Use new macro UCS_ORPHAN_CELL for broken double-cell characters.

UCS_ORPHAN_CELL is currently defined as U+0020 SPACE, which was
already used before this macro, so the behaviour does not change,
but the code seems clearer now.

I searched for ' ' and 32 and 0x20 and \x20, and replaced with
UCS_ORPHAN_CELL wherever UCS_NO_CHAR was involved.  However,
some BFU widgets first draw spaces and then overwrite with text;
those will require a more complex fix if UCS_ORPHAN_CELL is ever
changed to some other character.

src/bfu/menu.c

@@ -483,7 +483,7 @@ utf8:
 						  UCS_NO_CHAR, 0, hk_color);
 				} else {
 					draw_char(term, xbase + x - 1, y,
-						  ' ', 0, hk_color);
+						  UCS_ORPHAN_CELL, 0, hk_color);
 				}
 			} else {
 #ifdef CONFIG_DEBUG
@@ -508,7 +508,7 @@ utf8:
 						  y, UCS_NO_CHAR, 0, color);
 				} else {
 					draw_char(term, xbase + x - !!hk_state,
-						  y, ' ', 0, color);
+						  y, UCS_ORPHAN_CELL, 0, color);
 				}
 			} else {
 				draw_char(term, xbase + x - !!hk_state,
@@ -1171,10 +1171,10 @@ display_mainmenu(struct terminal *term, struct menu *menu)
 			/* Is second cell of double-width char on the place where
 			 * first char of the R_MAINMENU_SPACE will be displayed? */
 			if (schar->data == UCS_NO_CHAR) {
-				/* Replace double-width char with ' '. */
+				/* Replace double-width char with UCS_ORPHAN_CELL. */
 				schar++;
 				draw_char_data(term, term->width - R_MAINMENU_SPACE - 1,
-						0, ' ');
+					       0, UCS_ORPHAN_CELL);
 			}
 		}
 #endif

src/intl/charsets.h

@@ -16,6 +16,17 @@ typedef uint32_t unicode_val_T;
  * for the second cell of a double-cell character.  */
 #define UCS_NO_CHAR ((unicode_val_T) 0xFFFFFFFD)
 
+/* If ELinks should display a double-cell character but there is only
+ * one cell available, it displays this character instead.  This must
+ * be a single-cell character but need not be unique.  Possible values
+ * might be U+0020 SPACE or U+303F IDEOGRAPHIC HALF FILL SPACE.
+ *
+ * Some BFU widgets (at least input fields and list boxes) currently
+ * ignore this setting and use U+0020 instead.  (They first draw spaces
+ * and then overwrite with text; look for utf8_cells2bytes calls.)
+ * We should fix that if we ever change the value.  */
+#define UCS_ORPHAN_CELL ((unicode_val_T) 0x20)
+
 /*   replacement character. See u2cp(). */
 #define NBSP_CHAR ((unsigned char) 1)
 #define NBSP_CHAR_STRING "\001"

src/terminal/draw.c

@@ -160,7 +160,7 @@ draw_line(struct terminal *term, int x, int y, int l, struct screen_char *line)
 
 			sc = line;
 			data_save = sc->data;
-			sc->data = ' ';
+			sc->data = UCS_ORPHAN_CELL;
 			copy_screen_chars(screen_char, line, 1);
 			sc->data = data_save;
 			size--;
@@ -169,13 +169,13 @@ draw_line(struct terminal *term, int x, int y, int l, struct screen_char *line)
 
 		}
 		/* Instead of displaying double-width character at last column
-		 * display only space. */
+		 * display only UCS_ORPHAN_CELL. */
 		if (size - 1 > 0 && unicode_to_cell(line[size - 1].data) == 2) {
 			unicode_val_T data_save;
 
 			sc = &line[size - 1];
 			data_save = sc->data;
-			sc->data = ' ';
+			sc->data = UCS_ORPHAN_CELL;
 			copy_screen_chars(screen_char, line, size);
 			sc->data = data_save;
 		} else {
@@ -256,7 +256,7 @@ draw_border(struct terminal *term, struct box *box,
 
 #ifdef CONFIG_UTF8
 /* Checks cells left and right to the box for broken double-width chars.
- * Replace it with ' '.
+ * Replace it with UCS_ORPHAN_CELL.
  * 1+---+3
  * 1|box|##4
  * 1|   |##4
@@ -284,7 +284,7 @@ fix_dwchar_around_box(struct terminal *term, struct box *box, int border,
 		schar = get_char(term, x, y);
 		for (;height--; schar += term->width)
 			if (unicode_to_cell(schar->data) == 2)
-				schar->data = ' ';
+				schar->data = UCS_ORPHAN_CELL;
 	}
 
 	/* 2 */
@@ -296,7 +296,7 @@ fix_dwchar_around_box(struct terminal *term, struct box *box, int border,
 		schar = get_char(term, x, y);
 		for (;height--; schar += term->width)
 			if (unicode_to_cell(schar->data) == 2)
-				schar->data = ' ';
+				schar->data = UCS_ORPHAN_CELL;
 	}
 
 	/* 3 */
@@ -308,7 +308,7 @@ fix_dwchar_around_box(struct terminal *term, struct box *box, int border,
 		schar = get_char(term, x, y);
 		for (;height--; schar += term->width)
 			if (schar->data == UCS_NO_CHAR)
-				schar->data = ' ';
+				schar->data = UCS_ORPHAN_CELL;
 	}
 
 	/* 4 */
@@ -320,7 +320,7 @@ fix_dwchar_around_box(struct terminal *term, struct box *box, int border,
 		schar = get_char(term, x, y);
 		for (;height--; schar += term->width)
 			if (schar->data == UCS_NO_CHAR)
-				schar->data = ' ';
+				schar->data = UCS_ORPHAN_CELL;
 	}
 }
 #endif
@@ -437,7 +437,7 @@ draw_text_utf8(struct terminal *term, int x, int y,
 	}
 
 	if (start->data == UCS_NO_CHAR && x - 1 > 0)
-		draw_char_data(term, x - 1, y, ' ');
+		draw_char_data(term, x - 1, y, UCS_ORPHAN_CELL);
 
 	pos = start;
 
@@ -451,7 +451,7 @@ draw_text_utf8(struct terminal *term, int x, int y,
 			pos->data = UCS_NO_CHAR;
 			pos->attr = 0;
 		} else {
-			pos->data = (unicode_val_T)' ';
+			pos->data = UCS_ORPHAN_CELL;
 		}
 	} else {
 		pos->data = data;
@@ -474,7 +474,7 @@ draw_text_utf8(struct terminal *term, int x, int y,
 				pos->data = UCS_NO_CHAR;
 				pos->attr = 0;
 			} else {
-				pos->data = (unicode_val_T)' ';
+				pos->data = UCS_ORPHAN_CELL;
 			}
 		} else {
 			pos->data = data;

src/viewer/text/form.c

@@ -412,13 +412,14 @@ retry_after_scroll:
 				}
 
 				/* The character does not fit completely.
-				 * Write spaces to the cells that do fit.  */
+				 * Write UCS_ORPHAN_CELL to the cells that
+				 * do fit.  */
 				for (cell = 0; cell < cells; cell++) {
 					if (col_is_in_box(box, x + i + cell)
 					    && i + cell < fc->size)
 						draw_char_data(term,
-							       x + i + cell,
+							       x + i + cell, y,
-							       y, ' ');
+							       UCS_ORPHAN_CELL);
 				}
 
 drew_char:
@@ -538,7 +539,7 @@ utf8_select:
 							data = UCS_NO_CHAR;
 							i++;
 						} else if (cell == 2) {
-							data = ' ';
+							data = UCS_ORPHAN_CELL;
 						}
 					} else
 						data = '_';