I now have several perl modules manipulating one or another SET of Berkeley DB databases.

To change the SET being manipulated at a particular time has meant modifying the top level code in each module to point to a different path ... (simple enough I guess, as long as one remembers to modify ALL the modules in unison).

Better to centralise this, I thought, with a head module reading the path that is desired right now from a simple file entry, setting up the Berkeley Environment with that, and Exporting it.

Now the modules which import (use) those exported variables compile ok, but the BerkeleyDB module chokes at runtime on various error conditions. A typical complaint is that some LITERAL-BAREWORD-CONSTANT which may be a flag to a sub-object "is not numerical" ... but there are others too. So I've had to go back and re-paste the same path and environment setup code into each module separately in order to make them work again.

The code-repitition seems like it should be unnecessary (and potentially dangerous!) but I am wondering what I can do about it?

Any ideas appreciated?

Xstack