4.1BSD ^Z/tty mode problem

[Jay Lepreau]

1. Yes, this problem exists on both 4.1 and 4.2 and I've seen it (in "ded").
2. Your analysis is pretty much what I came up with (but not quite so
clearly!).
3. I wouldn't really call it a kernel problem, though, because:
4. I solved it in my application by disabling t_*suspc when in a funny
tty mode in the child (B).  This allows the child to read the ^Z itself,
reset its tty modes, and then issue the TSTP to the pgrp, avoiding
having the kernel issue the TSTP to the whole pgrp, with the attendant
problems. (Note that this problem doesn't occur in raw mode for the same
reason my solution works, so all those raw mode screen applications
don't see it.  Must be cbreak.)  Anyone have other ways?

Jay Lepreau, lepreau at utah-cs, {ihnp4,decvax}!utah-cs!lepreau