/****************************************************************************

** 

**

** Implementation of QTextStream class

**

** Created : 940922

**

** Copyright (C) 1992-2000 Trolltech AS.  All rights reserved.

**

** This file is part of the tools module of the Qt GUI Toolkit.

**

** This file may be distributed under the terms of the Q Public License

** as defined by Trolltech AS of Norway and appearing in the file

** LICENSE.QPL included in the packaging of this file.

**

** This file may be distributed and/or modified under the terms of the

** GNU General Public License version 2 as published by the Free Software

** Foundation and appearing in the file LICENSE.GPL included in the

** packaging of this file.

**

** Licensees holding valid Qt Enterprise Edition or Qt Professional Edition

** licenses may use this file in accordance with the Qt Commercial License

** Agreement provided with the Software.

**

** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE

** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.

**

** See http://www.trolltech.com/pricing.html or email sales@trolltech.com for

**   information about Qt Commercial License Agreements.

** See http://www.trolltech.com/qpl/ for QPL licensing information.

** See http://www.trolltech.com/gpl/ for GPL licensing information.

**

** Contact info@trolltech.com if any conditions of this licensing are

** not clear to you.

**

**********************************************************************/

#include "qtextstream.h"

#ifndef QT_NO_TEXTSTREAM

#include "qtextcodec.h"

#include "qregexp.h"

#include "qbuffer.h"

#include "qfile.h"

#include <stdio.h>

#include <ctype.h>

#include <stdlib.h>

#if defined(_OS_WIN32_)

#include <windows.h>

#endif

// NOT REVISED

/*!

  \class QTextStream qtextstream.h

  \brief The QTextStream class provides basic functions for reading and

  writing text using a QIODevice.

  \ingroup io

  \define endl

  \define bin

  \define oct

  \define dec

  \define hex

  \define flush

  \define ws

  The text stream class has a functional interface that is very

  similar to that of the standard C++ iostream class.  The difference

  between iostream and QTextStream is that our stream operates on a

  QIODevice, which is easily subclassed, while iostream operates on

  FILE * pointers, which can not be subclassed.

  Qt provides several global functions similar to the ones in iostream:

  
 \c bin sets the QTextStream to read/write binary numbers

  
 \c oct sets the QTextStream to read/write octal numbers

  
 \c dec sets the QTextStream to read/write decimal numbers

  
 \c hex sets the QTextStream to read/write hexadecimal numbers

  
 \c endl forces a line break

  
 \c flush forces the QIODevice to flush any buffered data

  
 \c ws eats any available white space (on input)

  
 \c reset resets the QTextStream to its default mode (see reset()).

  \warning By default, QTextStream will automatically detect whether

  integers in the stream are in decimal, octal, hexadecimal or binary

  format when reading from the stream. In particular, a leading '0'

  signifies octal, ie. the sequence "0100" will be interpreted as

  64.

  The QTextStream class reads and writes text and it is not

  appropriate for dealing with binary data (but QDataStream is).

  By default output of Unicode text (ie. QString) is done using the

  local 8-bit encoding.  This can be changed using the setEncoding()

  method.  For input, the QTextStream will auto-detect standard

  Unicode "byte order marked" text files, but otherwise the local

  8-bit encoding is used.

  \sa QDataStream

*/

/*

  \class QTSManip qtextstream.h

  \brief The QTSManip class is an internal helper class for the

  QTextStream.

  It is generally a very bad idea to use this class directly in

  application programs.

  \internal

  This class makes it possible to give the QTextStream function objects

  with arguments, like this:

  \code

    QTextStream cout( stdout, IO_WriteOnly );

    cout << setprecision( 8 );		// QTSManip used here!

    cout << 3.14159265358979323846;

  \endcode

  The setprecision() function returns a QTSManip object.

  The QTSManip object contains a pointer to a member function in

  QTextStream and an integer argument.

  When serializing a QTSManip into a QTextStream, the function

  is executed with the argument.

*/

/*! \fn QTSManip::QTSManip (QTSMFI m, int a)

  Constructs a QTSManip object which will call \a m (a member function

  in QTextStream which accepts a single int) with argument \a a when

  QTSManip::exec() is called.  Used internally in e.g. endl:

  \code

    s << "some text" << endl << "more text";

  \endcode

*/

/*! \fn void QTSManip::exec (QTextStream& s)

  Calls the member function specified in the constructor, for object

  \a s.  Used internally in e.g. endl:

  \code

    s << "some text" << endl << "more text";

  \endcode

*/

/*****************************************************************************

  QTextStream member functions

 *****************************************************************************/

#if defined(CHECK_STATE)

#undef  CHECK_STREAM_PRECOND

#define CHECK_STREAM_PRECOND  if ( !dev ) {				\

				qWarning( "QTextStream: No device" );	\

				return *this; }

#else

#define CHECK_STREAM_PRECOND

#endif

#define I_SHORT		0x0010

#define I_INT		0x0020

#define I_LONG		0x0030

#define I_TYPE_MASK	0x00f0

#define I_BASE_2	QTS::bin

#define I_BASE_8	QTS::oct

#define I_BASE_10	QTS::dec

#define I_BASE_16	QTS::hex

#define I_BASE_MASK	(QTS::bin | QTS::oct | QTS::dec | QTS::hex)

#define I_SIGNED	0x0100

#define I_UNSIGNED	0x0200

#define I_SIGN_MASK	0x0f00

static const QChar QEOF = QChar((ushort)0xffff); //guaranteed not to be a character.

const int QTextStream::basefield   = I_BASE_MASK;

const int QTextStream::adjustfield = ( QTextStream::left |

				       QTextStream::right |

				       QTextStream::internal );

const int QTextStream::floatfield  = ( QTextStream::scientific |

				       QTextStream::fixed );

class QTextStreamPrivate {

public:

#ifndef QT_NO_TEXTCODEC

    QTextStreamPrivate() : decoder( 0 ), sourceType( NotSet ) {}

    ~QTextStreamPrivate() { delete decoder; }

    QTextDecoder *decoder;		//???

#else

    QTextStreamPrivate() : sourceType( NotSet ) {}

    ~QTextStreamPrivate() { }

#endif

    QString ungetcBuf;

    enum SourceType { NotSet, IODevice, String, ByteArray, File };

    SourceType sourceType;

};

// skips whitespace and returns the first non-whitespace character

QChar QTextStream::eat_ws()

{

    QChar c;

    do { c = ts_getc(); } while ( c != QEOF && ts_isspace(c) );

    return c;

}

void QTextStream::init()

{

    // ### ungetcBuf = QEOF;

    dev = 0;					// no device set

    fstrm = owndev = FALSE;

    mapper = 0;

    d = new QTextStreamPrivate;

    doUnicodeHeader = TRUE;	 //default to autodetect

    latin1 = TRUE;		 // ### should use local?

    internalOrder = QChar::networkOrdered(); //default to network order

}

/*!

  Constructs a data stream that has no IO device.

*/

QTextStream::QTextStream()

{

    init();

    setEncoding( Locale ); //###

    reset();

    d->sourceType = QTextStreamPrivate::NotSet;

}

/*!

  Constructs a text stream that uses the IO device \a iod.

*/

QTextStream::QTextStream( QIODevice *iod )

{

    init();

    setEncoding( Locale ); //###

    dev = iod;					// set device

    reset();

    d->sourceType = QTextStreamPrivate::IODevice;

}

// TODO: use special-case handling of this case in QTextStream, and

//	 simplify this class to only deal with QChar or QString data.

class QStringBuffer : public QIODevice {

public:

    QStringBuffer( QString* str );

    ~QStringBuffer();

    bool  open( int m );

    void  close();

    void  flush();

    uint  size() const;

    int   at()   const;

    bool  at( int pos );

    int   readBlock( char *p, uint len );

    int writeBlock( const char *p, uint len );

    int   getch();

    int   putch( int ch );

    int   ungetch( int ch );

protected:

    QString* s;

private:        // Disabled copy constructor and operator=

    QStringBuffer( const QStringBuffer & );

    QStringBuffer &operator=( const QStringBuffer & );

};

QStringBuffer::QStringBuffer( QString* str )

{

    s = str;

}

QStringBuffer::~QStringBuffer()

{

}

bool QStringBuffer::open( int m )

{

    if ( !s ) {

#if defined(CHECK_STATE)

	qWarning( "QStringBuffer::open: No string" );

#endif

	return FALSE;

    }

    if ( isOpen() ) {                           // buffer already open

#if defined(CHECK_STATE)

	qWarning( "QStringBuffer::open: Buffer already open" );

#endif

	return FALSE;

    }

    setMode( m );

    if ( m & IO_Truncate ) {                    // truncate buffer

	s->truncate( 0 );

    }

    if ( m & IO_Append ) {                      // append to end of buffer

	ioIndex = s->length()*(int)sizeof(QChar);

    } else {

	ioIndex = 0;

    }

    setState( IO_Open );

    setStatus( 0 );

    return TRUE;

}

void QStringBuffer::close()

{

    if ( isOpen() ) {

	setFlags( IO_Direct );

	ioIndex = 0;

    }

}

void QStringBuffer::flush()

{

}

uint QStringBuffer::size() const

{

    return s ? s->length()*(int)sizeof(QChar) : 0;

}

int  QStringBuffer::at()   const

{

    return ioIndex;

}

bool QStringBuffer::at( int pos )

{

#if defined(CHECK_STATE)

    if ( !isOpen() ) {

	qWarning( "QStringBuffer::at: Buffer is not open" );

	return FALSE;

    }

#endif

    if ( (uint)pos >= s->length()*2 ) {

#if defined(CHECK_RANGE)

	qWarning( "QStringBuffer::at: Index %d out of range", pos );

#endif

	return FALSE;

    }

    ioIndex = pos;

    return TRUE;

}

int  QStringBuffer::readBlock( char *p, uint len )

{

#if defined(CHECK_STATE)

    CHECK_PTR( p );

    if ( !isOpen() ) {                          // buffer not open

	qWarning( "QStringBuffer::readBlock: Buffer not open" );

	return -1;

    }

    if ( !isReadable() ) {                      // reading not permitted

	qWarning( "QStringBuffer::readBlock: Read operation not permitted" );

	return -1;

    }

#endif

    if ( (uint)ioIndex + len > s->length()*sizeof(QChar) ) {

	// overflow

	if ( (uint)ioIndex >= s->length()*sizeof(QChar) ) {

	    setStatus( IO_ReadError );

	    return -1;

	} else {

	    len = s->length()*2 - (uint)ioIndex;

	}

    }

    memcpy( p, ((const char*)(s->unicode()))+ioIndex, len );

    ioIndex += len;

    return len;

}

int QStringBuffer::writeBlock( const char *p, uint len )

{

#if defined(CHECK_NULL)

    if ( p == 0 && len != 0 )

	qWarning( "QStringBuffer::writeBlock: Null pointer error" );

#endif

#if defined(CHECK_STATE)

    if ( !isOpen() ) {                          // buffer not open

	qWarning( "QStringBuffer::writeBlock: Buffer not open" );

	return -1;

    }

    if ( !isWritable() ) {                      // writing not permitted

	qWarning( "QStringBuffer::writeBlock: Write operation not permitted" );

	return -1;

    }

    if ( ioIndex&1 ) {

	qWarning( "QStringBuffer::writeBlock: non-even index - non Unicode" );

	return -1;

    }

    if ( len&1 ) {

	qWarning( "QStringBuffer::writeBlock: non-even length - non Unicode" );

	return -1;

    }

#endif

    s->replace(ioIndex/2, len/2, (QChar*)p, len/2);

    ioIndex += len;

    return len;

}

int QStringBuffer::getch()

{

#if defined(CHECK_STATE)

    if ( !isOpen() ) {                          // buffer not open

	qWarning( "QStringBuffer::getch: Buffer not open" );

	return -1;

    }

    if ( !isReadable() ) {                      // reading not permitted

	qWarning( "QStringBuffer::getch: Read operation not permitted" );

	return -1;

    }

#endif

    if ( (uint)ioIndex >= s->length()*2 ) {           // overflow

	setStatus( IO_ReadError );

	return -1;

    }

    return *((char*)s->unicode() + ioIndex++);

}

int QStringBuffer::putch( int ch )

{

    char c = ch;

    if ( writeBlock(&c,1) < 0 )

	return -1;

    else

	return ch;

}

int QStringBuffer::ungetch( int ch )

{

#if defined(CHECK_STATE)

    if ( !isOpen() ) {                          // buffer not open

	qWarning( "QStringBuffer::ungetch: Buffer not open" );

	return -1;

    }

    if ( !isReadable() ) {                      // reading not permitted

	qWarning( "QStringBuffer::ungetch: Read operation not permitted" );

	return -1;

    }

#endif

    if ( ch != -1 ) { // something to do with eof

	if ( ioIndex )

	    ioIndex--;

	else

	    ch = -1;

    }

    return ch;

}

/*!

  Constructs a text stream that operates on a Unicode QString through an

  internal device.

  If you set an encoding or codec with setEncoding() or setCodec(), this

  setting is ignored for text streams that operate on QString.

  Example:

  \code

    QString str;

    QTextStream ts( &str, IO_WriteOnly );

    ts << "pi = " << 3.14;			// str == "pi = 3.14"

  \endcode

  Writing data to the text stream will modify the contents of the string.

  The string will be expanded when data is written beyond the end of the

  string. Note that the string will not be truncated:

  \code

    QString str = "pi = 3.14";

    QTextStream ts( &str, IO_WriteOnly );

    ts <<  "2+2 = " << 2+2; 		// str == "2+2 = 414"

  \endcode

  Note that since QString is Unicode, you should not use readRawBytes()

  or writeRawBytes() on such a stream.

*/

QTextStream::QTextStream( QString* str, int filemode )

{

    // TODO: optimize for this case as it becomes more common

    //        (see QStringBuffer above)

    init();

    dev = new QStringBuffer( str );

    ((QStringBuffer *)dev)->open( filemode );

    owndev = TRUE;

    setEncoding(RawUnicode);

    reset();

    d->sourceType = QTextStreamPrivate::String;

}

/*! \obsolete

  This constructor is equivalent to the constructor taking a QString*

  parameter.

*/

QTextStream::QTextStream( QString& str, int filemode )

{

    init();

    dev = new QStringBuffer( &str );

    ((QStringBuffer *)dev)->open( filemode );

    owndev = TRUE;

    setEncoding(RawUnicode);

    reset();

    d->sourceType = QTextStreamPrivate::String;

}

/*!

  Constructs a text stream that operates on a byte array through an

  internal QBuffer device.

  Example:

  \code

    QByteArray array;

    QTextStream ts( array, IO_WriteOnly );

    ts << "pi = " << 3.14 << '\0';		// array == "pi = 3.14"

  \endcode

  Writing data to the text stream will modify the contents of the array.

  The array will be expanded when data is written beyond the end of the

  string.

  Same example, using a QBuffer:

  \code

    QByteArray array;

    QBuffer buf( array );

    buf.open( IO_WriteOnly );

    QTextStream ts( &buf );

    ts << "pi = " << 3.14 << '\0';		// array == "pi = 3.14"

    buf.close();

  \endcode

*/

QTextStream::QTextStream( QByteArray a, int mode )

{

    init();

    dev = new QBuffer( a );

    ((QBuffer *)dev)->open( mode );

    owndev = TRUE;

    setEncoding( Locale ); //### Locale???

    reset();

    d->sourceType = QTextStreamPrivate::ByteArray;

}

/*!

  Constructs a text stream that operates on an existing file handle \e fh

  through an internal QFile device.

  Example:

  \code

    QTextStream cout( stdout, IO_WriteOnly );

    QTextStream cin ( stdin,  IO_ReadOnly );

    QTextStream cerr( stderr, IO_WriteOnly );

 \endcode

*/

QTextStream::QTextStream( FILE *fh, int mode )

{

    init();

    setEncoding( Locale ); //###

    dev = new QFile;

    ((QFile *)dev)->open( mode, fh );

    fstrm = owndev = TRUE;

    reset();

    d->sourceType = QTextStreamPrivate::File;

}

/*!

  Destructs the text stream.

  The destructor does not affect the current IO device.

*/

QTextStream::~QTextStream()

{

    if ( owndev )

	delete dev;

    delete d;

}

/*!

  Positions the read pointer at the first non-whitespace character.

*/

void QTextStream::skipWhiteSpace()

{

    ts_ungetc( eat_ws() );

}

/*!

  \fn Encoding QTextStream::encoding() const

  Returns the encoding mode of the stream.

  \sa setEncoding()

*/

/*!

  Tries to read len characters from the stream and stores them in \a buf.

  Returns the number of characters really read.

  Attention: There will no QEOF appended if the read reaches the end of

  the file. EOF is reached when the return value does not equal \a len.

*/

uint QTextStream::ts_getbuf( QChar* buf, uint len )

{

    if( len<1 )

	return 0;

    uint rnum=0;   // the number of QChars really read

    if ( d && d->ungetcBuf.length() ) {

	while( rnum < len && rnum < d->ungetcBuf.length() ) {

	    buf[rnum] = d->ungetcBuf.constref(rnum);

	    rnum++;

	}

	d->ungetcBuf = d->ungetcBuf.mid( rnum );

	if ( rnum >= len )

	    return rnum;

    }

    // we use dev->ungetch() for one of the bytes of the unicode

    // byte-order mark, but a local unget hack for the other byte:

    int ungetHack = EOF;

    if ( doUnicodeHeader ) {

	doUnicodeHeader = FALSE; //only at the top

	int c1 = dev->getch();

	if ( c1 == EOF )

	    return rnum;

	int c2 = dev->getch();

	if ( c1 == 0xfe && c2 == 0xff ) {

	    mapper = 0;

	    latin1 = FALSE;

	    internalOrder = QChar::networkOrdered();   //network order

	} else if ( c1 == 0xff && c2 == 0xfe ) {

	    mapper = 0;

	    latin1 = FALSE;

	    internalOrder = !QChar::networkOrdered();   //reverse network order

	} else {

	    if ( c2 != EOF ) {

	 	dev->ungetch( c2 );

		ungetHack = c1;

	    } else {

	 	dev->ungetch( c1 );

		// note that a small possible bug might hide here

		// here, if only the first byte of a file has made it

		// so far, and that first byte is half of the

		// byte-order mark, then the utfness will not be

		// detected.  whether or not this is a bug depends on

		// taste.  I can't really decide.

	    }

	}

    }

#ifndef QT_NO_TEXTCODEC

    if ( mapper ) {

	bool shortRead = FALSE;

	if ( !d->decoder )

	    d->decoder = mapper->makeDecoder();

	while( rnum < len ) {

	    QString s;

	    bool readBlock = !( len == 1+rnum );

	    while ( TRUE ) {

		// for efficiency: normally read a whole block

		if ( readBlock ) {

		    // guess buffersize; this may be wrong (too small or too

		    // big). But we can handle this (either iterate reading

		    // or use ungetcBuf).

		    // Note that this might cause problems for codecs where

		    // one byte can result in >1 Unicode Characters if bytes

		    // are written to the stream in the meantime (loss of

		    // synchronicity).

		    uint rlen = len - rnum;

		    char *cbuf = new char[ rlen ];

		    if ( ungetHack != EOF ) {

			rlen = 1+dev->readBlock( cbuf+1, rlen-1 );

			cbuf[0] = (char)ungetHack;

			ungetHack = EOF;

		    } else {

			rlen = dev->readBlock( cbuf, rlen );

		    }

		    s  += d->decoder->toUnicode( cbuf, rlen );

		    delete[] cbuf;

		    // use buffered reading only for the first time, because we

		    // have to get the stream synchronous again (this is easier

		    // with single character reading)

		    readBlock = FALSE;

		}

		// get stream (and codec) in sync

		int c;

		if ( ungetHack == EOF ) {

		    c = dev->getch();

		} else {

		    c = ungetHack;

		    ungetHack = EOF;

		}

		if ( c == EOF ) {

		    shortRead = TRUE;

		    break;

		}

		char b = c;

		uint lengthBefore = s.length();

		s  += d->decoder->toUnicode( &b, 1 );

		if ( s.length() > lengthBefore )

		    break; // it seems we are in sync now

	    }

	    uint i = 0;

	    while( rnum < len && i < s.length() )

		buf[rnum++] = s.constref(i++);

	    if ( s.length() > i )

		// could be = but append is clearer

		d->ungetcBuf.append( s.mid( i ) );

	    if ( shortRead )

		return rnum;

	}

    } else

#endif

    if ( latin1 ) {

	if ( len == 1+rnum ) {

	    // use this method for one character because it is more efficient

	    // (arnt doubts whether it makes a difference, but lets it stand)

	    int c = (ungetHack == EOF) ? dev->getch() : ungetHack;

	    if ( c != EOF )

		buf[rnum++] = (char)c;

	} else {

	    if ( ungetHack != EOF ) {

		buf[rnum++] = (char)ungetHack;

		ungetHack = EOF;

	    }

	    char *cbuf = new char[len - rnum];

	    while ( !dev->atEnd() && rnum < len ) {

		uint rlen = len - rnum;

		rlen = dev->readBlock( cbuf, rlen );

		uint i = 0;

		while( i < rlen )

		    buf[rnum++] = cbuf[i++];

	    }

	    delete[] cbuf;

	}

    } else { // UCS-2 or UTF-16

	if ( len == 1+rnum ) {

	    int c1 = (ungetHack == EOF) ? dev->getch() : ungetHack;

	    if ( c1 == EOF )

		return rnum;

	    int c2 = dev->getch();

	    if ( c2 == EOF )

		return rnum;

	    if ( isNetworkOrder() )

		buf[rnum++] = QChar( c2, c1 );

	    else

		buf[rnum++] = QChar( c1, c2 );

	} else {

	    char *cbuf = new char[ 2*( len - rnum ) ]; // for paranoids: overflow possible

	    while ( !dev->atEnd() && rnum < len ) {

		uint rlen = 2 * ( len-rnum );

		if ( ungetHack != EOF ) {

		    rlen = 1+dev->readBlock( cbuf+1, rlen-1 );

		    cbuf[0] = (char)ungetHack;

		    ungetHack = EOF;

		} else {

		    rlen = dev->readBlock( cbuf, rlen );

		}

		// We can't use an odd number of bytes, so put it back. But

		// do it only if we are capable of reading more -- normally

		// there should not be an odd number, but the file might be

		// truncated or not in UTF-16...

		if ( (rlen & 1) == 1 )

		    if ( !dev->atEnd() )

			dev->ungetch( cbuf[--rlen] );

		uint i = 0;

		if ( isNetworkOrder() ) {

		    while( i < rlen ) {

			buf[rnum++] = QChar( cbuf[i+1], cbuf[i] );

			i+=2;

		    }

		} else {

		    while( i < rlen ) {

			buf[rnum++] = QChar( cbuf[i], cbuf[i+1] );

			i+=2;

		    }

		}

	    }

	    delete[] cbuf;

	}

    }

    return rnum;

}

/*!

  Puts one character to the stream.

*/

void QTextStream::ts_putc( QChar c )

{

#ifndef QT_NO_TEXTCODEC

    if ( mapper ) {

	int len = 1;

	QString s = c;

	QCString block = mapper->fromUnicode( s, len );

	dev->writeBlock( block, len );

    } else

#endif

    if ( latin1 ) {

	if( c.row() )

	    dev->putch( '?' ); //######unknown character???

	else

	    dev->putch( c.cell() );

    } else {

	if ( doUnicodeHeader ) {

	    doUnicodeHeader = FALSE;

	    ts_putc( QChar::byteOrderMark );

	}

	if ( internalOrder ) {

	    dev->writeBlock( (char*)&c, sizeof(QChar) );

	} else if ( isNetworkOrder() ) {

	    dev->putch(c.row());

	    dev->putch(c.cell());

	} else {

	    dev->putch(c.cell());

	    dev->putch(c.row());

	}

    }

}

/*!

  Puts one character to the stream.

*/

void QTextStream::ts_putc(int ch)

{

    ts_putc(QChar((ushort)ch));

}

bool QTextStream::ts_isdigit(QChar c)

{

    return c.isDigit();

}

bool QTextStream::ts_isspace( QChar c )

{

    return c.isSpace();

}

void QTextStream::ts_ungetc( QChar c )

{

    if ( c.unicode() == 0xffff )

	return;

    d->ungetcBuf.prepend( c );

}

/*!

  Reads \e len bytes from the stream into \e e s and returns a reference to

  the stream.

  The buffer \e s must be preallocated.

  Note that no encoding is done by this function.

  \warning The behaviour of this function is undefined unless the

  stream's encoding is set to Unicode or Latin1.