kurl.cpp

monqueror一个很具有参考价值的源玛

/* This file is part of the KDE libraries
    Copyright (C) 1999 Torben Weis <[e-mail]>

    This library is free software; you can redistribute it and/or
    modify it under the terms of the GNU Library General Public
    License as published by the Free Software Foundation; either
    version 2 of the License, or (at your option) any later version.

    This library is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    Library General Public License for more details.

    You should have received a copy of the GNU Library General Public License
    along with this library; see the file COPYING.LIB.  If not, write to
    the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
    Boston, MA 02111-1307, USA.
*/

#include "kurl.h"
#include <kdebug.h>

#include <stdio.h>
#include <assert.h>
#include <ctype.h>
#include <stdlib.h>

#include <qurl.h>
#include <qdir.h>
#include <qstringlist.h>

#include <qtextcodec.h>

#define _CODE_BY_YMWEI 1

#if !_CODE_BY_YMWEI
static QTextCodec * codecForHint( int encoding_hint /* not 0 ! */ )
{
    // Get the charset name from encoding_hint - but KCharsets doesn't
    // know about "unicode"
    QString charsetName =
        (encoding_hint == QFont::Unicode) ? "utf8" :
        KGlobal::charsets()->name( (QFont::CharSet) encoding_hint );

    bool ok;
    QTextCodec * textCodec = KGlobal::charsets()->codecForName( charsetName, ok );
    return ok ? textCodec : 0L;
}
#endif

static
QString encode( const QString& segment, bool encode_slash, int encoding_hint )
{
  char encode_extra = encode_slash ? '/' : 0;
  QCString local;
  if (encoding_hint==0)
    local = segment.local8Bit();
  else
  {
#if _CODE_BY_YMWEI
      local = segment.local8Bit();
#else
      QTextCodec * textCodec = codecForHint( encoding_hint );
      if (!textCodec)
          local = segment.local8Bit();
      else
          local = textCodec->fromUnicode( segment );
#endif
  }

  int old_length = local.length();

  if ( !old_length )
    return QString::null;

  // a worst case approximation
  QChar *new_segment = new QChar[ old_length * 3 + 1 ];
  int new_length = 0;

  for ( int i = 0; i < old_length; i++ )
  {
    // 'unsave' and 'reserved' characters
    // according to RFC 1738,
    // 2.2. URL Character Encoding Issues (pp. 3-4)
    // WABA: Added non-ascii
    unsigned char character = local[i];
    if ( (character <= 32) || (character >= 127) ||
         strchr("<>#@\"&%$:,;?={}|^~[]\'`\\", character) ||
         (character == encode_extra) )
    {
      new_segment[ new_length++ ] = '%';

      unsigned int c = character / 16;
      c += (c > 9) ? ('A' - 10) : '0';
      new_segment[ new_length++ ] = c;

      c = character % 16;
      c += (c > 9) ? ('A' - 10) : '0';
      new_segment[ new_length++ ] = c;

    }
    else
      new_segment[ new_length++ ] = local[i];
  }

  QString result = QString(new_segment, new_length);
  delete [] new_segment;
  return result;
}

static char hex2int( unsigned int _char )
{
  if ( _char >= 'A' && _char <='F')
    return _char - 'A' + 10;
  if ( _char >= 'a' && _char <='f')
    return _char - 'a' + 10;
  if ( _char >= '0' && _char <='9')
    return _char - '0';
  return -1;
}

// WABA: The result of lazy_encode isn't usable for a URL which
// needs to satisfies RFC requirements. However, the following
// operation will make it usable again:
//      encode(decode(...))
//
// As a result one can see that url.prettyURL() does not result in
// a RFC compliant URL but that the following sequence does:
//      KURL(url.prettyURL()).url()


static QString lazy_encode( const QString& segment )
{
  int old_length = segment.length();

  if ( !old_length )
    return QString::null;

  // a worst case approximation
  QChar *new_segment = new QChar[ old_length * 3 + 1 ];
  int new_length = 0;

  for ( int i = 0; i < old_length; i++ )
  {
    unsigned int character = segment[i].unicode(); // Don't use latin1()
                                                   // It returns 0 for non-latin1 values
    // Small set of really ambiguous chars
    if ((character < 32) ||  // Low ASCII
        ((character == '%') && // The escape character itself
           (i+2 < old_length) && // But only if part of a valid escape sequence!
          (hex2int(segment[i+1].unicode())!= -1) &&
          (hex2int(segment[i+2].unicode())!= -1)) ||
        (character == '?') || // Start of query delimiter
        (character == '#') || // Start of reference delimiter
        ((character == 32) && (i+1 == old_length))) // A trailing space
    {
      new_segment[ new_length++ ] = '%';

      unsigned int c = character / 16;
      c += (c > 9) ? ('A' - 10) : '0';
      new_segment[ new_length++ ] = c;

      c = character % 16;
      c += (c > 9) ? ('A' - 10) : '0';
      new_segment[ new_length++ ] = c;
    }
    else
    new_segment[ new_length++ ] = segment[i];
  }

  QString result = QString(new_segment, new_length);
  delete [] new_segment;
  return result;
}


static QString decode( const QString& segment, bool *keepEncoded=0, int encoding_hint=0 )
{
  bool isUnicode = false; // This detects utf-16, not utf-8
  bool isLocal = false;
  bool isAscii = true;
  int old_length = segment.length();
  if ( !old_length )
    return QString::null;

  int new_length = 0;

  // make a copy of the old one
  char *new_segment = new char[ old_length + 1];
  QChar *new_usegment = new QChar[ old_length + 1 ];

  int i = 0;
  while( i < old_length )
  {
    unsigned int character = segment[ i++ ].unicode();
    if ( (character == '%' ) &&
         ( i+1 < old_length) ) // Must have at least two chars left!
    {
      char a = hex2int( segment[i].latin1() );
      char b = hex2int( segment[i+1].latin1() );
      if ((a != -1) && (b != -1)) // Only replace if sequence is valid
      {
         character = a * 16 + b; // Replace with value of %dd
         i += 2; // Skip dd
         if (character > 127)
            isLocal = true;
      }
    }
    new_segment [ new_length ] = character;
    new_usegment [ new_length ] = character;
    new_length++;
    if (character > 127)
    {
       isAscii = false;
       if (character > 255)
          isUnicode = true;
    }
  }
  new_segment [ new_length ] = 0;
  QString result;
  // Encoding specified
  if ( encoding_hint )
  {
#if _CODE_BY_YMWEI
      result = QString::fromLocal8Bit(new_segment, new_length);
#else
      QTextCodec * textCodec = codecForHint( encoding_hint );
      if (textCodec)
      {
          QByteArray array;
          array.setRawData(new_segment, new_length);
          result = textCodec->toUnicode( array, new_length );
          array.resetRawData(new_segment, new_length);
      }
      else
          result = QString::fromLocal8Bit(new_segment, new_length);
#endif

      // No idea about keepEncoded... Hmm, it's unused anyway (!)
  }
  // Guess the encoding, if not specified
  else if ((!isAscii && !isUnicode) || isLocal)
  {
     result = QString::fromLocal8Bit(new_segment, new_length);
     if (keepEncoded)
       *keepEncoded = true;
  }
  else
  {
     result = QString( new_usegment, new_length);
     if (keepEncoded)
       *keepEncoded = false;
  }
  delete [] new_segment;
  delete [] new_usegment;
  return result;
}

bool KURL::isRelativeURL(const QString &_url)
{
  int len = _url.length();
  if (!len) return true; // Very short relative URL.
  const QChar *str = _url.unicode();

  // Absolute URL must start with alpha-character
  if (!isalpha(str[0].latin1()))
     return true; // Relative URL

  for(int i = 1; i < len; i++)
  {
     char c = str[i].latin1(); // Note: non-latin1 chars return 0!
     if (c == ':')
     {
        // URL starts with "xxx:/" -> absolute URL
        i++;
        if ((i < len) && (str[i].latin1() == '/'))
           return false;
        // Make an expection for mailto:user@host
        if ((i == 7) && (_url.left(7).lower() == "mailto:"))
           return false;
        // Make another expection for news:comp.os.newsgroup
        if ((i == 5) && (_url.left(5).lower() == "news:"))
           return false;
        // Make another expection for man:(3)
        if ((i == 4) && (_url.left(4).lower() == "man:"))
           return false;
        return true; // "xxx:" or "xxx::yyy"
     }

     // Protocol part may only contain alpha, digit, + or -
     if (!isalpha(c) && !isdigit(c) && (c != '+') && (c != '-'))
        return true; // Relative URL
  }
  // URL did not contain ':'
  return true; // Relative URL
}

KURL::List::List(const QStringList &list)
{
  for (QStringList::ConstIterator it = list.begin();
       it != list.end();
       it++)
    {
      append( KURL(*it) );
    }
}

QStringList KURL::List::toStringList() const
{
  QStringList lst;
   for( KURL::List::ConstIterator it = begin();
        it != end();
        it++)
   {
      lst.append( (*it).url() );
   }
   return lst;
}


KURL::KURL()
{
  reset();
}

KURL::~KURL()
{
}


KURL::KURL( const QString &url, int encoding_hint )
{
  reset();
  parse( url, encoding_hint );
}

KURL::KURL( const char * url, int encoding_hint )
{
  reset();
  parse( QString::fromLatin1(url), encoding_hint );
}

KURL::KURL( const KURL& _u )
{
  m_strProtocol = _u.m_strProtocol;
  m_strUser = _u.m_strUser;
  m_strPass = _u.m_strPass;
  m_strHost = _u.m_strHost;
  m_strPath = _u.m_strPath;
  m_strPath_encoded = _u.m_strPath_encoded;
  m_strQuery_encoded = _u.m_strQuery_encoded;
  m_strRef_encoded = _u.m_strRef_encoded;
  m_bIsMalformed = _u.m_bIsMalformed;
  m_iPort = _u.m_iPort;
}

QDataStream & operator<< (QDataStream & s, const KURL & a)
{
    s << a.m_strProtocol << a.m_strUser << a.m_strPass << a.m_strHost
      << a.m_strPath << a.m_strPath_encoded << a.m_strQuery_encoded << a.m_strRef_encoded
      << Q_INT8(a.m_bIsMalformed ? 1 : 0) << a.m_iPort;
    return s;
}

QDataStream & operator>> (QDataStream & s, KURL & a)
{
    Q_INT8 malf;
    s >> a.m_strProtocol >> a.m_strUser >> a.m_strPass >> a.m_strHost
      >> a.m_strPath >> a.m_strPath_encoded >> a.m_strQuery_encoded >> a.m_strRef_encoded
      >> malf >> a.m_iPort;
    a.m_bIsMalformed = (malf != 0);

    if ( a.m_strQuery_encoded.isEmpty() )
      a.m_strQuery_encoded = QString::null;

    return s;
}

KURL::KURL( const QUrl &u )
{
  m_strProtocol = u.protocol();
  m_strUser = u.user();
  m_strPass = u.password();
  m_strHost = u.host();
  m_strPath = u.path( FALSE );
  m_strPath_encoded = QString::null;
  m_strQuery_encoded = u.query();
  m_strRef_encoded = u.ref();
  m_bIsMalformed = !u.isValid();
  m_iPort = u.port();
}

KURL::KURL( const KURL& _u, const QString& _rel_url, int encoding_hint )
{
  // WORKAROUND THE RFC 1606 LOOPHOLE THAT ALLOWS
  // http:/index.html AS A VALID SYNTAX FOR RELATIVE
  // URLS. ( RFC 2396 section 5.2 item # 3 )
  QString rUrl = _rel_url;
  int len = _u.m_strProtocol.length();
  if ( _u.hasHost() && rUrl.length() != 0 &&
       rUrl.left( len ).lower() == _u.m_strProtocol.lower() &&
       rUrl[len] == ':' && (rUrl[len+1] != '/' ||
       (rUrl[len+1] == '/' && rUrl[len+2] != '/')) )
  {
                rUrl.remove( 0, rUrl.find( ':' ) + 1 );
  }

  if ( rUrl[0] == '#' )
  {
    *this = _u;
    setHTMLRef( decode(rUrl.mid(1), 0, encoding_hint) );
  }
  else if ( isRelativeURL( rUrl) )
  {
    *this = _u;
    m_strQuery_encoded = QString::null;
    m_strRef_encoded = QString::null;
    if ( rUrl[0] == '/')
    {
        m_strPath = QString::null;
        m_strPath_encoded = QString::null;
    }
    else
    {
       int pos = m_strPath.findRev( '/' );
       if (pos >= 0)
          m_strPath.truncate(pos);
       m_strPath += '/';
       if (!m_strPath_encoded.isEmpty())
       {
          pos = m_strPath_encoded.findRev( '/' );
          if (pos >= 0)
             m_strPath_encoded.truncate(pos);
          m_strPath_encoded += '/';
       }
    }
    KURL tmp( url() + rUrl, encoding_hint);
    *this = tmp;
    cleanPath();
  }
  else
  {
    KURL tmp( rUrl, encoding_hint);
    *this = tmp;
  }
}

void KURL::reset()
{
  m_strProtocol = QString::null;
  m_strUser = QString::null;
  m_strPass = QString::null;
  m_strHost = QString::null;
  m_strPath = QString::null;
  m_strPath_encoded = QString::null;
  m_strRef_encoded = QString::null;
  m_bIsMalformed = true;
  m_iPort = 0;
}

bool KURL::isEmpty() const
{
  return (m_strPath.isEmpty() && m_strProtocol.isEmpty());
}

KURL KURL::completeURL (const DOM::DOMString & _url, const DOM::DOMString & _baseUrl)
{
    QString url = _url.string();
    QString baseUrl = _baseUrl.string();
    KURL orig(baseUrl);
    if(_url[(unsigned int)0] != '/')
    {
        KURL u( orig, url );
        return u;
    }
    orig.setEncodedPathAndQuery(url);
    return orig;
}

void KURL::parse( const QString& _url, int encoding_hint )
{
  // Return immediately whenever the given url
  // is empty or null.
  if ( _url.length() == 0  )
  {
    m_strProtocol = _url;
    return;
  }

  // This is wrong!! All URLs should be deemed invalid until
  // they have been correctly parsed. Specially now since ::url()
  // returns the given url even if it is malformed. (DA)
  // m_bIsMalformed = false;

  QString port;
  int start = 0;