I am trying to live with the new string class. Looking at Plauger's Pre-San
Jose notes and the Borland 4.0 implementation, I am somewhat surprised that
NPOS seems to be a part of the global name space rather than the namespace
of string. Is that the final word? Is it a feature that dynarray and string
share that value???

Cay
horstman@cs.sjsu.edu

In article <1994Jan12.180911.27689@sjsumcs.sjsu.edu>, horstman@sjsumcs.sjsu.edu (Cay Horstmann) writes:
> I am trying to live with the new string class. Looking at Plauger's Pre-San
> Jose notes and the Borland 4.0 implementation, I am somewhat surprised that
> NPOS seems to be a part of the global name space rather than the namespace
> of string. Is that the final word? Is it a feature that dynarray and string
> share that value???

The standard header <defines> contains a number of common things
used throughout the library, including the definition

 const size_t NPOS = (size_t)(-1);

This is described in section 17.3.1 of Plauger's most recent library
draft.  Such definitions will ultimately be in the namespace for the
standard library ("iso_standard_library"), not global.

--Larry