Re: ONLINE Key problems

In article <47v5r6$4kk@arcadia.informatik.uni-muenchen.de> spitz@GANS2X.ana.med.uni-muenchen.de (Richard Spitz) writes: >From: spitz@GANS2X.ana.med.uni-muenchen.de (Richard Spitz) >Subject: Re: ONLINE Key problems >Date: 10 Nov 1995 09:25:58 GMT >Rick.Walters@ccmail.irs.gov wrote: >: Has anyone out there encountered a problem with the INFORMIX-OnLine 5.02 >: Engine running on a Pyramid where the arrow Keys are accepted >: intermittently after pushing them more than a few times? >Your problem sounds like it might be related to the time interval >between the "Escape" character that is the first char. of the multi- >char keycode sent by your cursor keys. >Since "Escape" is also the default "Accept key" in Informix 4GL (maybe >also in ISQL, never tried this). So the Informix system must be able >to distinguish whether the user pressed the "Escape" key or a cursor >key, and it does this by waiting for a predefined time interval for >additional characters to follow the "Escape" code. If there is no >additional input within this interval, "Escape" is assumed, otherwise, >a cursor or other control key is assumed. >If you have a slow network connection or a slow terminal line, or the >kernel is too busy and cannot pick up the characters fast enough from >the interface, then the interval is exceeded and "normal" key input >is assumed by Informix. That's why you are getting strange input >behavior. >There is an undocumented environment variable "DBESCWT", which sets the >waiting interval to the number of seconds assigned to it. "DBESCWT=1" >sets it to 1 second etc. Try if this helps you. I agree except I'd try setting it to 2 (I think the default is 1 anyway). The explanation for this is that the system call, alarm() which is often used to do the timing gives an interval between n-1 and n seconds. If you use n = 1 this means that the timing interval may approximate to zero! With DBESCWT you're guaranteed at least 1 second. I have had this symptom from s different cause. I started off thinking it must be DBESCWT but even really silly settings didn't help. I drove myself and tech support nuts (yeah, I know, I havn't recovered) before I realised the problem only happened between log-in and my first use of vi. I then discovered my xterm vt100 emulation was set to keypad mode (which sends different character strings for various keys) on log-in. Further investigation showed it was a call to elm -z in the log-in script. The cure was sending a magic incantation with tput. It might be worth checking your emulation mode. Ian