/* parser.c -- HTML Parser (c) 1998-2007 (W3C) MIT, ERCIM, Keio University See tidy.h for the copyright notice. */ #include "tidy-int.h" #include "lexer.h" #include "parser.h" #include "message.h" #include "clean.h" #include "tags.h" #include "tmbstr.h" #ifdef _MSC_VER #include "sprtf.h" #endif #ifndef SPRTF #define SPRTF printf #endif #ifdef AUTO_INPUT_ENCODING #include "charsets.h" #endif /* Issue #72 - Need to know to avoid error-reporting - no warning only if --show-body-only yes Issue #132 - likewise avoid warning if showing body only */ #define showingBodyOnly(doc) (cfgAutoBool(doc,TidyBodyOnly) == TidyYesState) ? yes : no Bool TY_(CheckNodeIntegrity)(Node *node) { #ifndef NO_NODE_INTEGRITY_CHECK Node *child; if (node->prev) { if (node->prev->next != node) return no; } if (node->next) { if (node->next == node || node->next->prev != node) return no; } if (node->parent) { if (node->prev == NULL && node->parent->content != node) return no; if (node->next == NULL && node->parent->last != node) return no; } for (child = node->content; child; child = child->next) if ( child->parent != node || !TY_(CheckNodeIntegrity)(child) ) return no; #endif return yes; } /* used to determine how attributes without values should be printed this was introduced to deal with user defined tags e.g. Cold Fusion */ Bool TY_(IsNewNode)(Node *node) { if (node && node->tag) { return (node->tag->model & CM_NEW); } return yes; } void TY_(CoerceNode)(TidyDocImpl* doc, Node *node, TidyTagId tid, Bool obsolete, Bool unexpected) { const Dict* tag = TY_(LookupTagDef)(tid); Node* tmp = TY_(InferredTag)(doc, tag->id); if (obsolete) TY_(ReportWarning)(doc, node, tmp, OBSOLETE_ELEMENT); else if (unexpected) TY_(ReportError)(doc, node, tmp, REPLACING_UNEX_ELEMENT); else TY_(ReportNotice)(doc, node, tmp, REPLACING_ELEMENT); TidyDocFree(doc, tmp->element); TidyDocFree(doc, tmp); node->was = node->tag; node->tag = tag; node->type = StartTag; node->implicit = yes; TidyDocFree(doc, node->element); node->element = TY_(tmbstrdup)(doc->allocator, tag->name); } /* extract a node and its children from a markup tree */ Node *TY_(RemoveNode)(Node *node) { if (node->prev) node->prev->next = node->next; if (node->next) node->next->prev = node->prev; if (node->parent) { if (node->parent->content == node) node->parent->content = node->next; if (node->parent->last == node) node->parent->last = node->prev; } node->parent = node->prev = node->next = NULL; return node; } /* remove node from markup tree and discard it */ Node *TY_(DiscardElement)( TidyDocImpl* doc, Node *element ) { Node *next = NULL; if (element) { next = element->next; TY_(RemoveNode)(element); TY_(FreeNode)( doc, element); } return next; } /* insert "node" into markup tree as the firt element of content of "element" */ void TY_(InsertNodeAtStart)(Node *element, Node *node) { node->parent = element; if (element->content == NULL) element->last = node; else element->content->prev = node; node->next = element->content; node->prev = NULL; element->content = node; } /* insert "node" into markup tree as the last element of content of "element" */ void TY_(InsertNodeAtEnd)(Node *element, Node *node) { node->parent = element; node->prev = element->last; if (element->last != NULL) element->last->next = node; else element->content = node; element->last = node; } /* insert "node" into markup tree in place of "element" which is moved to become the child of the node */ static void InsertNodeAsParent(Node *element, Node *node) { node->content = element; node->last = element; node->parent = element->parent; element->parent = node; if (node->parent->content == element) node->parent->content = node; if (node->parent->last == element) node->parent->last = node; node->prev = element->prev; element->prev = NULL; if (node->prev) node->prev->next = node; node->next = element->next; element->next = NULL; if (node->next) node->next->prev = node; } /* insert "node" into markup tree before "element" */ void TY_(InsertNodeBeforeElement)(Node *element, Node *node) { Node *parent; parent = element->parent; node->parent = parent; node->next = element; node->prev = element->prev; element->prev = node; if (node->prev) node->prev->next = node; if (parent->content == element) parent->content = node; } /* insert "node" into markup tree after "element" */ void TY_(InsertNodeAfterElement)(Node *element, Node *node) { Node *parent; parent = element->parent; node->parent = parent; /* AQ - 13 Jan 2000 fix for parent == NULL */ if (parent != NULL && parent->last == element) parent->last = node; else { node->next = element->next; /* AQ - 13 Jan 2000 fix for node->next == NULL */ if (node->next != NULL) node->next->prev = node; } element->next = node; node->prev = element; } static Bool CanPrune( TidyDocImpl* doc, Node *element ) { if ( !cfgBool(doc, TidyDropEmptyElems) ) return no; if ( TY_(nodeIsText)(element) ) return yes; if ( element->content ) return no; if ( element->tag == NULL ) return no; if ( element->tag->model & CM_BLOCK && element->attributes != NULL ) return no; if ( nodeIsA(element) && element->attributes != NULL ) return no; if ( nodeIsP(element) && !cfgBool(doc, TidyDropEmptyParas) ) return no; if ( element->tag->model & CM_ROW ) return no; if ( element->tag->model & CM_EMPTY ) return no; if ( nodeIsAPPLET(element) ) return no; if ( nodeIsOBJECT(element) ) return no; if ( nodeIsSCRIPT(element) && attrGetSRC(element) ) return no; if ( nodeIsTITLE(element) ) return no; /* #433359 - fix by Randy Waki 12 Mar 01 */ if ( nodeIsIFRAME(element) ) return no; /* fix for bug 770297 */ if (nodeIsTEXTAREA(element)) return no; /* fix for ISSUE #7 https://github.com/w3c/tidy-html5/issues/7 */ if (nodeIsCANVAS(element)) return no; if (nodeIsPROGRESS(element)) return no; if ( attrGetID(element) || attrGetNAME(element) ) return no; /* fix for bug 695408; a better fix would look for unknown and */ /* known proprietary attributes that make the element significant */ if (attrGetDATAFLD(element)) return no; /* fix for bug 723772, don't trim new-...-tags */ if (element->tag->id == TidyTag_UNKNOWN) return no; if (nodeIsBODY(element)) return no; if (nodeIsCOLGROUP(element)) return no; /* HTML5 - do NOT drop empty option if it has attributes */ if ( nodeIsOPTION(element) && element->attributes != NULL ) return no; /* fix for #103 - don't drop empty dd tags lest document not validate */ if (nodeIsDD(element)) return no; return yes; } /* return next element */ Node *TY_(TrimEmptyElement)( TidyDocImpl* doc, Node *element ) { if ( CanPrune(doc, element) ) { if (element->type != TextNode) TY_(ReportNotice)(doc, element, NULL, TRIM_EMPTY_ELEMENT); return TY_(DiscardElement)(doc, element); } return element->next; } Node* TY_(DropEmptyElements)(TidyDocImpl* doc, Node* node) { Node* next; while (node) { next = node->next; if (node->content) TY_(DropEmptyElements)(doc, node->content); if (!TY_(nodeIsElement)(node) && !(TY_(nodeIsText)(node) && !(node->start < node->end))) { node = next; continue; } next = TY_(TrimEmptyElement)(doc, node); node = next; } return node; } /* errors in positioning of form start or end tags generally require human intervention to fix Issue #166 - repeated
element also uses this flag to indicate duplicates, discarded */ static void BadForm( TidyDocImpl* doc ) { doc->badForm |= flg_BadForm; /* doc->errors++; */ } /* This maps hello world to hello world If last child of element is a text node then trim trailing white space character moving it to after element's end tag. */ static void TrimTrailingSpace( TidyDocImpl* doc, Node *element, Node *last ) { Lexer* lexer = doc->lexer; byte c; if (TY_(nodeIsText)(last)) { if (last->end > last->start) { c = (byte) lexer->lexbuf[ last->end - 1 ]; if ( c == ' ' #ifdef COMMENT_NBSP_FIX || c == 160 #endif ) { #ifdef COMMENT_NBSP_FIX /* take care with   */ if ( c == 160 && ( element->tag == doc->tags.tag_td || element->tag == doc->tags.tag_th ) ) { if (last->end > last->start + 1) last->end -= 1; } else #endif { last->end -= 1; if ( (element->tag->model & CM_INLINE) && !(element->tag->model & CM_FIELD) ) lexer->insertspace = yes; } } } } } #if 0 static Node *EscapeTag(Lexer *lexer, Node *element) { Node *node = NewNode(lexer->allocator, lexer); node->start = lexer->lexsize; AddByte(lexer, '<'); if (element->type == EndTag) AddByte(lexer, '/'); if (element->element) { char *p; for (p = element->element; *p != '\0'; ++p) AddByte(lexer, *p); } else if (element->type == DocTypeTag) { uint i; AddStringLiteral( lexer, "!DOCTYPE " ); for (i = element->start; i < element->end; ++i) AddByte(lexer, lexer->lexbuf[i]); } if (element->type == StartEndTag) AddByte(lexer, '/'); AddByte(lexer, '>'); node->end = lexer->lexsize; return node; } #endif /* 0 */ /* Only true for text nodes. */ Bool TY_(IsBlank)(Lexer *lexer, Node *node) { Bool isBlank = TY_(nodeIsText)(node); if ( isBlank ) isBlank = ( node->end == node->start || /* Zero length */ ( node->end == node->start+1 /* or one blank. */ && lexer->lexbuf[node->start] == ' ' ) ); return isBlank; } /* This maps

hello world to

hello world Trims initial space, by moving it before the start tag, or if this element is the first in parent's content, then by discarding the space */ static void TrimInitialSpace( TidyDocImpl* doc, Node *element, Node *text ) { Lexer* lexer = doc->lexer; Node *prev, *node; if ( TY_(nodeIsText)(text) && lexer->lexbuf[text->start] == ' ' && text->start < text->end ) { if ( (element->tag->model & CM_INLINE) && !(element->tag->model & CM_FIELD) ) { prev = element->prev; if (TY_(nodeIsText)(prev)) { if (prev->end == 0 || lexer->lexbuf[prev->end - 1] != ' ') lexer->lexbuf[(prev->end)++] = ' '; ++(element->start); } else /* create new node */ { node = TY_(NewNode)(lexer->allocator, lexer); node->start = (element->start)++; node->end = element->start; lexer->lexbuf[node->start] = ' '; TY_(InsertNodeBeforeElement)(element ,node); #if !defined(NDEBUG) && defined(_MSC_VER) SPRTF("TrimInitialSpace: Created text node, inserted before <%s>\n", (element->element ? element->element : "unknown")); #endif } } /* discard the space in current node */ ++(text->start); } } static Bool IsPreDescendant(Node* node) { Node *parent = node->parent; while (parent) { if (parent->tag && parent->tag->parser == TY_(ParsePre)) return yes; parent = parent->parent; } return no; } static Bool CleanTrailingWhitespace(TidyDocImpl* doc, Node* node) { Node* next; if (!TY_(nodeIsText)(node)) return no; if (node->parent->type == DocTypeTag) return no; if (IsPreDescendant(node)) return no; if (node->parent->tag && node->parent->tag->parser == TY_(ParseScript)) return no; next = node->next; /*

...

*/ if (!next && !TY_(nodeHasCM)(node->parent, CM_INLINE)) return yes; /*
...

...

*/ if (!next && node->parent->next && !TY_(nodeHasCM)(node->parent->next, CM_INLINE)) return yes; if (!next) return no; if (nodeIsBR(next)) return yes; if (TY_(nodeHasCM)(next, CM_INLINE)) return no; /* ...

...

*/ if (next->type == StartTag) return yes; /* ...
*/ if (next->type == StartEndTag) return yes; /* evil adjacent text nodes, Tidy should not generate these :-( */ if (TY_(nodeIsText)(next) && next->start < next->end && TY_(IsWhite)(doc->lexer->lexbuf[next->start])) return yes; return no; } static Bool CleanLeadingWhitespace(TidyDocImpl* ARG_UNUSED(doc), Node* node) { if (!TY_(nodeIsText)(node)) return no; if (node->parent->type == DocTypeTag) return no; if (IsPreDescendant(node)) return no; if (node->parent->tag && node->parent->tag->parser == TY_(ParseScript)) return no; /*

...
......

*/ if (nodeIsBR(node->prev)) return yes; /*

...

*/ if (node->prev == NULL && !TY_(nodeHasCM)(node->parent, CM_INLINE)) return yes; /*

...

... */ if (node->prev && !TY_(nodeHasCM)(node->prev, CM_INLINE) && TY_(nodeIsElement)(node->prev)) return yes; /*

...

*/ if (!node->prev && !node->parent->prev && !TY_(nodeHasCM)(node->parent->parent, CM_INLINE)) return yes; return no; } static void CleanSpaces(TidyDocImpl* doc, Node* node) { Node* next; while (node) { next = node->next; if (TY_(nodeIsText)(node) && CleanLeadingWhitespace(doc, node)) while (node->start < node->end && TY_(IsWhite)(doc->lexer->lexbuf[node->start])) ++(node->start); if (TY_(nodeIsText)(node) && CleanTrailingWhitespace(doc, node)) while (node->end > node->start && TY_(IsWhite)(doc->lexer->lexbuf[node->end - 1])) --(node->end); if (TY_(nodeIsText)(node) && !(node->start < node->end)) { TY_(RemoveNode)(node); TY_(FreeNode)(doc, node); node = next; continue; } if (node->content) CleanSpaces(doc, node->content); node = next; } } /* Move initial and trailing space out. This routine maps: hello world to hello world and hello world to hello world */ static void TrimSpaces( TidyDocImpl* doc, Node *element) { Node* text = element->content; if (nodeIsPRE(element) || IsPreDescendant(element)) return; if (TY_(nodeIsText)(text)) TrimInitialSpace(doc, element, text); text = element->last; if (TY_(nodeIsText)(text)) TrimTrailingSpace(doc, element, text); } static Bool DescendantOf( Node *element, TidyTagId tid ) { Node *parent; for ( parent = element->parent; parent != NULL; parent = parent->parent ) { if ( TagIsId(parent, tid) ) return yes; } return no; } static Bool InsertMisc(Node *element, Node *node) { if (node->type == CommentTag || node->type == ProcInsTag || node->type == CDATATag || node->type == SectionTag || node->type == AspTag || node->type == JsteTag || node->type == PhpTag ) { TY_(InsertNodeAtEnd)(element, node); return yes; } if ( node->type == XmlDecl ) { Node* root = element; while ( root && root->parent ) root = root->parent; if ( root && !(root->content && root->content->type == XmlDecl)) { TY_(InsertNodeAtStart)( root, node ); return yes; } } /* Declared empty tags seem to be slipping through ** the cracks. This is an experiment to figure out ** a decent place to pick them up. */ if ( node->tag && TY_(nodeIsElement)(node) && TY_(nodeCMIsEmpty)(node) && TagId(node) == TidyTag_UNKNOWN && (node->tag->versions & VERS_PROPRIETARY) != 0 ) { TY_(InsertNodeAtEnd)(element, node); return yes; } return no; } static void ParseTag( TidyDocImpl* doc, Node *node, GetTokenMode mode ) { Lexer* lexer = doc->lexer; if (node->tag == NULL) /* [i_a]2 prevent crash for active content (php, asp) docs */ return; /* Fix by GLP 2000-12-21. Need to reset insertspace if this is both a non-inline and empty tag (base, link, meta, isindex, hr, area). */ if (node->tag->model & CM_EMPTY) { lexer->waswhite = no; if (node->tag->parser == NULL) return; } else if (!(node->tag->model & CM_INLINE)) lexer->insertspace = no; if (node->tag->parser == NULL) return; if (node->type == StartEndTag) return; lexer->parent = node; /* [i_a]2 added this - not sure why - CHECKME: */ (*node->tag->parser)( doc, node, mode ); } /* the doctype has been found after other tags, and needs moving to before the html element */ static void InsertDocType( TidyDocImpl* doc, Node *element, Node *doctype ) { Node* existing = TY_(FindDocType)( doc ); if ( existing ) { TY_(ReportError)(doc, element, doctype, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, doctype ); } else { TY_(ReportError)(doc, element, doctype, DOCTYPE_AFTER_TAGS ); while ( !nodeIsHTML(element) ) element = element->parent; TY_(InsertNodeBeforeElement)( element, doctype ); } } /* move node to the head, where element is used as starting point in hunt for head. normally called during parsing */ static void MoveToHead( TidyDocImpl* doc, Node *element, Node *node ) { Node *head; TY_(RemoveNode)( node ); /* make sure that node is isolated */ if ( TY_(nodeIsElement)(node) ) { TY_(ReportError)(doc, element, node, TAG_NOT_ALLOWED_IN ); head = TY_(FindHEAD)(doc); assert(head != NULL); TY_(InsertNodeAtEnd)(head, node); if ( node->tag->parser ) ParseTag( doc, node, IgnoreWhitespace ); } else { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node ); } } /* moves given node to end of body element */ static void MoveNodeToBody( TidyDocImpl* doc, Node* node ) { Node* body = TY_(FindBody)( doc ); if ( body ) { TY_(RemoveNode)( node ); TY_(InsertNodeAtEnd)( body, node ); } } static void AddClassNoIndent( TidyDocImpl* doc, Node *node ) { ctmbstr sprop = "padding-left: 2ex; margin-left: 0ex" "; margin-top: 0ex; margin-bottom: 0ex"; if ( !cfgBool(doc, TidyDecorateInferredUL) ) return; if ( cfgBool(doc, TidyMakeClean) ) TY_(AddStyleAsClass)( doc, node, sprop ); else TY_(AddStyleProperty)( doc, node, sprop ); } /* element is node created by the lexer upon seeing the start tag, or by the parser when the start tag is inferred */ void TY_(ParseBlock)( TidyDocImpl* doc, Node *element, GetTokenMode mode) { #if !defined(NDEBUG) && defined(_MSC_VER) static int in_parse_block = 0; static int parse_block_cnt = 0; #endif Lexer* lexer = doc->lexer; Node *node; Bool checkstack = yes; uint istackbase = 0; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block++; parse_block_cnt++; SPRTF("Entering ParseBlock %d... %d %s\n",in_parse_block,parse_block_cnt, ((element && element->element) ? element->element : "")); #endif if ( element->tag->model & CM_EMPTY ) { #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlockL 1 %d...\n",in_parse_block); #endif return; } if ( nodeIsFORM(element) && DescendantOf(element, TidyTag_FORM) ) TY_(ReportError)(doc, element, NULL, ILLEGAL_NESTING ); /* InlineDup() asks the lexer to insert inline emphasis tags currently pushed on the istack, but take care to avoid propagating inline emphasis inside OBJECT or APPLET. For these elements a fresh inline stack context is created and disposed of upon reaching the end of the element. They thus behave like table cells in this respect. */ if (element->tag->model & CM_OBJECT) { istackbase = lexer->istackbase; lexer->istackbase = lexer->istacksize; } if (!(element->tag->model & CM_MIXED)) TY_(InlineDup)( doc, NULL ); /*\ * Issue #212 - If it is likely that it may be necessary * to move a leading space into a text node before this * element, then keep the mode MixedContent to keep any * leading space \*/ if ( !(element->tag->model & CM_INLINE) || (element->tag->model & CM_FIELD ) ) { mode = IgnoreWhitespace; } else if (mode == IgnoreWhitespace) { /* Issue #212 - Further fix in case ParseBlock() is called with 'IgnoreWhitespace' when such a leading space may need to be inserted before this element to preverve the browser view */ mode = MixedContent; } while ((node = TY_(GetToken)(doc, mode /*MixedContent*/)) != NULL) { /* end tag for this element */ if (node->type == EndTag && node->tag && (node->tag == element->tag || element->was == node->tag)) { TY_(FreeNode)( doc, node ); if (element->tag->model & CM_OBJECT) { /* pop inline stack */ while (lexer->istacksize > lexer->istackbase) TY_(PopInline)( doc, NULL ); lexer->istackbase = istackbase; } element->closed = yes; TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 2 %d...\n",in_parse_block); #endif return; } #if OBSOLETE /* Issue #380 Kill this code! But leave in src, just in case! */ if ( nodeIsBODY( node ) && DescendantOf( element, TidyTag_HEAD )) { /* If we're in the HEAD, close it before proceeding. This is an extremely rare occurance, but has been observed. **************************************************************** Issue #380 - This can cause an INFINITE loop! This code was added to SF CVS Tidy revision 1.121 by lpassey, Wed Jul 28 18:08:06 2004 UTC **************************************************************** */ TY_(UngetToken)( doc ); break; } #endif /* #if OBSOLETE */ if ( nodeIsHTML(node) || nodeIsHEAD(node) || nodeIsBODY(node) ) { if ( TY_(nodeIsElement)(node) ) TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } if (node->type == EndTag) { if (node->tag == NULL) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } else if ( nodeIsBR(node) ) node->type = StartTag; else if ( nodeIsP(node) ) { /* Cannot have a block inside a paragraph, so no checking for an ancestor is necessary -- but we _can_ have paragraphs inside a block, so change it to an implicit empty paragraph, to be dealt with according to the user's options */ node->type = StartEndTag; node->implicit = yes; #if OBSOLETE TY_(CoerceNode)(doc, node, TidyTag_BR, no, no); TY_(FreeAttrs)( doc, node ); /* discard align attribute etc. */ TY_(InsertNodeAtEnd)( element, node ); node = InferredTag(doc, TidyTag_BR); #endif } else if (DescendantOf( element, node->tag->id )) { /* if this is the end tag for an ancestor element then infer end tag for this element */ TY_(UngetToken)( doc ); break; #if OBSOLETE Node *parent; for ( parent = element->parent; parent != NULL; parent = parent->parent ) { if (node->tag == parent->tag) { if (!(element->tag->model & CM_OPT)) TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE ); TY_(UngetToken)( doc ); if (element->tag->model & CM_OBJECT) { /* pop inline stack */ while (lexer->istacksize > lexer->istackbase) TY_(PopInline)( doc, NULL ); lexer->istackbase = istackbase; } TrimSpaces( doc, element ); return; } } #endif } else { /* special case etc. for stuff moved in front of table */ if ( lexer->exiled && (TY_(nodeHasCM)(node, CM_TABLE) || nodeIsTABLE(node)) ) { TY_(UngetToken)( doc ); TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 2 %d...\n",in_parse_block); #endif return; } } } /* mixed content model permits text */ if (TY_(nodeIsText)(node)) { if ( checkstack ) { checkstack = no; if (!(element->tag->model & CM_MIXED)) { if ( TY_(InlineDup)(doc, node) > 0 ) continue; } } TY_(InsertNodeAtEnd)(element, node); mode = MixedContent; /* HTML4 strict doesn't allow mixed content for elements with %block; as their content model */ /* But only body, map, blockquote, form and noscript have content model %block; */ if ( nodeIsBODY(element) || nodeIsMAP(element) || nodeIsBLOCKQUOTE(element) || nodeIsFORM(element) || nodeIsNOSCRIPT(element) ) TY_(ConstrainVersion)( doc, ~VERS_HTML40_STRICT ); continue; } if ( InsertMisc(element, node) ) continue; /* allow PARAM elements? */ if ( nodeIsPARAM(node) ) { if ( TY_(nodeHasCM)(element, CM_PARAM) && TY_(nodeIsElement)(node) ) { TY_(InsertNodeAtEnd)(element, node); continue; } /* otherwise discard it */ TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } /* allow AREA elements? */ if ( nodeIsAREA(node) ) { if ( nodeIsMAP(element) && TY_(nodeIsElement)(node) ) { TY_(InsertNodeAtEnd)(element, node); continue; } /* otherwise discard it */ TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } /* ignore unknown start/end tags */ if ( node->tag == NULL ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } /* Allow CM_INLINE elements here. Allow CM_BLOCK elements here unless lexer->excludeBlocks is yes. LI and DD are special cased. Otherwise infer end tag for this element. */ if ( !TY_(nodeHasCM)(node, CM_INLINE) ) { if ( !TY_(nodeIsElement)(node) ) { if ( nodeIsFORM(node) ) BadForm( doc ); TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } /* #427671 - Fix by Randy Waki - 10 Aug 00 */ /* If an LI contains an illegal FRAME, FRAMESET, OPTGROUP, or OPTION start tag, discard the start tag and let the subsequent content get parsed as content of the enclosing LI. This seems to mimic IE and Netscape, and avoids an infinite loop: without this check, ParseBlock (which is parsing the LI's content) and ParseList (which is parsing the LI's parent's content) repeatedly defer to each other to parse the illegal start tag, each time inferring a missing or
  • respectively. NOTE: This check is a bit fragile. It specifically checks for the four tags that happen to weave their way through the current series of tests performed by ParseBlock and ParseList to trigger the infinite loop. */ if ( nodeIsLI(element) ) { if ( nodeIsFRAME(node) || nodeIsFRAMESET(node) || nodeIsOPTGROUP(node) || nodeIsOPTION(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); /* DSR - 27Apr02 avoid memory leak */ continue; } } if ( nodeIsTD(element) || nodeIsTH(element) ) { /* if parent is a table cell, avoid inferring the end of the cell */ if ( TY_(nodeHasCM)(node, CM_HEAD) ) { MoveToHead( doc, element, node ); continue; } if ( TY_(nodeHasCM)(node, CM_LIST) ) { TY_(UngetToken)( doc ); node = TY_(InferredTag)(doc, TidyTag_UL); AddClassNoIndent(doc, node); lexer->excludeBlocks = yes; } else if ( TY_(nodeHasCM)(node, CM_DEFLIST) ) { TY_(UngetToken)( doc ); node = TY_(InferredTag)(doc, TidyTag_DL); lexer->excludeBlocks = yes; } /* infer end of current table cell */ if ( !TY_(nodeHasCM)(node, CM_BLOCK) ) { TY_(UngetToken)( doc ); TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 3 %d...\n",in_parse_block); #endif return; } } else if ( TY_(nodeHasCM)(node, CM_BLOCK) ) { if ( lexer->excludeBlocks ) { if ( !TY_(nodeHasCM)(element, CM_OPT) ) TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE ); TY_(UngetToken)( doc ); if ( TY_(nodeHasCM)(element, CM_OBJECT) ) lexer->istackbase = istackbase; TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 4 %d...\n",in_parse_block); #endif return; } } else /* things like list items */ { if (node->tag->model & CM_HEAD) { MoveToHead( doc, element, node ); continue; } /* special case where a form start tag occurs in a tr and is followed by td or th */ if ( nodeIsFORM(element) && nodeIsTD(element->parent) && element->parent->implicit ) { if ( nodeIsTD(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } if ( nodeIsTH(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); node = element->parent; TidyDocFree(doc, node->element); node->element = TY_(tmbstrdup)(doc->allocator, "th"); node->tag = TY_(LookupTagDef)( TidyTag_TH ); continue; } } if ( !TY_(nodeHasCM)(element, CM_OPT) && !element->implicit ) TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE ); TY_(UngetToken)( doc ); if ( TY_(nodeHasCM)(node, CM_LIST) ) { if ( element->parent && element->parent->tag && element->parent->tag->parser == TY_(ParseList) ) { TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 5 %d...\n",in_parse_block); #endif return; } node = TY_(InferredTag)(doc, TidyTag_UL); AddClassNoIndent(doc, node); } else if ( TY_(nodeHasCM)(node, CM_DEFLIST) ) { if ( nodeIsDL(element->parent) ) { TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 6 %d...\n",in_parse_block); #endif return; } node = TY_(InferredTag)(doc, TidyTag_DL); } else if ( TY_(nodeHasCM)(node, CM_TABLE) || TY_(nodeHasCM)(node, CM_ROW) ) { /* http://tidy.sf.net/issue/1316307 */ /* In exiled mode, return so table processing can continue. */ if (lexer->exiled) { #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 7 %d...\n",in_parse_block); #endif return; } node = TY_(InferredTag)(doc, TidyTag_TABLE); } else if ( TY_(nodeHasCM)(element, CM_OBJECT) ) { /* pop inline stack */ while ( lexer->istacksize > lexer->istackbase ) TY_(PopInline)( doc, NULL ); lexer->istackbase = istackbase; TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 8 %d...\n",in_parse_block); #endif return; } else { TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 9 %d...\n",in_parse_block); #endif return; } } } /*\ * Issue #307 - an tag to ends any open element * Like #427827 - fixed by Randy Waki and Bjoern Hoehrmann 23 Aug 00 * in ParseInline(), fix copied HERE to ParseBlock() * href: http://www.w3.org/TR/html-markup/a.html * The interactive element a must not appear as a descendant of the a element. \*/ if ( nodeIsA(node) && !node->implicit && (nodeIsA(element) || DescendantOf(element, TidyTag_A)) ) { if (node->type != EndTag && node->attributes == NULL && cfgBool(doc, TidyCoerceEndTags) ) { node->type = EndTag; TY_(ReportError)(doc, element, node, COERCE_TO_ENDTAG); TY_(UngetToken)( doc ); continue; } if (nodeIsA(element)) { TY_(UngetToken)( doc ); } TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE); if (!(mode & Preformatted)) TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 9b %d...\n",in_parse_block); #endif return; } /* parse known element */ if (TY_(nodeIsElement)(node)) { if (node->tag->model & CM_INLINE) { if (checkstack && !node->implicit) { checkstack = no; if (!(element->tag->model & CM_MIXED)) /* #431731 - fix by Randy Waki 25 Dec 00 */ { if ( TY_(InlineDup)(doc, node) > 0 ) continue; } } mode = MixedContent; } else { checkstack = yes; mode = IgnoreWhitespace; } /* trim white space before
    */ if ( nodeIsBR(node) ) TrimSpaces( doc, element ); TY_(InsertNodeAtEnd)(element, node); if (node->implicit) TY_(ReportError)(doc, element, node, INSERTING_TAG ); /* Issue #212 - WHY is this hard coded to 'IgnoreWhitespace' while an effort has been made above to set a 'MixedContent' mode in some cases? WHY IS THE 'mode' VARIABLE NOT USED HERE???? */ ParseTag( doc, node, IgnoreWhitespace /*MixedContent*/ ); continue; } /* discard unexpected tags */ if (node->type == EndTag) TY_(PopInline)( doc, node ); /* if inline end tag */ TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } if (!(element->tag->model & CM_OPT)) TY_(ReportError)(doc, element, node, MISSING_ENDTAG_FOR); if (element->tag->model & CM_OBJECT) { /* pop inline stack */ while ( lexer->istacksize > lexer->istackbase ) TY_(PopInline)( doc, NULL ); lexer->istackbase = istackbase; } TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_block--; SPRTF("Exit ParseBlock 10 %d...\n",in_parse_block); #endif } /* [i_a] svg / math */ struct MatchingDescendantData { Node *found_node; Bool *passed_marker_node; /* input: */ TidyTagId matching_tagId; Node *node_to_find; Node *marker_node; }; static NodeTraversalSignal FindDescendant_cb(TidyDocImpl* ARG_UNUSED(doc), Node* node, void *propagate) { struct MatchingDescendantData *cb_data = (struct MatchingDescendantData *)propagate; if (TagId(node) == cb_data->matching_tagId) { /* make sure we match up 'unknown' tags exactly! */ if (cb_data->matching_tagId != TidyTag_UNKNOWN || (node->element != NULL && cb_data->node_to_find != NULL && cb_data->node_to_find->element != NULL && 0 == TY_(tmbstrcmp)(cb_data->node_to_find->element, node->element))) { cb_data->found_node = node; return ExitTraversal; } } if (cb_data->passed_marker_node && node == cb_data->marker_node) *cb_data->passed_marker_node = yes; return VisitParent; } /* Search the parent chain (from 'parent' upwards up to the root) for a node matching the given 'node'. When the search passes beyond the 'marker_node' (which is assumed to sit in the parent chain), this will be flagged by setting the boolean referenced by 'is_parent_of_marker' to yes. 'is_parent_of_marker' and 'marker_node' are optional parameters and may be NULL. */ static Node *FindMatchingDescendant( Node *parent, Node *node, Node *marker_node, Bool *is_parent_of_marker ) { struct MatchingDescendantData cb_data = { 0 }; cb_data.matching_tagId = TagId(node); cb_data.node_to_find = node; cb_data.marker_node = marker_node; assert(node); if (is_parent_of_marker) *is_parent_of_marker = no; TY_(TraverseNodeTree)(NULL, parent, FindDescendant_cb, &cb_data); return cb_data.found_node; } /* Act as a generic XML (sub)tree parser: collect each node and add it to the DOM, without any further validation. TODO : add schema- or other-hierarchy-definition-based validation of the subtree here... */ void TY_(ParseNamespace)(TidyDocImpl* doc, Node *basenode, GetTokenMode mode) { Lexer* lexer = doc->lexer; Node *node; Node *parent = basenode; uint istackbase; AttVal* av; /* #130 MathML attr and entity fix! */ /* a la : defer popping elements off the inline stack */ TY_(DeferDup)( doc ); istackbase = lexer->istackbase; lexer->istackbase = lexer->istacksize; mode = OtherNamespace; /* Preformatted; IgnoreWhitespace; */ while ((node = TY_(GetToken)(doc, mode)) != NULL) { /* fix check to skip action in InsertMisc for regular/empty nodes, which we don't want here... The way we do it here is by checking and processing everything and only what remains goes into InsertMisc() */ /* is this a close tag? And does it match the current parent node? */ if (node->type == EndTag) { /* to prevent end tags flowing from one 'alternate namespace' we check this in two phases: first we check if the tag is a descendant of the current node, and when it is, we check whether it is the end tag for a node /within/ or /outside/ the basenode. */ Bool outside; Node *mp = FindMatchingDescendant(parent, node, basenode, &outside); if (mp != NULL) { /* when mp != parent as we might expect, infer end tags until we 'hit' the matched parent or the basenode */ Node *n; for (n = parent; n != NULL && n != basenode->parent && n != mp; n = n->parent) { /* n->implicit = yes; */ n->closed = yes; TY_(ReportError)(doc, n->parent, n, MISSING_ENDTAG_BEFORE); } /* Issue #369 - Since 'assert' is DEBUG only, and there are simple cases where these can be fired, removing them pending feedback from the original author! assert(outside == no ? n == mp : 1); assert(outside == yes ? n == basenode->parent : 1); =================================================== */ if (outside == no) { /* EndTag for a node within the basenode subtree. Roll on... */ n->closed = yes; TY_(FreeNode)(doc, node); node = n; parent = node->parent; } else { /* EndTag for a node outside the basenode subtree: let the caller handle that. */ TY_(UngetToken)( doc ); node = basenode; parent = node->parent; } /* when we've arrived at the end-node for the base node, it's quitting time */ if (node == basenode) { lexer->istackbase = istackbase; assert(basenode->closed == yes); return; } } else { /* unmatched close tag: report an error and discard */ /* TY_(ReportError)(doc, parent, node, NON_MATCHING_ENDTAG); Issue #308 - Seems wrong warning! */ TY_(ReportError)(doc, parent, node, DISCARDING_UNEXPECTED); assert(parent); /* assert(parent->tag != node->tag); Issue #308 - Seems would always be true! */ TY_(FreeNode)( doc, node); /* Issue #308 - Discard unexpected end tag memory */ } } else if (node->type == StartTag) { /* #130 MathML attr and entity fix! care if it has attributes, and 'accidently' any of those attributes match known */ for ( av = node->attributes; av; av = av->next ) { av->dict = 0; /* does something need to be freed? */ } /* add another child to the current parent */ TY_(InsertNodeAtEnd)(parent, node); parent = node; } else { /* #130 MathML attr and entity fix! care if it has attributes, and 'accidently' any of those attributes match known */ for ( av = node->attributes; av; av = av->next ) { av->dict = 0; /* does something need to be freed? */ } TY_(InsertNodeAtEnd)(parent, node); } } TY_(ReportError)(doc, basenode->parent, basenode, MISSING_ENDTAG_FOR); } void TY_(ParseInline)( TidyDocImpl* doc, Node *element, GetTokenMode mode ) { #if !defined(NDEBUG) && defined(_MSC_VER) static int in_parse_inline = 0; #endif Lexer* lexer = doc->lexer; Node *node, *parent; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline++; SPRTF("Entering ParseInline %d...\n",in_parse_inline); #endif if (element->tag->model & CM_EMPTY) { #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 1 %d...\n",in_parse_inline); #endif return; } /* ParseInline is used for some block level elements like H1 to H6 For such elements we need to insert inline emphasis tags currently on the inline stack. For Inline elements, we normally push them onto the inline stack provided they aren't implicit or OBJECT/APPLET. This test is carried out in PushInline and PopInline, see istack.c InlineDup(...) is not called for elements with a CM_MIXED (inline and block) content model, e.g. or , otherwise constructs like

    111222333444555

    111222333444555

    111222333444555

    will get corrupted. */ if ((TY_(nodeHasCM)(element, CM_BLOCK) || nodeIsDT(element)) && !TY_(nodeHasCM)(element, CM_MIXED)) TY_(InlineDup)(doc, NULL); else if (TY_(nodeHasCM)(element, CM_INLINE)) TY_(PushInline)(doc, element); if ( nodeIsNOBR(element) ) doc->badLayout |= USING_NOBR; else if ( nodeIsFONT(element) ) doc->badLayout |= USING_FONT; /* Inline elements may or may not be within a preformatted element */ if (mode != Preformatted) mode = MixedContent; while ((node = TY_(GetToken)(doc, mode)) != NULL) { /* end tag for current element */ if (node->tag == element->tag && node->type == EndTag) { if (element->tag->model & CM_INLINE) TY_(PopInline)( doc, node ); TY_(FreeNode)( doc, node ); if (!(mode & Preformatted)) TrimSpaces(doc, element); /* if a font element wraps an anchor and nothing else then move the font element inside the anchor since otherwise it won't alter the anchor text color */ if ( nodeIsFONT(element) && element->content && element->content == element->last ) { Node *child = element->content; if ( nodeIsA(child) ) { child->parent = element->parent; child->next = element->next; child->prev = element->prev; element->next = NULL; element->prev = NULL; element->parent = child; element->content = child->content; element->last = child->last; child->content = element; TY_(FixNodeLinks)(child); TY_(FixNodeLinks)(element); } } element->closed = yes; TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 2 %d...\n",in_parse_inline); #endif return; } /* ... map 2nd to if 1st is explicit */ /* (see additional conditions below) */ /* otherwise emphasis nesting is probably unintentional */ /* big, small, sub, sup have cumulative effect to leave them alone */ if ( node->type == StartTag && node->tag == element->tag && TY_(IsPushed)( doc, node ) && !node->implicit && !element->implicit && node->tag && (node->tag->model & CM_INLINE) && !nodeIsA(node) && !nodeIsFONT(node) && !nodeIsBIG(node) && !nodeIsSMALL(node) && !nodeIsSUB(node) && !nodeIsSUP(node) && !nodeIsQ(node) && !nodeIsSPAN(node) && cfgBool(doc, TidyCoerceEndTags) ) { /* proceeds only if "node" does not have any attribute and follows a text node not finishing with a space */ if (element->content != NULL && node->attributes == NULL && TY_(nodeIsText)(element->last) && !TY_(TextNodeEndWithSpace)(doc->lexer, element->last) ) { TY_(ReportWarning)(doc, element, node, COERCE_TO_ENDTAG_WARN); node->type = EndTag; TY_(UngetToken)(doc); continue; } if (node->attributes == NULL || element->attributes == NULL) TY_(ReportWarning)(doc, element, node, NESTED_EMPHASIS); } else if ( TY_(IsPushed)(doc, node) && node->type == StartTag && nodeIsQ(node) ) { /*\ * Issue #215 - such nested quotes are NOT a problem if HTML5, so * only issue this warning if NOT HTML5 mode. \*/ if (TY_(HTMLVersion)(doc) != HT50) { TY_(ReportWarning)(doc, element, node, NESTED_QUOTATION); } } if ( TY_(nodeIsText)(node) ) { /* only called for 1st child */ if ( element->content == NULL && !(mode & Preformatted) ) TrimSpaces( doc, element ); if ( node->start >= node->end ) { TY_(FreeNode)( doc, node ); continue; } TY_(InsertNodeAtEnd)(element, node); continue; } /* mixed content model so allow text */ if (InsertMisc(element, node)) continue; /* deal with HTML tags */ if ( nodeIsHTML(node) ) { if ( TY_(nodeIsElement)(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node ); continue; } /* otherwise infer end of inline element */ TY_(UngetToken)( doc ); if (!(mode & Preformatted)) TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 3 %d...\n",in_parse_inline); #endif return; } /* within
    or
     map 

    to
    */ if ( nodeIsP(node) && node->type == StartTag && ( (mode & Preformatted) || nodeIsDT(element) || DescendantOf(element, TidyTag_DT ) ) ) { node->tag = TY_(LookupTagDef)( TidyTag_BR ); TidyDocFree(doc, node->element); node->element = TY_(tmbstrdup)(doc->allocator, "br"); TrimSpaces(doc, element); TY_(InsertNodeAtEnd)(element, node); continue; } /*

    allowed within

    in HTML 4.01 Transitional */ if ( nodeIsP(node) && node->type == StartTag && nodeIsADDRESS(element) ) { TY_(ConstrainVersion)( doc, ~VERS_HTML40_STRICT ); TY_(InsertNodeAtEnd)(element, node); (*node->tag->parser)( doc, node, mode ); continue; } /* ignore unknown and PARAM tags */ if ( node->tag == NULL || nodeIsPARAM(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node ); continue; } if ( nodeIsBR(node) && node->type == EndTag ) node->type = StartTag; if ( node->type == EndTag ) { /* coerce
    to
    */ if ( nodeIsBR(node) ) node->type = StartTag; else if ( nodeIsP(node) ) { /* coerce unmatched

    to

    */ if ( !DescendantOf(element, TidyTag_P) ) { TY_(CoerceNode)(doc, node, TidyTag_BR, no, no); TrimSpaces( doc, element ); TY_(InsertNodeAtEnd)( element, node ); node = TY_(InferredTag)(doc, TidyTag_BR); TY_(InsertNodeAtEnd)( element, node ); /* todo: check this */ continue; } } else if ( TY_(nodeHasCM)(node, CM_INLINE) && !nodeIsA(node) && !TY_(nodeHasCM)(node, CM_OBJECT) && TY_(nodeHasCM)(element, CM_INLINE) ) { /* allow any inline end tag to end current element */ /* http://tidy.sf.net/issue/1426419 */ /* but, like the browser, retain an earlier inline element. This is implemented by setting the lexer into a mode where it gets tokens from the inline stack rather than from the input stream. Check if the scenerio fits. */ if ( !nodeIsA(element) && (node->tag != element->tag) && TY_(IsPushed)( doc, node ) && TY_(IsPushed)( doc, element ) ) { /* we have something like bold bold and italic italics */ if ( TY_(SwitchInline)( doc, element, node ) ) { TY_(ReportError)(doc, element, node, NON_MATCHING_ENDTAG); TY_(UngetToken)( doc ); /* put this back */ TY_(InlineDup1)( doc, NULL, element ); /* dupe the , after */ if (!(mode & Preformatted)) TrimSpaces( doc, element ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 4 %d...\n",in_parse_inline); #endif return; /* close , but will re-open it, after */ } } TY_(PopInline)( doc, element ); if ( !nodeIsA(element) ) { if ( nodeIsA(node) && node->tag != element->tag ) { TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE ); TY_(UngetToken)( doc ); } else { TY_(ReportError)(doc, element, node, NON_MATCHING_ENDTAG); TY_(FreeNode)( doc, node); } if (!(mode & Preformatted)) TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 5 %d...\n",in_parse_inline); #endif return; } /* if parent is then discard unexpected inline end tag */ TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* special case etc. for stuff moved in front of table */ else if ( lexer->exiled && (TY_(nodeHasCM)(node, CM_TABLE) || nodeIsTABLE(node)) ) { TY_(UngetToken)( doc ); TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 6 %d...\n",in_parse_inline); #endif return; } } /* allow any header tag to end current header */ if ( TY_(nodeHasCM)(node, CM_HEADING) && TY_(nodeHasCM)(element, CM_HEADING) ) { if ( node->tag == element->tag ) { TY_(ReportError)(doc, element, node, NON_MATCHING_ENDTAG ); TY_(FreeNode)( doc, node); } else { TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE ); TY_(UngetToken)( doc ); } if (!(mode & Preformatted)) TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 7 %d...\n",in_parse_inline); #endif return; } /* an tag to ends any open element but is mapped to */ /* #427827 - fix by Randy Waki and Bjoern Hoehrmann 23 Aug 00 */ /* if (node->tag == doc->tags.tag_a && !node->implicit && TY_(IsPushed)(doc, node)) */ if ( nodeIsA(node) && !node->implicit && (nodeIsA(element) || DescendantOf(element, TidyTag_A)) ) { /* coerce to unless it has some attributes */ /* #427827 - fix by Randy Waki and Bjoern Hoehrmann 23 Aug 00 */ /* other fixes by Dave Raggett */ /* if (node->attributes == NULL) */ if (node->type != EndTag && node->attributes == NULL && cfgBool(doc, TidyCoerceEndTags) ) { node->type = EndTag; TY_(ReportError)(doc, element, node, COERCE_TO_ENDTAG); /* TY_(PopInline)( doc, node ); */ TY_(UngetToken)( doc ); continue; } TY_(UngetToken)( doc ); TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE); /* TY_(PopInline)( doc, element ); */ if (!(mode & Preformatted)) TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 8 %d...\n",in_parse_inline); #endif return; } if (element->tag->model & CM_HEADING) { if ( nodeIsCENTER(node) || nodeIsDIV(node) ) { if (!TY_(nodeIsElement)(node)) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } TY_(ReportError)(doc, element, node, TAG_NOT_ALLOWED_IN); /* insert center as parent if heading is empty */ if (element->content == NULL) { InsertNodeAsParent(element, node); continue; } /* split heading and make center parent of 2nd part */ TY_(InsertNodeAfterElement)(element, node); if (!(mode & Preformatted)) TrimSpaces(doc, element); element = TY_(CloneNode)( doc, element ); TY_(InsertNodeAtEnd)(node, element); continue; } if ( nodeIsHR(node) ) { if ( !TY_(nodeIsElement)(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } TY_(ReportError)(doc, element, node, TAG_NOT_ALLOWED_IN); /* insert hr before heading if heading is empty */ if (element->content == NULL) { TY_(InsertNodeBeforeElement)(element, node); continue; } /* split heading and insert hr before 2nd part */ TY_(InsertNodeAfterElement)(element, node); if (!(mode & Preformatted)) TrimSpaces(doc, element); element = TY_(CloneNode)( doc, element ); TY_(InsertNodeAfterElement)(node, element); continue; } } if ( nodeIsDT(element) ) { if ( nodeIsHR(node) ) { Node *dd; if ( !TY_(nodeIsElement)(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } TY_(ReportError)(doc, element, node, TAG_NOT_ALLOWED_IN); dd = TY_(InferredTag)(doc, TidyTag_DD); /* insert hr within dd before dt if dt is empty */ if (element->content == NULL) { TY_(InsertNodeBeforeElement)(element, dd); TY_(InsertNodeAtEnd)(dd, node); continue; } /* split dt and insert hr within dd before 2nd part */ TY_(InsertNodeAfterElement)(element, dd); TY_(InsertNodeAtEnd)(dd, node); if (!(mode & Preformatted)) TrimSpaces(doc, element); element = TY_(CloneNode)( doc, element ); TY_(InsertNodeAfterElement)(dd, element); continue; } } /* if this is the end tag for an ancestor element then infer end tag for this element */ if (node->type == EndTag) { for (parent = element->parent; parent != NULL; parent = parent->parent) { if (node->tag == parent->tag) { if (!(element->tag->model & CM_OPT) && !element->implicit) TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE); if( TY_(IsPushedLast)( doc, element, node ) ) TY_(PopInline)( doc, element ); TY_(UngetToken)( doc ); if (!(mode & Preformatted)) TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 9 %d...\n",in_parse_inline); #endif return; } } } /* block level tags end this element */ if (!(node->tag->model & CM_INLINE) && !(element->tag->model & CM_MIXED)) { if ( !TY_(nodeIsElement)(node) ) { TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* HTML5 */ if (nodeIsDATALIST(element)) { TY_(ConstrainVersion)( doc, ~VERS_HTML5 ); } else if (!(element->tag->model & CM_OPT)) TY_(ReportError)(doc, element, node, MISSING_ENDTAG_BEFORE); if (node->tag->model & CM_HEAD && !(node->tag->model & CM_BLOCK)) { MoveToHead(doc, element, node); continue; } /* prevent anchors from propagating into block tags except for headings h1 to h6 */ if ( nodeIsA(element) ) { if (node->tag && !(node->tag->model & CM_HEADING)) TY_(PopInline)( doc, element ); else if (!(element->content)) { TY_(DiscardElement)( doc, element ); TY_(UngetToken)( doc ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 10 %d...\n",in_parse_inline); #endif return; } } TY_(UngetToken)( doc ); if (!(mode & Preformatted)) TrimSpaces(doc, element); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 11 %d...\n",in_parse_inline); #endif return; } /* parse inline element */ if (TY_(nodeIsElement)(node)) { if (node->implicit) TY_(ReportError)(doc, element, node, INSERTING_TAG); /* trim white space before
    */ if ( nodeIsBR(node) ) TrimSpaces(doc, element); TY_(InsertNodeAtEnd)(element, node); ParseTag(doc, node, mode); continue; } /* discard unexpected tags */ TY_(ReportError)(doc, element, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node ); continue; } if (!(element->tag->model & CM_OPT)) TY_(ReportError)(doc, element, node, MISSING_ENDTAG_FOR); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_inline--; SPRTF("Exit ParseInline 12 %d...\n",in_parse_inline); #endif } void TY_(ParseEmpty)(TidyDocImpl* doc, Node *element, GetTokenMode mode) { Lexer* lexer = doc->lexer; if ( lexer->isvoyager ) { Node *node = TY_(GetToken)( doc, mode); if ( node ) { if ( !(node->type == EndTag && node->tag == element->tag) ) { /* TY_(ReportError)(doc, element, node, ELEMENT_NOT_EMPTY); */ TY_(UngetToken)( doc ); } else { TY_(FreeNode)( doc, node ); } } } } void TY_(ParseDefList)(TidyDocImpl* doc, Node *list, GetTokenMode mode) { Lexer* lexer = doc->lexer; Node *node, *parent; if (list->tag->model & CM_EMPTY) return; lexer->insert = NULL; /* defer implicit inline start tags */ while ((node = TY_(GetToken)( doc, IgnoreWhitespace)) != NULL) { if (node->tag == list->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); list->closed = yes; return; } /* deal with comments etc. */ if (InsertMisc(list, node)) continue; if (TY_(nodeIsText)(node)) { TY_(UngetToken)( doc ); node = TY_(InferredTag)(doc, TidyTag_DT); TY_(ReportError)(doc, list, node, MISSING_STARTTAG); } if (node->tag == NULL) { TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* if this is the end tag for an ancestor element then infer end tag for this element */ if (node->type == EndTag) { Bool discardIt = no; if ( nodeIsFORM(node) ) { BadForm( doc ); TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node ); continue; } for (parent = list->parent; parent != NULL; parent = parent->parent) { /* Do not match across BODY to avoid infinite loop between ParseBody and this parser, See http://tidy.sf.net/bug/1098012. */ if (nodeIsBODY(parent)) { discardIt = yes; break; } if (node->tag == parent->tag) { TY_(ReportError)(doc, list, node, MISSING_ENDTAG_BEFORE); TY_(UngetToken)( doc ); return; } } if (discardIt) { TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } } /* center in a dt or a dl breaks the dl list in two */ if ( nodeIsCENTER(node) ) { if (list->content) TY_(InsertNodeAfterElement)(list, node); else /* trim empty dl list */ { TY_(InsertNodeBeforeElement)(list, node); /* #540296 tidy dumps with empty definition list */ #if 0 TY_(DiscardElement)(list); #endif } /* #426885 - fix by Glenn Carroll 19 Apr 00, and Gary Dechaines 11 Aug 00 */ /* ParseTag can destroy node, if it finds that * this
    is followed immediately by
    . * It's awkward but necessary to determine if this * has happened. */ parent = node->parent; /* and parse contents of center */ lexer->excludeBlocks = no; ParseTag( doc, node, mode); lexer->excludeBlocks = yes; /* now create a new dl element, * unless node has been blown away because the * center was empty, as above. */ if (parent->last == node) { list = TY_(InferredTag)(doc, TidyTag_DL); TY_(InsertNodeAfterElement)(node, list); } continue; } if ( !(nodeIsDT(node) || nodeIsDD(node)) ) { TY_(UngetToken)( doc ); if (!(node->tag->model & (CM_BLOCK | CM_INLINE))) { TY_(ReportError)(doc, list, node, TAG_NOT_ALLOWED_IN); return; } /* if DD appeared directly in BODY then exclude blocks */ if (!(node->tag->model & CM_INLINE) && lexer->excludeBlocks) return; node = TY_(InferredTag)(doc, TidyTag_DD); TY_(ReportError)(doc, list, node, MISSING_STARTTAG); } if (node->type == EndTag) { TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* node should be
    or
    */ TY_(InsertNodeAtEnd)(list, node); ParseTag( doc, node, IgnoreWhitespace); } TY_(ReportError)(doc, list, node, MISSING_ENDTAG_FOR); } static Bool FindLastLI( Node *list, Node **lastli ) { Node *node; *lastli = NULL; for ( node = list->content; node ; node = node->next ) if ( nodeIsLI(node) && node->type == StartTag ) *lastli=node; return *lastli ? yes:no; } void TY_(ParseList)(TidyDocImpl* doc, Node *list, GetTokenMode ARG_UNUSED(mode)) { #if !defined(NDEBUG) && defined(_MSC_VER) static int in_parse_list = 0; #endif Lexer* lexer = doc->lexer; Node *node, *parent, *lastli; Bool wasblock; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_list++; SPRTF("Entering ParseList %d...\n",in_parse_list); #endif if (list->tag->model & CM_EMPTY) { #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_list--; SPRTF("Exit ParseList 1 %d... CM_EMPTY\n",in_parse_list); #endif return; } lexer->insert = NULL; /* defer implicit inline start tags */ while ((node = TY_(GetToken)( doc, IgnoreWhitespace)) != NULL) { if (node->tag == list->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); list->closed = yes; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_list--; SPRTF("Exit ParseList 2 %d... Endtag\n",in_parse_list); #endif return; } /* deal with comments etc. */ if (InsertMisc(list, node)) continue; if (node->type != TextNode && node->tag == NULL) { TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* if this is the end tag for an ancestor element then infer end tag for this element */ if (node->type == EndTag) { if ( nodeIsFORM(node) ) { BadForm( doc ); TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node ); continue; } if (TY_(nodeHasCM)(node,CM_INLINE)) { TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(PopInline)( doc, node ); TY_(FreeNode)( doc, node); continue; } for ( parent = list->parent; parent != NULL; parent = parent->parent ) { /* Do not match across BODY to avoid infinite loop between ParseBody and this parser, See http://tidy.sf.net/bug/1053626. */ if (nodeIsBODY(parent)) break; if (node->tag == parent->tag) { TY_(ReportError)(doc, list, node, MISSING_ENDTAG_BEFORE); TY_(UngetToken)( doc ); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_list--; SPRTF("Exit ParseList 3 %d... No End Tag\n",in_parse_list); #endif return; } } TY_(ReportError)(doc, list, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } if ( nodeIsLI(node) || TY_(IsHTML5Mode)(doc)) { /* node is
  • Issue #396 - A
      can have Zero or more li elements */ TY_(InsertNodeAtEnd)(list,node); } else { TY_(UngetToken)( doc ); if (TY_(nodeHasCM)(node,CM_BLOCK) && lexer->excludeBlocks) { TY_(ReportError)(doc, list, node, MISSING_ENDTAG_BEFORE); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_list--; SPRTF("Exit ParseList 4 %d... No End Tag\n",in_parse_list); #endif return; } /* http://tidy.sf.net/issue/1316307 */ /* In exiled mode, return so table processing can continue. */ else if ( lexer->exiled && (TY_(nodeHasCM)(node, CM_TABLE|CM_ROWGRP|CM_ROW) || nodeIsTABLE(node)) ) { #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_list--; SPRTF("Exit ParseList 5 %d... exiled\n",in_parse_list); #endif return; } /* http://tidy.sf.net/issue/836462 If "list" is an unordered list, insert the next tag within the last
    • to preserve the numbering to match the visual rendering of most browsers. */ if ( nodeIsOL(list) && FindLastLI(list, &lastli) ) { /* Create a node for error reporting */ node = TY_(InferredTag)(doc, TidyTag_LI); TY_(ReportError)(doc, list, node, MISSING_STARTTAG ); TY_(FreeNode)( doc, node); node = lastli; } else { /* Add an inferred
    • */ wasblock = TY_(nodeHasCM)(node,CM_BLOCK); node = TY_(InferredTag)(doc, TidyTag_LI); /* Add "display: inline" to avoid a blank line after
    • with Internet Explorer. See http://tidy.sf.net/issue/836462 */ TY_(AddStyleProperty)( doc, node, wasblock ? "list-style: none; display: inline" : "list-style: none" ); TY_(ReportError)(doc, list, node, MISSING_STARTTAG ); TY_(InsertNodeAtEnd)(list,node); } } ParseTag( doc, node, IgnoreWhitespace); } TY_(ReportError)(doc, list, node, MISSING_ENDTAG_FOR); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_list--; SPRTF("Exit ParseList 6 %d... missing end tag\n",in_parse_list); #endif } /* unexpected content in table row is moved to just before the table in accordance with Netscape and IE. This code assumes that node hasn't been inserted into the row. */ static void MoveBeforeTable( TidyDocImpl* ARG_UNUSED(doc), Node *row, Node *node ) { Node *table; /* first find the table element */ for (table = row->parent; table; table = table->parent) { if ( nodeIsTABLE(table) ) { TY_(InsertNodeBeforeElement)( table, node ); return; } } /* No table element */ TY_(InsertNodeBeforeElement)( row->parent, node ); } /* if a table row is empty then insert an empty cell this practice is consistent with browser behavior and avoids potential problems with row spanning cells */ static void FixEmptyRow(TidyDocImpl* doc, Node *row) { Node *cell; if (row->content == NULL) { cell = TY_(InferredTag)(doc, TidyTag_TD); TY_(InsertNodeAtEnd)(row, cell); TY_(ReportError)(doc, row, cell, MISSING_STARTTAG); } } void TY_(ParseRow)(TidyDocImpl* doc, Node *row, GetTokenMode ARG_UNUSED(mode)) { Lexer* lexer = doc->lexer; Node *node; Bool exclude_state; if (row->tag->model & CM_EMPTY) return; while ((node = TY_(GetToken)(doc, IgnoreWhitespace)) != NULL) { if (node->tag == row->tag) { if (node->type == EndTag) { TY_(FreeNode)( doc, node); row->closed = yes; FixEmptyRow( doc, row); return; } /* New row start implies end of current row */ TY_(UngetToken)( doc ); FixEmptyRow( doc, row); return; } /* if this is the end tag for an ancestor element then infer end tag for this element */ if ( node->type == EndTag ) { if ( (TY_(nodeHasCM)(node, CM_HTML|CM_TABLE) || nodeIsTABLE(node)) && DescendantOf(row, TagId(node)) ) { TY_(UngetToken)( doc ); return; } if ( nodeIsFORM(node) || TY_(nodeHasCM)(node, CM_BLOCK|CM_INLINE) ) { if ( nodeIsFORM(node) ) BadForm( doc ); TY_(ReportError)(doc, row, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } if ( nodeIsTD(node) || nodeIsTH(node) ) { TY_(ReportError)(doc, row, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } } /* deal with comments etc. */ if (InsertMisc(row, node)) continue; /* discard unknown tags */ if (node->tag == NULL && node->type != TextNode) { TY_(ReportError)(doc, row, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* discard unexpected
  • element */ if ( nodeIsTABLE(node) ) { TY_(ReportError)(doc, row, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* THEAD, TFOOT or TBODY */ if ( TY_(nodeHasCM)(node, CM_ROWGRP) ) { TY_(UngetToken)( doc ); return; } if (node->type == EndTag) { TY_(ReportError)(doc, row, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* if text or inline or block move before table if head content move to head */ if (node->type != EndTag) { if ( nodeIsFORM(node) ) { TY_(UngetToken)( doc ); node = TY_(InferredTag)(doc, TidyTag_TD); TY_(ReportError)(doc, row, node, MISSING_STARTTAG); } else if ( TY_(nodeIsText)(node) || TY_(nodeHasCM)(node, CM_BLOCK | CM_INLINE) ) { MoveBeforeTable( doc, row, node ); TY_(ReportError)(doc, row, node, TAG_NOT_ALLOWED_IN); lexer->exiled = yes; exclude_state = lexer->excludeBlocks; lexer->excludeBlocks = no; if (node->type != TextNode) ParseTag( doc, node, IgnoreWhitespace); lexer->exiled = no; lexer->excludeBlocks = exclude_state; continue; } else if (node->tag->model & CM_HEAD) { TY_(ReportError)(doc, row, node, TAG_NOT_ALLOWED_IN); MoveToHead( doc, row, node); continue; } } if ( !(nodeIsTD(node) || nodeIsTH(node)) ) { TY_(ReportError)(doc, row, node, TAG_NOT_ALLOWED_IN); TY_(FreeNode)( doc, node); continue; } /* node should be
    or */ TY_(InsertNodeAtEnd)(row, node); exclude_state = lexer->excludeBlocks; lexer->excludeBlocks = no; ParseTag( doc, node, IgnoreWhitespace); lexer->excludeBlocks = exclude_state; /* pop inline stack */ while ( lexer->istacksize > lexer->istackbase ) TY_(PopInline)( doc, NULL ); } } void TY_(ParseRowGroup)(TidyDocImpl* doc, Node *rowgroup, GetTokenMode ARG_UNUSED(mode)) { Lexer* lexer = doc->lexer; Node *node, *parent; if (rowgroup->tag->model & CM_EMPTY) return; while ((node = TY_(GetToken)(doc, IgnoreWhitespace)) != NULL) { if (node->tag == rowgroup->tag) { if (node->type == EndTag) { rowgroup->closed = yes; TY_(FreeNode)( doc, node); return; } TY_(UngetToken)( doc ); return; } /* if
    infer end tag */ if ( nodeIsTABLE(node) && node->type == EndTag ) { TY_(UngetToken)( doc ); return; } /* deal with comments etc. */ if (InsertMisc(rowgroup, node)) continue; /* discard unknown tags */ if (node->tag == NULL && node->type != TextNode) { TY_(ReportError)(doc, rowgroup, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* if TD or TH then infer if text or inline or block move before table if head content move to head */ if (node->type != EndTag) { if ( nodeIsTD(node) || nodeIsTH(node) ) { TY_(UngetToken)( doc ); node = TY_(InferredTag)(doc, TidyTag_TR); TY_(ReportError)(doc, rowgroup, node, MISSING_STARTTAG); } else if ( TY_(nodeIsText)(node) || TY_(nodeHasCM)(node, CM_BLOCK|CM_INLINE) ) { MoveBeforeTable( doc, rowgroup, node ); TY_(ReportError)(doc, rowgroup, node, TAG_NOT_ALLOWED_IN); lexer->exiled = yes; if (node->type != TextNode) ParseTag(doc, node, IgnoreWhitespace); lexer->exiled = no; continue; } else if (node->tag->model & CM_HEAD) { TY_(ReportError)(doc, rowgroup, node, TAG_NOT_ALLOWED_IN); MoveToHead(doc, rowgroup, node); continue; } } /* if this is the end tag for ancestor element then infer end tag for this element */ if (node->type == EndTag) { if ( nodeIsFORM(node) || TY_(nodeHasCM)(node, CM_BLOCK|CM_INLINE) ) { if ( nodeIsFORM(node) ) BadForm( doc ); TY_(ReportError)(doc, rowgroup, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } if ( nodeIsTR(node) || nodeIsTD(node) || nodeIsTH(node) ) { TY_(ReportError)(doc, rowgroup, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } for ( parent = rowgroup->parent; parent != NULL; parent = parent->parent ) { if (node->tag == parent->tag) { TY_(UngetToken)( doc ); return; } } } /* if THEAD, TFOOT or TBODY then implied end tag */ if (node->tag->model & CM_ROWGRP) { if (node->type != EndTag) { TY_(UngetToken)( doc ); return; } } if (node->type == EndTag) { TY_(ReportError)(doc, rowgroup, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } if ( !nodeIsTR(node) ) { node = TY_(InferredTag)(doc, TidyTag_TR); TY_(ReportError)(doc, rowgroup, node, MISSING_STARTTAG); TY_(UngetToken)( doc ); } /* node should be */ TY_(InsertNodeAtEnd)(rowgroup, node); ParseTag(doc, node, IgnoreWhitespace); } } void TY_(ParseColGroup)(TidyDocImpl* doc, Node *colgroup, GetTokenMode ARG_UNUSED(mode)) { Node *node, *parent; if (colgroup->tag->model & CM_EMPTY) return; while ((node = TY_(GetToken)(doc, IgnoreWhitespace)) != NULL) { if (node->tag == colgroup->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); colgroup->closed = yes; return; } /* if this is the end tag for an ancestor element then infer end tag for this element */ if (node->type == EndTag) { if ( nodeIsFORM(node) ) { BadForm( doc ); TY_(ReportError)(doc, colgroup, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } for ( parent = colgroup->parent; parent != NULL; parent = parent->parent ) { if (node->tag == parent->tag) { TY_(UngetToken)( doc ); return; } } } if (TY_(nodeIsText)(node)) { TY_(UngetToken)( doc ); return; } /* deal with comments etc. */ if (InsertMisc(colgroup, node)) continue; /* discard unknown tags */ if (node->tag == NULL) { TY_(ReportError)(doc, colgroup, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } if ( !nodeIsCOL(node) ) { TY_(UngetToken)( doc ); return; } if (node->type == EndTag) { TY_(ReportError)(doc, colgroup, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* node should be */ TY_(InsertNodeAtEnd)(colgroup, node); ParseTag(doc, node, IgnoreWhitespace); } } void TY_(ParseTableTag)(TidyDocImpl* doc, Node *table, GetTokenMode ARG_UNUSED(mode)) { #if !defined(NDEBUG) && defined(_MSC_VER) static int in_parse_table = 0; #endif Lexer* lexer = doc->lexer; Node *node, *parent; uint istackbase; TY_(DeferDup)( doc ); istackbase = lexer->istackbase; lexer->istackbase = lexer->istacksize; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_table++; SPRTF("Entering ParseTableTag %d...\n",in_parse_table); #endif while ((node = TY_(GetToken)(doc, IgnoreWhitespace)) != NULL) { if (node->tag == table->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); lexer->istackbase = istackbase; table->closed = yes; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_table--; SPRTF("Exit ParseTableTag 1 %d... EndTag\n",in_parse_table); #endif return; } /* deal with comments etc. */ if (InsertMisc(table, node)) continue; /* discard unknown tags */ if (node->tag == NULL && node->type != TextNode) { TY_(ReportError)(doc, table, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* if TD or TH or text or inline or block then infer */ if (node->type != EndTag) { if ( nodeIsTD(node) || nodeIsTH(node) || nodeIsTABLE(node) ) { TY_(UngetToken)( doc ); node = TY_(InferredTag)(doc, TidyTag_TR); TY_(ReportError)(doc, table, node, MISSING_STARTTAG); } else if ( TY_(nodeIsText)(node) ||TY_(nodeHasCM)(node,CM_BLOCK|CM_INLINE) ) { TY_(InsertNodeBeforeElement)(table, node); TY_(ReportError)(doc, table, node, TAG_NOT_ALLOWED_IN); lexer->exiled = yes; if (node->type != TextNode) ParseTag(doc, node, IgnoreWhitespace); lexer->exiled = no; continue; } else if (node->tag->model & CM_HEAD) { MoveToHead(doc, table, node); continue; } } /* if this is the end tag for an ancestor element then infer end tag for this element */ if (node->type == EndTag) { if ( nodeIsFORM(node) ) { BadForm( doc ); TY_(ReportError)(doc, table, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* best to discard unexpected block/inline end tags */ if ( TY_(nodeHasCM)(node, CM_TABLE|CM_ROW) || TY_(nodeHasCM)(node, CM_BLOCK|CM_INLINE) ) { TY_(ReportError)(doc, table, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } for ( parent = table->parent; parent != NULL; parent = parent->parent ) { if (node->tag == parent->tag) { TY_(ReportError)(doc, table, node, MISSING_ENDTAG_BEFORE ); TY_(UngetToken)( doc ); lexer->istackbase = istackbase; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_table--; SPRTF("Exit ParseTableTag 2 %d... missing EndTag\n",in_parse_table); #endif return; } } } if (!(node->tag->model & CM_TABLE)) { TY_(UngetToken)( doc ); TY_(ReportError)(doc, table, node, TAG_NOT_ALLOWED_IN); lexer->istackbase = istackbase; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_table--; SPRTF("Exit ParseTableTag 3 %d... CM_TABLE\n",in_parse_table); #endif return; } if (TY_(nodeIsElement)(node)) { TY_(InsertNodeAtEnd)(table, node); ParseTag(doc, node, IgnoreWhitespace); continue; } /* discard unexpected text nodes and end tags */ TY_(ReportError)(doc, table, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); } TY_(ReportError)(doc, table, node, MISSING_ENDTAG_FOR); lexer->istackbase = istackbase; #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_table--; SPRTF("Exit ParseTableTag 4 %d... missing end\n",in_parse_table); #endif } /* acceptable content for pre elements */ static Bool PreContent( TidyDocImpl* ARG_UNUSED(doc), Node* node ) { /* p is coerced to br's, Text OK too */ if ( nodeIsP(node) || TY_(nodeIsText)(node) ) return yes; if ( node->tag == NULL || nodeIsPARAM(node) || !TY_(nodeHasCM)(node, CM_INLINE|CM_NEW) ) return no; return yes; } void TY_(ParsePre)( TidyDocImpl* doc, Node *pre, GetTokenMode ARG_UNUSED(mode) ) { Node *node; if (pre->tag->model & CM_EMPTY) return; TY_(InlineDup)( doc, NULL ); /* tell lexer to insert inlines if needed */ while ((node = TY_(GetToken)(doc, Preformatted)) != NULL) { if ( node->type == EndTag && (node->tag == pre->tag || DescendantOf(pre, TagId(node))) ) { if (nodeIsBODY(node) || nodeIsHTML(node)) { TY_(ReportError)(doc, pre, node, DISCARDING_UNEXPECTED); TY_(FreeNode)(doc, node); continue; } if (node->tag == pre->tag) { TY_(FreeNode)(doc, node); } else { TY_(ReportError)(doc, pre, node, MISSING_ENDTAG_BEFORE ); TY_(UngetToken)( doc ); } pre->closed = yes; TrimSpaces(doc, pre); return; } if (TY_(nodeIsText)(node)) { TY_(InsertNodeAtEnd)(pre, node); continue; } /* deal with comments etc. */ if (InsertMisc(pre, node)) continue; if (node->tag == NULL) { TY_(ReportError)(doc, pre, node, DISCARDING_UNEXPECTED); TY_(FreeNode)(doc, node); continue; } /* strip unexpected tags */ if ( !PreContent(doc, node) ) { Node *newnode; /* fix for http://tidy.sf.net/bug/772205 */ if (node->type == EndTag) { /* http://tidy.sf.net/issue/1590220 */ if ( doc->lexer->exiled && (TY_(nodeHasCM)(node, CM_TABLE) || nodeIsTABLE(node)) ) { TY_(UngetToken)(doc); TrimSpaces(doc, pre); return; } TY_(ReportError)(doc, pre, node, DISCARDING_UNEXPECTED); TY_(FreeNode)(doc, node); continue; } /* http://tidy.sf.net/issue/1590220 */ else if (TY_(nodeHasCM)(node, CM_TABLE|CM_ROW) || nodeIsTABLE(node) ) { if (!doc->lexer->exiled) /* No missing close warning if exiled. */ TY_(ReportError)(doc, pre, node, MISSING_ENDTAG_BEFORE); TY_(UngetToken)(doc); return; } /* This is basically what Tidy 04 August 2000 did and far more accurate with respect to browser behaivour than the code commented out above. Tidy could try to propagate the
     into each disallowed child where
                  
     is allowed in order to replicate some browsers behaivour, but
                  there are a lot of exceptions, e.g. Internet Explorer does not propagate
                  
     into table cells while Mozilla does. Opera 6 never propagates
                  
     into blocklevel elements while Opera 7 behaves much like Mozilla.
    
                  Tidy behaves thus mostly like Opera 6 except for nested 
     elements
                  which are handled like Mozilla takes them (Opera6 closes all 
     after
                  the first 
    ). There are similar issues like replacing

    in

     with 
    , for example
    ...

    ...

    (Input)
    ...
    ...
    (Tidy)
    ...
    ...
    (Opera 7 and Internet Explorer)
    ...

    ...
    (Opera 6 and Mozilla)
    ...

    ...

    ...
    (Input)
    ...
    ......
    (Tidy, BUG!)
    ...
    ...
    ...
    (Internet Explorer)
    ...

    ...

    ...
    (Mozilla, Opera 6)
    ...
    ...

    ...
    (Opera 7) or something similar, they could also be closing the
     and propagate
                  the 
     into the newly opened 

    . Todo: IMG, OBJECT, APPLET, BIG, SMALL, SUB, SUP, FONT, and BASEFONT are dissallowed in

    , Tidy neither detects this nor does it perform any
                  cleanup operation. Tidy should at least issue a warning if it encounters
                  such constructs.
    
                  Todo: discarding 

    is abviously a bug, it should be replaced by
    . */ TY_(InsertNodeAfterElement)(pre, node); TY_(ReportError)(doc, pre, node, MISSING_ENDTAG_BEFORE); ParseTag(doc, node, IgnoreWhitespace); newnode = TY_(InferredTag)(doc, TidyTag_PRE); TY_(ReportError)(doc, pre, newnode, INSERTING_TAG); pre = newnode; TY_(InsertNodeAfterElement)(node, pre); continue; } if ( nodeIsP(node) ) { if (node->type == StartTag) { TY_(ReportError)(doc, pre, node, USING_BR_INPLACE_OF); /* trim white space before

    in

    */
                    TrimSpaces(doc, pre);
                
                    /* coerce both 

    and

    to
    */ TY_(CoerceNode)(doc, node, TidyTag_BR, no, no); TY_(FreeAttrs)( doc, node ); /* discard align attribute etc. */ TY_(InsertNodeAtEnd)( pre, node ); } else { TY_(ReportError)(doc, pre, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); } continue; } if ( TY_(nodeIsElement)(node) ) { /* trim white space before
    */ if ( nodeIsBR(node) ) TrimSpaces(doc, pre); TY_(InsertNodeAtEnd)(pre, node); ParseTag(doc, node, Preformatted); continue; } /* discard unexpected tags */ TY_(ReportError)(doc, pre, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); } TY_(ReportError)(doc, pre, node, MISSING_ENDTAG_FOR); } void TY_(ParseOptGroup)(TidyDocImpl* doc, Node *field, GetTokenMode ARG_UNUSED(mode)) { Lexer* lexer = doc->lexer; Node *node; lexer->insert = NULL; /* defer implicit inline start tags */ while ((node = TY_(GetToken)(doc, IgnoreWhitespace)) != NULL) { if (node->tag == field->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); field->closed = yes; TrimSpaces(doc, field); return; } /* deal with comments etc. */ if (InsertMisc(field, node)) continue; if ( node->type == StartTag && (nodeIsOPTION(node) || nodeIsOPTGROUP(node)) ) { if ( nodeIsOPTGROUP(node) ) TY_(ReportError)(doc, field, node, CANT_BE_NESTED); TY_(InsertNodeAtEnd)(field, node); ParseTag(doc, node, MixedContent); continue; } /* discard unexpected tags */ TY_(ReportError)(doc, field, node, DISCARDING_UNEXPECTED ); TY_(FreeNode)( doc, node); } } void TY_(ParseSelect)(TidyDocImpl* doc, Node *field, GetTokenMode ARG_UNUSED(mode)) { #if !defined(NDEBUG) && defined(_MSC_VER) static int in_parse_select = 0; #endif Lexer* lexer = doc->lexer; Node *node; lexer->insert = NULL; /* defer implicit inline start tags */ #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_select++; SPRTF("Entering ParseSelect %d...\n",in_parse_select); #endif while ((node = TY_(GetToken)(doc, IgnoreWhitespace)) != NULL) { if (node->tag == field->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); field->closed = yes; TrimSpaces(doc, field); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_select--; SPRTF("Exit ParseSelect 1 %d...\n",in_parse_select); #endif return; } /* deal with comments etc. */ if (InsertMisc(field, node)) continue; if ( node->type == StartTag && ( nodeIsOPTION(node) || nodeIsOPTGROUP(node) || nodeIsDATALIST(node) || nodeIsSCRIPT(node)) ) { TY_(InsertNodeAtEnd)(field, node); ParseTag(doc, node, IgnoreWhitespace); continue; } /* discard unexpected tags */ TY_(ReportError)(doc, field, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); } TY_(ReportError)(doc, field, node, MISSING_ENDTAG_FOR); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_select--; SPRTF("Exit ParseSelect 2 %d...\n",in_parse_select); #endif } /* HTML5 */ void TY_(ParseDatalist)(TidyDocImpl* doc, Node *field, GetTokenMode ARG_UNUSED(mode)) { #if !defined(NDEBUG) && defined(_MSC_VER) static int in_parse_datalist = 0; #endif Lexer* lexer = doc->lexer; Node *node; lexer->insert = NULL; /* defer implicit inline start tags */ #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_datalist++; SPRTF("Entering ParseDatalist %d...\n",in_parse_datalist); #endif while ((node = TY_(GetToken)(doc, IgnoreWhitespace)) != NULL) { if (node->tag == field->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); field->closed = yes; TrimSpaces(doc, field); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_datalist--; SPRTF("Exit ParseDatalist 1 %d...\n",in_parse_datalist); #endif return; } /* deal with comments etc. */ if (InsertMisc(field, node)) continue; if ( node->type == StartTag && ( nodeIsOPTION(node) || nodeIsOPTGROUP(node) || nodeIsDATALIST(node) || nodeIsSCRIPT(node)) ) { TY_(InsertNodeAtEnd)(field, node); ParseTag(doc, node, IgnoreWhitespace); continue; } /* discard unexpected tags */ TY_(ReportError)(doc, field, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); } TY_(ReportError)(doc, field, node, MISSING_ENDTAG_FOR); #if !defined(NDEBUG) && defined(_MSC_VER) in_parse_datalist--; SPRTF("Exit ParseDatalist 2 %d...\n",in_parse_datalist); #endif } void TY_(ParseText)(TidyDocImpl* doc, Node *field, GetTokenMode mode) { Lexer* lexer = doc->lexer; Node *node; lexer->insert = NULL; /* defer implicit inline start tags */ if ( nodeIsTEXTAREA(field) ) mode = Preformatted; else mode = MixedContent; /* kludge for font tags */ while ((node = TY_(GetToken)(doc, mode)) != NULL) { if (node->tag == field->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); field->closed = yes; TrimSpaces(doc, field); return; } /* deal with comments etc. */ if (InsertMisc(field, node)) continue; if (TY_(nodeIsText)(node)) { /* only called for 1st child */ if (field->content == NULL && !(mode & Preformatted)) TrimSpaces(doc, field); if (node->start >= node->end) { TY_(FreeNode)( doc, node); continue; } TY_(InsertNodeAtEnd)(field, node); continue; } /* for textarea should all cases of < and & be escaped? */ /* discard inline tags e.g. font */ if ( node->tag && node->tag->model & CM_INLINE && !(node->tag->model & CM_FIELD)) /* #487283 - fix by Lee Passey 25 Jan 02 */ { TY_(ReportError)(doc, field, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* terminate element on other tags */ if (!(field->tag->model & CM_OPT)) TY_(ReportError)(doc, field, node, MISSING_ENDTAG_BEFORE); TY_(UngetToken)( doc ); TrimSpaces(doc, field); return; } if (!(field->tag->model & CM_OPT)) TY_(ReportError)(doc, field, node, MISSING_ENDTAG_FOR); } void TY_(ParseTitle)(TidyDocImpl* doc, Node *title, GetTokenMode ARG_UNUSED(mode)) { Node *node; while ((node = TY_(GetToken)(doc, MixedContent)) != NULL) { if (node->tag == title->tag && node->type == StartTag && cfgBool(doc, TidyCoerceEndTags) ) { TY_(ReportError)(doc, title, node, COERCE_TO_ENDTAG); node->type = EndTag; TY_(UngetToken)( doc ); continue; } else if (node->tag == title->tag && node->type == EndTag) { TY_(FreeNode)( doc, node); title->closed = yes; TrimSpaces(doc, title); return; } if (TY_(nodeIsText)(node)) { /* only called for 1st child */ if (title->content == NULL) TrimInitialSpace(doc, title, node); if (node->start >= node->end) { TY_(FreeNode)( doc, node); continue; } TY_(InsertNodeAtEnd)(title, node); continue; } /* deal with comments etc. */ if (InsertMisc(title, node)) continue; /* discard unknown tags */ if (node->tag == NULL) { TY_(ReportError)(doc, title, node, DISCARDING_UNEXPECTED); TY_(FreeNode)( doc, node); continue; } /* pushback unexpected tokens */ TY_(ReportError)(doc, title, node, MISSING_ENDTAG_BEFORE); TY_(UngetToken)( doc ); TrimSpaces(doc, title); return; } TY_(ReportError)(doc, title, node, MISSING_ENDTAG_FOR); } /* This isn't quite right for CDATA content as it recognises tags within the content and parses them accordingly. This will unfortunately screw up scripts which include < + letter, < + !, < + ? or < + / + letter */ void TY_(ParseScript)(TidyDocImpl* doc, Node *script, GetTokenMode ARG_UNUSED(mode)) { Node *node; doc->lexer->parent = script; node = TY_(GetToken)(doc, CdataContent); doc->lexer->parent = NULL; if (node) { TY_(InsertNodeAtEnd)(script, node); } else { /* handle e.g. a document like "