[BUG] Can't store densities of simple math distributions in a deterministic node

[davidcerny]

Describe the bug
For simple math distributions like dnExp and dnNorm, the output of the .probability() and .lnProbability() methods can't be stored in a deterministic node. Interestingly, it is possible to do this for more complex distributions like dnBDP.

To reproduce

> seed(12345)
> a ~ dnExp(1.0)
> b := a.probability()
> c := a.lnProbability()
> a
   0.07298359
> a.probability()
   0.9296161
> b
   0.9296161
> a.lnProbability()
   -0.07298359
> c
   -0.07298359
> a ~ dnExp(1.0)
> a
   0.11636
> a.probability()
   0.8901547
> b
   0.9296161
> a.lnProbability()
   -0.11636
> c
   -0.07298359

Expected behavior
In the code above, the output of a.probability() and b (as well as of a.lnProbability() and c) should be identical after each redraw.

Computer info
macOS 12.7.1, RevBayes 1.2.2, latest version of the development branch (rapture-2386-g54c6ba).