wrong higher-order function error msg from parser

[yizhang-yiz]

Summary:

Wrong error message with non-existent higher-order function.

Description:

If a higher-order function does not exist, parser error message points to its functor argument instead of the higher-order function itself.

Reproducible Steps:

The following variant of ODE example from the manual, replacing integrate_ode_bdf with non-existent foobar

functions {
real[] sho(real t,        // time
           real[] y,      // state
           real[] theta,  // parameters
           real[] x_r,    // data (real)
           int[] x_i) {   // data (integer)
  real dydt[2];
  dydt[1] = y[2];
  dydt[2] = -y[1] - theta[1] * y[2];
   return dydt;
}
  
}

data {
  int<lower=1> T;
  real y[T*2];
  real t0;
  real ts[T];
}
transformed data {
  real x_r[0];
  int x_i[0]; }
parameters {
  real y0[2];
  vector<lower=0>[2] sigma;
  real theta[1];
}
model {
  real y_hat[T,2];
  sigma ~ cauchy(0, 2.5);
  theta ~ normal(0, 1);
  y0 ~ normal(0, 1);
  y_hat = foobar(sho, y0, t0, ts, theta, x_r,
                             x_i);
  for (t in 1:T)
    y[t] ~ normal(y_hat[t][1], sigma);
  for (t in 1:T)
    y[t+T] ~ normal(y_hat[t][2], sigma);
}

Current Output:

variable "sho" does not exist.

Expected Output:

Function foobar not found.

Additional Information:

Provide any additional information here.

Current Version:

v2.17.1

[VMatthijs]

stanc3 currently says

Semantic error at file "test.stan", line 34, characters 10-80:
   -------------------------------------------------
    32:    theta ~ normal(0, 1);
    33:    y0 ~ normal(0, 1);
    34:    y_hat = foobar(sho, y0, t0, ts, theta, x_r,
                   ^
    35:                               x_i);
    36:    for (t in 1:T)
   -------------------------------------------------

A returning function was expected but an undeclared identifier foobar was supplied.

Is that any better?

[bob-carpenter]

Yes, this is better---we should be treating the higher-order functions as functions. In the existing parser, they're hacked special expressions, so require all their own custom warning messages. I think "returning function" is confusing. I just think we should say "A function was expected" or that "'foobar' is an unknown function". If they supply a function without a return value, then we could say something about expecting a return value.

…

On Dec 13, 2018, at 8:48 AM, Matthijs Vákár ***@***.***> wrote: stanc3 currently says Semantic error at file "test.stan", line 34, characters 10-80: ------------------------------------------------- 32: theta ~ normal(0, 1); 33: y0 ~ normal(0, 1); 34: y_hat = foobar(sho, y0, t0, ts, theta, x_r, ^ 35: x_i); 36: for (t in 1:T) ------------------------------------------------- A returning function was expected but an undeclared identifier foobar was supplied. Is that any better? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.