objective-c - 使用 NSDirectoryEnumerator 为文件系统建模

Question

我正在尝试从给定的起始路径对文件系统的结构进行建模。目标是从该路径开始创建文件系统的标准 NSOutlineView。

我有一个名为 fileSystemItem 的模型对象。它具有以下(非常标准的)关系和属性:

• parentItem(指向另一个fileSystemItem对象)
• isLeaf(YES 对于文件；NO 对于文件夹)
• childrenItems(其他 fileSystemItems 的数组)
• fullPath(NSString；对象的文件路径)

我的问题是:如何使用 NSDirectoryEnumerator 构建模型？如果我这样做:

// NOTE: can't do "while (file = [dirEnum nextObject]) {...} because that sets 
// file to an auto-released string that doesn't get released until after ALL 
// iterations of the loop are complete. For large directories, that means our 
// memory use spikes to hundreds of MBs. So we do this instead to ensure that 
// the "file" string is released at the end of each iteration and our overall 
// memory footprint stays low.

NSDirectoryEnumerator *dirEnum = [aFileManager enumeratorAtPath:someStartingPath];
BOOL keepRunning = YES;
while (keepRunning)
{
    NSAutoreleasePool *innerPool = [[NSAutoreleasePool alloc] init];

    NSString *file = [dirEnum nextObject];
    if (file == nil) break;

    // ... examine "file". Create a fileSystemItem object to represent this item.
    // If it's a folder, we need to create a fileSystemItem for each item in the folder
    // and each fileSystemItem's "parentItem" relationship needs to be set to the 
    // fileSystemItem we're creating right here for "file." How can I do this inside
    // the directoryEnumerator, because as soon as we go to the next iteration of the    
    // loop (to handle the first item in "file" if "file" is a folder), we lose the  
    // reference to the fileSystemItem we created in THIS iteration of the loop for 
    // "file". Hopefully that makes sense... 

    [innerPool drain];
}

如果我编写一个递归函数来查看 startingPath 中的每个项目，并且如果该项目是一个文件夹，则在该文件夹上再次调用自身，依此类推，我可以了解如何构建模型。但是如何使用 NSDirectoryEnumerator 构建模型呢？我的意思是，据说这就是这个类存在的原因，对吗？

Answer 1

可以使用另一种目录枚举:

enumeratorAtURL:includePropertiesForKeys:options:errorHandler:

此枚举器具有其他有用的选项，并允许使用预取属性迭代 NSURL 实例，例如 NSURLNameKey、NSURLIsDirectoryKey、NSURLParentDirectoryURLKey 等。 ..它可以帮助摆脱递归的使用。