decoder.cpp

khtml在gtk上的移植版本

/*
    This file is part of the KDE libraries

    Copyright (C) 1999 Lars Knoll ([e-mail])
    Copyright (C) 2003 Apple Computer, Inc.

    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.
*/
//----------------------------------------------------------------------------
//
// KDE HTML Widget -- decoder for input stream

#undef DECODE_DEBUG
//#define DECODE_DEBUG

#include "decoder.h"
using namespace khtml;

#include "htmlhashes.h"

#include <qregexp.h>
#include <qtextcodec.h>

#include <kglobal.h>
#include <kcharsets.h>

#include <ctype.h>
#include <kdebug.h>
#include <klocale.h>

class KanjiCode
{
public:
    enum Type {ASCII, JIS, EUC, SJIS, UNICODE, UTF8 };
    static enum Type judge(const char *str, int length);
    static const int ESC;
    static const int _SS2_;
    static const unsigned char kanji_map_sjis[];
    static int ISkanji(int code)
    {
	if (code >= 0x100)
		    return 0;
	return (kanji_map_sjis[code & 0xff] & 1);
    }

    static int ISkana(int code)
    {
	if (code >= 0x100)
		    return 0;
	return (kanji_map_sjis[code & 0xff] & 2);
    }

};

const int KanjiCode::ESC = 0x1b;
const int KanjiCode::_SS2_ = 0x8e;

const unsigned char KanjiCode::kanji_map_sjis[] =
{
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
    0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
    2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0
};

/*
 * EUC-JP is
 *     [0xa1 - 0xfe][0xa1 - 0xfe]
 *     0x8e[0xa1 - 0xfe](SS2)
 *     0x8f[0xa1 - 0xfe][0xa1 - 0xfe](SS3)
 *
 * Shift_Jis is
 *     [0x81 - 0x9f, 0xe0 - 0xef(0xfe?)][0x40 - 0x7e, 0x80 - 0xfc]
 *
 * Shift_Jis Hankaku Kana is
 *     [0xa1 - 0xdf]
 */

/*
 * KanjiCode::judge() is based on judge_jcode() from jvim
 *     http://hp.vector.co.jp/authors/VA003457/vim/
 *
 * Special Thanks to Kenichi Tsuchida
 */

/*
 * Maybe we should use QTextCodec::heuristicContentMatch()
 * But it fails detection. It's not useful.
 */

enum KanjiCode::Type KanjiCode::judge(const char *str, int size)
{
    enum Type code;
    int i;
    int bfr = FALSE;		/* Kana Moji */
    int bfk = 0;		/* EUC Kana */
    int sjis = 0;
    int euc = 0;

    const unsigned char *ptr = (const unsigned char *) str;

    code = ASCII;

    i = 0;
    while (i < size) {
	if (ptr[i] == ESC && (size - i >= 3)) {
	    if ((ptr[i + 1] == '$' && ptr[i + 2] == 'B')
	    || (ptr[i + 1] == '(' && ptr[i + 2] == 'B')) {
		code = JIS;
		goto breakBreak;
	    } else if ((ptr[i + 1] == '$' && ptr[i + 2] == '@')
		    || (ptr[i + 1] == '(' && ptr[i + 2] == 'J')) {
		code = JIS;
		goto breakBreak;
	    } else if (ptr[i + 1] == '(' && ptr[i + 2] == 'I') {
		code = JIS;
		i += 3;
	    } else if (ptr[i + 1] == ')' && ptr[i + 2] == 'I') {
		code = JIS;
		i += 3;
	    } else {
		i++;
	    }
	    bfr = FALSE;
	    bfk = 0;
	} else {
	    if (ptr[i] < 0x20) {
		bfr = FALSE;
		bfk = 0;
		/* ?? check kudokuten ?? && ?? hiragana ?? */
		if ((i >= 2) && (ptr[i - 2] == 0x81)
			&& (0x41 <= ptr[i - 1] && ptr[i - 1] <= 0x49)) {
		    code = SJIS;
		    sjis += 100;	/* kudokuten */
		} else if ((i >= 2) && (ptr[i - 2] == 0xa1)
			&& (0xa2 <= ptr[i - 1] && ptr[i - 1] <= 0xaa)) {
		    code = EUC;
		    euc += 100;		/* kudokuten */
		} else if ((i >= 2) && (ptr[i - 2] == 0x82) && (0xa0 <= ptr[i - 1])) {
		    sjis += 40;		/* hiragana */
		} else if ((i >= 2) && (ptr[i - 2] == 0xa4) && (0xa0 <= ptr[i - 1])) {
		    euc += 40;	/* hiragana */
		}
	    } else {
		/* ?? check hiragana or katana ?? */
		if ((size - i > 1) && (ptr[i] == 0x82) && (0xa0 <= ptr[i + 1])) {
		    sjis++;	/* hiragana */
		} else if ((size - i > 1) && (ptr[i] == 0x83)
			 && (0x40 <= ptr[i + 1] && ptr[i + 1] <= 0x9f)) {
		    sjis++;	/* katakana */
		} else if ((size - i > 1) && (ptr[i] == 0xa4) && (0xa0 <= ptr[i + 1])) {
		    euc++;	/* hiragana */
		} else if ((size - i > 1) && (ptr[i] == 0xa5) && (0xa0 <= ptr[i + 1])) {
		    euc++;	/* katakana */
		}
		if (bfr) {
		    if ((i >= 1) && (0x40 <= ptr[i] && ptr[i] <= 0xa0) && ISkanji(ptr[i - 1])) {
			code = SJIS;
			goto breakBreak;
		    } else if ((i >= 1) && (0x81 <= ptr[i - 1] && ptr[i - 1] <= 0x9f) && ((0x40 <= ptr[i] && ptr[i] < 0x7e) || (0x7e < ptr[i] && ptr[i] <= 0xfc))) {
			code = SJIS;
			goto breakBreak;
		    } else if ((i >= 1) && (0xfd <= ptr[i] && ptr[i] <= 0xfe) && (0xa1 <= ptr[i - 1] && ptr[i - 1] <= 0xfe)) {
			code = EUC;
			goto breakBreak;
		    } else if ((i >= 1) && (0xfd <= ptr[i - 1] && ptr[i - 1] <= 0xfe) && (0xa1 <= ptr[i] && ptr[i] <= 0xfe)) {
			code = EUC;
			goto breakBreak;
		    } else if ((i >= 1) && (ptr[i] < 0xa0 || 0xdf < ptr[i]) && (0x8e == ptr[i - 1])) {
			code = SJIS;
			goto breakBreak;
		    } else if (ptr[i] <= 0x7f) {
			code = SJIS;
			goto breakBreak;
		    } else {
			if (0xa1 <= ptr[i] && ptr[i] <= 0xa6) {
			    euc++;	/* sjis hankaku kana kigo */
			} else if (0xa1 <= ptr[i] && ptr[i] <= 0xdf) {
			    ;	/* sjis hankaku kana */
			} else if (0xa1 <= ptr[i] && ptr[i] <= 0xfe) {
			    euc++;
			} else if (0x8e == ptr[i]) {
			    euc++;
			} else if (0x20 <= ptr[i] && ptr[i] <= 0x7f) {
			    sjis++;
			}
			bfr = FALSE;
			bfk = 0;
		    }
		} else if (0x8e == ptr[i]) {
		    if (size - i <= 1) {
			;
		    } else if (0xa1 <= ptr[i + 1] && ptr[i + 1] <= 0xdf) {
			/* EUC KANA or SJIS KANJI */
			if (bfk == 1) {
			    euc += 100;
			}
			bfk++;
			i++;
		    } else {
			/* SJIS only */
			code = SJIS;
			goto breakBreak;
		    }
		} else if (0x81 <= ptr[i] && ptr[i] <= 0x9f) {
		    /* SJIS only */
		    code = SJIS;
		    if ((size - i >= 1)
			    && ((0x40 <= ptr[i + 1] && ptr[i + 1] <= 0x7e)
			    || (0x80 <= ptr[i + 1] && ptr[i + 1] <= 0xfc))) {
			goto breakBreak;
		    }
		} else if (0xfd <= ptr[i] && ptr[i] <= 0xfe) {
		    /* EUC only */
		    code = EUC;
		    if ((size - i >= 1)
			    && (0xa1 <= ptr[i + 1] && ptr[i + 1] <= 0xfe)) {
			goto breakBreak;
		    }
		} else if (ptr[i] <= 0x7f) {
		    ;
		} else {
		    bfr = TRUE;
		    bfk = 0;
		}
	    }
	    i++;
	}
    }
    if (code == ASCII) {
	if (sjis > euc) {
	    code = SJIS;
	} else if (sjis < euc) {
	    code = EUC;
	}
    }
breakBreak:
    return (code);
}

Decoder::Decoder() 
{
    _refCount = 1;
    // latin1
    m_codec = QTextCodec::codecForName("iso8859-1");
    m_decoder = m_codec->makeDecoder();
    enc = 0;
    m_type = DefaultEncoding;
    body = false;
    beginning = true;
    visualRTL = false;
}
Decoder::~Decoder()
{
    assert(_refCount == 0);
    delete m_decoder;
}

void Decoder::setEncoding(const char *_encoding, EncodingType type)
{
#ifdef DECODE_DEBUG
    kdDebug(6005) << "setEncoding " << _encoding << " " << force << endl;
#endif
    enc = _encoding;
#ifdef DECODE_DEBUG
    kdDebug(6005) << "old encoding is:" << m_codec->name() << endl;
#endif
    enc = enc.lower();
#ifdef DECODE_DEBUG
    kdDebug(6005) << "requesting:" << enc << endl;
#endif
    if(enc.isNull() || enc.isEmpty())
        return;

#if APPLE_CHANGES
    QTextCodec *codec = (type == EncodingFromMetaTag || type == EncodingFromXMLHeader)
        ? QTextCodec::codecForNameEightBitOnly(enc)
        : QTextCodec::codecForName(enc);
    if (codec) {
        enc = codec->name();
        visualRTL = codec->usesVisualOrdering();
    }
#else
    if(enc == "visual") // hebrew visually ordered
        enc = "iso8859-8";
    bool b;
    QTextCodec *codec = KGlobal::charsets()->codecForName(enc, b);
    if (!b)
        codec = 0;

    if (codec && codec->mibEnum() == 11)  {
        // visually ordered unless one of the following
        if( !(enc == "iso-8859-8-i" || enc == "iso_8859-8-i"
                || enc == "csiso88598i" || enc == "logical") )
            visualRTL = true;
    }
#endif

    if( codec ) { // in case the codec didn't exist, we keep the old one (fixes some sites specifying invalid codecs)
        m_codec = codec;
        m_type = type;
        delete m_decoder;
        m_decoder = m_codec->makeDecoder();
    }
    
#ifdef DECODE_DEBUG
    kdDebug(6005) << "Decoder::encoding used is" << m_codec->name() << endl;
#endif
}

const char *Decoder::encoding() const
{
    return enc;
}

// Other browsers allow comments in the head section, so we need to also.
// It's important not to look for tags inside the comments.
static void skipComment(const char *&ptr, const char *pEnd)
{
    const char *p = ptr;
    // Allow <!-->; other browsers do.
    if (*p == '>') {
        p++;
    } else {
        while (p != pEnd) {
            if (*p == '-') {
                // This is the real end of comment, "-->".
                if (p[1] == '-' && p[2] == '>') {
                    p += 3;
                    break;
                }