selector for load/save config.
More prefs:
- CTCP VERSION response? Do we need it?
- Ignore list (maybe).
- Connect macro (log in to bot).

Remember, the config has to be <= 512 bytes!