|
|
/*++
Copyright (C) Microsoft Corporation, 1997 - 1999 All rights reserved.
Module Name:
parse.cxx
Abstract:
Command line parser.
Author:
Steve Kiraly (SteveKi) 04-Mar-1997
Revision History:
--*/ #include "precomp.hxx"
#pragma hdrstop
#include "parser.hxx"
/*++
Routine Name:
Memory allocations.
Routine Description:
Uses new and delete when a C++ file. Arguments:
Normal allocation routines.
Return Value:
N/A
--*/ static inline PVOID private_alloc( UINT size ) { return new BYTE [ size ]; }
static inline VOID private_free( PVOID pvoid ) { delete [] pvoid; }
/*++
Routine Name:
StringToArgv
Routine Description:
Coverts a commandline string to an array of pointers to string. Arguments:
TCHAR *string - pointer to command line string. TCHAR *pac - pointer to an int on return will contain the number of strins in the arrary of pointes. TCHAR **string - pointer to where to return a pointer to the array of pointers to strings.
Return Value:
TRUE if success, FALSE error occurred.
Notes:
--*/ BOOL StringToArgv( const TCHAR *str, UINT *pac, TCHAR ***pppav, BOOL bParseExeName ) { TCHAR *word = NULL; TCHAR *w = NULL; TCHAR *program_name = NULL; TCHAR **av = NULL; LPTSTR string = NULL; UINT ac = 0; UINT numslash; BOOL inquote; BOOL copychar; BOOL retval = TRUE; TString strString;
//
// The pointers passed in must be valid.
//
if( !str || !pac || !pppav ) { SPLASSERT( FALSE ); return FALSE; }
//
// Process any file redirection.
//
vProcessCommandFileRedirection( str, strString );
//
// Cast the string to a usable string pointer.
//
string = const_cast<LPTSTR>( static_cast<LPCTSTR>( strString ) );
//
// Allocate the word buffer, this the maximum size in case there is
// just one extreamly long argument.
//
word = (LPTSTR) private_alloc( sizeof (TCHAR) * (lstrlen (string) + 1) );
if( word ) { //
// If we are to get the parse the program name from the provided command line.
//
if( bParseExeName ) { retval = IsolateProgramName( string, word, &string );
//
// If program name found.
//
if( retval ) { //
// Add the program name to the argv list.
//
retval = AddStringToArgv( &av, word );
if( retval ) { ac++; } } } } else { retval = FALSE; }
for ( ; retval && *string; ac++ ) { //
// Skip any leading spaces.
//
for (; *string && _istspace(*string); string++);
//
// Get a word out of the string
//
for (copychar = 1, inquote = 0, w = word; *string; copychar = 1) { //
// Rules: 2N backslashes + " ==> N backslashes and begin/end quote
// 2N+1 backslashes + " ==> N backslashes + literal "
// N backslashes ==> N backslashes
//
for( numslash = 0; *string == TEXT('\\'); string++ ) { //
// count number of backslashes for use below
//
numslash++; }
if (*string == TEXT('"')) { //
// if 2N backslashes before, start/end quote, otherwise
// copy literally
//
if (numslash % 2 == 0) { if (inquote) { //
// Double quote inside quoted string
// skip first quote char and copy second
//
if (*(string+1) == TEXT('"') ) { string++; } else { copychar = 0; } } else { //
// don't copy quote
//
copychar = 0; }
inquote = !inquote;
}
//
// Divide numslash by two
//
numslash /= 2;
}
//
// Copy the slashes
//
for( ; numslash; numslash--) { *w++ = TEXT('\\'); }
//
// If at the end of the command line or
// a space has been found.
//
if( !*string || (!inquote && _istspace(*string))) { break; }
//
// copy character into argument
//
if (copychar) { *w++ = *string; }
string++; }
//
// Terminate the word
//
*w = 0;
retval = AddStringToArgv( &av, word );
}
//
// Free up the word buffer.
//
if( word ) { private_free( word ); }
if( retval ) { *pac = ac; *pppav = av; }
#if DBG
vDumpArgList( ac, av ); #endif
return retval; }
/*++
Routine Name:
ReleaseArgv
Routine Description:
Releases the argv pointer to an array of pointers to strings. Arguments:
TCHAR **av - pointer to an arrary of pointers to strings.
Return Value:
Nothing.
Notes:
This routine releases the strings as well as the arrary of pointers to strings.
--*/ VOID ReleaseArgv( TCHAR **av ) { TCHAR **tav;
if( av ) { for( tav = av; *tav; tav++ ) { if( *tav ) { private_free( *tav ); } } private_free( av ); } }
/*++
Routine Name:
IsolateProgramName
Routine Description:
Extracts the program name from the specified string. This routing is used because the program name follows a differenct parsing technique from the other command line arguments. Arguments:
TCHAR *p - pointer to a string which contains the program name. TCHAR *program_name - pointer buffer where to place the extracted program. TCHAR **string - pointer to where to return a pointer where the program name ended. Used as the start of the string for nay remaing command line arguments.
Return Value:
TRUE if success, FALSE error occurred.
Notes:
This routine is very specific to the NTFS file nameing format.
--*/ BOOL IsolateProgramName( const TCHAR *p, TCHAR *program_name, TCHAR **string ) { BOOL retval = FALSE; TCHAR c; //
// All the arguments must be valid.
//
if( p && program_name && *string ) { retval = TRUE; } //
// Skip any leading spaces.
//
for( ; retval && *p && _istspace(*p); p++ ); //
// A quoted program name is handled here. The handling is much
// simpler than for other arguments. Basically, whatever lies
// between the leading double-quote and next one, or a terminal null
// character is simply accepted. Fancier handling is not required
// because the program name must be a legal NTFS/HPFS file name.
//
if (retval && *p == TEXT('"')) { //
// scan from just past the first double-quote through the next
// double-quote, or up to a null, whichever comes first
//
while ((*(++p) != TEXT('"')) && (*p != TEXT('\0'))) { *program_name++ = *p; }
//
// append the terminating null
//
*program_name++ = TEXT('\0');
//
// if we stopped on a double-quote (usual case), skip over it
//
if (*p == TEXT('"')) { p++; } } else { //
// Not a quoted program name
//
do { *program_name++ = *p; c = (TCHAR) *p++;
} while (c > TEXT(' '));
if (c == TEXT('\0')) { p--; } else { *(program_name - 1) = TEXT('\0'); } }
if( retval ) { *string = const_cast<TCHAR *>( p ); }
return retval;
}
/*++
Routine Name:
AddStringToArgv
Routine Description:
Adds a string to a array of pointers to strings. Arguments:
TCHAR ***argv - pointer where to return the new array of pointer to strings. TCHAR *word - pointer to word or string to add to the array.
Return Value:
TRUE if success, FALSE error occurred.
Notes:
Realloc is not used here because a user of this code may decide to do allocations with new and delete rather that malloc free. Both the new array pointer and string pointer must be valid for this routine to do anything. The array pointer can point to null but the pointer cannot be null.
--*/ BOOL AddStringToArgv( TCHAR ***argv, TCHAR *word ) { BOOL retval = FALSE; UINT count; TCHAR **targv = NULL; TCHAR **tmp = NULL; TCHAR **nargv = NULL; TCHAR *w = NULL;
if( argv && word ) { //
// Allocate the word buffer plus the null.
//
w = (TCHAR *)private_alloc( ( lstrlen( word ) + 1 ) * sizeof(TCHAR) );
if( w ) { //
// Copy the word.
//
lstrcpy( w, word );
//
// Count the current size of the argv.
//
for( count = 1, targv = *argv; targv && *targv; targv++, count++ );
nargv = (TCHAR **)private_alloc( ( count + 1 ) * sizeof(TCHAR *) );
if( nargv ) { //
// Copy the orig argv to the new argv.
//
for( tmp = nargv, targv = *argv; targv && *targv; *nargv++ = *targv++ );
//
// Set the new string pointer.
//
*nargv++ = w;
//
// Mark the end.
//
*nargv = 0;
//
// Free the original argv
//
private_free( *argv );
//
// Set the return pointer.
//
*argv = tmp;
retval = TRUE; } } }
if( !retval ) { if( w ) private_free(w);
if( nargv ) private_free(nargv); }
return retval; }
/********************************************************************
Code for commad file redirection.
********************************************************************/
/*++
Routine Name:
vProcessCommandFileRedirection
Routine Description:
Arguments:
Return Value:
Nothing.
--*/ VOID vProcessCommandFileRedirection( IN LPCTSTR pszCmdLine, IN OUT TString &strCmdLine ) { DBGMSG( DBG_TRACE, ( "vProcessCommandFileRedirection\n" ) ); DBGMSG( DBG_TRACE, ( "Pre vProcessCommandFileRedirection "TSTR"\n", pszCmdLine ) );
SPLASSERT( pszCmdLine );
TCHAR szFilename [MAX_PATH]; TCHAR szBuffer [MAX_PATH]; LPTSTR pSrc = const_cast<LPTSTR>( pszCmdLine ); LPTSTR pDst = szBuffer; UINT nBufferSize = COUNTOF( szBuffer ); TStatusB bStatus;
for( ; pSrc && *pSrc; ) { //
// Look for the escape character. If found - look for
// the special characters.
//
if( *pSrc == TEXT('\\') && *(pSrc+1) == TEXT('@') ) { //
// Copy the special character skipping the escape.
//
*pDst++ = *(pSrc+1);
//
// Skip two characters in the source buffer
//
pSrc += 2;
//
// Flush the buffer if it's full.
//
bStatus DBGCHK = bFlushBuffer( szBuffer, nBufferSize, &pDst, strCmdLine, FALSE ); } //
// Look for an redirected file sentinal.
//
else if( *pSrc == TEXT('@') ) { //
// Flush the current working buffer.
//
bStatus DBGCHK = bFlushBuffer( szBuffer, nBufferSize, &pDst, strCmdLine, TRUE );
//
// Isolate the file name following the sentinal, note this
// requires special parsing for NTFS file names.
//
if( IsolateProgramName( pSrc+1, szFilename, &pSrc ) ) { AFileInfo FileInfo;
//
// Read the file into the a single buffer.
//
if( bReadCommandFileRedirection( szFilename, &FileInfo ) ) { //
// Tack on the file information to the command line.
//
bStatus DBGCHK = strCmdLine.bCat( FileInfo.pszOffset ); }
//
// Release the file information.
//
if( FileInfo.pData ) { private_free( FileInfo.pData ); } } } else { //
// Copy the string to working buffer.
//
*pDst++ = *pSrc++;
//
// Flush the buffer if it's full.
//
bStatus DBGCHK = bFlushBuffer( szBuffer, nBufferSize, &pDst, strCmdLine, FALSE ); } }
//
// Flush any remaining items into the command line string.
//
bStatus DBGCHK = bFlushBuffer( szBuffer, nBufferSize, &pDst, strCmdLine, TRUE );
DBGMSG( DBG_TRACE, ( "Post vProcessCommandFileRedirection "TSTR"\n", (LPCTSTR)strCmdLine ) ); }
/*++
Routine Name:
bFlushBuffer
Routine Description:
Flushed the working buffer for builing the complete redirected command string.
Arguments:
pszBuffer - Pointer to working buffer. nSize - Total size in bytes of working buffer. pszCurrent - Current pointer into working buffer, strDestination - Reference to destination string buffer. bForceFlush - TRUE force buffer flush, FALSE flush if full.
Return Value:
TRUE buffer flushed successfully, FALSE error occurred.
--*/ BOOL bFlushBuffer( IN LPCTSTR pszBuffer, IN UINT nSize, IN OUT LPTSTR *pszCurrent, IN OUT TString &strDestination, IN BOOL bForceFlush ) { TStatusB bStatus; bStatus DBGNOCHK = TRUE;
//
// If something is in the buffer.
//
if( *pszCurrent != pszBuffer ) { //
// If the destination buffer is full.
// nsize - 2 because the pointers are zero base and the
// we need one extra slot for the null terminator.
//
if( ( *pszCurrent > ( pszBuffer + nSize - 2 ) ) || bForceFlush ) { //
// Null terminate the working buffer.
//
**pszCurrent = 0;
//
// Tack on the string to the destination string.
//
bStatus DBGCHK = strDestination.bCat( pszBuffer );
//
// Reset the current buffer pointer.
//
*pszCurrent = const_cast<LPTSTR>( pszBuffer ); } } return bStatus; }
/*++
Routine Name:
bReadCommandFileRedirection
Routine Description:
Read the filename into a single string. This in my opinion is a hack, but does the job. I made the assumption the commad file will always be less than 16k in size. Because the file size assumption was made it became feasable to just read the file into one potentialy huge buffer. Note the file is not broken up into separate lines the new lines are modified inplace to spaces. The StringToArgv will parse the line into an argv list ignoring spaces where appropriate.
Arguments: szFilename - pointer to the redirected filename. pFileInfo - pointe to file information structure.
Return Value:
TRUE function complete ok. FALSE error occurred.
--*/ BOOL bReadCommandFileRedirection( IN LPCTSTR szFilename, IN AFileInfo *pFileInfo ) { DBGMSG( DBG_TRACE, ( "bReadCommandFileRedirection\n" ) ); DBGMSG( DBG_TRACE, ( "szFilename "TSTR"\n", szFilename ) );
HANDLE hFile = INVALID_HANDLE_VALUE; BOOL bReturn = FALSE; DWORD dwBytesRead = 0; DWORD dwFileSize = 0; LPTSTR pszUnicode = NULL;
#ifdef UNICODE
enum { kByteOrderMark = 0xFEFF, kMaxFileSize = 16*1024 };
memset( pFileInfo, 0, sizeof( AFileInfo ) );
//
// Open the redirected file.
//
hFile = CreateFile ( szFilename, GENERIC_READ, FILE_SHARE_READ, NULL, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL );
if( hFile != INVALID_HANDLE_VALUE ) { //
// Validate the file size.
//
dwFileSize = GetFileSize( hFile, NULL ); //
// We fail the call if the file is bigger than a reasonable
// file size. This was a decision made to make
// reading the file into a contiguous buffer a reasonable
// approach.
//
if( dwFileSize <= kMaxFileSize ) { //
// Allocate buffer to hold entire file plus a null terminator.
//
pFileInfo->pData = (PVOID)private_alloc( dwFileSize + sizeof( TCHAR ) );
if( pFileInfo->pData ) { //
// Read the file into one huge buffer.
//
bReturn = ReadFile( hFile, pFileInfo->pData, dwFileSize, &dwBytesRead, NULL );
//
// If the read succeeded.
//
if( bReturn && dwBytesRead == dwFileSize ) { //
// Assume failure.
//
bReturn = FALSE;
//
// Check for the byte order mark, This mark allows use to
// read both ansi and unicode file.
//
if( *(LPWORD)pFileInfo->pData != kByteOrderMark ) { DBGMSG( DBG_TRACE, ( "Ansi file found\n" ) );
//
// Null terminate the file.
//
*((LPBYTE)pFileInfo->pData + dwBytesRead) = 0;
pszUnicode = (LPTSTR)private_alloc( ( strlen( (LPSTR)pFileInfo->pData ) + 1 ) * sizeof(TCHAR) );
if( pszUnicode ) { if( AnsiToUnicodeString( (LPSTR)pFileInfo->pData, pszUnicode, 0 ) ) { //
// Release the ansi string, and assign the unicode string.
// then set the offset.
//
private_free( pFileInfo->pData ); pFileInfo->pData = pszUnicode; pFileInfo->pszOffset = pszUnicode; bReturn = TRUE; } } } else { DBGMSG( DBG_TRACE, ( "Unicode file found\n" ) );
//
// Null terminate file string.
//
*((LPBYTE)pFileInfo->pData + dwBytesRead) = 0; *((LPBYTE)pFileInfo->pData + dwBytesRead+1) = 0;
//
// Set the file offset to just after the byte mark.
//
pFileInfo->pszOffset = (LPTSTR)((LPBYTE)pFileInfo->pData + sizeof( WORD ) ); bReturn = TRUE; }
if( bReturn ) { //
// Replace carriage returns and line feeds with spaces.
// The spaces will be removed when the command line is
// converted to an argv list.
//
vReplaceNewLines( pFileInfo->pszOffset ); } } } } else { DBGMSG( DBG_WARN, ( "Redirected file too large %d.\n", dwFileSize ) ); } CloseHandle( hFile ); } //
// If something failed release all allocated resources.
//
if( !bReturn ) { private_free( pFileInfo->pData ); private_free( pszUnicode ); pFileInfo->pData = NULL; pFileInfo->pszOffset = NULL; }
#endif
return bReturn; }
/*++
Routine Name:
vReplaseCarriageReturnAndLineFeed
Routine Description:
Replace carriage returns and line feeds with spaces.
Arguments:
pszLine - pointer to line buffer where to replace carriage returns and line feed characters.
Return Value:
Nothing.
--*/ VOID vReplaceNewLines( IN LPTSTR pszLine ) { DBGMSG( DBG_TRACE, ( "vReplaceNewLines\n" ) );
for( LPTSTR p = pszLine; p && *p; ++p ) { if( *p == TEXT('\n') || *p == TEXT('\r') ) { *p = TEXT(' '); } } }
/*++
Routine Name:
AnsiToUnicodeString
Routine Description:
Converts an ansi string to unicode.
Arguments:
pAnsi - Ansi string to convert. pUnicode - Unicode buffer where to place the converted string. StringLength - Ansi string argument length, this may be null then length will be calculated.
Return Value:
Return value from MultiByteToWideChar.
--*/ INT AnsiToUnicodeString( IN LPSTR pAnsi, IN LPWSTR pUnicode, IN DWORD StringLength OPTIONAL ) { INT iReturn;
if( StringLength == 0 ) { StringLength = strlen( pAnsi ); }
iReturn = MultiByteToWideChar(CP_ACP, MB_PRECOMPOSED, pAnsi, -1, pUnicode, StringLength + 1 ); //
// Ensure NULL termination.
//
pUnicode[StringLength] = 0;
return iReturn; }
/*++
Routine Name:
vDumpArgList
Routine Description:
Dumps the arg list to the debugger.
Arguments:
ac - Count of strings in the argv list. av - Pointer to an array of pointers to strings, which I call an argv list.
Return Value:
Return value from MultiByteToWideChar.
--*/
#if DBG
VOID vDumpArgList( UINT ac, TCHAR **av ) { DBGMSG( DBG_TRACE, ( "vDumpArgList\n" ) );
for( UINT i = 0; ac; --ac, ++av, ++i ) { DBGMSG( DBG_TRACE, ( "%d - "TSTR"\n", i, *av ) ); } }
#endif
|