NA/NaN/Inf in foreign function call (arg 6) error from coxph function

The NaN/Inf message is almost certainly from a singular fit.
  I would be interested in a copy of the data set to test it: data sets that cause this 
are rare, and help me to tune the convergence/signularity criteria in the program.

  Note: it is much easier to use coxph(survobj ~ therapy + ReceptorA + ReceptorB, 
data=sample.data) than to put "sample.data$" in front of every variable name; and easier 
to read as well.

Terry Therneau (author of coxph function)

I am trying to fit a multivariate Cox proportional hazards model,
modelling survival outcome as a function of treatment and receptor
status. The data look like below: