You're definitely right...sorry about that. In this case the zener is a horrible addition, but you can still move the cap behind the fixed resistor. That will provide a minimum of protection against an overcharged cap discharging directly into the laser diode.
The answers to your questions are yes and no.
1. The LM317 tries to output a voltage on the Out pin that makes the voltage on the Adj pin equal to 1.25V. With the Adj pin connected directly to the laser the laser will never see more than 1.25V unless the LM317 is broken. So, using the LM317...
This LM317 circuit is a great current regulator for diodes, but I would make 2 small changes that will solve the disconnected diode problem.
1. Replace the silicon diode with a zener diode (Digi-Key - MAZ40240LFCT-ND (Manufacturer - MAZ40240LF)). This will clamp the output voltage to no more...