Closed andrjohns closed 1 year ago
@hsbadr
The 2.26 version of stanc has a bug where int[,] arguments are not treated as valid for lpmf UDFs, causing parse errors like the following:
int[,]
lpmf
> testcode <- " + functions { + real Dirichlet_Multinomial_lpmf(int[,] X, vector eta) { + return 0; + } + } + " > rstan::stanc(model_code = testcode) Error in rstan::stanc(model_code = testcode) : 0 Semantic error in 'string', line 2, column 2 to line 4, column 3: Probability mass functions require integer variates (first argument). Instead found type int[,].
This PR updates the stanc.js with a patch for allowing nested arrays up to 3 dims
This is needed for the MADPop reverse dependency with 2.26.
MADPop
@hsbadr
The 2.26 version of stanc has a bug where
int[,]
arguments are not treated as valid forlpmf
UDFs, causing parse errors like the following:This PR updates the stanc.js with a patch for allowing nested arrays up to 3 dims
This is needed for the
MADPop
reverse dependency with 2.26.