fixed #2061

include/gpac/utf.h

@@ -84,14 +84,15 @@ u32 gf_utf8_wcslen(const unsigned short *s);
 /*!
 \brief returns a UTF8 string from a string started with BOM
 
-Returns the length in character of a wide-char string
+Returns UTF8 from data
 \param data the string or wide-char string
 \param size of the data buffer
   size of the data buffer
-\param out_ptr set to an allocated buffer if needed for conversion, shall be destroyed by caller
-\return the UTF8 string corresponding
+\param out_ptr set to an allocated buffer if needed for conversion, shall be destroyed by caller. Must not be NULL
+\param result set to resulting UTF8 string. Must not be NULL
+\return error if any: GF_IO_ERR if UTF decode error or GF_BAD_PARAM
  */
-char *gf_utf_get_utf8_string_from_bom(u8 *data, u32 size, char **out_ptr);
+GF_Err gf_utf_get_utf8_string_from_bom(const u8 *data, u32 size, char **out_ptr, char **result);
 
 /*!
 \brief Checks validity of a UTF8 string

src/filters/load_bt_xmt.c

@@ -787,16 +787,17 @@ static const char *ctxload_probe_data(const u8 *probe_data, u32 size, GF_FilterP
 {
 	const char *mime_type = NULL;
 	char *dst = NULL;
-	u8 *res;
+	GF_Err e;
+	char *res=NULL;
 
 	/* check gzip magic header */
 	if ((size>2) && (probe_data[0] == 0x1f) && (probe_data[1] == 0x8b)) {
 		*score = GF_FPROBE_EXT_MATCH;
 		return "btz|bt.gz|xmt.gz|xmtz|wrl.gz|x3dv.gz|x3dvz|x3d.gz|x3dz";
 	}
 
-	res = gf_utf_get_utf8_string_from_bom((char *)probe_data, size, &dst);
-	if (res) probe_data = res;
+	e = gf_utf_get_utf8_string_from_bom(probe_data, size, &dst, &res);
+	if (e) return NULL;
 
 	//strip all spaces and \r\n
 	while (probe_data[0] && strchr("\n\r\t ", (char) probe_data[0]))

src/filters/load_text.c

@@ -3530,11 +3530,10 @@ void txtin_finalize(GF_Filter *filter)
 static const char *txtin_probe_data(const u8 *data, u32 data_size, GF_FilterProbeScore *score)
 {
 	char *dst = NULL;
-	u8 *res;
-
-	res = gf_utf_get_utf8_string_from_bom((char *)data, data_size, &dst);
-	if (res) data = res;
-
+	char *res=NULL;
+	GF_Err e = gf_utf_get_utf8_string_from_bom((char *)data, data_size, &dst, &res);
+	if (e) return NULL;
+
 #define PROBE_OK(_score, _mime) \
 		*score = _score;\
 		if (dst) gf_free(dst);\

src/utils/utf.c

@@ -648,36 +648,47 @@ u32 gf_utf8_mbstowcs(unsigned short* dest, size_t len, const char** srcp)
 
 
 GF_EXPORT
-char *gf_utf_get_utf8_string_from_bom(u8 *data, u32 size, char **out_ptr)
+GF_Err gf_utf_get_utf8_string_from_bom(const u8 *data, u32 size, char **out_ptr, char **result)
 {
 	u32 unicode_type = 0;
+	if (!out_ptr || !result || !data) return GF_BAD_PARAM;
 	*out_ptr = NULL;
+	*result = (char *) data;
 
 	if (size>=5) {
 		/*0: no unicode, 1: UTF-16BE, 2: UTF-16LE*/
 		if ((data[0]==0xFF) && (data[1]==0xFE)) {
 			if (!data[2] && !data[3]) {
-				return NULL;
+				return GF_OK;
 			} else {
 				unicode_type = 2;
 			}
 		} else if ((data[0]==0xFE) && (data[1]==0xFF)) {
 			if (!data[2] && !data[3]) {
-				return NULL;
+				return GF_OK;
 			} else {
 				unicode_type = 1;
 			}
 		} else if ((data[0]==0xEF) && (data[1]==0xBB) && (data[2]==0xBF)) {
-			return data+4;
+			*result = (char *) (data+4);
+			return GF_OK;
 		}
 	}
 
-	if (!unicode_type) return data;
+	if (!unicode_type) {
+		*result = (char *) data;
+		return GF_OK;
+	}
 
 	if (size%2) size--;
 	u16 *str_wc = gf_malloc(size+2);
+	if (!str_wc) return GF_OUT_OF_MEM;
 	u16 *srcwc;
 	char *dst = gf_malloc(size+2);
+	if (!dst) {
+		gf_free(str_wc);
+		return GF_OUT_OF_MEM;
+	}
 	*out_ptr = dst;
 	u32 i;
 	for (i=0; i<size; i+=2) {
@@ -704,10 +715,15 @@ char *gf_utf_get_utf8_string_from_bom(u8 *data, u32 size, char **out_ptr)
 	}
 	str_wc[i/2] = 0;
 	srcwc = str_wc;
-	gf_utf8_wcstombs(dst, size, (const unsigned short **) &srcwc);
+	u32 res = gf_utf8_wcstombs(dst, size, (const unsigned short **) &srcwc);
 	gf_free(str_wc);
-
-	return dst;
+	if (res==GF_UTF8_FAIL) {
+		gf_free(dst);
+		*out_ptr = NULL;
+		return GF_IO_ERR;
+	}
+	*result = dst;
+	return GF_OK;
 }