> Van: Nigel Orr- Underwater Acoustics Group > Aan: PICLIST@MITVMA.MIT.EDU > Onderwerp: Re: Another MPLAB USART question! > Datum: donderdag 24 september 1998 16:38 [Snip] > >They have "solved" that by inventing a method to inject byte-data into the > >program. I do not use the simulator myself (95% implementation of the > >controller ? you can bet on it that all the important features are in the > >last 5%), but it seems to be possible to create a file that contains data > >about where, when & what to inject. It's not very well documented, but > >that's the way to go. > > I know about that, and it really doesn't bother me that I have to write a > file to manipulate port pins. What is difficult is trying to simulate a > USART when it appears that it is not simulated at all- interrupts are not > generated and registers are not handled correctly- do you know of a > solution to that? Is it my mistake? No, It's not your mistake at all. It's MicroChip's mistake to believe that a kludgy solution like writing (with external means / not launchable from within the Development environment) a "injection" file would appeal to it's customers. And as I said, the important features are allways in the last unimplemented few percents. (as is apperantly the USART) Do you ever think about testing a serial protocol (like RS232 or I2C) by changing bits on a I/O port ? Forget it. You can do it, but only one bit at a time (changing for example DTR & DSR together is not possible ..). Another thing is that the injection of this bit-stream can only be triggered by reaching a certain adres (like a breakpoint) in the program. It _can not_ be triggered by any kind of user intervention (like pressing a button) or other events. So, easy testing of non (program) syngronous signals is not possible. As I see it a program that should _Emulate/Simulate_ something but implements not all features is worth less than a simple disassembler with single-step capabilities. All other behavour can be tested in the windowed version ... Greetz, Rudy Wieser