ed@sfucmpt.UUCP (05/16/84)
Greetings: I have a bunch of emacs users here who have a problem sending an EOT (^D to the rest of you) to a subprocess, such as a shell, since there appears to be a deficiency (bug) in the pseudo-tty stuff in Unix 4.2. Exactly, a write of a zero length record on the master pty does not cause an EOT to be seen on the slave pty (which is connected to a csh, typically). For further info, see TIOCREMOTE in the manual entry for pty(4). The fact that it is documented in pty(4) doesn't help me out much. Is there a hack to emacs, the pty driver, etc which will solve this problem. Thanks in advance, Ed Bryant Systems Programmer Lab for Computer and Communications Research Simon Fraser University (604)291-4430 / (604)291-4704 ... uw-beaver!ubc-vision!sfucmpt!ed
chris@umcp-cs.UUCP (05/20/84)
Don't use TIOCREMOTE. Instead, write the current EOF character to the PTY. (Of course there may not *be* a current EOF character, but the alternative is to fix TIOCREMOTE....) -- In-Real-Life: Chris Torek, Univ of MD Comp Sci (301) 454-7690 UUCP: {seismo,allegra,brl-bmd}!umcp-cs!chris CSNet: chris@umcp-cs ARPA: chris@maryland