Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!samsung!emory!att!cbfsb!mark From: mark@cbfsb.att.com (Mark Horton) Newsgroups: comp.windows.open-look Subject: Re: ctrl h in text windows and text panels Keywords: Openlook Problems and Annoyances Message-ID: <1990Dec7.203935.20647@cbfsb.att.com> Date: 7 Dec 90 20:39:35 GMT References: <353@shuksan.UUCP> Organization: AT&T Bell Laboratories Lines: 81 In article <353@shuksan.UUCP>, scott@shuksan.UUCP (Scott Moody) writes: I've faced some of these. I called 800-USA-4-SUN and had good answers to most questions within a couple days. > Problem 1: > > I can't figure out how to get CTRL H to equal backspace in > text windows, or text panel items. If it is in the xmodmap stuff > I can't see it. ctrl-h works in shelltool since it is a regular > termcap. This one drove me nuts too. Turns out if you add Keyboard.DeleteChar: ^H to your .Xdefaults it starts to work almost everywhere. The exception I've found is in the header panel in mailtool, where ^H is still ignored. > Problem 2: > on a b/w sparcstation running openlook 2, inserting characters (in vi) > results in the character under the cursor that is moving to the > right disappearing (after about 20 characters). > I then have to do a refresh to get it back. > Unfortunately this doesn't happen all the time. I haven't found this, and I use vi pretty heavily. One thing I *have* found is that the default shell tool uses a 35x80 window, but all the termcaps in the world think sun and sun-cmd have 34 lines. If you rlogin to a machine running an older TCP/IP that won't pass the window size, vi messes up. I finally changed my .openwin-menu to explicitly say "-size 593 444" which forces a 34x80 window. You didn't say if you're rlogined to another host, or if you're in shell tool or xterm or what, and what your window size is. I seem to recall a bug in vi that messed up tab stops if the screen width (normally 80) wasn't a multiple of the tab stop (normally 8) with lines to long to fit on one physical screen line. > Major Annoyance: > Moving from kernal based sunview to network based xview, I notice lots > of lost events from the keyboard if the system is loaded. Like I start > a compile in one window, move to another and start typing (without > click to type). Three possible things occur: > (1) the typing goes to the old window > (2) the typing goes to the bitbucket > (3) the typing appears in new window - sometimes delayed. > I had heard this was a problem with X but for it to happen all the > time really sucks! We are getting faster machines but slower window > systems. I agree. I think what's happening here is that there is no serialization of keystrokes and mouse motions like sunview does. This can cause them to be interpreted in the wrong order. I find that when I move the mouse I have to wait for the cursor to go solid; if I don't my keystrokes will often go to the previous window. This can take awhile on a Sun 3 with 20-25 windows up on it, which is what I'm looking at right now. It's probably a lot better on a Sun 4. I do try to keep my windows down to a minimum, but at the moment I really need all these dials and gages. One thing that does seem to help is getting in the habit of letting the fileserver do the work and the workstation be an X terminal. I've put menu items like "X Term fsa..." exec rsh cbfsa DISPLAY=`hostname`:0 $OPENWINHOME/demo/xterm -title cbfsa -name cbfsa -bg yellow -ls -sb "FSA Mail Tool..." exec rsh cbfsa DISPLAY=`hostname`:0 $OPENWINHOME/bin/xview/mailtool" in my .openwin-menu; this lets the fileserver cbfsa do the work, and response is much faster. Now if only I could figure out how to make xterm give me a bigger font! (Also, can I change the label in an xterm window with an escape sequence?) > Annoyance 3: > the text panel items show little arrows on the left and right meaning > there are more characters in those directions. Unfortunately you have > to click the arrow one at a time! There is no hold arrow down and > scroll. I assume you're talking about mailtool. I haven't found a workaround for this one either. Mark