If I had to guess I would presume that the idea is that
the Dummy Adapter is only really useful for debugging/diagnostic purposes
so it's not enabled by default and if you hit a problem
that needs it then you configure and build OpenOCD yourself with it enabled?
I have had issues in the past with Tcl and general questions about OpenOCD
which did not depend on a particular debug adapter, and I am guessing that most
users may want to report and/or help pin-point problems without needing to
download the OpenOCD source code and its dependencies and learning how to build
it from scratch, which is not straightforward. That is probably the reason why
the BUGS file suggests trying to reproduce an eventual problem with the dummy
adapter.
Most packagers will not be aware of that indication, so they probably wouldn't
manually enable the dummy adapter on first thought. I also think that the dummy
adapter will be very lightweight indeed.
Therefore, I would turn on support for the dummy adapter in the configure
script
by default, even if that means compiling support for 'bitbang' by default too.
[...]
I'm curious about your mention of "bitbang" above -
what specifically are you referring to there?
This is what I mean in file configure.ac:
AS_IF([test "x$ADAPTER_VAR([dummy])" = "xyes"], [
build_bitbang=yes
])
Therefore, I would turn on support for the dummy adapter
I guess that you could submit a patch to do this and
it would get discussed and reviewed as per the usual change processes?
I have submitted patches before. Would you be willing to review such a patch
then? No point doing the work if nobody here thinks it is a good idea after all.
Regards,
rdiez
