Issue #17835 has been updated by hsbt (Hiroshi SHIBATA).

Assignee set to naruse (Yui NARUSE)
Status changed from Open to Assigned

https://github.com/ruby/net-http/pull/26 was merged for Ruby 3.1.

But I'm not sure that we should backport this to Ruby 3.0 or old versions.

----------------------------------------
Bug #17835: Net::HTTP should automatically add Accept-Encoding request header on HEAD request as well
https://bugs.ruby-lang.org/issues/17835#change-94860

* Author: jrsyo (Shohei Maeda)
* Status: Assigned
* Priority: Normal
* Assignee: naruse (Yui NARUSE)
* Backport: 2.6: UNKNOWN, 2.7: UNKNOWN, 3.0: UNKNOWN
----------------------------------------
Currently, Net::HTTP adds the Accept-Encoding request header by default except for HEAD method.
https://github.com/ruby/ruby/blob/v3_0_1/lib/net/http.rb#L1489-L1493
https://github.com/ruby/ruby/blob/v3_0_1/lib/net/http/generic_request.rb#L34-L44

While this unlikely to cause any real-world problems, this isn't an ideal implementation according to the RFC below:
https://tools.ietf.org/html/rfc7230#section-3.3.2
> A server MAY send a Content-Length header field in a response to a
> HEAD request (Section 4.3.2 of [RFC7231]); a server MUST NOT send
> Content-Length in such a response unless its field-value equals the
> decimal number of octets that would have been sent in the payload
> body of a response if the same request had used the GET method.

Since server MAY (can) send the Content-Length header in response to a HEAD request, the client would see a Content-Length value of an uncompressed response if we don't send the Accept-Encoding request header.

```
$ curl -svo /dev/null 'https://www.ruby-lang.org/en/' -I --compressed 2>&1 | grep -i content-length
< content-length: 3380

$ curl -svo /dev/null 'https://www.ruby-lang.org/en/' -I 2>&1 | grep -i content-length
< content-length: 10932
```
 
Say, if the purpose of sending HEAD is to know the content-length beforehand (as an example) doesn¡Çt it make sense to also add the Accept-Encoding request header on HEAD requests as well? In fact, users almost certainly request compressed objects in the subsequent requests (e.g., GET) because of the auto-addition logic above anyway.

https://github.com/smaeda-ks/ruby/commit/020469cfcf5f910a22e9bbb24ba18ba4e147ff19

Thanks,
Shohei



-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>