"Arkadiy Vertleyb" <vertleyb@hotmail.com> writes:
"David Abrahams" <dave@boost-consulting.com> wrote
IOW, such templates in different translation units will be absolutely the same, although in each translation unit the compiler will have to instantiate different intermediate classes in order to produce them.
Right.
Do you still think this should be a problem?
"_Should_ be?" no, I never did. I just think it's technically a violation of the rules.
I have to admit that the solution with automatic ID generation would probably be non-conformant.
I guess the next question would be, if we require to manually provide a unique integer for each group of types/templates, how practical such system would be, especially considering the fact that an expression may contain types defined in multiple libraries, as well as user-defined types...
Are there any known examples of libraries that requre users to provide unique compile-time integers?
Can't we do something with a unique name per header file using typeof? #define BOOST_TYPEOF_HEADER boost_lambda_whatever We could use that to define a uniquely-named type and/or synthesize identifiers using __LINE__. -- Dave Abrahams Boost Consulting http://www.boost-consulting.com