[project @ 2005-04-16 05:09:32 by jmb]

Split out UTF-8 handling functions.
Submit URL-encoded forms in sensible encoding:
  * First entry in accept-charset list, if present
  * Document encoding, otherwise
  We may want to explicitly look for UTF-8, to save converting.
Convert cnv_str_local_enc/cnv_local_enc_str to use iconv (they're now veneers for utf8_[to/from]_enc).
Provide mechanism for looking up local system charset (derived from system alphabet, under RISC OS)

svn path=/import/netsurf/; revision=1647

1 files changed, 28 insertions, 2 deletions

diff --git a/render/box_construct.c b/render/box_construct.c
index 9723b4de1..88a432e67 100644
--- a/render/box_construct.c
+++ b/render/box_construct.c
@@ -1575,7 +1575,7 @@ bool box_iframe(BOX_SPECIAL_PARAMS)
 
 bool box_form(BOX_SPECIAL_PARAMS)
 {
-	char *action, *method, *enctype;
+	char *action, *method, *enctype, *charset;
 	form_method fmethod;
 	struct form *form;
 
@@ -1598,9 +1598,35 @@ bool box_form(BOX_SPECIAL_PARAMS)
 		xmlFree(method);
 	}
 
-	form = form_new(action, fmethod);
+	/* acceptable encoding(s) for form data */
+	if ((charset = (char *) xmlGetProp(n, (const xmlChar *) "accept-charset"))) {
+		char *comma = strchr(charset, ',');
+		if (!comma)
+			/* only one => use it */
+			comma = strdup(charset);
+		else
+			/* multiple => use first */
+			comma = strndup(charset, comma - charset);
+
+		xmlFree(charset);
+		charset = comma;
+	}
+	else if (content->data.html.encoding)
+		/* none specified => try document encoding */
+		charset = strdup(content->data.html.encoding);
+	else
+		/* none specified and no document encoding => 8859-1 */
+		charset = strdup("ISO-8859-1");
+
+	if (!charset) {
+		xmlFree(action);
+		return false;
+	}
+
+	form = form_new(action, fmethod, charset);
 	if (!form) {
 		xmlFree(action);
+		free(charset);
 		return false;
 	}
 	form->prev = content->data.html.forms;