Internal URL parser diverges from the spec in some parts

[d3lm]

Version

v14.16.0

Platform

Darwin Mac-Book-Pro.local 20.5.0 Darwin Kernel Version 20.5.0: Sat May 8 05:10:33 PDT 2021; root:xnu-7195.121.3~9/RELEASE_X86_64 x86_64

Subsystem

url › node_url.cc

What steps will reproduce the bug?

It's not directly a bug I have encountered but while I was delving into Node's internal URL parser I have noticed that it diverges from the spec in some parts, specifically for kHost and kHostname. Here's an excerpt from the spec:

And here's what Node.js does:

Notice how the 2. step from the spec is all the way at the bottom which means that the host was already parsed and assigned to the URL whereas in the spec, if stateOverride === kHostname then it returns early.

I wanted to ask if there is a reason why Node diverges from the spec here or if this is a bug?

How often does it reproduce? Is there a required condition?

/

What is the expected behavior?

/

What do you see instead?

/

Additional information

/

[jasnell]

Likely just a bug :-)

[RaisinTen]

A manual backport of #38742 to v14.x-staging will fix this

[RaisinTen]

Since #38742 is marked as dont-land-on-v14.x because of #37877 (comment), we can't backport the required fix to v14.x. Hence, I'm closing this issue as a wontfix. Please feel free to reopen if you think otherwise.