BUG: Analyze Always Updates aaaaaaaa.cnt

MORE INFORMATION

SourceSafe database files follow an order that allows every log file in the data directory to be associated with a number. The number can be derived using the following algorithm:

   __int64 PhyToNum(char *szFile)
   {
   __int64 lFileNum = 0;   /* Counter for file Number */ 
   int  i;              /* Loop Variable */ 

   for( i=7; i >= 0; i-- ) /* Start from the right of the filename */ 
    {
     lFileNum *= 26;  /* Multiply by the appropriate power of 26 */ 
     lFileNum += (__int64)(toupper(szFile[i]) - 'A');
                      /* Convert the value */ 
    }
   return (lFileNum);
   }
				

Please note that the __int64 is a Microsoft-specific variable type for a 64- bit integer. This code snippet may not work with all compilers.

If Analyze detects a filename that was put into one of the subdirectories of the data directory by mistake, it can evaluate to a number higher than any other file in the database. Analyze allocates memory for every file from the first to the highest number it finds, even if some of those files do not physically exist in the data directory. This can cause Analyze to ask for more memory than the operating system can make available. As a result, the "unable to create filemapping" message appears.

If the aaaaaaaa.cnt is set to an incorrectly large value and files and projects are then added to SourceSafe, it might be impossible to return the database to a usable state without deleting and purging the files that represent the large values in their physical names. This can be very difficult if a large number of files and projects have been added since the aaaaaaaa.cnt file was set to a large value.