probit transform integer return code?

[hkershaw-brown]

why use an integer return code? For different types of failures?
The oks are logicals.
logical, allocatable :: obs_probit_trans_ok(:), state_probit_trans_ok(:)

Originally posted by @hkershaw-brown in #794 (comment)

[hkershaw-brown]

Comment on lines +509 to +510

This is related to the question about the error code being an integer, if you had the error be a logical you could do:
call transform_to_probit(........., state_probit_trans_ok(i))

that way you there is less chance of a future person, putting something in the middle of these two statements.

call transform_to_probit(....., ierr)
call something_else_mathsie(.....,ierr)
state_probit_trans_ok(i) = (ierr == 0)

[hkershaw-brown]

#794 (comment)

The added if(obs_probit_trans_ok(i)) & at line 511 gives you the same answer as before, because of the bnrh

fail -> ! Just return the original ensemble
I guess this is for algorithm readability?