Merge with /srv/git/elinks.git

AUTHORS

@@ -331,6 +331,9 @@ Len Lattanzi <Len_Lattanzi@StanfordAlumni.org>
 M. K. Srikant <srix@vsnl.com>
 	Small fix in forms
 
+M. Levinson <levinsm@users.sourceforge.net>
+	Python scripting fixes
+
 Marco Bodrato <bodrato@linuz.sns.it>
 	Twinterm support
 
@@ -420,6 +423,9 @@ Omar Khayam <omark@cyentec.com>
 <otte@duke.edu>
 	Fix stdin reading on Mac OS X
 
+Pavol Babincak <scroolik@gmail.com>
+	Improved UTF-8 support with double-width chars
+
 Peder Stray <peder@ifi.uio.no>
 	Fix handling of key presses turning up as key prefixes
 

Makefile.config.in

@@ -156,6 +156,7 @@ CONFIG_SPIDERMONKEY = @CONFIG_SPIDERMONKEY@
 CONFIG_SSL = @CONFIG_SSL@
 CONFIG_SYSMOUSE = @CONFIG_SYSMOUSE@
 CONFIG_URI_REWRITE = @CONFIG_URI_REWRITE@
+CONFIG_UTF_8 = @CONFIG_UTF_8@
 CONFIG_XBEL_BOOKMARKS = @CONFIG_XBEL_BOOKMARKS@
 CONFIG_XMLTO = @CONFIG_XMLTO@
 CONFIG_GSSAPI = @CONFIG_GSSAPI@

config/m4/features.m4

@@ -19,7 +19,7 @@ AC_DEFUN([EL_LOG_CONFIG],
 	[msgdots2="`echo $about | sed 's/[0-9]/./g'`"]
 	[msgdots1="`echo $msgdots2 | sed 's/[a-z]/./g'`"]
 	[msgdots0="`echo $msgdots1 | sed 's/[A-Z]/./g'`"]
-	[msgdots="`echo $msgdots0 | sed 's/[_ ()]/./g'`"]
+	[msgdots="`echo $msgdots0 | sed 's/[-_ ()]/./g'`"]
 	DOTS="................................"
 	dots=`echo $DOTS | sed "s/$msgdots//"`
 

configure.in

@@ -168,6 +168,7 @@ AC_HEADER_STDC
 AC_HEADER_SYS_WAIT
 AC_HEADER_TIME
 
+AC_CHECK_HEADERS(wctype.h)
 AC_CHECK_HEADERS(fcntl.h limits.h time.h unistd.h)
 AC_CHECK_HEADERS(sigaction.h)
 AC_CHECK_HEADERS(arpa/inet.h)
@@ -447,7 +448,7 @@ EL_CONFIG_OPTIONAL_LIBRARY(CONFIG_IDN, idn, idna.h, idn, stringprep_check_versio
 
 if test "x{with_gc}" != xno; then
 	EL_CONFIG_OPTIONAL_LIBRARY(CONFIG_GC, gc, gc.h, gc, GC_init,
-		[  --with-gc            enable Boehm's garbage collector])
+		[  --with-gc               enable Boehm's garbage collector])
 fi
 
 EL_ARG_ENABLE(CONFIG_LZMA, lzma, [lzma],
@@ -770,50 +771,32 @@ dnl Check for Python
 dnl ===================================================================
 enable_python="no";
 
-AC_ARG_WITH(python, [  --with-python=[prefix]       enable Python support],
+AC_ARG_WITH(python, [  --with-python=[DIR]     enable Python support],
-            [
+            [ if test "x$withval" != xno; then enable_python=yes; fi ])
-if test "$withval" != no; then
-	# FIXME: If withval is a valid directory append it to PATH
-	# so that you can specify one of several Python installations.
-	if test "$withval" != yes; then
-		python_prefix="$withval"
-	else
-		python_prefix=""
-	fi
-	enable_python=yes
-	cat <<EOF
-***********************************************************************
-The Python support is incomplete and not so well integrated to ELinks
-yet. That means, e.g.., that you have no Python console and there might
-not be all the necessary hooks. Also, the Python interface is not too
-well tested (success stories heartily welcomed!).
-***********************************************************************
-EOF
-fi
-	    ])
-
-
-cf_result=no
 
 EL_SAVE_FLAGS
+cf_result=no
+
+AC_MSG_CHECKING([for Python])
 
 if test "$enable_python" = "yes"; then
-	if test -n "$python_prefix" && test -d "$python_prefix/bin"; then
+	AC_MSG_RESULT(yes);
-		PYTHON_PATH="$python_prefix/bin:$PATH"
+
+	if test -d "$withval"; then
+		PYTHON_PATH="$withval:$PATH"
 	else
 		PYTHON_PATH="$PATH"
 	fi
+
 	AC_PATH_PROG(PYTHON, python, no, $PYTHON_PATH)
-	if test "$PYTHON" = "no" ; then
+
-		cf_result=no
+	if test "$PYTHON" != no; then
-	else
+		cf_result="yes";
-		PYTHON_CFLAGS="-I`$PYTHON -c 'from distutils import sysconfig; print sysconfig.get_python_inc()' 2> /dev/null`"
+
-		PYTHON_LIBS="-lpython`$PYTHON -c 'from distutils import sysconfig; print sysconfig.get_config_var("VERSION")' 2> /dev/null`"
+		PYTHON_CFLAGS="`$PYTHON -c 'from distutils import sysconfig; print "-I%s -I%s" % (sysconfig.get_python_inc(), sysconfig.get_python_inc(plat_specific=True))'`"
-		if test -n "$python_prefix" && test -d "$python_prefix/lib"; then
+		PYTHON_LIBS="`$PYTHON -c 'from distutils import sysconfig; var = sysconfig.get_config_var; print "%s %s %s -L%s -lpython%s" % (var("LINKFORSHARED"), var("LIBS"), var("SYSLIBS"), var("LIBPL"), var("VERSION"))'`"
-			PYTHON_LIBS="-L$python_prefix/lib $PYTHON_LIBS"
-		fi
 		LIBS="$PYTHON_LIBS $LIBS"
-		CFLAGS="$PYTHON_CFLAGS $CFLAGS"
+		CPPFLAGS="$CPPFLAGS $PYTHON_CFLAGS"
 		AC_TRY_LINK([#include <Python.h>],
 		    [Py_Initialize();],
 		    cf_result=yes, cf_result=no)
@@ -822,17 +805,28 @@ if test "$enable_python" = "yes"; then
 			EL_RESTORE_FLAGS
 		else
 			EL_CONFIG(CONFIG_SCRIPTING_PYTHON, [Python])
-
-			CFLAGS="$CFLAGS_X"
 			AC_SUBST(PYTHON_LIBS)
 			AC_SUBST(PYTHON_CFLAGS)
+			cat <<EOF
+***********************************************************************
+The Python support is incomplete and not so well integrated to ELinks
+yet. That means, e.g.., that you have no Python console and there might
+not be all the necessary hooks. Also, the Python interface is not too
+well tested (success stories heartily welcomed!).
+***********************************************************************
+EOF
+		fi
+	else
+		if test -n "$withval" && test "x$withval" != xno; then
+			AC_MSG_ERROR([Python not found])
+		else
+			AC_MSG_WARN([Python support disabled])
 		fi
 	fi
+else
+	AC_MSG_RESULT(no);
 fi
 
-AC_MSG_CHECKING([for Python])
-if test "$cf_result"; then AC_MSG_RESULT($cf_result); fi
-
 
 dnl ===================================================================
 dnl Check for Lua, optional even if installed.
@@ -1283,6 +1277,10 @@ EL_ARG_ENABLE(CONFIG_OWN_LIBC, own-libc, [Own libc stubs],
 EL_ARG_ENABLE(CONFIG_SMALL, small, [Small binary],
 	      [  --enable-small          reduce binary size as far as possible (but see the bottom of doc/small.txt!)])
 
+EL_ARG_ENABLE(CONFIG_UTF_8, utf-8, [UTF-8],
+	      [  --enable-utf-8          enable UTF-8 support])
+
+
 AC_ARG_ENABLE(weehoofooboomookerchoo,
               [
     Also check out the features.conf file for more information about features!
@@ -1323,6 +1321,10 @@ AC_SUBST(CONFDIR)
 
 # Create LOCALEDIR #define for config.h
 LOCALEDIR=`eval echo "$datadir/locale"`
+while echo "$LOCALEDIR" | grep "\\$"
+do
+	LOCALEDIR=`eval echo "$LOCALEDIR"`
+done > /dev/null 2> /dev/null
 AC_DEFINE_UNQUOTED(LOCALEDIR, "$LOCALEDIR", [Directory containing locales])
 AC_SUBST(LOCALEDIR)
 

contrib/python/README.Python

@@ -1,13 +1,8 @@
-If you want to use Python scripting with ELinks add
+If you want to use Python scripting with ELinks, add --with-python to the
---with-python to the configure invocation copy hooks.py to ~/.elinks
+configure invocation and copy hooks.py to your ~/.elinks directory.
-When your Python installation is your own build, you could give prefix
-to the configure, eg.
---with-python=/usr/local when Python binary is placed in /usr/local/bin, etc.
 
-When 'configure' cannot find -lpython make symbolic link to the appropriate
+If configure cannot find Python you can supply a path, e.g.
-library, eg.
+--with-python=/usr/local/bin if your Python binary is in /usr/local/bin, etc.
-# cd /usr/local/lib
-# ln -s libpython2.4.so.1.0 libpython.so
 
 For the present hooks.py is not very usable. You are welcome to make it better.
 Good Luck!

features.conf

@@ -601,6 +601,27 @@ CONFIG_OWN_LIBC=no
 CONFIG_SMALL=no
 
 
+### Unicode UTF-8 support
+#
+# By enabling this option you get better Unicode support. At present only some
+# parts of ELinks are influenced with this. It includes DOM, plain, HTML
+# renderer and user interface. Beside normal Unicode characters there is
+# support for double-width characters (like Japanese, etc.).
+#
+# Some features of Unicode are not handled at all. Combining characters is
+# most visible absence.
+# Some features are partially supported. Like line breaking between
+# double-width characters. There is no other detection for determining when to
+# break or not.
+#
+# Note: This UTF-8 support is experimental.
+#
+# Default: disabled
+
+CONFIG_UTF_8=no
+
+
+
 ### Back-trace Printing
 #
 # Once upon a time, a disaster happens and ELinks crashes. That is a very sad

po/pl.po

@@ -2802,7 +2802,7 @@ msgstr "Rozr
 msgid ""
 "Whether the search should match the document text while maintaining\n"
 "case sensitivity."
-msgstr "Czy zaznaczać tekst różniący się wielkością liter od poszukiwanego?"
+msgstr "Czy rozróżniać wielkość liter przy szukaniu."
 
 #: src/config/options.inc:393
 msgid "Regular expressions"

src/bfu/button.c

@@ -77,18 +77,35 @@ add_dlg_button_do(struct dialog *dlg, unsigned char *text, int flags,
 	}
 }
 
+#ifdef CONFIG_UTF_8
+static void
+buttons_width(struct widget_data *widget_data, int n,
+	      int *minwidth, int *maxwidth, int utf8)
+#else
 static void
 buttons_width(struct widget_data *widget_data, int n,
 	      int *minwidth, int *maxwidth)
+#endif /* CONFIG_UTF_8 */
 {
 	int maxw = -BUTTON_HSPACING;
+#ifdef CONFIG_UTF_8
+	int button_lr_len = utf8_ptr2cells(BUTTON_LEFT, NULL)
+			  + utf8_ptr2cells(BUTTON_RIGHT, NULL);
+#endif /* CONFIG_UTF_8 */
 
 	assert(n > 0);
 	if_assert_failed return;
 
 	while (n--) {
-		int minw = (widget_data++)->widget->info.button.textlen
+		int minw;
-			   + BUTTON_HSPACING + BUTTON_LR_LEN;
+#ifdef CONFIG_UTF_8
+		if (utf8)
+			minw = utf8_ptr2cells((widget_data++)->widget->text, NULL)
+			       + BUTTON_HSPACING + button_lr_len;
+		else
+#endif /* CONFIG_UTF_8 */
+			minw = (widget_data++)->widget->info.button.textlen
+				+ BUTTON_HSPACING + BUTTON_LR_LEN;
 
 		maxw += minw;
 		if (minwidth) int_lower_bound(minwidth, minw);
@@ -100,7 +117,7 @@ buttons_width(struct widget_data *widget_data, int n,
 void
 dlg_format_buttons(struct terminal *term,
 		   struct widget_data *widget_data, int n,
-		   int x, int *y, int w, int *rw, enum format_align align)
+		   int x, int *y, int w, int *rw, enum format_align align, int format_only)
 {
 	int i1 = 0;
 
@@ -111,24 +128,45 @@ dlg_format_buttons(struct terminal *term,
 
 		while (i2 < n) {
 			mw = 0;
+#ifdef CONFIG_UTF_8
+			buttons_width(widget_data1, i2 - i1 + 1, NULL, &mw,
+				      term->utf8);
+#else
 			buttons_width(widget_data1, i2 - i1 + 1, NULL, &mw);
+#endif /* CONFIG_UTF_8 */
 			if (mw <= w) i2++;
 			else break;
 		}
 
 		mw = 0;
+#ifdef CONFIG_UTF_8
+		buttons_width(widget_data1, i2 - i1, NULL, &mw, term->utf8);
+#else
 		buttons_width(widget_data1, i2 - i1, NULL, &mw);
+#endif /* CONFIG_UTF_8 */
 		if (rw) int_bounds(rw, mw, w);
 
-		if (term) {
+		if (!format_only) {
 			int i;
 			int p = x + (align == ALIGN_CENTER ? (w - mw) / 2 : 0);
+#ifdef CONFIG_UTF_8
+			int button_lr_len = utf8_ptr2cells(BUTTON_LEFT, NULL)
+					  + utf8_ptr2cells(BUTTON_RIGHT, NULL);
+#endif /* CONFIG_UTF_8 */
 
 			for (i = i1; i < i2; i++) {
-				set_box(&widget_data[i].box,
+#ifdef CONFIG_UTF_8
-					 p, *y,
+				if (term->utf8)
-					 widget_data[i].widget->info.button.textlen
+					set_box(&widget_data[i].box,
-					 + BUTTON_LR_LEN, BUTTON_HEIGHT);
+						p, *y,
+						utf8_ptr2cells(widget_data[i].widget->text, NULL)
+						+ button_lr_len, BUTTON_HEIGHT);
+				else
+#endif /* CONFIG_UTF_8 */
+					set_box(&widget_data[i].box,
+						p, *y,
+						widget_data[i].widget->info.button.textlen
+						+ BUTTON_LR_LEN, BUTTON_HEIGHT);
 
 				p += widget_data[i].box.width + BUTTON_HSPACING;
 			}
@@ -145,8 +183,7 @@ display_button(struct dialog_data *dlg_data, struct widget_data *widget_data)
 	struct terminal *term = dlg_data->win->term;
 	struct color_pair *color, *shortcut_color;
 	struct box *pos = &widget_data->box;
-	int len = widget_data->box.width - BUTTON_LR_LEN;
+	int len, x;
-	int x = pos->x + BUTTON_LEFT_LEN;
 	int sel = is_selected_widget(dlg_data, widget_data);
 
 	if (sel) {
@@ -158,6 +195,23 @@ display_button(struct dialog_data *dlg_data, struct widget_data *widget_data)
 	}
 	if (!color || !shortcut_color) return EVENT_PROCESSED;
 
+#ifdef CONFIG_UTF_8
+	if (term->utf8) {
+		int button_left_len = utf8_ptr2cells(BUTTON_LEFT, NULL);
+		int button_right_len = utf8_ptr2cells(BUTTON_RIGHT, NULL);
+
+		x = pos->x + button_left_len;
+		len = widget_data->box.width -
+			(button_left_len + button_right_len);
+
+	} else
+#endif /* CONFIG_UTF_8 */
+	{
+		x = pos->x + BUTTON_LEFT_LEN;
+		len = widget_data->box.width - BUTTON_LR_LEN;
+	}
+
+
 	draw_text(term, pos->x, pos->y, BUTTON_LEFT, BUTTON_LEFT_LEN, 0, color);
 	if (len > 0) {
 		unsigned char *text = widget_data->widget->text;
@@ -167,6 +221,52 @@ display_button(struct dialog_data *dlg_data, struct widget_data *widget_data)
 		attr = get_opt_bool("ui.dialogs.underline_button_shortcuts")
 		     ? SCREEN_ATTR_UNDERLINE : 0;
 
+#ifdef CONFIG_UTF_8
+		if (term->utf8) {
+			if (hk_pos >= 0) {
+				int hk_bytes = utf8charlen(&text[hk_pos+1]);
+				int cells_to_hk = utf8_ptr2cells(text,
+						&text[hk_pos]);
+				int right = widget_data->widget->info.button.truetextlen
+					- hk_pos
+					- hk_bytes;
+
+				int hk_cells = utf8_char2cells(&text[hk_pos
+								      + 1],
+								NULL);
+
+				if (hk_pos)
+					draw_text(term, x, pos->y,
+						  text, hk_pos, 0, color);
+
+				draw_text(term, x + cells_to_hk, pos->y,
+					  &text[hk_pos + 1], hk_bytes,
+					  attr, shortcut_color);
+
+				if (right > 1)
+					draw_text(term, x+cells_to_hk+hk_cells,
+						  pos->y,
+						  &text[hk_pos + hk_bytes + 1],
+						  right - 1, 0, color);
+
+			} else {
+				int hk_width = utf8_char2cells(text, NULL);
+				int hk_len = utf8charlen(text);
+				int len_to_display =
+					utf8_cells2bytes(&text[hk_len],
+							 len - hk_width,
+							 NULL);
+
+				draw_text(term, x, pos->y,
+					  text, hk_len,
+					  attr, shortcut_color);
+
+				draw_text(term, x + hk_width, pos->y,
+					  &text[hk_len], len_to_display,
+					  0, color);
+			}
+		} else
+#endif /* CONFIG_UTF_8 */
 		if (hk_pos >= 0) {
 			int right = widget_data->widget->info.button.truetextlen - hk_pos - 1;
 
@@ -185,8 +285,17 @@ display_button(struct dialog_data *dlg_data, struct widget_data *widget_data)
 			draw_text(term, x + 1, pos->y, &text[1], len - 1, 0, color);
 		}
 	}
-	draw_text(term, x + len, pos->y, BUTTON_RIGHT, BUTTON_RIGHT_LEN, 0, color);
+#ifdef CONFIG_UTF_8
+	if (term->utf8) {
+		int text_cells = utf8_ptr2cells(widget_data->widget->text, NULL);
+		int hk = (widget_data->widget->info.button.hotkey_pos >= 0);
 
+		draw_text(term, x + text_cells - hk, pos->y,
+			  BUTTON_RIGHT, BUTTON_RIGHT_LEN, 0, color);
+	} else
+#endif /* CONFIG_UTF_8 */
+		draw_text(term, x + len, pos->y, BUTTON_RIGHT,
+			  BUTTON_RIGHT_LEN, 0, color);
 	if (sel) {
 		set_cursor(term, x, pos->y, 1);
 		set_window_ptr(dlg_data->win, pos->x, pos->y);

src/bfu/button.h

@@ -48,6 +48,6 @@ void add_dlg_button_do(struct dialog *dlg, unsigned char *text, int flags, widge
 #endif
 
 extern struct widget_ops button_ops;
-void dlg_format_buttons(struct terminal *, struct widget_data *, int, int, int *, int, int *, enum format_align);
+void dlg_format_buttons(struct terminal *, struct widget_data *, int, int, int *, int, int *, enum format_align, int);
 
 #endif

src/bfu/checkbox.c

@@ -39,7 +39,7 @@ void
 dlg_format_checkbox(struct terminal *term,
 		    struct widget_data *widget_data,
 		    int x, int *y, int w, int *rw,
-		    enum format_align align)
+		    enum format_align align, int format_only)
 {
 	unsigned char *text = widget_data->widget->text;
 
@@ -52,7 +52,7 @@ dlg_format_checkbox(struct terminal *term,
 		dlg_format_text_do(term, text, x + CHECKBOX_LS, y,
 				   w - CHECKBOX_LS, rw,
 				   get_bfu_color(term, "dialog.checkbox-label"),
-				   align);
+				   align, format_only);
 		if (rw) *rw += CHECKBOX_LS;
 	}
 }

src/bfu/checkbox.h

@@ -33,7 +33,7 @@ void
 dlg_format_checkbox(struct terminal *term,
 		    struct widget_data *widget_data,
 		    int x, int *y, int w, int *rw,
-		    enum format_align align);
+		    enum format_align align, int format_only);
 
 #define widget_has_group(widget_data)	((widget_data)->widget->type == WIDGET_CHECKBOX \
 					  ? (widget_data)->widget->info.checkbox.gid : -1)

src/bfu/dialog.c

@@ -12,6 +12,7 @@
 #include "bfu/dialog.h"
 #include "config/kbdbind.h"
 #include "config/options.h"
+#include "intl/charsets.h"
 #include "intl/gettext/libintl.h"
 #include "terminal/draw.h"
 #include "main/timer.h"
@@ -96,13 +97,32 @@ redraw_dialog(struct dialog_data *dlg_data, int layout)
 		title_color = get_bfu_color(term, "dialog.title");
 		if (title_color && box.width > 2) {
 			unsigned char *title = dlg_data->dlg->title;
-			int titlelen = int_min(box.width - 2, strlen(title));
+			int titlelen = strlen(title);
-			int x = (box.width - titlelen) / 2 + box.x;
+			int titlecells = titlelen;
-			int y = box.y - 1;
+			int x, y;
+
+#ifdef CONFIG_UTF_8
+			if (term->utf8)
+				titlecells = utf8_ptr2cells(title,
+							    &title[titlelen]);
+#endif /* CONFIG_UTF_8 */
+
+			titlecells = int_min(box.width - 2, titlecells);
+
+#ifdef CONFIG_UTF_8
+			if (term->utf8)
+				titlelen = utf8_cells2bytes(title, titlecells,
+							    NULL);
+#endif /* CONFIG_UTF_8 */
+
+			x = (box.width - titlecells) / 2 + box.x;
+			y = box.y - 1;
+
 
 			draw_text(term, x - 1, y, " ", 1, 0, title_color);
 			draw_text(term, x, y, title, titlelen, 0, title_color);
-			draw_text(term, x + titlelen, y, " ", 1, 0, title_color);
+			draw_text(term, x + titlecells, y, " ", 1, 0,
+				  title_color);
 		}
 	}
 
@@ -499,7 +519,7 @@ clear_dialog(struct dialog_data *dlg_data, struct widget_data *xxx)
 
 static void
 format_widgets(struct terminal *term, struct dialog_data *dlg_data,
-	       int x, int *y, int w, int h, int *rw)
+	       int x, int *y, int w, int h, int *rw, int format_only)
 {
 	struct widget_data *wdata = dlg_data->widgets_data;
 	int widgets = dlg_data->number_of_widgets;
@@ -509,15 +529,18 @@ format_widgets(struct terminal *term, struct dialog_data *dlg_data,
 		switch (wdata->widget->type) {
 		case WIDGET_FIELD_PASS:
 		case WIDGET_FIELD:
-			dlg_format_field(term, wdata, x, y, w, rw, ALIGN_LEFT);
+			dlg_format_field(term, wdata, x, y, w, rw, ALIGN_LEFT,
+					 format_only);
 			break;
 
 		case WIDGET_LISTBOX:
-			dlg_format_listbox(term, wdata, x, y, w, h, rw, ALIGN_LEFT);
+			dlg_format_listbox(term, wdata, x, y, w, h, rw,
+					   ALIGN_LEFT, format_only);
 			break;
 
 		case WIDGET_TEXT:
-			dlg_format_text(term, wdata, x, y, w, rw, h);
+			dlg_format_text(term, wdata, x, y, w, rw, h,
+					format_only);
 			break;
 
 		case WIDGET_CHECKBOX:
@@ -535,14 +558,16 @@ format_widgets(struct terminal *term, struct dialog_data *dlg_data,
 						break;
 				}
 
-				dlg_format_group(term, wdata, size, x, y, w, rw);
+				dlg_format_group(term, wdata, size, x, y, w, rw,
+						 format_only);
 				wdata += size - 1;
 
 			} else {
 
 				/* No horizontal space between checkboxes belonging to
 				 * the same group. */
-				dlg_format_checkbox(term, wdata, x, y, w, rw, ALIGN_LEFT);
+				dlg_format_checkbox(term, wdata, x, y, w, rw,
+						    ALIGN_LEFT, format_only);
 				if (widgets > 1
 				    && group == widget_has_group(&wdata[1]))
 					(*y)--;
@@ -554,7 +579,7 @@ format_widgets(struct terminal *term, struct dialog_data *dlg_data,
 		 * of the dialog. */
 		case WIDGET_BUTTON:
 			dlg_format_buttons(term, wdata, widgets,
-					   x, y, w, rw, ALIGN_CENTER);
+					   x, y, w, rw, ALIGN_CENTER, format_only);
 			return;
 		}
 	}
@@ -566,11 +591,17 @@ generic_dialog_layouter(struct dialog_data *dlg_data)
 	struct terminal *term = dlg_data->win->term;
 	int w = dialog_max_width(term);
 	int height = dialog_max_height(term);
-	int rw = int_min(w, strlen(dlg_data->dlg->title));
+	int x = 0, y, rw;
-	int y = dlg_data->dlg->layout.padding_top ? 0 : -1;
-	int x = 0;
 
-	format_widgets(NULL, dlg_data, x, &y, w, height, &rw);
+#ifdef CONFIG_UTF_8
+	if (term->utf8)
+		rw = int_min(w, utf8_ptr2cells(dlg_data->dlg->title, NULL));
+	else
+#endif /* CONFIG_UTF_8 */
+		rw = int_min(w, strlen(dlg_data->dlg->title));
+	y = dlg_data->dlg->layout.padding_top ? 0 : -1;
+
+	format_widgets(term, dlg_data, x, &y, w, height, &rw, 1);
 
 	/* Update the width to respond to the required minimum width */
 	if (dlg_data->dlg->layout.fit_datalen) {
@@ -585,7 +616,7 @@ generic_dialog_layouter(struct dialog_data *dlg_data)
 	y = dlg_data->box.y + DIALOG_TB + dlg_data->dlg->layout.padding_top;
 	x = dlg_data->box.x + DIALOG_LB;
 
-	format_widgets(term, dlg_data, x, &y, w, height, NULL);
+	format_widgets(term, dlg_data, x, &y, w, height, NULL, 0);
 }
 
 
@@ -607,7 +638,15 @@ draw_dialog(struct dialog_data *dlg_data, int width, int height)
 		/* Draw shadow */
 		draw_shadow(term, &dlg_data->box,
 			    get_bfu_color(term, "dialog.shadow"), 2, 1);
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			fix_dwchar_around_box(term, &dlg_data->box, 0, 2, 1);
+#endif /* CONFIG_UTF_8 */
 	}
+#ifdef CONFIG_UTF_8
+	else if(term->utf8)
+		fix_dwchar_around_box(term, &dlg_data->box, 0, 0, 0);
+#endif /* CONFIG_UTF_8 */
 }
 
 static void

src/bfu/group.c

@@ -16,10 +16,13 @@
 #include "terminal/terminal.h"
 #include "util/color.h"
 
+/* Same as in src/bfu/checkbox.c */
+#define CHECKBOX_LEN 3  /* "[X]" or "(X)" */
+
 void
 dlg_format_group(struct terminal *term,
 		 struct widget_data *widget_data,
-		 int n, int x, int *y, int w, int *rw)
+		 int n, int x, int *y, int w, int *rw, int format_only)
 {
 	int space_between_widgets = 1;
 	int line_width = 0;
@@ -33,13 +36,31 @@ dlg_format_group(struct terminal *term,
 		int widget_width;
 		int width;
 		unsigned char *text = widget_data->widget->text;
-		int label_length = (text && *text) ? strlen(text) : 0;
+		int label_length;
-		int label_padding = (label_length > 0);
+		int label_padding;
+
+#ifdef CONFIG_UTF_8
+		if (term->utf8) {
+			if (text && *text)
+				label_length = utf8_ptr2cells(text, NULL);
+			else
+				label_length = 0;
+		} else
+#endif /* CONFIG_UTF_8 */
+			label_length = (text && *text) ? strlen(text) : 0;
+
+		label_padding = (label_length > 0);
 
 		if (widget_data->widget->type == WIDGET_CHECKBOX) {
-			width = 3;
+			width = CHECKBOX_LEN;
 		} else if (widget_is_textfield(widget_data)) {
-			width = widget_data->widget->datalen;
+#ifdef CONFIG_UTF_8
+			if (term->utf8) {
+				width = utf8_ptr2cells(widget_data->widget->data,
+						       NULL);
+			} else
+#endif /* CONFIG_UTF_8 */
+				width = widget_data->widget->datalen;
 		} else {
 			/* TODO: handle all widget types. */
 			widget_data++;
@@ -56,22 +77,52 @@ dlg_format_group(struct terminal *term,
 
 		xpos = x + line_width;
 
-		if (term) {
+		if (!format_only) {
 			if (widget_data->widget->type == WIDGET_CHECKBOX) {
 				/* Draw text at right of checkbox. */
-				if (label_length)
+				if (label_length) {
-					draw_text(term, xpos + width + label_padding, *y,
+#ifdef CONFIG_UTF_8
-						  text, label_length,
+					if (term->utf8) {
-						  0, color);
+						int lb = utf8_cells2bytes(
+								text,
+								label_length,
+								NULL);
+						draw_text(term, xpos + width
+								+ label_padding,
+							  *y, text, lb, 0,
+							  color);
+					} else
+#endif /* CONFIG_UTF_8 */
+					{
+						draw_text(term, xpos + width
+								+ label_padding,
+							  *y, text,
+							  label_length, 0,
+							  color);
+					}
+				}
 
 				set_box(&widget_data->box, xpos, *y, width, 1);
 
 			} else if (widget_is_textfield(widget_data)) {
 				/* Draw label at left of widget. */
-				if (label_length)
+				if (label_length) {
-					draw_text(term, xpos, *y,
+#ifdef CONFIG_UTF_8
-						  text, label_length,
+					if (term->utf8) {
-						  0, color);
+						int lb = utf8_cells2bytes(
+								text,
+								label_length,
+								NULL);
+						draw_text(term, xpos, *y,
+							  text, lb, 0, color);
+					} else
+#endif /* CONFIG_UTF_8 */
+					{
+						draw_text(term, xpos, *y,
+							  text, label_length,
+							  0, color);
+					}
+				}
 
 				set_box(&widget_data->box,
 					xpos + label_padding + label_length, *y,
@@ -93,16 +144,23 @@ group_layouter(struct dialog_data *dlg_data)
 {
 	struct terminal *term = dlg_data->win->term;
 	int w = dialog_max_width(term);
-	int rw = int_min(w, strlen(dlg_data->dlg->title));
+	int rw;
 	int y = 0;
 	int n = dlg_data->number_of_widgets - 2;
 
-	dlg_format_group(NULL, dlg_data->widgets_data, n,
+#ifdef CONFIG_UTF_8
-			 0, &y, w, &rw);
+	if (term->utf8)
+		rw = int_min(w, utf8_ptr2cells(dlg_data->dlg->title, NULL));
+	else
+#endif /* CONFIG_UTF_8 */
+		rw = int_min(w, strlen(dlg_data->dlg->title));
+
+	dlg_format_group(term, dlg_data->widgets_data, n,
+			 0, &y, w, &rw, 1);
 
 	y++;
-	dlg_format_buttons(NULL, dlg_data->widgets_data + n, 2, 0, &y, w,
+	dlg_format_buttons(term, dlg_data->widgets_data + n, 2, 0, &y, w,
-			   &rw, ALIGN_CENTER);
+			   &rw, ALIGN_CENTER, 1);
 
 	w = rw;
 
@@ -110,9 +168,9 @@ group_layouter(struct dialog_data *dlg_data)
 
 	y = dlg_data->box.y + DIALOG_TB + 1;
 	dlg_format_group(term, dlg_data->widgets_data, n,
-			 dlg_data->box.x + DIALOG_LB, &y, w, NULL);
+			 dlg_data->box.x + DIALOG_LB, &y, w, NULL, 0);
 
 	y++;
 	dlg_format_buttons(term, dlg_data->widgets_data + n, 2,
-			   dlg_data->box.x + DIALOG_LB, &y, w, &rw, ALIGN_CENTER);
+			   dlg_data->box.x + DIALOG_LB, &y, w, &rw, ALIGN_CENTER, 0);
 }

src/bfu/group.h

@@ -7,7 +7,7 @@ struct widget_data;
 
 void dlg_format_group(struct terminal *term,
 		 struct widget_data *widget_data,
-		 int n, int x, int *y, int w, int *rw);
+		 int n, int x, int *y, int w, int *rw, int format_only);
 
 void group_layouter(struct dialog_data *);
 

src/bfu/inpfield.c

@@ -17,6 +17,7 @@
 #include "bfu/msgbox.h"
 #include "bfu/text.h"
 #include "config/kbdbind.h"
+#include "intl/charsets.h"
 #include "intl/gettext/libintl.h"
 #include "osdep/osdep.h"
 #include "session/session.h"
@@ -104,7 +105,7 @@ check_nonempty(struct dialog_data *dlg_data, struct widget_data *widget_data)
 void
 dlg_format_field(struct terminal *term,
 		 struct widget_data *widget_data,
-		 int x, int *y, int w, int *rw, enum format_align align)
+		 int x, int *y, int w, int *rw, enum format_align align, int format_only)
 {
 	static int max_label_width;
 	static int *prev_y; /* Assert the uniqueness of y */	/* TODO: get rid of this !! --Zas */
@@ -129,9 +130,9 @@ dlg_format_field(struct terminal *term,
 	}
 
 	if (label && *label) {
-		if (term) text_color = get_bfu_color(term, "dialog.text");
+		if (!format_only) text_color = get_bfu_color(term, "dialog.text");
 
-		dlg_format_text_do(term, label, x, y, w, rw, text_color, ALIGN_LEFT);
+		dlg_format_text_do(term, label, x, y, w, rw, text_color, ALIGN_LEFT, format_only);
 	}
 
 	/* XXX: We want the field and label on the same line if the terminal
@@ -141,7 +142,7 @@ dlg_format_field(struct terminal *term,
 			(*y) -= INPUTFIELD_HEIGHT;
 			dlg_format_text_do(term, INPUTFIELD_FLOAT_SEPARATOR,
 					   x + label_width, y, w, rw,
-					   text_color, ALIGN_LEFT);
+					   text_color, ALIGN_LEFT, format_only);
 			w -= INPUTFIELD_FLOAT_SEPARATOR_LEN + INPUTFIELD_FLOATLABEL_PADDING;
 			x += INPUTFIELD_FLOAT_SEPARATOR_LEN + INPUTFIELD_FLOATLABEL_PADDING;
 		}
@@ -265,11 +266,27 @@ display_field_do(struct dialog_data *dlg_data, struct widget_data *widget_data,
 	struct terminal *term = dlg_data->win->term;
 	struct color_pair *color;
 	int sel = is_selected_widget(dlg_data, widget_data);
+#ifdef CONFIG_UTF_8
+	int len = 0, left = 0;
+#endif /* CONFIG_UTF_8 */
 
-	int_bounds(&widget_data->info.field.vpos,
+#ifdef CONFIG_UTF_8
+	if (term->utf8) {
+		unsigned char *t = widget_data->cdata;
+		int p = widget_data->info.field.cpos;
+
+		len = utf8_ptr2cells(t, &t[p]);
+		int_bounds(&left, len - widget_data->box.width + 1, len);
+		int_lower_bound(&left, 0);
+		widget_data->info.field.vpos = utf8_cells2bytes(t, left, NULL);
+	} else
+#endif /* CONFIG_UTF_8 */
+	{
+		int_bounds(&widget_data->info.field.vpos,
 		   widget_data->info.field.cpos - widget_data->box.width + 1,
 		   widget_data->info.field.cpos);
-	int_lower_bound(&widget_data->info.field.vpos, 0);
+		int_lower_bound(&widget_data->info.field.vpos, 0);
+	}
 
 	color = get_bfu_color(term, "dialog.field");
 	if (color)
@@ -277,13 +294,26 @@ display_field_do(struct dialog_data *dlg_data, struct widget_data *widget_data,
 
 	color = get_bfu_color(term, "dialog.field-text");
 	if (color) {
-		int len = strlen(widget_data->cdata + widget_data->info.field.vpos);
+		unsigned char *text = widget_data->cdata + widget_data->info.field.vpos;
-		int w = int_min(len, widget_data->box.width);
+		int len, w;
+
+#ifdef CONFIG_UTF_8
+		if (term->utf8 && !hide)
+			len = utf8_ptr2cells(text, NULL);
+		else if (term->utf8)
+			len = utf8_ptr2chars(text, NULL);
+		else
+#endif /* CONFIG_UTF_8 */
+			len = strlen(text);
+		w = int_min(len, widget_data->box.width);
 
 		if (!hide) {
+#ifdef CONFIG_UTF_8
+			if (term->utf8)
+				w = utf8_cells2bytes(text, w, NULL);
+#endif /* CONFIG_UTF_8 */
 			draw_text(term, widget_data->box.x, widget_data->box.y,
-				  widget_data->cdata + widget_data->info.field.vpos, w,
+				  text, w, 0, color);
-				  0, color);
 		} else {
 			struct box box;
 
@@ -295,7 +325,14 @@ display_field_do(struct dialog_data *dlg_data, struct widget_data *widget_data,
 	}
 
 	if (sel) {
-		int x = widget_data->box.x + widget_data->info.field.cpos - widget_data->info.field.vpos;
+		int x;
+
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			x = widget_data->box.x + len - left;
+		else
+#endif /* CONFIG_UTF_8 */
+			x = widget_data->box.x + widget_data->info.field.cpos - widget_data->info.field.vpos;
 
 		set_cursor(term, x, widget_data->box.y, 0);
 		set_window_ptr(dlg_data->win, widget_data->box.x, widget_data->box.y);
@@ -435,13 +472,39 @@ kbd_field(struct dialog_data *dlg_data, struct widget_data *widget_data)
 			break;
 
 		case ACT_EDIT_RIGHT:
-			if (widget_data->info.field.cpos < strlen(widget_data->cdata))
+			if (widget_data->info.field.cpos < strlen(widget_data->cdata)) {
-				widget_data->info.field.cpos++;
+#ifdef CONFIG_UTF_8
+				if (term->utf8) {
+					unsigned char *next = widget_data->cdata + widget_data->info.field.cpos;
+					unsigned char *end = strchr(next, '\0');
+
+					utf_8_to_unicode(&next, end);
+					widget_data->info.field.cpos = (int)(next - widget_data->cdata);
+				} else
+#endif /* CONFIG_UTF_8 */
+				{
+					widget_data->info.field.cpos++;
+				}
+			}
 			goto display_field;
 
 		case ACT_EDIT_LEFT:
 			if (widget_data->info.field.cpos > 0)
 				widget_data->info.field.cpos--;
+#ifdef CONFIG_UTF_8
+			if (widget_data->info.field.cpos && term->utf8) {
+				unsigned char *t = widget_data->cdata;
+				unsigned char *t2 = t;
+				int p = widget_data->info.field.cpos;
+				unsigned char tmp = t[p];
+
+				t[p] = '\0';
+				strlen_utf8(&t2);
+				t[p] = tmp;
+				widget_data->info.field.cpos = (int)(t2 - t);
+
+			}
+#endif /* CONFIG_UTF_8 */
 			goto display_field;
 
 		case ACT_EDIT_HOME:
@@ -453,6 +516,34 @@ kbd_field(struct dialog_data *dlg_data, struct widget_data *widget_data)
 			goto display_field;
 
 		case ACT_EDIT_BACKSPACE:
+#ifdef CONFIG_UTF_8
+			if (widget_data->info.field.cpos && term->utf8) {
+				/* XXX: stolen from src/viewer/text/form.c */
+				/* FIXME: This isn't nice. We remove last byte
+				 *        from UTF-8 character to detect
+				 *        character before it. */
+				unsigned char *text = widget_data->cdata;
+				unsigned char *end = widget_data->cdata + widget_data->info.field.cpos - 1;
+				unicode_val_T data;
+				int old = widget_data->info.field.cpos;
+
+				while(1) {
+					data = utf_8_to_unicode(&text, end);
+					if (data == UCS_NO_CHAR)
+						break;
+				}
+
+				widget_data->info.field.cpos = (int)(text - widget_data->cdata);
+				if (old != widget_data->info.field.cpos) {
+					int length;
+
+					text = widget_data->cdata;
+					length = strlen(text + old) + 1;
+					memmove(text + widget_data->info.field.cpos, text + old, length);
+				}
+				goto display_field;
+			}
+#endif /* CONFIG_UTF_8 */
 			if (widget_data->info.field.cpos) {
 				memmove(widget_data->cdata + widget_data->info.field.cpos - 1,
 					widget_data->cdata + widget_data->info.field.cpos,
@@ -467,6 +558,20 @@ kbd_field(struct dialog_data *dlg_data, struct widget_data *widget_data)
 
 				if (widget_data->info.field.cpos >= cdata_len) goto display_field;
 
+#ifdef CONFIG_UTF_8
+				if (term->utf8) {
+					unsigned char *end = widget_data->cdata + cdata_len;
+					unsigned char *text = widget_data->cdata + widget_data->info.field.cpos;
+					unsigned char *old = text;
+
+					utf_8_to_unicode(&text, end);
+					if (old != text) {
+						memmove(old, text,
+								(int)(end - text) + 1);
+					}
+					goto display_field;
+				}
+#endif /* CONFIG_UTF_8 */
 				memmove(widget_data->cdata + widget_data->info.field.cpos,
 					widget_data->cdata + widget_data->info.field.cpos + 1,
 					cdata_len - widget_data->info.field.cpos + 1);
@@ -585,7 +690,22 @@ kbd_field(struct dialog_data *dlg_data, struct widget_data *widget_data)
 
 				memmove(text + 1, text, textlen + 1);
 				*text = get_kbd_key(ev);
+#ifdef CONFIG_UTF_8
+				if (term->utf8) {
+					static unsigned char buf[7];
+					unsigned char *t = buf;
+					static int i = 0;
+					unicode_val_T data;
 
+					buf[i++] = *text;
+					buf[i] = '\0';
+					data = utf_8_to_unicode(&t, buf + i);
+					if (i == 6) i = 0;
+					if (data == UCS_NO_CHAR)
+						return EVENT_PROCESSED;
+					else i = 0;
+				}
+#endif /* CONFIG_UTF_8 */
 				goto display_field;
 			}
 	}
@@ -641,7 +761,7 @@ input_line_layouter(struct dialog_data *dlg_data)
 		- ses->status.show_tabs_bar;
 
 	dlg_format_field(win->term, dlg_data->widgets_data, 0,
-			 &y, win->term->width, NULL, ALIGN_LEFT);
+			 &y, win->term->width, NULL, ALIGN_LEFT, 0);
 }
 
 static widget_handler_status_T

src/bfu/inpfield.h

@@ -62,7 +62,7 @@ extern struct widget_ops field_pass_ops;
 widget_handler_status_T check_number(struct dialog_data *, struct widget_data *);
 widget_handler_status_T check_nonempty(struct dialog_data *, struct widget_data *);
 
-void dlg_format_field(struct terminal *, struct widget_data *, int, int *, int, int *, enum format_align);
+void dlg_format_field(struct terminal *, struct widget_data *, int, int *, int, int *, enum format_align, int format_only);
 
 void input_field(struct terminal *, struct memory_list *, int, unsigned char *,
 		 unsigned char *, unsigned char *, unsigned char *, void *,

src/bfu/listbox.c

@@ -43,7 +43,7 @@ get_listbox_widget_data(struct widget_data *widget_data)
 void
 dlg_format_listbox(struct terminal *term, struct widget_data *widget_data,
 	           int x, int *y, int w, int max_height, int *rw,
-	           enum format_align align)
+	           enum format_align align, int format_only)
 {
 	int min, optimal_h, height;
 
@@ -453,6 +453,7 @@ display_listbox_item(struct listbox_item *item, void *data_, int *offset)
 	} else {
 		unsigned char *text;
 		struct listbox_ops *ops = data->box->ops;
+		int len_bytes;
 
 		assert(ops && ops->get_info);
 
@@ -461,8 +462,14 @@ display_listbox_item(struct listbox_item *item, void *data_, int *offset)
 
 		len = strlen(text);
 		int_upper_bound(&len, int_max(0, data->widget_data->box.width - depth * 5));
+#ifdef CONFIG_UTF_8
+		if (data->term->utf8)
+			len_bytes = utf8_cells2bytes(text, len, NULL);
+		else
+#endif /* CONFIG_UTF_8 */
+			len_bytes = len;
 
-		draw_text(data->term, x, y, text, len, 0, text_color);
+		draw_text(data->term, x, y, text, len_bytes, 0, text_color);
 
 		mem_free(text);
 	}

src/bfu/listbox.h

@@ -133,7 +133,7 @@ struct listbox_item {
 
 extern struct widget_ops listbox_ops;
 
-void dlg_format_listbox(struct terminal *, struct widget_data *, int, int *, int, int, int *, enum format_align);
+void dlg_format_listbox(struct terminal *, struct widget_data *, int, int *, int, int, int *, enum format_align, int format_only);
 
 struct listbox_item *traverse_listbox_items_list(struct listbox_item *, struct listbox_data *, int, int, int (*)(struct listbox_item *, void *, int *), void *);
 

src/bfu/menu.c

@@ -200,7 +200,14 @@ get_menuitem_text_width(struct terminal *term, struct menu_item *mi)
 
 	if (!text[0]) return 0;
 
-	return L_TEXT_SPACE + strlen(text) - !!mi->hotkey_pos + R_TEXT_SPACE;
+#ifdef CONFIG_UTF_8
+	if (term->utf8)
+		return L_TEXT_SPACE + utf8_ptr2cells(text, NULL)
+		       - !!mi->hotkey_pos + R_TEXT_SPACE;
+	else
+#endif /* CONFIG_UTF_8 */
+		return L_TEXT_SPACE + strlen(text)
+		       - !!mi->hotkey_pos + R_TEXT_SPACE;
 }
 
 /* Get desired width for right text in menu item, accounting spacing. */
@@ -360,17 +367,31 @@ set_menu_selection(struct menu *menu, int pos)
 	int_bounds(&menu->first, 0, menu->size - height);
 }
 
+/* width - number of standard terminal cells to be displayed (text + whitespace
+ *         separators). For double-width glyph width == 2.
+ * len - length of text in bytes */
 static inline void
 draw_menu_left_text(struct terminal *term, unsigned char *text, int len,
 		    int x, int y, int width, struct color_pair *color)
 {
 	int w = width - (L_TEXT_SPACE + R_TEXT_SPACE);
+	int max_len;
 
 	if (w <= 0) return;
 
 	if (len < 0) len = strlen(text);
 	if (!len) return;
-	if (len > w) len = w;
+
+#ifdef CONFIG_UTF_8
+	if (term->utf8) {
+		max_len = utf8_cells2bytes(text, w, NULL);
+		if (max_len <= 0)
+			return;
+	} else
+#endif /* CONFIG_UTF_8 */
+		max_len = w;
+
+	if (len > max_len) len = max_len;
 
 	draw_text(term, x + L_TEXT_SPACE, y, text, len, 0, color);
 }
@@ -389,6 +410,10 @@ draw_menu_left_text_hk(struct terminal *term, unsigned char *text,
 	int xbase = x + L_TEXT_SPACE;
 	int w = width - (L_TEXT_SPACE + R_TEXT_SPACE);
 	int hk_state = 0;
+#ifdef CONFIG_UTF_8
+	unsigned char *text2, *end;
+#endif
+
 #ifdef CONFIG_DEBUG
 	/* For redundant hotkeys highlighting. */
 	int double_hk = 0;
@@ -405,6 +430,10 @@ draw_menu_left_text_hk(struct terminal *term, unsigned char *text,
 		hk_color_sel = tmp;
 	}
 
+#ifdef CONFIG_UTF_8
+	if (term->utf8) goto utf8;
+#endif /* CONFIG_UTF_8 */
+
 	for (x = 0; x - !!hk_state < w && (c = text[x]); x++) {
 		if (!hk_state && x == hotkey_pos - 1) {
 			hk_state = 1;
@@ -417,12 +446,78 @@ draw_menu_left_text_hk(struct terminal *term, unsigned char *text,
 				  (double_hk ? hk_color_sel : hk_color));
 #else
 			draw_char(term, xbase + x - 1, y, c, hk_attr, hk_color);
-#endif
+#endif /* CONFIG_DEBUG */
 			hk_state = 2;
 		} else {
 			draw_char(term, xbase + x - !!hk_state, y, c, 0, color);
 		}
 	}
+	return;
+
+#ifdef CONFIG_UTF_8
+utf8:
+	end = strchr(text, '\0');
+	text2 = text;
+	for (x = 0; x - !!hk_state < w && *text2; x++) {
+		unicode_val_T data;
+
+		data = utf_8_to_unicode(&text2, end);
+		if (!hk_state && (int)(text2 - text) == hotkey_pos) {
+			hk_state = 1;
+			continue;
+		}
+		if (hk_state == 1) {
+			if (unicode_to_cell(data) == 2) {
+				if (x < w && xbase + x < term->width) {
+#ifdef CONFIG_DEBUG
+					draw_char(term, xbase + x - 1, y,
+						  data, hk_attr,
+						  (double_hk ? hk_color_sel
+						             : hk_color));
+#else
+					draw_char(term, xbase + x - 1, y,
+						  data, hk_attr, hk_color);
+#endif /* CONFIG_DEBUG */
+					x++;
+					draw_char(term, xbase + x - 1, y,
+						  UCS_NO_CHAR, 0, hk_color);
+				} else {
+					draw_char(term, xbase + x - 1, y,
+						  ' ', 0, hk_color);
+				}
+			} else {
+#ifdef CONFIG_DEBUG
+				draw_char(term, xbase + x - 1, y,
+					  data, hk_attr,
+					  (double_hk ? hk_color_sel
+					   	     : hk_color));
+#else
+				draw_char(term, xbase + x - 1, y,
+					  data, hk_attr, hk_color);
+#endif /* CONFIG_DEBUG */
+			}
+			hk_state = 2;
+		} else {
+			if (unicode_to_cell(data) == 2) {
+				if (x - !!hk_state + 1 < w &&
+				    xbase + x - !!hk_state + 1 < term->width) {
+					draw_char(term, xbase + x - !!hk_state,
+						  y, data, 0, color);
+					x++;
+					draw_char(term, xbase + x - !!hk_state,
+						  y, UCS_NO_CHAR, 0, color);
+				} else {
+					draw_char(term, xbase + x - !!hk_state,
+						  y, ' ', 0, color);
+				}
+			} else {
+				draw_char(term, xbase + x - !!hk_state,
+					  y, data, 0, color);
+			}
+		}
+
+	}
+#endif /* CONFIG_UTF_8 */
 }
 
 static inline void
@@ -465,7 +560,15 @@ display_menu(struct terminal *term, struct menu *menu)
 		/* Draw shadow */
 		draw_shadow(term, &menu->box,
 			    get_bfu_color(term, "dialog.shadow"), 2, 1);
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			fix_dwchar_around_box(term, &box, 1, 2, 1);
+#endif /* CONFIG_UTF_8 */
 	}
+#ifdef CONFIG_UTF_8
+	else if(term->utf8)
+		fix_dwchar_around_box(term, &box, 1, 0, 0);
+#endif /* CONFIG_UTF_8 */
 
 	menu_height = box.height;
 	box.height = 1;
@@ -1005,18 +1108,33 @@ display_mainmenu(struct terminal *term, struct menu *menu)
 		int l = mi->hotkey_pos;
 		int textlen;
 		int selected = (i == menu->selected);
+		int screencnt;
 
 		if (mi_text_translate(mi))
 			text = _(text, term);
 
 		textlen = strlen(text) - !!l;
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			screencnt = utf8_ptr2cells(text, NULL) - !!l;
+		else
+#endif /* CONFIG_UTF_8 */
+			screencnt = textlen;
 
 		if (selected) {
 			color = selected_color;
 			box.x = p;
-			box.width = L_MAINTEXT_SPACE + L_TEXT_SPACE
+#ifdef CONFIG_UTF_8
-				    + textlen
+			if (term->utf8)
-				    + R_TEXT_SPACE + R_MAINTEXT_SPACE;
+				box.width = L_MAINTEXT_SPACE + L_TEXT_SPACE
+					+ screencnt
+					+ R_TEXT_SPACE + R_MAINTEXT_SPACE;
+			else
+#endif /* CONFIG_UTF_8 */
+				box.width = L_MAINTEXT_SPACE + L_TEXT_SPACE
+					+ textlen
+					+ R_TEXT_SPACE + R_MAINTEXT_SPACE;
+
 			draw_box(term, &box, ' ', 0, color);
 			set_cursor(term, p, 0, 1);
 			set_window_ptr(menu->win, p, 1);
@@ -1034,7 +1152,7 @@ display_mainmenu(struct terminal *term, struct menu *menu)
 					    color);
 		}
 
-		p += textlen;
+		p += screencnt;
 
 		if (p >= term->width - R_MAINMENU_SPACE)
 			break;
@@ -1045,6 +1163,22 @@ display_mainmenu(struct terminal *term, struct menu *menu)
 	menu->last = i - 1;
 	int_lower_bound(&menu->last, menu->first);
 	if (menu->last < menu->size - 1) {
+#ifdef CONFIG_UTF_8
+		if (term->utf8) {
+			struct screen_char *schar;
+
+			schar = get_char(term, term->width - R_MAINMENU_SPACE, 0);
+			/* 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 ' '. */
+				schar++;
+				draw_char_data(term, term->width - R_MAINMENU_SPACE - 1,
+						0, ' ');
+			}
+		}
+#endif
+
 		set_box(&box,
 			term->width - R_MAINMENU_SPACE, 0,
 			R_MAINMENU_SPACE, 1);

src/bfu/text.c

@@ -19,6 +19,7 @@
 #include "terminal/terminal.h"
 #include "util/color.h"
 
+/* FIXME: For UTF-8 strings we need better function than isspace. */
 #define is_unsplitable(pos) (*(pos) && *(pos) != '\n' && !isspace(*(pos)))
 
 void
@@ -35,48 +36,107 @@ add_dlg_text(struct dialog *dlg, unsigned char *text,
 }
 
 /* Returns length of substring (from start of @text) before a split. */
+#ifdef CONFIG_UTF_8
 static inline int
-split_line(unsigned char *text, int max_width)
+split_line(unsigned char *text, int max_width, int *cells, int utf8)
+#else
+static inline int
+split_line(unsigned char *text, int max_width, int *cells)
+#endif /* CONFIG_UTF_8 */
 {
 	unsigned char *split = text;
+	int cells_save = *cells;
 
 	if (max_width <= 0) return 0;
 
 	while (*split && *split != '\n') {
-		unsigned char *next_split = split + 1;
+		unsigned char *next_split;
 
-		while (is_unsplitable(next_split))
+#ifdef CONFIG_UTF_8
-			next_split++;
+		if (utf8) {
+			unsigned char *next_char_begin = split
+							 + utf8charlen(split);
 
-		if (next_split - text > max_width) {
+			next_split = split;
+
+			*cells += utf8_char2cells(split, NULL);
+			while (*next_split && next_split != next_char_begin)
+				next_split++;
+
+			next_char_begin = next_split;
+			while (is_unsplitable(next_split))
+			{
+				if (next_split < next_char_begin) {
+					next_split++;
+					continue;
+				}
+				*cells += utf8_char2cells(next_split, NULL);
+				next_char_begin += utf8charlen(next_split);
+			}
+		} else
+#endif /* CONFIG_UTF_8 */
+		{
+			next_split = split + 1;
+
+			while (is_unsplitable(next_split))
+				next_split++;
+			*cells = next_split - text;
+		}
+
+		if (*cells > max_width) {
 			/* Force a split if no position was found yet,
 			 * meaning there's no splittable substring under
 			 * requested width. */
 			if (split == text) {
-				split = &text[max_width];
+#ifdef CONFIG_UTF_8
+				if (utf8) {
+					int m_bytes = utf8_cells2bytes(text,
+								       max_width,
+								       NULL);
+					split = &text[m_bytes];
+				} else
+#endif /* CONFIG_UTF_8 */
+					split = &text[max_width];
 
-				/* Give preference to split on a punctuation
+
-				 * if any. Note that most of the time
+				/* FIXME: Function ispunct won't work correctly
-				 * punctuation char is followed by a space so
+				 * with UTF-8 characters. We need some similar
-				 * this rule will not match often.
+				 * function for UTF-8 characters. */
-				 * We match dash and quotes too. */
+#ifdef CONFIG_UTF_8
-				while (--split != text) {
+				if (!utf8)
-					if (!ispunct(*split)) continue;
+#endif /* CONFIG_UTF_8 */
-					split++;
+				{
-					break;
+					/* Give preference to split on a
+					 * punctuation if any. Note that most
+					 * of the time punctuation char is
+					 * followed by a space so this rule
+					 * will not match often. We match dash
+					 * and quotes too. */
+					cells_save--;
+					while (--split != text) {
+						cells_save--;
+						if (!ispunct(*split)) continue;
+						split++;
+						cells_save++;
+						break;
+					}
 				}
 
 				/* If no way to do a clean split, just return
 				 * requested maximal width. */
-				if (split == text)
+				if (split == text) {
+					*cells = max_width;
 					return max_width;
+				}
 			}
 			break;
 		}
 
+		cells_save = *cells;
 		split = next_split;
 	}
 
+	*cells = cells_save;
 	return split - text;
 }
 
@@ -86,8 +146,13 @@ split_line(unsigned char *text, int max_width)
 #define realloc_lines(x, o, n) mem_align_alloc(x, o, n, LINES_GRANULARITY)
 
 /* Find the start of each line with the current max width */
+#ifdef CONFIG_UTF_8
+static unsigned char **
+split_lines(struct widget_data *widget_data, int max_width, int utf8)
+#else
 static unsigned char **
 split_lines(struct widget_data *widget_data, int max_width)
+#endif /* CONFIG_UTF_8 */
 {
 	unsigned char *text = widget_data->widget->text;
 	unsigned char **lines = (unsigned char **) widget_data->cdata;
@@ -100,19 +165,27 @@ split_lines(struct widget_data *widget_data, int max_width)
 
 	while (*text) {
 		int width;
+		int cells = 0;
 
 		/* Skip first leading \n or space. */
 		if (isspace(*text)) text++;
 		if (!*text) break;
 
-		width = split_line(text, max_width);
+#ifdef CONFIG_UTF_8
+		width = split_line(text, max_width, &cells, utf8);
+#else
+		width = split_line(text, max_width, &cells);
+#endif
 
 		/* split_line() may return 0. */
 		if (width < 1) {
 			width = 1; /* Infinite loop prevention. */
 		}
+		if (cells < 1) {
+			cells = 1; /* Infinite loop prevention. */
+		}
 
-		int_lower_bound(&widget_data->box.width, width);
+		int_lower_bound(&widget_data->box.width, cells);
 
 		if (!realloc_lines(&lines, line, line + 1))
 			break;
@@ -134,13 +207,15 @@ split_lines(struct widget_data *widget_data, int max_width)
 void
 dlg_format_text_do(struct terminal *term, unsigned char *text,
 		int x, int *y, int width, int *real_width,
-		struct color_pair *color, enum format_align align)
+		struct color_pair *color, enum format_align align,
+		int format_only)
 {
 	int line_width;
 	int firstline = 1;
 
 	for (; *text; text += line_width, (*y)++) {
 		int shift;
+		int cells = 0;
 
 		/* Skip first leading \n or space. */
 		if (!firstline && isspace(*text))
@@ -149,7 +224,11 @@ dlg_format_text_do(struct terminal *term, unsigned char *text,
 			firstline = 0;
 		if (!*text) break;
 
-		line_width = split_line(text, width);
+#ifdef CONFIG_UTF_8
+		line_width = split_line(text, width, &cells, term->utf8);
+#else
+		line_width = split_line(text, width, &cells);
+#endif /* CONFIG_UTF_8 */
 
 		/* split_line() may return 0. */
 		if (line_width < 1) {
@@ -157,18 +236,18 @@ dlg_format_text_do(struct terminal *term, unsigned char *text,
 			continue;
 		}
 
-		if (real_width) int_lower_bound(real_width, line_width);
+		if (real_width) int_lower_bound(real_width, cells);
-		if (!term || !line_width) continue;
+		if (format_only || !line_width) continue;
 
 		/* Calculate the number of chars to indent */
 		if (align == ALIGN_CENTER)
-			shift = (width - line_width) / 2;
+			shift = (width - cells) / 2;
 		else if (align == ALIGN_RIGHT)
-			shift = width - line_width;
+			shift = width - cells;
 		else
 			shift = 0;
 
-		assert(line_width <= width && shift < width);
+		assert(cells <= width && shift < width);
 
 		draw_text(term, x + shift, *y, text, line_width, 0, color);
 	}
@@ -176,7 +255,8 @@ dlg_format_text_do(struct terminal *term, unsigned char *text,
 
 void
 dlg_format_text(struct terminal *term, struct widget_data *widget_data,
-		int x, int *y, int width, int *real_width, int max_height)
+		int x, int *y, int width, int *real_width, int max_height,
+		int format_only)
 {
 	unsigned char *text = widget_data->widget->text;
 	unsigned char saved = 0;
@@ -202,9 +282,15 @@ dlg_format_text(struct terminal *term, struct widget_data *widget_data,
 
 		/* Ensure that the current split is valid but don't
 		 * split if we don't have to */
+#ifdef CONFIG_UTF_8
+		if (widget_data->box.width != width
+		    && !split_lines(widget_data, width, term->utf8))
+			return;
+#else
 		if (widget_data->box.width != width
 		    && !split_lines(widget_data, width))
 			return;
+#endif
 
 		lines = (unsigned char **) widget_data->cdata;
 
@@ -246,7 +332,7 @@ dlg_format_text(struct terminal *term, struct widget_data *widget_data,
 	dlg_format_text_do(term, text,
 		x, y, width, real_width,
 		get_bfu_color(term, "dialog.text"),
-		widget_data->widget->info.text.align);
+		widget_data->widget->info.text.align, format_only);
 
 	if (widget_data->widget->info.text.is_label) (*y)--;
 
@@ -333,7 +419,7 @@ format_and_display_text(struct widget_data *widget_data,
 
 	dlg_format_text(term, widget_data,
 			widget_data->box.x, &y, widget_data->box.width, NULL,
-			height);
+			height, 0);
 
 	display_text(dlg_data, widget_data);
 	redraw_from_window(dlg_data->win);

src/bfu/text.h

@@ -47,11 +47,11 @@ void add_dlg_text(struct dialog *dlg, unsigned char *text,
 extern struct widget_ops text_ops;
 void dlg_format_text_do(struct terminal *term,
 		    unsigned char *text, int x, int *y, int w, int *rw,
-		    struct color_pair *scolor, enum format_align align);
+		    struct color_pair *scolor, enum format_align align, int format_only);
 
 void
 dlg_format_text(struct terminal *term, struct widget_data *widget_data,
-		int x, int *y, int dlg_width, int *real_width, int height);
+		int x, int *y, int dlg_width, int *real_width, int height, int format_only);
 
 #define text_is_scrollable(widget_data) \
 	((widget_data)->widget->info.text.is_scrollable \

src/bookmarks/dialogs.c

@@ -377,7 +377,7 @@ do_move_bookmark(struct bookmark *dest, int insert_as_child,
 				if (box2->top == bm->box_item)
 					listbox_sel_move(widget_data, 1);
 			}
-				
+
 			del_from_list(bm->box_item);
 			del_from_list(bm);
 			if (insert_as_child) {

src/config/options.c

@@ -516,7 +516,7 @@ add_opt(struct option *tree, unsigned char *path, unsigned char *capt,
 			return NULL;
 		}
 	}
-	
+
 	add_opt_rec(tree, path, option);
 	return option;
 }

src/dialogs/download.c

@@ -145,15 +145,20 @@ download_dialog_layouter(struct dialog_data *dlg_data)
 		mem_free(msg);
 		return;
 	}
-	decode_uri_for_display(url);
+#ifdef CONFIG_UTF_8
+	if (term->utf8)
+		decode_uri(url);
+	else
+#endif /* CONFIG_UTF_8 */
+		decode_uri_for_display(url);
 	url_len = strlen(url);
 
 	if (show_meter) {
 		int_lower_bound(&w, DOWN_DLG_MIN);
 	}
 
-	dlg_format_text_do(NULL, url, 0, &y, w, &rw,
+	dlg_format_text_do(term, url, 0, &y, w, &rw,
-			dialog_text_color, ALIGN_LEFT);
+			dialog_text_color, ALIGN_LEFT, 1);
 
 	y++;
 	if (show_meter) y += 2;
@@ -161,13 +166,13 @@ download_dialog_layouter(struct dialog_data *dlg_data)
 #if CONFIG_BITTORRENT
 	if (bittorrent) y += 2;
 #endif
-	dlg_format_text_do(NULL, msg, 0, &y, w, &rw,
+	dlg_format_text_do(term, msg, 0, &y, w, &rw,
-			dialog_text_color, ALIGN_LEFT);
+			dialog_text_color, ALIGN_LEFT, 1);
 
 	y++;
-	dlg_format_buttons(NULL, dlg_data->widgets_data,
+	dlg_format_buttons(term, dlg_data->widgets_data,
 			   dlg_data->number_of_widgets, 0, &y, w,
-			   &rw, ALIGN_CENTER);
+			   &rw, ALIGN_CENTER, 1);
 
 	draw_dialog(dlg_data, w, y);
 
@@ -186,7 +191,7 @@ download_dialog_layouter(struct dialog_data *dlg_data)
 	y = dlg_data->box.y + DIALOG_TB + 1;
 	x = dlg_data->box.x + DIALOG_LB;
 	dlg_format_text_do(term, url, x, &y, w, NULL,
-			dialog_text_color, ALIGN_LEFT);
+			dialog_text_color, ALIGN_LEFT, 0);
 
 	if (show_meter) {
 		y++;
@@ -203,12 +208,12 @@ download_dialog_layouter(struct dialog_data *dlg_data)
 #endif
 	y++;
 	dlg_format_text_do(term, msg, x, &y, w, NULL,
-			dialog_text_color, ALIGN_LEFT);
+			dialog_text_color, ALIGN_LEFT, 0);
 
 	y++;
 	dlg_format_buttons(term, dlg_data->widgets_data,
 			   dlg_data->number_of_widgets, x, &y, w,
-			   NULL, ALIGN_CENTER);
+			   NULL, ALIGN_CENTER, 0);
 
 	mem_free(url);
 	mem_free(msg);
@@ -296,7 +301,14 @@ get_file_download_text(struct listbox_item *item, struct terminal *term)
 	unsigned char *uristring;
 
 	uristring = get_uri_string(file_download->uri, URI_PUBLIC);
-	if (uristring) decode_uri_for_display(uristring);
+	if (uristring) {
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			decode_uri(uristring);
+		else
+#endif /* CONFIG_UTF_8 */
+			decode_uri_for_display(uristring);
+	}
 
 	return uristring;
 }
@@ -369,7 +381,7 @@ draw_file_download(struct listbox_item *item, struct listbox_context *context,
 
 	color = get_bfu_color(context->term, stylename);
 
-	text = get_file_download_text(item, NULL);
+	text = get_file_download_text(item, context->term);
 	if (!text) return;
 
 	length = strlen(text);

src/dialogs/menu.c

@@ -581,7 +581,12 @@ query_file(struct session *ses, struct uri *uri, void *data,
 	add_mime_filename_to_string(&def, uri);
 
 	/* Remove the %-ugliness for display */
-	decode_uri_string_for_display(&def);
+#ifdef CONFIG_UTF_8
+	if (ses->tab->term->utf8)
+		decode_uri_string(&def);
+	else
+#endif /* CONFIG_UTF_8 */
+		decode_uri_string_for_display(&def);
 
 	if (interactive) {
 		input_dialog(ses->tab->term, NULL,

src/dialogs/options.c

@@ -59,8 +59,10 @@ charset_list(struct terminal *term, void *xxx, void *ses_)
 		unsigned char *name = get_cp_name(i);
 
 		if (!name) break;
-		if (is_cp_special(i))
+
-			continue;
+#ifndef CONFIG_UTF_8
+		if (is_cp_utf8(i)) continue;
+#endif /* CONFIG_UTF_8 */
 
 		items++;
 		add_to_menu(&mi, name, NULL, ACT_MAIN_NONE,

src/dialogs/status.c

@@ -150,15 +150,8 @@ get_current_link_info_and_title(struct session *ses,
 
 	link_title = get_current_link_title(doc_view);
 	if (link_title) {
-		unsigned char *src;
-
 		assert(*link_title);
 
-		/* Remove illicit chars. */
-		for (src = link_title; *src; src++)
-			if (!isprint(*src) || iscntrl(*src))
-				*src = '*';
-
 		ret = straconcat(link_info, " - ", link_title, NULL);
 		mem_free(link_info);
 		mem_free(link_title);
@@ -424,11 +417,25 @@ display_title_bar(struct session *ses, struct terminal *term)
 
 	if (document->title) {
 		int maxlen = int_max(term->width - 4 - buflen, 0);
-		int titlelen = int_min(strlen(document->title), maxlen);
+		int titlelen, titlewidth;
+
+#ifdef CONFIG_UTF_8
+		if (term->utf8) {
+			titlewidth = utf8_ptr2cells(document->title, NULL);
+			titlewidth = int_min(titlewidth, maxlen);
+
+			titlelen = utf8_cells2bytes(document->title,
+							titlewidth, NULL);
+		} else
+#endif /* CONFIG_UTF_8 */
+		{
+			titlewidth = int_min(strlen(document->title), maxlen);
+			titlelen = titlewidth;
+		}
 
 		add_bytes_to_string(&title, document->title, titlelen);
 
-		if (titlelen == maxlen)
+		if (titlewidth == maxlen)
 			add_bytes_to_string(&title, "...", 3);
 	}
 
@@ -436,7 +443,16 @@ display_title_bar(struct session *ses, struct terminal *term)
 		add_bytes_to_string(&title, buf, buflen);
 
 	if (title.length) {
-		int x = int_max(term->width - 1 - title.length, 0);
+		int x;
+#ifdef CONFIG_UTF_8
+		if (term->utf8) {
+			x = int_max(term->width - 1
+				    - utf8_ptr2cells(title.source,
+						     title.source
+						     + title.length), 0);
+		} else
+#endif /* CONFIG_UTF_8 */
+			x = int_max(term->width - 1 - title.length, 0);
 
 		draw_text(term, x, 0, title.source, title.length, 0,
 			  get_bfu_color(term, "title.title-text"));

src/document/css/apply.c

@@ -119,8 +119,7 @@ static css_applier_T css_appliers[CSS_PT_LAST] = {
 static void
 examine_element(struct html_context *html_context, struct css_selector *base,
 		enum css_selector_type seltype, enum css_selector_relation rel,
-                struct list_head *selectors, struct html_element *element,
+                struct list_head *selectors, struct html_element *element)
-                struct list_head *html_stack)
 {
 	struct css_selector *selector;
 	unsigned char *code;
@@ -138,7 +137,8 @@ examine_element(struct html_context *html_context, struct css_selector *base,
 		dbginfo(sel, type, base); \
 		merge_css_selectors(base, sel); \
 		/* Ancestor matches? */ \
-		if ((struct list_head *) element->next != html_stack) { \
+		if ((struct list_head *) element->next \
+		     != &html_context->stack) { \
 			struct html_element *ancestor; \
 			/* This is less effective than doing reverse iterations,
 			 * first over sel->leaves and then over the HTML stack,
@@ -147,21 +147,20 @@ examine_element(struct html_context *html_context, struct css_selector *base,
 			 * have to duplicate the whole examine_element(), so if
 			 * profiles won't show it really costs... */ \
 			for (ancestor = element->next; \
-			     (struct list_head *) ancestor != html_stack;\
+			     (struct list_head *) ancestor \
+			      != &html_context->stack;\
 			     ancestor = ancestor->next) \
 				examine_element(html_context, base, \
 						CST_ELEMENT, CSR_ANCESTOR, \
-						&sel->leaves, ancestor, \
+						&sel->leaves, ancestor); \
-						html_stack); \
 			examine_element(html_context, base, \
 			                CST_ELEMENT, CSR_PARENT, \
-			                &sel->leaves, element->next, \
+			                &sel->leaves, element->next); \
-			                html_stack); \
 		} \
 		/* More specific matches? */ \
 		examine_element(html_context, base, type + 1, \
 		                CSR_SPECIFITY, \
-		                &sel->leaves, element, html_stack); \
+		                &sel->leaves, element); \
 	}
 
 	if (seltype <= CST_ELEMENT && element->namelen) {
@@ -235,7 +234,7 @@ get_css_selector_for_element(struct html_context *html_context,
 #endif
 
 	examine_element(html_context, selector, CST_ELEMENT, CSR_ROOT,
-	                &css->selectors, element, html_stack);
+	                &css->selectors, element);
 
 #ifdef DEBUG_CSS
 	DBG("Element %.*s applied.", element->namelen, element->name);

src/document/document.h

@@ -118,13 +118,19 @@ struct link {
 #define get_link_name(link) \
 	(!link_is_form(link) ? (link)->data.name : NULL)
 
-
+#ifdef CONFIG_UTF_8
+struct search {
+	int x, y;
+	signed int n;		/* RAM is cheap nowadays */
+	unicode_val_T c;
+};
+#else
 struct search {
 	int x, y;
 	signed int n:24;	/* This structure is size-critical */
 	unsigned char c;
 };
-
+#endif
 
 struct document {
 	OBJECT_HEAD(struct document);
@@ -176,6 +182,10 @@ struct document {
 	struct search **slines1;
 	struct search **slines2;
 
+#ifdef CONFIG_UTF_8
+	unsigned char buf[7];
+	unsigned char buf_length;
+#endif
 	unsigned int id; /* Used to check cache entries. */
 
 	int cp;

src/document/dom/renderer.c

@@ -260,7 +260,12 @@ render_dom_line(struct dom_renderer *renderer, struct screen_char *template,
 	struct document *document = renderer->document;
 	struct conv_table *convert = renderer->convert_table;
 	enum convert_string_mode mode = renderer->convert_mode;
-	int x;
+	int x, charlen;
+#ifdef CONFIG_UTF_8
+	int utf8 = document->options.utf8;
+	unsigned char *end;
+#endif /* CONFIG_UTF_8 */
+
 
 	assert(renderer && template && string && length);
 
@@ -275,12 +280,15 @@ render_dom_line(struct dom_renderer *renderer, struct screen_char *template,
 
 	add_search_node(renderer, length);
 
-	for (x = 0; x < length; x++, renderer->canvas_x++) {
+#ifdef CONFIG_UTF_8
-		unsigned char data = string[x];
+	end = string + length;
+#endif /* CONFIG_UTF_8 */
+	for (x = 0, charlen = 1; x < length;x += charlen, renderer->canvas_x++) {
+		unsigned char *text = &string[x];
 
 		/* This is mostly to be able to break out so the indentation
 		 * level won't get to high. */
-		switch (data) {
+		switch (*text) {
 		case ASCII_TAB:
 		{
 			int tab_width = 7 - (X(renderer) & 7);
@@ -295,15 +303,33 @@ render_dom_line(struct dom_renderer *renderer, struct screen_char *template,
 			 * ``main loop'' add the actual tab char. */
 			for (; tab_width-- > 0; renderer->canvas_x++)
 				copy_screen_chars(POS(renderer), template, 1);
+			charlen = 1;
 			break;
 		}
 		default:
-			template->data = isscreensafe(data) ? data : '.';
+#ifdef CONFIG_UTF_8
+			if (utf8) {
+				unicode_val_T data;
+				charlen = utf8charlen(text);
+				data = utf_8_to_unicode(&text, end);
+
+				template->data = (unicode_val_T)data;
+
+				if (unicode_to_cell(data) == 2) {
+					copy_screen_chars(POS(renderer),
+							template, 1);
+
+					X(renderer)++;
+					template->data = UCS_NO_CHAR;
+				}
+
+			} else
+#endif /* CONFIG_UTF_8 */
+				template->data = isscreensafe(*text) ? *text:'.';
 		}
 
 		copy_screen_chars(POS(renderer), template, 1);
 	}
-
 	mem_free(string);
 }
 
@@ -1024,6 +1050,9 @@ render_dom_document(struct cache_entry *cached, struct document *document,
 	init_dom_renderer(&renderer, document, buffer, convert_table);
 
 	document->bgcolor = document->options.default_bg;
+#ifdef CONFIG_UTF_8
+	document->options.utf8 = is_cp_utf8(document->options.cp);
+#endif /* CONFIG_UTF_8 */
 
 	if (document->options.plain)
 		parser_type = SGML_PARSER_STREAM;

src/document/html/parser/forms.c

@@ -495,7 +495,13 @@ end_parse:
 	max_width = 0;
 	for (i = 0; i < order; i++) {
 		if (!labels[i]) continue;
-		int_lower_bound(&max_width, strlen(labels[i]));
+#ifdef CONFIG_UTF_8
+		if (html_context->options->utf8)
+			int_lower_bound(&max_width,
+					utf8_ptr2cells(labels[i], NULL));
+		else
+#endif /* CONFIG_UTF_8 */
+			int_lower_bound(&max_width, strlen(labels[i]));
 	}
 
 	for (i = 0; i < max_width; i++)

src/document/html/renderer.c

@@ -23,6 +23,7 @@
 #include "document/refresh.h"
 #include "document/renderer.h"
 #include "intl/charsets.h"
+#include "osdep/types.h"
 #include "protocol/uri.h"
 #include "session/session.h"
 #include "terminal/color.h"
@@ -209,6 +210,10 @@ realloc_spaces(struct part *part, int length)
 
 	if (!ALIGN_SPACES(&part->spaces, part->spaces_len, length))
 		return -1;
+#ifdef CONFIG_UTF_8
+	if (!ALIGN_SPACES(&part->char_width, part->spaces_len, length))
+		return -1;
+#endif
 
 	part->spaces_len = length;
 
@@ -374,6 +379,155 @@ get_format_screen_char(struct html_context *html_context,
 	return &schar_cache;
 }
 
+#ifdef CONFIG_UTF_8
+/* First possibly do the format change and then find out what coordinates
+ * to use since sub- or superscript might change them */
+static inline int
+set_hline(struct html_context *html_context, unsigned char *chars, int charslen,
+	  enum link_state link_state)
+{
+	struct part *part = html_context->part;
+	struct screen_char *schar = get_format_screen_char(html_context,
+	                                                   link_state);
+	int x = part->cx;
+	int y = part->cy;
+	int x2 = x;
+	int len = charslen;
+	int utf8 = html_context->options->utf8;
+
+	assert(part);
+	if_assert_failed return len;
+
+	assert(charslen >= 0);
+
+	if (realloc_spaces(part, x + charslen))
+		return 0;
+
+	if (part->document) {
+		if (realloc_line(html_context, part->document,
+		                 Y(y), X(x) + charslen - 1))
+			return 0;
+		if (utf8) {
+			unsigned char *end = chars + charslen;
+			unicode_val_T data;
+
+			if (part->document->buf_length) {
+				/* previous char was broken in the middle */
+				int length = utf8charlen(part->document->buf);
+				unsigned char i;
+				unsigned char *buf_ptr = part->document->buf;
+
+				for (i = part->document->buf_length; i < length && chars < end;) {
+					part->document->buf[i++] = *chars++;
+				}
+				part->document->buf_length = i;
+				part->document->buf[i] = '\0';
+				data = utf_8_to_unicode(&buf_ptr, buf_ptr + i);
+				if (data != UCS_NO_CHAR) {
+					part->document->buf_length = 0;
+					goto good_char;
+				} else {
+					/* Still not full char */
+					return 0;
+				}
+			}
+
+			for (; chars < end; x++) {
+				if (*chars == NBSP_CHAR) {
+					schar->data = ' ';
+					part->spaces[x] = html_context->options->wrap_nbsp;
+					part->char_width[x] = 1;
+					chars++;
+				} else {
+					part->spaces[x] = (*chars == ' ');
+					data = utf_8_to_unicode(&chars, end);
+					if (data == UCS_NO_CHAR) {
+						if (charslen == 1) {
+							/* HR */
+							unsigned char attr = schar->attr;
+
+							schar->data = *chars++;
+							schar->attr = SCREEN_ATTR_FRAME;
+							copy_screen_chars(&POS(x, y), schar, 1);
+							schar->attr = attr;
+							part->char_width[x] = 0;
+							continue;
+						} else {
+							unsigned char i;
+broken_char:							/* broken char */
+							for (i = 0; chars < end;i++) {
+								part->document->buf[i] = *chars++;
+							}
+							part->document->buf_length = i;
+							return x - x2;
+						}
+					} else {
+good_char:
+						if (unicode_to_cell(data) == 2) {
+							schar->data = (unicode_val_T)data;
+							part->char_width[x] = 2;
+							copy_screen_chars(&POS(x++, y), schar, 1);
+							data = utf_8_to_unicode(&chars, end);
+							if (data == UCS_NO_CHAR) goto broken_char;
+							schar->data = (unicode_val_T)data;
+							part->spaces[x] = 0;
+							part->char_width[x] = 0;
+						} else {
+							part->char_width[x] = unicode_to_cell(data);
+							schar->data = (unicode_val_T)data;
+						}
+					}
+				}
+				copy_screen_chars(&POS(x, y), schar, 1);
+			}
+		} else {
+			for (; charslen > 0; charslen--, x++, chars++) {
+				part->char_width[x] = 1;
+				if (*chars == NBSP_CHAR) {
+					schar->data = ' ';
+					part->spaces[x] = html_context->options->wrap_nbsp;
+				} else {
+					part->spaces[x] = (*chars == ' ');
+					schar->data = *chars;
+				}
+				copy_screen_chars(&POS(x, y), schar, 1);
+			}
+
+		}
+		len = x - x2;
+	} else {
+		if (utf8) {
+			unsigned char *end;
+
+			for (end = chars + charslen; chars < end; x++) {
+				unicode_val_T data;
+
+				part->spaces[x] = (*chars == ' ');
+				data = utf_8_to_unicode(&chars, end);
+				part->char_width[x] = unicode_to_cell(data);
+				if (part->char_width[x] == 2) {
+					x++;
+					part->spaces[x] = 0;
+					part->char_width[x] = 0;
+					data = utf_8_to_unicode(&chars, end);
+				}
+				if (data == UCS_NO_CHAR) {
+					/* this is at the end only */
+					return x - x2;
+				}
+			}
+			len = x - x2;
+		} else {
+			for (; charslen > 0; charslen--, x++, chars++) {
+				part->spaces[x] = (*chars == ' ');
+				part->char_width[x] = 1;
+			}
+		}
+	}
+	return len;
+}
+#else
+
 /* First possibly do the format change and then find out what coordinates
  * to use since sub- or superscript might change them */
 static inline void
@@ -405,7 +559,6 @@ set_hline(struct html_context *html_context, unsigned char *chars, int charslen,
 				part->spaces[x] = (*chars == ' ');
 				schar->data = *chars;
 			}
-
 			copy_screen_chars(&POS(x, y), schar, 1);
 		}
 	} else {
@@ -414,6 +567,7 @@ set_hline(struct html_context *html_context, unsigned char *chars, int charslen,
 		}
 	}
 }
+#endif /* CONFIG_UTF_8 */
 
 static void
 move_links(struct html_context *html_context, int xf, int yf, int xt, int yt)
@@ -645,29 +799,53 @@ split_line_at(struct html_context *html_context, int width)
 	if (part->document) {
 		assert(part->document->data);
 		if_assert_failed return 0;
-		assertm(POS(width, part->cy).data == ' ',
+#ifdef CONFIG_UTF_8
-			"bad split: %c", POS(width, part->cy).data);
+		if (html_context->options->utf8
-		move_chars(html_context, width + 1, part->cy, par_format.leftmargin, part->cy + 1);
+		    && width < part->spaces_len && part->char_width[width] == 2) {
-		del_chars(html_context, width, part->cy);
+			move_chars(html_context, width, part->cy, par_format.leftmargin, part->cy + 1);
+			del_chars(html_context, width, part->cy);
+		} else
+#endif
+		{
+			assertm(POS(width, part->cy).data == ' ',
+					"bad split: %c", POS(width, part->cy).data);
+			move_chars(html_context, width + 1, part->cy, par_format.leftmargin, part->cy + 1);
+			del_chars(html_context, width, part->cy);
+
+		}
 	}
 
-	width++; /* Since we were using (x + 1) only later... */
+#ifdef CONFIG_UTF_8
+	if (!(html_context->options->utf8
+	      && width < part->spaces_len
+	      && part->char_width[width] == 2))
+#endif
+		width++; /* Since we were using (x + 1) only later... */
 
 	tmp = part->spaces_len - width;
 	if (tmp > 0) {
 		/* 0 is possible and I'm paranoid ... --Zas */
 		memmove(part->spaces, part->spaces + width, tmp);
+#ifdef CONFIG_UTF_8
+		memmove(part->char_width, part->char_width + width, tmp);
+#endif
 	}
 
 	assert(tmp >= 0);
 	if_assert_failed tmp = 0;
 	memset(part->spaces + tmp, 0, width);
+#ifdef CONFIG_UTF_8
+	memset(part->char_width + tmp, 0, width);
+#endif
 
 	if (par_format.leftmargin > 0) {
 		tmp = part->spaces_len - par_format.leftmargin;
 		assertm(tmp > 0, "part->spaces_len - par_format.leftmargin == %d", tmp);
 		/* So tmp is zero, memmove() should survive that. Don't recover. */
 		memmove(part->spaces + par_format.leftmargin, part->spaces, tmp);
+#ifdef CONFIG_UTF_8
+		memmove(part->char_width + par_format.leftmargin, part->char_width, tmp);
+#endif
 	}
 
 	part->cy++;
@@ -705,13 +883,38 @@ split_line(struct html_context *html_context)
 	assert(part);
 	if_assert_failed return 0;
 
-	for (x = overlap(par_format); x >= par_format.leftmargin; x--)
+#ifdef CONFIG_UTF_8
-		if (x < part->spaces_len && part->spaces[x])
+	if (html_context->options->utf8) {
-			return split_line_at(html_context, x);
+		for (x = overlap(par_format); x >= par_format.leftmargin; x--) {
 
-	for (x = par_format.leftmargin; x < part->cx ; x++)
+			if (x < part->spaces_len && (part->spaces[x]
-		if (x < part->spaces_len && part->spaces[x])
+			    || (part->char_width[x] == 2
-			return split_line_at(html_context, x);
+				/* Ugly hack. If we haven't place for
+				 * double-width characters we print two
+				 * double-width characters. */
+				&& x != par_format.leftmargin)))
+				return split_line_at(html_context, x);
+		}
+
+		for (x = par_format.leftmargin; x < part->cx ; x++) {
+			if (x < part->spaces_len && (part->spaces[x]
+			    || (part->char_width[x] == 2
+				/* We want to break line after _second_
+				 * double-width character. */
+				&& x > par_format.leftmargin)))
+				return split_line_at(html_context, x);
+		}
+	} else
+#endif
+	{
+		for (x = overlap(par_format); x >= par_format.leftmargin; x--)
+			if (x < part->spaces_len && part->spaces[x])
+				return split_line_at(html_context, x);
+
+		for (x = par_format.leftmargin; x < part->cx ; x++)
+			if (x < part->spaces_len && part->spaces[x])
+				return split_line_at(html_context, x);
+	}
 
 	/* Make sure that we count the right margin to the total
 	 * actual box width. */
@@ -1140,9 +1343,15 @@ done_link_state_info(void)
 	       sizeof(renderer_context.link_state_info));
 }
 
+#ifdef CONFIG_UTF_8
 static inline void
 process_link(struct html_context *html_context, enum link_state link_state,
-	     unsigned char *chars, int charslen)
+	     unsigned char *chars, int charslen, int cells)
+#else
+static inline void
+process_link(struct html_context *html_context, enum link_state link_state,
+		   unsigned char *chars, int charslen)
+#endif /* CONFIG_UTF_8 */
 {
 	struct part *part = html_context->part;
 	struct link *link;
@@ -1194,6 +1403,9 @@ process_link(struct html_context *html_context, enum link_state link_state,
 		if (x_offset) {
 			charslen -= x_offset;
 			chars += x_offset;
+#ifdef CONFIG_UTF_8
+			cells -= x_offset;
+#endif /* CONFIG_UTF_8 */
 		}
 
 		link = new_link(html_context, chars, charslen);
@@ -1208,14 +1420,26 @@ process_link(struct html_context *html_context, enum link_state link_state,
 	}
 
 	/* Add new canvas positions to the link. */
-	if (realloc_points(link, link->npoints + charslen)) {
+#ifdef CONFIG_UTF_8
+	if (realloc_points(link, link->npoints + cells))
+#else
+	if (realloc_points(link, link->npoints + charslen))
+#endif /* CONFIG_UTF_8 */
+	{
 		struct point *point = &link->points[link->npoints];
 		int x = X(part->cx) + x_offset;
 		int y = Y(part->cy);
 
+#ifdef CONFIG_UTF_8
+		link->npoints += cells;
+
+		for (; cells > 0; cells--, point++, x++)
+#else
 		link->npoints += charslen;
 
-		for (; charslen > 0; charslen--, point++, x++) {
+		for (; charslen > 0; charslen--, point++, x++)
+#endif /* CONFIG_UTF_8 */
+		{
 			point->x = x;
 			point->y = y;
 		}
@@ -1266,6 +1490,9 @@ put_chars(struct html_context *html_context, unsigned char *chars, int charslen)
 {
 	enum link_state link_state;
 	struct part *part;
+#ifdef CONFIG_UTF_8
+	int cells;
+#endif /* CONFIG_UTF_8 */
 
 	assert(html_context);
 	if_assert_failed return;
@@ -1324,18 +1551,55 @@ put_chars(struct html_context *html_context, unsigned char *chars, int charslen)
 		else if (html_context->options->links_numbering)
 			put_link_number(html_context);
 	}
-
+#ifdef CONFIG_UTF_8
-	set_hline(html_context, chars, charslen, link_state);
+	cells =
+#endif /* CONFIG_UTF_8 */
+		set_hline(html_context, chars, charslen, link_state);
 
 	if (link_state != LINK_STATE_NONE) {
+#if 0
+/* This all code is from utf8 branch */
+#define is_drawing_subs_or_sups() \
+		((format.style.attr & AT_SUBSCRIPT \
+		  && html_context->options->display_subs) \
+		 || (format.style.attr & AT_SUPERSCRIPT \
+		     && html_context->options->display_sups))
+
+		/* We need to update the current @link_state because <sub> and
+		 * <sup> tags will output to the canvas using an inner
+		 * put_chars() call which results in their process_link() call
+		 * will ``update'' the @link_state. */
+		if (link_state == LINK_STATE_NEW
+		    && (is_drawing_subs_or_sups()
+			|| update_after_subscript != renderer_context.subscript)) {
+			link_state = get_link_state(html_context);
+		}
+
+#undef is_drawing_subs_or_sups
+#endif
+
+#ifdef CONFIG_UTF_8
+		process_link(html_context, link_state, chars, charslen,
+			     cells);
+#else
 		process_link(html_context, link_state, chars, charslen);
+#endif /* CONFIG_UTF_8 */
 	}
 
+#ifdef CONFIG_UTF_8
 	if (renderer_context.nowrap
-	    && part->cx + charslen > overlap(par_format))
+	    && part->cx + cells > overlap(par_format))
+		return;
+
+	part->cx += cells;
+#else
+	if (renderer_context.nowrap
+			&& part->cx + charslen > overlap(par_format))
 		return;
 
 	part->cx += charslen;
+#endif /* CONFIG_UTF_8 */
+
 	renderer_context.nobreak = 0;
 
 	if (!(html_context->options->wrap || html_is_preformatted())) {
@@ -1351,7 +1615,11 @@ put_chars(struct html_context *html_context, unsigned char *chars, int charslen)
 	}
 
 	assert(charslen > 0);
+#ifdef CONFIG_UTF_8
+	part->xa += cells;
+#else
 	part->xa += charslen;
+#endif /* CONFIG_UTF_8 */
 	int_lower_bound(&part->max_width, part->xa
 			+ par_format.leftmargin + par_format.rightmargin
 			- (chars[charslen - 1] == ' '
@@ -1410,6 +1678,9 @@ end:
 	part->cx = -1;
 	part->xa = 0;
    	memset(part->spaces, 0, part->spaces_len);
+#ifdef CONFIG_UTF_8
+   	memset(part->char_width, 0, part->spaces_len);
+#endif
 }
 
 static void
@@ -1829,6 +2100,9 @@ format_html_part(struct html_context *html_context,
 
 	done_link_state_info();
 	mem_free_if(part->spaces);
+#ifdef CONFIG_UTF_8
+	mem_free_if(part->char_width);
+#endif
 
 	if (document) {
 		struct node *node = document->nodes.next;
@@ -1912,6 +2186,9 @@ render_html_document(struct cache_entry *cached, struct document *document,
 							   &document->cp,
 							   &document->cp_status,
 							   document->options.hard_assume);
+#ifdef CONFIG_UTF_8
+	html_context->options->utf8 = is_cp_utf8(document->options.cp);
+#endif /* CONFIG_UTF_8 */
 
 	if (title.length) {
 		document->title = convert_string(renderer_context.convert_table,

src/document/html/renderer.h

@@ -23,6 +23,10 @@ struct part {
 
 	unsigned char *spaces;
 	int spaces_len;
+#ifdef CONFIG_UTF_8
+	unsigned char *char_width;
+#endif
+
 
 	struct box box;
 

src/document/options.h

@@ -101,6 +101,9 @@ struct document_options {
 	unsigned int no_cache:1;
 	unsigned int gradual_rerendering:1;
 
+#ifdef CONFIG_UTF_8
+	unsigned int utf8:1;
+#endif /* CONFIG_UTF_8 */
 	/* Active link coloring */
 	/* This is mostly here to make use of this option cache so link
 	 * drawing is faster. --jonas */

src/document/plain/renderer.c

@@ -175,12 +175,12 @@ get_uri_length(unsigned char *line, int length)
 static int
 print_document_link(struct plain_renderer *renderer, int lineno,
 		    unsigned char *line, int line_pos, int width,
-		    int expanded, struct screen_char *pos)
+		    int expanded, struct screen_char *pos, int cells)
 {
 	struct document *document = renderer->document;
 	unsigned char *start = &line[line_pos];
 	int len = get_uri_length(start, width - line_pos);
-	int screen_column = line_pos + expanded;
+	int screen_column = cells + expanded;
 	struct link *new_link;
 	int link_end = line_pos + len;
 	unsigned char saved_char;
@@ -226,209 +226,6 @@ print_document_link(struct plain_renderer *renderer, int lineno,
 	return len;
 }
 
-enum mode_16_256 {
-	COLOR_NONE,
-	COLOR_16_FOREGROUND,
-	COLOR_16_BACKGROUND,
-	COLOR_38,
-	COLOR_48,
-	COLOR_256_FOREGROUND,
-	COLOR_256_BACKGROUND
-};
-
-static int
-change_colors(struct screen_char *template, unsigned char *line, int line_pos, int width, struct document *document)
-{
-	unsigned char fg, bg, bold;
-	enum mode_16_256 color = COLOR_NONE;
-	unsigned char value = 0;
-	int start = 0;
-
-#if defined(CONFIG_88_COLORS) || defined(CONFIG_256_COLORS)
-	fg = template->color[0];
-	bg = template->color[1];
-#else
-	fg = template->color[0] & 15;
-	bg = template->color[0] >> 4;
-#endif
-	bold = template->attr & SCREEN_ATTR_BOLD;
-	for (; line_pos < width; line_pos++) {
-		unsigned char ch = line[line_pos];
-
-		switch (color) {
-		case COLOR_NONE:
-			switch (ch) {
-			case '0':
-				bold = 0;
-				break;
-			case '1':
-				bold = SCREEN_ATTR_BOLD;
-				break;
-			case '3':
-				color = COLOR_16_FOREGROUND;
-				break;
-			case '4':
-				color = COLOR_16_BACKGROUND;
-				break;
-			case 'm':
-				goto end;
-			default:
-				break;
-			}
-			break;
-
-		case COLOR_16_FOREGROUND:
-			switch (ch) {
-			case '9':
-				fg = 7;
-				color = COLOR_NONE;
-				break;
-			case '8':
-				color = COLOR_38;
-				break;
-			case 'm':
-				goto end;
-			default:
-				if (ch >= '0' && ch <= '7') fg = ch - '0';
-				else goto end;
-				color = COLOR_NONE;
-				break;
-			}
-			break;
-
-		case COLOR_16_BACKGROUND:
-			switch (ch) {
-			case '9':
-				bg = 0;
-				color = COLOR_NONE;
-				break;
-			case '8':
-				color = COLOR_48;
-				break;
-			case 'm':
-				goto end;
-			default:
-				if (ch >= '0' && ch <= '7') bg = ch - '0';
-				else goto end;
-				color = COLOR_NONE;
-				break;
-			}
-			break;
-
-		case COLOR_38:
-			switch (ch) {
-			case '5':
-				start = 1;
-				value = 0;
-				color = COLOR_256_FOREGROUND;
-				break;
-			case 'm':
-				goto end;
-			case ';':
-				break;
-			default:
-				color = COLOR_NONE;
-				break;
-			}
-			break;
-
-		case COLOR_48:
-			switch (ch) {
-			case '5':
-				start = 1;
-				value = 0;
-				color = COLOR_256_BACKGROUND;
-				break;
-			case 'm':
-				goto end;
-			case ';':
-				break;
-			default:
-				color = COLOR_NONE;
-				break;
-			}
-			break;
-
-		case COLOR_256_FOREGROUND:
-			switch (ch) {
-			case ';':
-			case 'm':
-				if (start) {
-					start = 0;
-					value = 0;
-				} else {
-#if defined(CONFIG_88_COLORS) || defined(CONFIG_256_COLORS)
-#ifdef CONFIG_88_COLORS
-					if (document->options.color_mode == COLOR_MODE_88) fg = value;
-#endif
-#ifdef CONFIG_256_COLORS
-					if (document->options.color_mode == COLOR_MODE_256) fg = value;
-#endif
-#endif
-					color = COLOR_NONE;
-					value = 0;
-				}
-				if (ch == 'm') goto end;
-				break;
-			default:
-				if (ch >= '0' && ch <= '9') value *= 10 + ch - '0';
-				else goto end;
-				break;
-			}
-			break;
-
-		case COLOR_256_BACKGROUND:
-			switch (ch) {
-			case ';':
-			case 'm':
-				if (start) {
-					start = 0;
-					value = 0;
-				} else {
-#if defined(CONFIG_88_COLORS) || defined(CONFIG_256_COLORS)
-#ifdef CONFIG_88_COLORS
-					if (document->options.color_mode == COLOR_MODE_88) bg = value;
-#endif
-#ifdef CONFIG_256_COLORS
-					if (document->options.color_mode == COLOR_MODE_256) bg = value;
-#endif
-#endif
-					color = COLOR_NONE;
-					value = 0;
-				}
-				if (ch == 'm') goto end;
-				break;
-			default:
-				if (ch >= '0' && ch <= '9') value *= 10 + ch - '0';
-				else goto end;
-				break;
-			}
-			break;
-		}
-	}
-end:
-#if defined(CONFIG_88_COLORS) || defined(CONFIG_256_COLORS)
-#ifdef CONFIG_88_COLORS
-	if (document->options.color_mode == COLOR_MODE_88) {
-		TERM_COLOR_FOREGROUND(template->color) = fg;
-		TERM_COLOR_BACKGROUND(template->color) = bg;
-	}
-#endif
-#ifdef CONFIG_256_COLORS
-	if (document->options.color_mode == COLOR_MODE_256) {
-		TERM_COLOR_FOREGROUND(template->color) = fg;
-		TERM_COLOR_BACKGROUND(template->color) = bg;
-	}
-#endif
-#endif
-
-	if (document->options.color_mode == COLOR_MODE_16) {
-		fg |= bold;
-		set_term_color16(template, document->options.color_flags, fg, bg);
-	}
-	return line_pos;
-}
-
 static inline int
 add_document_line(struct plain_renderer *renderer,
 		  unsigned char *line, int line_width)
@@ -437,6 +234,10 @@ add_document_line(struct plain_renderer *renderer,
 	struct screen_char *template = &renderer->template;
 	struct screen_char saved_renderer_template = *template;
 	struct screen_char *pos, *startpos;
+#ifdef CONFIG_UTF_8
+	int utf8 = document->options.utf8;
+#endif /* CONFIG_UTF_8 */
+	int cells = 0;
 	int lineno = renderer->lineno;
 	int expanded = 0;
 	int width = line_width;
@@ -448,13 +249,31 @@ add_document_line(struct plain_renderer *renderer,
 	if (!line) return 0;
 
 	/* Now expand tabs */
-	for (line_pos = 0; line_pos < width; line_pos++) {
+	for (line_pos = 0; line_pos < width;) {
 		unsigned char line_char = line[line_pos];
+		int charlen = 1;
+		int cell = 1;
+#ifdef CONFIG_UTF_8
+		unicode_val_T data;
+
+		if (utf8) {
+			unsigned char *line_char2 = &line[line_pos];
+			charlen = utf8charlen(&line_char);
+			data = utf_8_to_unicode(&line_char2, &line[width]);
+
+			if (data == UCS_NO_CHAR) {
+				line_pos += charlen;
+				continue;
+			}
+
+			cell = unicode_to_cell(data);
+		}
+#endif /* CONFIG_UTF_8 */
 
 		if (line_char == ASCII_TAB
-		    && (line_pos + 1 == width
+		    && (line_pos + charlen == width
-			|| line[line_pos + 1] != ASCII_BS)) {
+		      	|| line[line_pos + charlen] != ASCII_BS)) {
-			int tab_width = 7 - ((line_pos + expanded) & 7);
+		  	int tab_width = 7 - ((cells + expanded) & 7);
 
 			expanded += tab_width;
 		} else if (line_char == ASCII_BS) {
@@ -475,6 +294,8 @@ add_document_line(struct plain_renderer *renderer,
 				expanded--;
 #endif
 		}
+		line_pos += charlen;
+		cells += cell;
 	}
 
 	assert(expanded >= 0);
@@ -485,28 +306,55 @@ add_document_line(struct plain_renderer *renderer,
 		return 0;
 	}
 
+	cells = 0;
 	expanded = 0;
-	for (line_pos = 0; line_pos < width; line_pos++) {
+	for (line_pos = 0; line_pos < width;) {
 		unsigned char line_char = line[line_pos];
 		unsigned char next_char, prev_char;
+		int charlen = 1;
+		int cell = 1;
+#ifdef CONFIG_UTF_8
+		unicode_val_T data;
 
-		prev_char = line_pos > 0 ? line[line_pos - 1] : '\0';
+		if (utf8) {
-		next_char = (line_pos + 1 < width) ? line[line_pos + 1]
+			unsigned char *line_char2 = &line[line_pos];
-						   : '\0';
+			charlen = utf8charlen(&line_char);
+			data = utf_8_to_unicode(&line_char2, &line[width]);
 
-		switch (line_char) {
+			if (data == UCS_NO_CHAR) {
-		case 27:
+				line_pos += charlen;
-			if (next_char != '[') goto normal;
-			line_pos += 2;
-			line_pos = change_colors(&saved_renderer_template, line, line_pos, width, document);
-			*template = saved_renderer_template;
-			break;
-		case ASCII_BS:
-			if (!(expanded + line_pos)) {
-				/* We've backspaced to the start of the line */
 				continue;
 			}
 
+			cell = unicode_to_cell(data);
+		}
+#endif /* CONFIG_UTF_8 */
+
+		prev_char = line_pos > 0 ? line[line_pos - 1] : '\0';
+		next_char = (line_pos + charlen < width) ?
+		  		line[line_pos + charlen] : '\0';
+
+		/* Do not expand tabs that precede back-spaces; this saves the
+		 * back-space code some trouble. */
+		if (line_char == ASCII_TAB && next_char != ASCII_BS) {
+			int tab_width = 7 - ((cells + expanded) & 7);
+
+			expanded += tab_width;
+
+			template->data = ' ';
+			do
+				copy_screen_chars(pos++, template, 1);
+			while (tab_width--);
+
+			*template = saved_renderer_template;
+
+		} else if (line_char == ASCII_BS) {
+#if 0
+			if (!(expanded + cells)) {
+				/* We've backspaced to the start of the line */
+				continue;
+			}
+#endif
 			if (pos > startpos)
 				pos--;  /* Backspace */
 
@@ -516,8 +364,8 @@ add_document_line(struct plain_renderer *renderer,
 				/* x^H_ becomes _^Hx */
 				if (line_pos - 1 >= 0)
 					line[line_pos - 1] = next_char;
-				if (line_pos + 1 < width)
+				if (line_pos + charlen < width)
-					line[line_pos + 1] = prev_char;
+					line[line_pos + charlen] = prev_char;
 
 				/* Go back and reparse the swapped characters */
 				if (line_pos - 2 >= 0)
@@ -562,62 +410,69 @@ add_document_line(struct plain_renderer *renderer,
 			/* Handle _^Hx^Hx as both bold and underlined */
 			if (template->attr)
 				template->attr |= pos->attr;
-			break;
+		} else {
-		case ASCII_TAB:
+			int added_chars = 0;
-			/* Do not expand tabs that precede back-spaces; this saves the
-			 * back-space code some trouble. */
-			if (next_char != ASCII_BS) {
-				int tab_width = 7 - ((line_pos + expanded) & 7);
 
-				expanded += tab_width;
+			if (document->options.plain_display_links
-
+			    && isalpha(line_char) && isalpha(next_char)) {
-				template->data = ' ';
+				/* We only want to check for a URI if there are
-				do
+				 * at least two consecutive alphabetic
-					copy_screen_chars(pos++, template, 1);
+				 * characters, or if we are at the very start of
-				while (tab_width--);
+				 * the line.  It improves performance a bit.
-
+				 * --Zas */
-				*template = saved_renderer_template;
+				added_chars = print_document_link(renderer,
-				break;
-			}
-		default:
-normal:
-			{
-				int added_chars = 0;
-
-				if (document->options.plain_display_links
-				    && isalpha(line_char) && isalpha(next_char)) {
-					/* We only want to check for a URI if there are
-				 	* at least two consecutive alphabetic
-				 	* characters, or if we are at the very start of
-				 	* the line.  It improves performance a bit.
-				 	* --Zas */
-					added_chars = print_document_link(renderer,
 								  lineno, line,
 								  line_pos,
 								  width,
 								  expanded,
-								  pos);
+								  pos, cells);
-				}
+			}
 
-				if (added_chars) {
+			if (added_chars) {
-					line_pos += added_chars - 1;
+				line_pos += added_chars - 1;
-					pos += added_chars;
+				pos += added_chars;
-				} else {
+			} else {
-					if (!isscreensafe(line_char) && line_char != 27)
+#ifdef CONFIG_UTF_8
+				if (utf8) {
+					unsigned char *text = &line[line_pos];
+					unicode_val_T data =
+						utf_8_to_unicode(&text,
+								&line[width]);
+
+					if (data == UCS_NO_CHAR) {
+						line_pos += charlen;
+						continue;
+					}
+
+					template->data = (unicode_val_T)data;
+					copy_screen_chars(pos++, template, 1);
+
+					if (unicode_to_cell(data) == 2) {
+						template->data = UCS_NO_CHAR;
+						copy_screen_chars(pos++,
+								  template, 1);
+						cell++;
+					}
+				} else
+#endif /* CONFIG_UTF_8 */
+				{
+					if (!isscreensafe(line_char))
 						line_char = '.';
 					template->data = line_char;
 					copy_screen_chars(pos++, template, 1);
 
-					/* Detect copy of nul chars to screen, this
+					/* Detect copy of nul chars to screen,
-				 	* should not occur. --Zas */
+					 * this should not occur. --Zas */
 					assert(line_char);
 				}
-
-				*template = saved_renderer_template;
 			}
-		}
-	}
 
+			*template = saved_renderer_template;
+		}
+
+		line_pos += charlen;
+		cells += cell;
+	}
 	mem_free(line);
 
 	realloc_line(document, pos - startpos, lineno);
@@ -664,19 +519,20 @@ add_document_lines(struct plain_renderer *renderer)
 	int length = renderer->length;
 	int was_empty_line = 0;
 	int was_wrapped = 0;
-
+#ifdef CONFIG_UTF_8
+	int utf8 = is_cp_utf8(renderer->document->cp);
+#endif
 	for (; length > 0; renderer->lineno++) {
 		unsigned char *xsource;
 		int width, added, only_spaces = 1, spaces = 0, was_spaces = 0;
 		int last_space = 0;
 		int tab_spaces = 0;
 		int step = 0;
+ 		int cells = 0;
 
 		/* End of line detection: We handle \r, \r\n and \n types. */
-		for (width = 0;
+ 		for (width = 0; (width < length) &&
-		     width + tab_spaces < renderer->max_width
+ 				(cells < renderer->max_width);) {
-		      && width < length;
-		     width++) {
 			if (source[width] == ASCII_CR)
 				step++;
 			if (source[width + step] == ASCII_LF)
@@ -695,6 +551,22 @@ add_document_lines(struct plain_renderer *renderer)
 				only_spaces = 0;
 				was_spaces = 0;
 			}
+#ifdef CONFIG_UTF_8
+			if (utf8) {
+				unsigned char *text = &source[width];
+				unicode_val_T data = utf_8_to_unicode(&text,
+							&source[length]);
+
+				if (data == UCS_NO_CHAR) return;
+
+				cells += unicode_to_cell(data);
+				width += utf8charlen(&source[width]);
+			} else
+#endif /* CONFIG_UTF_8 */
+			{
+				cells++;
+				width++;
+			}
 		}
 
 		if (only_spaces && step) {
@@ -722,6 +594,7 @@ add_document_lines(struct plain_renderer *renderer)
 				width -= was_spaces;
 				step += was_spaces;
 			}
+
 			if (!step && (width < length) && last_space) {
 				width = last_space;
 				step = 1;
@@ -777,6 +650,9 @@ render_plain_document(struct cache_entry *cached, struct document *document,
 
 	document->bgcolor = document->options.default_bg;
 	document->width = 0;
+#ifdef CONFIG_UTF_8
+	document->options.utf8 = is_cp_utf8(document->options.cp);
+#endif /* CONFIG_UTF_8 */
 
 	/* Setup the style */
 	init_template(&renderer.template, &document->options);

src/document/renderer.c

@@ -345,8 +345,14 @@ render_document(struct view_state *vs, struct document_view *doc_view,
 			}
 
 			document->title = get_uri_string(document->uri, components);
-			if (document->title)
+			if (document->title) {
-				decode_uri_for_display(document->title);
+#ifdef CONFIG_UTF_8
+				if (doc_view->document->options.utf8)
+					decode_uri(document->title);
+				else
+#endif /* CONFIG_UTF_8 */
+					decode_uri_for_display(document->title);
+			}
 		}
 
 #ifdef CONFIG_CSS

src/intl/charsets.c

@@ -140,6 +140,14 @@ u2cp_(unicode_val_T u, int to, int no_nbsp_hack)
 	int s;
 
 	if (u < 128) return strings[u];
+
+	to &= ~SYSTEM_CHARSET_FLAG;
+
+#ifdef CONFIG_UTF_8
+	if (codepages[to].table == table_utf_8)
+		return encode_utf_8(u);
+#endif /* CONFIG_UTF_8 */
+
 	/* To mark non breaking spaces, we use a special char NBSP_CHAR. */
 	if (u == 0xa0) return no_nbsp_hack ? " " : NBSP_CHAR_STRING;
 	if (u == 0xad) return "";
@@ -151,7 +159,6 @@ u2cp_(unicode_val_T u, int to, int no_nbsp_hack)
 		return u2cp_(strange, to, no_nbsp_hack);
 	}
 
-	to &= ~SYSTEM_CHARSET_FLAG;
 
 	for (j = 0; codepages[to].table[j].c; j++)
 		if (codepages[to].table[j].u == u)
@@ -165,8 +172,13 @@ u2cp_(unicode_val_T u, int to, int no_nbsp_hack)
 
 static unsigned char utf_buffer[7];
 
+#ifdef CONFIG_UTF_8
+inline unsigned char *
+encode_utf_8(unicode_val_T u)
+#else
 static unsigned char *
 encode_utf_8(unicode_val_T u)
+#endif /* CONFIG_UTF_8 */
 {
 	memset(utf_buffer, 0, 7);
 
@@ -200,6 +212,252 @@ encode_utf_8(unicode_val_T u)
 	return utf_buffer;
 }
 
+#ifdef CONFIG_UTF_8
+/* Number of bytes utf8 character indexed by first byte. Illegal bytes are
+ * equal ones and handled different. */
+static char utf8char_len_tab[256] = {
+	1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,
+	1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,
+	1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,
+	1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,
+	1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,
+	1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1, 1,1,1,1,1,1,1,1,
+	2,2,2,2,2,2,2,2, 2,2,2,2,2,2,2,2, 2,2,2,2,2,2,2,2, 2,2,2,2,2,2,2,2,
+	3,3,3,3,3,3,3,3, 3,3,3,3,3,3,3,3, 4,4,4,4,4,4,4,4, 5,5,5,5,6,6,1,1,
+};
+
+inline int utf8charlen(const unsigned char *p)
+{
+	return p ? utf8char_len_tab[*p] : 0;
+}
+
+inline int
+strlen_utf8(unsigned char **str)
+{
+	unsigned char *s = *str;
+	unsigned char *end = strchr(s, '\0');
+	int x;
+	int len;
+
+	for (x = 0;; x++, s += len) {
+		len = utf8charlen(s);
+		if (s + len > end) break;
+	}
+	*str = s;
+	return x;
+}
+
+#define utf8_issingle(p) (((p) & 0x80) == 0)
+#define utf8_islead(p) (utf8_issingle(p) || ((p) & 0xc0) == 0xc0)
+
+/* Start from @current and move back to @pos char. This pointer return. The
+ * most left pointer is @start. */
+inline unsigned char *
+utf8_prevchar(unsigned char *current, int pos, unsigned char *start)
+{
+	if (current == NULL || start == NULL || pos < 0)
+		return NULL;
+	while (pos > 0 && current != start) {
+		current--;
+		if (utf8_islead(*current))
+			pos--;
+	}
+	return current;
+}
+
+/* Count number of standard terminal cells needed for displaying UTF-8
+ * character. */
+int
+utf8_char2cells(unsigned char *utf8_char, unsigned char *end)
+{
+	unicode_val_T u;
+
+	if (end == NULL)
+		end = strchr(utf8_char, '\0');
+
+	if(!utf8_char || !end)
+		return -1;
+
+	u = utf_8_to_unicode(&utf8_char, end);
+
+	return unicode_to_cell(u);
+}
+
+/* Count number of standard terminal cells needed for displaying string
+ * with UTF-8 characters. */
+int
+utf8_ptr2cells(unsigned char *string, unsigned char *end)
+{
+	int charlen, cell, cells = 0;
+
+	if (end == NULL)
+		end = strchr(string, '\0');
+
+	if(!string || !end)
+		return -1;
+
+	do {
+		charlen = utf8charlen(string);
+		if (string + charlen > end)
+			break;
+
+		cell = utf8_char2cells(string, end);
+		if  (cell < 0)
+			return -1;
+
+		cells += cell;
+		string += charlen;
+	} while (1);
+
+	return cells;
+}
+
+/* Count number of characters in string. */
+int
+utf8_ptr2chars(unsigned char *string, unsigned char *end)
+{
+	int charlen, chars = 0;
+
+	if (end == NULL)
+		end = strchr(string, '\0');
+
+	if(!string || !end)
+		return -1;
+
+	do {
+		charlen = utf8charlen(string);
+		if (string + charlen > end)
+			break;
+
+		chars++;
+		string += charlen;
+	} while (1);
+
+	return chars;
+}
+
+/*
+ * Count number of bytes from begining of the string needed for displaying
+ * specified number of cells.
+ */
+int
+utf8_cells2bytes(unsigned char *string, int max_cells, unsigned char *end)
+{
+	unsigned int bytes = 0, cells = 0;
+
+	assert(max_cells>=0);
+
+	if (end == NULL)
+		end = strchr(string, '\0');
+
+	if(!string || !end)
+		return -1;
+
+	do {
+		int cell = utf8_char2cells(&string[bytes], end);
+		if (cell < 0)
+			return -1;
+
+		cells += cell;
+		if (cells > max_cells)
+			break;
+
+		bytes += utf8charlen(&string[bytes]);
+
+		if (string + bytes > end) {
+			bytes = end - string;
+			break;
+		}
+	} while(1);
+
+	return bytes;
+}
+
+/*
+ * Find out number of standard terminal collumns needed for displaying symbol
+ * (glyph) which represents Unicode character c.
+ * TODO: Use wcwidth when it is available.
+ *
+ * @return	2 for double-width glyph, 1 for others.
+ * 		TODO: May be extended to return 0 for zero-width glyphs
+ * 		(like composing, maybe unprintable too).
+ */
+inline int
+unicode_to_cell(unicode_val_T c)
+{
+	if (c >= 0x1100
+		&& (c <= 0x115f			/* Hangul Jamo */
+		|| c == 0x2329
+		|| c == 0x232a
+		|| (c >= 0x2e80 && c <= 0xa4cf
+			&& c != 0x303f)		/* CJK ... Yi */
+		|| (c >= 0xac00 && c <= 0xd7a3)	/* Hangul Syllables */
+		|| (c >= 0xf900 && c <= 0xfaff)	/* CJK Compatibility
+								Ideographs */
+		|| (c >= 0xfe30 && c <= 0xfe6f)	/* CJK Compatibility Forms */
+		|| (c >= 0xff00 && c <= 0xff60)	/* Fullwidth Forms */
+		|| (c >= 0xffe0 && c <= 0xffe6)
+		|| (c >= 0x20000 && c <= 0x2fffd)
+		|| (c >= 0x30000 && c <= 0x3fffd)))
+		return 2;
+
+	return 1;
+}
+
+inline unicode_val_T
+utf_8_to_unicode(unsigned char **string, unsigned char *end)
+{
+	unsigned char *str = *string;
+	unicode_val_T u;
+	int length;
+
+	length = utf8char_len_tab[str[0]];
+
+	if (str + length > end) {
+		return UCS_NO_CHAR;
+	}
+
+	switch (length) {
+		case 1:
+			u = str[0];
+			break;
+		case 2:
+			u = (str[0] & 0x1f) << 6;
+			u += (str[1] & 0x3f);
+			break;
+		case 3:
+			u = (str[0] & 0x0f) << 12;
+			u += ((str[1] & 0x3f) << 6);
+			u += (str[2] & 0x3f);
+			break;
+		case 4:
+			u = (str[0] & 0x0f) << 18;
+			u += ((str[1] & 0x3f) << 12);
+			u += ((str[2] & 0x3f) << 6);
+			u += (str[3] & 0x3f);
+			break;
+		case 5:
+			u = (str[0] & 0x0f) << 24;
+			u += ((str[1] & 0x3f) << 18);
+			u += ((str[2] & 0x3f) << 12);
+			u += ((str[3] & 0x3f) << 6);
+			u += (str[4] & 0x3f);
+			break;
+		case 6:
+		default:
+			u = (str[0] & 0x01) << 30;
+			u += ((str[1] & 0x3f) << 24);
+			u += ((str[2] & 0x3f) << 18);
+			u += ((str[3] & 0x3f) << 12);
+			u += ((str[4] & 0x3f) << 6);
+			u += (str[5] & 0x3f);
+			break;
+	}
+	*string = str + length;
+	return u;
+}
+#endif /* CONFIG_UTF_8 */
+
 /* This slow and ugly code is used by the terminal utf_8_io */
 unsigned char *
 cp2utf_8(int from, int c)
@@ -430,11 +688,18 @@ get_entity_string(const unsigned char *str, const int strlen, int encoding)
 	static struct entity_cache entity_cache[ENTITY_CACHE_MAXLEN][ENTITY_CACHE_SIZE];
 	static unsigned int nb_entity_cache[ENTITY_CACHE_MAXLEN];
 	static int first_time = 1;
-	unsigned int slen;
+	unsigned int slen = 0;
 	unsigned char *result = NULL;
 
 	if (strlen <= 0) return NULL;
 
+#ifdef CONFIG_UTF_8
+	/* TODO: caching UTF-8 */
+	encoding &= ~SYSTEM_CHARSET_FLAG;
+	if (codepages[encoding].table == table_utf_8)
+		goto skip;
+#endif /* CONFIG_UTF_8 */
+
 	if (first_time) {
 		memset(&nb_entity_cache, 0, ENTITY_CACHE_MAXLEN * sizeof(unsigned int));
 		first_time = 0;
@@ -488,7 +753,9 @@ get_entity_string(const unsigned char *str, const int strlen, int encoding)
 		fprintf(stderr, "miss\n");
 #endif
 	}
-
+#ifdef CONFIG_UTF_8
+skip:
+#endif /* CONFIG_UTF_8 */
 	if (*str == '#') { /* Numeric entity. */
 		int l = (int) strlen;
 		unsigned char *st = (unsigned char *) str;
@@ -540,6 +807,11 @@ get_entity_string(const unsigned char *str, const int strlen, int encoding)
 		if (element) result = u2cp(element->c, encoding);
 	}
 
+#ifdef CONFIG_UTF_8
+	if (codepages[encoding].table == table_utf_8) {
+		return result;
+	}
+#endif /* CONFIG_UTF_8 */
 end:
 	/* Take care of potential buffer overflow. */
 	if (strlen < sizeof(entity_cache[slen][0].str)) {
@@ -865,7 +1137,7 @@ get_cp_mime_name(int cp_index)
 }
 
 int
-is_cp_special(int cp_index)
+is_cp_utf8(int cp_index)
 {
 	cp_index &= ~SYSTEM_CHARSET_FLAG;
 	return codepages[cp_index].table == table_utf_8;

src/intl/charsets.h

@@ -51,8 +51,21 @@ unsigned char *convert_string(struct conv_table *convert_table,
 int get_cp_index(unsigned char *);
 unsigned char *get_cp_name(int);
 unsigned char *get_cp_mime_name(int);
-int is_cp_special(int);
+int is_cp_utf8(int);
 void free_conv_table(void);
+#ifdef CONFIG_UTF_8
+inline unsigned char *encode_utf_8(unicode_val_T);
+inline unsigned char *utf8_prevchar(unsigned char *, int, unsigned char *);
+inline int utf8charlen(const unsigned char *);
+int utf8_char2cells(unsigned char *, unsigned char *);
+int utf8_ptr2cells(unsigned char *, unsigned char *);
+int utf8_ptr2chars(unsigned char *, unsigned char *);
+int utf8_cells2bytes(unsigned char *, int, unsigned char *);
+inline int unicode_to_cell(unicode_val_T);
+inline int strlen_utf8(unsigned char **);
+inline unicode_val_T utf_8_to_unicode(unsigned char **, unsigned char *);
+#endif /* CONFIG_UTF_8 */
+
 unsigned char *cp2utf_8(int, int);
 
 unsigned char *u2cp_(unicode_val_T, int, int no_nbsp_hack);

src/main/version.c

@@ -118,6 +118,9 @@ get_dyn_full_version(struct terminal *term, int more)
 #endif
 #ifndef CONFIG_MOUSE
 		comma, _("No mouse", term),
+#endif
+#ifdef CONFIG_UTF_8
+		comma, "UTF-8",
 #endif
 		comma,
 		NULL

src/protocol/bittorrent/dialogs.c

@@ -574,7 +574,12 @@ bittorrent_message_dialog(struct session *ses, void *data)
 
 	uristring = get_uri_string(message->uri, URI_PUBLIC);
 	if (uristring) {
-		decode_uri_for_display(uristring);
+#ifdef CONFIG_UTF_8
+		if (ses->tab->term->utf8)
+			decode_uri(uristring);
+		else
+#endif /* CONFIG_UTF_8 */
+			decode_uri_for_display(uristring);
 		add_format_to_string(&string,
 			_("Unable to retrieve %s", ses->tab->term),
 			uristring);
@@ -719,7 +724,12 @@ bittorrent_query_callback(void *data, enum connection_state state,
 
 		/* Let's make the filename pretty for display & save */
 		/* TODO: The filename can be the empty string here. See bug 396. */
-		decode_uri_string_for_display(&filename);
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			decode_uri_string(&filename);
+		else
+#endif /* CONFIG_UTF_8 */
+			decode_uri_string_for_display(&filename);
 	}
 
 	add_format_to_string(&msg,

src/protocol/gopher/gopher.c

@@ -649,6 +649,10 @@ init_gopher_index_cache_entry(struct connection *conn)
 		return S_OUT_OF_MEM;
 
 	where = get_uri_string(conn->uri, URI_PUBLIC);
+
+	/* TODO: Use different function when using UTF-8
+	 * in terminal (decode_uri_for_display replaces
+	 * bytes of UTF-8 characters width '*'). */
 	if (where) decode_uri_for_display(where);
 
 	add_format_to_string(&buffer,

src/scripting/python/core.c

@@ -4,7 +4,6 @@
 #include "config.h"
 #endif
 
-#include "scripting/python/core.h"
 #include <Python.h>
 
 #include <stdio.h>
@@ -14,24 +13,91 @@
 
 #include "config/home.h"
 #include "main/module.h"
+#include "scripting/scripting.h"
+#include "scripting/python/core.h"
+#include "scripting/python/python.h"
 #include "util/env.h"
 #include "util/file.h"
 #include "util/string.h"
 
 
-PyObject *pDict, *pModule;
+PyObject *pDict = NULL, *pModule = NULL;
+
+/* Error reporting. */
+
+void
+alert_python_error(struct session *ses)
+{
+	unsigned char *msg = "(no traceback available)";
+	PyObject *err_type = NULL, *err_value = NULL, *err_traceback = NULL;
+	PyObject *tb_module = NULL;
+	PyObject *tb_dict;
+	PyObject *format_function;
+	PyObject *msg_list = NULL;
+	PyObject *empty_string = NULL;
+	PyObject *join_method = NULL;
+	PyObject *msg_string = NULL;
+	unsigned char *temp;
+
+	/*
+	 * Retrieve the current error indicator and use the format_exception()
+	 * function in Python's traceback module to produce an informative
+	 * error message. It returns a list of Python string objects.
+	 */
+	PyErr_Fetch(&err_type, &err_value, &err_traceback);
+	PyErr_NormalizeException(&err_type, &err_value, &err_traceback);
+	if (!err_traceback) goto end;
+
+	tb_module = PyImport_ImportModule("traceback");
+	if (!tb_module) goto end;
+
+	tb_dict = PyModule_GetDict(tb_module);
+	format_function = PyDict_GetItemString(tb_dict, "format_exception");
+	if (!format_function || !PyCallable_Check(format_function)) goto end;
+
+	msg_list = PyObject_CallFunction(format_function, "OOO",
+					 err_type, err_value, err_traceback);
+	if (!msg_list) goto end;
+
+	/*
+	 * Use the join() method of an empty Python string to join the list
+	 * of strings into one Python string containing the entire error
+	 * message. Then get the contents of the Python string.
+	 */
+	empty_string = PyString_FromString("");
+	if (!empty_string) goto end;
+
+	join_method = PyObject_GetAttrString(empty_string, "join");
+	if (!join_method || !PyCallable_Check(join_method)) goto end;
+
+	msg_string = PyObject_CallFunction(join_method, "O", msg_list);
+	if (!msg_string) goto end;
+
+	temp = (unsigned char *)PyString_AsString(msg_string);
+	if (temp) msg = temp;
+
+end:
+	report_scripting_error(&python_scripting_module, ses, msg);
+
+	Py_XDECREF(err_type);
+	Py_XDECREF(err_value);
+	Py_XDECREF(err_traceback);
+	Py_XDECREF(tb_module);
+	Py_XDECREF(msg_list);
+	Py_XDECREF(empty_string);
+	Py_XDECREF(join_method);
+	Py_XDECREF(msg_string);
+
+	/* In case another error occurred while reporting the original error: */
+	PyErr_Clear();
+}
 
 void
 cleanup_python(struct module *module)
 {
 	if (Py_IsInitialized()) {
-		if (pModule) {
+		Py_XDECREF(pDict);
-			Py_DECREF(pModule);
+		Py_XDECREF(pModule);
-		}
-		if (PyErr_Occurred()) {
-			PyErr_Print();
-			PyErr_Clear();
-		}
 		Py_Finalize();
 	}
 }
@@ -44,15 +110,25 @@ init_python(struct module *module)
 	if (!python_path) return;
 	env_set("PYTHONPATH", python_path, -1);
 	mem_free(python_path);
+
+	/* Treat warnings as errors so they can be caught and handled;
+	 * otherwise they would be printed to stderr.
+	 *
+	 * NOTE: PySys_ResetWarnOptions() and PySys_AddWarnOption() have been
+	 * available and stable for many years but they're not officially
+	 * documented as part of Python's public API, so in theory these two
+	 * functions might no longer be available in some hypothetical future
+	 * version of Python. */
+	PySys_ResetWarnOptions();
+	PySys_AddWarnOption("error");
+
 	Py_Initialize();
 	pModule = PyImport_ImportModule("hooks");
 
 	if (pModule) {
 		pDict = PyModule_GetDict(pModule);
+		Py_INCREF(pDict);
 	} else {
-		if (PyErr_Occurred()) {
+		alert_python_error(NULL);
-			PyErr_Print();
-			PyErr_Clear();
-		}
 	}
 }

src/scripting/python/core.h

@@ -3,7 +3,9 @@
 #define EL__SCRIPTING_PYTHON_CORE_H
 
 struct module;
+struct session;
 
+void alert_python_error(struct session *ses);
 void init_python(struct module *module);
 void cleanup_python(struct module *module);
 

src/scripting/python/hooks.c

@@ -32,29 +32,29 @@ do_script_hook_goto_url(struct session *ses, unsigned char **url)
 
 	if (pFunc && PyCallable_Check(pFunc)) {
 		PyObject *pValue;
-		unsigned char *str;
+		unsigned char *current_url;
 
 		if (!ses || !have_location(ses)) {
-			str = NULL;
+			current_url = NULL;
 		} else {
-			str = struri(cur_loc(ses)->vs.uri);
+			current_url = struri(cur_loc(ses)->vs.uri);
 		}
 
-		pValue = PyObject_CallFunction(pFunc, "ss", *url, str);
+		pValue = PyObject_CallFunction(pFunc, "ss", *url, current_url);
-		if (pValue && (pValue != Py_None)) {
+		if (pValue) {
-			const unsigned char *res = PyString_AsString(pValue);
+			if (pValue != Py_None) {
+				const unsigned char *str;
+				unsigned char *new_url;
 
-			if (res) {
+				str = PyString_AsString(pValue);
-				unsigned char *new_url = stracpy((unsigned char *)res);
+				if (str) {
-
+					new_url = stracpy((unsigned char *)str);
-				if (new_url) mem_free_set(url, new_url);
+					if (new_url) mem_free_set(url, new_url);
+				}
 			}
 			Py_DECREF(pValue);
 		} else {
-			if (PyErr_Occurred()) {
+			alert_python_error(ses);
-				PyErr_Print();
-				PyErr_Clear();
-			}
 		}
 	}
 }
@@ -72,26 +72,26 @@ script_hook_goto_url(va_list ap, void *data)
 }
 
 static void
-do_script_hook_follow_url(unsigned char **url)
+do_script_hook_follow_url(struct session *ses, unsigned char **url)
 {
 	PyObject *pFunc = PyDict_GetItemString(pDict, "follow_url_hook");
 
 	if (pFunc && PyCallable_Check(pFunc)) {
 		PyObject *pValue = PyObject_CallFunction(pFunc, "s", *url);
-		if (pValue && (pValue != Py_None)) {
+		if (pValue) {
-			const unsigned char *str = PyString_AsString(pValue);
+			if (pValue != Py_None) {
-			unsigned char *new_url;
+				const unsigned char *str;
+				unsigned char *new_url;
 
-			if (str) {
+				str = PyString_AsString(pValue);
-				new_url = stracpy((unsigned char *)str);
+				if (str) {
-				if (new_url) mem_free_set(url, new_url);
+					new_url = stracpy((unsigned char *)str);
+					if (new_url) mem_free_set(url, new_url);
+				}
 			}
 			Py_DECREF(pValue);
 		} else {
-			if (PyErr_Occurred()) {
+			alert_python_error(ses);
-				PyErr_Print();
-				PyErr_Clear();
-			}
 		}
 	}
 }
@@ -100,15 +100,17 @@ static enum evhook_status
 script_hook_follow_url(va_list ap, void *data)
 {
 	unsigned char **url = va_arg(ap, unsigned char **);
+	struct session *ses = va_arg(ap, struct session *);
 
 	if (pDict && *url)
-		do_script_hook_follow_url(url);
+		do_script_hook_follow_url(ses, url);
 
 	return EVENT_HOOK_STATUS_NEXT;
 }
 
 static void
-do_script_hook_pre_format_html(unsigned char *url, struct cache_entry *cached,
+do_script_hook_pre_format_html(struct session *ses, unsigned char *url,
+			       struct cache_entry *cached,
 			       struct fragment *fragment)
 {
 	PyObject *pFunc = PyDict_GetItemString(pDict, "pre_format_html_hook");
@@ -118,21 +120,21 @@ do_script_hook_pre_format_html(unsigned char *url, struct cache_entry *cached,
 		                                         fragment->data,
 		                                         fragment->length);
 
-		if (pValue && (pValue != Py_None)) {
+		if (pValue) {
-			const unsigned char *str = PyString_AsString(pValue);
+			if (pValue != Py_None) {
+				const unsigned char *str;
+				int len;
 
-			if (str) {
+				str = PyString_AsString(pValue);
-				int len = PyString_Size(pValue); /* strlen(str); */
+				if (str) {
-
+					len = PyString_Size(pValue);
-				add_fragment(cached, 0, str, len);
+					add_fragment(cached, 0, str, len);
-				normalize_cache_entry(cached, len);
+					normalize_cache_entry(cached, len);
+				}
 			}
 			Py_DECREF(pValue);
 		} else {
-			if (PyErr_Occurred()) {
+			alert_python_error(ses);
-				PyErr_Print();
-				PyErr_Clear();
-			}
 		}
 	}
 }
@@ -146,7 +148,7 @@ script_hook_pre_format_html(va_list ap, void *data)
 	unsigned char *url = struri(cached->uri);
 
 	if (pDict && ses && url && cached->length && *fragment->data)
-		do_script_hook_pre_format_html(url, cached, fragment);
+		do_script_hook_pre_format_html(ses, url, cached, fragment);
 
 	return EVENT_HOOK_STATUS_NEXT;
 }
@@ -159,20 +161,21 @@ do_script_hook_get_proxy(unsigned char **new_proxy_url, unsigned char *url)
 	if (pFunc && PyCallable_Check(pFunc)) {
 		PyObject *pValue = PyObject_CallFunction(pFunc, "s", url);
 
-		if (pValue && (pValue != Py_None)) {
+		if (pValue) {
-			const unsigned char *str = PyString_AsString(pValue);
+			if (pValue != Py_None) {
+				const unsigned char *str;
+				unsigned char *new_url;
 
-			if (str) {
+				str = PyString_AsString(pValue);
-				unsigned char *new_url = stracpy((unsigned char *)str);
+				if (str) {
-
+					new_url = stracpy((unsigned char *)str);
-				if (new_url) mem_free_set(new_proxy_url, new_url);
+					if (new_url) mem_free_set(new_proxy_url,
+								  new_url);
+				}
 			}
 			Py_DECREF(pValue);
 		} else {
-			if (PyErr_Occurred()) {
+			alert_python_error(NULL);
-				PyErr_Print();
-				PyErr_Clear();
-			}
 		}
 	}
 }
@@ -198,14 +201,9 @@ do_script_hook_quit(void)
 		PyObject *pValue = PyObject_CallFunction(pFunc, NULL);
 
 		if (pValue) {
-			if (pValue != Py_None) {
+			Py_DECREF(pValue);
-				Py_DECREF(pValue);
-			}
 		} else {
-			if (PyErr_Occurred()) {
+			alert_python_error(NULL);
-				PyErr_Print();
-				PyErr_Clear();
-			}
 		}
 	}
 }

src/scripting/scripting.c

@@ -24,10 +24,14 @@
 #include "scripting/ruby/ruby.h"
 #include "scripting/smjs/smjs.h"
 
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+
 
 /* Error reporting. */
 
-#if defined(CONFIG_SCRIPTING_RUBY) || defined(CONFIG_SCRIPTING_SPIDERMONKEY)
+#if defined(CONFIG_SCRIPTING_RUBY) || defined(CONFIG_SCRIPTING_SPIDERMONKEY) || defined(CONFIG_SCRIPTING_PYTHON)
 void
 report_scripting_error(struct module *module, struct session *ses,
 		       unsigned char *msg)
@@ -38,6 +42,7 @@ report_scripting_error(struct module *module, struct session *ses,
 	if (!ses) {
 		if (list_empty(terminals)) {
 			usrerror("[%s error] %s", module->name, msg);
+			sleep(3);
 			return;
 		}
 

src/session/download.c

@@ -1134,7 +1134,12 @@ do_type_query(struct type_query *type_query, unsigned char *ct, struct mime_hand
 
 		/* Let's make the filename pretty for display & save */
 		/* TODO: The filename can be the empty string here. See bug 396. */
-		decode_uri_string_for_display(&filename);
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			decode_uri_string(&filename);
+		else
+#endif /* CONFIG_UTF_8 */
+			decode_uri_string_for_display(&filename);
 	}
 
 	text = get_dialog_offset(dlg, TYPE_QUERY_WIDGETS_COUNT);

src/session/session.c

@@ -268,7 +268,12 @@ print_error_dialog(struct session *ses, enum connection_state state,
 
 	uristring = uri ? get_uri_string(uri, URI_PUBLIC) : NULL;
 	if (uristring) {
-		decode_uri_for_display(uristring);
+#ifdef CONFIG_UTF_8
+		if (ses->tab->term->utf8)
+			decode_uri(uristring);
+		else
+#endif /* CONFIG_UTF_8 */
+			decode_uri_for_display(uristring);
 		add_format_to_string(&msg,
 			_("Unable to retrieve %s", ses->tab->term),
 			uristring);

src/terminal/draw.c

@@ -7,6 +7,7 @@
 #include "elinks.h"
 
 #include "config/options.h"
+#include "intl/charsets.h"
 #include "terminal/color.h"
 #include "terminal/draw.h"
 #include "terminal/screen.h"
@@ -102,13 +103,30 @@ draw_char_color(struct terminal *term, int x, int y, struct color_pair *color)
 }
 
 void
+#ifdef CONFIG_UTF_8
+draw_char_data(struct terminal *term, int x, int y, unicode_val_T data)
+#else
 draw_char_data(struct terminal *term, int x, int y, unsigned char data)
+#endif /* CONFIG_UTF_8 */
 {
 	struct screen_char *screen_char = get_char(term, x, y);
 
 	if (!screen_char) return;
 
 	screen_char->data = data;
+
+#ifdef CONFIG_UTF_8
+#ifdef CONFIG_DEBUG
+	/* Detect attempt to draw double-width char on the last
+	 * column of terminal. */
+	if (unicode_to_cell(data) == 2 && x + 1 > term->width)
+		INTERNAL("Attempt to draw double-width glyph on last column!");
+#endif /* CONFIG_DEBUG */
+
+	if (data == UCS_NO_CHAR)
+		screen_char->attr = 0;
+#endif /* CONFIG_UTF_8 */
+
 	set_screen_dirty(term->screen, y, y);
 }
 
@@ -127,7 +145,39 @@ draw_line(struct terminal *term, int x, int y, int l, struct screen_char *line)
 	size = int_min(l, term->width - x);
 	if (size == 0) return;
 
-	copy_screen_chars(screen_char, line, size);
+#ifdef CONFIG_UTF_8
+	if (term->utf8) {
+		struct screen_char *sc;
+
+		if (line->data == UCS_NO_CHAR && x == 0) {
+			unicode_val_T data_save;
+
+			sc = line;
+			data_save = sc->data;
+			sc->data = ' ';
+			copy_screen_chars(screen_char, line, 1);
+			sc->data = data_save;
+			size--;
+			line++;
+			screen_char++;
+
+		}
+		/* Instead of displaying double-width character at last column
+		 * display only space. */
+		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 = ' ';
+			copy_screen_chars(screen_char, line, size);
+			sc->data = data_save;
+		} else {
+			copy_screen_chars(screen_char, line, size);
+		}
+	} else
+#endif
+		copy_screen_chars(screen_char, line, size);
 	set_screen_dirty(term->screen, y, y);
 }
 
@@ -198,10 +248,88 @@ draw_border(struct terminal *term, struct box *box,
 	set_screen_dirty(term->screen, borderbox.y, borderbox.y + borderbox.height);
 }
 
+#ifdef CONFIG_UTF_8
+/* Checks cells left and right to the box for broken double-width chars.
+ * Replace it with ' '.
+ * 1+---+3
+ * 1|box|##4
+ * 1|   |##4
+ * 1|   |##4
+ * 1+---+##4
+ *   2#####4
+ * 1,2,3,4 - needs to be checked, # - shadow , +,-,| - border
+ */
+void
+fix_dwchar_around_box(struct terminal *term, struct box *box, int border,
+		     int shadow_width, int shadow_height)
+{
+	struct screen_char *schar;
+	int height, x, y;
+
+	if (!term->utf8)
+		return;
+
+	/* 1 */
+	x = box->x - border - 1;
+	if (x > 0) {
+		y = box->y - border;
+		height = box->height + 2 * border;
+
+		schar = get_char(term, x, y);
+		for (;height--; schar += term->width)
+			if (unicode_to_cell(schar->data) == 2)
+				schar->data = ' ';
+	}
+
+	/* 2 */
+	x = box->x - border + shadow_width - 1;
+	if (x > 0 && x < term->width) {
+		y = box->y + border + box->height;
+		height = shadow_height;
+
+		schar = get_char(term, x, y);
+		for (;height--; schar += term->width)
+			if (unicode_to_cell(schar->data) == 2)
+				schar->data = ' ';
+	}
+
+	/* 3 */
+	x = box->x + box->width + border;
+	if (x < term->width) {
+		y = box->y - border;
+		height = shadow_height;
+
+		schar = get_char(term, x, y);
+		for (;height--; schar += term->width)
+			if (schar->data == UCS_NO_CHAR)
+				schar->data = ' ';
+	}
+
+	/* 4 */
+	x = box->x + box->width + border + shadow_width;
+	if (x < term->width) {
+		y = box->y - border + shadow_height;
+		height = box->height + 2 * border;
+
+		schar = get_char(term, x, y);
+		for (;height--; schar += term->width)
+			if (schar->data == UCS_NO_CHAR)
+				schar->data = ' ';
+	}
+}
+#endif
+
+#ifdef CONFIG_UTF_8
 void
 draw_char(struct terminal *term, int x, int y,
-	  unsigned char data, enum screen_char_attr attr,
+	  unicode_val_T data, enum screen_char_attr attr,
 	  struct color_pair *color)
+#else
+void
+draw_char(struct terminal *term, int x, int y,
+		unsigned char data, enum screen_char_attr attr,
+	  struct color_pair *color)
+#endif /* CONFIG_UTF_8 */
 {
 	struct screen_char *screen_char = get_char(term, x, y);
 
@@ -277,6 +405,80 @@ draw_shadow(struct terminal *term, struct box *box,
 	draw_box(term, &dbox, ' ', 0, color);
 }
 
+#ifdef CONFIG_UTF_8
+static void
+draw_text_utf8(struct terminal *term, int x, int y,
+	       unsigned char *text, int length,
+	       enum screen_char_attr attr, struct color_pair *color)
+{
+	struct screen_char *start, *pos;
+	unsigned char *end = text + length;
+	unicode_val_T data;
+
+	assert(text && length >= 0);
+	if_assert_failed return;
+
+	if (length <= 0) return;
+	if (x >= term->width) return;
+
+	data = utf_8_to_unicode(&text, end);
+	if (data == UCS_NO_CHAR) return;
+	start = get_char(term, x, y);
+	if (color) {
+		start->attr = attr;
+		set_term_color(start, color, 0,
+			       get_opt_int_tree(term->spec, "colors"));
+	}
+
+	if (start->data == UCS_NO_CHAR && x - 1 > 0)
+		draw_char_data(term, x - 1, y, ' ');
+
+	pos = start;
+
+	if (unicode_to_cell(data) == 2) {
+		/* Is there enough room for whole double-width char? */
+		if (x + 1 < term->width) {
+			pos->data = data;
+			pos++;
+			x++;
+
+			pos->data = UCS_NO_CHAR;
+			pos->attr = 0;
+		} else {
+			pos->data = (unicode_val_T)' ';
+		}
+	} else {
+		pos->data = data;
+	}
+	pos++;
+	x++;
+
+	for (; x < term->width; x++, pos++) {
+		data = utf_8_to_unicode(&text, end);
+		if (data == UCS_NO_CHAR) break;
+		if (color) copy_screen_chars(pos, start, 1);
+
+		if (unicode_to_cell(data) == 2) {
+			/* Is there enough room for whole double-width char? */
+			if (x + 1 < term->width) {
+				pos->data = data;
+
+				x++;
+				pos++;
+				pos->data = UCS_NO_CHAR;
+				pos->attr = 0;
+			} else {
+				pos->data = (unicode_val_T)' ';
+			}
+		} else {
+			pos->data = data;
+		}
+	}
+	set_screen_dirty(term->screen, y, y);
+
+}
+#endif /* CONFIG_UTF_8 */
+
 void
 draw_text(struct terminal *term, int x, int y,
 	  unsigned char *text, int length,
@@ -288,6 +490,13 @@ draw_text(struct terminal *term, int x, int y,
 	assert(text && length >= 0);
 	if_assert_failed return;
 
+#ifdef CONFIG_UTF_8
+	if (term->utf8) {
+		draw_text_utf8(term, x, y, text, length, attr, color);
+		return;
+	}
+#endif /* CONFIG_UTF_8 */
+
 	if (length <= 0) return;
 	pos = get_char(term, x, y);
 	if (!pos) return;

src/terminal/draw.h

@@ -1,6 +1,8 @@
 #ifndef EL__TERMINAL_DRAW_H
 #define EL__TERMINAL_DRAW_H
 
+#include "intl/charsets.h" /* unicode_val_T */
+
 struct color_pair;
 struct box;
 struct terminal;
@@ -19,7 +21,11 @@ enum screen_char_attr {
 /* One position in the terminal screen's image. */
 struct screen_char {
 	/* Contains either character value or frame data. */
+#ifdef CONFIG_UTF_8
+	unicode_val_T data;
+#else
 	unsigned char data;
+#endif /* CONFIG_UTF_8 */
 
 	/* Attributes are screen_char_attr bits. */
 	unsigned char attr;
@@ -202,7 +208,11 @@ void draw_char_color(struct terminal *term, int x, int y,
 		     struct color_pair *color);
 
 /* Sets the data of a screen position. */
+#ifdef CONFIG_UTF_8
+void draw_char_data(struct terminal *term, int x, int y, unicode_val_T data);
+#else
 void draw_char_data(struct terminal *term, int x, int y, unsigned char data);
+#endif /* CONFIG_UTF_8 */
 
 /* Sets the data to @border and of a screen position. */
 void draw_border_char(struct terminal *term, int x, int y,
@@ -213,9 +223,15 @@ void draw_border_cross(struct terminal *, int x, int y,
 		       enum border_cross_direction, struct color_pair *color);
 
 /* Draws a char. */
+#ifdef CONFIG_UTF_8
+void draw_char(struct terminal *term, int x, int y,
+	       unicode_val_T data, enum screen_char_attr attr,
+	       struct color_pair *color);
+#else
 void draw_char(struct terminal *term, int x, int y,
 	       unsigned char data, enum screen_char_attr attr,
 	       struct color_pair *color);
+#endif /* CONFIG_UTF_8 */
 
 /* Draws area defined by @box using the same colors and attributes. */
 void draw_box(struct terminal *term, struct box *box,
@@ -230,6 +246,11 @@ void draw_shadow(struct terminal *term, struct box *box,
 void draw_border(struct terminal *term, struct box *box,
 		 struct color_pair *color, int width);
 
+#ifdef CONFIG_UTF_8
+void fix_dwchar_around_box(struct terminal *term, struct box *box, int border,
+			   int shadow_width, int shadow_height);
+#endif /* CONFIG_UTF_8 */
+
 /* Draws @length chars from @text. */
 void draw_text(struct terminal *term, int x, int y,
 	       unsigned char *text, int length,

src/terminal/event.c

@@ -162,6 +162,12 @@ check_terminal_name(struct terminal *term, struct terminal_info *info)
 	object_unlock(term->spec);
 	term->spec = get_opt_rec(config_options, name);
 	object_lock(term->spec);
+#ifdef CONFIG_UTF_8
+	/* Probably not best place for set this. But now we finally have
+	 * term->spec and term->utf8 should be set before decode session info.
+	 * --Scrool */
+	term->utf8 = get_opt_bool_tree(term->spec, "utf_8_io");
+#endif /* CONFIG_UTF_8 */
 }
 
 #ifdef CONFIG_MOUSE
@@ -259,9 +265,13 @@ handle_interlink_event(struct terminal *term, struct term_event *ev)
 			return 0;
 		}
 
-		if (interlink->utf_8.len) {
+#ifdef CONFIG_UTF_8
-			utf8_io = get_opt_bool_tree(term->spec, "utf_8_io");
+		utf8_io = !!term->utf8;
+#else
+		utf8_io = get_opt_bool_tree(term->spec, "utf_8_io");
+#endif /* CONFIG_UTF_8 */
 
+		if (interlink->utf_8.len) {
 			if ((key & 0xC0) == 0x80 && utf8_io) {
 				interlink->utf_8.ucs <<= 6;
 				interlink->utf_8.ucs |= key & 0x3F;
@@ -280,11 +290,7 @@ handle_interlink_event(struct terminal *term, struct term_event *ev)
 			}
 		}
 
-		if (key < 0x80 || key > 0xFF
+		if (key < 0x80 || key > 0xFF || !utf8_io) {
-		    || (utf8_io == -1
-			? !get_opt_bool_tree(term->spec, "utf_8_io")
-			: !utf8_io)) {
-
 			term_send_event(term, ev);
 			break;
 

src/terminal/screen.c

@@ -31,6 +31,7 @@
 unsigned char frame_dumb[48] =	"   ||||++||++++++--|-+||++--|-+----++++++++     ";
 static unsigned char frame_vt100[48] =	"aaaxuuukkuxkjjjkmvwtqnttmlvwtqnvvwwmmllnnjla    ";
 
+#ifndef CONFIG_UTF_8
 /* For UTF8 I/O */
 static unsigned char frame_vt100_u[48] = {
 	177, 177, 177, 179, 180, 180, 180, 191,
@@ -40,6 +41,7 @@ static unsigned char frame_vt100_u[48] = {
 	193, 194, 194, 192, 192, 218, 218, 197,
 	197, 217, 218, 177,  32, 32,  32,  32
 };
+#endif /* CONFIG_UTF_8 */
 
 static unsigned char frame_freebsd[48] = {
 	130, 138, 128, 153, 150, 150, 150, 140,
@@ -79,6 +81,13 @@ static struct string m11_hack_frame_seqs[] = {
 	/* begin border: */	TERM_STRING("\033[11m"),
 };
 
+#ifdef CONFIG_UTF_8
+static struct string utf8_linux_frame_seqs[] = {
+	/* end border: */	TERM_STRING("\033[10m\033%G"),
+	/* begin border: */	TERM_STRING("\033%@\033[11m"),
+};
+#endif /* CONFIG_UTF_8 */
+
 static struct string vt100_frame_seqs[] = {
 	/* end border: */	TERM_STRING("\x0f"),
 	/* begin border: */	TERM_STRING("\x0e"),
@@ -100,10 +109,12 @@ struct screen_driver {
 	 * uniquely identify the screen_driver. */
 	enum term_mode_type type;
 
+#ifndef CONFIG_UTF_8
 	/* Charsets when doing UTF8 I/O. */
 	/* [0] is the common charset and [1] is the frame charset.
 	 * Test wether to use UTF8 I/O using the use_utf8_io() macro. */
 	int charsets[2];
+#endif /* CONFIG_UTF_8 */
 
 	/* The frame translation table. May be NULL. */
 	unsigned char *frame;
@@ -120,6 +131,11 @@ struct screen_driver {
 	/* These are directly derived from the terminal options. */
 	unsigned int transparent:1;
 
+#ifdef CONFIG_UTF_8
+	/* UTF-8 I/O */
+	unsigned int utf8:1;
+#endif /* CONFIG_UTF_8 */
+
 	/* The terminal._template_ name. */
 	unsigned char name[1]; /* XXX: Keep last! */
 };
@@ -127,56 +143,81 @@ struct screen_driver {
 static struct screen_driver dumb_screen_driver = {
 				NULL_LIST_HEAD,
 	/* type: */		TERM_DUMB,
+#ifndef CONFIG_UTF_8
 	/* charsets: */		{ -1, -1 },	/* No UTF8 I/O */
+#endif /* CONFIG_UTF_8 */
 	/* frame: */		frame_dumb,
 	/* frame_seqs: */	NULL,
 	/* underline: */	underline_seqs,
 	/* color_mode: */	COLOR_MODE_16,
 	/* transparent: */	1,
+#ifdef CONFIG_UTF_8
+	/* utf-8: */		0,
+#endif /* CONFIG_UTF_8 */
 };
 
 static struct screen_driver vt100_screen_driver = {
 				NULL_LIST_HEAD,
 	/* type: */		TERM_VT100,
+#ifndef CONFIG_UTF_8
 	/* charsets: */		{ -1, -1 },	/* No UTF8 I/O */
+#endif /* CONFIG_UTF_8 */
 	/* frame: */		frame_vt100,
 	/* frame_seqs: */	vt100_frame_seqs,
 	/* underline: */	underline_seqs,
 	/* color_mode: */	COLOR_MODE_16,
 	/* transparent: */	1,
+#ifdef CONFIG_UTF_8
+	/* utf-8: */		0,
+#endif /* CONFIG_UTF_8 */
 };
 
 static struct screen_driver linux_screen_driver = {
 				NULL_LIST_HEAD,
 	/* type: */		TERM_LINUX,
+#ifndef CONFIG_UTF_8
 	/* charsets: */		{ -1, -1 },	/* No UTF8 I/O */
+#endif /* CONFIG_UTF_8 */
 	/* frame: */		NULL,		/* No restrict_852 */
 	/* frame_seqs: */	NULL,		/* No m11_hack */
 	/* underline: */	underline_seqs,
 	/* color_mode: */	COLOR_MODE_16,
 	/* transparent: */	1,
+#ifdef CONFIG_UTF_8
+	/* utf-8: */		0,
+#endif /* CONFIG_UTF_8 */
 };
 
 static struct screen_driver koi8_screen_driver = {
 				NULL_LIST_HEAD,
 	/* type: */		TERM_KOI8,
+#ifndef CONFIG_UTF_8
 	/* charsets: */		{ -1, -1 },	/* No UTF8 I/O */
+#endif /* CONFIG_UTF_8 */
 	/* frame: */		frame_koi,
 	/* frame_seqs: */	NULL,
 	/* underline: */	underline_seqs,
 	/* color_mode: */	COLOR_MODE_16,
 	/* transparent: */	1,
+#ifdef CONFIG_UTF_8
+	/* utf-8: */		0,
+#endif /* CONFIG_UTF_8 */
 };
 
 static struct screen_driver freebsd_screen_driver = {
 				NULL_LIST_HEAD,
 	/* type: */		TERM_FREEBSD,
+#ifndef CONFIG_UTF_8
 	/* charsets: */		{ -1, -1 },	/* No UTF8 I/O */
+#endif /* CONFIG_UTF_8 */
 	/* frame: */		frame_freebsd,
 	/* frame_seqs: */	NULL,		/* No m11_hack */
 	/* underline: */	underline_seqs,
 	/* color_mode: */	COLOR_MODE_16,
 	/* transparent: */	1,
+#ifdef CONFIG_UTF_8
+	/* utf-8: */		0,
+#endif /* CONFIG_UTF_8 */
 };
 
 /* XXX: Keep in sync with enum term_mode_type. */
@@ -188,13 +229,22 @@ static struct screen_driver *screen_drivers[] = {
 	/* TERM_FREEBSD: */	&freebsd_screen_driver,
 };
 
+#ifdef CONFIG_UTF_8
+#define use_utf8_io(driver)	((driver)->utf8)
+#else
+#define use_utf8_io(driver)	((driver)->charsets[0] != -1)
+#endif /* CONFIG_UTF_8 */
 
 static INIT_LIST_HEAD(active_screen_drivers);
 
 static void
 update_screen_driver(struct screen_driver *driver, struct option *term_spec)
 {
+#ifdef CONFIG_UTF_8
+	driver->utf8 = get_opt_bool_tree(term_spec, "utf_8_io");
+#else
 	int utf8_io = get_opt_bool_tree(term_spec, "utf_8_io");
+#endif /* CONFIG_UTF_8 */
 
 	driver->color_mode = get_opt_int_tree(term_spec, "colors");
 	driver->transparent = get_opt_bool_tree(term_spec, "transparency");
@@ -205,6 +255,25 @@ update_screen_driver(struct screen_driver *driver, struct option *term_spec)
 		driver->underline = NULL;
 	}
 
+#ifdef CONFIG_UTF_8
+	if (driver->type == TERM_LINUX) {
+		if (get_opt_bool_tree(term_spec, "restrict_852"))
+			driver->frame = frame_restrict;
+
+		if (get_opt_bool_tree(term_spec, "m11_hack"))
+			driver->frame_seqs = m11_hack_frame_seqs;
+
+		if (driver->utf8)
+			driver->frame_seqs = utf8_linux_frame_seqs;
+
+	} else if (driver->type == TERM_FREEBSD) {
+		if (get_opt_bool_tree(term_spec, "m11_hack"))
+			driver->frame_seqs = m11_hack_frame_seqs;
+
+	} else if (driver->type == TERM_VT100) {
+		driver->frame = frame_vt100;
+	}
+#else
 	if (utf8_io) {
 		driver->charsets[0] = get_opt_codepage_tree(term_spec, "charset");
 		if (driver->type == TERM_LINUX) {
@@ -239,11 +308,11 @@ update_screen_driver(struct screen_driver *driver, struct option *term_spec)
 		} else if (driver->type == TERM_FREEBSD) {
 			if (get_opt_bool_tree(term_spec, "m11_hack"))
 				driver->frame_seqs = m11_hack_frame_seqs;
-
 		} else if (driver->type == TERM_VT100) {
 			driver->frame = frame_vt100;
 		}
-	}
+ 	}
+#endif /* CONFIG_UTF_8 */
 }
 
 static int
@@ -282,6 +351,10 @@ add_screen_driver(enum term_mode_type type, struct terminal *term, int env_len)
 
 	term->spec->change_hook = screen_driver_change_hook;
 
+#ifdef CONFIG_UTF_8
+	term->utf8 = use_utf8_io(driver);
+#endif /* CONFIG_UTF_8 */
+
 	return driver;
 }
 
@@ -300,6 +373,9 @@ get_screen_driver(struct terminal *term)
 		/* Some simple probably useless MRU ;) */
 		move_to_top_of_list(active_screen_drivers, driver);
 
+#ifdef CONFIG_UTF_8
+		term->utf8 = use_utf8_io(driver);
+#endif /* CONFIG_UTF_8 */
 		return driver;
 	}
 
@@ -366,11 +442,15 @@ struct screen_state {
 #define compare_bg_color(a, b)	(TERM_COLOR_BACKGROUND(a) == TERM_COLOR_BACKGROUND(b))
 #define compare_fg_color(a, b)	(TERM_COLOR_FOREGROUND(a) == TERM_COLOR_FOREGROUND(b))
 
-#define use_utf8_io(driver)	((driver)->charsets[0] != -1)
+#ifdef CONFIG_UTF_8
-
+static inline void
+add_char_data(struct string *screen, struct screen_driver *driver,
+	      unicode_val_T data, unsigned char border)
+#else
 static inline void
 add_char_data(struct string *screen, struct screen_driver *driver,
 	      unsigned char data, unsigned char border)
+#endif /* CONFIG_UTF_8 */
 {
 	if (!isscreensafe(data)) {
 		add_char_to_string(screen, ' ');
@@ -381,13 +461,21 @@ add_char_data(struct string *screen, struct screen_driver *driver,
 		data = driver->frame[data - 176];
 
 	if (use_utf8_io(driver)) {
+#ifdef CONFIG_UTF_8
+		if (border)
+			add_char_to_string(screen, (unsigned char)data);
+		else
+			if (data != UCS_NO_CHAR)
+				add_to_string(screen, encode_utf_8(data));
+#else
 		int charset = driver->charsets[!!border];
 
 		add_to_string(screen, cp2utf_8(charset, data));
+#endif /* CONFIG_UTF_8 */
 		return;
 	}
 
-	add_char_to_string(screen, data);
+	add_char_to_string(screen, (unsigned char)data);
 }
 
 /* Time critical section. */
@@ -399,17 +487,32 @@ add_char16(struct string *screen, struct screen_driver *driver,
 	unsigned char underline = (ch->attr & SCREEN_ATTR_UNDERLINE);
 	unsigned char bold = (ch->attr & SCREEN_ATTR_BOLD);
 
-	if (border != state->border && driver->frame_seqs) {
+	if (
+#ifdef CONFIG_UTF_8
+	    (!use_utf8_io(driver) || ch->data != UCS_NO_CHAR) &&
+#endif /* CONFIG_UTF_8 */
+	    border != state->border && driver->frame_seqs
+	   ) {
 		state->border = border;
 		add_term_string(screen, driver->frame_seqs[!!border]);
 	}
 
-	if (underline != state->underline && driver->underline) {
+	if (
+#ifdef CONFIG_UTF_8
+	    (!use_utf8_io(driver) || ch->data != UCS_NO_CHAR) &&
+#endif /* CONFIG_UTF_8 */
+	    underline != state->underline && driver->underline
+	   ) {
 		state->underline = underline;
 		add_term_string(screen, driver->underline[!!underline]);
 	}
 
-	if (bold != state->bold) {
+	if (
+#ifdef CONFIG_UTF_8
+	    (!use_utf8_io(driver) || ch->data != UCS_NO_CHAR) &&
+#endif /* CONFIG_UTF_8 */
+	    bold != state->bold
+	   ) {
 		state->bold = bold;
 		if (bold) {
 			add_bytes_to_string(screen, "\033[1m", 4);
@@ -419,7 +522,12 @@ add_char16(struct string *screen, struct screen_driver *driver,
 		}
 	}
 
-	if (!compare_color(ch->color, state->color)) {
+	if (
+#ifdef CONFIG_UTF_8
+	    (!use_utf8_io(driver) || ch->data != UCS_NO_CHAR) &&
+#endif /* CONFIG_UTF_8 */
+	    !compare_color(ch->color, state->color)
+	   ) {
 		copy_color(state->color, ch->color);
 
 		add_bytes_to_string(screen, "\033[0", 3);
@@ -521,7 +629,12 @@ add_char256(struct string *screen, struct screen_driver *driver,
 {
 	unsigned char attr_delta = (ch->attr ^ state->attr);
 
-	if (attr_delta) {
+	if (
+#ifdef CONFIG_UTF_8
+	    (!use_utf8_io(driver) || ch->data != UCS_NO_CHAR) &&
+#endif /* CONFIG_UTF_8 */
+	    attr_delta
+	   ) {
 		if ((attr_delta & SCREEN_ATTR_FRAME) && driver->frame_seqs) {
 			state->border = !!(ch->attr & SCREEN_ATTR_FRAME);
 			add_term_string(screen, driver->frame_seqs[state->border]);
@@ -544,7 +657,12 @@ add_char256(struct string *screen, struct screen_driver *driver,
 		state->attr = ch->attr;
 	}
 
-	if (!compare_color(ch->color, state->color)) {
+	if (
+#ifdef CONFIG_UTF_8
+	    (!use_utf8_io(driver) || ch->data != UCS_NO_CHAR) &&
+#endif /* CONFIG_UTF_8 */
+	    !compare_color(ch->color, state->color)
+	   ) {
 		copy_color(state->color, ch->color);
 
 		add_foreground_color(screen, color256_seqs, ch);

src/terminal/terminal.h

@@ -126,6 +126,11 @@ struct terminal {
 	 * work and even maintaining these structures ;-). */
 	unsigned int master:1;
 
+#ifdef CONFIG_UTF_8
+	/* Indicates whether UTF-8 I/O is used */
+	unsigned int utf8:1;
+#endif /* CONFIG_UTF_8 */
+
 	/* The current tab number. */
 	int current_tab;
 

src/viewer/dump/dump.c

@@ -26,6 +26,7 @@
 #include "document/options.h"
 #include "document/renderer.h"
 #include "document/view.h"
+#include "intl/charsets.h"
 #include "intl/gettext/libintl.h"
 #include "main/select.h"
 #include "main/main.h"
@@ -346,6 +347,11 @@ add_document_to_string(struct string *string, struct document *document)
 	assert(string && document);
 	if_assert_failed return NULL;
 
+#ifdef CONFIG_UTF_8
+	if (is_cp_utf8(document->options.cp))
+		goto utf_8;
+#endif /* CONFIG_UTF_8 */
+
 	for (y = 0; y < document->height; y++) {
 		int white = 0;
 		int x;
@@ -378,7 +384,45 @@ add_document_to_string(struct string *string, struct document *document)
 
 		add_char_to_string(string, '\n');
 	}
+#ifdef CONFIG_UTF_8
+	goto end;
+utf_8:
+	for (y = 0; y < document->height; y++) {
+		struct screen_char *pos = document->data[y].chars;
+		int white = 0;
+		int x;
 
+		for (x = 0; x < document->data[y].length; x++) {
+			unicode_val_T data = pos->data;
+			unsigned int frame = (pos->attr & SCREEN_ATTR_FRAME);
+
+			if (!isscreensafe(data)) {
+				white++;
+				continue;
+			} else if (frame && data >= 176 && data < 224) {
+				data = frame_dumb[data - 176];
+
+				if (data <= ' ') {
+					/* Count spaces. */
+					white++;
+				} else {
+					/* Print spaces if any. */
+					if (white) {
+						add_xchar_to_string(string, ' ', white);
+						white = 0;
+					}
+					if (frame)
+						add_char_to_string(string, data);
+					else
+						add_to_string(string, encode_utf_8(data));
+				}
+			}
+		}
+
+		add_char_to_string(string, '\n');
+	}
+end:
+#endif /* CONFIG_UTF_8 */
 	return string;
 }
 
@@ -666,6 +710,11 @@ dump_to_file(struct document *document, int fd)
 
 	if (!buf) return -1;
 
+#ifdef CONFIG_UTF_8
+	if (is_cp_utf8(document->options.cp))
+		goto utf_8;
+#endif /* CONFIG_UTF_8 */
+
 	for (y = 0; y < document->height; y++) {
 		int white = 0;
 		int x;
@@ -702,13 +751,64 @@ dump_to_file(struct document *document, int fd)
 		if (write_char('\n', fd, buf, &bptr))
 			goto fail;
 	}
+#ifdef CONFIG_UTF_8
+	goto ref;
+utf_8:
+	for (y = 0; y < document->height; y++) {
+		int white = 0;
+		int x;
+
+		for (x = 0; x < document->data[y].length; x++) {
+			unicode_val_T c;
+			unsigned char attr = document->data[y].chars[x].attr;
+
+			c = document->data[y].chars[x].data;
+
+			if ((attr & SCREEN_ATTR_FRAME)
+			    && c >= 176 && c < 224)
+				c = frame_dumb[c - 176];
+			else {
+				unsigned char *utf8_buf = encode_utf_8(c);
+
+				while (*utf8_buf) {
+					if (write_char(*utf8_buf++,
+						fd, buf, &bptr)) goto fail;
+				}
+				continue;
+			}
+
+			if (c <= ' ') {
+				/* Count spaces. */
+				white++;
+				continue;
+			}
+
+			/* Print spaces if any. */
+			while (white) {
+				if (write_char(' ', fd, buf, &bptr))
+					goto fail;
+				white--;
+			}
+
+			/* Print normal char. */
+			if (write_char(c, fd, buf, &bptr))
+				goto fail;
+		}
+
+		/* Print end of line. */
+		if (write_char('\n', fd, buf, &bptr))
+			goto fail;
+	}
+#endif /* CONFIG_UTF_8 */
 
 	if (hard_write(fd, buf, bptr) != bptr) {
 fail:
 		mem_free(buf);
 		return -1;
 	}
-
+#ifdef CONFIG_UTF_8
+ref:
+#endif /* CONFIG_UTF_8 */
 	if (document->nlinks && get_opt_bool("document.dump.references")) {
 		int x;
 		unsigned char *header = "\nReferences\n\n   Visible links\n";

src/viewer/text/form.c

@@ -163,11 +163,22 @@ init_form_state(struct form_control *fc, struct form_state *fs)
 		case FC_TEXTAREA:
 			fs->value = stracpy(fc->default_value);
 			fs->state = strlen(fc->default_value);
+#ifdef CONFIG_UTF_8
+			if (fc->type == FC_TEXT)
+				fs->state_cell = utf8_ptr2cells(fs->value, NULL);
+			if (fc->type == FC_PASSWORD)
+				fs->state_cell = utf8_ptr2chars(fs->value, NULL);
+			if (fc->type == FC_TEXTAREA)
+				fs->state_cell = 0;
+#endif /* CONFIG_UTF_8 */
 			fs->vpos = 0;
 			break;
 		case FC_FILE:
 			fs->value = stracpy("");
 			fs->state = 0;
+#ifdef CONFIG_UTF_8
+			fs->state_cell = 0;
+#endif /* CONFIG_UTF_8 */
 			fs->vpos = 0;
 			break;
 		case FC_SELECT:
@@ -330,21 +341,27 @@ draw_form_entry(struct terminal *term, struct document_view *doc_view,
 	dy = box->y - vs->y;
 	switch (fc->type) {
 		unsigned char *s;
+#ifdef CONFIG_UTF_8
+		unsigned char *text, *end;
+#endif /* CONFIG_UTF_8 */
 		int len;
 		int i, x, y;
 
 		case FC_TEXT:
 		case FC_PASSWORD:
 		case FC_FILE:
-			int_bounds(&fs->vpos, fs->state - fc->size + 1, fs->state);
 			if (!link->npoints) break;
 
 			y = link->points[0].y + dy;
 			if (!row_is_in_box(box, y))
 				break;
 
-			len = strlen(fs->value) - fs->vpos;
 			x = link->points[0].x + dx;
+#ifdef CONFIG_UTF_8
+			if (term->utf8) goto utf_8;
+#endif /* CONFIG_UTF_8 */
+			int_bounds(&fs->vpos, fs->state - fc->size + 1, fs->state);
+			len = strlen(fs->value) - fs->vpos;
 
 			for (i = 0; i < fc->size; i++, x++) {
 				unsigned char data;
@@ -360,6 +377,43 @@ draw_form_entry(struct terminal *term, struct document_view *doc_view,
 				draw_char_data(term, x, y, data);
 			}
 			break;
+#ifdef CONFIG_UTF_8
+utf_8:
+			end = NULL; /* Shut up the compiler. */
+			int_bounds(&fs->vpos, fs->state_cell - fc->size + 1, fs->state_cell);
+			if (fc->type == FC_PASSWORD)
+				len = utf8_ptr2chars(fs->value + fs->vpos, NULL);
+			else
+				len = utf8_ptr2cells(fs->value + fs->vpos, NULL);
+			text = fs->value;
+			end = strchr(text, '\0');
+
+
+			for (i = 0; i < fc->size; i++, x++) {
+				unicode_val_T data;
+
+				if (!col_is_in_box(box, x)) continue;
+
+				if (fs->value && i >= -fs->vpos && i < len) {
+					if (fc->type != FC_PASSWORD)
+						data = utf_8_to_unicode(&text, end);
+					else
+						data = '*';
+				} else
+					data = '_';
+
+				if (unicode_to_cell(data) == 2) {
+					if (i + 1 < fc->size) {
+						draw_char_data(term, x++, y, data);
+						data = UCS_NO_CHAR;
+						i++;
+					} else
+						data = ' ';
+				}
+				draw_char_data(term, x, y, data);
+			}
+			break;
+#endif /* CONFIG_UTF_8 */
 		case FC_TEXTAREA:
 			draw_textarea(term, fs, doc_view, link);
 			break;
@@ -378,6 +432,9 @@ draw_form_entry(struct terminal *term, struct document_view *doc_view,
 			else
 				/* XXX: when can this happen? --pasky */
 				s = "";
+#ifdef CONFIG_UTF_8
+			if (term->utf8) goto utf_8_select;
+#endif /* CONFIG_UTF_8 */
 			len = s ? strlen(s) : 0;
 			for (i = 0; i < link->npoints; i++) {
 				x = link->points[i].x + dx;
@@ -386,6 +443,36 @@ draw_form_entry(struct terminal *term, struct document_view *doc_view,
 					draw_char_data(term, x, y, i < len ? s[i] : '_');
 			}
 			break;
+#ifdef CONFIG_UTF_8
+utf_8_select:
+			text = s;
+			end = strchr(s, '\0');
+			len = utf8_ptr2cells(text, end);
+			for (i = 0; i < link->npoints; i++) {
+				x = link->points[i].x + dx;
+				y = link->points[i].y + dy;
+				if (is_in_box(box, x, y)) {
+					unicode_val_T data;
+					if (i < len) {
+						int cell;
+
+						data = utf_8_to_unicode(&s, end);
+						cell = unicode_to_cell(data);
+						if (i + 1 < len && cell == 2) {
+							draw_char_data(term, x++, y, data);
+
+							data = UCS_NO_CHAR;
+							i++;
+						} else if (cell == 2) {
+							data = ' ';
+						}
+					} else
+						data = '_';
+					draw_char_data(term, x, y, data);
+				}
+			}
+			break;
+#endif /* CONFIG_UTF_8 */
 		case FC_SUBMIT:
 		case FC_IMAGE:
 		case FC_RESET:
@@ -1195,6 +1282,9 @@ field_op(struct session *ses, struct document_view *doc_view,
 	unsigned char *text;
 	int length;
 	enum frame_event_status status = FRAME_EVENT_REFRESH;
+#ifdef CONFIG_UTF_8
+	int utf8 = ses->tab->term->utf8;
+#endif /* CONFIG_UTF_8 */
 
 	assert(ses && doc_view && link && ev);
 	if_assert_failed return FRAME_EVENT_OK;
@@ -1214,49 +1304,140 @@ field_op(struct session *ses, struct document_view *doc_view,
 
 	switch (action_id) {
 		case ACT_EDIT_LEFT:
-			fs->state = int_max(fs->state - 1, 0);
+#ifdef CONFIG_UTF_8
+			if (fc->type == FC_TEXTAREA) {
+				status = textarea_op_left(fs, fc, utf8);
+				break;
+			}
+			if (utf8) {
+				int old_state = fs->state;
+				unsigned char *new_value;
+
+				new_value = utf8_prevchar(fs->value + fs->state, 1, fs->value);
+				fs->state = new_value - fs->value;
+
+				if (old_state != fs->state) {
+					if (fc->type == FC_PASSWORD)
+						fs->state_cell = int_max(fs->state_cell - 1, 0);
+					else
+						fs->state_cell = int_max(utf8_char2cells(new_value, NULL) - 1, 0);
+				}
+			} else
+#endif /* CONFIG_UTF_8 */
+				fs->state = int_max(fs->state - 1, 0);
 			break;
 		case ACT_EDIT_RIGHT:
-			fs->state = int_min(fs->state + 1, strlen(fs->value));
+#ifdef CONFIG_UTF_8
+			if (fc->type == FC_TEXTAREA) {
+				status = textarea_op_right(fs, fc, utf8);
+				break;
+			}
+			if (utf8) {
+				unsigned char *text = fs->value + fs->state;
+				unsigned char *end = strchr(text, '\0');
+				int old_state = fs->state;
+				unicode_val_T data = utf_8_to_unicode(&text, end);
+
+				fs->state = (int)(text - fs->value);
+				if (old_state != fs->state) {
+					if (fc->type == FC_PASSWORD)
+						fs->state_cell = int_min(fs->state_cell + 1,
+								         utf8_ptr2cells(fs->value, NULL));
+					else
+						fs->state_cell += unicode_to_cell(data);
+				}
+			} else
+#endif /* CONFIG_UTF_8 */
+				fs->state = int_min(fs->state + 1, strlen(fs->value));
 			break;
 		case ACT_EDIT_HOME:
+#ifdef CONFIG_UTF_8
+			if (fc->type == FC_TEXTAREA) {
+				status = textarea_op_home(fs, fc, utf8);
+			} else {
+				fs->state = 0;
+				fs->state_cell = 0;
+			}
+#else
 			if (fc->type == FC_TEXTAREA) {
 				status = textarea_op_home(fs, fc);
 			} else {
 				fs->state = 0;
 			}
+
+#endif /* CONFIG_UTF_8 */
 			break;
 		case ACT_EDIT_UP:
 			if (fc->type != FC_TEXTAREA)
 				status = FRAME_EVENT_IGNORED;
 			else
+#ifdef CONFIG_UTF_8
+				status = textarea_op_up(fs, fc, utf8);
+#else
 				status = textarea_op_up(fs, fc);
+#endif /* CONFIG_UTF_8 */
 			break;
 		case ACT_EDIT_DOWN:
 			if (fc->type != FC_TEXTAREA)
 				status = FRAME_EVENT_IGNORED;
 			else
+#ifdef CONFIG_UTF_8
+				status = textarea_op_down(fs, fc, utf8);
+#else
 				status = textarea_op_down(fs, fc);
+#endif /* CONFIG_UTF_8 */
 			break;
 		case ACT_EDIT_END:
 			if (fc->type == FC_TEXTAREA) {
+#ifdef CONFIG_UTF_8
+				status = textarea_op_end(fs, fc, utf8);
+#else
 				status = textarea_op_end(fs, fc);
+#endif /* CONFIG_UTF_8 */
 			} else {
 				fs->state = strlen(fs->value);
+#ifdef CONFIG_UTF_8
+				if (utf8 && fc->type != FC_PASSWORD)
+					fs->state_cell = utf8_ptr2cells(fs->value,
+									fs->value + fs->state);
+				else if(utf8)
+					fs->state_cell = utf8_ptr2chars(fs->value,
+									fs->value + fs->state);
+#endif /* CONFIG_UTF_8 */
 			}
 			break;
 		case ACT_EDIT_BEGINNING_OF_BUFFER:
 			if (fc->type == FC_TEXTAREA) {
+#ifdef CONFIG_UTF_8
+				status = textarea_op_bob(fs, fc, utf8);
+#else
 				status = textarea_op_bob(fs, fc);
+#endif /* CONFIG_UTF_8 */
 			} else {
 				fs->state = 0;
 			}
+#ifdef CONFIG_UTF_8
+				fs->state_cell = 0;
+#endif /* CONFIG_UTF_8 */
 			break;
 		case ACT_EDIT_END_OF_BUFFER:
 			if (fc->type == FC_TEXTAREA) {
+#ifdef CONFIG_UTF_8
+				status = textarea_op_eob(fs, fc, utf8);
+#else
 				status = textarea_op_eob(fs, fc);
+#endif /* CONFIG_UTF_8 */
 			} else {
 				fs->state = strlen(fs->value);
+#ifdef CONFIG_UTF_8
+				if (utf8 && fc->type != FC_PASSWORD)
+					fs->state_cell = utf8_ptr2cells(fs->value,
+									fs->value + fs->state);
+				else if(utf8)
+					fs->state_cell = utf8_ptr2chars(fs->value,
+									fs->value + fs->state);
+
+#endif /* CONFIG_UTF_8 */
 			}
 			break;
 		case ACT_EDIT_OPEN_EXTERNAL:
@@ -1274,6 +1455,9 @@ field_op(struct session *ses, struct document_view *doc_view,
 			if (!form_field_is_readonly(fc))
 				fs->value[0] = 0;
 			fs->state = 0;
+#ifdef CONFIG_UTF_8
+			fs->state_cell = 0;
+#endif /* CONFIG_UTF_8 */
 			break;
 		case ACT_EDIT_PASTE_CLIPBOARD:
 			if (form_field_is_readonly(fc)) break;
@@ -1289,13 +1473,27 @@ field_op(struct session *ses, struct document_view *doc_view,
 					fs->value = v;
 					memmove(v, text, length + 1);
 					fs->state = strlen(fs->value);
+#ifdef CONFIG_UTF_8
+					if(utf8 && fc->type == FC_PASSWORD)
+						fs->state_cell = utf8_ptr2chars(fs->value,
+										fs->value + fs->state);
+					else if (utf8 && fc->type == FC_TEXTAREA)
+						fs->state_cell = 0;
+					else if (utf8)
+						fs->state_cell = utf8_ptr2cells(fs->value,
+										fs->value + fs->state);
+#endif /* CONFIG_UTF_8 */
 				}
 			}
 			mem_free(text);
 			break;
 		case ACT_EDIT_ENTER:
 			if (fc->type == FC_TEXTAREA) {
+#ifdef CONFIG_UTF_8
+				status = textarea_op_enter(fs, fc, utf8);
+#else
 				status = textarea_op_enter(fs, fc);
+#endif /* CONFIG_UTF_8 */
 				break;
 			}
 
@@ -1320,12 +1518,33 @@ field_op(struct session *ses, struct document_view *doc_view,
 				status = FRAME_EVENT_OK;
 				break;
 			}
+#ifdef CONFIG_UTF_8
+			if (utf8) {
+				int old_state = fs->state;
+				unsigned char *new_value;
 
-			length = strlen(fs->value + fs->state) + 1;
+				new_value = utf8_prevchar(fs->value + fs->state, 1, fs->value);
-			text = fs->value + fs->state;
+				fs->state = new_value - fs->value;
 
-			memmove(text - 1, text, length);
+				if (old_state != fs->state) {
-			fs->state--;
+					if (fc->type == FC_TEXTAREA)
+						fs->state_cell = 0;
+					else if (fc->type == FC_PASSWORD)
+						fs->state_cell = int_max(fs->state_cell - 1, 0);
+					else
+						fs->state_cell -= utf8_char2cells(new_value, NULL);
+					length = strlen(fs->value + old_state) + 1;
+					memmove(new_value, fs->value + old_state, length);
+				}
+			} else
+#endif /* CONFIG_UTF_8 */
+			{
+				length = strlen(fs->value + fs->state) + 1;
+				text = fs->value + fs->state;
+
+				memmove(text - 1, text, length);
+				fs->state--;
+			}
 			break;
 		case ACT_EDIT_DELETE:
 			if (form_field_is_readonly(fc)) {
@@ -1338,7 +1557,20 @@ field_op(struct session *ses, struct document_view *doc_view,
 				status = FRAME_EVENT_OK;
 				break;
 			}
+#ifdef CONFIG_UTF_8
+			if (utf8) {
+				unsigned char *end = fs->value + length;
+				unsigned char *text = fs->value + fs->state;
+				unsigned char *old = text;
 
+				utf_8_to_unicode(&text, end);
+				if (old != text) {
+					memmove(old, text,
+						(int)(end - text) + 1);
+				}
+				break;
+			}
+#endif /* CONFIG_UTF_8 */
 			text = fs->value + fs->state;
 
 			memmove(text, text + 1, length - fs->state);
@@ -1368,6 +1600,18 @@ field_op(struct session *ses, struct document_view *doc_view,
 			memmove(text, fs->value + fs->state, length);
 
 			fs->state = (int) (text - fs->value);
+#ifdef CONFIG_UTF_8
+			if (utf8) {
+				if(fc->type == FC_PASSWORD)
+					fs->state_cell = utf8_ptr2cells(fs->value,
+									fs->value + fs->state);
+				else if (fc->type == FC_TEXTAREA)
+					fs->state_cell = 0;
+				else
+					fs->state_cell = utf8_ptr2cells(fs->value,
+									fs->value + fs->state);
+			}
+#endif /* CONFIG_UTF_8 */
 			break;
 		case ACT_EDIT_KILL_TO_EOL:
 			if (form_field_is_readonly(fc)) {
@@ -1422,7 +1666,7 @@ field_op(struct session *ses, struct document_view *doc_view,
 			while (fs->state > 0
 			       && isspace(fs->value[fs->state - 1]))
 				--fs->state;
-			while (fs->state > 0 
+			while (fs->state > 0
 			       && !isspace(fs->value[fs->state - 1]))
 				--fs->state;
 			break;
@@ -1466,13 +1710,55 @@ field_op(struct session *ses, struct document_view *doc_view,
 			}
 
 			if (form_field_is_readonly(fc)
-			    || strlen(fs->value) >= fc->maxlength
+					|| strlen(fs->value) >= fc->maxlength
-			    || !insert_in_string(&fs->value, fs->state, "?", 1)) {
+#ifndef CONFIG_UTF_8
+					|| !insert_in_string(&fs->value, fs->state, "?", 1)
+#endif /* CONFIG_UTF_8 */
+					)
+					{
 				status = FRAME_EVENT_OK;
 				break;
 			}
+#ifdef CONFIG_UTF_8
+			if (utf8) {
+				static unsigned char buf[7];
+				static int i = 0;
+				unicode_val_T data;
+				unsigned char *t;
 
+				t = buf;
+				buf[i++] = get_kbd_key(ev);
+				buf[i] = 0;
+				data = utf_8_to_unicode(&t, buf + i);
+				if (data != UCS_NO_CHAR) {
+					if (!insert_in_string(&fs->value, fs->state, buf, i)) {
+					    	i = 0;
+						return FRAME_EVENT_OK;
+					}
+					fs->state += i;
+					if (fc->type == FC_PASSWORD)
+						fs->state_cell++;
+					else if (fc->type == FC_TEXTAREA)
+						fs->state_cell = 0;
+					else
+						fs->state_cell += unicode_to_cell(data);
+					i = 0;
+					break;
+				}
+
+				if (i == 6) {
+					i = 0;
+				}
+				return FRAME_EVENT_OK;
+
+			} else {
+				if (!insert_in_string(&fs->value, fs->state, "?", 1))
+					return FRAME_EVENT_OK;
+				fs->value[fs->state++] = get_kbd_key(ev);
+			}
+#else
 			fs->value[fs->state++] = get_kbd_key(ev);
+#endif /* CONFIG_UTF_8 */
 			break;
 	}
 

src/viewer/text/form.h

@@ -39,6 +39,9 @@ struct form_state {
 
 	unsigned char *value;
 	int state;
+#ifdef CONFIG_UTF_8
+	int state_cell;
+#endif /* CONFIG_UTF_8 */
 	int vpos;
 	int vypos;
 

src/viewer/text/link.c

@@ -114,6 +114,9 @@ get_link_cursor_offset(struct document_view *doc_view, struct link *link)
 {
 	struct form_control *fc;
 	struct form_state *fs;
+#ifdef CONFIG_UTF_8
+	int utf8 = doc_view->document->options.utf8;
+#endif /* CONFIG_UTF_8 */
 
 	switch (link->type) {
 		case LINK_CHECKBOX:
@@ -125,12 +128,21 @@ get_link_cursor_offset(struct document_view *doc_view, struct link *link)
 		case LINK_FIELD:
 			fc = get_link_form_control(link);
 			fs = find_form_state(doc_view, fc);
-			return fs ? fs->state - fs->vpos : 0;
+#ifdef CONFIG_UTF_8
+			if (utf8) {
+				return fs ? fs->state_cell - fs->vpos : 0;
+			} else
+#endif /* CONFIG_UTF_8 */
+				return fs ? fs->state - fs->vpos : 0;
 
 		case LINK_AREA:
 			fc = get_link_form_control(link);
 			fs = find_form_state(doc_view, fc);
+#ifdef CONFIG_UTF_8
+			return fs ? area_cursor(fc, fs, utf8) : 0;
+#else
 			return fs ? area_cursor(fc, fs) : 0;
+#endif /* CONFIG_UTF_8 */
 
 		case LINK_HYPERTEXT:
 		case LINK_MAP:
@@ -1342,7 +1354,7 @@ end:
 	do_menu(term, mi, ses, 1);
 }
 
-/* Return current link's title. Pretty trivial. */
+/* Return current link's title. */
 unsigned char *
 get_current_link_title(struct document_view *doc_view)
 {
@@ -1356,7 +1368,29 @@ get_current_link_title(struct document_view *doc_view)
 
 	link = get_current_link(doc_view);
 
-	return (link && link->title && *link->title) ? stracpy(link->title) : NULL;
+	if (link && link->title && *link->title) {
+		unsigned char *link_title, *src;
+		struct conv_table *convert_table;
+
+		convert_table = get_translation_table(doc_view->document->cp,
+						      doc_view->document->options.cp);
+
+		link_title = convert_string(convert_table, link->title,
+					    strlen(link->title),
+					    doc_view->document->options.cp,
+					    CSM_DEFAULT, NULL, NULL, NULL);
+		/* Remove illicit chars. */
+#ifdef CONFIG_UTF_8
+		if (link_title && !doc_view->document->options.utf8)
+#endif /* CONFIG_UTF_8 */
+			for (src = link_title; *src; src++)
+				if (!isprint(*src) || iscntrl(*src))
+					*src = '*';
+
+		return link_title;
+	}
+
+	return NULL;
 }
 
 unsigned char *
@@ -1401,7 +1435,12 @@ get_current_link_info(struct session *ses, struct document_view *doc_view)
 			add_char_to_string(&str, ')');
 		}
 
-		decode_uri_string_for_display(&str);
+#ifdef CONFIG_UTF_8
+		if (term->utf8)
+			decode_uri_string(&str);
+		else
+#endif /* CONFIG_UTF_8 */
+			decode_uri_string_for_display(&str);
 		return str.source;
 	}
 

src/viewer/text/link.h

@@ -11,6 +11,7 @@ struct session;
 struct term_event;
 struct terminal;
 struct uri;
+struct conv_table;
 
 void set_link(struct document_view *doc_view);
 void clear_link(struct terminal *term, struct document_view *doc_view);

src/viewer/text/search.c

@@ -9,6 +9,11 @@
 #endif
 
 #include <ctype.h> /* tolower(), isprint() */
+
+#if defined(CONFIG_UTF_8) && defined(HAVE_WCTYPE_H)
+#include <wctype.h>
+#endif
+
 #include <sys/types.h> /* FreeBSD needs this before regex.h */
 #ifdef HAVE_REGEX_H
 #include <regex.h>
@@ -22,6 +27,7 @@
 #include "config/kbdbind.h"
 #include "document/document.h"
 #include "document/view.h"
+#include "intl/charsets.h"
 #include "intl/gettext/libintl.h"
 #include "main/event.h"
 #include "main/module.h"
@@ -44,9 +50,15 @@
 
 static INIT_INPUT_HISTORY(search_history);
 
+#undef UCHAR
+#ifdef CONFIG_UTF_8
+#define UCHAR unicode_val_T
+#else
+#define UCHAR unsigned char
+#endif
 
 static inline void
-add_srch_chr(struct document *document, unsigned char c, int x, int y, int nn)
+add_srch_chr(struct document *document, UCHAR c, int x, int y, int nn)
 {
 	assert(document);
 	if_assert_failed return;
@@ -146,7 +158,7 @@ get_srch(struct document *document)
 			     x++);
 
 			for (; x < width; x++) {
-				unsigned char c = document->data[y].chars[x].data;
+				UCHAR c = document->data[y].chars[x].data;
 				int count = 0;
 				int xx;
 
@@ -441,18 +453,58 @@ is_in_range_regex(struct document *document, int y, int height,
 }
 #endif /* HAVE_REGEX_H */
 
-/* Returns an allocated string which is a lowered copy of passed one. */
+static UCHAR *
-static unsigned char *
+memacpy_u(unsigned char *text, int textlen, int utf8)
-lowered_string(unsigned char *text, int textlen)
 {
-	unsigned char *ret;
+#ifdef CONFIG_UTF_8
+	UCHAR *mem = mem_alloc((textlen + 1) * sizeof(UCHAR));
 
-	if (textlen < 0) textlen = strlen(text);
+	if (!mem) return NULL;
+	if (utf8) {
+		int i;
 
-	ret = mem_calloc(1, textlen + 1);
+		for (i = 0; i < textlen; i++)
+			mem[i] = utf_8_to_unicode(&text, text + 7);
+	} else {
+		int i;
+
+		for (i = 0; i < textlen; i++)
+			mem[i] = text[i];
+	}
+	mem[textlen] = 0;
+	return mem;
+#else
+	return memacpy(text, textlen);
+#endif
+}
+
+static int
+strlen_u(unsigned char *text, int utf8)
+{
+#ifdef CONFIG_UTF_8
+	if (utf8)
+		return strlen_utf8(&text);
+#endif
+	return strlen(text);
+
+}
+
+/* Returns an allocated string which is a lowered copy of passed one. */
+static UCHAR *
+lowered_string(unsigned char *text, int textlen, int utf8)
+{
+	UCHAR *ret;
+
+	if (textlen < 0) textlen = strlen_u(text, utf8);
+
+	ret = memacpy_u(text, textlen, utf8);
 	if (ret && textlen) {
 		do {
-			ret[textlen] = tolower(text[textlen]);
+#if defined(CONFIG_UTF_8) && defined(HAVE_WCTYPE_H)
+			ret[textlen] = utf8 ? towlower(ret[textlen]) : tolower(ret[textlen]);
+#else
+			ret[textlen] = tolower(ret[textlen]);
+#endif
 		} while (textlen--);
 	}
 
@@ -463,14 +515,14 @@ static int
 is_in_range_plain(struct document *document, int y, int height,
 		  unsigned char *text, int textlen,
 		  int *min, int *max,
-		  struct search *s1, struct search *s2)
+		  struct search *s1, struct search *s2, int utf8)
 {
 	int yy = y + height;
-	unsigned char *txt;
+	UCHAR *txt;
 	int found = 0;
 	int case_sensitive = get_opt_bool("document.browse.search.case");
 
-	txt = case_sensitive ? stracpy(text) : lowered_string(text, textlen);
+	txt = case_sensitive ? memacpy_u(text, textlen, utf8) : lowered_string(text, textlen, utf8);
 	if (!txt) return -1;
 
 	/* TODO: This is a great candidate for nice optimizations. Fresh CS
@@ -478,9 +530,11 @@ is_in_range_plain(struct document *document, int y, int height,
 	 * trivial, probably a starter; very fast as well) or Turbo-BM (or
 	 * maybe some other Boyer-Moore variant, I don't feel that strong in
 	 * this area), hmm?  >:) --pasky */
-
+#if defined(CONFIG_UTF_8) && defined(HAVE_WCTYPE_H)
+#define maybe_tolower(c) (case_sensitive ? (c) : utf8 ? towlower(c) : tolower(c))
+#else
 #define maybe_tolower(c) (case_sensitive ? (c) : tolower(c))
-
+#endif
 	for (; s1 <= s2; s1++) {
 		int i;
 
@@ -513,18 +567,23 @@ srch_failed:
 	return found;
 }
 
+
 static int
 is_in_range(struct document *document, int y, int height,
 	    unsigned char *text, int *min, int *max)
 {
 	struct search *s1, *s2;
 	int textlen;
+	int utf8 = 0;
 
 	assert(document && text && min && max);
 	if_assert_failed return -1;
 
+#ifdef CONFIG_UTF_8
+	utf8 = document->options.utf8;
+#endif
 	*min = INT_MAX, *max = 0;
-	textlen = strlen(text);
+	textlen = strlen_u(text, utf8);
 
 	if (get_range(document, y, height, textlen, &s1, &s2))
 		return 0;
@@ -535,7 +594,7 @@ is_in_range(struct document *document, int y, int height,
 					 min, max, s1, s2);
 #endif
 	return is_in_range_plain(document, y, height, text, textlen,
-				 min, max, s1, s2);
+				 min, max, s1, s2, utf8);
 }
 
 #define realloc_points(pts, size) \
@@ -543,24 +602,28 @@ is_in_range(struct document *document, int y, int height,
 
 static void
 get_searched_plain(struct document_view *doc_view, struct point **pt, int *pl,
-		   int l, struct search *s1, struct search *s2)
+		   int l, struct search *s1, struct search *s2, int utf8)
 {
-	unsigned char *txt;
+	UCHAR *txt;
 	struct point *points = NULL;
 	struct box *box;
 	int xoffset, yoffset;
 	int len = 0;
 	int case_sensitive = get_opt_bool("document.browse.search.case");
 
-	txt = case_sensitive ? stracpy(*doc_view->search_word)
+	txt = case_sensitive ? memacpy_u(*doc_view->search_word, l, utf8)
-			     : lowered_string(*doc_view->search_word, l);
+			     : lowered_string(*doc_view->search_word, l, utf8);
 	if (!txt) return;
 
 	box = &doc_view->box;
 	xoffset = box->x - doc_view->vs->x;
 	yoffset = box->y - doc_view->vs->y;
 
+#if defined(CONFIG_UTF_8) && defined(HAVE_WCTYPE_H)
+#define maybe_tolower(c) (case_sensitive ? (c) : utf8 ? towlower(c) : tolower(c))
+#else
 #define maybe_tolower(c) (case_sensitive ? (c) : tolower(c))
+#endif
 
 	for (; s1 <= s2; s1++) {
 		int i;
@@ -671,7 +734,7 @@ get_searched_regex(struct document_view *doc_view, struct point **pt, int *pl,
 #endif /* HAVE_REGEX_H */
 
 static void
-get_searched(struct document_view *doc_view, struct point **pt, int *pl)
+get_searched(struct document_view *doc_view, struct point **pt, int *pl, int utf8)
 {
 	struct search *s1, *s2;
 	int l;
@@ -683,7 +746,7 @@ get_searched(struct document_view *doc_view, struct point **pt, int *pl)
 		return;
 
 	get_search_data(doc_view->document);
-	l = strlen(*doc_view->search_word);
+	l = strlen_u(*doc_view->search_word, utf8);
 	if (get_range(doc_view->document, doc_view->vs->y,
 		      doc_view->box.height, l, &s1, &s2)) {
 		*pt = NULL;
@@ -697,7 +760,7 @@ get_searched(struct document_view *doc_view, struct point **pt, int *pl)
 		get_searched_regex(doc_view, pt, pl, l, s1, s2);
 	else
 #endif
-		get_searched_plain(doc_view, pt, pl, l, s1, s2);
+		get_searched_plain(doc_view, pt, pl, l, s1, s2, utf8);
 }
 
 /* Highlighting of searched strings. */
@@ -706,6 +769,7 @@ draw_searched(struct terminal *term, struct document_view *doc_view)
 {
 	struct point *pt = NULL;
 	int len = 0;
+	int utf8 = 0;
 
 	assert(term && doc_view);
 	if_assert_failed return;
@@ -713,7 +777,10 @@ draw_searched(struct terminal *term, struct document_view *doc_view)
 	if (!has_search_word(doc_view))
 		return;
 
-	get_searched(doc_view, &pt, &len);
+#ifdef CONFIG_UTF_8
+	utf8 = doc_view->document->options.utf8;
+#endif
+	get_searched(doc_view, &pt, &len, utf8);
 	if (len) {
 		int i;
 		struct color_pair *color = get_bfu_color(term, "searched");
@@ -856,6 +923,11 @@ point_intersect(struct point *p1, int l1, struct point *p2, int l2)
 static int
 find_next_link_in_search(struct document_view *doc_view, int direction)
 {
+	int utf8 = 0;
+#ifdef CONFIG_UTF_8
+	utf8 = doc_view->document->options.utf8;
+#endif
+
 	assert(doc_view && doc_view->vs);
 	if_assert_failed return 0;
 
@@ -879,7 +951,7 @@ find_next_link_in_search(struct document_view *doc_view, int direction)
 
 nt:
 		link = &doc_view->document->links[doc_view->vs->current_link];
-		get_searched(doc_view, &pt, &len);
+		get_searched(doc_view, &pt, &len, utf8);
 		if (point_intersect(pt, len, link->points, link->npoints)) {
 			mem_free(pt);
 			return 0;
@@ -1174,7 +1246,7 @@ fixup_typeahead_match(struct session *ses, struct document_view *doc_view)
 	doc_view->box.height += 1;
 }
 
-static inline unsigned char
+static inline UCHAR
 get_document_char(struct document *document, int x, int y)
 {
 	return (document->height > y && document->data[y].length > x)
@@ -1196,7 +1268,7 @@ draw_typeahead_match(struct terminal *term, struct document_view *doc_view,
 	for (i = 0, j = 0; text[j] && i < end; i++, j++) {
 		int x = link->points[i].x;
 		int y = link->points[i].y;
-		unsigned char data = get_document_char(doc_view->document, x, y);
+		UCHAR data = get_document_char(doc_view->document, x, y);
 
 		/* Text wrapping might remove space chars from the link
 		 * position array so try to align the matched typeahead text
@@ -1311,7 +1383,6 @@ text_typeahead_handler(struct input_line *line, int action_id)
 				 * clears the last search. */
 				search_for_do(ses, buffer, direction, 0);
 			}
-			goto_current_link(ses, doc_view, 0);
 			return INPUT_LINE_CANCEL;
 
 		case ACT_EDIT_PREVIOUS_ITEM:

src/viewer/text/textarea.c

@@ -40,6 +40,11 @@
 struct line_info {
 	int start;
 	int end;
+#ifdef CONFIG_UTF_8
+	int last_char_width;
+	int split_prev:1;
+	int split_next:1;
+#endif /* CONFIG_UTF_8 */
 };
 
 /* We add two extra entries to the table so the ending info can be added
@@ -47,6 +52,90 @@ struct line_info {
 #define realloc_line_info(info, size) \
 	mem_align_alloc(info, size, (size) + 3, 0xFF)
 
+#ifdef CONFIG_UTF_8
+/* Allocates a line_info table describing the layout of the textarea buffer.
+ *
+ * @width	is max width and the offset at which text will be wrapped
+ * @wrap	controls how the wrapping of text is performed
+ * @format	is non zero the @text will be modified to make it suitable for
+ *		encoding it for form posting
+ */
+static struct line_info *
+format_textutf8(unsigned char *text, int width, enum form_wrap wrap, int format)
+{
+	struct line_info *line = NULL;
+	int line_number = 0;
+	int begin = 0;
+	int pos = 0;
+	int skip;
+	unsigned char *wrappos=NULL;
+	int chars_cells=0; /* Number of console chars on line */
+
+	assert(text);
+	if_assert_failed return NULL;
+
+	/* Allocate the ending entries */
+	if (!realloc_line_info(&line, 0))
+		return NULL;
+
+	while (text[pos]) {
+		int char_cells = utf8_char2cells(&text[pos], NULL);
+
+		if (text[pos] == ' ')
+			wrappos = &text[pos];
+
+		if (text[pos] == '\n') {
+			skip = 1;
+
+		} else if (wrap == FORM_WRAP_NONE || chars_cells + char_cells < width) {
+			pos += utf8charlen(&text[pos]);
+			chars_cells += char_cells;
+			continue;
+
+		} else {
+			if (wrappos) {
+				/* When formatting text for form submitting we
+				 * have to apply the wrapping mode. */
+				if (wrap == FORM_WRAP_HARD && format)
+					*wrappos = '\n';
+				pos = wrappos - text;
+			}
+			skip = !!wrappos;
+		}
+
+		if (!realloc_line_info(&line, line_number)) {
+			mem_free_if(line);
+			return NULL;
+		}
+
+		line[line_number].last_char_width = char_cells;
+		line[line_number].split_next = !skip;
+		line[line_number].start = begin;
+		line[line_number++].end = pos;
+		line[line_number].split_prev = !skip;
+
+		begin = pos += skip;
+
+		chars_cells = 0;
+		wrappos = NULL;
+	}
+
+	line[line_number].split_next = 0;
+
+	/* Flush the last text before the loop ended */
+	line[line_number].start = begin;
+	line[line_number++].end = pos;
+
+	/* Add end marker */
+	line[line_number].start = line[line_number].end = -1;
+
+	line[line_number].split_next = line[line_number].split_prev = 0;
+	line[0].split_prev = 0;
+
+	return line;
+}
+#endif /* CONFIG_UTF_8 */
+
 /* Allocates a line_info table describing the layout of the textarea buffer.
  *
  * @width	is max width and the offset at which text will be wrapped
@@ -137,6 +226,59 @@ get_textarea_line_number(struct line_info *line, int cursor_position)
 
 /* Fixes up the vpos and vypos members of the form_state. Returns the
  * logical position in the textarea view. */
+#ifdef CONFIG_UTF_8
+int
+area_cursor(struct form_control *fc, struct form_state *fs, int utf8)
+{
+	struct line_info *line;
+	int x, y;
+
+	assert(fc && fs);
+	if_assert_failed return 0;
+
+	if (utf8)
+		line = format_textutf8(fs->value, fc->cols, fc->wrap, 0);
+	else
+		line = format_text(fs->value, fc->cols, fc->wrap, 0);
+
+	if (!line) return 0;
+
+	if (fs->state_cell)
+		y = get_textarea_line_number(line, fs->state_cell);
+	else
+		y = get_textarea_line_number(line, fs->state);
+
+	if (y == -1) {
+		mem_free(line);
+		return 0;
+	}
+
+	if (utf8) {
+		if (fs->state_cell) {
+			x = utf8_ptr2cells(fs->value + line[y].start,
+					   fs->value + fs->state_cell);
+			x += line[y].last_char_width;
+		} else
+			x = utf8_ptr2cells(fs->value + line[y].start,
+					   fs->value + fs->state);
+	} else {
+		x = fs->state - line[y].start;
+		if (fc->wrap && x == fc->cols) x--;
+	}
+
+	mem_free(line);
+
+	int_bounds(&fs->vpos, x - fc->cols + 1, x);
+	int_bounds(&fs->vypos, y - fc->rows + 1, y);
+
+	x -= fs->vpos;
+	y -= fs->vypos;
+
+	return y * fc->cols + x;
+}
+
+#else
+
 int
 area_cursor(struct form_control *fc, struct form_state *fs)
 {
@@ -169,6 +311,91 @@ area_cursor(struct form_control *fc, struct form_state *fs)
 
 	return y * fc->cols + x;
 }
+#endif
+
+#ifdef CONFIG_UTF_8
+static void
+draw_textarea_utf8(struct terminal *term, struct form_state *fs,
+	      struct document_view *doc_view, struct link *link)
+{
+	struct line_info *line, *linex;
+	struct form_control *fc;
+	struct box *box;
+	int vx, vy;
+	int sl, ye;
+	int x, xbase, y;
+
+	assert(term && doc_view && doc_view->document && doc_view->vs && link);
+	if_assert_failed return;
+	fc = get_link_form_control(link);
+	assertm(fc, "link %d has no form control", (int) (link - doc_view->document->links));
+	if_assert_failed return;
+
+	box = &doc_view->box;
+	vx = doc_view->vs->x;
+	vy = doc_view->vs->y;
+
+	if (!link->npoints) return;
+	area_cursor(fc, fs, 1);
+	linex = format_textutf8(fs->value, fc->cols, fc->wrap, 0);
+	if (!linex) return;
+	line = linex;
+	sl = fs->vypos;
+	while (line->start != -1 && sl) sl--, line++;
+
+	xbase = link->points[0].x + box->x - vx;
+	y = link->points[0].y + box->y - vy;
+	ye = y + fc->rows;
+
+	for (; line->start != -1 && y < ye; line++, y++) {
+		int i;
+		unsigned char *text, *end;
+
+		text = fs->value + line->start;
+		end = fs->value + line->end;
+
+		text += utf8_cells2bytes(text, fs->vpos, end);
+
+		if (!row_is_in_box(box, y)) continue;
+
+		for (i = 0, x = xbase; i < fc->cols; i++, x++) {
+			unicode_val_T data;
+
+			if (!col_is_in_box(box, x))
+				continue;
+
+			if (i >= -fs->vpos && text < end) {
+				int cell;
+
+				data = utf_8_to_unicode(&text, end);
+				cell = unicode_to_cell(data);
+				if (cell == 2) {
+					draw_char_data(term, x++, y, data);
+					i++;
+					data = UCS_NO_CHAR;
+				}
+
+			} else
+				data = '_';
+
+			draw_char_data(term, x, y, data);
+		}
+	}
+
+	for (; y < ye; y++) {
+		int i;
+
+		if (!row_is_in_box(box, y)) continue;
+
+		for (i = 0, x = xbase; i < fc->cols; i++, x++) {
+			if (col_is_in_box(box, x))
+				draw_char_data(term, x, y, '_');
+		}
+	}
+
+	mem_free(linex);
+}
+#endif /* CONFIG_UTF_8 */
 
 void
 draw_textarea(struct terminal *term, struct form_state *fs,
@@ -183,6 +410,13 @@ draw_textarea(struct terminal *term, struct form_state *fs,
 
 	assert(term && doc_view && doc_view->document && doc_view->vs && link);
 	if_assert_failed return;
+
+#ifdef CONFIG_UTF_8
+	if (term->utf8) {
+		draw_textarea_utf8(term, fs, doc_view, link);
+		return;
+	}
+#endif /* CONFIG_UTF_8 */
 	fc = get_link_form_control(link);
 	assertm(fc, "link %d has no form control", (int) (link - doc_view->document->links));
 	if_assert_failed return;
@@ -192,7 +426,11 @@ draw_textarea(struct terminal *term, struct form_state *fs,
 	vy = doc_view->vs->y;
 
 	if (!link->npoints) return;
+#ifdef CONFIG_UTF_8
+	area_cursor(fc, fs, 0);
+#else
 	area_cursor(fc, fs);
+#endif /* CONFIG_UTF_8 */
 	linex = format_text(fs->value, fc->cols, fc->wrap, 0);
 	if (!linex) return;
 	line = linex;
@@ -257,6 +495,7 @@ encode_textarea(struct submitted_value *sv)
 
 	/* We need to reformat text now if it has to be wrapped hard, just
 	 * before encoding it. */
+	/* TODO: Do we need here UTF-8 format or not? --scrool */
 	blabla = format_text(sv->value, fc->cols, fc->wrap, 1);
 	mem_free_if(blabla);
 
@@ -447,6 +686,39 @@ menu_textarea_edit(struct terminal *term, void *xxx, void *ses_)
 	textarea_edit(0, term, fs, doc_view, link);
 }
 
+#ifdef CONFIG_UTF_8
+static enum frame_event_status
+textarea_op(struct form_state *fs, struct form_control *fc, int utf8,
+	    int (*do_op)(struct form_state *, struct line_info *, int, int))
+{
+	struct line_info *line;
+	int current, state;
+	int state_cell;
+
+	assert(fs && fs->value && fc);
+	if_assert_failed return FRAME_EVENT_OK;
+
+	if (utf8)
+		line = format_textutf8(fs->value, fc->cols, fc->wrap, 0);
+	else
+		line = format_text(fs->value, fc->cols, fc->wrap, 0);
+	if (!line) return FRAME_EVENT_OK;
+
+	current = get_textarea_line_number(line, fs->state);
+	state = fs->state;
+	state_cell = fs->state_cell;
+	if (do_op(fs, line, current, utf8)) {
+		mem_free(line);
+		return FRAME_EVENT_IGNORED;
+	}
+
+	mem_free(line);
+	return (fs->state == state && fs->state_cell == state_cell)
+		? FRAME_EVENT_OK : FRAME_EVENT_REFRESH;
+}
+
+#else
+
 static enum frame_event_status
 textarea_op(struct form_state *fs, struct form_control *fc,
 	    int (*do_op)(struct form_state *, struct line_info *, int))
@@ -471,14 +743,88 @@ textarea_op(struct form_state *fs, struct form_control *fc,
 
 	return fs->state == state ? FRAME_EVENT_OK : FRAME_EVENT_REFRESH;
 }
+#endif /* CONFIG_UTF_8 */
+
+#ifdef CONFIG_UTF_8
+void
+new_pos(struct form_state *fs, struct line_info *line, int current, int max_cells)
+{
+	unsigned char *text = fs->value + line[current].start;
+	unsigned char *end = fs->value + line[current].end;
+	int cells = 0;
+
+	while(cells < max_cells) {
+		unicode_val_T data = utf_8_to_unicode(&text, end);
+
+		if (data == UCS_NO_CHAR) break;
+		cells += unicode_to_cell(data);
+	}
+	fs->state = (int)(text - fs->value);
+}
+#endif /* CONFIG_UTF_8 */
 
 static int
+#ifdef CONFIG_UTF_8
+do_op_home(struct form_state *fs, struct line_info *line, int current, int utf8)
+#else
 do_op_home(struct form_state *fs, struct line_info *line, int current)
+#endif /* CONFIG_UTF_8 */
 {
-	if (current != -1) fs->state = line[current].start;
+	if (current == -1)
+		return 0;
+#ifdef CONFIG_UTF_8
+	if (utf8)
+		fs->state = line[current - !!fs->state_cell].start;
+	else
+#endif /* CONFIG_UTF_8 */
+		fs->state = line[current].start;
 	return 0;
 }
 
+#ifdef CONFIG_UTF_8
+static int
+do_op_up(struct form_state *fs, struct line_info *line, int current, int utf8)
+{
+	int old_state;
+
+	if (current == -1) return 0;
+
+	if (!(current - !!fs->state_cell)) return 1;
+
+	if (!utf8) {
+		fs->state -= line[current].start - line[current-1].start;
+		int_upper_bound(&fs->state, line[current-1].end);
+		return 0;
+	}
+
+	old_state = fs->state;
+	if (fs->state_cell) {
+		int len = utf8_ptr2cells(fs->value + line[current - 1].start,
+			                 fs->value + fs->state_cell);
+
+		new_pos(fs, line, current - 2, len + line[current - 1].last_char_width);
+	} else {
+		int len = utf8_ptr2cells(fs->value + line[current].start,
+					 fs->value + fs->state);
+
+		new_pos(fs, line, current - 1, len);
+	}
+
+	if (old_state != fs->state ) {
+		if (fs->state_cell && fs->state == line[current - 1].start) {
+			unsigned char *new_value;
+
+			new_value = utf8_prevchar(fs->value + fs->state, 1, fs->value);
+			fs->state_cell = new_value - fs->value;
+		} else
+			fs->state_cell = 0;
+	}
+
+	return 0;
+}
+
+#else
+
 static int
 do_op_up(struct form_state *fs, struct line_info *line, int current)
 {
@@ -489,6 +835,49 @@ do_op_up(struct form_state *fs, struct line_info *line, int current)
 	int_upper_bound(&fs->state, line[current-1].end);
 	return 0;
 }
+#endif /* CONFIG_UTF_8 */
+
+#ifdef CONFIG_UTF_8
+static int
+do_op_down(struct form_state *fs, struct line_info *line, int current, int utf8)
+{
+	int old_state;
+
+	if (current == -1) return 0;
+
+	if (line[current + 1 - !!fs->state_cell].start == -1) return 1;
+
+	if (!utf8) {
+		fs->state += line[current+1].start - line[current].start;
+		int_upper_bound(&fs->state, line[current+1].end);
+		return 0;
+	}
+
+	old_state = fs->state;
+	if (fs->state_cell) {
+		int len = utf8_ptr2cells(fs->value + line[current - 1].start,
+			                 fs->value + fs->state_cell);
+
+		new_pos(fs, line, current, len + line[current - 1].last_char_width);
+	} else {
+		int len = utf8_ptr2cells(fs->value + line[current].start,
+					 fs->value + fs->state);
+
+		new_pos(fs, line, current + 1, len);
+	}
+	if (old_state != fs->state ) {
+		if (fs->state_cell && fs->state == line[current+1].start) {
+			unsigned char *new_value;
+
+			new_value = utf8_prevchar(fs->value + fs->state, 1, fs->value);
+			fs->state_cell = new_value - fs->value;
+		} else
+			fs->state_cell = 0;
+	}
+	return 0;
+}
+
+#else
 
 static int
 do_op_down(struct form_state *fs, struct line_info *line, int current)
@@ -500,6 +889,39 @@ do_op_down(struct form_state *fs, struct line_info *line, int current)
 	int_upper_bound(&fs->state, line[current+1].end);
 	return 0;
 }
+#endif /* CONFIG_UTF_8 */
+
+#ifdef CONFIG_UTF_8
+static int
+do_op_end(struct form_state *fs, struct line_info *line, int current, int utf8)
+{
+	if (current == -1) {
+		fs->state = strlen(fs->value);
+		return 0;
+	}
+
+	if (!utf8) {
+		int wrap = line[current + 1].start == line[current].end;
+
+		/* Don't jump to next line when wrapping. */
+		fs->state = int_max(0, line[current].end - wrap);
+		return 0;
+	}
+
+	current -= !!fs->state_cell;
+	fs->state = line[current].end;
+	if (line[current].split_next) {
+		unsigned char *new_value;
+
+		new_value = utf8_prevchar(fs->value + fs->state, 1, fs->value);
+		fs->state_cell = new_value - fs->value;
+	} else {
+		fs->state_cell = 0;
+	}
+	return 0;
+}
+
+#else
 
 static int
 do_op_end(struct form_state *fs, struct line_info *line, int current)
@@ -515,9 +937,14 @@ do_op_end(struct form_state *fs, struct line_info *line, int current)
 	}
 	return 0;
 }
+#endif /* CONFIG_UTF_8 */
 
 static int
+#ifdef CONFIG_UTF_8
+do_op_bob(struct form_state *fs, struct line_info *line, int current, int utf8)
+#else
 do_op_bob(struct form_state *fs, struct line_info *line, int current)
+#endif /* CONFIG_UTF_8 */
 {
 	if (current == -1) return 0;
 
@@ -527,7 +954,11 @@ do_op_bob(struct form_state *fs, struct line_info *line, int current)
 }
 
 static int
+#ifdef CONFIG_UTF_8
+do_op_eob(struct form_state *fs, struct line_info *line, int current, int utf8)
+#else
 do_op_eob(struct form_state *fs, struct line_info *line, int current)
+#endif /* CONFIG_UTF_8 */
 {
 	if (current == -1) {
 		fs->state = strlen(fs->value);
@@ -543,50 +974,102 @@ do_op_eob(struct form_state *fs, struct line_info *line, int current)
 	return 0;
 }
 
+#ifdef CONFIG_UTF_8
+enum frame_event_status
+textarea_op_home(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_home);
+}
+#else
 enum frame_event_status
 textarea_op_home(struct form_state *fs, struct form_control *fc)
 {
 	return textarea_op(fs, fc, do_op_home);
 }
+#endif /* CONFIG_UTF_8 */
 
+#ifdef CONFIG_UTF_8
+enum frame_event_status
+textarea_op_up(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_up);
+}
+#else
 enum frame_event_status
 textarea_op_up(struct form_state *fs, struct form_control *fc)
 {
 	return textarea_op(fs, fc, do_op_up);
 }
+#endif /* CONFIG_UTF_8 */
 
+#ifdef CONFIG_UTF_8
+enum frame_event_status
+textarea_op_down(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_down);
+}
+#else
 enum frame_event_status
 textarea_op_down(struct form_state *fs, struct form_control *fc)
 {
 	return textarea_op(fs, fc, do_op_down);
 }
+#endif /* CONFIG_UTF_8 */
 
+#ifdef CONFIG_UTF_8
+enum frame_event_status
+textarea_op_end(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_end);
+}
+#else
 enum frame_event_status
 textarea_op_end(struct form_state *fs, struct form_control *fc)
 {
 	return textarea_op(fs, fc, do_op_end);
 }
+#endif /* CONFIG_UTF_8 */
 
 /* Set the form state so the cursor is on the first line of the buffer.
  * Preserve the column if possible. */
+#ifdef CONFIG_UTF_8
+enum frame_event_status
+textarea_op_bob(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_bob);
+}
+#else
 enum frame_event_status
 textarea_op_bob(struct form_state *fs, struct form_control *fc)
 {
 	return textarea_op(fs, fc, do_op_bob);
 }
+#endif /* CONFIG_UTF_8 */
 
 /* Set the form state so the cursor is on the last line of the buffer. Preserve
  * the column if possible. This is done by getting current and last line and
  * then shifting the state by the delta of both lines start position bounding
  * the whole thing to the end of the last line. */
+#ifdef CONFIG_UTF_8
+enum frame_event_status
+textarea_op_eob(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_eob);
+}
+#else
 enum frame_event_status
 textarea_op_eob(struct form_state *fs, struct form_control *fc)
 {
 	return textarea_op(fs, fc, do_op_eob);
 }
+#endif /* CONFIG_UTF_8 */
 
 enum frame_event_status
+#ifdef CONFIG_UTF_8
+textarea_op_enter(struct form_state *fs, struct form_control *fc, int utf8)
+#else
 textarea_op_enter(struct form_state *fs, struct form_control *fc)
+#endif /* CONFIG_UTF_8 */
 {
 	assert(fs && fs->value && fc);
 	if_assert_failed return FRAME_EVENT_OK;
@@ -600,6 +1083,85 @@ textarea_op_enter(struct form_state *fs, struct form_control *fc)
 	return FRAME_EVENT_REFRESH;
 }
 
+#ifdef CONFIG_UTF_8
+static int
+do_op_left(struct form_state *fs, struct line_info *line, int current, int utf8)
+{
+	int old_state;
+	int new_state;
+	unsigned char *new_value;
+
+	if (!utf8) {
+		fs->state = int_max(fs->state - 1, 0);
+		return 0;
+	}
+
+	if (fs->state_cell) {
+		fs->state = fs->state_cell;
+		fs->state_cell = 0;
+		return 0;
+	}
+
+	old_state = fs->state;
+	new_value = utf8_prevchar(fs->value + fs->state, 1, fs->value);
+	new_state = new_value - fs->value;
+
+	if (old_state != new_state) {
+		if (old_state == line[current].start && line[current].split_prev)
+			fs->state_cell = new_state;
+		else
+			fs->state = new_state;
+	}
+	return 0;
+}
+
+static int
+do_op_right(struct form_state *fs, struct line_info *line, int current, int utf8)
+{
+	unsigned char *text, *end;
+	int old_state;
+
+	if (!utf8) {
+		/* TODO: zle */
+		fs->state = int_min(fs->state + 1, strlen(fs->value));
+		return 0;
+	}
+
+	if (fs->state_cell) {
+		fs->state_cell = 0;
+		return 0;
+	}
+
+	text = fs->value + fs->state;
+	end = strchr(text, '\0');
+	old_state = fs->state;
+	utf_8_to_unicode(&text, end);
+
+	fs->state = text - fs->value;
+
+	if (old_state != fs->state && line[current].split_next) {
+		fs->state_cell = (fs->state == line[current].end) ? old_state:0;
+	} else if (!line[current].split_next) {
+		fs->state_cell = 0;
+	}
+
+	return 0;
+}
+#endif /* CONFIG_UTF_8 */
+
+#ifdef CONFIG_UTF_8
+enum frame_event_status
+textarea_op_left(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_left);
+}
+
+enum frame_event_status
+textarea_op_right(struct form_state *fs, struct form_control *fc, int utf8)
+{
+	return textarea_op(fs, fc, utf8, do_op_right);
+}
+#endif /* CONFIG_UTF_8 */
 
 void
 set_textarea(struct document_view *doc_view, int direction)
@@ -607,6 +1169,9 @@ set_textarea(struct document_view *doc_view, int direction)
 	struct form_control *fc;
 	struct form_state *fs;
 	struct link *link;
+#ifdef CONFIG_UTF_8
+	int utf8 = doc_view->document->options.utf8;
+#endif /* CONFIG_UTF_8 */
 
 	assert(doc_view && doc_view->vs && doc_view->document);
 	assert(direction == 1 || direction == -1);
@@ -627,8 +1192,15 @@ set_textarea(struct document_view *doc_view, int direction)
 
 	/* Depending on which way we entered the textarea move cursor so that
 	 * it is available at end or start. */
+#ifdef CONFIG_UTF_8
+	if (direction == 1)
+		textarea_op_eob(fs, fc, utf8);
+	else
+		textarea_op_bob(fs, fc, utf8);
+#else
 	if (direction == 1)
 		textarea_op_eob(fs, fc);
 	else
 		textarea_op_bob(fs, fc);
+#endif /* CONFIG_UTF_8 */
 }

src/viewer/text/textarea.h

@@ -13,7 +13,11 @@ struct link;
 struct session;
 struct terminal;
 
+#ifdef CONFIG_UTF_8
+int area_cursor(struct form_control *fc, struct form_state *fs, int utf8);
+#else
 int area_cursor(struct form_control *fc, struct form_state *fs);
+#endif /* CONFIG_UTF_8 */
 void draw_textarea(struct terminal *term, struct form_state *fs, struct document_view *doc_view, struct link *link);
 unsigned char *encode_textarea(struct submitted_value *sv);
 
@@ -21,6 +25,17 @@ extern int textarea_editor;
 void textarea_edit(int, struct terminal *, struct form_state *, struct document_view *, struct link *);
 void menu_textarea_edit(struct terminal *term, void *xxx, void *ses_);
 
+#ifdef CONFIG_UTF_8
+enum frame_event_status textarea_op_home(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_up(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_down(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_end(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_bob(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_eob(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_enter(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_left(struct form_state *fs, struct form_control *fc, int utf8);
+enum frame_event_status textarea_op_right(struct form_state *fs, struct form_control *fc, int utf8);
+#else
 enum frame_event_status textarea_op_home(struct form_state *fs, struct form_control *fc);
 enum frame_event_status textarea_op_up(struct form_state *fs, struct form_control *fc);
 enum frame_event_status textarea_op_down(struct form_state *fs, struct form_control *fc);
@@ -28,6 +43,7 @@ enum frame_event_status textarea_op_end(struct form_state *fs, struct form_contr
 enum frame_event_status textarea_op_bob(struct form_state *fs, struct form_control *fc);
 enum frame_event_status textarea_op_eob(struct form_state *fs, struct form_control *fc);
 enum frame_event_status textarea_op_enter(struct form_state *fs, struct form_control *fc);
+#endif /* CONFIG_UTF_8 */
 
 void set_textarea(struct document_view *doc_view, int direction);