However, it is not true that testing is necessary for a deterrent to be maintained. The deterrence results from an enemy's fear of a nuclear counterattack. Only if the enemy is sure that the nuclear weapons arrayed against it will not work will the deterrence effect be lifted. It does not rely on the level of assurance that the deterring nation has in its stockpile.
As nuclear weapons degrade, what becomes uncertain is the magnitude of the explosive force, not whether there will be fission and fusion. To test whether chain reaction fission will take place, one only need test the chemical explosion and prefission radiation injection stages. Along with measurements of the amount of fissile material, this is sufficient to determine whether there will be a nuclear explosion.
Such tests are non-nuclear and can assure the testor that the nuclear weapons will still maintain their deterrent capability.
Testing is more often done for increasing or maintaining war-fighting instead of war-deterring capabilities. The testing of new weapon designs falls into this category, as does the exact measurment of the forces released from a nuclear explosion. As the only legitimate reasons for maintaining nuclear arsenals is deterrence (according to international treaties), such purposes are invalid.
Thus there is no possible military benefit to France for restarting testing.
Politically, such testing would be disasterous. The world is now trying to limit the number of nuclear states, trying to persuade non-nuclear countries (N. Korea, Iran, Iraq, Libya, ...) not to develop and test warheads. If the nuclear powers go ahead and test, especially if they are involved in vertical proliferation, the moral strength of arguments against horizontal proliferation evaporates. In this scheme of things, if you don't wish N. Korea or Iraq to test nuclear weapons, you should similarly oppose French testing.