How does Stata compute R-square in an OLS with no constant and a dependent variable with no variation?

[Aditya Kashikar]

To replicate this, please see the code below:

Code:
clear
set seed 123
set obs 1000
gen x=rnormal()
gen one=1
reg one x, noconstant

I can understand the beta coeffients as the least squares estimate in this setup. But R-square has SS(Total) in the denominator, which is 0 in this setup. Then, what is stata outputting here?
There are some other related issues I have with this output but I think I can resolve them if I understand what Stata is doing to compute R-square.

[Joao Santos Silva]

Dear Aditya Kashikar,

Because the model does not include a constant, SS(Total) is 1000, not 0. Note that the R2 does not have the usual interpretation in models without a constant.

Best wishes,

Joao

[Aditya Kashikar]

Thank you Joao Santos Silva