[GenABEL-dev] [Genabel-commits] r664 - pkg/MixABEL/src/MXlib

[Yurii Aulchenko]

Hi William,

Thanks for fixing the code -- now running 'example(FastMixedModel)'
produces perfectly valid test stat values!

Looking at your commit diff at genabel-commits

http://lists.r-forge.r-project.org/pipermail/genabel-commits/2011-February/000047.html

I see a couple of changes, the first of which is

> -      //pchisq[it]=nulldev-ChildTwoVarCompModel.MinimiseDeviance(&altminimand);
> +      pchisq[it]=nulldev-ChildTwoVarCompModel.MinimiseDeviance(&altminimand);


looks like you are removing comment to a line, which does look like
chi-sq is computed at. So, this eventually is the fix to the problem,
right? Looks like an easy one (when you know what to look for :) )

At other place it is not too obvious to me what has happened -- it
looks like you changed the format, but the code itself stayed the same
(below). Correct?

>   if(ppRvPx_mat[0])
> -       gsl_matrix_free(ppRvPx_mat[0]);
> -      if(ppRvPx_mat[1])
> -       gsl_matrix_free(ppRvPx_mat[1]);
> -      if(ppPVPx_mat[0])
> -       gsl_matrix_free(ppPVPx_mat[0]);
> -      if(ppPVPx_mat[1])
> -       gsl_matrix_free(ppPVPx_mat[1]);
> -      if(ppxPVPx_mat[0])
> -       gsl_matrix_free(ppxPVPx_mat[0]);
> -      if(ppxPVPx_mat[1])
> -       gsl_matrix_free(ppxPVPx_mat[1]);
> -      if(ppxPVPy_vec[0])
> -       gsl_vector_free(ppxPVPy_vec[0]);
> -      if(ppxPVPy_vec[1])
> -       gsl_vector_free(ppxPVPy_vec[1]);
> +    gsl_matrix_free(ppRvPx_mat[0]);
> +  if(ppRvPx_mat[1])
> +    gsl_matrix_free(ppRvPx_mat[1]);
> +  if(ppPVPx_mat[0])
> +    gsl_matrix_free(ppPVPx_mat[0]);
> +  if(ppPVPx_mat[1])
> +    gsl_matrix_free(ppPVPx_mat[1]);
> +  if(ppxPVPx_mat[0])
> +    gsl_matrix_free(ppxPVPx_mat[0]);
> +  if(ppxPVPx_mat[1])
> +    gsl_matrix_free(ppxPVPx_mat[1]);
> +  if(ppxPVPy_vec[0])
> +    gsl_vector_free(ppxPVPy_vec[0]);
> +  if(ppxPVPy_vec[1])
> +    gsl_vector_free(ppxPVPy_vec[1]);
>
>  }