strbuffer.cpp

ncbi源码

/*
 * ===========================================================================
 * PRODUCTION $Log: strbuffer.cpp,v $
 * PRODUCTION Revision 1000.3  2004/06/01 19:40:32  gouriano
 * PRODUCTION PRODUCTION: UPGRADED [GCC34_MSVC7] Dev-tree R1.42
 * PRODUCTION
 * ===========================================================================
 */

/*  $Id: strbuffer.cpp,v 1000.3 2004/06/01 19:40:32 gouriano Exp $
* ===========================================================================
*
*                            PUBLIC DOMAIN NOTICE
*               National Center for Biotechnology Information
*
*  This software/database is a "United States Government Work" under the
*  terms of the United States Copyright Act.  It was written as part of
*  the author's official duties as a United States Government employee and
*  thus cannot be copyrighted.  This software/database is freely available
*  to the public for use. The National Library of Medicine and the U.S.
*  Government have not placed any restriction on its use or reproduction.
*
*  Although all reasonable efforts have been taken to ensure the accuracy
*  and reliability of the software and data, the NLM and the U.S.
*  Government do not and cannot warrant the performance or results that
*  may be obtained by using this software or data. The NLM and the U.S.
*  Government disclaim all warranties, express or implied, including
*  warranties of performance, merchantability or fitness for any particular
*  purpose.
*
*  Please cite the author in any work or product based on this material.
*
* ===========================================================================
*
* Author: Eugene Vasilchenko
*
* File Description:
*   Input buffer
*
* ---------------------------------------------------------------------------
* $Log: strbuffer.cpp,v $
* Revision 1000.3  2004/06/01 19:40:32  gouriano
* PRODUCTION: UPGRADED [GCC34_MSVC7] Dev-tree R1.42
*
* Revision 1.42  2004/05/24 18:13:01  gouriano
* In text output files make indentation optional
*
* Revision 1.41  2004/05/17 21:06:02  gorelenk
* Added include of PCH ncbi_pch.hpp
*
* Revision 1.40  2004/05/10 14:51:13  gouriano
* Work with unsigned number (division/modulus ops) when writing an integer
*
* Revision 1.39  2003/12/31 20:51:33  gouriano
* added possibility to seek (when possible) in CByteSourceReader
*
* Revision 1.38  2003/11/19 15:41:11  vasilche
* Pushback unused data in destructor of CIStreamBuffer.
*
* Revision 1.37  2003/09/25 12:49:14  kuznets
* Change to allowed more than one subsource collector
*
* Revision 1.36  2003/03/21 17:15:36  vasilche
* Avoid unnecessary buffer growth in GetChars().
*
* Revision 1.35  2003/02/26 21:32:00  gouriano
* modify C++ exceptions thrown by this library
*
* Revision 1.34  2002/11/18 19:49:37  grichenk
* More details in error messages
*
* Revision 1.33  2002/11/04 21:29:22  grichenk
* Fixed usage of const CRef<> and CRef<> constructor
*
* Revision 1.32  2002/10/22 20:22:55  gouriano
* undo the prev change
*
* Revision 1.31  2002/10/22 19:01:36  gouriano
* replaced THROW0_TRACE by throw in CIStreamBuffer::FillBuffer
*
* Revision 1.30  2002/01/29 16:01:21  grichenk
* COStreamBuffer destructor fixed - no exceptions thrown
*
* Revision 1.29  2001/08/08 18:35:55  grichenk
* CIStreamBuffer::FindChar() -- fixed bug with buffer pointers
*
* Revision 1.28  2001/06/20 17:36:19  grichenk
* Updated FillBuffer() to work with GCC 3.0
*
* Revision 1.27  2001/06/06 15:50:01  grichenk
* Fixed auto-wrapping of long lines
*
* Revision 1.26  2001/05/17 15:07:15  lavr
* Typos corrected
*
* Revision 1.25  2001/05/11 13:58:52  grichenk
* Removed "while" reading loop in CIStreamBuffer::FillBuffer()
*
* Revision 1.24  2001/04/17 21:47:29  vakatov
* COStreamBuffer::Flush() -- try to flush the underlying output stream
* regardless of its state (clear the state temporarily before flushing,
* restore it afterwards).
*
* Revision 1.23  2001/03/14 17:59:26  vakatov
* COStreamBuffer::  renamed GetFreeSpace() -> GetAvailableSpace()
* to avoid clash with MS-Win system headers' #define
*
* Revision 1.22  2001/01/05 20:09:05  vasilche
* Added util directory for various algorithms and utility classes.
*
* Revision 1.21  2000/12/15 15:38:46  vasilche
* Added support of Int8 and long double.
* Enum values now have type Int4 instead of long.
*
* Revision 1.20  2000/10/20 15:51:44  vasilche
* Fixed data error processing.
* Added interface for constructing container objects directly into output stream.
* object.hpp, object.inl and object.cpp were split to
* objectinfo.*, objecttype.*, objectiter.* and objectio.*.
*
* Revision 1.19  2000/10/13 20:59:21  vasilche
* Avoid using of ssize_t absent on some compilers.
*
* Revision 1.18  2000/10/13 20:22:57  vasilche
* Fixed warnings on 64 bit compilers.
* Fixed missing typename in templates.
*
* Revision 1.17  2000/08/15 19:44:51  vasilche
* Added Read/Write hooks:
* CReadObjectHook/CWriteObjectHook for objects of specified type.
* CReadClassMemberHook/CWriteClassMemberHook for specified members.
* CReadChoiceVariantHook/CWriteChoiceVariant for specified choice variants.
* CReadContainerElementHook/CWriteContainerElementsHook for containers.
*
* Revision 1.16  2000/07/03 18:42:47  vasilche
* Added interface to typeinfo via CObjectInfo and CConstObjectInfo.
* Reduced header dependency.
*
* Revision 1.15  2000/06/16 20:01:26  vasilche
* Avoid use of unexpected_exception() which is unimplemented on Mac.
*
* Revision 1.14  2000/06/01 19:07:05  vasilche
* Added parsing of XML data.
*
* Revision 1.13  2000/05/24 20:08:50  vasilche
* Implemented XML dump.
*
* Revision 1.12  2000/05/03 14:38:14  vasilche
* SERIAL: added support for delayed reading to generated classes.
* DATATOOL: added code generation for delayed reading.
*
* Revision 1.11  2000/04/28 16:58:14  vasilche
* Added classes CByteSource and CByteSourceReader for generic reading.
* Added delayed reading of choice variants.
*
* Revision 1.10  2000/04/13 14:50:28  vasilche
* Added CObjectIStream::Open() and CObjectOStream::Open() for easier use.
*
* Revision 1.9  2000/04/06 16:11:01  vasilche
* Fixed bug with iterators in choices.
* Removed unneeded calls to ReadExternalObject/WriteExternalObject.
* Added output buffering to text ASN.1 data.
*
* Revision 1.8  2000/03/29 15:55:29  vasilche
* Added two versions of object info - CObjectInfo and CConstObjectInfo.
* Added generic iterators by class -
* 	CTypeIterator<class>, CTypeConstIterator<class>,
* 	CStdTypeIterator<type>, CStdTypeConstIterator<type>,
* 	CObjectsIterator and CObjectsConstIterator.
*
* Revision 1.7  2000/03/10 21:16:47  vasilche
* Removed EOF workaround code.
*
* Revision 1.6  2000/03/10 17:59:21  vasilche
* Fixed error reporting.
* Added EOF bug workaround on MIPSpro compiler (not finished).
*
* Revision 1.5  2000/03/07 14:06:24  vasilche
* Added stream buffering to ASN.1 binary input.
* Optimized class loading/storing.
* Fixed bugs in processing OPTIONAL fields.
* Added generation of reference counted objects.
*
* Revision 1.4  2000/02/17 20:02:45  vasilche
* Added some standard serialization exceptions.
* Optimized text/binary ASN.1 reading.
* Fixed wrong encoding of StringStore in ASN.1 binary format.
* Optimized logic of object collection.
*
* Revision 1.3  2000/02/11 17:10:25  vasilche
* Optimized text parsing.
*
* Revision 1.2  2000/02/02 19:07:41  vasilche
* Added THROWS_NONE to constructor/destructor of exception.
*
* Revision 1.1  2000/02/01 21:47:23  vasilche
* Added CGeneratedChoiceTypeInfo for generated choice classes.
* Added buffering to CObjectIStreamAsn.
* Removed CMemberInfo subclasses.
* Added support for DEFAULT/OPTIONAL members.
*
*
* ===========================================================================
*/

#include <ncbi_pch.hpp>
#include <corelib/ncbistre.hpp>
#include <corelib/ncbi_limits.hpp>
#include <util/strbuffer.hpp>
#include <util/bytesrc.hpp>
#include <algorithm>

BEGIN_NCBI_SCOPE

static const size_t KInitialBufferSize = 4096;

static inline
size_t BiggerBufferSize(size_t size) THROWS1_NONE
{
    return size * 2;
}

CIStreamBuffer::CIStreamBuffer(void)
    THROWS1((bad_alloc))
    : m_Error(0), m_BufferOffset(0),
      m_BufferSize(KInitialBufferSize), m_Buffer(new char[KInitialBufferSize]),
      m_CurrentPos(m_Buffer), m_DataEndPos(m_Buffer),
      m_Line(1),
      m_CollectPos(0)
{
}

CIStreamBuffer::~CIStreamBuffer(void)
{
    try {
        Close();
    }
    catch ( exception& exc ) {
        ERR_POST(Warning <<
                 "~CIStreamBuffer: exception while closing: " << exc.what());
    }
    delete[] m_Buffer;
    
}

void CIStreamBuffer::Open(CByteSourceReader& reader)
{
    Close();
    m_Input = &reader;
    m_Error = 0;
}

void CIStreamBuffer::Close(void)
{
    if ( m_Input ) {
        size_t unused = m_DataEndPos - m_CurrentPos;
        if ( unused ) {
            m_Input->Pushback(m_CurrentPos, unused);
        }
        m_Input.Reset();
    }
    m_BufferOffset = 0;
    m_CurrentPos = m_Buffer;
    m_DataEndPos = m_Buffer;
    m_Line = 1;
    m_Error = 0;
}

void CIStreamBuffer::StartSubSource(void)
{
    _ASSERT(!m_CollectPos);

    m_CollectPos = m_CurrentPos;
    m_Collector = m_Input->SubSource(m_DataEndPos - m_CurrentPos, m_Collector);
}

CRef<CByteSource> CIStreamBuffer::EndSubSource(void)
{
    _ASSERT(m_Collector);
    _ASSERT(m_CollectPos);

    _ASSERT(m_CollectPos <= m_CurrentPos);
    if ( m_CurrentPos != m_CollectPos )
        m_Collector->AddChunk(m_CollectPos, m_CurrentPos - m_CollectPos);

    CRef<CByteSource> source = m_Collector->GetSource();

    m_CollectPos = 0;
    m_Collector.Reset();

    return source;
}

// this method is highly optimized
char CIStreamBuffer::SkipSpaces(void)
    THROWS1((CIOException))
{
    // cache pointers
    char* pos = m_CurrentPos;
    char* end = m_DataEndPos;
    // make sure thire is at least one char in buffer
    if ( pos == end ) {
        // fill buffer
        pos = FillBuffer(pos);
        // cache m_DataEndPos
        end = m_DataEndPos;
    }
    // main cycle
    // at the beginning:
    //     pos == m_CurrentPos
    //     end == m_DataEndPos
    //     pos < end
    for (;;) {
        // we use do{}while() cycle because
        // condition is true at the beginning ( pos < end )
        do {
            // cache current char
            char c = *pos;
            if ( c != ' ' ) { // it's not space (' ')
                // point m_CurrentPos to first non space char
                m_CurrentPos = pos;
                // return char value
                return c;
            }
            // skip space char
        } while ( ++pos < end );
        // here pos == end == m_DataEndPos
        // point m_CurrentPos to end of buffer
        m_CurrentPos = pos;
        // fill next portion
        pos = FillBuffer(pos);
        // cache m_DataEndPos
        end = m_DataEndPos;
    }
}

// this method is highly optimized
void CIStreamBuffer::FindChar(char c)
    THROWS1((CIOException))
{
    // cache pointers
    char* pos = m_CurrentPos;
    char* end = m_DataEndPos;
    // make sure thire is at least one char in buffer
    if ( pos == end ) {
        // fill buffer
        pos = FillBuffer(pos);
        // cache m_DataEndPos
        end = m_DataEndPos;
    }
    // main cycle
    // at the beginning:
    //     pos == m_CurrentPos
    //     end == m_DataEndPos
    //     pos < end
    for (;;) {
        char* found = static_cast<char*>(memchr(pos, c, end - pos));
        if ( found ) {
            m_CurrentPos = found;
            return;
        }
        // point m_CurrentPos to end of buffer
        m_CurrentPos = end;
        // fill next portion
        pos = FillBuffer(end);
        // cache m_DataEndPos
        end = m_DataEndPos;
    }
}

// this method is highly optimized
size_t CIStreamBuffer::PeekFindChar(char c, size_t limit)
    THROWS1((CIOException))
{
    _ASSERT(limit > 0);
    PeekCharNoEOF(limit - 1);
    // cache pointers
    char* pos = m_CurrentPos;
    size_t bufferSize = m_DataEndPos - pos;
    if ( bufferSize != 0 ) {
        char* found =
            static_cast<char*>(memchr(pos, c, min(limit, bufferSize)));
        if ( found )
            return found - pos;
    }
    return limit;
}

char* CIStreamBuffer::FillBuffer(char* pos, bool noEOF)
    THROWS1((CIOException, bad_alloc))
{
    _ASSERT(pos >= m_DataEndPos);
    // remove unused portion of buffer at the beginning
    _ASSERT(m_CurrentPos >= m_Buffer);
    size_t erase = m_CurrentPos - m_Buffer;
    if ( erase > 0 ) {
        char* newPos = m_CurrentPos - erase;
        if ( m_Collector ) {
            _ASSERT(m_CollectPos);
            size_t count = m_CurrentPos - m_CollectPos;
            if ( count > 0 )
                m_Collector->AddChunk(m_CollectPos, count);
            m_CollectPos = newPos;
        }
        size_t copy_count = m_DataEndPos - m_CurrentPos;
        if ( copy_count )
            memmove(newPos, m_CurrentPos, copy_count);
        m_CurrentPos = newPos;
        m_DataEndPos -= erase;
        m_BufferOffset += erase;
        pos -= erase;
    }
    size_t dataSize = m_DataEndPos - m_Buffer;
    size_t newPosOffset = pos - m_Buffer;
    if ( newPosOffset >= m_BufferSize ) {
        // reallocate buffer
        size_t newSize = BiggerBufferSize(m_BufferSize);
        while ( newPosOffset >= newSize ) {
            newSize = BiggerBufferSize(newSize);
        }
        char* newBuffer = new char[newSize];
        memcpy(newBuffer, m_Buffer, dataSize);
        m_CurrentPos = newBuffer + (m_CurrentPos - m_Buffer);
        if ( m_CollectPos )
            m_CollectPos = newBuffer + (m_CollectPos - m_Buffer);
        pos = newBuffer + newPosOffset;
        m_DataEndPos = newBuffer + dataSize;
        delete[] m_Buffer;
        m_Buffer = newBuffer;
        m_BufferSize = newSize;
    }
    size_t load = m_BufferSize - dataSize;
    while ( load > 0  &&  pos >= m_DataEndPos ) {
        size_t count = m_Input->Read(m_DataEndPos, load);
        if ( count == 0 ) {
            if ( pos < m_DataEndPos )
                return pos;
            if ( m_Input->EndOfData() ) {
                if ( noEOF ) {
                    // ignore EOF
                    _ASSERT(m_Buffer <= m_CurrentPos);
                    _ASSERT(m_CurrentPos <= pos);
                    _ASSERT(m_DataEndPos <= m_Buffer + m_BufferSize);
                    _ASSERT(!m_CollectPos ||
                            (m_CollectPos>=m_Buffer &&
                             m_CollectPos<=m_CurrentPos));
                    return pos;
                }
                m_Error = "end of file";
//                THROW0_TRACE(CEofException());
                NCBI_THROW(CEofException,eEof,m_Error);
            }
            else {
                m_Error = "read fault";
//                THROW1_TRACE(CIOException, "read fault");
                NCBI_THROW(CIOException,eRead,m_Error);
            }
        }
        m_DataEndPos += count;
        load -= count;
    }
    _ASSERT(m_Buffer <= m_CurrentPos);
    _ASSERT(m_CurrentPos <= pos);
    _ASSERT(pos < m_DataEndPos);
    _ASSERT(m_DataEndPos <= m_Buffer + m_BufferSize);
    _ASSERT(!m_CollectPos || (m_CollectPos>=m_Buffer && m_CollectPos<=m_CurrentPos));
    return pos;
}

char CIStreamBuffer::FillBufferNoEOF(char* pos)
    THROWS1((CIOException, bad_alloc))
{
    pos = FillBuffer(pos, true);
    if ( pos >= m_DataEndPos )
        return 0;
    else
        return *pos;
}

void CIStreamBuffer::GetChars(char* buffer, size_t count)
    THROWS1((CIOException))
{
    // cache pos
    char* pos = m_CurrentPos;
    for ( ;; ) {
        size_t c = m_DataEndPos - pos;
        if ( c >= count ) {
            // all data is already in buffer -> copy it
            memcpy(buffer, pos, count);
            m_CurrentPos = pos + count;
            return;
        }
        else {
            memcpy(buffer, pos, c);
            buffer += c;
            count -= c;
            m_CurrentPos = pos += c;
            pos = FillBuffer(pos);
        }
    }
}

void CIStreamBuffer::GetChars(size_t count)
    THROWS1((CIOException))