[R] GLM using truncated lognormal distribution

[Marc Girondot]

Dear honorable list-members,

I know how to fit a truncated lognormal distribution (or Gaussian) 
(example here:
http://max2.ese.u-psud.fr/epc/conservation/Girondot/Publications/Blog_r/Entrees/2012/5/24_Adjust_a_truncated_lognormal_distribution.html 
) but I would like to use it in the context of glm.

Rather than using family=gaussian(), ideally I would like to have a 
family=truncated_gaussian().
I see using fix(gaussian) how is organized the gaussian() function. It 
is not 100% clear now but I think I could manage to change it to do a 
family=truncated_gaussian().

But before to do it, perhaps it exists already.

I find the package truncnorm but it does not do this function.

Thanks a lot for any advice,

Marc