I am currently using a case tool with a code generator to develop systems on a VAX. 90% of the source code that we will be placing into Perforce is generated automatically by the code generator from specifications maintained in the metadata (database). The ideal situation would be to check in the metadata files at the same time source code is checked in and have the system only keep the differences so that if one has to return to a prior version the metadata can be regenerated. Can Perforce meet this "ideal" need? If not, then I can maintain copies of the databases separately, but it gets really ugly.

Any comments or experiences you have in tracking changes to data files would be greatly appreciated.

Thanks.