Added some configure logic to auto-detect the system shift-jis charset alias

May fix issue #81

2 files changed, 76 insertions, 22 deletions

diff --git a/gmime/gmime-charset.c b/gmime/gmime-charset.c
index 567de1bf..938de4fd 100644
--- a/gmime/gmime-charset.c
+++ b/gmime/gmime-charset.c
@@ -60,6 +60,7 @@
 #define ICONV_ISO_INT_FORMAT "iso-%u-%u"
 #define ICONV_ISO_STR_FORMAT "iso-%u-%s"
 #endif /* __aix__, __irix__, __sun__ */
+#define ICONV_SHIFT_JIS "shift-jis"
 #define ICONV_10646 "iso-10646"
 #endif /* USE_ICONV_DETECT */
 
@@ -133,12 +134,14 @@ static struct {
 	/* Japanese charsets */
 	{ "eucjp-0",         "eucJP"  	  },  /* should this map to "EUC-JP" instead? */
 	{ "ujis-0",          "ujis"  	  },  /* we might want to map this to EUC-JP */
-	{ "jisx0208.1983-0", "SJIS"       },
-	{ "jisx0212.1990-0", "SJIS"       },
-	{ "pck",	     "SJIS"       },
 	{ NULL,              NULL         }
 };
 
+static const char *shiftjis_aliases[] = {
+	"shift-jis", "shift_jis", "sjis", "shift_jis-2004", "shift_jisx0213",
+	"jisx0208.1983-0", "jisx0212.1990-0", "pck", NULL
+};
+
 /* map CJKR charsets to their language code */
 /* NOTE: only support charset names that will be returned by
  * g_mime_charset_iconv_name() so that we don't have to keep track of
@@ -147,21 +150,20 @@ static struct {
 	const char *charset;
 	const char *lang;
 } cjkr_lang_map[] = {
-	{ "Big5",        "zh" },
-	{ "BIG5HKSCS",   "zh" },
-	{ "gb2312",      "zh" },
-	{ "gb18030",     "zh" },
-	{ "gbk",         "zh" },
-	{ "euc-tw",      "zh" },
-	{ "iso-2022-jp", "ja" },
-	{ "Shift-JIS",   "ja" },
-	{ "sjis",        "ja" },
-	{ "ujis",        "ja" },
-	{ "eucJP",       "ja" },
-	{ "euc-jp",      "ja" },
-	{ "euc-kr",      "ko" },
-	{ "koi8-r",      "ru" },
-	{ "koi8-u",      "uk" }
+	{ "Big5",          "zh" },
+	{ "BIG5HKSCS",     "zh" },
+	{ "gb2312",        "zh" },
+	{ "gb18030",       "zh" },
+	{ "gbk",           "zh" },
+	{ "euc-tw",        "zh" },
+	{ "iso-2022-jp",   "ja" },
+	{ ICONV_SHIFT_JIS, "ja" },
+	{ "ujis",          "ja" },
+	{ "eucJP",         "ja" },
+	{ "euc-jp",        "ja" },
+	{ "euc-kr",        "ko" },
+	{ "koi8-r",        "ru" },
+	{ "koi8-u",        "uk" }
 };
 
 static GHashTable *iconv_charsets = NULL;
@@ -409,6 +411,18 @@ g_mime_charset_language (const char *charset)
 	return NULL;
 }
 
+static gboolean
+is_shift_jis (const char *name)
+{
+	int i;
+	
+	for (i = 0; shiftjis_aliases[i] != NULL; i++) {
+		if (!strcmp (name, shiftjis_aliases[i]))
+			return TRUE;
+	}
+	
+	return FALSE;
+}
 
 static const char *
 strdown (char *str)
@@ -483,8 +497,7 @@ g_mime_charset_iconv_name (const char *charset)
 								      iso, codepage);
 			} else {
 				/* codepage is a string - probably iso-2022-jp or something */
-				iconv_name = g_strdup_printf (ICONV_ISO_STR_FORMAT,
-							      iso, p);
+				iconv_name = g_strdup_printf (ICONV_ISO_STR_FORMAT, iso, p);
 			}
 		} else {
 			/* p == buf, which probably means we've
@@ -503,6 +516,8 @@ g_mime_charset_iconv_name (const char *charset)
 			buf += 2;
 		
 		iconv_name = g_strdup_printf ("CP%s", buf);
+	} else if (is_shift_jis (name)) {
+		iconv_name = g_strdup (ICONV_SHIFT_JIS);
 	} else {
 		/* assume charset name is ok as is? */
 		iconv_name = g_strdup (charset);
diff --git a/iconv-detect.c b/iconv-detect.c
index c033fd07..16333b8b 100644
--- a/iconv-detect.c
+++ b/iconv-detect.c
@@ -1,6 +1,6 @@
 /* -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8 -*- */
 /*  GMime
- *  Copyright (C) 1999-2017 Jeffrey Stedfast
+ *  Copyright (C) 1999-2020 Jeffrey Stedfast
  *
  *  This library is free software; you can redistribute it and/or
  *  modify it under the terms of the GNU Lesser General Public License
@@ -47,6 +47,12 @@ enum {
 	ISO_DASH_UINT_DASH_STR_LOWER   = (1 << 0),
 	ISO_DASH_UINT_DASH_STR         = (1 << 1),
 	ISO_UINT_DASH_STR              = (1 << 2),
+
+	/* shift-jis */
+	SHIFT_DASH_JIS                 = (1 << 0),
+	SHIFT_UNDERSCORE_JIS           = (1 << 1),
+	SJIS                           = (1 << 2),
+	SHIFT_JISX0213                 = (1 << 3),
 };
 
 
@@ -76,6 +82,15 @@ static CharInfo iso2022_tests[] = {
 
 static int num_iso2022_tests = sizeof (iso2022_tests) / sizeof (CharInfo);
 
+static CharInfo shiftjis_tests[] = {
+	{ "shift-jis",      "shift-jis",      SHIFT_DASH_JIS },
+	{ "shift_jis",      "shift_jis",      SHIFT_UNDERSCORE_JIS },
+	{ "sjis",           "sjis",           SJIS },
+	{ "shift_jisx0213", "shift_jisx0213", SHIFT_JISX0213 },
+};
+
+static int num_shiftjis_tests = sizeof (shiftjis_tests) / sizeof (CharInfo);
+
 static CharInfo iso10646_tests[] = {
 	{ "iso-10646-1", "iso-%u-%u",  ISO_DASH_UINT_DASH_UINT_LOWER },
 	{ "ISO-10646-1", "ISO-%u-%u",  ISO_DASH_UINT_DASH_UINT },
@@ -91,7 +106,7 @@ static int num_iso10646_tests = sizeof (iso10646_tests) / sizeof (CharInfo);
 
 int main (int argc, char **argv)
 {
-	unsigned int iso8859, iso2022, iso10646;
+	unsigned int iso8859, iso2022, iso10646, shiftjis;
 	CharInfo *info;
 	iconv_t cd;
 	FILE *fp;
@@ -150,6 +165,30 @@ int main (int argc, char **argv)
 		fprintf (fp, "#define ICONV_ISO_STR_FORMAT \"%s\"\n", info[i].format);
 	}
 	
+	shiftjis = ISO_UNSUPPORTED;
+	info = shiftjis_tests;
+	/*printf ("#define ISO_2022_FORMAT(iso,codepage)\t");*/
+	for (i = 0; i < num_shiftjis_tests; i++) {
+		cd = iconv_open (info[i].charset, "UTF-8");
+		if (cd != (iconv_t) -1) {
+			iconv_close (cd);
+			/*printf ("(\"%s\", (iso), (codepage))\n", info[i].format);*/
+			fprintf (stderr, "System prefers %s\n", info[i].charset);
+			shiftjis = info[i].id;
+			break;
+		}
+	}
+	
+	if (shiftjis == ISO_UNSUPPORTED) {
+		fprintf (stderr, "System doesn't support any SHIFT-JIS aliases\n");
+		fprintf (fp, "#define ICONV_SHIFT_JIS \"%s\"\n", info[0].format);
+#ifdef CONFIGURE_IN
+		return EXIT_FAILURE;
+#endif
+	} else {
+		fprintf (fp, "#define ICONV_SHIFT_JIS \"%s\"\n", info[i].format);
+	}
+	
 	iso10646 = ISO_UNSUPPORTED;
 	info = iso10646_tests;
 	/*printf ("#define ISO_10646_FORMAT(iso,codepage)\t");*/