DOM Strings are now capable of containing either UTF-8 or UTF-16 encoded data.

The charset used for strings within a document is specified at document creation time. Whilst it is possible to mix charsets within a document, it's not recommended.

Things that need fixing:

+ dom_string_get_data() doesn't return the charset. Better would be to permit
  the client to request a charset for the data to be returned in.
+ Interned node name strings will break if the document is UTF-16 
  (dom_document_create()). In fact, these could quite happily be globals,
  rather than allocating a set for each document.
+ Other usage of dom string constructors need checking for sanity
+ DOM Strings need to gain more utility APIs (such as getting the character
  length of a string, string concatenation etc).


svn path=/trunk/dom/; revision=3614

1 files changed, 3 insertions, 0 deletions

diff --git a/bindings/xml/xmlparser.c b/bindings/xml/xmlparser.c
index 6f1516b..9541a7c 100644
--- a/bindings/xml/xmlparser.c
+++ b/bindings/xml/xmlparser.c
@@ -182,6 +182,7 @@ xml_parser *xml_parser_create(const char *enc, const char *int_enc,
 
 	/* Create key for user data registration */
 	err = dom_string_create_from_ptr_no_doc((dom_alloc) alloc, pw,
+			DOM_STRING_UTF8, 
 			(const uint8_t *) "__xmlnode", SLEN("__xmlnode"),
 			&parser->udkey);
 	if (err != DOM_NO_ERR) {
@@ -194,6 +195,7 @@ xml_parser *xml_parser_create(const char *enc, const char *int_enc,
 	/* Get DOM implementation */
 	/* Create a string representation of the features we want */
 	err = dom_string_create_from_ptr_no_doc((dom_alloc) alloc, pw,
+			DOM_STRING_UTF8,
 			(const uint8_t *) "XML", SLEN("XML"), &features);
 	if (err != DOM_NO_ERR) {
 		dom_string_unref(parser->udkey);
@@ -327,6 +329,7 @@ void xml_parser_start_document(void *ctx)
 			/* qname */ NULL,
 			/* doctype */ NULL,
 			&doc,
+			DOM_STRING_UTF8,
 			(dom_alloc) parser->alloc,
 			parser->pw);
 	if (err != DOM_NO_ERR) {