add support for utf-8

[matkoniecz]

There is no support for UTF-8 in libroxml for now.

#63 (comment)

[baemhai]

Until this issue is taken up in earnest, I would like to suggest improvement steps :

1. make the limitation to ASCII and thus 7-bit char input explicit, in docs and doxygen. What is considered _"minimum, easy-to-use" _ is not as intuitive today as it was a few years back.

2. while standards are not satisfied, libroxml could put in safeguards with little patches like this

--- a/src/roxml_parser.c
+++ b/src/roxml_parser.c
@@ -17,7 +17,7 @@
 #define ROXML_PARSER_NCELLS	256
 #define ROXML_PARSER_NCELLS_MAX	512
 
-#define ROXML_PARSER_ITEM(parser, item)	(parser + item)
+#define ROXML_PARSER_ITEM(parser, item)	(parser + (unsigned char)item)
 
 ROXML_INT int roxml_parser_multiple(roxml_parser_item_t *parser, char *chunk, void *data)
 {

[kitech]

v3.0.2 crash attr with non ascii value,

<?xml version="1.0" encoding="UTF-8"?>
<items>
  <item>
    <name title="日本語">日本語</name>
    <description>これはテストです</description>
  </item>
  <item>
    <name text="中文">中文</name>
    <description>这是一个测试</description>
  </item>
</items>