2000-08-11-11:59:47 Steve Simmons:
> There's a strong implication here that we've just made package
> namespaces heirarchical.
The possibility of doing so was raised during discussion of ways to
handle needs for one running program to have two different versions
of the same support module.
Once the possibility was raised, discussion began about possible
problems. One problem proposed was the occasional need for really
truly module global flag variables, and I proposed that in the face
of heirarchical package namespaces, a path that explicitly roots
itself with main:: might be a working global available identically
to all packages.
> Modules (packages) should have the option of declaring themselves
> to be either rooted or relative in the name space.
That provides only weak support for the problem that launched this
thread; it means that the heirarchical package namespace hack will
only solve the need for multiple versions of the same package to
co-exist, if they decline to force themselves to be global. Having
this choice at the package level is far less helpful for the
motivating problem than having it at the individual variable level.
-Bennett
PGP signature