On Fri, Mar 03, 2017 at 02:47:09PM +0100, Sebastian Benoit wrote:
> i thought i had fixed this in 2012, but apparently i never commited that
> diff even though i had oks for it. And i remember we had discussions about
> this in the past.
> 

Well, if it is not in the tree, it does not exist :)
Another anedote: I once had a half-working driver for zyd(4), I called
it zy(4) but I never committed it, somebody else came and wrote a new
driver instead, and I accidentally deleted my code.

> The question here is: do we need relayd to block this to protect whatever
> application is behind it? Do we gain anything from blocking this request?
> 

Yes, I want to keep relayd as strict as possible.  So we can change
and extend the list if there is enough evidence but I'm not going to
open it up completely.

> Anecdotal evidence(*) suggests that no one should rely on DELETE having a 
> body.
> 

OK, I fixed it with the proposed diff.

Reyk

Reply via email to