C++ PATCH for c++/91338 - P1161R3: Deprecate a[b,c]

[Marek Polacek <polacek at redhat dot com>]

This patch implements P1161R3, Deprecate uses of the comma operator in
subscripting expressions:
<http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1161r3.html>
which made its way to C++20.  New [depr.comma.subscript] shows:

  void f(int *a, int b, int c) {
    a[b,c];                     // deprecated
    a[(b,c)];                   // OK
  }

I've added a new option, -Wcomma-subscript, so that this warning can be
selectively turned off in c++2a, where it is enabled by default, and turned
on in lower standards, to check if your codebase has any occurrences of it,
without needing to change the -std=.

As for implementation, I've extended cp_parser_skip_to_closing_square_bracket_1
so that we can check for a non-nested token inside [ ].

Bootstrapped/regtested on x86_64-linux, ok for trunk?

2019-08-03  Marek Polacek  <polacek@redhat.com>

	PR c++/91338 - Implement P1161R3: Deprecate a[b,c].
	* c-opts.c (c_common_post_options): Enable -Wcomma-subscript by
	default for C++2a.
	* c.opt (Wcomma-subscript): New warning.

	* parser.c (cp_parser_postfix_open_square_expression): Warn about uses
	of a comma operator within a subscripting expression.
	(cp_parser_skip_to_closing_square_bracket_1): New function, made out
	of...
	(cp_parser_skip_to_closing_square_bracket): ...this.

	* doc/invoke.texi: Document -Wcomma-subscript.

	* g++.dg/cpp2a/comma1.C: New test.
	* g++.dg/cpp2a/comma2.C: New test.
	* g++.dg/cpp2a/comma3.C: New test.

diff --git gcc/c-family/c-opts.c gcc/c-family/c-opts.c
index e97bbdf5c6f..e43172598f3 100644
--- gcc/c-family/c-opts.c
+++ gcc/c-family/c-opts.c
@@ -916,6 +916,10 @@ c_common_post_options (const char **pfilename)
   if (!global_options_set.x_warn_register)
     warn_register = cxx_dialect >= cxx17;
 
+  /* -Wcomma-subscript is enabled by default in C++20.  */
+  if (!global_options_set.x_warn_comma_subscript)
+    warn_comma_subscript = cxx_dialect >= cxx2a;
+
   /* Declone C++ 'structors if -Os.  */
   if (flag_declone_ctor_dtor == -1)
     flag_declone_ctor_dtor = optimize_size;
diff --git gcc/c-family/c.opt gcc/c-family/c.opt
index 4c8b0026000..257cadfa5f1 100644
--- gcc/c-family/c.opt
+++ gcc/c-family/c.opt
@@ -428,6 +428,10 @@ Wclobbered
 C ObjC C++ ObjC++ Var(warn_clobbered) Warning EnabledBy(Wextra)
 Warn about variables that might be changed by \"longjmp\" or \"vfork\".
 
+Wcomma-subscript
+C++ ObjC++ Var(warn_comma_subscript) Warning
+Warn about uses of a comma operator within a subscripting expression.
+
 Wcomment
 C ObjC C++ ObjC++ CPP(warn_comments) CppReason(CPP_W_COMMENTS) Var(cpp_warn_comment) Init(0) Warning LangEnabledBy(C ObjC C++ ObjC++,Wall)
 Warn about possibly nested block comments, and C++ comments spanning more than one physical line.
diff --git gcc/cp/parser.c gcc/cp/parser.c
index ebeffdb775f..1a5ae147b84 100644
--- gcc/cp/parser.c
+++ gcc/cp/parser.c
@@ -2669,6 +2669,8 @@ static bool cp_parser_init_statement_p
   (cp_parser *);
 static bool cp_parser_skip_to_closing_square_bracket
   (cp_parser *);
+static int cp_parser_skip_to_closing_square_bracket_1
+  (cp_parser *, enum cpp_ttype);
 
 /* Concept-related syntactic transformations */
 
@@ -7522,7 +7524,33 @@ cp_parser_postfix_open_square_expression (cp_parser *parser,
 	  index = cp_parser_braced_list (parser, &expr_nonconst_p);
 	}
       else
-	index = cp_parser_expression (parser);
+	{
+	  /* [depr.comma.subscript]: A comma expression appearing as
+	     the expr-or-braced-init-list of a subscripting expression
+	     is deprecated.  A parenthesized comma expression is not
+	     deprecated.  */
+	  if (warn_comma_subscript)
+	    {
+	      /* Save tokens so that we can put them back.  */
+	      cp_lexer_save_tokens (parser->lexer);
+
+	      /* Look for ',' that is not nested in () or {}.  */
+	      if (cp_parser_skip_to_closing_square_bracket_1 (parser,
+							      CPP_COMMA) == -1)
+		{
+		  auto_diagnostic_group d;
+		  warning_at (cp_lexer_peek_token (parser->lexer)->location,
+			      OPT_Wcomma_subscript,
+			      "top-level comma expression in array subscript "
+			      "is deprecated");
+		}
+
+	      /* Roll back the tokens we skipped.  */
+	      cp_lexer_rollback_tokens (parser->lexer);
+	    }
+
+	  index = cp_parser_expression (parser);
+	}
     }
 
   parser->greater_than_is_operator_p = saved_greater_than_is_operator_p;
@@ -22857,16 +22885,25 @@ cp_parser_braced_list (cp_parser* parser, bool* non_constant_p)
 }
 
 /* Consume tokens up to, and including, the next non-nested closing `]'.
-   Returns true iff we found a closing `]'.  */
+   Returns 1 iff we found a closing `]'.  Returns -1 if OR_TTYPE is not
+   CPP_EOF and we found an unnested token of that type.  */
 
-static bool
-cp_parser_skip_to_closing_square_bracket (cp_parser *parser)
+static int
+cp_parser_skip_to_closing_square_bracket_1 (cp_parser *parser,
+					    enum cpp_ttype or_ttype)
 {
   unsigned square_depth = 0;
+  unsigned paren_depth = 0;
+  unsigned brace_depth = 0;
 
   while (true)
     {
-      cp_token * token = cp_lexer_peek_token (parser->lexer);
+      cp_token *token = cp_lexer_peek_token (parser->lexer);
+
+      /* Have we found what we're looking for before the closing square?  */
+      if (token->type == or_ttype && or_ttype != CPP_EOF
+	  && brace_depth == 0 && paren_depth == 0 && square_depth == 0)
+	return -1;
 
       switch (token->type)
 	{
@@ -22876,20 +22913,38 @@ cp_parser_skip_to_closing_square_bracket (cp_parser *parser)
 	  /* FALLTHRU */
 	case CPP_EOF:
 	  /* If we've run out of tokens, then there is no closing `]'.  */
-	  return false;
+	  return 0;
 
         case CPP_OPEN_SQUARE:
           ++square_depth;
           break;
 
         case CPP_CLOSE_SQUARE:
-	  if (!square_depth--)
+	  if (square_depth-- == 0)
 	    {
 	      cp_lexer_consume_token (parser->lexer);
-	      return true;
+	      return 1;
 	    }
 	  break;
 
+	case CPP_OPEN_BRACE:
+	  ++brace_depth;
+	  break;
+
+	case CPP_CLOSE_BRACE:
+	  if (brace_depth-- == 0)
+	    return 0;
+	  break;
+
+	case CPP_OPEN_PAREN:
+	  ++paren_depth;
+	  break;
+
+	case CPP_CLOSE_PAREN:
+	  if (paren_depth-- == 0)
+	    return 0;
+	  break;
+
 	default:
 	  break;
 	}
@@ -22899,6 +22954,15 @@ cp_parser_skip_to_closing_square_bracket (cp_parser *parser)
     }
 }
 
+/* Consume tokens up to, and including, the next non-nested closing `]'.
+   Returns true iff we found a closing `]'.  */
+
+static bool
+cp_parser_skip_to_closing_square_bracket (cp_parser *parser)
+{
+  return cp_parser_skip_to_closing_square_bracket_1 (parser, CPP_EOF) == 1;
+}
+
 /* Return true if we are looking at an array-designator, false otherwise.  */
 
 static bool
diff --git gcc/doc/invoke.texi gcc/doc/invoke.texi
index 7b3c77b8033..01aab60f895 100644
--- gcc/doc/invoke.texi
+++ gcc/doc/invoke.texi
@@ -230,7 +230,8 @@ in the following sections.
 -fvisibility-inlines-hidden @gol
 -fvisibility-ms-compat @gol
 -fext-numeric-literals @gol
--Wabi=@var{n}  -Wabi-tag  -Wconversion-null  -Wctor-dtor-privacy @gol
+-Wabi=@var{n}  -Wabi-tag  -Wcomma-subscript  -Wconversion-null @gol
+-Wctor-dtor-privacy @gol
 -Wdelete-non-virtual-dtor  -Wdeprecated-copy  -Wdeprecated-copy-dtor @gol
 -Wliteral-suffix @gol
 -Wmultiple-inheritance  -Wno-init-list-lifetime @gol
@@ -3037,6 +3038,24 @@ Warn when a type with an ABI tag is used in a context that does not
 have that ABI tag.  See @ref{C++ Attributes} for more information
 about ABI tags.
 
+@item -Wcomma-subscript @r{(C++ and Objective-C++ only)}
+@opindex Wcomma-subscript
+@opindex Wno-comma-subscript
+Warn about uses of a comma expression within a subscripting expression.
+This usage was deprecated in C++2a.  However, a comma expression wrapped
+in @code{( )} is not deprecated.  Example:
+
+@smallexample
+@group
+void f(int *a, int b, int c) @{
+    a[b,c];     // deprecated
+    a[(b,c)];   // OK
+@}
+@end group
+@end smallexample
+
+Enabled by default with @option{-std=c++2a}.
+
 @item -Wctor-dtor-privacy @r{(C++ and Objective-C++ only)}
 @opindex Wctor-dtor-privacy
 @opindex Wno-ctor-dtor-privacy
diff --git gcc/testsuite/g++.dg/cpp2a/comma1.C gcc/testsuite/g++.dg/cpp2a/comma1.C
new file mode 100644
index 00000000000..5de48b4cdb9
--- /dev/null
+++ gcc/testsuite/g++.dg/cpp2a/comma1.C
@@ -0,0 +1,26 @@
+// PR c++/91338 - P1161R3: Deprecate a[b,c].
+// { dg-do compile { target c++11 } }
+
+struct S {
+  int operator,(int) { return 42; }
+};
+
+void
+fn (int *a, int b, int c)
+{
+  a[b,c]; // { dg-warning "top-level comma expression in array subscript is deprecated" "" { target c++2a } }
+  a[(b,c)];
+
+  a[(void) b, c]; // { dg-warning "top-level comma expression in array subscript is deprecated" "" { target c++2a } }
+  a[((void) b, c)];
+
+  a[(void) b, (void) c, (void) b, b]; // { dg-warning "top-level comma expression in array subscript is deprecated" "" { target c++2a } }
+  a[((void) b, (void) c, (void) b, b)];
+
+  a[S(), 10]; // { dg-warning "top-level comma expression in array subscript is deprecated" "" { target c++2a } }
+  a[(S(), 10)];
+
+  a[int{(1,2)}];
+  a[int{(1,2)}, int{}]; // { dg-warning "top-level comma expression in array subscript is deprecated" "" { target c++2a } }
+  a[(int{(1,2)}, int{})];
+}
diff --git gcc/testsuite/g++.dg/cpp2a/comma2.C gcc/testsuite/g++.dg/cpp2a/comma2.C
new file mode 100644
index 00000000000..383079ce4f8
--- /dev/null
+++ gcc/testsuite/g++.dg/cpp2a/comma2.C
@@ -0,0 +1,27 @@
+// PR c++/91338 - P1161R3: Deprecate a[b,c].
+// { dg-do compile { target c++2a } }
+// { dg-options "-Wno-comma-subscript" }
+
+struct S {
+  int operator,(int) { return 42; }
+};
+
+void
+fn (int *a, int b, int c)
+{
+  a[b,c]; // { dg-bogus "top-level comma expression in array subscript is deprecated" }
+  a[(b,c)];
+
+  a[(void) b, c]; // { dg-bogus "top-level comma expression in array subscript is deprecated" }
+  a[((void) b, c)];
+
+  a[(void) b, (void) c, (void) b, b]; // { dg-bogus "top-level comma expression in array subscript is deprecated" }
+  a[((void) b, (void) c, (void) b, b)];
+
+  a[S(), 10]; // { dg-bogus "top-level comma expression in array subscript is deprecated" }
+  a[(S(), 10)];
+
+  a[int{(1,2)}];
+  a[int{(1,2)}, int{}]; // { dg-bogus "top-level comma expression in array subscript is deprecated" }
+  a[(int{(1,2)}, int{})];
+}
diff --git gcc/testsuite/g++.dg/cpp2a/comma3.C gcc/testsuite/g++.dg/cpp2a/comma3.C
new file mode 100644
index 00000000000..59a518382d7
--- /dev/null
+++ gcc/testsuite/g++.dg/cpp2a/comma3.C
@@ -0,0 +1,27 @@
+// PR c++/91338 - P1161R3: Deprecate a[b,c].
+// { dg-do compile { target c++11 } }
+// { dg-options "-Wcomma-subscript" }
+
+struct S {
+  int operator,(int) { return 42; }
+};
+
+void
+fn (int *a, int b, int c)
+{
+  a[b,c]; // { dg-warning "top-level comma expression in array subscript is deprecated" }
+  a[(b,c)];
+
+  a[(void) b, c]; // { dg-warning "top-level comma expression in array subscript is deprecated" }
+  a[((void) b, c)];
+
+  a[(void) b, (void) c, (void) b, b]; // { dg-warning "top-level comma expression in array subscript is deprecated" }
+  a[((void) b, (void) c, (void) b, b)];
+
+  a[S(), 10]; // { dg-warning "top-level comma expression in array subscript is deprecated" }
+  a[(S(), 10)];
+
+  a[int{(1,2)}];
+  a[int{(1,2)}, int{}]; // { dg-warning "top-level comma expression in array subscript is deprecated" }
+  a[(int{(1,2)}, int{})];
+}