debugging with mkII PDI in Studio 6

I just tried out your scenario. A couple of things - the PDI is not a hardware debugger. Atmel offeres a ICE but it's $425 and slow and molasses on - well - ICE. I used to have a JTAG connector on the board but I took it off in v7.

But all is not lost. I do almost all my debugging in the simulator2. It has some pros and cons. It's hard to emulate real-time actions like serial IO or limit switches. But it's portable because it doesn't require any HW and I use it on the road all the time. Another down-side is that it's slow to emulate very long sequences.

I have some tools int eh code to help with this. In tinyg.h there are some defines that take out a lot of the startup stuff like messages and NVM writes. There is also a way to feed a command to the serial buffer on startup. See test.c

I like the simulator on Studio4 better than studio6. There are some workarounds I documented here:

https://github.com/synthetos/TinyG/wiki/Developer-Notes

Lastly, if you are going to play with the code I recommend using the current push in edge (build 367.04). It's a release candidate for version 0.95

Alden

One last thing - I fired up the debug in Studio6 and it worked - it didn.t fail like it did for you. You may want to grab the Studio6 project files from the /support directory in the edge and see if you get better results

I was able to compile and start a debugging session in Simulator 2. This is normally what happens - I have not experienced the failure you noted. Setting up an existing project in 6 is a bear. See the developer notes link from a few posts back.

I use 4 for simulation as the interface is actually better (less broken). The compiler and libs used in 6 generate better code, though. Also, I almost always simulate in -O0 optimization as that way you see every instruction and declared variable instead of jumping around and having to look at the register set to see what's going on. I takes longer to run, though as there are more instructions to execute.

Alden