Skip to content

Latest commit

 

History

History
30 lines (17 loc) · 1.68 KB

v3-LIMITS.md

File metadata and controls

30 lines (17 loc) · 1.68 KB

Known differences

As of 3.x release

  • Topics such as Cross-Origin, Content Security Policy, Mixed Content, Service Workers are ignored, given our server-side context.

  • On the upside, there are no forbidden headers.

  • res.url contains the final url when following redirects.

  • For convenience, res.body is a Node.js Readable stream, so decoding can be handled independently.

  • Similarly, req.body can either be null, a buffer or a Readable stream.

  • Also, you can handle rejected fetch requests through checking err.type and err.code. See ERROR-HANDLING.md for more info.

  • Only support res.text(), res.json(), res.blob(), res.arraybuffer(), res.buffer()

  • There is currently no built-in caching, as server-side caching varies by use-cases.

  • Current implementation lacks server-side cookie store, you will need to extract Set-Cookie headers manually.

  • If you are using res.clone() and writing an isomorphic app, note that stream on Node.js has a smaller internal buffer size (16Kb, aka highWaterMark) from client-side browsers (>1Mb, not consistent across browsers). Learn how to get around this.

  • Because Node.js stream doesn't expose a disturbed property like Stream spec, using a consumed stream for new Response(body) will not set bodyUsed flag correctly.