Reading view

How to tell where a file or folder is

23 January 2026 at 15:30

It might appear obvious how to tell where a file or folder is on your Mac: all you should need to do is select the item in the Finder and Get Info for it, where the full path is revealed. As I’m sure you’ll have guessed, that’s sometimes accurate, but sometimes conforms to the illusions the Finder creates. Allow me to demonstrate using my free utility Precize, which aims for complete accuracy.

We’ll start with an easy one, any of the apps you have installed in your main /Applications folder.

I picked Alfred, whose path is given as being in the main /Applications folder, which is on the Data volume, where I could install it. Look at the two lines below its path, and you’ll see its inode number, allocated by the file system, as the last on both lines, here 13587531. That’s relatively low, as we’ll see later, so confirms that it’s on the Data volume.

Now repeat that with one of the apps bundled with macOS.

The App Store app has a different path, starting from /System, indicating that, although it might appear to be in the main /Applications folder, in fact that app is in the Signed System Volume (SSV) rather than the Data volume. The Finder agrees with this, and there’s added confirmation in the inode number of 1152921500311879708, which really is vast, and much larger than that of the Alfred app. As this app is part of the System, we can’t copy, move or remove it.

Now try Safari, another of the apps bundled with macOS.

Although the Finder tries to pretend this is alongside Alfred in the main /Applications folder, the path given by Precize is way off that, and gives its own clue that Safari is actually in the apps.dmg cryptex, which is here grafted into a folder in the Preboot volume, so in neither the System or Data volume. However, it has a very low inode number of 534152 indicating it has been grafted into that path. The other number, given before the inode number, is also different, and corresponds to the Preboot volume, rather than either of the volumes in the boot volume group.

For a finale, repeat this with one of your files stored in iCloud Drive.

While the Finder will continue its pretence that it’s in something called iCloud Drive, in fact it’s tucked away in a nest of folders inside your Home ~/Library folder. This means it’s on the Data volume, and it too has a low inode number of 341568 to confirm that, and the same volume reference number as Alfred and the App Store apps, but different from that of Safari. Of course, there’s also a copy of that file’s details and its data held in iCloud Drive, to which you don’t have direct access.

This isn’t just a trick or an exercise, though. The real location of a file or folder is important, as it tells you what you can do with it. You know that, subject to permissions, you should be able to move or remove items from your Data volume, including your iCloud Drive, but those in the SSV and cryptexes are immutable. You can run copies of apps from your Data volume, but not those in the SSV or a cryptex, which can’t be copied directly, and whose copies won’t launch. And if you need to access items in iCloud Drive from Terminal, you won’t find anything named iCloud Drive there, just a long path from your Home ~/Library folder.

You’ll never become lost in the file system again.

回调之 Node.js VS 串行之 Python

中原驿站

胡中元

12 April 2018 at 18:46

Node 与 Python，都是脚本语言，有着类似的使用场景，所以在各个地方早已经互相 pk、比较过无数回了。虽然我知道编程语言之间的 VS 是一个很 low 的行为，因为他们必定是各有优势的。但今天我还是特别的想说说自己的心得体会。

Node 是我曾经特别喜欢，也是非常熟练的编程语言。Python 我还处于学习阶段，不敢说深入了解。

使用场景

如果要评论手机 App 开发，自然是有着一堆框架的 Node 胜出，而如果站在科学计算领域，那胜利者绝对是 Python。所以本文主要还是对简单的日常场景进行对比 ———— 代替 Shell 的那些操作、作为服务器中间件与数据库打交道等等。

Python 脚本

前几个周我写了数个 Python 脚本，这便是其中一个。将 NAS 中我收集的漂亮壁纸的分辨率、时间信息记录到 SQLite 数据库中，再定时从数据库中随机取出近期未使用过的壁纸，让桌面壁纸换一换。

Python 对于这种事情确实在行。另外我根据爬虫创建的翻墙规则 Shadowrocket-ADBlock-Rules 也是使用 Python 生成的，5 线程爬虫协同工作，代码清晰简洁。

Node 脚本

Node 上面提到的那些，相对而言更适合较大的项目。比如我开发的 XSYU-GMS 教务系统，就是一个前后端全栈 Javascript 项目。

对于 web 中间件，例如监听 443 端口提供数据相关的 API 调用服务。Node 原生可作为一个守护进程保持运行，灵活维护一些常量（Python 也是这样，只是感觉工作量会更大一点），作为脚本语言的高维护性体现出来，并且 Node “以事务驱动单进程” 的特性在极限情况下可以实现超高的性能。

Node 成败均在回调

继续上文。虽说 Node “以事务驱动单进程” 的特性在极限情况下可以实现超高的性能。但是对于大多数情况下，Python 也是可以撑住的。低负载的情况下，Node 引以为傲的 “事务驱动” 就变成了可怕的 “回调地狱”，在性能提升不大的情况下，严重影响了开发效率，甚至是提高了程序复杂度所导致的出错率。

我曾经在用 Node 开发 RSS 爬虫时，认真思考每一步之间的相互依赖关系，哪些可以并行，哪些必须等待。然后精确地实现，代码看起来非常复杂。实际上的效率提升其实真不是很重要，全部都用串行实现，岂不一样可以达到目的。

毕竟，要最追求效率那我应该选择 C 甚至是汇编，既然选择了 Node 这样的脚本语言，那就是为了开发方便。实际上，计算机编程语言的发展，就是在不断地用运行效率换取开发效率。当然，不会因为开发效率彻底代替运行效率。

Promise 与 Async Function

这大概就是拯救 Node 回调地狱的存在吧，确实，程序写起来体验好太多了。但是…… 这又导致了 await 地狱…… >o<

下面是我今天的代码，功能从数据库中获取部分 Email 地址，给他们发送邮件。

(async function() {
    let maxTimeStr = new Date().toUTCString();

    ret = await sqlP(`select username from user WHERE 
        d=0 AND status=1 AND created_at line.username);


    for(addr of emailList) {
        try {
            await sendMailForTK(addr);
        }
        catch(e) {
            console.log('sendMail failed', addr, e);
            await sleep(61);
            contine;
        }

        // send Done
        sqlP('update user SET d=88 WHERE username="' + addr + '"');

        await sleep(6);
    }

    process.exit();
})();

/**
三步串行操作：
1. 从数据库取得邮箱地址
2. 发送邮件
3，将记录写回数据库
*/

代码中的 sqlP(), sleep(), sendMailForTK()，都是我自己封装的 Promise Object，虽然封装不是个麻烦事，但我有一种感觉 —— 以后几乎所有的 Node 开发，都要经历这一步了。同时必须要经过的一步那就是在调用的时候用上 await，这就是 await 地狱。

于是问题来了，既然我们 80% 需要的程序逻辑都是串行，那么把这 80% 的代码特殊处理还真是一键麻烦的没必要的事情。像 Python 那样默认为串行，在 20% 的时候再使用多线程，在我看来是一种更合适、更通用的模式。

事件驱动是 JS 的核心，所以 Node 自然以后也只能保持这种模式了。

对 Node 的总结

我认为，Node 的最佳使用场景仅限于需要其特点：”异步回调” 的时候，可以带来高性能。而其他时候，这只会带来麻烦。

当然，Node 超级给力的模块机制，以及与简单易学的 JS 搭配实现的 web 开发语言前后端同一等等，使得 JS 永远充满魅力。

Python 的缺点

就并行串行而言，Python 我更偏好，但是这门语言我还是有一些需要吐槽的地方。

1. 编程风格

对于初学的我来说，完全不能赞同 Python 那些语法就是 “优雅的编程”，而 C 风格的大括号、分号就有多难看。在 class 里面，最多只能一个空行，而不能 2 个。使用 Python 编程让我有种被限制的感觉（也许还是因为我不熟悉吧）

总之我认为 “非 C 风格编程” 是一个缺点。

2. 相比 Node 包管理机制更弱

这不是 Python 太差，而是 Node 太叼了。Node 只提供底层 API，提倡使用第三方包完成你的工作（这也是导致 JS 项目层出不穷的原因），而 Python 并没有这种提倡。