On 7/29/25 3:45 AM, Jakub Jelinek wrote:
Hi!
The P2843R3 Preprocessing is never undefined paper contains comments
that various compilers handle comma operators in preprocessor expressions
incorrectly and I think they are right.
In both C and C++ the grammar uses constant-expression non-terminal
for #if/#elif and in both C and C++ that NT is conditional-expression,
so
#if 1, 2
is IMHO clearly wrong in both languages.
C89 then says for constant-expression
"Constant expressions shall not contain assignment, increment, decrement,
function-call, or comma operators, except when they are contained within the
operand of a sizeof operator."
Because all the remaining identifiers in the #if/#elif expression are
replaced with 0 I think assignments, increment, decrement and function-call
aren't that big deal because (0 = 1) or ++4 etc. are all invalid, but
for comma expressions I think it matters. In r0-56429 PR456 Joseph has
added !CPP_OPTION (pfile, c99) to handle that correctly.
Then C99 changed that to:
"Constant expressions shall not contain assignment, increment, decrement,
function-call,
or comma operators, except when they are contained within a subexpression that
is not
evaluated."
That made for C99+
#if 1 || (1, 2)
etc. valid but
#if (1, 2)
is still invalid, ditto
#if 1 ? 1, 2 : 3
In C++ I can't find anything like that though, and as can be seen on say
int a[(1, 2)];
int b[1 ? 1, 2 : 3];
being accepted by C++ and rejected by C while
int c[1, 2];
int d[1 ? 2 : 3, 4];
being rejected in both C and C++, so I think for C++ it is indeed just the
grammar that prevents #if 1, 2. When it is the second operand of ?: or
inside of () the grammar just uses expression and that allows comma
operator.
So, the following patch uses different decisions for C++ when to diagnose
comma operator in preprocessor expressions, for C++ tracks if it is inside
of () (obviously () around #embed clauses don't count unless one uses
limit ((1, 2)) etc.) or inside of the second ?: operand and allows comma
operator there and disallows elsewhere.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
OK.
BTW, I wonder if anything in the standard disallows <=> in the preprocessor
expressions. Say
#if (0 <=> 1) < 0
etc.
#include <compare>
constexpr int a = (0 <=> 1) < 0;
is valid (but not valid without #include <compare>) and the expressions
don't use any identifiers.
I don't see anything to disallow it, but it also seems useless.
2025-07-29 Jakub Jelinek <ja...@redhat.com>
PR c++/120778
* internal.h (struct lexer_state): Add comma_ok member.
* expr.cc (_cpp_parse_expr): Initialize it to 0, increment on
CPP_OPEN_PAREN and CPP_QUERY, decrement on CPP_CLOSE_PAREN
and CPP_COLON.
(num_binary_op): For C++ pedwarn on comma operator if
pfile->state.comma_ok is 0 instead of !c99 or skip_eval.
* g++.dg/cpp/if-comma-1.C: New test.
--- libcpp/internal.h.jj 2025-07-27 23:31:11.218987636 +0200
+++ libcpp/internal.h 2025-07-28 17:28:18.162037320 +0200
@@ -282,6 +282,9 @@ struct lexer_state
/* Nonzero to skip evaluating part of an expression. */
unsigned int skip_eval;
+ /* Nonzero if CPP_COMMA is valid in expression in C++. */
+ unsigned int comma_ok;
+
/* Nonzero when tokenizing a deferred pragma. */
unsigned char in_deferred_pragma;
--- libcpp/expr.cc.jj 2025-04-08 14:09:46.899507170 +0200
+++ libcpp/expr.cc 2025-07-28 17:57:55.904095906 +0200
@@ -1460,6 +1460,7 @@ _cpp_parse_expr (cpp_reader *pfile, cons
location_t virtual_location = 0;
pfile->state.skip_eval = 0;
+ pfile->state.comma_ok = 0;
/* Set up detection of #if ! defined(). */
pfile->mi_ind_cmacro = 0;
@@ -1521,6 +1522,10 @@ _cpp_parse_expr (cpp_reader *pfile, cons
lex_count--;
continue;
+ case CPP_OPEN_PAREN:
+ pfile->state.comma_ok++;
+ break;
+
default:
if ((int) op.op <= (int) CPP_EQ || (int) op.op >= (int) CPP_PLUS_EQ)
SYNTAX_ERROR2_AT (op.loc,
@@ -1574,13 +1579,16 @@ _cpp_parse_expr (cpp_reader *pfile, cons
case CPP_CLOSE_PAREN:
if (pfile->state.in_directive == 3 && top == pfile->op_stack)
goto embed_done;
+ pfile->state.comma_ok--;
continue;
case CPP_OR_OR:
if (!num_zerop (top->value))
pfile->state.skip_eval++;
break;
- case CPP_AND_AND:
case CPP_QUERY:
+ pfile->state.comma_ok++;
+ /* FALLTHRU */
+ case CPP_AND_AND:
if (num_zerop (top->value))
pfile->state.skip_eval++;
break;
@@ -1592,6 +1600,8 @@ _cpp_parse_expr (cpp_reader *pfile, cons
pfile->state.skip_eval++;
else
pfile->state.skip_eval--;
+ pfile->state.comma_ok--;
+ break;
default:
break;
}
@@ -2209,8 +2219,10 @@ num_binary_op (cpp_reader *pfile, cpp_nu
/* Comma. */
default: /* case CPP_COMMA: */
- if (CPP_PEDANTIC (pfile) && (!CPP_OPTION (pfile, c99)
- || !pfile->state.skip_eval))
+ if (CPP_PEDANTIC (pfile)
+ && (CPP_OPTION (pfile, cplusplus)
+ ? !pfile->state.comma_ok
+ : (!CPP_OPTION (pfile, c99) || !pfile->state.skip_eval)))
cpp_pedwarning (pfile, CPP_W_PEDANTIC,
"comma operator in operand of #%s",
pfile->state.in_directive == 3
--- gcc/testsuite/g++.dg/cpp/if-comma-1.C.jj 2025-07-28 18:14:51.530977617
+0200
+++ gcc/testsuite/g++.dg/cpp/if-comma-1.C 2025-07-28 18:14:41.180111370
+0200
@@ -0,0 +1,42 @@
+// PR c++/120778
+// { dg-do preprocess }
+// { dg-options "-pedantic-errors" }
+
+#if (1, 2)
+#define M1 1
+#else
+#error
+#endif
+#if 1 ? 2, 3 : 4
+#define M2 2
+#else
+#error
+#endif
+#if 0 ? 2, 0 : 1
+#define M3 3
+#else
+#error
+#endif
+#if 0 || (1, 2)
+#define M4 4
+#else
+#error
+#endif
+#if 1 || (1, 2)
+#define M5 5
+#else
+#error
+#endif
+#if (1, 2) && 1
+#define M6 6
+#else
+#error
+#endif
+#if 1 && (1, 2)
+#define M7 7
+#else
+#error
+#endif
+#if M1 + M2 + M3 + M4 + M5 + M6 + M7 != 28
+#error
+#endif
Jakub
