Chris, see below.
> > > My guess is: > > > > jtag.py leaves the reset line (or another one crucial for > > starting) applied, if it succeeds. If there is an error, > > the reset (or other) line is released, which starts the MSP. > > (or something alike) > > nope. > the cpu runs, but its not correctly reset to start the user app. > at least with my version of the library Not true with my version. As I stated in an earlier mail today, I checked what happens on the reset pin and proved my findings (somehow): If this error msg occures, the reset line will be released and the cpu starts. With no error, the reset line stays low... I started looking at the jtag.py source code and found that the call to _parjtag.release() (done by jtag.close()) is somehow responsible. At least, if I delete the regarding line, every works fine (i. e. reconnect is possible). But regardless of calling jtag.close() or not: if I start insight after jtag.py I have no connect problem. By the way... This leads me to another issue with jtag.py/proxy/insight. See corresponding thread. Cheers, Ralf
