为何JSON.parse未报错却忽略数据？npm包接口readme字段异常排查

Great question! Let me break down exactly what's going on here:

The Root Cause #

The npm registry API serves different versions of package metadata based on the Accept header included in your request:

• When you load the URL directly in your browser, your browser sends an Accept header that includes values like text/html or application/json. The registry responds with the full package metadata, including the complete readme content you see.
• When you use a default fetch() call, the browser sends a broader Accept header (typically */*). The npm registry interprets this as a request for the minified, install-focused version of the metadata—this is the format npm uses internally when installing packages, so non-essential fields like readme are emptied to reduce payload size.

How to Fix It #

To retrieve the full metadata (including the complete readme), explicitly set the Accept header to application/json in your fetch request:

fetch('http://registry.npmjs.org/core-js', {
  headers: {
    'Accept': 'application/json'
  }
})
.then(response => response.json())
.then(data => console.log(data.readme)); // Now you'll see the full README content!

Quick Way to Verify #

You can confirm this behavior using your browser's dev tools:

1. Load the URL directly and check the Accept request header in the Network tab—you'll see the full readme in the response.
2. Run your original fetch code, inspect the request headers, and notice the response has an empty readme field.
3. Run the modified fetch code with the explicit Accept header, and you'll see the full readme loads correctly.

内容的提问来源于stack exchange，提问作者Juan Picado