attribute.cxx

SP是一个基于GNU C++编译器

// Copyright (c) 1994 James Clark
// See the file COPYING for copying permission.

#ifdef __GNUG__
#pragma implementation
#endif
#include "splib.h"
#include "Attribute.h"
#include "MessageArg.h"
#include "macros.h"
#include "ParserMessages.h"
#include "StringVectorMessageArg.h"
#include "Syntax.h"
#include "Entity.h"
#include "Notation.h"

#ifdef SP_NAMESPACE
namespace SP_NAMESPACE {
#endif

DeclaredValue::DeclaredValue()
{
}

DeclaredValue::~DeclaredValue()
{
}

AttributeValue *DeclaredValue::makeValueFromToken(Text &text,
						  AttributeContext &context,
						  const StringC &name,
						  unsigned &specLength) const
{
  return makeValue(text, context, name, specLength);
}

AttributeSemantics *DeclaredValue::makeSemantics(const TokenizedAttributeValue &,
						 AttributeContext &,
						 const StringC &,
						 unsigned &,
						 unsigned &) const
{
  return 0;
}

Boolean DeclaredValue::containsToken(const StringC &) const
{
  return 0;
}

Boolean DeclaredValue::isNotation() const
{
  return 0;
}

Boolean DeclaredValue::isEntity() const
{
  return 0;
}

Boolean DeclaredValue::isId() const
{
  return 0;
}

Boolean DeclaredValue::isIdref() const
{
  return 0;
}

const Vector<StringC> *DeclaredValue::getTokens() const
{
  return 0;
}


CdataDeclaredValue::CdataDeclaredValue()
{
}

Boolean CdataDeclaredValue::tokenized() const
{
  return 0;
}

AttributeValue *CdataDeclaredValue::makeValue(Text &text, AttributeContext &context,
					      const StringC &,
					      unsigned &specLength) const
{
  const Syntax &syntax = context.attributeSyntax();
  size_t normsep = syntax.normsep();
  size_t normalizedLength = text.normalizedLength(normsep);
  specLength += normalizedLength;
  size_t litlen = syntax.litlen();
  // A length error will already have been given if
  // length > litlen - normsep.
  if (litlen >= normsep && text.size() <= litlen - normsep
      && normalizedLength > litlen)
    context.message(ParserMessages::normalizedAttributeValueLength,
		    NumberMessageArg(litlen),
		    NumberMessageArg(normalizedLength));
  return new CdataAttributeValue(text);
}

void CdataDeclaredValue::buildDesc(AttributeDefinitionDesc &desc) const
{
  desc.declaredValue = AttributeDefinitionDesc::cdata;
}

DeclaredValue *CdataDeclaredValue::copy() const
{
  return new CdataDeclaredValue(*this);
}

TokenizedDeclaredValue::TokenizedDeclaredValue(TokenType type,
					       Boolean isList)
: type_(type), isList_(isList)
{
  switch (type) {
  case name:
  case entityName:
    initialCategories_ = Syntax::nameStartCategory;
    subsequentCategories_ = (Syntax::nameStartCategory|Syntax::digitCategory
			     | Syntax::otherNameCategory);
    break;
  case number:
    initialCategories_ = Syntax::digitCategory;
    subsequentCategories_ = Syntax::digitCategory;
    break;
  case nameToken:
    initialCategories_ = (Syntax::nameStartCategory|Syntax::digitCategory
			  | Syntax::otherNameCategory);
    subsequentCategories_ = initialCategories_;
    break;
  case numberToken:
    initialCategories_ = Syntax::digitCategory;
    subsequentCategories_ = (Syntax::nameStartCategory|Syntax::digitCategory
			     | Syntax::otherNameCategory);
    break;
  }
}

Boolean TokenizedDeclaredValue::tokenized() const
{
  return 1;
}

AttributeValue *TokenizedDeclaredValue::makeValue(Text &text,
						  AttributeContext &context,
						  const StringC &str,
						  unsigned &specLength) const
{
  return makeTokenizedValue(text, context, str, specLength);
}

TokenizedAttributeValue *
TokenizedDeclaredValue::makeTokenizedValue(Text &text,
					   AttributeContext &context,
					   const StringC &name,
					   unsigned &specLength) const
{
  Vector<size_t> spaceIndex;
  const Syntax &syntax = context.attributeSyntax();
  Char space = syntax.space();
  text.subst(*(type_ == entityName
	       ? syntax.entitySubstTable()
	       : syntax.generalSubstTable()),
	     space);
  const StringC &value = text.string();
  size_t i = 0;
  size_t length = value.size();

  for (;;) {
    if (i >= length) {
      // ends with a space (which would have to have been entered
      // via a numeric character reference)
      if (context.validate())
	context.message(ParserMessages::attributeValueSyntax);
      break;
    }
    size_t startIndex = i;
    if (context.validate()) {
      if (!(syntax.charCategory(value[i]) & initialCategories_)) {
        context.Messenger::setNextLocation(text.charLocation(i));
        Char c = value[i];
	if (!(syntax.charCategory(value[i]) & subsequentCategories_))
	  context.message(ParserMessages::attributeValueChar,
	                  StringMessageArg(StringC(&c, 1)),
	                  StringMessageArg(name));
	else if (initialCategories_ == Syntax::digitCategory)
	  context.message(ParserMessages::attributeValueNumberToken,
	                  StringMessageArg(StringC(&c, 1)),
	                  StringMessageArg(name));
	else
	  context.message(ParserMessages::attributeValueName,
	                  StringMessageArg(StringC(&c, 1)),
	                  StringMessageArg(name));
      }
      else {
	for (++i;
             i < length
	     && (syntax.charCategory(value[i]) & subsequentCategories_);
	     i++)
	  ;
	if (i < length && value[i] != space) {
	  Char c = value[i];
	  // character value[i] is not allowed anywhere in the value
	  context.Messenger::setNextLocation(text.charLocation(i));
	  context.message(ParserMessages::attributeValueChar,
		          StringMessageArg(StringC(&c, 1)),
		          StringMessageArg(name));
	}
      }
    }
    while (i < length && value[i] != space)
      i++;
    if (i - startIndex > syntax.namelen()) {
      context.Messenger::setNextLocation(text.charLocation(i));
      context.message(ParserMessages::nameTokenLength,
		      NumberMessageArg(syntax.namelen()));
    }
    if (i == length)
      break;
    if (!isList_ && context.validate() && spaceIndex.size() == 0) {
      context.Messenger::setNextLocation(text.charLocation(i));
      context.message(ParserMessages::attributeValueMultiple,
		      StringMessageArg(name));
    }
    spaceIndex.push_back(i);
    i++;
  }
  size_t normsep = syntax.normsep();
  size_t litlen = syntax.litlen();
  size_t normalizedLength = normsep + length;
  // should we count CDATA and SDATA entities here?
  if (isList_) {
    normalizedLength += 1;
    // length is now the number of characters in each token in the list
    // + 1 for each token in the list; so add normsep - 1 for each
    // token in the list.
    if (normsep > 0)
      normalizedLength += (normsep - 1)*(spaceIndex.size() + 1);
    else
      normalizedLength -= spaceIndex.size() + 1;
  }
  specLength += normalizedLength;
  // A length error will already have been given if
  // length > litlen - normsep.
  if (litlen >= normsep && length <= litlen - normsep
      && normalizedLength > litlen)
    context.message(ParserMessages::normalizedAttributeValueLength,
		    NumberMessageArg(litlen),
		    NumberMessageArg(normalizedLength));
  return new TokenizedAttributeValue(text, spaceIndex);
}

Boolean TokenizedAttributeValue::recoverUnquoted(const StringC &str,
						 const Location &strLoc,
						 AttributeContext &context,
						 const StringC &name)
{
  TextIter iter(text_);
  TextItem::Type type;
  const Char *s;
  size_t len;
  const Location *loc;
  if (iter.next(type, s, len, loc)
      && type == TextItem::data
      && len == text_.size()
      && loc->origin().pointer() == strLoc.origin().pointer()
      && loc->index() + len == strLoc.index()
      && !iter.next(type, s, len, loc)) {
    context.Messenger::setNextLocation(strLoc);
    context.message(ParserMessages::attributeValueChar,
		    StringMessageArg(StringC(str.data(), 1)),
		    StringMessageArg(name));
    return 1;
  }
  return 0;
}

void TokenizedDeclaredValue::buildDesc(AttributeDefinitionDesc &desc) const
{
  desc.declaredValue = AttributeDefinitionDesc::DeclaredValue(
    type_ - name + (isList_
		    ? AttributeDefinitionDesc::names
		    : AttributeDefinitionDesc::name));
}

DeclaredValue *TokenizedDeclaredValue::copy() const
{
  return new TokenizedDeclaredValue(*this);
}

GroupDeclaredValue::GroupDeclaredValue(TokenType type,
				       Vector<StringC> &vec)
: TokenizedDeclaredValue(type, 0)
{
  vec.swap(allowedValues_);
}

void GroupDeclaredValue::buildDesc(AttributeDefinitionDesc &desc) const
{
  desc.allowedValues = allowedValues_;
}

DeclaredValue *GroupDeclaredValue::copy() const
{
  return new GroupDeclaredValue(*this);
}

AttributeValue *GroupDeclaredValue::makeValue(Text &text,
					      AttributeContext &context,
					      const StringC &name,
					      unsigned &specLength) const
{
  TokenizedAttributeValue *val = makeTokenizedValue(text, context, name,
						    specLength);
  if (!val || !context.validate())
    return val;
  for (size_t i = 0; i < allowedValues_.size(); i++)
    if (val->string() == allowedValues_[i])
      return val;
  context.message(ParserMessages::attributeValueNotInGroup,
		  StringMessageArg(val->string()),
		  StringMessageArg(name),
		  StringVectorMessageArg(allowedValues_));
  return val;
}

AttributeValue *GroupDeclaredValue::makeValueFromToken(Text &text,
						       AttributeContext &context,
						       const StringC &,
						       unsigned &specLength)
     const
{
  const Syntax &syntax = context.attributeSyntax();
  size_t litlen = syntax.litlen();
  size_t normsep = syntax.normsep();
  if (normsep > litlen || text.size() >  litlen - normsep)
    context.message(ParserMessages::normalizedAttributeValueLength,
		    NumberMessageArg(litlen),
		    NumberMessageArg(text.size() + normsep));
  specLength += text.size() + normsep;
  return new TokenizedAttributeValue(text, Vector<size_t>());
}

Boolean GroupDeclaredValue::containsToken(const StringC &token) const
{
  for (size_t i = 0; i < allowedValues_.size(); i++)
    if (allowedValues_[i] == token)
      return 1;
  return 0;
}

const Vector<StringC> *GroupDeclaredValue::getTokens() const
{
  return &allowedValues_;
}

NameTokenGroupDeclaredValue::NameTokenGroupDeclaredValue(Vector<StringC> &vec)
: GroupDeclaredValue(nameToken, vec)
{
}

void NameTokenGroupDeclaredValue::buildDesc(AttributeDefinitionDesc &desc) const
{
  GroupDeclaredValue::buildDesc(desc);
  desc.declaredValue = AttributeDefinitionDesc::nameTokenGroup;
}

DeclaredValue *NameTokenGroupDeclaredValue::copy() const
{
  return new NameTokenGroupDeclaredValue(*this);
}

NotationDeclaredValue::NotationDeclaredValue(Vector<StringC> &vec)
: GroupDeclaredValue(name, vec)
{
}

Boolean NotationDeclaredValue::isNotation() const
{
  return 1;
}

AttributeSemantics *
NotationDeclaredValue::makeSemantics(const TokenizedAttributeValue &value,
				     AttributeContext &context,
				     const StringC &,
				     unsigned &,
				     unsigned &) const
{
  ConstPtr<Notation> notation
    = context.getAttributeNotation(value.string(),
				   value.tokenLocation(0));
  if (notation.isNull()) {
    if (context.validate()) {
      context.setNextLocation(value.tokenLocation(0));
      context.message(ParserMessages::invalidNotationAttribute,
		      StringMessageArg(value.string()));
    }
    return 0;
  }
  return new NotationAttributeSemantics(notation);
}

void NotationDeclaredValue::buildDesc(AttributeDefinitionDesc &desc) const
{
  GroupDeclaredValue::buildDesc(desc);
  desc.declaredValue = AttributeDefinitionDesc::notation;
}

DeclaredValue *NotationDeclaredValue::copy() const
{
  return new NotationDeclaredValue(*this);
}

EntityDeclaredValue::EntityDeclaredValue(Boolean isList)
: TokenizedDeclaredValue(entityName, isList)
{
}

Boolean EntityDeclaredValue::isEntity() const
{
  return 1;
}

AttributeSemantics *
EntityDeclaredValue::makeSemantics(const TokenizedAttributeValue &value,
				   AttributeContext &context,
				   const StringC &,
				   unsigned &,
				   unsigned &nEntityNames) const
{
  Boolean valid = 1;
  size_t nTokens = value.nTokens();
  nEntityNames += nTokens;
  Vector<ConstPtr<Entity> > entities(nTokens);
  for (size_t i = 0; i < nTokens; i++) {
    entities[i] = context.getAttributeEntity(value.token(i),
					     value.tokenLocation(i));
    if (entities[i].isNull()) {
      if (context.validate()) {
	context.setNextLocation(value.tokenLocation(i));
	context.message(ParserMessages::invalidEntityAttribute,
		        StringMessageArg(value.token(i)));
      }
      valid = 0;
    }
    else if (!entities[i]->isDataOrSubdoc()) {
      if (context.validate()) {
        context.Messenger::setNextLocation(value.tokenLocation(i));
        context.message(ParserMessages::notDataOrSubdocEntity,
		        StringMessageArg(value.token(i)));
      }