Ticket #1529 (remove search converter overhead)

* add: void tty_print_anychar (int c)
    * remove superfluous search converter in convert_from_utf_to_current_c,
      convert_from_8bit_to_utf_c add converter to WEdit struct
    * fix: init defaulf codesets
    * fix: segfault on replace if LANG=C

Signed-off-by: Ilia Maslakov <il.smind@gmail.com>

edit/edit-widget.h

@@ -132,6 +132,7 @@ struct WEdit {
     /* user map stuff */
     const edit_key_map_type *user_map;
     const edit_key_map_type *ext_map;
+    GIConv converter;
 
     int extmod;
 

edit/edit.c

@@ -285,19 +285,9 @@ edit_load_file_fast (WEdit *edit, const char *filename)
 {
     long buf, buf2;
     int file = -1;
-#ifdef HAVE_CHARSET
-    const char *cp_id;
-#endif
-
     edit->curs2 = edit->last_byte;
     buf2 = edit->curs2 >> S_EDIT_BUF_SIZE;
     edit->utf8 = 0;
-#ifdef HAVE_CHARSET
-    cp_id = get_codepage_id (source_codepage);
-
-    if (cp_id != NULL)
-        edit->utf8 = str_isutf8 (cp_id);
-#endif
     if ((file = mc_open (filename, O_RDONLY | O_BINARY)) == -1) {
 	GString *errmsg = g_string_new(NULL);
 	g_string_sprintf(errmsg, _(" Cannot open %s for reading "), filename);
@@ -723,6 +713,26 @@ edit_init (WEdit *edit, int lines, int columns, const char *filename,
 	    g_free (edit);
 	return 0;
     }
+    edit->utf8 = 0;
+    edit->converter = str_cnv_from_term;
+#ifdef HAVE_CHARSET
+    const char *cp_id = NULL;
+    cp_id = get_codepage_id (source_codepage >= 0 ?
+                            source_codepage : display_codepage);
+
+    if (cp_id != NULL) {
+        GIConv conv;
+        conv = str_crt_conv_from (cp_id);
+        if (conv != INVALID_CONV) {
+            if (edit->converter != str_cnv_from_term)
+                str_close_conv (edit->converter);
+            edit->converter = conv;
+        }
+    }
+    if (cp_id != NULL)
+        edit->utf8 = str_isutf8 (cp_id);
+#endif
+
     edit->loading_done = 1;
     edit->modified = 0;
     edit->locked = 0;
@@ -1092,7 +1102,9 @@ edit_insert (WEdit *edit, int c)
 void
 edit_insert_over (WEdit * edit)
 {
-    for (int i = 0; i < edit->over_col; i++ ) {
+    int i;
+
+    for ( i = 0; i < edit->over_col; i++ ) {
         edit_insert (edit, ' ');
     }
     edit->over_col = 0;

edit/editcmd.c

@@ -2393,12 +2393,22 @@ void
 edit_select_codepage_cmd (WEdit *edit)
 {
 #ifdef HAVE_CHARSET
+    const char *cp_id = NULL;
     if (do_select_codepage ()) {
-	const char *cp_id;
-
 	cp_id = get_codepage_id (source_codepage >= 0 ?
 				    source_codepage : display_codepage);
 
+        if (cp_id != NULL) {
+            GIConv conv;
+            conv = str_crt_conv_from (cp_id);
+            if (conv != INVALID_CONV) {
+                if (edit->converter != str_cnv_from_term)
+                    str_close_conv (edit->converter);
+                edit->converter = conv;
+            }
+        }
+
+
 	if (cp_id != NULL)
 	    edit->utf8 = str_isutf8 (cp_id);
     }

edit/editdraw.c

@@ -332,18 +332,7 @@ print_to_widget (WEdit *edit, long row, int start_col, int start_col_real,
 		tty_lowlevel_setcolor (color);
 	    }
 	}
-	if ( textchar > 255 ) {
+	tty_print_anychar (textchar);
-            int res = g_unichar_to_utf8 (textchar, (char *)str);
-            if ( res == 0 ) {
-                str[0] = '.';
-                str[1] = '\0';
-            } else {
-                str[res] = '\0';
-            }
-            tty_print_string ((char *) str);
-        } else {
-            tty_print_char (textchar);
-        }
 	p++;
     }
 }
@@ -497,11 +486,11 @@ edit_draw_this_line (WEdit *edit, long b, long row, long start_col,
 #ifdef HAVE_CHARSET
 		    if ( utf8_display ) {
 		        if ( !edit->utf8 ) {
-		            c = convert_from_8bit_to_utf_c ((unsigned char) c);
+		            c = convert_from_8bit_to_utf_c ((unsigned char) c, edit->converter);
 		        }
 		    } else {
 		        if ( edit->utf8 ) {
-		            c = convert_from_utf_to_current_c (c);
+		            c = convert_from_utf_to_current_c (c, edit->converter);
 		        } else {
 #endif
 		            c = convert_to_display_c (c);

src/charsets.c

@@ -351,14 +351,12 @@ convert_from_utf_to_current (const char *str)
 }
 
 unsigned char
-convert_from_utf_to_current_c (const int input_char)
+convert_from_utf_to_current_c (const int input_char, GIConv conv)
 {
     unsigned char str[6 + 1];
     unsigned char buf_ch[6 + 1];
     unsigned char ch = '.';
-    const char *cp_from;
 
-    GIConv conv;
     int res = 0;
 
     res = g_unichar_to_utf8 (input_char, (char *)str);
@@ -367,59 +365,44 @@ convert_from_utf_to_current_c (const int input_char)
     }
     str[res] = '\0';
 
-    cp_from =  get_codepage_id ( source_codepage );
+    switch (str_translate_char (conv, (char *)str, -1, (char *)buf_ch, sizeof(buf_ch))) {
-    conv = str_crt_conv_from ( cp_from );
+    case ESTR_SUCCESS:
-
+        ch = buf_ch[0];
-    if (conv != INVALID_CONV) {
+        break;
-        switch (str_translate_char (conv, (char *)str, -1, (char *)buf_ch, sizeof(buf_ch))) {
+    case ESTR_PROBLEM:
-        case ESTR_SUCCESS:
+    case ESTR_FAILURE:
-            ch = buf_ch[0];
+        ch = '.';
-            break;
+        break;
-        case ESTR_PROBLEM:
-        case ESTR_FAILURE:
-            ch = '.';
-            break;
-        }
-        str_close_conv (conv);
     }
     return ch;
 }
 
 int
-convert_from_8bit_to_utf_c (const char input_char)
+convert_from_8bit_to_utf_c (const char input_char, GIConv conv)
 {
     unsigned char str[2];
     unsigned char buf_ch[6 + 1];
     int ch = '.';
     int res = 0;
-    GIConv conv;
-    const char *cp_from;
 
     str[0] = (unsigned char) input_char;
     str[1] = '\0';
 
-    cp_from = get_codepage_id ( source_codepage );
+    switch (str_translate_char (conv, (char *)str, -1, (char *)buf_ch, sizeof(buf_ch))) {
-    conv = str_crt_conv_from (cp_from);
+    case ESTR_SUCCESS:
-
+        res = g_utf8_get_char_validated ((char *)buf_ch, -1);
-    if (conv != INVALID_CONV) {
+        if ( res < 0 ) {
-        switch (str_translate_char (conv, (char *)str, -1, (char *)buf_ch, sizeof(buf_ch))) {
+            ch = buf_ch[0];
-        case ESTR_SUCCESS:
+        } else {
-            res = g_utf8_get_char_validated ((char *)buf_ch, -1);
+            ch = res;
-            if ( res < 0 ) {
-                ch = buf_ch[0];
-            } else {
-                ch = res;
-            }
-            break;
-        case ESTR_PROBLEM:
-        case ESTR_FAILURE:
-            ch = '.';
-            break;
         }
-        str_close_conv (conv);
+        break;
+    case ESTR_PROBLEM:
+    case ESTR_FAILURE:
+        ch = '.';
+        break;
     }
     return ch;
-
 }
 
 int

src/charsets.h

@@ -44,16 +44,16 @@ unsigned char convert_from_utf_to_current (const char *str);
  * param input_char, gunichar
  * return char in needle codepage (by global int source_codepage)
 */
-unsigned char convert_from_utf_to_current_c (const int input_char);
+unsigned char convert_from_utf_to_current_c (const int input_char, GIConv conv);
 /*
  * Converter from selected codepage 8-bit
- * param char input_char
+ * param char input_char, GIConv converter
  * return int utf char
 */
-int convert_from_8bit_to_utf_c (const char input_char);
+int convert_from_8bit_to_utf_c (const char input_char, GIConv conv);
 /*
  * Converter from display codepage 8-bit to utf-8
- * param char input_char
+ * param char input_char, GIConv converter
  * return int utf char
 */
 int convert_from_8bit_to_utf_c2 (const char input_char);

src/main.c

@@ -1313,7 +1313,6 @@ static void check_codeset()
         _system_codepage = str_detect_termencoding();
         _source_codepage = get_codepage_id (source_codepage);
         _display_codepage = get_codepage_id (display_codepage);
-
         if ( (strcmp (_system_codepage, _display_codepage)) ||
              (strcmp (_system_codepage, _source_codepage)) ) {
             if (quick_dialog (&ecs) == B_ENTER){
@@ -1327,6 +1326,8 @@ static void check_codeset()
                     source_codepage = display_codepage;
                     cp_source = cp_display;
                     profile_changed = 1;
+                } else {
+                    utf8_display = str_isutf8 (_system_codepage);
                 }
             } else {
                 if ( skip_check_codeset ) {

src/tty/tty-ncurses.c

@@ -6,6 +6,7 @@
 
    Written by:
    Andrew Borodin <aborodin@vmail.ru>, 2009.
+   Ilia Maslakov <il.smind@gmail.com>, 2009.
 
    This file is part of the Midnight Commander.
 
@@ -284,6 +285,25 @@ tty_print_char (int c)
     addch (c);
 }
 
+void
+tty_print_anychar (int c)
+{
+    unsigned char str[6 + 1];
+
+    if ( c > 255 ) {
+        int res = g_unichar_to_utf8 (c, (char *)str);
+        if ( res == 0 ) {
+            str[0] = '.';
+            str[1] = '\0';
+        } else {
+            str[res] = '\0';
+        }
+        addstr (str_term_form (s));
+    } else {
+        addch (c);
+    }
+}
+
 void
 tty_print_alt_char (int c)
 {

src/tty/tty-slang.c

@@ -471,6 +471,25 @@ tty_print_alt_char (int c)
     SLsmg_draw_object (SLsmg_get_row(), SLsmg_get_column(), c);
 }
 
+void
+tty_print_anychar (int c)
+{
+    unsigned char str[6 + 1];
+
+    if ( c > 255 ) {
+        int res = g_unichar_to_utf8 (c, (char *)str);
+        if ( res == 0 ) {
+            str[0] = '.';
+            str[1] = '\0';
+        } else {
+            str[res] = '\0';
+        }
+        SLsmg_write_string ((char *) str_term_form (str));
+    } else {
+        SLsmg_write_char ((SLwchar_Type) ((unsigned int) c));
+    }
+}
+
 void
 tty_print_string (const char *s)
 {

src/tty/tty.h

@@ -65,6 +65,7 @@ extern void tty_set_alt_charset (gboolean alt_charset);
 extern void tty_display_8bit (gboolean what);
 extern void tty_print_char(int c);
 extern void tty_print_alt_char(int c);
+extern void tty_print_anychar(int c);
 extern void tty_print_string(const char *s);
 extern void tty_printf(const char *s, ...);