Enhanced Machine Controller

lerman's change does fix the test case, and I've committed it to the CVS
version of emc. The fix will be in the next release of emc, whether it's
2.1.8 or 2.2.0.

I've added a test case to the CVS version of emc, and next i'll try
applying the change that lerman suggests.

Sorry about that. I've identified the problem and the cure (I think).

In file rs274ngc/interp_o_word.cc around line 492, change:
case O_return:
if(settings->call_level == 0)
{
// this is in the definition
break;
}

// if we were skipping, no longer
settings->skipping_o = 0;

// in a call -- must do a return
// restore old values of parameters
// restore file position from context

settings->call_level--;

for(i=0; i<INTERP_SUB_PARAMS; i++)
{
settings->parameters[i] =
settings->sub_context[settings->call_level].saved_params[i];
}

To:
case O_return:
if(settings->call_level == 0)
{
// this is in the definition
break;
}

// if we were skipping, no longer
settings->skipping_o = 0;

// in a call -- must do a return
// restore old values of parameters
// restore file position from context

settings->call_level--;

for(i=0; i<INTERP_SUB_PARAMS; i++)
{
settings->parameters[i+INTERP_FIRST_SUBROUTINE_PARAM] =
settings->sub_context[settings->call_level].saved_params[i];
}
[Unfortunately, I don't have a machine currently containing EMC2, so I
can't try this out. I also don't know how current the source I'm looking at
is, so the line number might be off. I expect to have a new machine on line
in a week or two. (lerman)]

Thank you for reporting the bug.