> I know about MAP.
> Who however can knows, how it is possible to compare of the 
> same name sections in libraries?

As I said earlier, this is possible for MACLIBs because they are printable
human-readable text. For TXTLIBs or LOADLIBs, the only comparison possible
using the native tools is a bit-by-bit comparison of the contents of the
members, which isn't really very useful. 

> Can there are any programs or procedures, which allow it to make?
> For example, how to check by means of loading in memory of 
> modules from different libraries.
> Or, to define(determine) address of a beginning of sections 
> in libraries and their length, and then as to check.
> Whether there are any ideas, as it to make how to write such program?

It might help to step back and tell us what the bigger problem you're trying
to solve is. I'm guessing that you're trying to find out what version of a
library is being used for applications, and remove some duplicate copies. Is
that guess correct? 

In the case of TXTLIBs, address symbol resolution and such are resolved at
the time you LOAD the module into storage, so there really isn't any such
usage information *in* the module other than the external symbol table,
which doesn't tell you much more than what the MAP option does on the *LIB
commands. LOADLIBs are similar in behavior.

You could write a modification to the CMS LOAD command that recorded the
information it used when linking the application, but that's a fairly major
project, and won't catch already-compiled modules - you'd have to rebuild
all your MODULE files for it to be effective. You'd also have to get deep
into the OS simulation stuff to catch things used there. 

I think CMAP (a commercial product) used to do this sort of thing, but
VM/ESA is no longer supported by most ISVs, so I doubt it is still
available. 

Reply via email to