On Wed, Aug 7, 2013 at 9:13 PM, David Korn <[email protected]> wrote: > cc: [email protected] > Subject: Re: Re: [ast-developers] AT&T Software Technology ast alpha software > download update > -------- > > I deferred all the namespace changes to a later release. > > I addition to types in namespaces, there is also the question of > nested namespaces. When I added namespaces I was envisioning > a nested namespace model and namepace names would be identifiers. > > The alternative, which Roland sent a patch for allows does not allow > for nesting, but allows namespaces to be named by a compound variable. > > Thus, resarch.att.com could be the name of a namespace, but that does > not mean that research.att is a namespace and that of I am in > research.att.com, that if there is a namespace named research.att, > that it will be searched. I just means that variables not found > in research.att.com will be searched for in the global namespace.
That would be the java approach which is IMO more sane. Otherwise you get into a lot of trouble for deeper-nested names such as namespace com.ge.healthcare.project.ravenna.util.net.http.core ... > > The issue with namespace types that needs to be worked out is > how the namespace for type names will be handled. Currently gobal types > are kept in .sh.type.typename. I suspect that types defined in > namespace foo, should be kept in .foo.sh.type.typename but I haven't > modified the code for this. Why do you need code for this? For now, with Roland's patch, the code worked fine except for arithmetic expressions. That was AFAIK the major issue which prevented us to exploit the feature. That may be also the only bug which should be fixed as *soon* as possible because it will enable people to test the feature in-depth and contribute feedback. Irek _______________________________________________ ast-developers mailing list [email protected] http://lists.research.att.com/mailman/listinfo/ast-developers
