fix code conversions, add conv_iconv_strdup_with_cd(), and use EUC-JP-MS for JIS to UTF-8 conversion if available.

git-svn-id: svn://sylpheed.sraoss.jp/sylpheed/trunk@7 ee746299-78ed-0310-b773-934348b2243d

8 files changed, 107 insertions, 100 deletions

diff --git a/src/codeconv.c b/src/codeconv.c
index 2ebbbee3..1440f80e 100644
--- a/src/codeconv.c
+++ b/src/codeconv.c
@@ -31,9 +31,7 @@
 #  include <locale.h>
 #endif
 
-#if HAVE_ICONV
-#  include <iconv.h>
-#endif
+#include <iconv.h>
 
 #include "intl.h"
 #include "codeconv.h"
@@ -375,9 +373,34 @@ void conv_sjistoeuc(gchar *outbuf, gint outlen, const gchar *inbuf)
 
 void conv_jistoutf8(gchar *outbuf, gint outlen, const gchar *inbuf)
 {
+	static iconv_t cd = (iconv_t)-1;
+	static gboolean iconv_ok = TRUE;
 	gchar *tmpstr;
+	gchar *eucstr;
+
+	Xalloca(eucstr, outlen, return);
+
+	conv_jistoeuc(eucstr, outlen, inbuf);
+
+	if (cd == (iconv_t)-1) {
+		if (!iconv_ok) {
+			strncpy2(outbuf, inbuf, outlen);
+			return;
+		}
+		cd = iconv_open(CS_UTF_8, CS_EUC_JP_MS);
+		if (cd == (iconv_t)-1) {
+			cd = iconv_open(CS_UTF_8, CS_EUC_JP);
+			if (cd == (iconv_t)-1) {
+				g_warning("conv_jistoutf8(): %s\n",
+					  g_strerror(errno));
+				iconv_ok = FALSE;
+				strncpy2(outbuf, inbuf, outlen);
+				return;
+			}
+		}
+	}
 
-	tmpstr = conv_iconv_strdup(inbuf, CS_ISO_2022_JP, CS_UTF_8);
+	tmpstr = conv_iconv_strdup_with_cd(eucstr, cd);
 	if (tmpstr) {
 		strncpy2(outbuf, tmpstr, outlen);
 		g_free(tmpstr);
@@ -864,9 +887,9 @@ CodeConvFunc conv_get_code_conv_func(const gchar *src_charset_str,
 
 	/* auto detection mode */
 	if (!src_charset_str && !dest_charset_str) {
-		//if (src_charset == C_EUC_JP || src_charset == C_SHIFT_JIS)
-		//	return conv_anytodisp;
-		//else
+		if (src_charset == C_EUC_JP || src_charset == C_SHIFT_JIS)
+			return conv_anytodisp;
+		else
 			return conv_noconv;
 	}
 
@@ -934,20 +957,12 @@ gchar *conv_iconv_strdup(const gchar *inbuf,
 			 const gchar *src_code, const gchar *dest_code)
 {
 	iconv_t cd;
-	const gchar *inbuf_p;
 	gchar *outbuf;
-	gchar *outbuf_p;
-	size_t in_size;
-	size_t in_left;
-	size_t out_size;
-	size_t out_left;
-	size_t n_conv;
-	size_t len;
 
 	if (!src_code)
 		src_code = conv_get_outgoing_charset_str();
 	if (!dest_code)
-		dest_code = conv_get_locale_charset_str();
+		dest_code = conv_get_internal_charset_str();
 
 	/* don't convert if current codeset is US-ASCII */
 	if (!strcasecmp(dest_code, CS_US_ASCII))
@@ -961,6 +976,25 @@ gchar *conv_iconv_strdup(const gchar *inbuf,
 	if (cd == (iconv_t)-1)
 		return NULL;
 
+	outbuf = conv_iconv_strdup_with_cd(inbuf, cd);
+
+	iconv_close(cd);
+
+	return outbuf;
+}
+
+gchar *conv_iconv_strdup_with_cd(const gchar *inbuf, iconv_t cd)
+{
+	const gchar *inbuf_p;
+	gchar *outbuf;
+	gchar *outbuf_p;
+	size_t in_size;
+	size_t in_left;
+	size_t out_size;
+	size_t out_left;
+	size_t n_conv;
+	size_t len;
+
 	inbuf_p = inbuf;
 	in_size = strlen(inbuf);
 	in_left = in_size;
@@ -981,6 +1015,7 @@ gchar *conv_iconv_strdup(const gchar *inbuf,
 	while ((n_conv = iconv(cd, (ICONV_CONST gchar **)&inbuf_p, &in_left,
 			       &outbuf_p, &out_left)) == (size_t)-1) {
 		if (EILSEQ == errno) {
+			g_print("iconv(): at %d: %s\n", in_size - in_left, g_strerror(errno));
 			inbuf_p++;
 			in_left--;
 			if (out_left == 0) {
@@ -1016,8 +1051,6 @@ gchar *conv_iconv_strdup(const gchar *inbuf,
 	outbuf = g_realloc(outbuf, len + 1);
 	outbuf[len] = '\0';
 
-	iconv_close(cd);
-
 	return outbuf;
 }
 
@@ -1070,6 +1103,7 @@ static const struct {
 	{C_ISO_2022_JP_3,	CS_ISO_2022_JP_3},
 	{C_EUC_JP,		CS_EUC_JP},
 	{C_EUC_JP,		CS_EUCJP},
+	{C_EUC_JP_MS,		CS_EUC_JP_MS},
 	{C_SHIFT_JIS,		CS_SHIFT_JIS},
 	{C_SHIFT_JIS,		CS_SHIFT__JIS},
 	{C_SHIFT_JIS,		CS_SJIS},
@@ -1459,6 +1493,7 @@ gboolean conv_is_multibyte_encoding(CharSet encoding)
 {
 	switch (encoding) {
 	case C_EUC_JP:
+	case C_EUC_JP_MS:
 	case C_EUC_KR:
 	case C_EUC_TW:
 	case C_EUC_CN:
@@ -1574,6 +1609,8 @@ void conv_encode_header(gchar *dest, gint len, const gchar *src,
 	guchar *destp = dest;
 	gboolean use_base64;
 
+	g_return_if_fail(g_utf8_validate(src, -1, NULL) == TRUE);
+
 	if (MB_CUR_MAX > 1) {
 		use_base64 = TRUE;
 		mimesep_enc = "?B?";
@@ -1582,9 +1619,7 @@ void conv_encode_header(gchar *dest, gint len, const gchar *src,
 		mimesep_enc = "?Q?";
 	}
 
-	cur_encoding = conv_get_locale_charset_str();
-	if (!strcmp(cur_encoding, CS_US_ASCII))
-		cur_encoding = CS_ISO_8859_1;
+	cur_encoding = conv_get_internal_charset_str();
 	out_encoding = conv_get_outgoing_charset_str();
 	if (!strcmp(out_encoding, CS_US_ASCII))
 		out_encoding = CS_ISO_8859_1;
@@ -1646,14 +1681,7 @@ void conv_encode_header(gchar *dest, gint len, const gchar *src,
 				if (addr_field && (*p == '(' || *p == ')'))
 					break;
 
-				if (MB_CUR_MAX > 1) {
-					mb_len = mblen(p, MB_CUR_MAX);
-					if (mb_len < 0) {
-						g_warning("conv_encode_header(): invalid multibyte character encountered\n");
-						mb_len = 1;
-					}
-				} else
-					mb_len = 1;
+				mb_len = g_utf8_skip[*p];
 
 				Xstrndup_a(part_str, srcp, cur_len + mb_len, );
 				out_str = conv_codeset_strdup
diff --git a/src/codeconv.h b/src/codeconv.h
index 6a498af6..71e0a20b 100644
--- a/src/codeconv.h
+++ b/src/codeconv.h
@@ -25,6 +25,7 @@
 #endif
 
 #include <glib.h>
+#include <iconv.h>
 
 typedef struct _CodeConverter	CodeConverter;
 
@@ -74,6 +75,7 @@ typedef enum
 	C_ISO_2022_JP_2,
 	C_ISO_2022_JP_3,
 	C_EUC_JP,
+	C_EUC_JP_MS,
 	C_SHIFT_JIS,
 	C_ISO_2022_KR,
 	C_EUC_KR,
@@ -145,6 +147,7 @@ struct _CodeConverter
 #define CS_ISO_2022_JP_3	"ISO-2022-JP-3"
 #define CS_EUC_JP		"EUC-JP"
 #define CS_EUCJP		"EUCJP"
+#define CS_EUC_JP_MS		"EUC-JP-MS"
 #define CS_SHIFT_JIS		"Shift_JIS"
 #define CS_SHIFT__JIS		"SHIFT-JIS"
 #define CS_SJIS			"SJIS"
@@ -208,6 +211,8 @@ CodeConvFunc conv_get_code_conv_func	(const gchar	*src_charset_str,
 gchar *conv_iconv_strdup		(const gchar	*inbuf,
 					 const gchar	*src_code,
 					 const gchar	*dest_code);
+gchar *conv_iconv_strdup_with_cd	(const gchar	*inbuf,
+					 iconv_t	 cd);
 
 const gchar *conv_get_charset_str		(CharSet	 charset);
 CharSet conv_get_charset_from_str		(const gchar	*charset);
diff --git a/src/compose.c b/src/compose.c
index 2cfa22cd..5a8cce4d 100644
--- a/src/compose.c
+++ b/src/compose.c
@@ -1559,7 +1559,9 @@ static gchar *compose_get_signature_str(Compose *compose)
 
 	if (sig_str) {
 		utf8_sig_str = conv_codeset_strdup
-			(sig_str, conv_get_locale_charset_str(), CS_UTF_8);
+			(sig_str,
+			 conv_get_locale_charset_str(),
+			 conv_get_internal_charset_str());
 		g_free(sig_str);
 	}
 
@@ -1593,7 +1595,8 @@ static void compose_insert_file(Compose *compose, const gchar *file)
 	while (fgets(buf, sizeof(buf), fp) != NULL) {
 		gchar *str;
 
-		str = conv_codeset_strdup(buf, cur_encoding, CS_UTF_8);
+		str = conv_codeset_strdup
+			(buf, cur_encoding, conv_get_internal_charset_str());
 		if (!str) continue;
 
 		/* strip <CR> if DOS/Windows file,
@@ -2755,7 +2758,7 @@ static gint compose_write_to_file(Compose *compose, const gchar *file,
 			encoding = ENC_BASE64;
 #endif
 
-		src_codeset = CS_UTF_8;
+		src_codeset = conv_get_internal_charset_str();
 
 		debug_print("src encoding = %s, out encoding = %s, transfer encoding = %s\n",
 			    src_codeset, out_codeset, procmime_get_encoding_str(encoding));
@@ -2942,8 +2945,9 @@ static gint compose_write_body_to_file(Compose *compose, const gchar *file)
 	gtk_text_buffer_get_end_iter(buffer, &end);
 	tmp = gtk_text_buffer_get_text(buffer, &start, &end, FALSE);
 
-	chars = conv_codeset_strdup
-		(tmp, CS_UTF_8, conv_get_locale_charset_str());
+	chars = conv_codeset_strdup(tmp,
+				    conv_get_internal_charset_str(),
+				    conv_get_locale_charset_str());
 
 	g_free(tmp);
 
@@ -3694,24 +3698,13 @@ static gint compose_redirect_write_headers(Compose *compose, FILE *fp)
 static void compose_convert_header(gchar *dest, gint len, gchar *src,
 				   gint header_len, gboolean addr_field)
 {
-	gchar *str;
-	const gchar *cur_encoding;
-
 	g_return_if_fail(src != NULL);
 	g_return_if_fail(dest != NULL);
 
 	if (len < 1) return;
 
 	g_strchomp(src);
-
-#warning FIXME_GTK2 redundant code conversion
-	cur_encoding = conv_get_locale_charset_str();
-	if (!strcmp(cur_encoding, CS_US_ASCII))
-		cur_encoding = CS_ISO_8859_1;
-	str = conv_codeset_strdup(src, CS_UTF_8, cur_encoding);
-	if (str)
-		conv_encode_header(dest, len, str, header_len, addr_field);
-	g_free(str);
+	conv_encode_header(dest, len, src, header_len, addr_field);
 }
 
 static void compose_generate_msgid(Compose *compose, gchar *buf, gint len)
diff --git a/src/imap.c b/src/imap.c
index c454e855..cbd430ad 100644
--- a/src/imap.c
+++ b/src/imap.c
@@ -350,8 +350,8 @@ static gchar *search_array_str			(GPtrArray	*array,
 static void imap_path_separator_subst		(gchar		*str,
 						 gchar		 separator);
 
-static gchar *imap_modified_utf7_to_locale	(const gchar	*mutf7_str);
-static gchar *imap_locale_to_modified_utf7	(const gchar	*from);
+static gchar *imap_modified_utf7_to_utf8	(const gchar	*mutf7_str);
+static gchar *imap_utf8_to_modified_utf7	(const gchar	*from);
 
 static GSList *imap_get_seq_set_from_msglist	(GSList		*msglist);
 static void imap_seq_set_free			(GSList		*seq_list);
@@ -1765,8 +1765,8 @@ static GSList *imap_parse_list(IMAPSession *session, const gchar *real_path,
 		name = g_basename(buf);
 		if (name[0] == '.') continue;
 
-		loc_name = imap_modified_utf7_to_locale(name);
-		loc_path = imap_modified_utf7_to_locale(buf);
+		loc_name = imap_modified_utf7_to_utf8(name);
+		loc_path = imap_modified_utf7_to_utf8(buf);
 		new_item = folder_item_new(loc_name, loc_path);
 		if (strcasestr(flags, "\\Noinferiors") != NULL)
 			new_item->no_sub = TRUE;
@@ -1889,7 +1889,7 @@ static FolderItem *imap_create_folder(Folder *folder, FolderItem *parent,
 
 	/* keep trailing directory separator to create a folder that contains
 	   sub folder */
-	imap_path = imap_locale_to_modified_utf7(dirpath);
+	imap_path = imap_utf8_to_modified_utf7(dirpath);
 	strtailchomp(dirpath, '/');
 	Xstrdup_a(new_name, name, {g_free(dirpath); return NULL;});
 	strtailchomp(new_name, '/');
@@ -1989,7 +1989,7 @@ static gint imap_rename_folder(Folder *folder, FolderItem *item,
 	} else
 		newpath = g_strdup(name);
 
-	real_newpath = imap_locale_to_modified_utf7(newpath);
+	real_newpath = imap_utf8_to_modified_utf7(newpath);
 	imap_path_separator_subst(real_newpath, separator);
 
 	ok = imap_cmd_rename(session, real_oldpath, real_newpath);
@@ -2432,7 +2432,7 @@ static gchar *imap_get_real_path(IMAPFolder *folder, const gchar *path)
 	g_return_val_if_fail(folder != NULL, NULL);
 	g_return_val_if_fail(path != NULL, NULL);
 
-	real_path = imap_locale_to_modified_utf7(path);
+	real_path = imap_utf8_to_modified_utf7(path);
 	separator = imap_get_path_separator(folder, path);
 	imap_path_separator_subst(real_path, separator);
 
@@ -3645,7 +3645,7 @@ static void imap_path_separator_subst(gchar *str, gchar separator)
 	}
 }
 
-static gchar *imap_modified_utf7_to_locale(const gchar *mutf7_str)
+static gchar *imap_modified_utf7_to_utf8(const gchar *mutf7_str)
 {
 	static iconv_t cd = (iconv_t)-1;
 	static gboolean iconv_ok = TRUE;
@@ -3660,15 +3660,16 @@ static gchar *imap_modified_utf7_to_locale(const gchar *mutf7_str)
 	if (!iconv_ok) return g_strdup(mutf7_str);
 
 	if (cd == (iconv_t)-1) {
-		cd = iconv_open(conv_get_locale_charset_str(), "UTF-7");
+		cd = iconv_open(conv_get_internal_charset_str(), CS_UTF_7);
 		if (cd == (iconv_t)-1) {
 			g_warning("iconv cannot convert UTF-7 to %s\n",
-				  conv_get_locale_charset_str());
+				  conv_get_internal_charset_str());
 			iconv_ok = FALSE;
 			return g_strdup(mutf7_str);
 		}
 	}
 
+	/* modified UTF-7 to normal UTF-7 conversion */
 	norm_utf7 = g_string_new(NULL);
 
 	for (p = mutf7_str; *p != '\0'; p++) {
@@ -3715,7 +3716,7 @@ static gchar *imap_modified_utf7_to_locale(const gchar *mutf7_str)
 	return to_str;
 }
 
-static gchar *imap_locale_to_modified_utf7(const gchar *from)
+static gchar *imap_utf8_to_modified_utf7(const gchar *from)
 {
 	static iconv_t cd = (iconv_t)-1;
 	static gboolean iconv_ok = TRUE;
@@ -3728,15 +3729,16 @@ static gchar *imap_locale_to_modified_utf7(const gchar *from)
 	if (!iconv_ok) return g_strdup(from);
 
 	if (cd == (iconv_t)-1) {
-		cd = iconv_open("UTF-7", conv_get_locale_charset_str());
+		cd = iconv_open(CS_UTF_7, conv_get_internal_charset_str());
 		if (cd == (iconv_t)-1) {
 			g_warning(_("iconv cannot convert %s to UTF-7\n"),
-				  conv_get_locale_charset_str());
+				  conv_get_internal_charset_str());
 			iconv_ok = FALSE;
 			return g_strdup(from);
 		}
 	}
 
+	/* UTF-8 to normal UTF-7 conversion */
 	Xstrdup_a(from_tmp, from, return g_strdup(from));
 	from_len = strlen(from);
 	norm_utf7_len = from_len * 5;
@@ -3760,18 +3762,13 @@ static gchar *imap_locale_to_modified_utf7(const gchar *from)
 			from_tmp++;
 			from_len--;
 		} else {
-			size_t mb_len = 0, conv_len = 0;
+			size_t conv_len = 0;
 
 			/* unprintable char: convert to UTF-7 */
 			p = from_tmp;
 			while (!IS_PRINT(*(guchar *)p) && conv_len < from_len) {
-				mb_len = mblen(p, MB_LEN_MAX);
-				if (mb_len <= 0) {
-					g_warning("wrong multibyte sequence\n");
-					return g_strdup(from);
-				}
-				conv_len += mb_len;
-				p += mb_len;
+				conv_len += g_utf8_skip[*(guchar *)p];
+				p += g_utf8_skip[*(guchar *)p];
 			}
 
 			from_len -= conv_len;
diff --git a/src/statusbar.c b/src/statusbar.c
index 38d1377f..8ddc9508 100644
--- a/src/statusbar.c
+++ b/src/statusbar.c
@@ -49,26 +49,12 @@ void statusbar_puts(GtkStatusbar *statusbar, const gchar *str)
 {
 	gint cid;
 	gchar *buf;
+	gchar *tmp;
 
-	buf = g_strdup(str);
-	strretchomp(buf);
-	if (strlen(buf) > 76) {
-		wchar_t *wbuf;
-
-		wbuf = strdup_mbstowcs(buf);
-
-		if (wcslen(wbuf) > 60) {
-			gchar *tmp;
-
-			g_free(buf);
-			wbuf[60] = (wchar_t)0;
-			tmp = strdup_wcstombs(wbuf);
-			buf = g_strconcat(tmp, "...", NULL);
-			g_free(tmp);
-		}
-
-		g_free(wbuf);
-	}
+	tmp = g_strdup(str);
+	strretchomp(tmp);
+	buf = trim_string(tmp, 76);
+	g_free(tmp);
 
 	cid = gtk_statusbar_get_context_id(statusbar, "Standard Output");
 	gtk_statusbar_pop(statusbar, cid);
diff --git a/src/textview.c b/src/textview.c
index 2b240b6b..da860bfe 100644
--- a/src/textview.c
+++ b/src/textview.c
@@ -1269,14 +1269,14 @@ void textview_scroll_one_line(TextView *textview, gboolean up)
 	if (!up) {
 		upper = vadj->upper - vadj->page_size;
 		if (vadj->value < upper) {
-			vadj->value += vadj->step_increment * 4;
+			vadj->value += vadj->step_increment;
 			vadj->value = MIN(vadj->value, upper);
 			g_signal_emit_by_name(G_OBJECT(vadj),
 					      "value_changed", 0);
 		}
 	} else {
 		if (vadj->value > 0.0) {
-			vadj->value -= vadj->step_increment * 4;
+			vadj->value -= vadj->step_increment;
 			vadj->value = MAX(vadj->value, 0.0);
 			g_signal_emit_by_name(G_OBJECT(vadj),
 					      "value_changed", 0);
@@ -1366,7 +1366,7 @@ static void textview_smooth_scroll_one_line(TextView *textview, gboolean up)
 		upper = vadj->upper - vadj->page_size;
 		if (vadj->value < upper) {
 			old_value = vadj->value;
-			last_value = vadj->value + vadj->step_increment * 4;
+			last_value = vadj->value + vadj->step_increment;
 			last_value = MIN(last_value, upper);
 
 			textview_smooth_scroll_do(textview, old_value,
@@ -1376,7 +1376,7 @@ static void textview_smooth_scroll_one_line(TextView *textview, gboolean up)
 	} else {
 		if (vadj->value > 0.0) {
 			old_value = vadj->value;
-			last_value = vadj->value - vadj->step_increment * 4;
+			last_value = vadj->value - vadj->step_increment;
 			last_value = MAX(last_value, 0.0);
 
 			textview_smooth_scroll_do(textview, old_value,
diff --git a/src/unmime.c b/src/unmime.c
index e7dce098..bb9e2822 100644
--- a/src/unmime.c
+++ b/src/unmime.c
@@ -112,7 +112,7 @@ void unmime_header(gchar *out, const gchar *str)
 			continue;
 		}
 
-		/* convert to locale encoding */
+		/* convert to UTF-8 */
 		conv_str = conv_codeset_strdup(decoded_text, charset, NULL);
 		if (conv_str) {
 			len = strlen(conv_str);
diff --git a/src/utils.c b/src/utils.c
index 493c9597..693f86ea 100644
--- a/src/utils.c
+++ b/src/utils.c
@@ -518,11 +518,9 @@ gint get_mbs_len(const gchar *s)
 		return -1;
 
 	while (*p != '\0') {
-		mb_len = mblen(p, MB_LEN_MAX);
+		mb_len = g_utf8_skip[*(guchar *)p];
 		if (mb_len == 0)
 			break;
-		else if (mb_len < 0)
-			return -1;
 		else
 			len++;
 
@@ -1340,13 +1338,13 @@ gchar *trim_string(const gchar *str, gint len)
 	if (!str) return NULL;
 	if (strlen(str) <= len)
 		return g_strdup(str);
+	if (g_utf8_validate(str, -1, NULL) == FALSE)
+		return g_strdup(str);
 
 	while (*p != '\0') {
-		mb_len = mblen(p, MB_LEN_MAX);
+		mb_len = g_utf8_skip[*(guchar *)p];
 		if (mb_len == 0)
 			break;
-		else if (mb_len < 0)
-			return g_strdup(str);
 		else if (new_len + mb_len > len)
 			break;
 
@@ -1367,13 +1365,13 @@ gchar *trim_string_before(const gchar *str, gint len)
 	if (!str) return NULL;
 	if ((new_len = strlen(str)) <= len)
 		return g_strdup(str);
+	if (g_utf8_validate(str, -1, NULL) == FALSE)
+		return g_strdup(str);
 
 	while (*p != '\0') {
-		mb_len = mblen(p, MB_LEN_MAX);
+		mb_len = g_utf8_skip[*(guchar *)p];
 		if (mb_len == 0)
 			break;
-		else if (mb_len < 0)
-			return g_strdup(str);
 
 		new_len -= mb_len;
 		p += mb_len;