Hi!
This is another case which changed from compile time undefined behavior
to ill-formed, diagnostic required. Now, we warn on this, so pedantically
that is good enough, maybe all we need is a testcase, but the following
patch changes it to a pedwarn for C++26.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
Or do you prefer a warning everywhere and just add a testcase?
2025-08-02 Jakub Jelinek <ja...@redhat.com>
PR preprocessor/120778
* macro.cc (stringify_arg): For C++26 emit a pedarn instead of warning
for \ at the end of stringification.
* g++.dg/DRs/dr2578.C: New test.
--- libcpp/macro.cc.jj 2025-08-01 15:55:51.131693172 +0200
+++ libcpp/macro.cc 2025-08-01 17:49:10.717143946 +0200
@@ -1003,7 +1003,10 @@ stringify_arg (cpp_reader *pfile, const
/* Ignore the final \ of invalid string literals. */
if (backslash_count & 1)
{
- cpp_error (pfile, CPP_DL_WARNING,
+ cpp_error (pfile,
+ CPP_OPTION (pfile, cplusplus)
+ && CPP_OPTION (pfile, lang) >= CLK_GNUCXX26
+ ? CPP_DL_PEDWARN : CPP_DL_WARNING,
"invalid string literal, ignoring final %<\\%>");
dest--;
}
--- gcc/testsuite/g++.dg/DRs/dr2578.C.jj 2025-08-01 18:03:58.297808256
+0200
+++ gcc/testsuite/g++.dg/DRs/dr2578.C 2025-08-01 18:13:01.174877937 +0200
@@ -0,0 +1,10 @@
+// DR 2578 - Undefined behavior when creating an invalid string literal via
stringizing
+// { dg-do preprocess }
+// { dg-options "-pedantic-errors" }
+
+#define A(a) #a
+#define B(a) A(a)
+#define C \\
+
+const char *x = B(C); // { dg-warning "invalid string literal, ignoring final
'\\\\'" "" { target c++23_down } }
+// { dg-error "invalid string literal, ignoring final '\\\\'" "" { target
c++26 } .-1 }
Jakub
