Support UTF-8 in the input bar

* introduce RP_IS_UTF8_{CHAR,CONT,START} macros.  Those yield non-zero
  only if the locale is UTF-8.
* use those macros in editor.c to properly handle UTF-8 multibyte
  characters.
* use them also in input.c:update_input_window, to draw the cursor

Reviewing and comments are welcome.
Patches for generic support of multibyte encodings are welcome too.
UTF-8 was chosen because of its processing simplicity and its wide use,
not because of any political opinion or religious belief. ;)

3 files changed, 150 insertions, 97 deletions

diff --git a/src/editor.c b/src/editor.c
index ac10d1f..6c17754 100644
--- a/src/editor.c
+++ b/src/editor.c
@@ -157,36 +157,48 @@ execute_edit_action (rp_input_line *line, KeySym ch, unsigned int modifier, char
 static edit_status
 editor_forward_char (rp_input_line *line)
 {
-  if (line->position < line->length)
+  if (line->position == line->length)
+    return EDIT_NO_OP;
+
+  if (RP_IS_UTF8_START (line->buffer[line->position]))
     {
-      line->position++;
-      return EDIT_MOVE;
+      do
+        line->position++;
+      while (RP_IS_UTF8_CONT (line->buffer[line->position]));
     }
   else
-    return EDIT_NO_OP;
+    line->position++;
 
+  return EDIT_MOVE;
 }
 
 static edit_status
 editor_backward_char (rp_input_line *line)
 {
-  if (line->position > 0)
-    {
-      line->position--;
-      return EDIT_MOVE;
-    }
-  else
+  if (line->position == 0)
     return EDIT_NO_OP;
+
+  do
+    line->position--;
+  while (line->position > 0 && RP_IS_UTF8_CONT (line->buffer[line->position]));
+
+  return EDIT_MOVE;
 }
 
 static edit_status
 editor_forward_word (rp_input_line *line)
 {
-  if (line->position < line->length)
-    {
-      for (; line->position < line->length && !isalnum (line->buffer[line->position]); line->position++);
-      for (; line->position < line->length && isalnum (line->buffer[line->position]); line->position++);
-    }
+  if (line->position == line->length)
+    return EDIT_NO_OP;
+
+  while (line->position < line->length
+	 && !isalnum (line->buffer[line->position]))
+    line->position++;
+
+  while (line->position < line->length
+	 && (isalnum (line->buffer[line->position])
+	     || RP_IS_UTF8_CHAR (line->buffer[line->position])))
+    line->position++;
 
   return EDIT_MOVE;
 }
@@ -194,11 +206,16 @@ editor_forward_word (rp_input_line *line)
 static edit_status
 editor_backward_word (rp_input_line *line)
 {
-  if (line->position > 0)
-    {
-      for (; line->position > 0 && !isalnum (line->buffer[line->position - 1]); line->position--);
-      for (; line->position > 0 && isalnum (line->buffer[line->position - 1]); line->position--);
-    }
+  if (line->position == 0)
+    return EDIT_NO_OP;
+
+  while (line->position > 0 && !isalnum (line->buffer[line->position]))
+    line->position--;
+
+  while (line->position > 0
+	 && (isalnum (line->buffer[line->position])
+	     || RP_IS_UTF8_CHAR (line->buffer[line->position])))
+    line->position--;
 
   return EDIT_MOVE;
 }
@@ -206,77 +223,100 @@ editor_backward_word (rp_input_line *line)
 static edit_status
 editor_beginning_of_line (rp_input_line *line)
 {
-  if (line->position > 0)
-    line->position = 0;
-
-  return EDIT_MOVE;
+  if (line->position == 0)
+    return EDIT_NO_OP;
+  else
+    {
+      line->position = 0;
+      return EDIT_MOVE;
+    }
 }
 
 static edit_status
 editor_end_of_line (rp_input_line *line)
 {
-  if (line->position < line->length)
-    line->position = line->length;
-
-  return EDIT_MOVE;
+  if (line->position == line->length)
+    return EDIT_NO_OP;
+  else
+    {
+      line->position = line->length;
+      return EDIT_MOVE;
+    }
 }
 
 static edit_status
 editor_delete_char (rp_input_line *line)
 {
-  int i;
+  size_t diff = 0;
 
-  if (line->position < line->length)
-    {
-      for (i = line->position; i < line->length; i++)
-        line->buffer[i] = line->buffer[i + 1];
+  if (line->position == line->length)
+    return EDIT_NO_OP;
 
-      line->length--;
-      return EDIT_DELETE;
+  if (RP_IS_UTF8_START (line->buffer[line->position]))
+    {
+      do
+        diff++;
+      while (RP_IS_UTF8_CONT (line->buffer[line->position + diff]));
     }
   else
-    return EDIT_NO_OP;
+    diff++;
+
+  memmove (&line->buffer[line->position],
+           &line->buffer[line->position + diff],
+           line->length - line->position + diff + 1);
+
+  line->length -= diff;
+
+  return EDIT_DELETE;
 }
 
 static edit_status
 editor_backward_delete_char (rp_input_line *line)
 {
-  int i;
-
-  if (line->position > 0)
-    {
-      for (i = line->position - 1; i < line->length; i++)
-        line->buffer[i] = line->buffer[i + 1];
+  size_t diff = 1;
 
-      line->position--;
-      line->length--;
-      return EDIT_DELETE;
-    }
-  else
+  if (line->position == 0)
     return EDIT_NO_OP;
+
+  while (line->position - diff > 0
+         && RP_IS_UTF8_CONT (line->buffer[line->position - diff]))
+    diff++;
+
+  memmove (&line->buffer[line->position - diff],
+           &line->buffer[line->position],
+           line->length - line->position + 1);
+
+  line->position -= diff;
+  line->length -= diff;
+
+  return EDIT_DELETE;
 }
 
 static edit_status
 editor_kill_word (rp_input_line *line)
 {
-  int i, diff;
+  size_t diff = 0;
 
+  if (line->position == line->length)
+    return EDIT_NO_OP;
 
-  if (line->position < line->length)
-    {
-      for (i = line->position; i < line->length && !isalnum (line->buffer[i]); i++);
-      for (; i < line->length && isalnum (line->buffer[i]); i++);
+  while (line->position + diff < line->length &&
+         !isalnum (line->buffer[line->position + diff]))
+    diff++;
 
-      diff = i - line->position;
+  while (line->position + diff < line->length
+	 && (isalnum (line->buffer[line->position + diff])
+	     || RP_IS_UTF8_CHAR (line->buffer[line->position + diff])))
+    diff++;
 
-      /* Add the word to the X11 selection. */
-      set_nselection (&line->buffer[line->position], diff);
+  /* Add the word to the X11 selection. */
+  set_nselection (&line->buffer[line->position], diff);
 
-      for (i = line->position; i <= line->length - diff; i++)
-        line->buffer[i] = line->buffer[i + diff];
+  memmove (&line->buffer[line->position],
+           &line->buffer[line->position + diff],
+           line->length - line->position + diff + 1);
 
-      line->length -= diff;
-    }
+  line->length -= diff;
 
   return EDIT_DELETE;
 }
@@ -284,25 +324,29 @@ editor_kill_word (rp_input_line *line)
 static edit_status
 editor_backward_kill_word (rp_input_line *line)
 {
-  int i, diff;
+  size_t diff = 1;
 
-  if (line->position > 0)
-    {
-      for (i = line->position; i > 0 && !isalnum (line->buffer[i - 1]); i--);
-      for (; i > 0 && isalnum (line->buffer[i - 1]); i--);
+  if (line->position == 0)
+    return EDIT_NO_OP;
 
-      diff = line->position - i;
+  while (line->position - diff > 0 &&
+         !isalnum (line->buffer[line->position - diff]))
+    diff++;
 
-      line->position = i;
+  while (line->position - diff > 0
+	 && (isalnum (line->buffer[line->position - diff])
+	     || RP_IS_UTF8_CHAR (line->buffer[line->position - diff])))
+    diff++;
 
-      /* Add the word to the X11 selection. */
-      set_nselection (&line->buffer[line->position], diff);
+  /* Add the word to the X11 selection. */
+  set_nselection (&line->buffer[line->position - diff], diff);
 
-      for (; i <= line->length - diff; i++)
-        line->buffer[i] = line->buffer[i + diff];
+  memmove (&line->buffer[line->position - diff],
+           &line->buffer[line->position],
+           line->length - line->position + 1);
 
-      line->length -= diff;
-    }
+  line->position -= diff;
+  line->length -= diff;
 
   return EDIT_DELETE;
 }
@@ -310,14 +354,14 @@ editor_backward_kill_word (rp_input_line *line)
 static edit_status
 editor_kill_line (rp_input_line *line)
 {
-  if (line->position < line->length)
-    {
-      /* Add the line to the X11 selection. */
-      set_selection (&line->buffer[line->position]);
+  if (line->position == line->length)
+    return EDIT_NO_OP;
 
-      line->length = line->position;
-      line->buffer[line->length] = 0;
-    }
+  /* Add the line to the X11 selection. */
+  set_selection (&line->buffer[line->position]);
+
+  line->length = line->position;
+  line->buffer[line->length] = '\0';
 
   return EDIT_DELETE;
 }
@@ -326,14 +370,9 @@ editor_kill_line (rp_input_line *line)
 static void
 backward_kill_line (rp_input_line *line)
 {
-  int i;
-
-  /* If we're at the beginning, we have nothing to do. */
-  if (line->position <= 0)
-    return;
-
-  for (i = line->position; i<= line->length; i++)
-    line->buffer[i - line->position] = line->buffer[i];
+  memmove (&line->buffer[0],
+           &line->buffer[line->position],
+           line->length - line->position + 1);
 
   line->length -= line->position;
   line->position = 0;
@@ -342,7 +381,7 @@ backward_kill_line (rp_input_line *line)
 static edit_status
 editor_backward_kill_line (rp_input_line *line)
 {
-  if (line->position <= 0)
+  if (line->position == 0)
     return EDIT_NO_OP;
 
   /* Add the line to the X11 selection. */
@@ -429,7 +468,7 @@ editor_no_action (rp_input_line *line UNUSED)
 static edit_status
 editor_insert (rp_input_line *line, char *keysym_buf)
 {
-  int nbytes;
+  size_t nbytes;
 
   PRINT_DEBUG (("keysym_buf: '%s'\n", keysym_buf));
 
@@ -441,8 +480,10 @@ editor_insert (rp_input_line *line, char *keysym_buf)
       line->buffer = xrealloc (line->buffer, line->size);
     }
 
-  memmove (&line->buffer[line->position + nbytes], &line->buffer[line->position], line->length - line->position);
-  strncpy (&line->buffer[line->position], keysym_buf, nbytes);
+  memmove (&line->buffer[line->position + nbytes],
+	   &line->buffer[line->position],
+	   line->length - line->position + 1);
+  memcpy (&line->buffer[line->position], keysym_buf, nbytes);
 
   line->length += nbytes;
   line->position += nbytes;
diff --git a/src/editor.h b/src/editor.h
index f3ffa31..b6e4b76 100644
--- a/src/editor.h
+++ b/src/editor.h
@@ -35,6 +35,11 @@ edit_status
   EDIT_NO_OP
 };
 
+/* UTF-8 handling macros */
+#define RP_IS_UTF8_CHAR(c) (defaults.utf8_locale && (c) & 0xC0)
+#define RP_IS_UTF8_START(c) (defaults.utf8_locale && ((c) & 0xC0) == 0xC0)
+#define RP_IS_UTF8_CONT(c) (defaults.utf8_locale && ((c) & 0xC0) == 0x80)
+
 /* Input line functions */
 rp_input_line *input_line_new (char *prompt, char *preinput, int history_id, completion_fn fn);
 void input_line_free (rp_input_line *line);
diff --git a/src/input.c b/src/input.c
index 86866f7..5757bf7 100644
--- a/src/input.c
+++ b/src/input.c
@@ -428,16 +428,23 @@ read_key (KeySym *keysym, unsigned int *modifiers, char *keysym_name, int len)
 static void
 update_input_window (rp_screen *s, rp_input_line *line)
 {
-  int prompt_width = rp_text_width (s, line->prompt, -1);
-  int input_width  = rp_text_width (s, line->buffer, line->length);
-  int   total_width;
+  int prompt_width, input_width, total_width;
+  int char_len = 0, height;
   GC lgc;
   XGCValues gcv;
-  int height;
 
+  prompt_width = rp_text_width (s, line->prompt, -1);
+  input_width  = rp_text_width (s, line->buffer, line->length);
   total_width = defaults.bar_x_padding * 2 + prompt_width + input_width + MAX_FONT_WIDTH (defaults.font);
   height = (FONT_HEIGHT (s) + defaults.bar_y_padding * 2);
 
+  if (RP_IS_UTF8_START (line->buffer[line->position]))
+    do
+      char_len++;
+    while (RP_IS_UTF8_CONT (line->buffer[line->position + char_len]));
+  else
+    char_len = 1;
+
   if (total_width < defaults.input_window_size + prompt_width)
     {
       total_width = defaults.input_window_size + prompt_width;
@@ -466,12 +473,12 @@ update_input_window (rp_screen *s, rp_input_line *line)
   gcv.foreground = s->fg_color ^ s->bg_color;
   lgc = XCreateGC (dpy, s->input_window, GCFunction | GCForeground, &gcv);
 
-  /* Draw a cheap-o cursor - MkII */
+  /* Draw a cheap-o cursor - MkIII */
   XFillRectangle (dpy, s->input_window, lgc,
 		  defaults.bar_x_padding + prompt_width +
 		  rp_text_width (s, line->buffer, line->position),
                   defaults.bar_y_padding,
-		  rp_text_width (s, &line->buffer[line->position], 1),
+		  rp_text_width (s, &line->buffer[line->position], char_len),
                   FONT_HEIGHT (s));
 
   XFlush (dpy);