Issue #166 - repeated main element.

With this fix introduced two new services, FindNodeById and
FindNodeWithId. The former does a total tree search for a TidyTagId.

Maybe there is a way to optimise this search...

Also change the uint badForm from an on/off to a bit field, so could be
extended to other document format errors.

src/localize.c

@@ -1603,7 +1603,9 @@ void TY_(ReportError)(TidyDocImpl* doc, Node *element, Node *node, uint code)
         break;
 
     case DISCARDING_UNEXPECTED:
-        /* Force error if in a bad form */
+        /* Force error if in a bad form, or 
+           Issue #166 - repeated <main> element
+        */
         messageNode(doc, doc->badForm ? TidyError : TidyWarning, node, fmt, nodedesc);
         break;
 
@@ -1739,7 +1741,7 @@ void TY_(ErrorSummary)( TidyDocImpl* doc )
       }
     }
 
-    if (doc->badForm)
+    if (doc->badForm & flg_BadForm) /* Issue #166 - changed to BIT flag to support other errors */
     {
         tidy_out(doc, "You may need to move one or both of the <form> and </form>\n");
         tidy_out(doc, "tags. HTML elements should be properly nested and form elements\n");
@@ -1749,6 +1751,13 @@ void TY_(ErrorSummary)( TidyDocImpl* doc )
         tidy_out(doc, "table! Note that one form can't be nested inside another!\n\n");
     }
 
+    if (doc->badForm & flg_BadMain) /* Issue #166 - repeated <main> element */
+    {
+        tidy_out(doc, "Only one <main> element is allowed in a document.\n");
+        tidy_out(doc, "Subsequent <main> elements have been discarded, which may\n");
+        tidy_out(doc, "render the document invalid.\n");
+    }
+    
     if (doc->badAccess)
     {
         /* Tidy "classic" accessibility tests */

src/parser.c

@@ -373,10 +373,12 @@ Node* TY_(DropEmptyElements)(TidyDocImpl* doc, Node* node)
 /* 
   errors in positioning of form start or end tags
   generally require human intervention to fix
+  Issue #166 - repeated <main> element also uses this flag
+  to indicate duplicates, discarded
 */
 static void BadForm( TidyDocImpl* doc )
 {
-    doc->badForm = yes;
+    doc->badForm |= flg_BadForm;
     /* doc->errors++; */
 }
 
@@ -3694,6 +3696,39 @@ void TY_(ParseHead)(TidyDocImpl* doc, Node *head, GetTokenMode ARG_UNUSED(mode))
 #endif
 }
 
+/*\ 
+ *  Issue #166 - repeated <main> element
+ *  But this service is generalised to check for other duplicate elements
+\*/
+Bool TY_(FindNodeWithId)( Node *node, TidyTagId tid )
+{
+    Node *content;
+    while (node)
+    {
+        if (TagIsId(node,tid))
+            return yes;
+        for (content = node->content; content; content = content->content)
+        {
+            if (TY_(FindNodeWithId)(content,tid))
+                return yes;
+        }
+        node = node->next;
+    }
+    return no;
+}
+
+
+/*\ 
+ *  Issue #166 - repeated <main> element
+ *  Do a global search for an element
+\*/
+Bool TY_(FindNodeById)( TidyDocImpl* doc, TidyTagId tid )
+{
+    Node *node = (doc ? doc->root.content : NULL);
+    return TY_(FindNodeWithId)(node,tid);
+}
+
+
 void TY_(ParseBody)(TidyDocImpl* doc, Node *body, GetTokenMode mode)
 {
     Lexer* lexer = doc->lexer;
@@ -3934,6 +3969,18 @@ void TY_(ParseBody)(TidyDocImpl* doc, Node *body, GetTokenMode mode)
 
         if (TY_(nodeIsElement)(node))
         {
+            if (nodeIsMAIN(node)) {
+                /*\ Issue #166 - repeated <main> element
+                 *  How to efficiently search for a previous main element?
+                \*/
+                if ( TY_(FindNodeById)(doc, TidyTag_MAIN) )
+                {
+                    doc->badForm |= flg_BadMain; /* this is an ERROR in format */
+                    TY_(ReportError)(doc, body, node, DISCARDING_UNEXPECTED);
+                    TY_(FreeNode)( doc, node);
+                    continue;
+                }
+            }
             /* Issue #20 - merging from Ger Hobbelt fork put back CM_MIXED, which had been
                removed to fix this issue - reverting to fix 880221e
              */

src/tidy-int.h

@@ -23,6 +23,15 @@
 #define MIN(a,b) (((a) < (b))?(a):(b))
 #endif
 
+/*\
+ *  Issue #166 - repeated <main> element
+ *  Change the previous on/off uint flag badForm
+ *  to a BIT flag to support other than <form>
+ *  errors. This could be extended more...
+\*/
+#define flg_BadForm     0x00000001
+#define flg_BadMain     0x00000002
+
 struct _TidyDocImpl
 {
     /* The Document Tree (and backing store buffer) */
@@ -63,7 +72,7 @@ struct _TidyDocImpl
     uint                badAccess;   /* for accessibility errors */
     uint                badLayout;   /* for bad style errors */
     uint                badChars;    /* for bad char encodings */
-    uint                badForm;     /* for badly placed form tags */
+    uint                badForm;     /* bit field, for badly placed form tags, or other format errors */
 
     /* Memory allocator */
     TidyAllocator*      allocator;