Content filtering is a different matter entirely, and usually happens in a way that is content-sensitive. The kind of "silent dropping" that's being discussed here is indiscriminate, with no consideration being given to the frame content. The fact of the matter is that silently dropping end-to-end frames without understanding why they've been transmitted is extremely dangerous. On Tue, Nov 12, 2013 at 2:25 AM, Nicolas Mailhot <nicolas.mailhot@laposte.net> wrote: > > Le Dim 10 novembre 2013 19:11, James M Snell a écrit : > >> -1 on the "may be dropped". As I've mentioned before, silently >> dropping end-to-end frames could significantly impact the semantics of >> the stream data and could have very bad unintended side effects. The >> result is that end-to-end extension frames become impossible to rely >> upon. The better (and more reliable) option is to require that >> end-to-end frames are either passed through untouched or the stream is >> closed with an RST_STREAM if the endpoint does not intend to pass them >> along. > > And the better option will live only as long as no one figures how to push > objectionable optional content such as ads, trackers or cookies with it, > then you'll see those frames dropped regardless of what the spec says. > > It's an unfortunate reality today that the people who control web sites > have not always the best interests of the people who consult them in mind, > and silently dropping parts of the stream at the browser or at any > intermediary level is here to stay. You can not specify morals in the spec > and giving an all-or-nothing power to one of the nodes involved in the > http/2 dialog is not going to work. > > Regards, > > -- > Nicolas Mailhot >Received on Tuesday, 12 November 2013 16:44:32 UTC
This archive was generated by hypermail 2.4.0 : Thursday, 2 February 2023 18:43:38 UTC