Provide informative error message with mismatch in data=list(...)

Summary:

On my system, if data provided through the rstan::stan(..., data=list(...) ) interface do not match the required types in various ways, I get the inscrutible error

trying deprecated constructor; please alert package maintainer
Error in new_CppObject_xp(fields$.module, fields$.pointer, ...) : 
  no valid constructor available for the argument list
failed to create the sampler; sampling not done

This took me a while to discover the true source of the error (the mismatch was not obvious). Best would be if checking was done that could say in which argument the mismatch appears; but it would at least be much better if the error message said something like:

  No valid constructor available for the argument list: are the provided data of the correct form?

Description:

For example, the following code:

scode <- "
data {
 real x[1];
}
parameters {
 real y[1];
} 
model {
 y[1] ~ normal(x[1], 1);
} 
"
fit1 <- stan(model_code = scode, data = list(x=1.0), iter = 10, verbose = FALSE, chains=1) 

results in the error above. Changing to data=list(x=matrix(1.0)) does not work either, but as @bgoodri pointed out below, data=list(x=array(1.0,1)) does work: obvious in retrospect but not at the time.

The same error occurs in other simple situations, like:

Length mismatch:

scode0 <- "
data {
 vector[4] x;
}
parameters {
 real y[4];
} 
model {
 y ~ normal(x, 1);
}
"
fit0 <- stan(model_code = scode0, data = list(x=rep(1.0,3)), iter = 10, verbose = verbose, chains=1)  

Missing variable:

scode4 <- "
data {
 real x[1];
 int K;
}
parameters {
 real y[K];
} 
model {
 y[1] ~ normal(x[1], 1);
}
"
fit4 <- stan(model_code = scode4, data = list(x=1.0), iter = 10, verbose = verbose, chains=1)  

R and RStan Version:

> packageVersion("rstan")
[1] ‘2.17.2’
> R.version.string
[1] "R version 3.4.3 (2017-11-30)"

Operating System:

Debian sid, updated.

Start at the rstan::stan(..., data=list(...)) entry point and reproduce the provided examples for type, length, and missing-variable mismatches. Trace how these inputs reach the constructor error, then make the reported failure identify the data mismatch or argument involved and verify the examples produce informative messages.