The precedent case for this law happened during the cold war, and was about a newspaper that published a counter-espionage report (Case: NJA 1988 s. 118²).
The military argued that the national safety was damaged by the publishing, but would not give any details because of the nature of doing national security work.
The court made two tests to decide the case. Did the information deserve to be classified under secrecy, and was the damages reported by the military believable. On the first, they said it was doubtful, and on the second they said a straight no. Case dismissed.
To me it sounds as very clear and public requirement, which would work fine in both Snowden's and Manning's cases.
1) Please note that this is only about publishing secrets. Breaking a contract or an oath can and is likely to be punished under different laws.
But regardless, that is a decent enough requirement system. The military refusing to present evidence in the case also hurt their argument, which is how it should be.
Anyway, the point is that if I don't know what Snowden leaked, I bet most people don't know either.
However, while Mr Snowden's conduct was admirable, there will be some people out there interested in state secrets, with the intent of malice. With the intent of doing harm to the government, the state, people working for the government/state or the people of that state.
I agree that secrecy is bad, but I feel we must at least have some secrecy, but clear and public requirements for when something can become (or remain) secret.