AVR-Chat

--- In AVR-Chat@yahoogroups.com, "Paul Sinclair" <paul@...> wrote:

> I'm really not familiar with the
> debugging of these chips "on a
> chip with a debug interface"

Maybe you've used a hosted software development environment, like 
Visual Studio or Elipse.  Working on a chip with a JTAG interface is 
very much like this, except that LEDs flash and motors spin.  You 
make a change to the code.  You punch the "build" button and the new 
code is compiled.  You set some breakpoints by right-clicking on a 
line of code.  You punch run.  There's a pause while the binary file 
is transferred to the target microcontroller, then off we go.  When 
the breakpoint is hit, execution stops and you can bring up various 
views of memory, registers, variables, ports, etc. to see what's 
going on.  Single step, step into functions, step out of functions 
and run again. I haven't used debugWIRE as much, but it's pretty much 
the same if a bit slower.

On the other hand, without a debugger, when you've re-built the 
program you have to pop up a separate tool to take the binary file 
and transfer it to the target.  When that's done, you disconnect from 
the target and reset it.  Then things either work or they don't.  If 
they don't, you have to figure out why either using pure brain power 
or with the aid of extra bits of code inserted to make the invisible 
visible.  For each trial, it's build, load, reset and ponder all over 
again.

Don't get me wrong, a debugger isn't a panacea. There are problems, 
particularly in hard real-time systems, that aren't helped a whole 
lot by a debugger.  But, I've done projects with them and projects 
without and by-and-large I much prefer to work with.

Graham.