> (patches welcome)
My patch is attached.
Screen capture:
http://www1.interq.or.jp/~deton/lynx-wcwidth/
Known issue:
* Highlight is incorrect on selecting wrapped long wide-char link text.
* Whereis highlight position is incorrect after wide-char text.
* Not support treating double-width for East Asian Ambiguous Width characters.
(for xterm -cjk_width)
TODO:
* Use mk_wcwidth_cjk() in http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
* Add cjk_width option to lynx
diff --git a/src/GridText.c b/src/GridText.c
index e7221bf..b0e30ad 100644
--- a/src/GridText.c
+++ b/src/GridText.c
@@ -74,6 +74,9 @@
static void HText_trimHightext(HText *text, int final, int stop_before);
+#define IS_UTF_FIRST(ch) (text->T.output_utf8 && \
+ (UCH((ch))&0xc0) == 0xc0)
+
#define IS_UTF_EXTRA(ch) (text->T.output_utf8 && \
(UCH((ch))&0xc0) == 0x80)
@@ -589,8 +592,16 @@ static int ctrl_chars_on_this_line = 0; /* num of ctrl chars in current line */
static int utfxtra_on_this_line = 0; /* num of UTF-8 extra bytes in line,
they *also* count as ctrl chars. */
+#ifdef UTFWIDEWIDTH_SUPPORT
+static int utfxtracells_on_this_line = 0; /* num of UTF-8 extra cells in line */
+static int utfextracells(const char *s);
+#endif
#ifdef WIDEC_CURSES
+# ifdef UTFWIDEWIDTH_SUPPORT /* TODO: support for !WIDEC_CURSES */
+#define UTFXTRA_ON_THIS_LINE utfxtracells_on_this_line
+# else
#define UTFXTRA_ON_THIS_LINE 0
+# endif
#else
#define UTFXTRA_ON_THIS_LINE utfxtra_on_this_line
#endif
@@ -2690,6 +2701,10 @@ static HTLine *insert_blanks_in_line(HTLine *line, int line_number,
break;
ioldc++;
pre = s + 1;
+#ifdef UTFWIDEWIDTH_SUPPORT
+ if (text && text->T.output_utf8 && IS_UTF_FIRST(*s))
+ ioldc += utfextracells(s);
+#endif
}
s++;
}
@@ -2832,6 +2847,9 @@ static void split_line(HText *text, unsigned split)
ctrl_chars_on_this_line = 0; /*reset since we are going to a new line */
utfxtra_on_this_line = 0; /*reset too, we'll count them */
+#ifdef UTFWIDEWIDTH_SUPPORT
+ utfxtracells_on_this_line = 0;
+#endif
text->LastChar = ' ';
#ifdef DEBUG_APPCH
@@ -2959,6 +2977,10 @@ static void split_line(HText *text, unsigned split)
ctrl_chars_on_this_line++;
} else if (IS_UTF_EXTRA(p[i])) {
utfxtra_on_this_line++;
+#ifdef UTFWIDEWIDTH_SUPPORT
+ } else if (IS_UTF_FIRST(p[i])) {
+ utfxtracells_on_this_line += utfextracells(&p[i]);
+#endif
}
if (p[i] == LY_SOFT_HYPHEN &&
(int) text->permissible_split < i)
@@ -4112,6 +4134,18 @@ void HText_appendCharacter(HText *text, int ch)
line->data[line->size] = '\0';
utfxtra_on_this_line++;
ctrl_chars_on_this_line++;
+#ifdef UTFWIDEWIDTH_SUPPORT
+ /* update utfxtracells_on_this_line on last byte of UTF-8 sequence */
+ {
+ /* find start position of UTF-8 sequence */
+ int utff = line->size - 2;
+ while (utff > 0 && IS_UTF_EXTRA(line->data[utff]))
+ utff--;
+ int utf_xlen = UTF_XLEN(line->data[utff]);
+ if (line->size - utff == utf_xlen + 1) /* have last byte */
+ utfxtracells_on_this_line += utfextracells(&(line->data[utff]));
+ }
+#endif
return;
} else if (ch & 0x80) { /* a first char of UTF-8 sequence - kw */
if ((line->size > (MAX_LINE - 7))) {
@@ -4331,6 +4365,9 @@ void HText_appendCharacter(HText *text, int ch)
(actual
+ (int) style->rightIndent
+ ((IS_CJK_TTY && text->kanji_buf) ? 1 : 0)
+#ifdef UTFWIDEWIDTH_SUPPORT
+ + utfxtracells_on_this_line
+#endif
) >= WRAP_COLS(text))
|| (text->T.output_utf8
&& ((actual
@@ -9176,6 +9213,12 @@ static int HText_TrueLineSize(HTLine *line, HText *text, int IgnoreSpaces)
UCH(line->data[i]) != HT_NON_BREAK_SPACE &&
UCH(line->data[i]) != HT_EN_SPACE) {
true_size++;
+#ifdef UTFWIDEWIDTH_SUPPORT
+ if (text && text->T.output_utf8 &&
+ IS_UTF_FIRST(line->data[i])) {
+ true_size += utfextracells(&(line->data[i]));
+ }
+#endif
}
}
} else {
@@ -9183,6 +9226,12 @@ static int HText_TrueLineSize(HTLine *line, HText *text, int IgnoreSpaces)
if (!IsSpecialAttrChar(line->data[i]) &&
IS_UTF8_EXTRA(line->data[i])) {
true_size++;
+#ifdef UTFWIDEWIDTH_SUPPORT
+ if (text && text->T.output_utf8 &&
+ IS_UTF_FIRST(line->data[i])) {
+ true_size += utfextracells(&(line->data[i]));
+ }
+#endif
}
}
}
@@ -14869,3 +14918,16 @@ GLOBALDEF HTProtocol LYLynxCache =
{"LYNXCACHE", LYHandleCache, 0};
#endif /* GLOBALDEF_IS_MACRO */
#endif /* USE_CACHEJAR */
+
+#ifdef UTFWIDEWIDTH_SUPPORT
+static int utfextracells(const char *s)
+{
+ UCode_t ucs = UCGetUniFromUtf8String(&s);
+ if (ucs <= 0)
+ return 0;
+ int cells = wcwidth(ucs);
+ if (cells <= 1)
+ return 0;
+ return cells - 1;
+}
+#endif
_______________________________________________
Lynx-dev mailing list
Lynx-dev@nongnu.org
https://lists.nongnu.org/mailman/listinfo/lynx-dev
