Issue #456 - Merge branch 'meta-charset' of tidy-html5-marco.

This pulls the work done by @marcoscaceres WIP #458 into the issue-456
branch, to complete the new add-meta-charset option.

include/tidyenum.h

@@ -648,6 +648,7 @@ typedef enum
     TidyXmlPIs,                  /**< If set to yes PIs must end with ?> */
     TidyXmlSpace,                /**< If set to yes adds xml:space attr as needed */
     TidyXmlTags,                 /**< Treat input as XML */
+    TidyMetaCharset,             /**< Adds/checks/fixes meta charset in the head, based on document type */
     N_TIDY_OPTIONS               /**< Must be last */
 } TidyOptionId;
 

src/attrs.h

@@ -184,6 +184,7 @@ Bool TY_(AttributeIsMismatched)(Node* node, AttVal* attval, TidyDocImpl* doc);
 #define attrIsBOTTOMMARGIN(av)      AttrIsId( av, TidyAttr_BOTTOMMARGIN  )
 #define attrIsCELLPADDING(av)       AttrIsId( av, TidyAttr_CELLPADDING  )
 #define attrIsCELLSPACING(av)       AttrIsId( av, TidyAttr_CELLSPACING  )
+#define attrIsCHARSET(av)           AttrIsId( av, TidyAttr_CHARSET  )
 #define attrIsCHAR(av)              AttrIsId( av, TidyAttr_CHAR  )
 #define attrIsCHAROFF(av)           AttrIsId( av, TidyAttr_CHAROFF  )
 #define attrIsCHARSET(av)           AttrIsId( av, TidyAttr_CHARSET  )
@@ -385,6 +386,7 @@ Bool TY_(AttributeIsMismatched)(Node* node, AttVal* attval, TidyDocImpl* doc);
 #define attrGetHEIGHT( nod )      TY_(AttrGetById)( nod, TidyAttr_HEIGHT  )
 #define attrGetFOR( nod )         TY_(AttrGetById)( nod, TidyAttr_FOR  )
 #define attrGetSELECTED( nod )    TY_(AttrGetById)( nod, TidyAttr_SELECTED  )
+#define attrGetCHARSET( nod )     TY_(AttrGetById)( nod, TidyAttr_CHARSET  )
 #define attrGetCHECKED( nod )     TY_(AttrGetById)( nod, TidyAttr_CHECKED  )
 #define attrGetLANG( nod )        TY_(AttrGetById)( nod, TidyAttr_LANG  )
 #define attrGetTARGET( nod )      TY_(AttrGetById)( nod, TidyAttr_TARGET  )

src/config.c

@@ -336,6 +336,7 @@ static const TidyOptionImpl option_defs[] =
     { TidyXmlPIs,                  MU, "assume-xml-procins",          BL, no,              ParseBool,         boolPicks       },
     { TidyXmlSpace,                MU, "add-xml-space",               BL, no,              ParseBool,         boolPicks       },
     { TidyXmlTags,                 MU, "input-xml",                   BL, no,              ParseBool,         boolPicks       },
+    { TidyMetaCharset,             MS, "add-meta-charset",            BL, no,              ParseBool,         boolPicks       }, /* 20161004 - Issue #456 */
     { N_TIDY_OPTIONS,              XX, NULL,                          XY, 0,               NULL,              NULL            }
 };
 

src/language_en.h

@@ -1508,7 +1508,18 @@ static languageDefinition language_en = { whichPluralForm_en, {
         "This option specifies if Tidy should use the XML parser rather than the "
         "error correcting HTML parser. "
     },
-
+    {/* Important notes for translators:
+        - Use only <code></code>, <var></var>, <em></em>, <strong></strong>, and
+          <br/>.
+        - Entities, tags, attributes, etc., should be enclosed in <code></code>.
+        - Option values should be enclosed in <var></var>.
+        - It's very important that <br/> be self-closing!
+        - The strings "Tidy" and "HTML Tidy" are the program name and must not
+          be translated. */
+      TidyMetaCharset,             0,
+        "This option adds a meta element and sets the charset attribute to the encoding of the document."
+        "Set this option to 'yes' if you want this."
+    },
 
     /********************************************
      ** TidyConfigCategory enumeration
@@ -1772,7 +1783,6 @@ static languageDefinition language_en = { whichPluralForm_en, {
         "https://github.com/htacg/tidy-html5/blob/master/README/LOCALIZE.md"
     },
     
-    
     /********************************************
      ** Report Output
      ** @remark enum source TidyStrings
@@ -1886,7 +1896,6 @@ static languageDefinition language_en = { whichPluralForm_en, {
     { ELEMENT_NOT_EMPTY,            0,   "%s element not empty or not closed"                                      }, /* ReportError, ReportAttrError */
     { UNEXPECTED_END_OF_FILE,       0,   "unexpected end of file %s"                                               }, /* ReportError, ReportAttrError */
     { UNEXPECTED_ENDTAG,            0,   "unexpected </%s>"                                                        }, /* ReportError, ReportFatal */
-
     
 #if SUPPORT_ACCESSIBILITY_CHECKS
     

src/lexer.c

@@ -1827,6 +1827,121 @@ Node *TY_(FindBody)( TidyDocImpl* doc )
     return node;
 }
 
+/* Check meta charset
+   1. if there is no meta charset, it adds one.
+   2. if there is a meta charset, it moves it to the top if HEAD.
+   3. if it doesn't match the output encoding, warn about that.
+   4. if there are duplicates, discard them.
+ */
+Bool TY_(TidyMetaCharset)(TidyDocImpl* doc)
+{
+    AttVal *charsetAttr;
+    AttVal *contentAttr;
+    AttVal *httpEquivAttr;
+    Bool charsetFound = no;
+    ctmbstr enc = TY_(GetEncodingNameFromTidyId)(cfg(doc, TidyOutCharEncoding));
+    Node *currentNode;
+    Node *head = TY_(FindHEAD)( doc );
+    Node *metaTag;
+    Node *prevNode;
+    TidyBuffer buf;
+    TidyBuffer charsetString;
+    tmbstr httpEquivAttrValue;
+    tmbstr lcontent;
+    tmbstr newValue;
+    /* We can't do anything we don't have a head or encoding is NULL */
+    if( !head || !enc )
+        return no;
+    tidyBufInit(&charsetString);
+    for (currentNode = head->content; currentNode; currentNode = currentNode->next)
+    {
+        if (!nodeIsMETA(currentNode))
+            continue;
+        charsetAttr = attrGetCHARSET(currentNode);
+        httpEquivAttr = attrGetHTTP_EQUIV(currentNode);
+        if(!charsetAttr && !httpEquivAttr)
+            continue;
+        /*
+        Meta charset comes in quite a few flavors:
+        1. <meta charset=value> - expected for (X)HTML5.
+        */
+        if (charsetAttr && !httpEquivAttr)
+        {
+            // we already found one, so remove the rest.
+            if(charsetFound)
+            {
+                prevNode = currentNode->prev;
+                TY_(ReportError)(doc, head, currentNode, DISCARDING_UNEXPECTED);
+                TY_(DiscardElement)( doc, currentNode );
+                currentNode = prevNode;
+                continue;
+            }
+            charsetFound = yes;
+            // Fix mismatched attribute value
+            if(TY_(tmbstrcmp)(TY_(tmbstrtolower)(charsetAttr->value), enc) != 0)
+            {
+                newValue = (tmbstr) TidyDocAlloc( doc, TY_(tmbstrlen)(enc) );
+                TY_(tmbstrcpy)( newValue, enc );
+                TY_(ReportAttrError)( doc, currentNode, charsetAttr, BAD_ATTRIBUTE_VALUE_REPLACED );
+                charsetAttr->value = newValue;
+            }
+            // Make sure it's the first element.
+            if ( currentNode != head->content->next ){
+                TY_(RemoveNode)( currentNode );
+                TY_(InsertNodeAtStart)( head, currentNode );
+            }
+            continue;
+        }
+        /*
+        2. <meta http-equiv="content-type" content="text/html; charset=UTF-8">
+        expected for HTML4. This is normally ok - but can clash.
+        */
+        if(httpEquivAttr && !charsetAttr)
+        {
+            tidyBufClear(&charsetString);
+            tidyBufAppend(&charsetString, "charset=", 8);
+            tidyBufAppend(&charsetString, (char*)enc, TY_(tmbstrlen)( enc ));
+            contentAttr = TY_(AttrGetById)(currentNode, TidyAttr_CONTENT);
+            httpEquivAttrValue = TY_(tmbstrtolower)(httpEquivAttr->value);
+
+            if(!contentAttr || TY_(tmbstrcmp)(httpEquivAttr->value, (tmbstr) "content-type") != 0)
+                continue;
+            lcontent = TY_(tmbstrtolower)(contentAttr->value);
+            if(TY_(tmbsubstr)(lcontent, (ctmbstr) &charsetString)){
+                printf("WARN ABOUT CLASH: %s \n", contentAttr->value);
+            }
+        }
+        /*
+        3. <meta charset="utf-8" http-equiv="Content-Type" content="...">
+        This is generally bad.
+        */
+        if(httpEquivAttr && charsetAttr)
+        {
+           printf("WARN ABOUT HTTP EQUIV AND CHARSET ATTR! \n");
+        }
+    }
+    if(charsetFound){
+        return yes;
+    }
+    metaTag = TY_(InferredTag)(doc, TidyTag_META);
+    switch(TY_(HTMLVersion)(doc))
+    {
+        case HT50:
+        case XH50:
+            TY_(AddAttribute)( doc, metaTag, "charset", enc);
+            break;
+        default:
+            tidyBufInit(&buf);
+            tidyBufAppend(&buf, "text/html; charset=", 19);
+            tidyBufAppend(&buf, (char*)enc, TY_(tmbstrlen)(enc));
+            TY_(AddAttribute)( doc, metaTag, "content", (char*)buf.bp);
+            tidyBufFree(&buf);
+    }
+    TY_(InsertNodeAtStart)( head, metaTag );
+    tidyBufFree(&charsetString);
+    return yes;
+}
+
 /* add meta element for Tidy */
 Bool TY_(AddGenerator)( TidyDocImpl* doc )
 {

src/lexer.h

@@ -491,6 +491,9 @@ Node* TY_(FindXmlDecl)(TidyDocImpl* doc);
 /* Returns containing block element, if any */
 Node* TY_(FindContainer)( Node* node );
 
+/* Adds meta element and sets the charset */
+Bool TY_(TidyMetaCharset)( TidyDocImpl* doc );
+
 /* add meta element for Tidy */
 Bool TY_(AddGenerator)( TidyDocImpl* doc );
 

src/tidylib.c

@@ -1992,6 +1992,7 @@ int         tidyDocCleanAndRepair( TidyDocImpl* doc )
     Bool xhtmlOut = cfgBool( doc, TidyXhtmlOut );
     Bool xmlDecl  = cfgBool( doc, TidyXmlDecl );
     Bool tidyMark = cfgBool( doc, TidyMark );
+    Bool tidyMetaCharset = cfgBool( doc, TidyMetaCharset);
     Bool tidyXmlTags = cfgBool( doc, TidyXmlTags );
     Bool wantNameAttr = cfgBool( doc, TidyAnchorAsName );
     Bool mergeEmphasis = cfgBool( doc, TidyMergeEmphasis );
@@ -2095,6 +2096,9 @@ int         tidyDocCleanAndRepair( TidyDocImpl* doc )
 
         if (tidyMark )
             TY_(AddGenerator)(doc);
+
+        if (tidyMetaCharset)
+            TY_(TidyMetaCharset)(doc);
     }
 
     /* ensure presence of initial <?xml version="1.0"?> */