Re: [fetch] Content-Length on HEAD (#135)

> To be clear - nothing's going to break if you put C-L: 0 on HEAD requests, but there's no good reason to require them.

The good reason is compatability, not copy-and-paste. In the real world, people don't read the specs, bugs happen. 7230 is not a good justification for behaviour changes - it suffers from the normal aspirational vs reality problems, whereas Fetch() & friends tries to reflect the latter.

The complaint here seems to be primarily with Chrome's comment, which may be poorly worded, but was our best guess at _why_ this was happening, not a statement as to why we were doing it (we were doing it for compatibility with servers and other clients; we presumed it was for X, but regardless, saw it as necessary)

As to the question of would we change, we're not opposed to removing workarounds, especially if the compat story has improved, but figuring out whether it's safe to remove the workarounds often takes 3-6 months and dedicated monitoring, and I don't know if anyone's signing up for that. At the least, you could file a bug in the Chrome bug tracker to see if someone soul will take it up and champion it to measure interop issues.

(For what it's worth, this issue goes back to the very first commit of Chromium; I haven't dug in further to the pre-opensource code to figure out the full metrics, but each of those decisions was not made lightly - Compat was the goal)

---
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/fetch/issues/135#issuecomment-144648596