Date: Sun, 7 Sep 1997 14:56:29 -0700 From: Ken McGlothlen <mcglk@yang.cpac.washington.edu> To: freebsd-questions@freebsd.org Subject: Re: tcsh, stty, and other oddities. Message-ID: <199709072156.OAA07555@yang.cpac.washington.edu>
next in thread | raw e-mail | index | archive | help
"Jay D. Nelson" <jdn@qiv.com> writes: | You have gone further into termcap than I ever did. (Decoding/debugging | termcap or terminfo is roughly akin to doing your own root canal as a hobby.) Oh, termcap is *easy*. If you really want to do your own root-canals, sendmail.cf is where it's at. :) | When a rational value for $TERM is set on the remote login and it doesn't | work on your xterm or console you have two choices: 1) add the remote termcap | entry to your termcap -- or -- 2) add your terminal definition to termcap on | the remote. That seems fair, except that I hate mucking around with FreeBSD's termcap when it's not necessary. I could just replace FreeBSD's termcap with SunOS's termcap for vt100, etc., but the next time I install FreeBSD, I'm gonna have to incorporate that. I'd rather fix the problem the first time. Furthermore, I'm not at all sure that it's the termcap file that's trashing the system's interpretation of ^C, ^Z, and so on. | On some systems, tset actually works right so for those, add a call to tset | in your startup script. When I'm masquerading as a vt100, tset works correctly to set the terminal type as a vt100, so that's not necessarily the problem. | OpenLook doesn't use xterm from what I remember The SunOS system is using X, not OpenLook. xterm entries exist (and work properly) on the SunOS system I'm using. | so add xterm, vt??? and cons25* termcap entries to the termcap on the | Sun. But the Sun system *works just fine*. When I go to the FreeBSD system using the identical terminal setup on my client, the problems start. The SunOS system works fine. cons25 termcaps don't really come into this, because none of my telnet clients can masquerade as a cons25 terminal. | The strangeness is coming from programs on the Sun reading _Sun's_ antique | termcap and sending those values to your more evolved software. Your FreeBSD | system is doing exactly what it is told by the Sun. Uh . . . perhaps I haven't made myself clear. The problem happens when I'm telnetting to a FreeBSD system from a Macintosh running NCSA Telnet 2.7b4, from a PC/Win95 system running vanilla Telnet, and from a FreeBSD box running XF86. It *doesn't* happen *at* *all* when I telnet to a Sun using exactly the same clients. It *also* doesn't seem to happen when I telnet to a FreeBSD box using Win95 QVT/Net 4.1.4, for reasons I don't understand at all. The only thing that seems to fix the tab-completion problem is setting -oxtabs on the FreeBSD box, which doesn't even make sense. This will fix it on the Mac, PC/vanilla and X clients, but screws up tab output on some clients (they come across as ^I rather than a number of spaces). I don't *think* it's the termcap, but I don't know. Modifying Sun's termcap file is irrelevant in this case. I can't imagine why ^C becomes suspend, why ^Z becomes ineffectual, and why Emacs flips out, on top of the tcsh problems. The odd thing is that I know of other people who have had similar problems with my ISP, where the main user box is a FreeBSD system. Any other ideas? As usual, please reply to one of the following addresses. And thanks. ---Ken McGlothlen mcglk@serv.net mcglk@cpac.washington.edu
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709072156.OAA07555>