P1703R1 Recognizing header unit imports requires full preprocessing

Author: jensmaurer

source/basic.tex

@@ -963,8 +963,9 @@
 \end{note}
 
 \pnum
-\begin{note} For point of instantiation of a template,
-see~\ref{temp.point}.\end{note}%
+\begin{note}
+For point of instantiation of a template, see~\ref{temp.point}.
+\end{note}
 \indextext{scope!declarations and|)}
 
 \rSec2[basic.scope.block]{Block scope}
@@ -2590,8 +2591,7 @@
 
 \pnum
 A token sequence beginning with
-\opt{\tcode{export}} \tcode{module} or
-\opt{\tcode{export}} \tcode{import}
+\opt{\tcode{export}} \tcode{module}
 and not immediately followed by \tcode{::}
 is never interpreted as the \grammarterm{declaration}
 of a \grammarterm{top-level-declaration}.

source/lex.tex

@@ -273,6 +273,7 @@
 \begin{bnf}
 \nontermdef{preprocessing-token}\br
     header-name\br
+    import-keyword\br
     identifier\br
     pp-number\br
     character-literal\br
@@ -338,17 +339,13 @@
 except that a \grammarterm{header-name}\iref{lex.header} is only formed
 \begin{itemize}
 \item
-within a \tcode{\#include} directive\iref{cpp.include},
+after the \tcode{include} or \tcode{import} preprocessing token in an
+\tcode{\#include}\iref{cpp.include} or
+\tcode{import}\iref{cpp.import} directive, or
 
 \item
-within a \grammarterm{has-include-expression}, or
+within a \grammarterm{has-include-expression}.
 
-\item
-outside of any preprocessing directive,
-if applying phase 4 of translation to the sequence
-of preprocessing tokens produced thus far
-is valid and
-results in an \grammarterm{import-seq}\iref{cpp.module}.
 \end{itemize}
 \end{itemize}
 
@@ -359,6 +356,11 @@
 \end{codeblock}
 \end{example}
 
+\pnum
+The \grammarterm{import-keyword} is produced
+by processing an \tcode{import} directive\iref{cpp.import} and
+has no associated grammar productions.
+
 \pnum
 \begin{example} The program fragment \tcode{0xe+foo} is parsed as a
 preprocessing number token (one that is not a valid floating or integer

source/modules.tex

@@ -411,9 +411,9 @@
 
 \begin{bnf}
 \nontermdef{module-import-declaration}\br
-    \opt{\keyword{export}} \keyword{import} module-name \opt{attribute-specifier-seq} \terminal{;}\br
-    \opt{\keyword{export}} \keyword{import} module-partition \opt{attribute-specifier-seq} \terminal{;}\br
-    \opt{\keyword{export}} \keyword{import} header-name \opt{attribute-specifier-seq} \terminal{;}
+    \opt{\keyword{export}} import-keyword module-name \opt{attribute-specifier-seq} \terminal{;}\br
+    \opt{\keyword{export}} import-keyword module-partition \opt{attribute-specifier-seq} \terminal{;}\br
+    \opt{\keyword{export}} import-keyword header-name \opt{attribute-specifier-seq} \terminal{;}
 \end{bnf}
 
 \pnum
@@ -478,7 +478,7 @@
 a \grammarterm{header-name} is also recognized by the
 preprocessor, and results in macros defined at the
 end of phase 4 of translation of the header unit
-being made visible as described in \ref{cpp.module}.
+being made visible as described in \ref{cpp.import}.
 \end{note}
 A declaration of a name with internal linkage is
 permitted within a header unit despite all

source/preprocessor.tex

@@ -11,10 +11,14 @@
 \pnum
 A \defn{preprocessing directive} consists of a sequence of preprocessing tokens
 that satisfies the following constraints:
-The first token in the sequence is a
-\tcode{\#}
-preprocessing token that (at the start of translation phase 4)
-either is the first character in the source file
+The first token in the sequence,
+referred to as a \defnadj{directive-introducing}{token},
+is a \tcode{\#} preprocessing token,
+an \tcode{import} preprocessing token, or
+an \tcode{export} preprocessing token
+immediately followed by an \tcode{import} preprocessing token,
+that (at the start of translation phase 4)
+either begins with the first character in the source file
 (optionally after white space containing no new-line characters)
 or follows white space containing at least one new-line character.
 The last token in the sequence is the first new-line character
@@ -50,6 +54,7 @@
 \begin{bnf}\obeyspaces
 \nontermdef{control-line}\br
     \terminal{\# include} pp-tokens new-line\br
+    \opt{\terminal{export}} \terminal{import} pp-tokens new-line\br
     \terminal{\# define } identifier replacement-list new-line\br
     \terminal{\# define } identifier lparen \opt{identifier-list} \terminal{)} replacement-list new-line\br
     \terminal{\# define } identifier lparen \terminal{... )} replacement-list new-line\br
@@ -149,9 +154,8 @@
 The only white-space characters that shall appear
 between preprocessing tokens
 within a preprocessing directive
-(from just after the introducing
-\tcode{\#}
-preprocessing token through just before the terminating new-line character)
+(from just after the directive-introducing token
+through just before the terminating new-line character)
 are space and horizontal-tab
 (including spaces that have replaced comments
 or possibly other white-space characters
@@ -243,7 +247,7 @@
 if the identifier is currently defined
 as a macro name
 (that is, if it is predefined
-or if it has one or more active macro definitions\iref{cpp.module},
+or if it has one or more active macro definitions\iref{cpp.import},
 for example because
 it has been the subject of a
 \tcode{\#define}
@@ -623,11 +627,11 @@
 denotes an importable header\iref{module.import},
 it is
 \impldef{whether source file inclusion of importable header
-is replaced with \tcode{import}}
-whether the preprocessing directive
-is instead replaced by the \grammarterm{preprocessing-token}{s}
+is replaced with \tcode{import} directive}
+whether the \tcode{\#include} preprocessing directive
+is instead replaced by an \tcode{import} directive\iref{cpp.import} of the form
 \begin{ncbnf}
-\terminal{import} header-name \terminal{;}
+\terminal{import} header-name \terminal{;} new-line
 \end{ncbnf}
 
 \pnum
@@ -665,104 +669,39 @@
 \end{codeblock}
 \end{example}
 
-\rSec1[cpp.glob.frag]{Global module fragment}
-
-\begin{bnf}
-\nontermdef{pp-global-module-fragment}\br
-    \terminal{module} \terminal{;} pp-balanced-token-seq \terminal{module}
-\end{bnf}
-
-\pnum
-If the first two preprocessing tokens at the start of phase 4 of translation
-are \tcode{module} \tcode{;},
-the result of preprocessing shall begin with
-a \grammarterm{pp-global-module-fragment}
-for which all \grammarterm{preprocessing-token}{s}
-in the \grammarterm{pp-balanced-token-seq}
-were produced directly or indirectly by source file inclusion\iref{cpp.include},
-and for which the second \tcode{module} \grammarterm{preprocessing-token}
-was not produced by source file inclusion or
-macro replacement\iref{cpp.replace}.
-Otherwise,
-the first two preprocessing tokens at the end of phase 4 of translation
-shall not be \tcode{module} \tcode{;}.
-
-\rSec1[cpp.module]{Header units}
+\rSec1[cpp.import]{Header unit importation}
 \indextext{header unit!preprocessing}%
 \indextext{macro import|(}%
 
-\begin{bnf}
-\nontermdef{import-seq}\br
-    \opt{top-level-token-seq} \opt{\terminal{export}} \terminal{import}
-\end{bnf}
-
-\begin{bnf}
-\nontermdef{top-level-token-seq}\br
-    \descr{any \grammarterm{pp-balanced-token-seq} ending in \terminal{;} or \terminal{\}}}
-\end{bnf}
-
 \begin{bnf}
 \nontermdef{pp-import}\br
-    \terminal{import} header-name \opt{pp-import-suffix} \terminal{;}\br
-    \terminal{import} header-name-tokens \opt{pp-import-suffix} \terminal{;}
-\end{bnf}
-
-\begin{bnf}
-\nontermdef{pp-import-suffix}\br
-    pp-import-suffix-token\br
-    pp-import-suffix pp-import-suffix-token
-\end{bnf}
-
-\begin{bnf}
-\nontermdef{pp-import-suffix-token}\br
-    \descr{any \grammarterm{pp-balanced-token} other than \terminal{;}}
-\end{bnf}
-
-\begin{bnf}
-\nontermdef{pp-balanced-token-seq}\br
-    pp-balanced-token\br
-    pp-balanced-token-seq pp-balanced-token
-\end{bnf}
-
-\begin{bnf}
-\nontermdef{pp-balanced-token}\br
-    pp-ldelim \opt{pp-balanced-token-seq} pp-rdelim\br
-    \descr{any \grammarterm{preprocessing-token} other than a \grammarterm{pp-ldelim} or \grammarterm{pp-rdelim}}
-\end{bnf}
-
-\begin{bnf}
-\nontermdef{pp-ldelim} \descr{one of}\br
-  \terminal{(    [    \{    <:    <\%}
-\end{bnf}
-
-\begin{bnf}
-\nontermdef{pp-rdelim} \descr{one of}\br
-  \terminal{)    ]    \}    :>    \%>}
+    \opt{\terminal{export}} \terminal{import} header-name \opt{pp-tokens} \terminal{;} new-line\br
+    \opt{\terminal{export}} \terminal{import} header-name-tokens \opt{pp-tokens} \terminal{;} new-line\br
+    \opt{\terminal{export}} \terminal{import} pp-tokens \terminal{;} new-line
 \end{bnf}
 
 \pnum
-A sequence of \grammarterm{preprocessing-token}{s}
-matching the form of a \grammarterm{pp-import}
+The preprocessing tokens after the \tcode{import} preprocessing token
+in the \tcode{import} \grammarterm{control-line}
+are processed just as in normal text
+(i.e., each identifier currently defined as a macro name
+is replaced by its replacement list of preprocessing tokens).
+An \tcode{import} directive
+matching the first two forms of a \grammarterm{pp-import}
 instructs the preprocessor to import macros
 from the header unit\iref{module.import}
 denoted by the \grammarterm{header-name}.
-A \grammarterm{pp-import} is only recognized
-when the sequence of tokens
-produced by phase 4 of translation
-up to the \tcode{import} token
-forms an \grammarterm{import-seq},
-and the \tcode{import} token is not within
-the \grammarterm{header-name-tokens} or \grammarterm{pp-import-suffix}
-of another \grammarterm{pp-import}.
-The \tcode{;} \grammarterm{preprocessing-token}
-terminating a \grammarterm{pp-import}
-shall not have been produced by
-macro replacement\iref{cpp.replace}.
-The \defnadj{point of}{macro import} for a \grammarterm{pp-import} is
-immediately after the \tcode{;} terminating the \grammarterm{pp-import}.
+The \defnadj{point of}{macro import} for the
+first two forms of \grammarterm{pp-import} is
+immediately after the \grammarterm{new-line} terminating
+the \grammarterm{pp-import}.
+The last form of \grammarterm{pp-import} is only considered
+if the first two forms did not match.
 
 \pnum
-In the second form of \grammarterm{pp-import},
+In all three forms of \grammarterm{pp-import},
+the \tcode{import} token is replaced by the \grammarterm{import-keyword} token.
+Additionally, in the second form of \grammarterm{pp-import},
 a \grammarterm{header-name} token is formed as if
 the \grammarterm{header-name-tokens}
 were the \grammarterm{pp-tokens} of a \tcode{\#include} directive.
@@ -847,6 +786,50 @@
 
 \indextext{macro import|)}%
 
+\rSec1[cpp.glob.frag]{Global module fragment}
+
+\begin{bnf}
+\nontermdef{pp-global-module-fragment}\br
+    \terminal{module} \terminal{;} pp-balanced-token-seq \terminal{module}
+\end{bnf}
+\begin{bnf}
+
+\nontermdef{pp-balanced-token-seq}\br
+    pp-balanced-token\br
+    pp-balanced-token-seq pp-balanced-token
+\end{bnf}
+
+\begin{bnf}
+\nontermdef{pp-balanced-token}\br
+    pp-ldelim \opt{pp-balanced-token-seq} pp-rdelim\br
+    \descr{any \grammarterm{preprocessing-token} other than a \grammarterm{pp-ldelim} or \grammarterm{pp-rdelim}}
+\end{bnf}
+
+\begin{bnf}
+\nontermdef{pp-ldelim} \descr{one of}\br
+  \terminal{(    [    \{    <:    <\%}
+\end{bnf}
+
+\begin{bnf}
+\nontermdef{pp-rdelim} \descr{one of}\br
+  \terminal{)    ]    \}    :>    \%>}
+\end{bnf}
+
+\pnum
+If the first two preprocessing tokens at the start of phase 4 of translation
+are \tcode{module} \tcode{;},
+the result of preprocessing shall begin with
+a \grammarterm{pp-global-module-fragment}
+for which all \grammarterm{preprocessing-token}{s}
+in the \grammarterm{pp-balanced-token-seq}
+were produced directly or indirectly by source file inclusion\iref{cpp.include},
+and for which the second \tcode{module} \grammarterm{preprocessing-token}
+was not produced by source file inclusion or
+macro replacement\iref{cpp.replace}.
+Otherwise,
+the first two preprocessing tokens at the end of phase 4 of translation
+shall not be \tcode{module} \tcode{;}.
+
 \rSec1[cpp.replace]{Macro replacement}%
 \indextext{macro!replacement|(}%
 \indextext{replacement!macro|see{macro, replacement}}%