For the serializing/deserializing effort the initial idea was to use the project Version Release Modification (VRM) to identify what version the serialized file is at. This would work fine for the actual releases of the project. However, since we frequently request user's to get the latest code from CVS we have a problem as during the development cycle there could be numerous changes to the serialization and we would still be using the same VRM identifier. One possible alternative to this is to use VRM + Serialization Level. The "Serialization Level" would be a #define value that needs to be updated every time a change is made to serialization. The serialization level could be stored in XercesVersion.hpp. Does this sound like a reasonable approach? Hopefully there isn't too much overhead with having to increment the Serialization Level when checking code in that has serialization changes. To keep track of this item I was considering creating a list of coding reminders that should be reviewed before checking code in (for example, make sure code uses new memory management, etc). Thanks in advance for your feedback. Regards, David A. Cargill --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
