Re: From wchar_t to char32_t

[Bruno Haible]

This patch migrates the 'mbswidth' module from wchar_t to char32_t.


2023-07-02  Bruno Haible  <bruno@clisp.org>

        mbswidth: Overcome wchar_t limitations.
        * lib/mbswidth.c: Include <uchar.h> instead of <wctype.h>.
        (mbsnwidth): Use mbrtoc32 instead of mbrtowc. Use c32width instead of
        wcwidth. Use c32iscntrl instead of iswcntrl.
        * modules/mbswidth (Depends-on): Remove wctype-h, mbrtowc, wcwidth. Add
        uchar, mbrtoc32, c32width, c32iscntrl.
        (Link): Add $(LIBC32CONV).

diff --git a/lib/mbswidth.c b/lib/mbswidth.c
index ca0a47d16f..da2d8030f3 100644
--- a/lib/mbswidth.c
+++ b/lib/mbswidth.c
@@ -29,11 +29,11 @@
 /* Get isprint().  */
 #include <ctype.h>
 
-/* Get mbstate_t, mbrtowc(), mbsinit(), wcwidth().  */
+/* Get mbstate_t, mbsinit().  */
 #include <wchar.h>
 
-/* Get iswcntrl().  */
-#include <wctype.h>
+/* Get char32_t, mbrtoc32(), c32iscntrl(), c32width().  */
+#include <uchar.h>
 
 /* Get INT_MAX.  */
 #include <limits.h>
@@ -97,11 +97,11 @@ mbsnwidth (const char *string, size_t nbytes, int flags)
                 memset (&mbstate, 0, sizeof mbstate);
                 do
                   {
-                    wchar_t wc;
+                    char32_t wc;
                     size_t bytes;
                     int w;
 
-                    bytes = mbrtowc (&wc, p, plimit - p, &mbstate);
+                    bytes = mbrtoc32 (&wc, p, plimit - p, &mbstate);
 
                     if (bytes == (size_t) -1)
                       /* An invalid multibyte sequence was encountered.  */
@@ -132,8 +132,10 @@ mbsnwidth (const char *string, size_t nbytes, int flags)
                     if (bytes == 0)
                       /* A null wide character was encountered.  */
                       bytes = 1;
+                    else if (bytes == (size_t) -3)
+                      bytes = 0;
 
-                    w = wcwidth (wc);
+                    w = c32width (wc);
                     if (w >= 0)
                       /* A printable multibyte character.  */
                       {
@@ -145,7 +147,7 @@ mbsnwidth (const char *string, size_t nbytes, int flags)
                       /* An unprintable multibyte character.  */
                       if (!(flags & MBSW_REJECT_UNPRINTABLE))
                         {
-                          if (!iswcntrl (wc))
+                          if (!c32iscntrl (wc))
                             {
                               if (width == INT_MAX)
                                 goto overflow;
diff --git a/modules/mbswidth b/modules/mbswidth
index f3fe2ed21b..4dd8c55ee4 100644
--- a/modules/mbswidth
+++ b/modules/mbswidth
@@ -10,10 +10,11 @@ m4/mbswidth.m4
 
 Depends-on:
 wchar
-wctype-h
-mbrtowc
+uchar
+mbrtoc32
 mbsinit
-wcwidth
+c32width
+c32iscntrl
 extensions
 
 configure.ac:
@@ -28,6 +29,7 @@ Include:
 Link:
 $(LTLIBUNISTRING) when linking with libtool, $(LIBUNISTRING) otherwise
 $(MBRTOWC_LIB)
+$(LTLIBC32CONV) when linking with libtool, $(LIBC32CONV) otherwise
 
 License:
 GPL