bool Package::Unpack() { PackagePtr sharedMe = shared_from_this(); // very basic sanity check xmlNodePtr root = xmlDocGetRootElement(_opf); string rootName(reinterpret_cast<const char*>(root->name)); rootName.tolower(); if ( rootName != "package" ) { HandleError(EPUBError::OPFInvalidPackageDocument); return false; // not an OPF file, innit? } if ( _getProp(root, "version").empty() ) { HandleError(EPUBError::OPFPackageHasNoVersion); } InstallPrefixesFromAttributeValue(_getProp(root, "prefix", ePub3NamespaceURI)); // go through children to determine the CFI index of the <spine> tag static const xmlChar* kSpineName = BAD_CAST "spine"; static const xmlChar* kManifestName = BAD_CAST "manifest"; static const xmlChar* kMetadataName = BAD_CAST "metadata"; _spineCFIIndex = 0; uint32_t idx = 0; xmlNodePtr child = xmlFirstElementChild(root); while ( child != nullptr ) { idx += 2; if ( xmlStrEqual(child->name, kSpineName) ) { _spineCFIIndex = idx; if ( _spineCFIIndex != 6 ) HandleError(EPUBError::OPFSpineOutOfOrder); } else if ( xmlStrEqual(child->name, kManifestName) && idx != 4 ) { HandleError(EPUBError::OPFManifestOutOfOrder); } else if ( xmlStrEqual(child->name, kMetadataName) && idx != 2 ) { HandleError(EPUBError::OPFMetadataOutOfOrder); } child = xmlNextElementSibling(child); } if ( _spineCFIIndex == 0 ) { HandleError(EPUBError::OPFNoSpine); return false; // spineless! } #if EPUB_COMPILER_SUPPORTS(CXX_INITIALIZER_LISTS) XPathWrangler xpath(_opf, {{"opf", OPFNamespace}, {"dc", DCNamespace}}); #else XPathWrangler::NamespaceList __m; __m["opf"] = OPFNamespace; __m["dc"] = DCNamespace; XPathWrangler xpath(_opf, __m); #endif // simple things: manifest and spine items xmlNodeSetPtr manifestNodes = nullptr; xmlNodeSetPtr spineNodes = nullptr; try { manifestNodes = xpath.Nodes("/opf:package/opf:manifest/opf:item"); spineNodes = xpath.Nodes("/opf:package/opf:spine/opf:itemref"); if ( manifestNodes == nullptr ) { HandleError(EPUBError::OPFNoManifestItems); } if ( spineNodes == nullptr ) { HandleError(EPUBError::OPFNoSpineItems); } for ( int i = 0; i < manifestNodes->nodeNr; i++ ) { auto p = std::make_shared<ManifestItem>(sharedMe); if ( p->ParseXML(p, manifestNodes->nodeTab[i]) ) { #if EPUB_HAVE(CXX_MAP_EMPLACE) _manifest.emplace(p->Identifier(), p); #else _manifest[p->Identifier()] = p; #endif StoreXMLIdentifiable(p); } else { // TODO: Need an error here } } // check fallback chains typedef std::map<string, bool> IdentSet; IdentSet idents; for ( auto &pair : _manifest ) { ManifestItemPtr item = pair.second; if ( item->FallbackID().empty() ) continue; idents[item->XMLIdentifier()] = true; while ( !item->FallbackID().empty() ) { if ( idents[item->FallbackID()] ) { HandleError(EPUBError::OPFFallbackChainCircularReference); break; } item = item->Fallback(); } idents.clear(); } SpineItemPtr cur; for ( int i = 0; i < spineNodes->nodeNr; i++ ) { auto next = std::make_shared<SpineItem>(sharedMe); if ( next->ParseXML(next, spineNodes->nodeTab[i]) == false ) { // TODO: need an error code here continue; } // validation of idref auto manifestFound = _manifest.find(next->Idref()); if ( manifestFound == _manifest.end() ) { HandleError(EPUBError::OPFInvalidSpineIdref, _Str(next->Idref(), " does not correspond to a manifest item")); continue; } // validation of spine resource type w/fallbacks ManifestItemPtr manifestItem = next->ManifestItem(); bool isContentDoc = false; do { if ( manifestItem->MediaType() == "application/xhtml+xml" || manifestItem->MediaType() == "image/svg" ) { isContentDoc = true; break; } } while ( (manifestItem = manifestItem->Fallback()) ); if ( !isContentDoc ) HandleError(EPUBError::OPFFallbackChainHasNoContentDocument); StoreXMLIdentifiable(next); if ( cur != nullptr ) { cur->SetNextItem(next); } else { _spine = next; } cur = next; } } catch (const std::system_error& exc) { if ( manifestNodes != nullptr ) xmlXPathFreeNodeSet(manifestNodes); if ( spineNodes != nullptr ) xmlXPathFreeNodeSet(spineNodes); if ( exc.code().category() == epub_spec_category() ) throw; return false; } catch (...) { if ( manifestNodes != nullptr ) xmlXPathFreeNodeSet(manifestNodes); if ( spineNodes != nullptr ) xmlXPathFreeNodeSet(spineNodes); return false; } xmlXPathFreeNodeSet(manifestNodes); xmlXPathFreeNodeSet(spineNodes); // now the metadata, which is slightly more involved due to extensions xmlNodeSetPtr metadataNodes = nullptr; xmlNodeSetPtr refineNodes = xmlXPathNodeSetCreate(nullptr); try { shared_ptr<PropertyHolder> holderPtr = std::dynamic_pointer_cast<PropertyHolder>(sharedMe); metadataNodes = xpath.Nodes("/opf:package/opf:metadata/*"); if ( metadataNodes == nullptr ) HandleError(EPUBError::OPFNoMetadata); bool foundIdentifier = false, foundTitle = false, foundLanguage = false, foundModDate = false; string uniqueIDRef = _getProp(root, "unique-identifier"); if ( uniqueIDRef.empty() ) HandleError(EPUBError::OPFPackageUniqueIDInvalid); for ( int i = 0; i < metadataNodes->nodeNr; i++ ) { xmlNodePtr node = metadataNodes->nodeTab[i]; PropertyPtr p; if ( node->ns != nullptr && xmlStrcmp(node->ns->href, BAD_CAST DCNamespace) == 0 ) { // definitely a main node p = std::make_shared<Property>(holderPtr); } else if ( _getProp(node, "name").size() > 0 ) { // it's an ePub2 item-- ignore it continue; } else if ( _getProp(node, "refines").empty() ) { // not refining anything, so it's a main node p = std::make_shared<Property>(holderPtr); } else { // by elimination it's refining something-- we'll process it later when we know we've got all the main nodes in there xmlXPathNodeSetAdd(refineNodes, node); } if ( p && p->ParseMetaElement(node) ) { switch ( p->Type() ) { case DCType::Identifier: { foundIdentifier = true; if ( !uniqueIDRef.empty() && uniqueIDRef != p->XMLIdentifier() ) HandleError(EPUBError::OPFPackageUniqueIDInvalid); break; } case DCType::Title: { foundTitle = true; break; } case DCType::Language: { foundLanguage = true; break; } case DCType::Custom: { if ( p->PropertyIdentifier() == MakePropertyIRI("modified", "dcterms") ) foundModDate = true; break; } default: break; } AddProperty(p); StoreXMLIdentifiable(p); } } if ( !foundIdentifier ) HandleError(EPUBError::OPFMissingIdentifierMetadata); if ( !foundTitle ) HandleError(EPUBError::OPFMissingTitleMetadata); if ( !foundLanguage ) HandleError(EPUBError::OPFMissingLanguageMetadata); if ( !foundModDate ) HandleError(EPUBError::OPFMissingModificationDateMetadata); for ( int i = 0; i < refineNodes->nodeNr; i++ ) { xmlNodePtr node = refineNodes->nodeTab[i]; string ident = _getProp(node, "refines"); if ( ident.empty() ) { HandleError(EPUBError::OPFInvalidRefinementAttribute, "Empty IRI for 'refines' attribute"); continue; } if ( ident[0] == '#' ) { ident = ident.substr(1); } else { // validation only right now IRI iri(ident); if ( iri.IsEmpty() ) { HandleError(EPUBError::OPFInvalidRefinementAttribute, _Str("#", ident, " is not a valid IRI")); } else if ( iri.IsRelative() == false ) { HandleError(EPUBError::OPFInvalidRefinementAttribute, _Str(iri.IRIString(), " is not a relative IRI")); } continue; } auto found = _xmlIDLookup.find(ident); if ( found == _xmlIDLookup.end() ) { HandleError(EPUBError::OPFInvalidRefinementTarget, _Str("#", ident, " does not reference an item in this document")); continue; } PropertyPtr prop = std::dynamic_pointer_cast<Property>(found->second); if ( prop ) { // it's a property, so this is an extension PropertyExtensionPtr extPtr = std::make_shared<PropertyExtension>(prop); if ( extPtr->ParseMetaElement(node) ) prop->AddExtension(extPtr); } else { // not a property, so treat this as a plain property shared_ptr<PropertyHolder> ptr = std::dynamic_pointer_cast<PropertyHolder>(found->second); if ( ptr ) { prop = std::make_shared<Property>(ptr); if ( prop->ParseMetaElement(node) ) ptr->AddProperty(prop); } } } // now look at the <spine> element for properties xmlNodePtr spineNode = xmlFirstElementChild(root); for ( uint32_t i = 2; i < _spineCFIIndex; i += 2 ) spineNode = xmlNextElementSibling(spineNode); string value = _getProp(spineNode, "page-progression-direction"); if ( !value.empty() ) { PropertyPtr prop = std::make_shared<Property>(holderPtr); prop->SetPropertyIdentifier(MakePropertyIRI("page-progression-direction")); prop->SetValue(value); AddProperty(prop); } } catch (std::system_error& exc) { if ( metadataNodes != nullptr ) xmlXPathFreeNodeSet(metadataNodes); if ( refineNodes != nullptr ) xmlXPathFreeNodeSet(refineNodes); if ( exc.code().category() == epub_spec_category() ) throw; return false; } catch (...) { if ( metadataNodes != nullptr ) xmlXPathFreeNodeSet(metadataNodes); if ( refineNodes != nullptr ) xmlXPathFreeNodeSet(refineNodes); return false; } xmlXPathFreeNodeSet(metadataNodes); xmlXPathFreeNodeSet(refineNodes); // now any content type bindings xmlNodeSetPtr bindingNodes = nullptr; try { bindingNodes = xpath.Nodes("/opf:package/opf:bindings/*"); if ( bindingNodes != nullptr ) { for ( int i = 0; i < bindingNodes->nodeNr; i++ ) { xmlNodePtr node = bindingNodes->nodeTab[i]; if ( xmlStrcasecmp(node->name, MediaTypeElementName) != 0 ) continue; //////////////////////////////////////////////////////////// // ePub Publications 3.0 §3.4.16: The `mediaType` Element // The media-type attribute is required. string mediaType = _getProp(node, "media-type"); if ( mediaType.empty() ) { HandleError(EPUBError::OPFBindingHandlerNoMediaType); throw false; } // Each child mediaType of a bindings element must define a unique // content type in its media-type attribute, and the media type // specified must not be a Core Media Type. if ( _contentHandlers[mediaType].empty() == false ) { // user shouldn't have added manual things yet, but for safety we'll look anyway for ( auto ptr : _contentHandlers[mediaType] ) { if ( typeid(*ptr) == typeid(MediaHandler) ) { HandleError(EPUBError::OPFMultipleBindingsForMediaType); } } } if ( CoreMediaTypes.find(mediaType) != CoreMediaTypes.end() ) { HandleError(EPUBError::OPFCoreMediaTypeBindingEncountered); } // The handler attribute is required string handlerID = _getProp(node, "handler"); if ( handlerID.empty() ) { HandleError(EPUBError::OPFBindingHandlerNotFound); } // The required handler attribute must reference the ID [XML] of an // item in the manifest of the default implementation for this media // type. The referenced item must be an XHTML Content Document. ManifestItemPtr handlerItem = ManifestItemWithID(handlerID); if ( !handlerItem ) { HandleError(EPUBError::OPFBindingHandlerNotFound); } if ( handlerItem->MediaType() != "application/xhtml+xml" ) { HandleError(EPUBError::OPFBindingHandlerInvalidType, _Str("Media handlers must be XHTML content documents, but referenced item has type '", handlerItem->MediaType(), "'.")); } // All XHTML Content Documents designated as handlers must have the // `scripted` property set in their manifest item's `properties` // attribute. if ( handlerItem->HasProperty(ItemProperties::HasScriptedContent) == false ) { HandleError(EPUBError::OPFBindingHandlerNotScripted); } // all good-- install it now _contentHandlers[mediaType].push_back(std::make_shared<MediaHandler>(sharedMe, mediaType, handlerItem->AbsolutePath())); } } } catch (std::exception& exc) { std::cerr << "Exception processing OPF file: " << exc.what() << std::endl; if ( bindingNodes != nullptr ) xmlXPathFreeNodeSet(bindingNodes); throw; } catch (...) { if ( bindingNodes != nullptr ) xmlXPathFreeNodeSet(bindingNodes); return false; } xmlXPathFreeNodeSet(bindingNodes); // now the navigation tables for ( auto item : _manifest ) { if ( !item.second->HasProperty(ItemProperties::Navigation) ) continue; NavigationList tables = NavTablesFromManifestItem(sharedMe, item.second); for ( auto table : tables ) { // have to dynamic_cast these guys to get the right pointer type shared_ptr<class NavigationTable> navTable = std::dynamic_pointer_cast<class NavigationTable>(table); #if EPUB_HAVE(CXX_MAP_EMPLACE) _navigation.emplace(navTable->Type(), navTable); #else _navigation[navTable->Type()] = navTable; #endif } } // lastly, let's set the media support information InitMediaSupport(); return true; }
/* * call-seq: * evaluate(search_path) * * Evaluate the +search_path+ returning an XML::XPath object. */ static VALUE evaluate(int argc, VALUE *argv, VALUE self) { VALUE search_path, xpath_handler; VALUE thing = Qnil; xmlXPathContextPtr ctx; xmlXPathObjectPtr xpath; xmlChar *query; Data_Get_Struct(self, xmlXPathContext, ctx); if(rb_scan_args(argc, argv, "11", &search_path, &xpath_handler) == 1) xpath_handler = Qnil; query = (xmlChar *)StringValuePtr(search_path); if(Qnil != xpath_handler) { /* FIXME: not sure if this is the correct place to shove private data. */ ctx->userData = (void *)xpath_handler; xmlXPathRegisterFuncLookup(ctx, lookup, (void *)xpath_handler); } xmlResetLastError(); xmlSetStructuredErrorFunc(NULL, xpath_exception_handler); /* For some reason, xmlXPathEvalExpression will blow up with a generic error */ /* when there is a non existent function. */ xmlSetGenericErrorFunc(NULL, xpath_generic_exception_handler); xpath = xmlXPathEvalExpression(query, ctx); xmlSetStructuredErrorFunc(NULL, NULL); xmlSetGenericErrorFunc(NULL, NULL); if(xpath == NULL) { VALUE xpath = rb_const_get(mNokogiriXml, rb_intern("XPath")); VALUE klass = rb_const_get(xpath, rb_intern("SyntaxError")); xmlErrorPtr error = xmlGetLastError(); rb_exc_raise(Nokogiri_wrap_xml_syntax_error(klass, error)); } assert(ctx->doc); assert(DOC_RUBY_OBJECT_TEST(ctx->doc)); switch(xpath->type) { case XPATH_STRING: thing = NOKOGIRI_STR_NEW2(xpath->stringval); break; case XPATH_NODESET: if(NULL == xpath->nodesetval) { thing = Nokogiri_wrap_xml_node_set(xmlXPathNodeSetCreate(NULL), DOC_RUBY_OBJECT(ctx->doc)); } else { thing = Nokogiri_wrap_xml_node_set(xpath->nodesetval, DOC_RUBY_OBJECT(ctx->doc)); } break; case XPATH_NUMBER: thing = rb_float_new(xpath->floatval); break; case XPATH_BOOLEAN: thing = xpath->boolval == 1 ? Qtrue : Qfalse; break; default: thing = Nokogiri_wrap_xml_node_set(xmlXPathNodeSetCreate(NULL), DOC_RUBY_OBJECT(ctx->doc)); } return thing; }
/** * Loop through array of operations and perform them */ static void edProcess(xmlDocPtr doc, const XmlEdAction* ops, int ops_count) { int k; xmlXPathContextPtr ctxt = xmlXPathNewContext(doc); /* NOTE: later registrations override earlier ones */ registerXstarNs(ctxt); /* variables */ previous_insertion = xmlXPathNodeSetCreate(NULL); registerXstarVariable(ctxt, "prev", xmlXPathWrapNodeSet(previous_insertion)); xmlDeregisterNodeDefault(&removeNodeFromPrev); #if HAVE_EXSLT_XPATH_REGISTER /* register extension functions */ exsltDateXpathCtxtRegister(ctxt, BAD_CAST "date"); exsltMathXpathCtxtRegister(ctxt, BAD_CAST "math"); exsltSetsXpathCtxtRegister(ctxt, BAD_CAST "set"); exsltStrXpathCtxtRegister(ctxt, BAD_CAST "str"); #endif /* namespaces from doc */ extract_ns_defs(doc, ctxt); /* namespaces from command line */ nsarr_xpath_register(ctxt); for (k = 0; k < ops_count; k++) { xmlXPathObjectPtr res; xmlNodeSetPtr nodes; /* NOTE: to make relative paths match as if from "/", set context to document; setting to root would match as if from "/node()/" */ ctxt->node = (xmlNodePtr) doc; if (ops[k].op == XML_ED_VAR) { res = xmlXPathEvalExpression(BAD_CAST ops[k].arg2, ctxt); xmlXPathRegisterVariable(ctxt, BAD_CAST ops[k].arg1, res); continue; } res = xmlXPathEvalExpression(BAD_CAST ops[k].arg1, ctxt); if (!res || res->type != XPATH_NODESET || !res->nodesetval) continue; nodes = res->nodesetval; switch (ops[k].op) { case XML_ED_DELETE: edDelete(doc, nodes); break; case XML_ED_MOVE: { xmlXPathObjectPtr res_to; ctxt->node = (xmlNodePtr) doc; res_to = xmlXPathEvalExpression(BAD_CAST ops[k].arg2, ctxt); if (!res_to || res_to->type != XPATH_NODESET || res_to->nodesetval->nodeNr != 1) { fprintf(stderr, "move destination is not a single node\n"); continue; } edMove(doc, nodes, res_to->nodesetval->nodeTab[0]); xmlXPathFreeObject(res_to); break; } case XML_ED_UPDATE: edUpdate(doc, nodes, ops[k].arg2, ops[k].type, ctxt); break; case XML_ED_RENAME: edRename(doc, nodes, ops[k].arg2, ops[k].type); break; case XML_ED_INSERT: edInsert(doc, nodes, ops[k].arg2, ops[k].arg3, ops[k].type, -1); break; case XML_ED_APPEND: edInsert(doc, nodes, ops[k].arg2, ops[k].arg3, ops[k].type, 1); break; case XML_ED_SUBNODE: edInsert(doc, nodes, ops[k].arg2, ops[k].arg3, ops[k].type, 0); break; default: break; } xmlXPathFreeObject(res); } /* NOTE: free()ing ctxt also free()s previous_insertion */ previous_insertion = NULL; xmlDeregisterNodeDefault(NULL); xmlXPathFreeContext(ctxt); }
/** * xsltInitCtxtKey: * @ctxt: an XSLT transformation context * @idoc: the document information (holds key values) * @keyDef: the key definition * * Computes the key tables this key and for the current input document. * * Returns: 0 on success, -1 on error */ int xsltInitCtxtKey(xsltTransformContextPtr ctxt, xsltDocumentPtr idoc, xsltKeyDefPtr keyDef) { int i, len, k; xmlNodeSetPtr matchList = NULL, keylist; xmlXPathObjectPtr matchRes = NULL, useRes = NULL; xmlChar *str = NULL; xsltKeyTablePtr table; xmlNodePtr oldInst, cur; xmlNodePtr oldContextNode; xsltDocumentPtr oldDocInfo; int oldXPPos, oldXPSize; xmlDocPtr oldXPDoc; int oldXPNsNr; xmlNsPtr *oldXPNamespaces; xmlXPathContextPtr xpctxt; #ifdef KEY_INIT_DEBUG fprintf(stderr, "xsltInitCtxtKey %s : %d\n", keyDef->name, ctxt->keyInitLevel); #endif if ((keyDef->comp == NULL) || (keyDef->usecomp == NULL)) return(-1); /* * Detect recursive keys */ if (ctxt->keyInitLevel > ctxt->nbKeys) { #ifdef WITH_XSLT_DEBUG_KEYS XSLT_TRACE(ctxt,XSLT_TRACE_KEYS, xsltGenericDebug(xsltGenericDebugContext, "xsltInitCtxtKey: key definition of %s is recursive\n", keyDef->name)); #endif xsltTransformError(ctxt, NULL, keyDef->inst, "Key definition for %s is recursive\n", keyDef->name); ctxt->state = XSLT_STATE_STOPPED; return(-1); } ctxt->keyInitLevel++; xpctxt = ctxt->xpathCtxt; idoc->nbKeysComputed++; /* * Save context state. */ oldInst = ctxt->inst; oldDocInfo = ctxt->document; oldContextNode = ctxt->node; oldXPDoc = xpctxt->doc; oldXPPos = xpctxt->proximityPosition; oldXPSize = xpctxt->contextSize; oldXPNsNr = xpctxt->nsNr; oldXPNamespaces = xpctxt->namespaces; /* * Set up contexts. */ ctxt->document = idoc; ctxt->node = (xmlNodePtr) idoc->doc; ctxt->inst = keyDef->inst; xpctxt->doc = idoc->doc; xpctxt->node = (xmlNodePtr) idoc->doc; /* TODO : clarify the use of namespaces in keys evaluation */ xpctxt->namespaces = keyDef->nsList; xpctxt->nsNr = keyDef->nsNr; /* * Evaluate the 'match' expression of the xsl:key. * TODO: The 'match' is a *pattern*. */ matchRes = xmlXPathCompiledEval(keyDef->comp, xpctxt); if (matchRes == NULL) { #ifdef WITH_XSLT_DEBUG_KEYS XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsltInitCtxtKey: %s evaluation failed\n", keyDef->match)); #endif xsltTransformError(ctxt, NULL, keyDef->inst, "Failed to evaluate the 'match' expression.\n"); ctxt->state = XSLT_STATE_STOPPED; goto error; } else { if (matchRes->type == XPATH_NODESET) { matchList = matchRes->nodesetval; #ifdef WITH_XSLT_DEBUG_KEYS if (matchList != NULL) XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsltInitCtxtKey: %s evaluates to %d nodes\n", keyDef->match, matchList->nodeNr)); #endif } else { /* * Is not a node set, but must be. */ #ifdef WITH_XSLT_DEBUG_KEYS XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsltInitCtxtKey: %s is not a node set\n", keyDef->match)); #endif xsltTransformError(ctxt, NULL, keyDef->inst, "The 'match' expression did not evaluate to a node set.\n"); ctxt->state = XSLT_STATE_STOPPED; goto error; } } if ((matchList == NULL) || (matchList->nodeNr <= 0)) goto exit; /** * Multiple key definitions for the same name are allowed, so * we must check if the key is already present for this doc */ table = (xsltKeyTablePtr) idoc->keys; while (table != NULL) { if (xmlStrEqual(table->name, keyDef->name) && (((keyDef->nameURI == NULL) && (table->nameURI == NULL)) || ((keyDef->nameURI != NULL) && (table->nameURI != NULL) && (xmlStrEqual(table->nameURI, keyDef->nameURI))))) break; table = table->next; } /** * If the key was not previously defined, create it now and * chain it to the list of keys for the doc */ if (table == NULL) { table = xsltNewKeyTable(keyDef->name, keyDef->nameURI); if (table == NULL) goto error; table->next = idoc->keys; idoc->keys = table; } /* * SPEC XSLT 1.0 (XSLT 2.0 does not clarify the context size!) * "...the use attribute of the xsl:key element is evaluated with x as " the current node and with a node list containing just x as the * current node list" */ xpctxt->contextSize = 1; xpctxt->proximityPosition = 1; for (i = 0; i < matchList->nodeNr; i++) { cur = matchList->nodeTab[i]; if (! IS_XSLT_REAL_NODE(cur)) continue; xpctxt->node = cur; /* * Process the 'use' of the xsl:key. * SPEC XSLT 1.0: * "The use attribute is an expression specifying the values of * the key; the expression is evaluated once for each node that * matches the pattern." */ if (useRes != NULL) xmlXPathFreeObject(useRes); useRes = xmlXPathCompiledEval(keyDef->usecomp, xpctxt); if (useRes == NULL) { xsltTransformError(ctxt, NULL, keyDef->inst, "Failed to evaluate the 'use' expression.\n"); ctxt->state = XSLT_STATE_STOPPED; break; } if (useRes->type == XPATH_NODESET) { if ((useRes->nodesetval != NULL) && (useRes->nodesetval->nodeNr != 0)) { len = useRes->nodesetval->nodeNr; str = xmlXPathCastNodeToString(useRes->nodesetval->nodeTab[0]); } else { continue; } } else { len = 1; if (useRes->type == XPATH_STRING) { /* * Consume the string value. */ str = useRes->stringval; useRes->stringval = NULL; } else { str = xmlXPathCastToString(useRes); } } /* * Process all strings. */ k = 0; while (1) { if (str == NULL) goto next_string; #ifdef WITH_XSLT_DEBUG_KEYS XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsl:key : node associated to ('%s', '%s')\n", keyDef->name, str)); #endif keylist = xmlHashLookup(table->keys, str); if (keylist == NULL) { keylist = xmlXPathNodeSetCreate(cur); if (keylist == NULL) goto error; xmlHashAddEntry(table->keys, str, keylist); } else { /* * TODO: How do we know if this function failed? */ xmlXPathNodeSetAdd(keylist, cur); } switch (cur->type) { case XML_ELEMENT_NODE: case XML_TEXT_NODE: case XML_CDATA_SECTION_NODE: case XML_PI_NODE: case XML_COMMENT_NODE: cur->psvi = keyDef; break; case XML_ATTRIBUTE_NODE: ((xmlAttrPtr) cur)->psvi = keyDef; break; case XML_DOCUMENT_NODE: case XML_HTML_DOCUMENT_NODE: ((xmlDocPtr) cur)->psvi = keyDef; break; default: break; } xmlFree(str); str = NULL; next_string: k++; if (k >= len) break; str = xmlXPathCastNodeToString(useRes->nodesetval->nodeTab[k]); } } exit: error: ctxt->keyInitLevel--; /* * Restore context state. */ xpctxt->doc = oldXPDoc; xpctxt->nsNr = oldXPNsNr; xpctxt->namespaces = oldXPNamespaces; xpctxt->proximityPosition = oldXPPos; xpctxt->contextSize = oldXPSize; ctxt->node = oldContextNode; ctxt->document = oldDocInfo; ctxt->inst = oldInst; if (str) xmlFree(str); if (useRes != NULL) xmlXPathFreeObject(useRes); if (matchRes != NULL) xmlXPathFreeObject(matchRes); return(0); }
/** * 'update' operation */ static void edUpdate(xmlDocPtr doc, xmlNodeSetPtr nodes, const char *val, XmlNodeType type, xmlXPathContextPtr ctxt) { int i; xmlXPathCompExprPtr xpath = NULL; if (type == XML_EXPR) { xpath = xmlXPathCompile((const xmlChar*) val); if (!xpath) return; } for (i = 0; i < nodes->nodeNr; i++) { /* update node */ if (type == XML_EXPR) { xmlXPathObjectPtr res; ctxt->node = nodes->nodeTab[i]; res = xmlXPathCompiledEval(xpath, ctxt); if (res->type == XPATH_NODESET || res->type == XPATH_XSLT_TREE) { int j; xmlNodePtr oldChild; xmlNodeSetPtr oldChildren = xmlXPathNodeSetCreate(NULL); /* NOTE: newChildren can be NULL for empty result set */ xmlNodeSetPtr newChildren = res->nodesetval; /* NOTE: nodes can be both oldChildren and newChildren */ /* unlink the old children */ for (oldChild = nodes->nodeTab[i]->children; oldChild; oldChild = oldChild->next) { xmlUnlinkNode(oldChild); /* we can't free it now because an oldChild can also be newChild! just put it in the list */ xmlXPathNodeSetAdd(oldChildren, oldChild); } /* add the new children */ for (j = 0; newChildren && j < newChildren->nodeNr; j++) { xmlNodePtr node = newChildren->nodeTab[j]; xmlAddChild(nodes->nodeTab[i], /* if node is linked to this doc we need to copy */ (node->doc == doc)? xmlDocCopyNode(node, doc, 1) : node); newChildren->nodeTab[j] = NULL; } newChildren->nodeNr = 0; /* NOTE: if any oldChildren were newChildren, they've been copied so we can free them all now */ for (j = 0; j < oldChildren->nodeNr; j++) { xmlFreeNode(oldChildren->nodeTab[j]); oldChildren->nodeTab[j] = NULL; } oldChildren->nodeNr = 0; xmlXPathFreeNodeSet(oldChildren); } else { res = xmlXPathConvertString(res); update_string(doc, nodes->nodeTab[i], res->stringval); } xmlXPathFreeObject(res); } else { update_string(doc, nodes->nodeTab[i], (const xmlChar*) val); } } xmlXPathFreeCompExpr(xpath); }
static void method_caller(xmlXPathParserContextPtr ctxt, int nargs) { const xmlChar * function; const xmlChar * functionURI; size_t i, count; xsltTransformContextPtr transform; xmlXPathObjectPtr xpath; VALUE obj; VALUE *args; VALUE result; transform = xsltXPathGetTransformContext(ctxt); function = ctxt->context->function; functionURI = ctxt->context->functionURI; obj = (VALUE)xsltGetExtData(transform, functionURI); count = (size_t)ctxt->valueNr; args = calloc(count, sizeof(VALUE *)); for(i = 0; i < count; i++) { VALUE thing; xpath = valuePop(ctxt); switch(xpath->type) { case XPATH_STRING: thing = NOKOGIRI_STR_NEW2(xpath->stringval); break; case XPATH_NODESET: if(NULL == xpath->nodesetval) { thing = Nokogiri_wrap_xml_node_set( xmlXPathNodeSetCreate(NULL), DOC_RUBY_OBJECT(ctxt->context->doc)); } else { thing = Nokogiri_wrap_xml_node_set(xpath->nodesetval, DOC_RUBY_OBJECT(ctxt->context->doc)); } break; default: rb_raise(rb_eRuntimeError, "do not handle type: %d", xpath->type); } args[i] = thing; } result = rb_funcall3(obj, rb_intern((const char *)function), (int)count, args); switch(TYPE(result)) { case T_FLOAT: case T_BIGNUM: case T_FIXNUM: xmlXPathReturnNumber(ctxt, NUM2DBL(result)); break; case T_STRING: xmlXPathReturnString( ctxt, xmlStrdup((xmlChar *)StringValuePtr(result)) ); break; case T_TRUE: xmlXPathReturnTrue(ctxt); break; case T_FALSE: xmlXPathReturnFalse(ctxt); break; case T_NIL: break; default: rb_raise(rb_eRuntimeError, "Invalid return type"); } }
bool Package::Unpack() { // very basic sanity check xmlNodePtr root = xmlDocGetRootElement(_opf); string rootName(reinterpret_cast<const char*>(root->name)); rootName.tolower(); if ( rootName != "package" ) return false; // not an OPF file, innit? InstallPrefixesFromAttributeValue(_getProp(root, "prefix", ePub3NamespaceURI)); // go through children to determine the CFI index of the <spine> tag static const xmlChar* kSpineName = BAD_CAST "spine"; _spineCFIIndex = 0; xmlNodePtr child = root->children; while ( child != nullptr ) { if ( child->type == XML_ELEMENT_NODE ) { _spineCFIIndex += 2; if ( xmlStrEqual(child->name, kSpineName) ) break; } child = child->next; } if ( _spineCFIIndex == 0 ) return false; // spineless! XPathWrangler xpath(_opf, {{"opf", OPFNamespace}, {"dc", DCNamespace}}); // simple things: manifest and spine items xmlNodeSetPtr manifestNodes = nullptr; xmlNodeSetPtr spineNodes = nullptr; try { manifestNodes = xpath.Nodes("/opf:package/opf:manifest/opf:item"); spineNodes = xpath.Nodes("/opf:package/opf:spine/opf:itemref"); if ( manifestNodes == nullptr || spineNodes == nullptr ) throw false; // looks invalid, or at least unusable, to me for ( int i = 0; i < manifestNodes->nodeNr; i++ ) { ManifestItem *p = new ManifestItem(manifestNodes->nodeTab[i], this); _manifest.emplace(p->Identifier(), p); } SpineItem* cur = nullptr; for ( int i = 0; i < spineNodes->nodeNr; i++ ) { SpineItem* next = new SpineItem(spineNodes->nodeTab[i], this); if ( cur != nullptr ) { cur->SetNextItem(next); } else { _spine.reset(next); } cur = next; } } catch (...) { if ( manifestNodes != nullptr ) xmlXPathFreeNodeSet(manifestNodes); if ( spineNodes != nullptr ) xmlXPathFreeNodeSet(spineNodes); return false; } xmlXPathFreeNodeSet(manifestNodes); xmlXPathFreeNodeSet(spineNodes); // now the metadata, which is slightly more involved due to extensions xmlNodeSetPtr metadataNodes = nullptr; xmlNodeSetPtr refineNodes = xmlXPathNodeSetCreate(nullptr); try { metadataNodes = xpath.Nodes("/opf:package/opf:metadata/*"); if ( metadataNodes == nullptr ) throw false; std::map<string, class Metadata*> metadataByID; for ( int i = 0; i < metadataNodes->nodeNr; i++ ) { xmlNodePtr node = metadataNodes->nodeTab[i]; class Metadata* p = nullptr; if ( node->ns != nullptr && xmlStrcmp(node->ns->href, BAD_CAST DCNamespace) == 0 ) { // definitely a main node p = new class Metadata(node, this); } else if ( _getProp(node, "name").size() > 0 ) { // it's an ePub2 item-- ignore it continue; } else if ( _getProp(node, "refines").empty() ) { // not refining anything, so it's a main node p = new class Metadata(node, this); } else { // by elimination it's refining something-- we'll process it later when we know we've got all the main nodes in there xmlXPathNodeSetAdd(refineNodes, node); } if ( p != nullptr ) { _metadata.push_back(p); if ( !p->Identifier().empty() ) metadataByID[p->Identifier()] = p; } } for ( int i = 0; i < refineNodes->nodeNr; i++ ) { xmlNodePtr node = refineNodes->nodeTab[i]; string ident = _getProp(node, "refines"); if ( ident.empty() ) continue; if ( ident[0] == '#' ) ident = ident.substr(1); auto found = metadataByID.find(ident); if ( found == metadataByID.end() ) continue; found->second->AddExtension(node, this); } } catch (...) { if ( metadataNodes != nullptr ) xmlXPathFreeNodeSet(metadataNodes); if ( refineNodes != nullptr ) xmlXPathFreeNodeSet(refineNodes); return false; } xmlXPathFreeNodeSet(metadataNodes); xmlXPathFreeNodeSet(refineNodes); // now any content type bindings xmlNodeSetPtr bindingNodes = nullptr; try { bindingNodes = xpath.Nodes("/opf:package/opf:bindings/*"); if ( bindingNodes != nullptr ) { for ( int i = 0; i < bindingNodes->nodeNr; i++ ) { xmlNodePtr node = bindingNodes->nodeTab[i]; if ( xmlStrcasecmp(node->name, MediaTypeElementName) != 0 ) continue; //////////////////////////////////////////////////////////// // ePub Publications 3.0 §3.4.16: The `mediaType` Element // The media-type attribute is required. string mediaType = _getProp(node, "media-type"); if ( mediaType.empty() ) { throw std::invalid_argument("mediaType element has missing or empty media-type attribute."); } // Each child mediaType of a bindings element must define a unique // content type in its media-type attribute, and the media type // specified must not be a Core Media Type. if ( _contentHandlers[mediaType].empty() == false ) { // user shouldn't have added manual things yet, but for safety we'll look anyway for ( auto ptr : _contentHandlers[mediaType] ) { if ( typeid(*ptr) == typeid(MediaHandler) ) { throw std::invalid_argument(_Str("Duplicate media handler found for type '", mediaType, "'.")); } } } if ( CoreMediaTypes.find(mediaType) != CoreMediaTypes.end() ) { throw std::invalid_argument("mediaType element specifies an EPUB Core Media Type."); } // The handler attribute is required string handlerID = _getProp(node, "handler"); if ( handlerID.empty() ) { throw std::invalid_argument("mediaType element has missing or empty handler attribute."); } // The required handler attribute must reference the ID [XML] of an // item in the manifest of the default implementation for this media // type. The referenced item must be an XHTML Content Document. const ManifestItem* handlerItem = ManifestItemWithID(handlerID); if ( handlerItem == nullptr ) { throw std::invalid_argument(_Str("mediaType element references non-existent handler with ID '", handlerID, "'.")); } if ( handlerItem->MediaType() != "application/xhtml+xml" ) { throw std::invalid_argument(_Str("Media handlers must be XHTML content documents, but referenced item has type '", handlerItem->MediaType(), "'.")); } // All XHTML Content Documents designated as handlers must have the // `scripted` property set in their manifest item's `properties` // attribute. if ( handlerItem->HasProperty(ItemProperties::HasScriptedContent) == false ) { throw std::invalid_argument("Media handlers must have the `scripted` property."); } // all good-- install it now _contentHandlers[mediaType].push_back(new MediaHandler(this, mediaType, handlerItem->AbsolutePath())); } } } catch (std::exception& exc) { std::cerr << "Exception processing OPF file: " << exc.what() << std::endl; if ( bindingNodes != nullptr ) xmlXPathFreeNodeSet(bindingNodes); return false; } catch (...) { if ( bindingNodes != nullptr ) xmlXPathFreeNodeSet(bindingNodes); return false; } xmlXPathFreeNodeSet(bindingNodes); // now the navigation tables for ( auto item : _manifest ) { if ( !item.second->HasProperty(ItemProperties::Navigation) ) continue; NavigationList tables = NavTablesFromManifestItem(item.second); for ( auto table : tables ) { // have to dynamic_cast these guys to get the right pointer type class NavigationTable* navTable = dynamic_cast<class NavigationTable*>(table); _navigation.emplace(navTable->Type(), navTable); } } return true; }
static xmlSecNodeSetPtr xmlSecXPathDataExecute(xmlSecXPathDataPtr data, xmlDocPtr doc, xmlNodePtr hereNode) { xmlXPathObjectPtr xpathObj = NULL; xmlSecNodeSetPtr nodes; xmlSecAssert2(data != NULL, NULL); xmlSecAssert2(data->expr != NULL, NULL); xmlSecAssert2(data->ctx != NULL, NULL); xmlSecAssert2(doc != NULL, NULL); xmlSecAssert2(hereNode != NULL, NULL); /* do not forget to set the doc */ data->ctx->doc = doc; /* here function works only on the same document */ if(hereNode->doc == doc) { xmlXPathRegisterFunc(data->ctx, (xmlChar *)"here", xmlSecXPathHereFunction); data->ctx->here = hereNode; data->ctx->xptr = 1; } /* execute xpath or xpointer expression */ switch(data->type) { case xmlSecXPathDataTypeXPath: case xmlSecXPathDataTypeXPath2: xpathObj = xmlXPathEvalExpression(data->expr, data->ctx); if(xpathObj == NULL) { xmlSecXmlError2("xmlXPathEvalExpression", NULL, "expr=%s", xmlSecErrorsSafeString(data->expr)); return(NULL); } break; case xmlSecXPathDataTypeXPointer: xpathObj = xmlXPtrEval(data->expr, data->ctx); if(xpathObj == NULL) { xmlSecXmlError2("xmlXPtrEval", NULL, "expr=%s", xmlSecErrorsSafeString(data->expr)); return(NULL); } break; } /* sometime LibXML2 returns an empty nodeset or just NULL, we want to reserve NULL for our own purposes so we simply create an empty node set here */ if(xpathObj->nodesetval == NULL) { xpathObj->nodesetval = xmlXPathNodeSetCreate(NULL); if(xpathObj->nodesetval == NULL) { xmlXPathFreeObject(xpathObj); xmlSecXmlError2("xmlXPathNodeSetCreate", NULL, "expr=%s", xmlSecErrorsSafeString(data->expr)); return(NULL); } } nodes = xmlSecNodeSetCreate(doc, xpathObj->nodesetval, data->nodeSetType); if(nodes == NULL) { xmlSecInternalError2("xmlSecNodeSetCreate", NULL, "type=%d", data->nodeSetType); xmlXPathFreeObject(xpathObj); return(NULL); } xpathObj->nodesetval = NULL; xmlXPathFreeObject(xpathObj); return(nodes); }
/** * xsltInitCtxtKey: * @ctxt: an XSLT transformation context * @doc: an XSLT document * @keyd: the key definition * * Computes the key tables this key and for the current input document. */ int xsltInitCtxtKey(xsltTransformContextPtr ctxt, xsltDocumentPtr doc, xsltKeyDefPtr keyd) { int i; xmlNodeSetPtr nodelist = NULL, keylist; xmlXPathObjectPtr res = NULL; xmlChar *str, **list; xsltKeyTablePtr table; int oldPos, oldSize; xmlNodePtr oldInst; xmlNodePtr oldNode; xsltDocumentPtr oldDoc; xmlDocPtr oldXDoc; int oldNsNr; xmlNsPtr *oldNamespaces; doc->nbKeysComputed++; /* * Evaluate the nodelist */ oldXDoc= ctxt->xpathCtxt->doc; oldPos = ctxt->xpathCtxt->proximityPosition; oldSize = ctxt->xpathCtxt->contextSize; oldNsNr = ctxt->xpathCtxt->nsNr; oldNamespaces = ctxt->xpathCtxt->namespaces; oldInst = ctxt->inst; oldDoc = ctxt->document; oldNode = ctxt->node; if (keyd->comp == NULL) goto error; if (keyd->usecomp == NULL) goto error; ctxt->document = doc; ctxt->xpathCtxt->doc = doc->doc; ctxt->xpathCtxt->node = (xmlNodePtr) doc->doc; ctxt->node = (xmlNodePtr) doc->doc; /* TODO : clarify the use of namespaces in keys evaluation */ ctxt->xpathCtxt->namespaces = keyd->nsList; ctxt->xpathCtxt->nsNr = keyd->nsNr; ctxt->inst = keyd->inst; res = xmlXPathCompiledEval(keyd->comp, ctxt->xpathCtxt); ctxt->xpathCtxt->contextSize = oldSize; ctxt->xpathCtxt->proximityPosition = oldPos; ctxt->inst = oldInst; if (res != NULL) { if (res->type == XPATH_NODESET) { nodelist = res->nodesetval; #ifdef WITH_XSLT_DEBUG_KEYS if (nodelist != NULL) XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsltInitCtxtKey: %s evaluates to %d nodes\n", keyd->match, nodelist->nodeNr)); #endif } else { #ifdef WITH_XSLT_DEBUG_KEYS XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsltInitCtxtKey: %s is not a node set\n", keyd->match)); #endif goto error; } } else { #ifdef WITH_XSLT_DEBUG_KEYS XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsltInitCtxtKey: %s evaluation failed\n", keyd->match)); #endif ctxt->state = XSLT_STATE_STOPPED; goto error; } /* * for each node in the list evaluate the key and insert the node */ if ((nodelist == NULL) || (nodelist->nodeNr <= 0)) goto error; /** * Multiple key definitions for the same name are allowed, so * we must check if the key is already present for this doc */ table = (xsltKeyTablePtr) doc->keys; while (table != NULL) { if (xmlStrEqual(table->name, keyd->name) && (((keyd->nameURI == NULL) && (table->nameURI == NULL)) || ((keyd->nameURI != NULL) && (table->nameURI != NULL) && (xmlStrEqual(table->nameURI, keyd->nameURI))))) break; table = table->next; } /** * If the key was not previously defined, create it now and * chain it to the list of keys for the doc */ if (table == NULL) { table = xsltNewKeyTable(keyd->name, keyd->nameURI); if (table == NULL) goto error; table->next = doc->keys; doc->keys = table; } for (i = 0;i < nodelist->nodeNr;i++) { if (IS_XSLT_REAL_NODE(nodelist->nodeTab[i])) { ctxt->node = nodelist->nodeTab[i]; list = xsltEvalXPathKeys(ctxt, keyd->usecomp, keyd); if (list != NULL) { int ix = 0; str = list[ix++]; while (str != NULL) { #ifdef WITH_XSLT_DEBUG_KEYS XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsl:key : node associated to(%s,%s)\n", keyd->name, str)); #endif keylist = xmlHashLookup(table->keys, str); if (keylist == NULL) { keylist = xmlXPathNodeSetCreate(nodelist->nodeTab[i]); xmlHashAddEntry(table->keys, str, keylist); } else { xmlXPathNodeSetAdd(keylist, nodelist->nodeTab[i]); } switch (nodelist->nodeTab[i]->type) { case XML_ELEMENT_NODE: case XML_TEXT_NODE: case XML_CDATA_SECTION_NODE: case XML_PI_NODE: case XML_COMMENT_NODE: nodelist->nodeTab[i]->psvi = keyd; break; case XML_ATTRIBUTE_NODE: { xmlAttrPtr attr = (xmlAttrPtr) nodelist->nodeTab[i]; attr->psvi = keyd; break; } case XML_DOCUMENT_NODE: case XML_HTML_DOCUMENT_NODE: { xmlDocPtr kdoc = (xmlDocPtr) nodelist->nodeTab[i]; kdoc->psvi = keyd; break; } default: break; } xmlFree(str); str = list[ix++]; } xmlFree(list); #ifdef WITH_XSLT_DEBUG_KEYS } else { XSLT_TRACE(ctxt,XSLT_TRACE_KEYS,xsltGenericDebug(xsltGenericDebugContext, "xsl:key : use %s failed to return strings\n", keyd->use)); #endif } } } error: ctxt->document = oldDoc; ctxt->xpathCtxt->doc = oldXDoc; ctxt->xpathCtxt->nsNr = oldNsNr; ctxt->xpathCtxt->namespaces = oldNamespaces; ctxt->node = oldNode; if (res != NULL) xmlXPathFreeObject(res); return(0); }