Path: utzoo!utgpu!watmath!uunet!tut.cis.ohio-state.edu!UUNET.UU.NET!mcvax!litp!rms
From: mcvax!litp!rms@UUNET.UU.NET (Richard M. STALLMAN)
Newsgroups: gnu.emacs.bug
Subject: Problem with X11 startup
Message-ID: <8903111642.AA12144@litp.univ-p6-7.fr>
Date: 11 Mar 89 16:42:17 GMT
Sender: daemon@tut.cis.ohio-state.edu
Reply-To: rms@wheaties.ai.mit.edu
Distribution: gnu
Organization: GNUs Not Usenet
Lines: 27


	 We have found locally that we need the following modification to
    lisp/term/x-win.el in order to get GNU emacs to start up properly with
    X11.  This is true on Sequent Symmetries, Sun3s, and Sun4s.

I tested Emacs 18.53 with X11 on a Sun 3 before release.
It seemed to work properly.

The change you propose only appears to work if you don't check it
completely.  It introduces a bug, that C-g fails to interrupt
a program.

The fact that you proposed this change suggests that you are using
a system which has bugs in the support for SIGIO.

If I could program Emacs to work correctly on such systems I would do
so.  But I believe this is impossible; certainly no one has told me
any way to do it.  I will not install a half-measure in Emacs which
fixes the problem only in certain cases; if I did so, then users would
think that Emacs is simply unreliable.  I would rather distribute an
Emacs that fails completely on systems with bugs than distribute one
which will only fail occasionally.

I know that some Unix system users would prefer I took a different
attitude, but remember that use on Unix is a side-benefit as far as I
am concerned.  The GNU kernel, when we have one, will not have this
bug.