Release 0.1 (Sun Jul 23 17:04:53 CEST 2006)

  * preparation for release
  * enhanced escape sequence processing (tho' doesn't work properly
    yet)
  * removed optional <pfd> parameter to 'open' command 
  * threaded version seems to work! It can run ppp.
  * more conversion to pthreads.
  * partial conversion to using pthreads.
  * readline working better, but:
  * look at using pthread_sigmask so that when talking to the
    CLI (through the raw console PFD) then CLI can detect
    CTRL-C as an interrupt) (the console itself of course
    can't and shouldn't because it is raw)
  	pppd now not checked
	after returning from subshells terminal screwed (this
		is readline interaction, I'm sure)
  * added subshell command ('!')
  * added USE_READLINE symbol
  * plumbing good enough for pppd to work! 
  * beginnings of readline and history support
  * stty_post() and stty_raw() added. Right thing happens with
    terminal when switching console between control and shell.
  * stty_echo() now takes a PFD as parameter; this will later
    be enhanced to toggle rawness, echoing and NLCR mapping.
  * CLI performs max and min argc checking, relieving per-command
    handlers of the responsibility of checking
  * commands may be abbreviated, provided unique
  * 'list' commands displays flags (which will include rawness,
    echoing and NLCR mapping)
  * added display in debug mode of hexified read() and write()en
    strings (this was to help trace why yes was outputting one
    newline, but this was getting interpreted as two)
  * open command will now open shell by default but also other
    stuff.
  * implemented proper CLI with command lookup and delegation
  * added many more commands (desc, debug, close, disconnect)
  * connect now takes one or two args (one connects console to
    the other port)
  * centralised prompting into one function 
  * switch between raw and cooked as change in and out of control
    mode. Works!
  * by running the shell in a subshell rather than exec()ing it, 
    it gives the child a chance to call close() after, and altough
    this *still* doesn't trigger an EOF on the parent side of th
    pty, it is enough to ensure that select() reads all data
    *before* SIGCHLD arrives.
  * switched to bool instead of boolean which vim highlights better
  * some function and variable renamings 
  * balanced connects and disconnectes.
  * got it working again!
  * actually broken at the moment! subshell echo's are not seen by
    select())
  * integration with ADE C library (being much overhauled simultaenously)
  * corrections for strigent gcc warning options
  * started from scratch pretty much, using multiplexed IO now I worked
    out the the purpose of parent and child were input and output processing
    (resp?) and the grandchild did the slave side. Now I do parent does
    multiplexed IO processing and the child does the slave side.
  * stripped a lot out
  * corrected indentation
  * un-abstracted bodies of functions called only once back into
    their calling functions
  * changed name of program to ptyplumber and added some svn:ignores