# Use the improper prior $p(v) \propto 1/v$ into Jags

I know that one can approximate this density ($p(v) \propto 1/v$) using its truncated version and implement it this way:

   B~dunif(log(BInf),log(BSup))
v<-exp(B)


but I would like to use the exact form (I checked that under this prior my posterior is proper). Is there any solution to achieve this ?