pjs/mailnews/base/util/nsMsgLineBuffer.cpp

/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 2 -*-
 *
 * The contents of this file are subject to the Netscape Public
 * License Version 1.1 (the "License"); you may not use this file
 * except in compliance with the License. You may obtain a copy of
 * the License at http://www.mozilla.org/NPL/
 *
 * Software distributed under the License is distributed on an "AS
 * IS" basis, WITHOUT WARRANTY OF ANY KIND, either express or
 * implied. See the License for the specific language governing
 * rights and limitations under the License.
 *
 * The Original Code is mozilla.org code.
 *
 * The Initial Developer of the Original Code is Netscape
 * Communications Corporation.  Portions created by Netscape are
 * Copyright (C) 1999 Netscape Communications Corporation. All
 * Rights Reserved.
 *
 * Contributor(s): 
 */

#include "msgCore.h"
#include "prlog.h"
#include "nsMsgLineBuffer.h"

#include "nsIInputStream.h" // used by nsMsgLineStreamBuffer

MOZ_DECL_CTOR_COUNTER(nsByteArray);

nsByteArray::nsByteArray()
{
	MOZ_COUNT_CTOR(nsByteArray);
	m_buffer = NULL;
	m_bufferSize = 0;
	m_bufferPos = 0;
}

nsByteArray::~nsByteArray()
{
	MOZ_COUNT_DTOR(nsByteArray);
	PR_FREEIF(m_buffer);
}

nsresult nsByteArray::GrowBuffer(PRUint32 desired_size, PRUint32 quantum)
{
	if (m_bufferSize < desired_size)
	{
		char *new_buf;
		PRUint32 increment = desired_size - m_bufferSize;
		if (increment < quantum) /* always grow by a minimum of N bytes */
			increment = quantum;


		new_buf = (m_buffer
				 ? (char *) PR_REALLOC (m_buffer, (m_bufferSize + increment))
				 : (char *) PR_MALLOC (m_bufferSize + increment));
		if (! new_buf)
			return NS_ERROR_OUT_OF_MEMORY;
		m_buffer = new_buf;
		m_bufferSize += increment;
	}
  return 0;
}

nsresult nsByteArray::AppendString(const char *string)
{
	PRUint32 strLength = (string) ? PL_strlen(string) : 0;
	return AppendBuffer(string, strLength);

}

nsresult nsByteArray::AppendBuffer(const char *buffer, PRUint32 length)
{
	nsresult ret = NS_OK;
	if (m_bufferPos + length > m_bufferSize)
		ret = GrowBuffer(m_bufferPos + length, 1024);
	if (ret == NS_OK)
	{
		memcpy(m_buffer + m_bufferPos, buffer, length);
		m_bufferPos += length;
	}
	return ret;
}

MOZ_DECL_CTOR_COUNTER(nsMsgLineBuffer);

nsMsgLineBuffer::nsMsgLineBuffer(nsMsgLineBufferHandler *handler, PRBool convertNewlinesP)
{
	MOZ_COUNT_CTOR(nsMsgLineBuffer);
	m_handler = handler;
	m_convertNewlinesP = convertNewlinesP;
    m_lookingForCRLF = PR_TRUE;
}

nsMsgLineBuffer::~nsMsgLineBuffer()
{
	MOZ_COUNT_DTOR(nsMsgLineBuffer);
}

void
nsMsgLineBuffer::SetLookingForCRLF(PRBool b)
{
  m_lookingForCRLF = b;
}

PRInt32	nsMsgLineBuffer::BufferInput(const char *net_buffer, PRInt32 net_buffer_size)
{
    int status = 0;
    if (m_bufferPos > 0 && m_buffer && m_buffer[m_bufferPos - 1] == CR &&
        net_buffer_size > 0 && net_buffer[0] != LF) {
        /* The last buffer ended with a CR.  The new buffer does not start
           with a LF.  This old buffer should be shipped out and discarded. */
        PR_ASSERT(m_bufferSize > m_bufferPos);
        if (m_bufferSize <= m_bufferPos) return -1;
        status = ConvertAndSendBuffer();
        if (status < 0) 
			return status;
        m_bufferPos = 0;
    }
    while (net_buffer_size > 0)
	{
        const char *net_buffer_end = net_buffer + net_buffer_size;
        const char *newline = 0;
        const char *s;
        
        
        for (s = net_buffer; s < net_buffer_end; s++)
		{
          if (m_lookingForCRLF) {
            /* Move forward in the buffer until the first newline.
               Stop when we see CRLF, CR, or LF, or the end of the buffer.
               *But*, if we see a lone CR at the *very end* of the buffer,
               treat this as if we had reached the end of the buffer without
               seeing a line terminator.  This is to catch the case of the
               buffers splitting a CRLF pair, as in "FOO\r\nBAR\r" "\nBAZ\r\n".
            */
            if (*s == CR || *s == LF) {
              newline = s;
              if (newline[0] == CR) {
                if (s == net_buffer_end - 1) {
                  /* CR at end - wait for the next character. */
                  newline = 0;
                  break;
                }
                else if (newline[1] == LF) {
                  /* CRLF seen; swallow both. */
                  newline++;
                }
              }
              newline++;
              break;
			}
          }
          else {
            /* if not looking for a CRLF, stop at CR or LF.  (for example, when parsing the newsrc file).  this fixes #9896, where we'd lose the last line of anything we'd parse that used CR as the line break. */
            if (*s == CR || *s == LF) {
              newline = s;
              newline++;
              break;
            }
          }
		}
        
        /* Ensure room in the net_buffer and append some or all of the current
           chunk of data to it. */
        {
            const char *end = (newline ? newline : net_buffer_end);
            PRUint32 desired_size = (end - net_buffer) + m_bufferPos + 1;
            
            if (desired_size >= m_bufferSize)
            {
                status = GrowBuffer (desired_size, 1024);
                if (status < 0) 
					return status;
            }
            memcpy (m_buffer + m_bufferPos, net_buffer, (end - net_buffer));
            m_bufferPos += (end - net_buffer);
        }
        
        /* Now m_buffer contains either a complete line, or as complete
           a line as we have read so far.
           
           If we have a line, process it, and then remove it from `m_buffer'.
           Then go around the loop again, until we drain the incoming data.
           */
        if (!newline)
            return 0;
        
        status = ConvertAndSendBuffer();
        if (status < 0) return status;
        
        net_buffer_size -= (newline - net_buffer);
        net_buffer = newline;
        m_bufferPos = 0;
	}
#ifdef DEBUG_bienvenu
	printf("returning from buffer input m_bufferPos = %ld\n", m_bufferPos);
#endif	
    return 0;
}

PRInt32 nsMsgLineBuffer::HandleLine(char *line, PRUint32 line_length)
{
	NS_ASSERTION(PR_FALSE, "must override this method if you don't provide a handler");
	return 0;
}

PRInt32 nsMsgLineBuffer::ConvertAndSendBuffer()
{
    /* Convert the line terminator to the native form.
     */

	char *buf = m_buffer;
	PRInt32 length = m_bufferPos;

    char* newline;
    
    PR_ASSERT(buf && length > 0);
    if (!buf || length <= 0) 
		return -1;
    newline = buf + length;
    
    PR_ASSERT(newline[-1] == CR || newline[-1] == LF);
    if (newline[-1] != CR && newline[-1] != LF)
		return -1;
    
    if (!m_convertNewlinesP)
	{
	}
#if (MSG_LINEBREAK_LEN == 1)
    else if ((newline - buf) >= 2 &&
             newline[-2] == CR &&
             newline[-1] == LF)
	{
        /* CRLF -> CR or LF */
        buf [length - 2] = MSG_LINEBREAK[0];
        length--;
	}
    else if (newline > buf + 1 &&
             newline[-1] != MSG_LINEBREAK[0])
	{
        /* CR -> LF or LF -> CR */
        buf [length - 1] = MSG_LINEBREAK[0];
	}
#else
    else if (((newline - buf) >= 2 && newline[-2] != CR) ||
             ((newline - buf) >= 1 && newline[-1] != LF))
	{
        /* LF -> CRLF or CR -> CRLF */
        length++;
        buf[length - 2] = MSG_LINEBREAK[0];
        buf[length - 1] = MSG_LINEBREAK[1];
	}
#endif
    
    return (m_handler) ? m_handler->HandleLine(buf, length) : HandleLine(buf, length);
}

// If there's still some data (non CRLF terminated) flush it out
PRInt32 nsMsgLineBuffer::FlushLastLine()
{
	char *buf = m_buffer + m_bufferPos;
	PRInt32 length = m_bufferPos - 1;
	if (length > 0)
		return (m_handler) ? m_handler->HandleLine(buf, length) : HandleLine(buf, length);
	else
		return 0;
}

///////////////////////////////////////////////////////////////////////////////////////////////////
// This is a utility class used to efficiently extract lines from an input stream by buffering
// read but unprocessed stream data in a buffer. 
///////////////////////////////////////////////////////////////////////////////////////////////////

nsMsgLineStreamBuffer::nsMsgLineStreamBuffer(PRUint32 aBufferSize, const char * aEndOfLineToken, 
											 PRBool aAllocateNewLines, PRBool aEatCRLFs) 
											: m_eatCRLFs(aEatCRLFs), m_allocateNewLines(aAllocateNewLines),
											  m_endOfLineToken(aEndOfLineToken)
{
	NS_PRECONDITION(aBufferSize > 0, "invalid buffer size!!!");
	m_dataBuffer = nsnull;
	m_startPos = 0;
    m_numBytesInBuffer = 0;

	// used to buffer incoming data by ReadNextLineFromInput
	if (aBufferSize > 0)
	{
		m_dataBuffer = (char *) PR_CALLOC(sizeof(char) * aBufferSize);
	}

	m_dataBufferSize = aBufferSize;
}

nsMsgLineStreamBuffer::~nsMsgLineStreamBuffer()
{
	PR_FREEIF(m_dataBuffer); // release our buffer...
}

// the design for this method has an inherit bug: if the length of the line is greater than the size of m_dataBufferSize, 
// then we'll never find the next line because we can't hold the whole line in memory. 
// aInputStream - the input stream we want to read a line from
// aPauseForMoreData is returned as PR_TRUE if the stream does not yet contain a line and we must wait for more
// data to come into the stream.

// Note to people wishing to modify this function: Be *VERY CAREFUL* this is a critical function used by all of
// our mail protocols including imap, nntp, and pop. If you screw it up, you could break a lot of stuff.....

char * nsMsgLineStreamBuffer::ReadNextLine(nsIInputStream * aInputStream, PRUint32 &aNumBytesInLine, PRBool &aPauseForMoreData)
{
	// try to extract a line from m_inputBuffer. If we don't have an entire line, 
	// then read more bytes out from the stream. If the stream is empty then wait
	// on the monitor for more data to come in.

	NS_PRECONDITION(m_dataBuffer && m_dataBufferSize > 0, "invalid input arguments for read next line from input");
	
	// initialize out values
	aPauseForMoreData = PR_FALSE;
	aNumBytesInLine = 0;
	char * endOfLine = nsnull;
    char * startOfLine = m_dataBuffer+m_startPos;

	if (m_numBytesInBuffer > 0) // any data in our internal buffer?
		endOfLine = PL_strstr(startOfLine, m_endOfLineToken); // see if we already have a line ending...

	// it's possible that we got here before the first time we receive data from the server
	// so aInputStream will be nsnull...
	if (!endOfLine && aInputStream) // get some more data from the server
	{
		PRUint32 numBytesInStream = 0;
		PRUint32 numBytesCopied = 0;
		aInputStream->Available(&numBytesInStream);
		// if the number of bytes we want to read from the stream, is greater than the number
		// of bytes left in our buffer, then we need to shift the start pos and its contents
		// down to the beginning of m_dataBuffer...
		PRUint32 numFreeBytesInBuffer = m_dataBufferSize - m_startPos - m_numBytesInBuffer;
		if (numBytesInStream >= numFreeBytesInBuffer)
		{
            if (m_numBytesInBuffer && m_startPos)
            {
                nsCRT::memmove(m_dataBuffer, startOfLine, m_numBytesInBuffer);
                m_dataBuffer[m_numBytesInBuffer] = '\0'; // make sure the end
                                                         // of the buffer is
                                                         // terminated
                m_startPos = 0;
                startOfLine = m_dataBuffer;
                numFreeBytesInBuffer = m_dataBufferSize - m_numBytesInBuffer;
//				printf("moving data in read line around because buffer filling up\n");
            }
            NS_ASSERTION(m_startPos == 0, "m_startPos should be 0 .....\n");
		}

		PRUint32 numBytesToCopy = PR_MIN(numFreeBytesInBuffer - 1 /* leave one for a null terminator */, numBytesInStream);
		// read the data into the end of our data buffer
		if (numBytesToCopy > 0)
		{
			aInputStream->Read(startOfLine + m_numBytesInBuffer,
                               numBytesToCopy, &numBytesCopied);
            m_numBytesInBuffer += numBytesCopied;
			m_dataBuffer[m_startPos + m_numBytesInBuffer] = '\0';
		}
        else if (!m_numBytesInBuffer)
        {
          aPauseForMoreData = PR_TRUE;
          return nsnull;
        }

		// okay, now that we've tried to read in more data from the stream, look for another end of line 
		// character 
		endOfLine = PL_strstr(startOfLine, m_endOfLineToken);

	}

	// okay, now check again for endOfLine.
	if (endOfLine)
	{
		if (!m_eatCRLFs)
			endOfLine += PL_strlen(m_endOfLineToken); // count for CRLF

        aNumBytesInLine = endOfLine - startOfLine;

		// PR_CALLOC zeros out the allocated line
		char* newLine = (char*) PR_CALLOC(aNumBytesInLine+1);
		if (!newLine)
        {
            aNumBytesInLine = 0;
            aPauseForMoreData = PR_TRUE;
			return nsnull;
        }

		nsCRT::memcpy(newLine, startOfLine, aNumBytesInLine); // copy the string into the new line buffer

		if (m_eatCRLFs)
			endOfLine += PL_strlen(m_endOfLineToken); // advance past CRLF if we haven't already done so...

		// now we need to update the data buffer to go past the line we just read out. 
        m_numBytesInBuffer -= (endOfLine - startOfLine);
        if (m_numBytesInBuffer)
            m_startPos = endOfLine - m_dataBuffer;
        else
            m_startPos = 0;

		return newLine;
	}

	aPauseForMoreData = PR_TRUE;
	return nsnull; // if we somehow got here. we don't have another line in the buffer yet...need to wait for more data...
}