FW: OT: broken/useless ansi - console driver??

[Jean-Pierre A. Radley]

John Esak propounded (on Tue, Oct 25, 2005 at 07:39:28PM -0400):
| I think you've lost me on this one... I have no idea what you are talking
| about. Production machine??  This is a developer who is writing new code on
| a development machine... and various connected via tcp/ip outlying clients.
| How do you develop if not by compiling and re-compiling and trying the new
| binary... over and over until it all works?  It's the way I do it. I don't
| know any other way. :-)

Obviously you can't tell if something works without running it.

But testing version N+1 of something where version N is in constant use
never requires that you overwrite, remove, destroy or cripple version N.

Beginning users of Unix are taught, very early on, that you will hang
from your own petard by typing 'rm -r' in any directory, worst case
being doing this as root while in /.

Unix does not protect you from that and other destructive commands
(like, say, 'chmod 0 *').  So you learn, "don't do that!".

Some do 'touch /-i' to guard against 'rm *'; it works up to a point.
Others alias rm (in shells which have aliases) to do something other
than just plain run /bin/rm.

Now suppose the 'text file busy' error had never been in effect.  I
postulate that early in one's Unix career, one would have been taught,
right along with the lesson about 'rm *', to "Do not copy a program on
top of one which may be in use, lest its current users see really nasty
effects".

So: as the doctor said, if it hurts to keep hitting your head against
the fireplace mantle, don't do it!

After you're satisfied with your testing of version N+1, use the
mv-before-cp commands I mentioned earlier to install version N+1.

-- 
JP