[lex.phases] Move dropping whitespace to end of phase 4

Phase 7 deals only with tokens, not whitespace.  For clarity
move the dropping of whitespace to the end of phase 4, after
preprocessing directives are deleted.  Newline characters are
significant in the grammar of preprocessing directives to
whitespace characters are clearly signicant until this point.
Conversely, it is good to recognise that string literal
preprocessing tokens are adjacted regardless of whitespace
in phases 5 and 6.

Author: AlisdairM

source/lex.tex

@@ -165,6 +165,7 @@
 expanded\iref{cpp.replace}, and \tcode{_Pragma} unary operator expressions are executed\iref{cpp.pragma.op}.
 A \tcode{\#include} preprocessing directive\iref{cpp.include} causes the named header or
 source file to be processed from phase 1 through phase 4, recursively.
+Whitespace characters separating preprocessing tokens are no longer significant.
 All preprocessing directives are then deleted.
 
 \item
@@ -180,7 +181,6 @@
 
 \item
 Each preprocessing token is converted into a token\iref{lex.token}.
-Whitespace characters separating tokens are no longer significant.
 The resulting tokens constitute a \defn{translation unit} and
 are syntactically and
 semantically analyzed as a \grammarterm{translation-unit}\iref{basic.link} and