parsesd.cxx

SP是一个基于GNU C++编译器

// Copyright (c) 1994, 1995, 1997 James Clark
// See the file COPYING for copying permission.

#include "splib.h"
#include "Parser.h"
#include "macros.h"
#include "SdFormalError.h"
#include "MessageBuilder.h"
#include "ParserMessages.h"
#include "MessageArg.h"
#include "CharsetRegistry.h"
#include "ISetIter.h"
#include "token.h"
#include "TokenMessageArg.h"
#include "constant.h"
#include "SdText.h"
#include "NumericCharRefOrigin.h"

#ifdef SP_NAMESPACE
namespace SP_NAMESPACE {
#endif

class CharSwitcher {
public:
  CharSwitcher();
  void addSwitch(WideChar from, WideChar to);
  SyntaxChar subst(WideChar c);
  size_t nSwitches() const;
  Boolean switchUsed(size_t i) const;
  WideChar switchFrom(size_t i) const;
  WideChar switchTo(size_t i) const;
private:
  Vector<PackedBoolean> switchUsed_;
  Vector<WideChar> switches_;
};

// Information about the SGML declaration being built.

struct SdBuilder {
  SdBuilder();
  void addFormalError(const Location &, const MessageType1 &, const StringC &);
  Ptr<Sd> sd;
  Ptr<Syntax> syntax;
  CharsetDecl syntaxCharsetDecl;
  CharsetInfo syntaxCharset;
  CharSwitcher switcher;
  Boolean externalSyntax;
  Boolean enr;
  Boolean www;
  Boolean valid;
  Boolean external;
  IList<SdFormalError> formalErrorList;
};

class CharsetMessageArg : public MessageArg {
public:
  CharsetMessageArg(const ISet<WideChar> &set);
  MessageArg *copy() const;
  void append(MessageBuilder &) const;
private:
  ISet<WideChar> set_;
};

struct SdParam {
  typedef unsigned char Type;
  enum {
    invalid,
    eE,
    minimumLiteral,
    mdc,
    minus,
    number,
    capacityName,
    name,
    paramLiteral,
    systemIdentifier,
    generalDelimiterName,
    referenceReservedName,
    quantityName,
    reservedName		// Sd::ReservedName is added to this
  };
  Type type;
  StringC token;
  Text literalText;
  String<SyntaxChar> paramLiteralText;
  union {
    Number n;
    Sd::Capacity capacityIndex;
    Syntax::Quantity quantityIndex;
    Syntax::ReservedName reservedNameIndex;
    Syntax::DelimGeneral delimGeneralIndex;
  };
};

class AllowedSdParams {
public:
  AllowedSdParams(SdParam::Type,
		  SdParam::Type = SdParam::invalid,
		  SdParam::Type = SdParam::invalid,
		  SdParam::Type = SdParam::invalid,
		  SdParam::Type = SdParam::invalid,
		  SdParam::Type = SdParam::invalid);
  Boolean param(SdParam::Type) const;
  SdParam::Type get(int i) const;
private:
  enum { maxAllow = 6 };
  SdParam::Type allow_[maxAllow];
};

class AllowedSdParamsMessageArg : public MessageArg {
public:
  AllowedSdParamsMessageArg(const AllowedSdParams &allow,
			    const ConstPtr<Sd> &sd);
  MessageArg *copy() const;
  void append(MessageBuilder &) const;
private:
  AllowedSdParams allow_;
  ConstPtr<Sd> sd_;
};

struct StandardSyntaxSpec {
  struct AddedFunction {
    const char *name;
    Syntax::FunctionClass functionClass;
    SyntaxChar syntaxChar;
  };
  const AddedFunction *addedFunction;
  size_t nAddedFunction;
  Boolean shortref;
};

static StandardSyntaxSpec::AddedFunction coreFunctions[] = {
  { "TAB", Syntax::cSEPCHAR, 9 },
};

static StandardSyntaxSpec coreSyntax = {
  coreFunctions, SIZEOF(coreFunctions), 0
};

static StandardSyntaxSpec refSyntax = {
  coreFunctions, SIZEOF(coreFunctions), 1
};

void Parser::doInit()
{
  if (cancelled()) {
    allDone();
    return;
  }
  // When document entity doesn't exist, don't give any errors
  // other than the cannot open error.
  if (currentInput()->get(messenger()) == InputSource::eE) {
    if (currentInput()->accessError()) {
      allDone();
      return;
    }
  }
  else
    currentInput()->ungetToken();
  const CharsetInfo &initCharset = sd().internalCharset();
  ISet<WideChar> missing;
  findMissingMinimum(initCharset, missing);
  if (!missing.isEmpty()) {
    message(ParserMessages::sdMissingCharacters, CharsetMessageArg(missing));
    giveUp();
    return;
  }
  Boolean found = 0;
  StringC systemId;
  if (scanForSgmlDecl(initCharset)) {
    if (options().warnExplicitSgmlDecl)
      message(ParserMessages::explicitSgmlDecl);
    found = 1;
  }
  else {
    currentInput()->ungetToken();
    if (entityCatalog().sgmlDecl(initCharset, messenger(), systemId)) {
      InputSource *in = entityManager().open(systemId,
					     sd().docCharset(),
					     InputSourceOrigin::make(),
					     0,
					     messenger());
      if (in) {
	pushInput(in);
	if (scanForSgmlDecl(initCharset))
	  found = 1;
	else {
	  message(ParserMessages::badDefaultSgmlDecl);
	  popInputStack();
	}
      }
    }
  }
  if (found) {
    if (startMarkup(eventsWanted().wantPrologMarkup(), currentLocation())) {
      size_t nS = currentInput()->currentTokenLength() - 6;
      for (size_t i = 0; i < nS; i++)
	currentMarkup()->addS(currentInput()->currentTokenStart()[i]);
      currentMarkup()->addDelim(Syntax::dMDO);
      currentMarkup()->addSdReservedName(Sd::rSGML,
					currentInput()->currentTokenStart()
					+ (currentInput()->currentTokenLength() - 4),
					4);
    }
    Syntax *syntaxp = new Syntax(sd());
    CharSwitcher switcher;
    if (!setStandardSyntax(*syntaxp, refSyntax, sd().internalCharset(), switcher, 0)) {
      giveUp();
      return;
    }
    syntaxp->implySgmlChar(sd());
    setSyntax(syntaxp);
    compileSdModes();
    ConstPtr<Sd> refSd(sdPointer());
    ConstPtr<Syntax> refSyntax(syntaxPointer());
    if (!parseSgmlDecl()) {
      giveUp();
      return;
    }
    // queue an SGML declaration event
    eventHandler().sgmlDecl(new (eventAllocator())
			    SgmlDeclEvent(sdPointer(),
					  syntaxPointer(),
					  instanceSyntaxPointer(),
					  refSd,
					  refSyntax,
					  currentInput()->nextIndex(),
					  systemId,
					  markupLocation(),
					  currentMarkup()));
    if (inputLevel() == 2) {
      // FIXME perhaps check for junk after SGML declaration
      popInputStack();
    }
  }
  else {
    if (!implySgmlDecl()) {
      giveUp();
      return;
    }
    currentInput()->willNotSetDocCharset();
    // queue an SGML declaration event
    eventHandler().sgmlDecl(new (eventAllocator())
			    SgmlDeclEvent(sdPointer(),
					  syntaxPointer()));
  }
						      
  // Now we have sd and syntax set up, prepare to parse the prolog.
  compilePrologModes();
  setPhase(prologPhase);
}

Boolean Parser::implySgmlDecl()
{
  Syntax *syntaxp = new Syntax(sd());
  const StandardSyntaxSpec *spec;
  if (options().shortref)
    spec = &refSyntax;
  else
    spec = &coreSyntax;
  CharSwitcher switcher;
  if (!setStandardSyntax(*syntaxp, *spec, sd().internalCharset(), switcher, 0))
    return 0;
  syntaxp->implySgmlChar(sd());
  for (int i = 0; i < Syntax::nQuantity; i++)
    syntaxp->setQuantity(i, options().quantity[i]);
  setSyntax(syntaxp);
  return 1;
}

Boolean Parser::setStandardSyntax(Syntax &syn,
				  const StandardSyntaxSpec &spec,
				  const CharsetInfo &internalCharset,
				  CharSwitcher &switcher,
				  Boolean www)
{
  static UnivCharsetDesc::Range syntaxCharsetRanges[] = {
    { 0, 128, 0 },
  };
  static UnivCharsetDesc syntaxCharsetDesc(syntaxCharsetRanges,
					   SIZEOF(syntaxCharsetRanges));
  static CharsetInfo syntaxCharset(syntaxCharsetDesc);

  Boolean valid = 1;
  if (!checkSwitches(switcher, syntaxCharset))
    valid = 0;
  size_t i;
  for (i = 0; i < switcher.nSwitches(); i++)
    if (switcher.switchTo(i) >= 128)
      message(ParserMessages::switchNotInCharset,
	      NumberMessageArg(switcher.switchTo(i)));
  static const Char shunchar[] = {
    0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
    16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,
    127, 255
    };
  
  for (i = 0; i < SIZEOF(shunchar); i++)
    syn.addShunchar(shunchar[i]);
  syn.setShuncharControls();
  static Syntax::StandardFunction standardFunctions[3] = {
    Syntax::fRE, Syntax::fRS, Syntax::fSPACE
  };
  static SyntaxChar functionChars[3] = { 13, 10, 32 };
  for (i = 0; i < 3; i++) {
    Char docChar;
    if (translateSyntax(switcher,
			syntaxCharset,
			internalCharset,
			functionChars[i],
			docChar)
	&& checkNotFunction(syn, docChar))
      syn.setStandardFunction(standardFunctions[i], docChar);
    else
      valid = 0;
  }
  for (i = 0; i < spec.nAddedFunction; i++) {
    Char docChar;
    if (translateSyntax(switcher,
			syntaxCharset,
			internalCharset,
			spec.addedFunction[i].syntaxChar,
			docChar)
	&& checkNotFunction(syn, docChar))
      syn.addFunctionChar(internalCharset.execToDesc(spec.addedFunction[i].name),
			  spec.addedFunction[i].functionClass,
			  docChar);
    else
      valid = 0;
  }
  
  static SyntaxChar nameChars[2] = { 45, 46 }; // '-' '.'
  ISet<Char> nameCharSet;
  for (i = 0; i < 2; i++) {
    Char docChar;
    if (translateSyntax(switcher,
			syntaxCharset,
			internalCharset,
			nameChars[i],
			docChar))
	nameCharSet.add(docChar);
    else
      valid = 0;
  }
  if (!checkNmchars(nameCharSet, syn))
    valid = 0;
  else
    syn.addNameCharacters(nameCharSet);
  syn.setNamecaseGeneral(1);
  syn.setNamecaseEntity(0);
  if (!setRefDelimGeneral(syn, syntaxCharset, internalCharset, switcher))
    valid = 0;
  setRefNames(syn, internalCharset, www);
  syn.enterStandardFunctionNames();
  if (spec.shortref
      && !addRefDelimShortref(syn, syntaxCharset, internalCharset, switcher))
    valid = 0;
  return valid;
}

Boolean Parser::setRefDelimGeneral(Syntax &syntax,
				   const CharsetInfo &syntaxCharset,
				   const CharsetInfo &internalCharset,
				   CharSwitcher &switcher)
{
  // Column 3 from Figure 3
  static const char delims[][2] = {
    { 38 },
    { 45, 45 },
    { 38, 35 },
    { 93 },
    { 91 },
    { 93 },
    { 91 },
    { 38 },
    { 60, 47 },
    { 41 },
    { 40 },
    { 0 }, // HCRO
    { 34 },
    { 39 },
    { 62 },
    { 60, 33 },
    { 45 },
    { 93, 93 },
    { 47 },
    { 47 }, // NESTC
    { 63 },
    { 124 },
    { 37 },
    { 62 },
    { 60, 63 },
    { 43 },
    { 59 },
    { 42 },
    { 35 },
    { 44 },
    { 60 },
    { 62 },
    { 61 },
  };
  Boolean valid = 1;
  ISet<WideChar> missing;
  for (int i = 0; i < Syntax::nDelimGeneral; i++)
    if (syntax.delimGeneral(i).size() == 0) {
      StringC delim;
      size_t j;
      for (j = 0; j < 2 && delims[i][j] != '\0'; j++) {
	UnivChar univChar = translateUniv(delims[i][j], switcher,
					  syntaxCharset);
	Char c;
	if (univToDescCheck(internalCharset, univChar, c))
	  delim += c;
	else {
	  missing += univChar;
	  valid = 0;
	}
      }
      if (delim.size() == j) {
	if (checkGeneralDelim(syntax, delim))
	  syntax.setDelimGeneral(i, delim);
	else
	  valid = 0;
      }
    }
  if (!missing.isEmpty())
    message(ParserMessages::missingSignificant646, CharsetMessageArg(missing));
  return valid;
}

void Parser::setRefNames(Syntax &syntax, const CharsetInfo &internalCharset,
			 Boolean www)
{
  static const char *const referenceNames[] = {
    "ALL",
    "ANY",
    "ATTLIST",
    "CDATA",
    "CONREF",
    "CURRENT",
    "DATA",
    "DEFAULT",
    "DOCTYPE",
    "ELEMENT",
    "EMPTY",
    "ENDTAG",
    "ENTITIES",
    "ENTITY",
    "FIXED",
    "ID",
    "IDLINK",
    "IDREF",
    "IDREFS",