even better handling of invalid Unicode characters

git-svn-id: svn://svn.savannah.gnu.org/nano/trunk/nano@2975 35c25a1d-7b9e-4130-9fde-d3aeb78583b8

diff --git a/ChangeLog b/ChangeLog
index 2321bc7afce69e8486ec73b1e4e2c8dbccf1885c..70764eff9922117fb19b38777a8313f16a09eb52 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -134,6 +134,10 @@ CVS code -
          get_key_buffer(), check_statusblank(), nanogetstr(),
          titlebar(), statusbar(), bottombars(), edit_refresh(),
          do_yesno(), and do_help(). (DLR)
+       - Treat the Unicode characters D800-DFFF and FFFE-FFFF as
+         invalid, since the C library's multibyte functions don't seem
+         to.  New function is_valid_unicode(); changes to mbrep() and
+         make_mbchar(). (DLR)
 - color.c:
        - Remove unneeded fcntl.h include. (DLR)
 - chars.c:
@@ -143,13 +147,6 @@ CVS code -
   mbrep()
        - New function, the equivalent of control_mbrep() for non-control
          characters. (DLR)
-       - Treat the Unicode characters D800-DFFF and FFFE-FFFF as
-         invalid, since the C library's multibyte functions don't seem
-         to. (DLR)
-  make_mbchar()
-       - Treat the Unicode characters D800-DFFF and FFFE-FFFF as
-         invalid, since the C library's multibyte functions don't seem
-         to. (DLR)
   parse_mbchar()
        - Remove now-unneeded bad_chr parameter. (DLR)
   mbstrchr()

diff --git a/src/chars.c b/src/chars.c
index f21b9d6f4886131655d338720ff7991b9d7b72f5..8703d9cc1fae756afa536f5a8a43f6a06e5e426c 100644 (file)
--- a/src/chars.c
+++ b/src/chars.c
@@ -255,10 +255,8 @@ char *mbrep(const char *c, char *crep, int *crep_len)
     if (ISSET(USE_UTF8)) {
        wchar_t wc;
 
-       /* Unicode D800-DFFF and FFFE-FFFF are invalid, even though
-        * they're parsed properly. */
-       if (mbtowc(&wc, c, MB_CUR_MAX) < 0 || ((0xD800 <= wc && wc <=
-               0xDFFF) || (0XFFFE <= wc && wc <= 0xFFFF))) {
+       /* Reject invalid Unicode characters. */
+       if (mbtowc(&wc, c, MB_CUR_MAX) < 0 || !is_valid_unicode(wc)) {
            mbtowc(NULL, NULL, 0);
            crep = (char *)bad_mbchar;
            *crep_len = bad_mbchar_len;
@@ -331,12 +329,10 @@ char *make_mbchar(int chr, int *chr_mb_len)
 #ifdef ENABLE_UTF8
     if (ISSET(USE_UTF8)) {
        chr_mb = charalloc(MB_CUR_MAX);
-       *chr_mb_len = wctomb(chr_mb, chr);
+       *chr_mb_len = wctomb(chr_mb, (wchar_t)chr);
 
-       /* Unicode D800-DFFF and FFFE-FFFF are invalid, even though
-        * they're parsed properly. */
-       if (*chr_mb_len < 0 || ((0xD800 <= chr && chr <= 0xDFFF) ||
-               (0XFFFE <= chr && chr <= 0xFFFF))) {
+       /* Reject invalid Unicode characters. */
+       if (*chr_mb_len < 0 || !is_valid_unicode((wchar_t)chr)) {
            wctomb(NULL, 0);
            *chr_mb_len = 0;
        }
@@ -887,6 +883,16 @@ bool has_blank_mbchars(const char *s)
 #endif /* ENABLE_NANORC */
 #endif /* !DISABLE_JUSTIFY */
 
+#ifdef ENABLE_UTF8
+/* Return TRUE if wc is valid Unicode (i.e, it's not negative or in the
+ * ranges D800-DFFF or FFFE-FFFF), and FALSE otherwise. */
+bool is_valid_unicode(wchar_t wc)
+{
+    return (0 <= wc && (wc <= 0xD7FF || 0xE000 <= wc) && (wc !=
+       0xFFFE && wc != 0xFFFF));
+}
+#endif
+
 #ifdef ENABLE_NANORC
 /* Check if the string s is a valid multibyte string.  Return TRUE if it
  * is, and FALSE otherwise. */

diff --git a/src/proto.h b/src/proto.h
index d3d5ee7cb1cf2706cd452b94f4837508930708d2..86ef80eedafb92bbf8a7f4e36bbefa79c2df9b50 100644 (file)
--- a/src/proto.h
+++ b/src/proto.h
@@ -197,6 +197,9 @@ bool has_blank_chars(const char *s);
 bool has_blank_mbchars(const char *s);
 #endif
 #endif
+#ifdef ENABLE_UTF8
+bool is_valid_unicode(wchar_t wc);
+#endif
 #ifdef ENABLE_NANORC
 bool is_valid_mbstring(const char *s);
 #endif