ftpdirectoryparser.cpp

linux下开源浏览器WebKit的源码,市面上的很多商用浏览器都是移植自WebKit

        if (result.modifiedTime.tm_mon != 0)
        {
          result.modifiedTime.tm_mon--;
          result.modifiedTime.tm_mday = atoi(tokens[0]+3);
          result.modifiedTime.tm_year = atoi(tokens[0]+6);
          if (result.modifiedTime.tm_year < 80)
            result.modifiedTime.tm_year += 100;
        }

        result.modifiedTime.tm_hour = atoi(tokens[1]+0);
        result.modifiedTime.tm_min = atoi(tokens[1]+3);
        if ((tokens[1][5]) == 'P' && result.modifiedTime.tm_hour < 12)
          result.modifiedTime.tm_hour += 12;

        /* the caller should do this (if dropping "." and ".." is desired)
        if (result.type == FTPDirectoryEntry && result.filename[0] == '.' &&
            (result.filenameLength == 1 || (result.filenameLength == 2 &&
                                      result.filename[1] == '.')))
          return FTPJunkEntry;
        */

        return result.type;  
      } /* if (lstyle == 'W' && (!state.listStyle || state.listStyle == lstyle)) */
    } /* if (!lstyle && (!state.listStyle || state.listStyle == 'W')) */
#endif

    /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */

#if defined(SUPPORT_OS2)
    if (!lstyle && (!state.listStyle || state.listStyle == 'O')) /* OS/2 test */
    {
      /* 220 server IBM TCP/IP for OS/2 - FTP Server ver 23:04:36 on Jan 15 1997 ready.
      * fixed position, space padded columns. I have only a vague idea 
      * of what the contents between col 18 and 34 might be: All I can infer
      * is that there may be attribute flags in there and there may be 
      * a " DIR" in there.
      *
      *          1         2         3         4         5         6
      *0123456789012345678901234567890123456789012345678901234567890123456789
      *----- size -------|??????????????? MM-DD-YY|  HH:MM| nnnnnnnnn....
      *                 0  DIR            04-11-95   16:26  .
      *                 0  DIR            04-11-95   16:26  ..
      *                 0  DIR            04-11-95   16:26  ADDRESS
      *               612  RHSA           07-28-95   16:45  air_tra1.bag
      *               195  A              08-09-95   10:23  Alfa1.bag
      *                 0  RHS   DIR      04-11-95   16:26  ATTACH
      *               372  A              08-09-95   10:26  Aussie_1.bag
      *            310992                 06-28-94   09:56  INSTALL.EXE
      *                            1         2         3         4
      *                  01234567890123456789012345678901234567890123456789
      * dirlist from the mirror.pl project, col positions from Mozilla.
      */
      p = &(line[toklen[0]]);
      /* \s(\d\d-\d\d-\d\d)\s+(\d\d:\d\d)\s */
      if (numtoks >= 4 && toklen[0] <= 18 && isASCIIDigit(*tokens[0]) &&
         (linelen - toklen[0]) >= (53-18)                        &&
         p[18-18] == ' ' && p[34-18] == ' '                      &&
         p[37-18] == '-' && p[40-18] == '-' && p[43-18] == ' '   &&
         p[45-18] == ' ' && p[48-18] == ':' && p[51-18] == ' '   &&
         isASCIIDigit(p[35-18]) && isASCIIDigit(p[36-18])        &&
         isASCIIDigit(p[38-18]) && isASCIIDigit(p[39-18])        &&
         isASCIIDigit(p[41-18]) && isASCIIDigit(p[42-18])        &&
         isASCIIDigit(p[46-18]) && isASCIIDigit(p[47-18])        &&
         isASCIIDigit(p[49-18]) && isASCIIDigit(p[50-18])
      )
      {
        lstyle = 'O'; /* OS/2 */
        if (!state.listStyle)
        {            
          for (pos = 1; lstyle && pos < toklen[0]; pos++)
          {
            if (!isASCIIDigit(tokens[0][pos]))
              lstyle = 0;
          }
        }
      }

      if (lstyle == 'O')
      {
        state.parsedOne = true;
        state.listStyle = lstyle;

        p = &(line[toklen[0]]);

        result.caseSensitive = true;
        result.filename = &p[53-18];
        result.filenameLength = (&(line[linelen_sans_wsp]))
                           - (result.filename);
        result.type = FTPFileEntry;

        /* I don't have a real listing to determine exact pos, so scan. */
        for (pos = (18-18); pos < ((35-18)-4); pos++)
        {
          if (p[pos+0] == ' ' && p[pos+1] == 'D' && 
              p[pos+2] == 'I' && p[pos+3] == 'R')
          {
            result.type = FTPDirectoryEntry;
            break;
          }
        }
    
        if (result.type != FTPDirectoryEntry)
        {
          pos = toklen[0];
          result.fileSize = String(tokens[0], pos);
        }  
    
        result.modifiedTime.tm_mon = atoi(&p[35-18]) - 1;
        result.modifiedTime.tm_mday = atoi(&p[38-18]);
        result.modifiedTime.tm_year = atoi(&p[41-18]);
        if (result.modifiedTime.tm_year < 80)
          result.modifiedTime.tm_year += 100;
        result.modifiedTime.tm_hour = atoi(&p[46-18]);
        result.modifiedTime.tm_min = atoi(&p[49-18]);
   
        /* the caller should do this (if dropping "." and ".." is desired)
        if (result.type == FTPDirectoryEntry && result.filename[0] == '.' &&
            (result.filenameLength == 1 || (result.filenameLength == 2 &&
                                      result.filename[1] == '.')))
          return FTPJunkEntry;
        */

        return result.type;
      } /* if (lstyle == 'O') */

    } /* if (!lstyle && (!state.listStyle || state.listStyle == 'O')) */
#endif

    /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
    
#if defined(SUPPORT_LSL)
    if (!lstyle && (!state.listStyle || state.listStyle == 'U')) /* /bin/ls & co. */
    {
      /* UNIX-style listing, without inum and without blocks
       * "-rw-r--r--   1 root     other        531 Jan 29 03:26 README"
       * "dr-xr-xr-x   2 root     other        512 Apr  8  1994 etc"
       * "dr-xr-xr-x   2 root     512 Apr  8  1994 etc"
       * "lrwxrwxrwx   1 root     other          7 Jan 25 00:17 bin -> usr/bin"
       * Also produced by Microsoft's FTP servers for Windows:
       * "----------   1 owner    group         1803128 Jul 10 10:18 ls-lR.Z"
       * "d---------   1 owner    group               0 May  9 19:45 Softlib"
       * Also WFTPD for MSDOS:
       * "-rwxrwxrwx   1 noone    nogroup      322 Aug 19  1996 message.ftp"
       * Hellsoft for NetWare:
       * "d[RWCEMFA] supervisor            512       Jan 16 18:53    login"
       * "-[RWCEMFA] rhesus             214059       Oct 20 15:27    cx.exe"
       * Newer Hellsoft for NetWare: (netlab2.usu.edu)
       * - [RWCEAFMS] NFAUUser               192 Apr 27 15:21 HEADER.html
       * d [RWCEAFMS] jrd                    512 Jul 11 03:01 allupdates
       * Also NetPresenz for the Mac:
       * "-------r--         326  1391972  1392298 Nov 22  1995 MegaPhone.sit"
       * "drwxrwxr-x               folder        2 May 10  1996 network"
       * Protected directory:
       * "drwx-wx-wt  2 root  wheel  512 Jul  1 02:15 incoming"
       * uid/gid instead of username/groupname:
       * "drwxr-xr-x  2 0  0  512 May 28 22:17 etc"
      */
    
      if (numtoks >= 6)
      {
        /* there are two perm formats (Hellsoft/NetWare and *IX strmode(3)).
         * Scan for size column only if the perm format is one or the other.
         */
        if (toklen[0] == 1 || (tokens[0][1]) == '[')
        {
          if (*tokens[0] == 'd' || *tokens[0] == '-')
          {
            pos = toklen[0]-1;
            p = tokens[0] + 1;
            if (pos == 0)
            {
              p = tokens[1];
              pos = toklen[1];
            }
            if ((pos == 9 || pos == 10)        && 
                (*p == '[' && p[pos-1] == ']') &&
                (p[1] == 'R' || p[1] == '-')   &&
                (p[2] == 'W' || p[2] == '-')   &&
                (p[3] == 'C' || p[3] == '-')   &&
                (p[4] == 'E' || p[4] == '-'))
            {
              /* rest is FMA[S] or AFM[S] */
              lstyle = 'U'; /* very likely one of the NetWare servers */
            }
          }
        }
        else if ((toklen[0] == 10 || toklen[0] == 11) 
                   && strchr("-bcdlpsw?DFam", *tokens[0]))
        {
          p = &(tokens[0][1]);
          if ((p[0] == 'r' || p[0] == '-') &&
              (p[1] == 'w' || p[1] == '-') &&
              (p[3] == 'r' || p[3] == '-') &&
              (p[4] == 'w' || p[4] == '-') &&
              (p[6] == 'r' || p[6] == '-') &&
              (p[7] == 'w' || p[7] == '-'))
            /* 'x'/p[9] can be S|s|x|-|T|t or implementation specific */
          {
            lstyle = 'U'; /* very likely /bin/ls */
          }
        }
      }
      if (lstyle == 'U') /* first token checks out */
      {
        lstyle = 0;
        for (pos = (numtoks-5); !lstyle && pos > 1; pos--)
        {
          /* scan for: (\d+)\s+([A-Z][a-z][a-z])\s+
           *  (\d\d\d\d|\d\:\d\d|\d\d\:\d\d|\d\:\d\d\:\d\d|\d\d\:\d\d\:\d\d)
           *  \s+(.+)$
          */
          if (isASCIIDigit(*tokens[pos]) /* size */
              /* (\w\w\w) */
           && toklen[pos+1] == 3 && isASCIIAlpha(*tokens[pos+1]) &&
              isASCIIAlpha(tokens[pos+1][1]) && isASCIIAlpha(tokens[pos+1][2])
              /* (\d|\d\d) */
           && isASCIIDigit(*tokens[pos+2]) &&
                (toklen[pos+2] == 1 || 
                  (toklen[pos+2] == 2 && isASCIIDigit(tokens[pos+2][1])))
           && toklen[pos+3] >= 4 && isASCIIDigit(*tokens[pos+3]) 
              /* (\d\:\d\d\:\d\d|\d\d\:\d\d\:\d\d) */
           && (toklen[pos+3] <= 5 || (
               (toklen[pos+3] == 7 || toklen[pos+3] == 8) &&
               (tokens[pos+3][toklen[pos+3]-3]) == ':'))
           && isASCIIDigit(tokens[pos+3][toklen[pos+3]-2])
           && isASCIIDigit(tokens[pos+3][toklen[pos+3]-1])
           && (
              /* (\d\d\d\d) */
                 ((toklen[pos+3] == 4 || toklen[pos+3] == 5) &&
                  isASCIIDigit(tokens[pos+3][1]) &&
                  isASCIIDigit(tokens[pos+3][2])  )
              /* (\d\:\d\d|\d\:\d\d\:\d\d) */
              || ((toklen[pos+3] == 4 || toklen[pos+3] == 7) && 
                  (tokens[pos+3][1]) == ':' &&
                  isASCIIDigit(tokens[pos+3][2]) && isASCIIDigit(tokens[pos+3][3]))
              /* (\d\d\:\d\d|\d\d\:\d\d\:\d\d) */
              || ((toklen[pos+3] == 5 || toklen[pos+3] == 8) && 
                  isASCIIDigit(tokens[pos+3][1]) && (tokens[pos+3][2]) == ':' &&
                  isASCIIDigit(tokens[pos+3][3]) && isASCIIDigit(tokens[pos+3][4])) 
              )
           )
          {
            lstyle = 'U'; /* assume /bin/ls or variant format */
            tokmarker = pos;

            /* check that size is numeric */
            p = tokens[tokmarker];
            for (pos = 0; lstyle && pos < toklen[tokmarker]; pos++)
            {
              if (!isASCIIDigit(*p++))
                lstyle = 0;
            }
            if (lstyle)
            {
              month_num = 0;
              p = tokens[tokmarker+1];
              for (pos = 0;pos < (12*3); pos+=3)
              {
                if (p[0] == month_names[pos+0] && 
                    p[1] == month_names[pos+1] && 
                    p[2] == month_names[pos+2])
                  break;
                month_num++;
              }
              if (month_num >= 12)
                lstyle = 0;
            }
          } /* relative position test */
        } /* while (pos+5) < numtoks */
      } /* if (numtoks >= 4) */

      if (lstyle == 'U')
      {
        state.parsedOne = true;
        state.listStyle = lstyle;
    
        result.caseSensitive = false;
        result.type = FTPJunkEntry;
        if (*tokens[0] == 'd' || *tokens[0] == 'D')
          result.type = FTPDirectoryEntry;
        else if (*tokens[0] == 'l')
          result.type = FTPLinkEntry;
        else if (*tokens[0] == '-' || *tokens[0] == 'F')
          result.type = FTPFileEntry; /* (hopefully a regular file) */

        if (result.type != FTPDirectoryEntry)
        {
          pos = toklen[tokmarker];
          result.fileSize = String(tokens[tokmarker], pos);
        }

        result.modifiedTime.tm_mon  = month_num;
        result.modifiedTime.tm_mday = atoi(tokens[tokmarker+2]);
        if (result.modifiedTime.tm_mday == 0)
          result.modifiedTime.tm_mday++;

        p = tokens[tokmarker+3];
        pos = (unsigned int)atoi(p);
        if (p[1] == ':') /* one digit hour */
          p--;
        if (p[2] != ':') /* year */
        {
          result.modifiedTime.tm_year = pos;
        }
        else
        {
          result.modifiedTime.tm_hour = pos;
          result.modifiedTime.tm_min  = atoi(p+3);
          if (p[5] == ':')
            result.modifiedTime.tm_sec = atoi(p+6);
       
          if (!state.now)
          {
            time_t now = time(NULL);
            state.now = now * 1000000.0;

            // FIXME: This code has the year 2038 bug
            gmtime_r(&now, &state.nowFTPTime);
            state.nowFTPTime.tm_year += 1900;
          }

          result.modifiedTime.tm_year = state.nowFTPTime.tm_year;
          if ( (( state.nowFTPTime.tm_mon << 5) + state.nowFTPTime.tm_mday) <
               ((result.modifiedTime.tm_mon << 5) + result.modifiedTime.tm_mday) )
            result.modifiedTime.tm_year--;
       
        } /* time/year */
        
        result.filename = tokens[tokmarker+4];
        result.filenameLength = (&(line[linelen_sans_wsp]))
                           - (result.filename);

        if (result.type == FTPLinkEntry && result.filenameLength > 4)
        {
          p = result.filename + 1;
          for (pos = 1; pos < (result.filenameLength - 4); pos++)
          {
            if (*p == ' ' && p[1] == '-' && p[2] == '>' && p[3] == ' ')
            {
              result.linkname = p + 4;
              result.linknameLength = (&(line[linelen_sans_wsp]))
                               - (result.linkname);
              result.filenameLength = pos;
              break;
            }
            p++;
          }
        }

#if defined(SUPPORT_LSLF) /* some (very rare) servers return ls -lF */
        if (result.filenameLength > 1)
        {
          p = result.filename[result.filenameLength-1];
          pos = result.type;
          if (pos == 'd') { 
             if (*p == '/') result.filenameLength--; /* directory */
          } else if (pos == 'l') { 
             if (*p == '@') result.filenameLength--; /* symlink */
          } else if (pos == 'f') { 
             if (*p == '*') result.filenameLength--; /* executable */
          } else if (*p == '=' || *p == '%' || *p == '|') {
            result.filenameLength--; /* socket, whiteout, fifo */
          }
        }
#endif
     
        /* the caller should do this (if dropping "." and ".." is desired)
        if (result.type == FTPDirectoryEntry && result.filename[0] == '.' &&
            (result.filenameLength == 1 || (result.filenameLength == 2 &&
                                      result.filename[1] == '.')))
          return FTPJunkEntry;
        */

        return result.type;  

      } /* if (lstyle == 'U') */

    } /* if (!lstyle && (!state.listStyle || state.listStyle == 'U')) */
#endif

    /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */

#if defined(SUPPORT_W16) /* 16bit Windows */
    if (!lstyle && (!state.listStyle || state.listStyle == 'w'))
    {       /* old SuperTCP suite FTP server for Win3.1 */
            /* old NetManage Chameleon TCP/IP suite FTP server for Win3.1 */
      /*
      * SuperTCP dirlist from the mirror.pl project
      * mon/day/year separator may be '/' or '-'.
      * .               <DIR>           11-16-94        17:16
      * ..              <DIR>           11-16-94        17:16
      * INSTALL         <DIR>           11-16-94        17:17
      * CMT             <DIR>           11-21-94        10:17
      * DESIGN1.DOC          11264      05-11-95        14:20
      * README.TXT            1045      05-10-95        11:01
      * WPKIT1.EXE          960338      06-21-95        17:01
      * CMT.CSV                  0      07-06-95        14:56
      *
      * Chameleon dirlist guessed from lynx
      * .               <DIR>      Nov 16 1994 17:16   
      * ..              <DIR>      Nov 16 1994 17:16   
      * INSTALL         <DIR>      Nov 16 1994 17:17
      * CMT             <DIR>      Nov 21 1994 10:17
      * DESIGN1.DOC     11264      May 11 1995 14:20   A
      * README.TXT       1045      May 10 1995 11:01
      * WPKIT1.EXE     960338      Jun 21 1995 17:01   R
      * CMT.CSV             0      Jul 06 1995 14:56   RHA