How to implement a system of ODEs in desolve with C compiled code and changing parameter size

Question

I am trying to implement a system of ODEs using the R package deSolve and compiled C code. I am using the recommendations provided with the following vignette https://cran.r-project.org/web/packages/deSolve/vignettes/compiledCode.pdf. In my model, I have one parameter for which I want to allocate the size dynamically. The question was almost addressed in a previous thread How to implement a system in deSolve, R, with N equations a N+m parameters?. More specifically I would like to declare the total size of the parameter vector as a parameter and not as a fixed value like it is done in the previous thread. Unfortunately, the compiler is not that happy with this option and I would like to know whether someone has any idea or hint on how to solve it.

For the example let me republish the one kindly published by @tpetzoldt in the previous thread and slightly modified to highlight my issue.

#include <R.h>
#define nA 6
#define nT 9

union parvec {
  struct {
    double r[3], a[nA];
  };
  double value[nT];
} p;

/* initializer  */
void initmod(void (* odeparms)(int *, double *))
{
    int N = nT; /* total number of parameters */
    odeparms(&N, p.value);
}

/* Derivatives */
void derivs (int *neq, double *t, double *y, double *ydot,
             double *yout, int *ip) {

  double y_sum = 0;
  for (int i = 0; i < *neq; i++) {
    y_sum = 0;
    for (int j = 0; j < *neq; j++) y_sum += p.a[i + *neq * j] * y[j];
    ydot[i] = p.r[i] * y[i] * (1 - y_sum);
  }
}

# file call_model.R
library(deSolve)

system("R CMD SHLIB model.c")
dyn.load("model.dll")

p <- c(r = c(0.1, 0.3, 0.04), A = c(0.2, 0.3, 0.3, 0.5, 0.4, 0.2))
y <- c(X = c(2, 2, 2))
times <- seq(0, 200, by = 0.1)

out <- ode(y, times, func = "derivs", parms = p,
           dllname = "model", initfunc = "initmod")
matplot.0D(out)

dyn.unload("model.dll")

I have looked at the vignette and specifically the section 3. Alternative way of passing parameters and data in compiled code but did succeed in solving my problem.