Dave Clark writes: > So, given that the other macros need to validate that the bit mask > flag was defined correctly, what would you say is the best method for > accomplishing that? ... a custom data type code different from the > IBM-defined standard ones? ...or, what?
I would be personally be quite hesitant about whether such a sophisticated solution is really required in this case. The HLASM internal flag definitions do not use a macro, and the macros to test, set and clear a flag are effectively one line each. Once a programmer understands the convention, it is all very simple. However, I would agree that a custom data type might be a suitable solution, for example a lower case letter. Normally, the type attribute has the advantage that forward references to definitions in open code will be resolved automatically by lookahead processing. However, if the flag is being generated by a macro, this advantage of the type attribute will be lost. Jonathan Scott, HLASM IBM Hursley, UK
