Question

I am mainly interested in the design decision behind this.

background information:
FileSystemInfo is base class to (and only to) FileInfo and DirectoryInfo.
Both classes implement GetAccessControl(), returning a FileSecurity or DirectorySecurity object respectively.
FileSecurity and DirectorySecurity both derive from FileSystemSecurity – and again are the only classes to do so.
Neither FileSecurity nor DirectorySecurity seem to declare any methods or properties of their own – apart from constructors.

Yet still, FileSystemInfo does not contain a public FileSystemSecurity GetAccessControl() method.

Question:
Can anybody shed some light onto why FileSystemInfo does not contain this method?

example code

public static void GrantFullControlToBuiltinUsers(this FileSystemInfo fileSystemInfo)
{
    FileSystemSecurity acFile;

    if(fileSystemInfo is DirectoryInfo)
        acFile = ((DirectoryInfo) fileSystemInfo).GetAccessControl();
    else
        acFile = ((FileInfo)fileSystemInfo).GetAccessControl();


    acFile.AddAccessRule(
        new FileSystemAccessRule(GetAccountNameBuiltinUsers(),
                                    FileSystemRights.FullControl,
                                    AccessControlType.Allow));


    if (fileSystemInfo is DirectoryInfo)
        ((DirectoryInfo)fileSystemInfo).SetAccessControl((DirectorySecurity)acFile);
    else
        ((FileInfo)fileSystemInfo).SetAccessControl((FileSecurity)acFile);
}

The code is far from beautiful with all the (unnecessary) casts in it and I wondered why the library was designed in this way.

Answer 1

My guess is that they wanted to have the two GetAccessControl methods return the proper, concrete types – FileSecurity and DirectorySecurity, respectively. If they inherited from a common GetAccessControl() method, they would be forced to return FileSystemSecurity, and the user would have to cast it manually.

It’s an aesthetic choice, mostly.