Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!elroy!mahendo!jplgodo!wlbr!pete From: pete@wlbr.EATON.COM (Pete Lyall) Newsgroups: comp.sys.m6809 Subject: Re: CoCo smart terminal emulators Message-ID: <12404@wlbr.EATON.COM> Date: 25 Apr 88 17:22:10 GMT References: <1547@vaxb.calgary.UUCP> <717@svax.cs.cornell.edu> <833@ndsuvax.UUCP> <5009@pucc.Princeton.EDU> Reply-To: pete@wlbr.UUCP (0000-Pete Lyall) Organization: Eaton IMS, Westlake Village, CA Lines: 40 In article <5009@pucc.Princeton.EDU> EWTILENI@pucc.Princeton.EDU writes: >>My second complaint is also with the keyboard driver. Has anyone figured >>out how to fix the shift key rollover problem? > >This is a hardware problem, as each of the CoCo's keys in the keyboard does >NOT have it's own input/diode/whatever. There was some discussion of this >on the COCO list (why aren't you on it?), and it's not an OS-9 related >problem - it occurs in RSDOS as well and is due to hardware. Hmmmm. Now mind you I spend little or no time under RSDOS per se, but my experience seems to conflict with yours. I have had no trouble whatsoever with RSDOS keyboard handling (nothing worse than the original COLOR BASIC 1.2 roms provided with their half-assed quick key-down check). The new code also patches out a key-down check in the disk roms when it effects the overlay, but again, this is noting serious. I looked rather closely at CC3IO and noted that there were two things that could be done to improve the keyboard handling. It has been a while, but I think one of them is testing to see if last key combination was SHIFT/CTRL + current key, and also to see if all keys have been up since that time. If not, ignore current key. They also need to add code that clears the last-key-down variable when the keyboard is scanned and no keys are pressed. This would save a lot of headaches. Also - the hardware is a SUPERSET of what was in the CC2... the only difference I have seen (in the schematic) is that in addition to all lines of the keyboard matrix being tied to PIA0 ports A and B, is that there is a large OR gate reading all the inbound keystrobe lines to be able to generate an interrupt on keypress. Perhaps you had better research the problem a little more the next time before offering a 'diagnosis'. -- Pete Lyall (OS9 Users Group VP)| DELPHI: OS9UGVP | Eaton Corp.(818)-706-5693 Compuserve: 76703,4230 (OS9 Sysop) OS9 (home): (805)-985-0632 (24hr./1200 baud) Internet: pete@wlbr.eaton.com UUCP: {ihnp4,scgvax,jplgodo,voder}!wlbr!pete