I'm currently using Test Director as my repository for all loadrunner scripts/scenarios/results data. However, the integration is pathetically slow, error prone, and just plain inneficient in other ways as well. For example, if I need to send a script to a customer to help him test our software, I can't just bundle up the script from TD. I have to first find where TD put it, copy it elsewhere, etc etc.

I've considered separating LR from TD, but I don't know a good way to replace the positive aspects of integration. I'm hoping for suggestions or models that others use.

Some questions I'm considering are:
Do you have a share that you save everything to? How do you link specific scripts/scenarios to different projects? How do you maintain version control for different releases? Is there a way to look at a set of results and easily tell what runs have been performed, and when they were done?

I've actually considered using our CVS repository, and checking all scripts in with the build, but this obviously wouldn't work for results, and seems like it would be problematic for scenarios as well.

Please reply with the model your company uses, and the advantages and disadvantages of it.