Take UTF-8 spaces into account for empty strings

2 files changed, 179 insertions, 15 deletions

diff --git a/smenu.c b/smenu.c
index 4a08024..3266f8e 100755
--- a/smenu.c
+++ b/smenu.c
@@ -83,6 +83,7 @@ int marked = -1; /* Index of the marked word or -1. */
 char *word_buffer;
 
 int (*my_isprint)(int);
+int (*my_isempty)(const unsigned char *);
 
 /* UTF-8 useful symbols. */
 /* """"""""""""""""""""" */
@@ -1692,20 +1693,164 @@ read:
 
 /* ====================================================== */
 /* Returns 1 if a string is empty or only made of spaces. */
-/* TODO: take all UTF-8 spaces into account.              */
+/* Non UTF-8 version.                                     */
 /* ====================================================== */
 int
-isempty(const char *s)
+isempty_non_utf8(const unsigned char *s)
 {
   while (*s != '\0')
   {
-    if (my_isprint(*s) && *s != ' ' && *s != '\t')
+    if (*s != ' ' && *s != '\t')
       return 0;
     s++;
   }
   return 1;
 }
 
+/* ====================================================== */
+/* Returns 1 if a string is empty or only made of spaces. */
+/* UTF-8 version.                                         */
+/* ====================================================== */
+int
+isempty_utf8(const unsigned char *s)
+{
+  unsigned char c, d;
+
+  while (*s != '\0')
+  {
+    if (*s == ' ' || *s == '\t') /* Normal ASCII spaces. */
+      goto next;
+
+    if (*s < 0xc2) /* Not an UTF-8 space -> return FALSE. */
+      return 0;
+
+    /* Scanning for a potential non UTF-8 spaces scanning. */
+    /* """"""""""""""""""""""""""""""""""""""""""""""""""" */
+    if ((c = *(s + 1)) != '\0')
+    {
+      if (*s == 0xc2 && (c == 0x85 || c == 0xa0))
+      {
+        s++;
+        goto next; /* Unnamed control character or NO-BREAK SPACE. */
+      }
+
+      if ((d = *(s + 2)) == '\0')
+        return 0;
+
+      if (*s == 0xe1 && c == 0x9a && d == 0x80)
+      {
+        s += 2;
+        goto next; /* OGHAM SPACE MARK. */
+      }
+
+      if (*s == 0xe1 && c == 0xa0 && d == 0x8e)
+      {
+        s += 2;
+        goto next; /* MONGOLIAN VOWEL SEPARATOR. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x80)
+      {
+        s += 2;
+        goto next; /* EN QUAD. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x81)
+      {
+        s += 2;
+        goto next; /* EM QUAD. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x82)
+      {
+        s += 2;
+        goto next; /* EN SPACE. */
+      }
+      if (*s == 0xe2 && c == 0x80 && d == 0x83)
+      {
+        s += 2;
+        goto next; /* EM SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x84)
+      {
+        s += 2;
+        goto next; /* THREE-PER-EM SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x85)
+      {
+        s += 2;
+        goto next; /* FOUR-PER-EM SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x86)
+      {
+        s += 2;
+        goto next; /* SIX-PER-EM SPACE. */
+      }
+      if (*s == 0xe2 && c == 0x80 && d == 0x87)
+      {
+        s += 2;
+        goto next; /* FIGURE SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x88)
+      {
+        s += 2;
+        goto next; /* PUNCTUATION SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x89)
+      {
+        s += 2;
+        goto next; /* THIN SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0x8a)
+      {
+        s += 2;
+        goto next; /* HAIR SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0xa8)
+      {
+        s += 2;
+        goto next; /* LINE SEPARATOR. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0xa9)
+      {
+        s += 2;
+        goto next; /* PARAGRAPH SEPARATOR. */
+      }
+
+      if (*s == 0xe2 && c == 0x80 && d == 0xaf)
+      {
+        s += 2;
+        goto next; /* NARROW NO-BREAK SPACE. */
+      }
+
+      if (*s == 0xe2 && c == 0x81 && d == 0x9f)
+      {
+        s += 2;
+        goto next; /* MEDIUM MATHEMATICAL SPACE. */
+      }
+
+      if (*s == 0xe3 && c == 0x80 && d == 0x80)
+      {
+        s += 2;
+        goto next; /* IDEOGRAPHIC SPACE. */
+      }
+
+      return 0;
+    }
+  next:
+    s++;
+  }
+  return 1;
+}
+
 /* ======================================================================== */
 /* Parse a regular expression based selector.                               */
 /* The string to parse is bounded by a delimiter so we must parse something */
@@ -3550,16 +3695,19 @@ get_word(FILE          *input,
            && ll_find(line_delims_list, utf8_buffer, buffer_cmp) == NULL)
       byte = get_bytes(input, utf8_buffer, zapped_glyphs_list, langinfo, misc);
 
-    if (langinfo->utf8 && utf8_get_length(utf8_buffer[0]) > 1)
+    if (byte != EOF)
     {
-      size_t pos;
+      if (langinfo->utf8 && utf8_get_length(utf8_buffer[0]) > 1)
+      {
+        size_t pos;
 
-      pos = strlen(utf8_buffer);
-      while (pos > 0)
-        my_ungetc(utf8_buffer[--pos], input);
+        pos = strlen(utf8_buffer);
+        while (pos > 0)
+          my_ungetc(utf8_buffer[--pos], input);
+      }
+      else
+        my_ungetc(byte, input);
     }
-    else
-      my_ungetc(byte, input);
   }
 
   /* Mark it as the last word of a record if its sequence matches a */
@@ -8833,14 +8981,24 @@ main(int argc, char *argv[])
   else
     langinfo.utf8 = 0;
 
-  /* my_isprint is a function pointer that points to the 7 or 8-bit */
-  /* version of isprint according to the content of UTF-8.          */
-  /* """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" */
+  /* my_isprint is a function pointers that points to   */
+  /* the 7 or 8-bit version of isprint according to the */
+  /* current locale.                                    */
+  /* """""""""""""""""""""""""""""""""""""""""""""""""" */
   if (langinfo.utf8 || langinfo.bits == 8)
     my_isprint = isprint8;
   else
     my_isprint = isprint7;
 
+  /* my_isempty is a function pointers that points to         */
+  /* the utf8 or non_utf8 version of isprint according to the */
+  /* current locale.                                          */
+  /* """""""""""""""""""""""""""""""""""""""""""""""""""""""" */
+  if (langinfo.utf8)
+    my_isempty = isempty_utf8;
+  else
+    my_isempty = isempty_non_utf8;
+
   /* Set terminal in noncanonical, noecho mode and  */
   /* if TERM is unset or unknown, vt100 is assumed. */
   /* """""""""""""""""""""""""""""""""""""""""""""" */
@@ -11322,7 +11480,7 @@ main(int argc, char *argv[])
         long *word_pos = xmalloc(sizeof(long));
         int   may_number;
 
-        if (!isempty(word->str))
+        if (!my_isempty(word->str))
         {
           *word_pos = wi;
 
@@ -11774,7 +11932,7 @@ main(int argc, char *argv[])
         else
           len = 0;
 
-        if (!isempty(word_a[wi + offset].str + len))
+        if (!my_isempty(word_a[wi + offset].str + len))
           break;
 
         /* Keep non selectable empty words to allow special effects. */
diff --git a/smenu.h b/smenu.h
index 9994b54..c242ab1 100755
--- a/smenu.h
+++ b/smenu.h
@@ -822,3 +822,9 @@ get_clicked_index(win_t  *win,
 
 void
 align_word(word_t *word, alignment_t alignment, size_t prerfix, char sp);
+
+int
+isempty_non_utf8(const unsigned char * s);
+
+int
+isempty_utf8(const unsigned char * s);