When priors are used on a particular parameter, a rescaling (by setScale) wrecks the fit completely, because suddenly the priors will pull toward a different goal in natural units (value*scale).
Of course the user can handle this by changing the priors manually, but it would be much safer if it was done internally, so that mean is multiplied by oldscale/newscale and sigma by abs(oldscale/newscale) as soon as setScale is applied.
When priors are used on a particular parameter, a rescaling (by setScale) wrecks the fit completely, because suddenly the priors will pull toward a different goal in natural units (value*scale). Of course the user can handle this by changing the priors manually, but it would be much safer if it was done internally, so that mean is multiplied by oldscale/newscale and sigma by abs(oldscale/newscale) as soon as setScale is applied.