McAllister@DOCKMASTER.ARPA (08/03/88)
Date: Wednesday, 3 August 1988 0500 edt From: Richard McAllister<McAllister@DOCKMASTER.ARPA> Subject: TIC/Interrupts and ProDOS Reference responses by David A. Lyons and Don Elton on 01 Aug 88 via INFO-APPLE@BRL.ARPA (now BRL.MIL) David Lyons did an excellent job of analyzing the handling of interrupts by the ProDOS MLI. Indeed the system global page change occurred in going from PRODOS 8 version 1.1.1 to version 1.2. I apologize to Don Elton for the incorrect attribution. TIC uses ProDOS standardly. The IIc System Clock, distributed by Applied Engineering, is a product of Creative Peripherals Unlimited (CPU). It attaches to either serial port of the IIc and is stimulated by sending a specific pattern via the command register (I suspect on the DTR pin). It responds with data on the DSR pin which must be read by carefully timed machine language loops with each status check providing one data bit. It reads 40 bits representing day of week, month, day of month, hours, minutes, seconds. I have not sorted how the clock is set. The year must be managed by software. Any interruption of the timing will result in bad data and the potential for interference with another device on the same port is obvious. What is not obvious is is why modem data is garbled when the clock is not on the same port. David suggests that the solution is to not read the clock while in terminal mode. ASCII EXPRESS using XON/XOFF operates that way and can time-date stamp ProDOS files. Both TIC and ProTERM periodically read the clock and modem data gets garbled. Until a better solution is found that's the best suggestion. David, can you pinpoint where interrupts are re-enabled after the JMP $DE00 ? I am sincerely thankful for your interest and effort.