usenet@scubed.UUCP (USENET News System) (05/11/89)
I am currently an X-10 release 4 user, running under BSD Unix v4.3 on a Sun 3/50. Approximately once a week I encounter a very annoying problem. When in the "vi" editor, some magic (and as yet, undetermined) combination of keystrokes toggles the caps lock on. Following that, every keystroke is displayed and interpreted as an uppercase character -- in or out of the "vi" editor and in any window. This is effectively a "crash" since I can do no more work; for all practical purposes all "vi" commands and Unix commands are lowercase. It is sufficient to exit X-Windows and fire X up again in order to correct the problem. It would, therefore, seem to be a problem with X and not my workstation. If there is a combination of keytrokes which puts me into caps lock mode, I assume there must be another combination (uppercase, hopefully) which will reset the session to lowercase. Presumably, any necessary "command" to fix this problem could be aliased to an uppercase command anyhow. Can anyone help me? Thanks, Tom Rankin
wdc@ATHENA.MIT.EDU (Bill Cattey) (05/11/89)
I got bit by this problem today with my Xibm server. We were able to telnet in, and after poking at the led settings with xset, the problem went away. Try: xset -led 2 (which is what worked for us.) I have no clue as to what causes this. -wdc
tsf@PROOF.ERGO.CS.CMU.EDU (Timothy Freeman) (05/15/89)
Well, when this arises on my Sun with X10R4, I fix it by pressing both shift keys down at once. There is an analogous problem with a virtual "meta-lock" becoming stuck down, which can be fixed by pressing both meta keys down at once. The problem is that sometimes the server doesn't hear the up transition on a shift key. So the server thinks there is one shift key down, but there are really zero. When this problem happens and you depress both shift keys, the count of the number of shift keys down goes to 3. The server is smart enough to know that there are really only two shift keys, so it reduces the count to 2 in this special case, and then when you lift the shift keys, the count goes back down to 0 like it should. I don't remember where I heard of this, it may have been on the sun-spots mailing list. I think code for the special case in the server that decrements the count from 3 to 2 has comments that explain the phenomenon. This is a common problem. Spread the word (but not by posting to comp.windows.x). -- Tim Freeman Arpanet: tsf@cs.cmu.edu Uucp: ...!seismo.css.gov!cs.cmu.edu!tsf --