On Fri, Dec 26, 2014 at 8:46 PM, Jonathan S. Shapiro <[email protected]> wrote:
> Given an object r of record type R, a field fld is accessible as r.fld. > The part that has always left me queasy is the promotion of fld into the > global namespace for construction purposes. I now understand that this is > limited to the context of record constructors, but it still violates my > sense of taste about such things. > I understand the issue now. I find the promotion is a common problem *especially* in constructors in Java/C#/C++ when member names get aliased by constructor or function arguments. We ended up with a generation of programmers "solving" the problem with naming conventions like "m_". I'd rather require "this.flr", or "self.flr", "
_______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
