Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!ZURICH.AI.MIT.EDU!cph From: cph@ZURICH.AI.MIT.EDU (Chris Hanson) Newsgroups: gnu.emacs.bug Subject: (none) Message-ID: <8908181759.AA13573@zurich.ai.mit.edu> Date: 18 Aug 89 17:59:08 GMT Sender: daemon@tut.cis.ohio-state.edu Distribution: gnu Organization: GNUs Not Usenet Lines: 60 Here's a couple of messages that should be of interest. From cph Fri Aug 18 13:36:45 1989 To: rjh@ai.mit.edu CC: hp-bugs@ai.mit.edu, system-hackers@ai.mit.edu, rjh@ai.mit.edu In-reply-to: Robert J. Hall's message of Fri, 18 Aug 89 13:05:26 edt <8908181705.AA23199@praz.ai.mit.edu> Subject: clisp-mode broken on HPs From: rjh@ai.mit.edu (Robert J. Hall) Date: Fri, 18 Aug 89 13:05:26 edt CLISP mode in gnu emacs used to work, but now something appears screwed up in the inter-process communication. Doing c-C c-E to the following form results in some erroneous error message. It looks like only so many characters are getting sent to the child process, then it starts interpreting the remainder as new input without processing any of the first characters. (setq barf '(a a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as a a a a a a a a a a a a a a aa a a as)) Delete one of the `a' lines from the above and it works. NOTE that it seems unlikely that this will fix itself when we get the new version of Lucid, as it seems to be an emacs problem. It appears to have started with the big changeover. -- Bob I wouldn't be surprised if this happens on the Suns also. We had a similar problem with the Scheme subprocess -- and the culprit turned out to be that nearly every version of unix had bugs in their PTY implementation. The fix we used for Scheme was to bind `process-connection-type' to nil when the subprocess was started, thus forcing the use of pipes for communication. Return-Path:From: rjh@ai.mit.edu (Robert J. Hall) Date: Fri, 18 Aug 89 13:58:08 edt To: cph In-Reply-To: Chris Hanson's message of Fri, 18 Aug 89 13:36:49 edt <8908181736.AA13561@zurich.ai.mit.edu> Subject: clisp-mode broken on HPs Your fix to clisp mode worked great. Thanks. -- Bob