On Mon, Jul 01, 2024 at 12:48:12AM +0200, Jörg Sonnenberger wrote:
>On 6/30/24 3:42 PM, Stuart Henderson wrote:
>>
>>There's also this, though:
>>
>>"However, a server MAY omit header fields for which a value is
>>determined only while generating the content"
>>
>>and specifically goes on to talk about HEAD.
>
>The intention here is to allow the server to avoid expensive 
>computations of the body when it is going to be thrown away anyway.
>A proxy certainly should not make any decisions in this regard and
>just preserve content-length if it is present.

I totally agree:

RFC 9110 9.3.2:
> However, a server MAY omit header fields for which a value is 
> determined only while generating the content. 

Relayd doas not generaty content. It has the Content-length header 
readily available.

Is there any good reason to drop the Content-length header from HEAD 
responses?
According to the commit message and comment it was removed because it 
would kind of lie about the content length:

Commit message:
> relay_read_http: strip out Content-Length if we strip the body too
> We should not forward Content-Length if the body is not also 
> forwarded.

Comment:
> Strip Content-Length header from HEAD responses since there is no 
> actual payload in the response.

I very much suspect, this was done under the assumption that a 
Content-length header in a HEAD response was wrong.

millert@, is there any good reason to keep this quirk for HEAD 
responses?


Christopher