fivedev/harbour-core
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

2011-02-11 14:46 UTC+0100 Przemyslaw Czerpak (druzus/at/priv.onet.pl)

Browse Source 
* harbour/src/rtl/base64d.c
    ! fixed destination buffer size calculation - it was too small
    ! fixed to work with platform/compilers where 'char' is unsigned type.
      Please NEVER use 'char' type in math calculation without explicit
      'signed' or 'unsigned' specifier. On some platforms (i,e. MIPS Linux
      ports or in all OpenWatcom builds) 'char' is unsigned type. I once
      cleaned whole code and fixed all wrong places. I do not want to
      repeat it - sometimes is hard to locate problematic places.
    ! fixed index to ASCII decode table verification - '{' was accepted
      as valid base64 char.
      BTW this code works only on ASCII based machines. It's not the
      one code with such limitation in Harbour SVN anyhow it would be
      good to mark such limitation - maybe someone will want to port
      Harbour to some IBM mainframe machines.
    % removed redundant code
This commit is contained in:
Przemyslaw Czerpak
2011-02-11 13:46:16 +00:00
parent ffaf84caa2
commit 84c23a8dcc
2 changed files with 61 additions and 82 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
harbour/ChangeLog
View File

@@ -16,6 +16,23 @@
The license applies to all entries newer than 2009-04-28.
*/
2011-02-11 14:46 UTC+0100 Przemyslaw Czerpak (druzus/at/priv.onet.pl)
* harbour/src/rtl/base64d.c
! fixed destination buffer size calculation - it was too small
! fixed to work with platform/compilers where 'char' is unsigned type.
Please NEVER use 'char' type in math calculation without explicit
'signed' or 'unsigned' specifier. On some platforms (i,e. MIPS Linux
ports or in all OpenWatcom builds) 'char' is unsigned type. I once
cleaned whole code and fixed all wrong places. I do not want to
repeat it - sometimes is hard to locate problematic places.
! fixed index to ASCII decode table verification - '{' was accepted
as valid base64 char.
BTW this code works only on ASCII based machines. It's not the
one code with such limitation in Harbour SVN anyhow it would be
good to mark such limitation - maybe someone will want to port
Harbour to some IBM mainframe machines.
% removed redundant code
2011-02-11 13:30 UTC+0100 Viktor Szakats (harbour.01 syenar.hu)
* src/rtl/base64d.c
! Fixed crash when empty or non-string parameter was passed to HB_BASE64DECODE().

126
harbour/src/rtl/base64d.c
View File

@@ -55,20 +55,11 @@
#include "hbapi.h"
typedef enum
{
step_a, step_b, step_c, step_d
} base64_decodestep;
/* Warning: this code works only on ASCII based machines */
typedef struct
static signed char base64_decode_value( char value_in )
{
base64_decodestep step;
char plainchar;
} base64_decodestate;
static int base64_decode_value( char value_in )
{
static const char s_decoding[] =
static const signed char s_decoding[] =
{
62, -1, -1, -1, 63, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, -1, -1, -1, -2, -1,
-1, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,
@@ -77,88 +68,59 @@ static int base64_decode_value( char value_in )
};
value_in -= 43;
if( value_in < 0 || value_in > ( char ) HB_SIZEOFARRAY( s_decoding ) )
if( value_in < 0 || value_in >= ( char ) HB_SIZEOFARRAY( s_decoding ) )
return -1;
return s_decoding[ ( int ) value_in ];
}
static int base64_decode_block( const char * code_in, const int length_in, char * pszPlainttextOut )
static HB_SIZE base64_decode_block( const char * code_in, const HB_SIZE length_in, char * pszPlainttextOut )
{
base64_decodestep step;
char plainchar;
const char * codechar = code_in;
const char * code_end = code_in + length_in;
char * pszPlainchar = pszPlainttextOut;
char fragment;
signed char fragment;
step = step_a;
plainchar = 0;
*pszPlainchar = plainchar;
switch( step )
for( ;; )
{
for( ;; )
do
{
case step_a:
do
{
if( codechar == code_in + length_in )
{
step = step_a;
plainchar = *pszPlainchar;
return pszPlainchar - pszPlainttextOut;
}
fragment = ( char ) base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar = ( fragment & 0x03f ) << 2;
case step_b:
do
{
if( codechar == code_in + length_in )
{
step = step_b;
plainchar = *pszPlainchar;
return pszPlainchar - pszPlainttextOut;
}
fragment = ( char ) base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar++ |= ( fragment & 0x030 ) >> 4;
*pszPlainchar = ( fragment & 0x00f ) << 4;
case step_c:
do
{
if( codechar == code_in + length_in )
{
step = step_c;
plainchar = *pszPlainchar;
return pszPlainchar - pszPlainttextOut;
}
fragment = ( char ) base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar++ |= ( fragment & 0x03c ) >> 2;
*pszPlainchar = ( fragment & 0x003 ) << 6;
case step_d:
do
{
if( codechar == code_in + length_in )
{
step = step_d;
plainchar = *pszPlainchar;
return pszPlainchar - pszPlainttextOut;
}
fragment = ( char ) base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar++ |= ( fragment & 0x03f );
if( codechar == code_end )
return pszPlainchar - pszPlainttextOut;
fragment = base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar = ( fragment & 0x03f ) << 2;
do
{
if( codechar == code_end )
return pszPlainchar - pszPlainttextOut;
fragment = base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar++ |= ( fragment & 0x030 ) >> 4;
*pszPlainchar = ( fragment & 0x00f ) << 4;
do
{
if( codechar == code_end )
return pszPlainchar - pszPlainttextOut;
fragment = base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar++ |= ( fragment & 0x03c ) >> 2;
*pszPlainchar = ( fragment & 0x003 ) << 6;
do
{
if( codechar == code_end )
return pszPlainchar - pszPlainttextOut;
fragment = base64_decode_value( *codechar++ );
}
while( fragment < 0 );
*pszPlainchar++ |= ( fragment & 0x03f );
}
/* control should not reach here */
return pszPlainchar - pszPlainttextOut;
}
HB_FUNC( HB_BASE64DECODE )
@@ -167,7 +129,7 @@ HB_FUNC( HB_BASE64DECODE )
if( len > 0 && len <= INT_MAX ) /* TOFIX */
{
char * code = ( char * ) hb_xgrab( ( ( ( ( len - 1 ) * 3 ) / 4 ) + 1 ) * sizeof( char ) );
char * code = ( char * ) hb_xgrab( ( ( ( len * 3 ) / 4 ) + 1 ) * sizeof( char ) );
HB_SIZE nSize = base64_decode_block( hb_parcx( 1 ), len, code );
hb_retclen_buffer( code, nSize );