Cursor如何使用Merkle树进行代码库索引

Engineer’s Codex is a publication about real-world software engineering.

工程师的法典是关于现实世界软件工程的出版物。

Cursor, the popular AI IDE that recently announced they hit $300M ARR, uses Merkle trees to index code fast. This post goes over exactly how.

Cursor，这个最近宣布达到 3 亿美元年经常性收入的热门 AI IDE，使用 Merkle 树快速索引代码。 本文详细介绍了具体方法。

Before diving into Cursor's implementation, let's first understand what a Merkle tree is.

在深入了解 Cursor 的实现之前，让我们先了解什么是 Merkle 树。

A Merkle tree is a tree structure in which every "leaf" node is labeled with the cryptographic hash of a data block, and every non-leaf node is labeled with the cryptographic hash of the labels of its child nodes. This creates a hierarchical structure where changes at any level can be efficiently detected by comparing hash values.

Merkle 树是一种树结构，其中每个“叶”节点标记为数据块的加密哈希，而每个非叶节点标记为其子节点标签的加密哈希。这创建了一个层次结构，在任何级别的变化都可以通过比较哈希值高效检测。

Think of them as a fingerprinting system for data:

将它们视为数据的指纹系统：

Each piece of data (like a file) gets its own unique fingerprint (hash)
每个数据片段（如文件）都有其独特的指纹（哈希）
Pairs of fingerprints are combined and given a new fingerprint
指纹对被组合并赋予新的指纹
This process continues until you have just one master fingerprint (the root hash)
此过程持续进行，直到您只有一个主指纹（根哈希）

The root hash summarizes all data contained in the individual pieces, serving as a cryptographic commitment to the entire dataset. The beauty of this approach is that if any single piece of data changes, it will change all the fingerprints above it, ultimately changing the root hash.

根哈希总结了包含在各个数据块中的所有数据，作为对整个数据集的加密承诺。这种方法的美在于，如果任何单个数据块发生变化，它将改变其上所有的指纹，最终改变根哈希。

[

Merkle Tree | Suman Kundu

](https://substackcdn.com/image/fetch/$s_!Lb0A!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff2e51d80-fe7d-4756-a47e-9b27b126bea4_772x467.jpeg)

Top software engineers know a lot. But how do you kn...