[PATCH 1/2] regex: new _REGEX_AVOID_UCHAR_H option

This is for GNU Emacs, where the regex code is only auxiliary,
probably not enough to justify the hassle of incorporating
Gnulib’s uchar-h module and its dependencies.
The following changes are in effect only if _REGEX_AVOID_UCHAR_H,
and cause the code to avoid the need for Gnulib’s uchar-h module.
* lib/regcomp.c: Don’t include localeinfo.h.
(re_set_fastmap_icase): Don’t use localeinfo.h functions,
as they rely on char32_t.
* lib/regex_internal.h: Just use of wchar_t, wctype_t and related
functions, instead of using char32_t and related functions.
---
 ChangeLog            | 14 ++++++++++++++
 lib/regcomp.c        |  4 ++--
 lib/regex_internal.h | 38 ++++++++++++++++++++++++++------------
 3 files changed, 42 insertions(+), 14 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 8892a3d09b..169ee053dd 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,17 @@
+2026-05-24  Paul Eggert  <[email protected]>
+
+	regex: new _REGEX_AVOID_UCHAR_H option
+	This is for GNU Emacs, where the regex code is only auxiliary,
+	probably not enough to justify the hassle of incorporating
+	Gnulib’s uchar-h module and its dependencies.
+	The following changes are in effect only if _REGEX_AVOID_UCHAR_H,
+	and cause the code to avoid the need for Gnulib’s uchar-h module.
+	* lib/regcomp.c: Don’t include localeinfo.h.
+	(re_set_fastmap_icase): Don’t use localeinfo.h functions,
+	as they rely on char32_t.
+	* lib/regex_internal.h: Just use of wchar_t, wctype_t and related
+	functions, instead of using char32_t and related functions.
+
 2026-05-22  Paul Eggert  <[email protected]>
 
 	regex-tests: omit unused parts
diff --git a/lib/regcomp.c b/lib/regcomp.c
index ab8716991c..ab3783f94c 100644
--- a/lib/regcomp.c
+++ b/lib/regcomp.c
@@ -23,7 +23,7 @@
 
 /* The localeinfo-related code fixes glibc bug 20381.
    Someday this fix should be merged into glibc.  */
-#ifndef _LIBC
+#if !defined _LIBC && !defined _REGEX_AVOID_UCHAR_H
 # include "localeinfo.h"
 #endif
 
@@ -285,7 +285,7 @@ re_set_fastmap (char *fastmap, unsigned char ch)
 static void
 re_set_fastmap_icase (char *fastmap, wchar_t wc, mbstate_t *mbs)
 {
-#ifdef _LIBC
+#if defined _LIBC || defined _REGEX_AVOID_UCHAR_H
   wchar_t folded[1] = {__towlower (wc)};
   int nfolded = folded[0] != wc;
 #else
diff --git a/lib/regex_internal.h b/lib/regex_internal.h
index 2409170b0a..bf6c9ba3b8 100644
--- a/lib/regex_internal.h
+++ b/lib/regex_internal.h
@@ -119,13 +119,18 @@
 #define NEWLINE_CHAR '\n'
 #define WIDE_NEWLINE_CHAR L'\n'
 
-/* Use Gnulib <uchar.h> when outside of glibc.  */
-#ifdef _LIBC
+/* Use Gnulib <uchar.h> if outside glibc and not avoided by the app.  */
+#if defined _LIBC || defined _REGEX_AVOID_UCHAR_H
 # include <wchar.h>
 # include <wctype.h>
 #else
 # include <uchar.h>
 # undef wctype_t
+# define wchar_t char32_t
+# define wctype_t c32_type_test_t
+#endif
+
+#ifndef _LIBC
 # undef __wctype
 # undef __iswalnum
 # undef __iswctype
@@ -135,17 +140,26 @@
 # undef __mbrtowc
 # undef __wcrtomb
 # undef __regfree
-# define wchar_t char32_t
-# define wctype_t c32_type_test_t
-# define __wctype c32_get_type_test
-# define __iswalnum c32isalnum
-# define __iswctype c32_apply_type_test
-# define __towlower c32tolower
-# define __towupper c32toupper
-# define __btowc btoc32
-# define __mbrtowc mbrtoc32
-# define __wcrtomb c32rtomb
 # define __regfree regfree
+# ifdef _REGEX_AVOID_UCHAR_H
+#  define __wctype wctype
+#  define __iswalnum iswalnum
+#  define __iswctype iswctype
+#  define __towlower towlower
+#  define __towupper towupper
+#  define __btowc btowc
+#  define __mbrtowc mbrtowc
+#  define __wcrtomb wcrtomb
+# else
+#  define __wctype c32_get_type_test
+#  define __iswalnum c32isalnum
+#  define __iswctype c32_apply_type_test
+#  define __towlower c32tolower
+#  define __towupper c32toupper
+#  define __btowc btoc32
+#  define __mbrtowc mbrtoc32
+#  define __wcrtomb c32rtomb
+# endif
 #endif /* not _LIBC */
 
 /* Types related to integers.  Unless protected by #ifdef _LIBC, the
-- 
2.54.0