This is how I remember how composite functions are defined.
'The range of the second is equal to or a subset of the domain of the first'.
I.e. If we had fog(x) then Ran g

dom f i.e. the range of g(x) fits inside the domain of f(x)
I.e. If we had gof(x) then Ran f

dom g i.e. the range of f(x) fits inside the domain of g(x)
For hog(x), the range of g(x) must be equal to or a subset of the domain of h(x).
So domain of h(x)=R\{0}
We need to restrict the domain of g(x) so that the range of g(x) is within/equal to R\{0}.
Currently the range of g(x) is
)
but we want it to be equal/a subest of R\{0}. Currently the range of g(x) is R
+, so we don't need to restrict it. Now the domain of hog(x) will be the domain of g(x). i.e. the domain of hog(x)=R
+So for goh(x), the range of h(x) must be equal to or within the domain of g(x)
So domain of g(x)=R
+We need to restrict the domain of h(x) so that the range of h(x) is within/equal to R
+.
Currently the range of h(x) is R
+, so we don't need to restrict it since it is equal to domain g(x)
Now the domain of goh(x) will be the domain of h(x). i.e. domain of goh(x)=R
+So for those two, you didn't need to restrict the domain because the range of the second was already equal/a subset of the domain of the first. Other cases you will need to restrict it.
After all that (sorry) I think the one bit of information that you needed was this: Once the domain has been restricted for the composite function, the domain of the composite function will be the domain of the second function.EDIT: To make it simpler to see, most people draw up a table like so
| Dom | Ran |
g(x) R
+
h(x) R\{0} R
+Then work out the retriction from there.