> On 15 Aug 2026, at 19:23, Dimitry Andric <[email protected]> wrote:
>
> On 14 Aug 2026, at 21:32, Faraz Vahedi <[email protected]> wrote:
>>
>> The branch main has been updated by kfv:
>>
>> URL:
>> https://cgit.FreeBSD.org/src/commit/?id=cd0727ec709bb54f8f82104f6113284a15dd3464
>>
>> commit cd0727ec709bb54f8f82104f6113284a15dd3464
>> Author: Faraz Vahedi <[email protected]>
>> AuthorDate: 2026-08-14 12:22:00 +0000
>> Commit: Faraz Vahedi <[email protected]>
>> CommitDate: 2026-08-14 19:31:54 +0000
>>
>> libc: Add <stdio.h> C23 feature test macro
>>
>> Define the __STDC_VERSION_STDIO_H__ feature test macro now that
>> the header fully conforms to C23.
>>
>> Reviewed by: fuz
>> Approved by: fuz (mentor)
>> MFC after: 1 month
>> Differential Revision: https://reviews.freebsd.org/D58842
> ...
>> diff --git a/include/stdio.h b/include/stdio.h
>> index ea016d65095d..50e3ff2d1c34 100644
>> --- a/include/stdio.h
>> +++ b/include/stdio.h
>> @@ -32,8 +32,8 @@
>> * SUCH DAMAGE.
>> */
>>
>> -#ifndef _STDIO_H_
>> -#define _STDIO_H_
>> +#ifndef __STDC_VERSION_STDIO_H__
>> +#define __STDC_VERSION_STDIO_H__ 202311L
>>
>> #include <sys/cdefs.h>
>> #include <sys/_null.h>
>
> This particular change appears to have broken the math/gmp port, which
> hasn't functionally changed since 2023.
>
> It has this construct to detect the existence of FILE:
>
> /* For reference, "defined(EOF)" cannot be used here. In g++ 2.95.4,
> <iostream> defines EOF but not FILE. */
> #if defined (FILE) \
> || defined (H_STDIO) \
> || defined (_H_STDIO) /* AIX */ \
> || defined (_STDIO_H) /* glibc, Sun, SCO */ \
> || defined (_STDIO_H_) /* BSD, OSF */ \
> || defined (__STDIO_H) /* Borland */ \
> || defined (__STDIO_H__) /* IRIX */ \
> || defined (_STDIO_INCLUDED) /* HPUX */ \
> || defined (__dj_include_stdio_h_) /* DJGPP */ \
> || defined (_FILE_DEFINED) /* Microsoft */ \
> || defined (__STDIO__) /* Apple MPW MrC */ \
> || defined (_MSL_STDIO_H) /* Metrowerks */ \
> || defined (_STDIO_H_INCLUDED) /* QNX4 */ \
> || defined (_ISO_STDIO_ISO_H) /* Sun C++ */ \
> || defined (__STDIO_LOADED) /* VMS */ \
> || defined (_STDIO) /* HPE NonStop */ \
> || defined (__DEFINED_FILE) /* musl */
> #define _GMP_H_HAVE_FILE 1
> #endif
>
> Now that _STDIO_H_ has been removed, the detection fails and another
> part of the port falls over due to an incorrect prototype.
>
> I guess we could patch the math/gmp port with another
> __STDC_VERSION_STDIO_H__ macro, but it might be good to do an exp-run to
> see if there are more GNU projects which rely on the existence of the
> _STDIO_H_ macro?
FWIW, I would prefer something like this:
diff --git a/include/stdio.h b/include/stdio.h
index 50e3ff2d1c34..07ed71035e84 100644
--- a/include/stdio.h
+++ b/include/stdio.h
@@ -32,8 +32,12 @@
* SUCH DAMAGE.
*/
+#ifndef _STDIO_H_
+#define _STDIO_H_
+
#ifndef __STDC_VERSION_STDIO_H__
#define __STDC_VERSION_STDIO_H__ 202311L
+#endif
#include <sys/cdefs.h>
#include <sys/_null.h>
@@ -540,4 +544,4 @@ extern int __isthreaded;
__END_DECLS
__NULLABILITY_PRAGMA_POP
-#endif /* !__STDC_VERSION_STDIO_H__ */
+#endif /* !_STDIO_H_ */
-Dimitry