FileTokenizer.cpp

Go to the documentation of this file.

/**
 * MOAB, a Mesh-Oriented datABase, is a software component for creating,
 * storing and accessing finite element mesh data.
 *
 * Copyright 2004 Sandia Corporation.  Under the terms of Contract
 * DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government
 * retains certain rights in this software.
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 */

#include "FileTokenizer.hpp"
#include "moab/ReadUtilIface.hpp"
#include "moab/ErrorHandler.hpp"

#include <cstring>
#include <cctype>
#include <string>
#include <cstdlib>

namespace moab
{

using namespace std;

FileTokenizer::FileTokenizer( FILE* file_ptr, ReadUtilIface* )
    : filePtr( file_ptr ), nextToken( buffer ), bufferEnd( buffer ), lineNumber( 1 ), lastChar( '\0' )
{
}

FileTokenizer::~FileTokenizer()
{
    fclose( filePtr );
}

bool FileTokenizer::eof() const
{
    return nextToken == bufferEnd && feof( filePtr );
}

const char* FileTokenizer::get_string()
{
    // If the whitespace character marking the end of the
    // last token was a newline, increment the line count.
    if( lastChar == '\n' ) ++lineNumber;

    // Loop until either found the start of a token to return or have
    // reached the end of the file.
    for( ;; )
    {
        // If the buffer is empty, read more.
        if( nextToken == bufferEnd )
        {
            size_t count = fread( buffer, 1, sizeof( buffer ) - 1, filePtr );
            if( 0 == count )
            {
                if( feof( filePtr ) )
                    return NULL;
                else
                    MB_SET_ERR_RET_VAL( "I/O Error", NULL );
            }

            nextToken = buffer;
            bufferEnd = buffer + count;
        }

        // If the current character is not a space, we've found a token.
        if( !isspace( *nextToken ) ) break;

        // If the current space character is a newline,
        // increment the line number count.
        if( *nextToken == '\n' ) ++lineNumber;
        ++nextToken;
    }

    // Store the start of the token in "result" and
    // advance "nextToken" to one past the end of the
    // token.
    char* result = nextToken;
    while( nextToken != bufferEnd && !isspace( static_cast< unsigned char >( *nextToken ) ) )
        ++nextToken;

    // If we have reached the end of the buffer without finding
    // a whitespace character terminating the token, we need to
    // read more from the file.  Only try once.  If the token is
    // too large to fit in the buffer, give up.
    if( nextToken == bufferEnd )
    {
        // Shift the (possibly) partial token to the start of the buffer.
        size_t remaining = bufferEnd - result;
        memmove( buffer, result, remaining );
        result    = buffer;
        nextToken = result + remaining;

        // Fill the remainder of the buffer after the token.
        size_t count = fread( nextToken, 1, sizeof( buffer ) - remaining - 1, filePtr );
        if( 0 == count && !feof( filePtr ) ) MB_SET_ERR_RET_VAL( "I/O Error", NULL );
        bufferEnd = nextToken + count;

        // Continue to advance nextToken until we find the space
        // terminating the token.
        while( nextToken != bufferEnd && !isspace( *nextToken ) )
            ++nextToken;

        if( nextToken == bufferEnd )
        {  // EOF
            *bufferEnd = '\0';
            ++bufferEnd;
        }
    }

    // Save terminating whitespace character (or NULL char if EOF).
    lastChar = *nextToken;
    // Put null in buffer to mark end of current token.
    *nextToken = '\0';
    // Advance nextToken to the next character to search next time.
    ++nextToken;

    return result;
}

bool FileTokenizer::get_double_internal( double& result )
{
    // Get a token
    const char *token_end, *token = get_string();
    if( !token ) return false;

    // Check for hex value -- on some platforms (e.g. Linux), strtod
    // will accept hex values, on others (e.g. Sun) it will not.  Force
    // failure on hex numbers for consistency.
    if( token[0] && token[1] && token[0] == '0' && toupper( token[1] ) == 'X' )
        MB_SET_ERR_RET_VAL( "Syntax error at line " << line_number() << ": expected number, got \"" << token << "\"",
                            false );

    // Parse token as double
    result = strtod( token, (char**)&token_end );

    // If the one past the last char read by strtod is
    // not the NULL character terminating the string,
    // then parse failed.
    if( *token_end )
        MB_SET_ERR_RET_VAL( "Syntax error at line " << line_number() << ": expected number, got \"" << token << "\"",
                            false );

    return true;
}

bool FileTokenizer::get_float_internal( float& result )
{
    double d;
    if( !get_double_internal( d ) ) return false;

    result = (float)d;

    return true;
}

bool FileTokenizer::get_long_int_internal( long& result )
{
    // Get a token
    const char *token_end, *token = get_string();
    if( !token ) return false;

    // Parse token as long
    result = strtol( token, (char**)&token_end, 0 );

    // If the one past the last char read by strtol is
    // not the NULL character terminating the string,
    // then parse failed.
    if( *token_end )
        MB_SET_ERR_RET_VAL( "Syntax error at line " << line_number() << ": expected number, got \"" << token << "\"",
                            false );

    return true;
}

bool FileTokenizer::get_byte_internal( unsigned char& result )
{
    long i;
    if( !get_long_int_internal( i ) ) return false;

    result = (unsigned char)i;
    if( i != (long)result ) MB_SET_ERR_RET_VAL( "Numeric overflow at line " << line_number(), false );

    return true;
}

bool FileTokenizer::get_short_int_internal( short& result )
{
    long i;
    if( !get_long_int_internal( i ) ) return false;

    result = (short)i;
    if( i != (long)result ) MB_SET_ERR_RET_VAL( "Numeric overflow at line " << line_number(), false );

    return true;
}

bool FileTokenizer::get_integer_internal( int& result )
{
    long i;
    if( !get_long_int_internal( i ) ) return false;

    result = (int)i;
    if( i != (long)result ) MB_SET_ERR_RET_VAL( "Numeric overflow at line " << line_number(), false );

    return true;
}

bool FileTokenizer::get_boolean_internal( bool& result )
{
    // Get a token
    const char* token = get_string();
    if( !token ) return false;

    if( token[1] || ( token[0] != '0' && token[0] != '1' ) )
        MB_SET_ERR_RET_VAL( "Syntax error at line " << line_number() << ": expected 0 or 1, got \"" << token << "\"",
                            false );

    result = token[0] == '1';

    return true;
}

bool FileTokenizer::get_floats( size_t count, float* array )
{
    for( size_t i = 0; i < count; ++i )
    {
        if( !get_float_internal( *array ) ) return false;
        ++array;
    }

    return true;
}

bool FileTokenizer::get_doubles( size_t count, double* array )
{
    for( size_t i = 0; i < count; ++i )
    {
        if( !get_double_internal( *array ) ) return false;
        ++array;
    }

    return true;
}

bool FileTokenizer::get_bytes( size_t count, unsigned char* array )
{
    for( size_t i = 0; i < count; ++i )
    {
        if( !get_byte_internal( *array ) ) return false;
        ++array;
    }

    return true;
}

bool FileTokenizer::get_short_ints( size_t count, short* array )
{
    for( size_t i = 0; i < count; ++i )
    {
        if( !get_short_int_internal( *array ) ) return false;
        ++array;
    }

    return true;
}

bool FileTokenizer::get_integers( size_t count, int* array )
{
    for( size_t i = 0; i < count; ++i )
    {
        if( !get_integer_internal( *array ) ) return false;
        ++array;
    }

    return true;
}

bool FileTokenizer::get_long_ints( size_t count, long* array )
{
    for( size_t i = 0; i < count; ++i )
    {
        if( !get_long_int_internal( *array ) ) return false;
        ++array;
    }

    return true;
}

bool FileTokenizer::get_booleans( size_t count, bool* array )
{
    for( size_t i = 0; i < count; ++i )
    {
        if( !get_boolean_internal( *array ) ) return false;
        ++array;
    }

    return true;
}

void FileTokenizer::unget_token()
{
    if( nextToken - buffer < 2 ) return;

    --nextToken;
    *nextToken = lastChar;
    --nextToken;
    while( nextToken > buffer && *nextToken )
        --nextToken;

    if( !*nextToken ) ++nextToken;

    lastChar = '\0';
}

bool FileTokenizer::match_token( const char* str, bool print_error )
{
    // Get a token
    const char* token = get_string();
    if( !token ) return false;

    // Check if it matches
    if( 0 == strcmp( token, str ) ) return true;

    // Construct error message
    if( print_error )
        MB_SET_ERR_CONT( "Syntax error at line " << line_number() << ": expected \"" << str << "\", got \"" << token
                                                 << "\"" );

    return false;
}

int FileTokenizer::match_token( const char* const* list, bool print_error )
{
    // Get a token
    const char* token = get_string();
    if( !token ) return 0;

    // Check if it matches any input string
    const char* const* ptr;
    for( ptr = list; *ptr; ++ptr )
    {
        if( 0 == strcmp( token, *ptr ) ) return ptr - list + 1;
    }

    if( !print_error ) return 0;

    // No match, constuct error message
    std::string message( "Parsing error at line " );
    char lineno[16];
    sprintf( lineno, "%d", line_number() );
    message += lineno;
    message += ": expected one of {";
    for( ptr = list; *ptr; ++ptr )
    {
        message += " ";
        message += *ptr;
    }
    message += " } got \"";
    message += token;
    message += "\"";
    MB_SET_ERR_CONT( message.c_str() );

    return 0;
}

bool FileTokenizer::get_newline( bool report_error )
{
    if( lastChar == '\n' )
    {
        lastChar = ' ';
        ++lineNumber;
        return true;
    }

    // Loop until either we a) find a newline, b) find a non-whitespace
    // character or c) reach the end of the file.
    for( ;; )
    {
        // If the buffer is empty, read more.
        if( nextToken == bufferEnd )
        {
            size_t count = fread( buffer, 1, sizeof( buffer ), filePtr );
            if( 0 == count )
            {
                if( eof() )
                    MB_SET_ERR_RET_VAL( "File truncated at line " << line_number(), false );
                else
                    MB_SET_ERR_RET_VAL( "I/O Error", false );
            }

            nextToken = buffer;
            bufferEnd = buffer + count;
        }

        // If the current character is not a space, the we've failed.
        if( !isspace( *nextToken ) )
            if( report_error ) MB_SET_ERR_RET_VAL( "Expected newline at line " << line_number(), false );

        // If the current space character is a newline,
        // increment the line number count.
        if( *nextToken == '\n' )
        {
            ++lineNumber;
            ++nextToken;
            lastChar = ' ';
            return true;
        }
        ++nextToken;
    }

    return false;
}

bool FileTokenizer::get_binary( size_t size, void* mem )
{
    // If data in buffer
    if( nextToken != bufferEnd )
    {
        // If requested size is less than buffer contents,
        // just pass back part of the buffer
        if( bufferEnd - nextToken <= (int)size )
        {
            memcpy( mem, nextToken, size );
            nextToken += size;
            return true;
        }

        // Copy buffer contents into memory and clear buffer
        memcpy( mem, nextToken, bufferEnd - nextToken );
        size -= bufferEnd - nextToken;
        mem       = reinterpret_cast< char* >( mem ) + ( bufferEnd - nextToken );
        nextToken = bufferEnd;
    }

    // Read any additional data from file
    return size == fread( mem, 1, size, filePtr );
}

}  // namespace moab