On Thu, Aug 20, 2015 at 5:32 PM, Nicolas George <geo...@nsup.org> wrote: > Le tridi 3 fructidor, an CCXXIII, Stephan Holljes a écrit : >> Should the same be done for the Host-parameter sent by the client too? > > It can be discussed. > > On one hand, the Host field has a special meaning for the protocol, unlike > User-Agent that is only used for logging, so it makes sense to give it a > special meaning.
In ffserver the User-Agent is used to decide how to deal with a client. (WMP being a special case.) > > On the other hand, if there is a mechanism to access all fields, there is no > reason to have a special case for it. > > Actually, I suspect you could have written the code for the AVDictionary > (with a hardcoded total size limit to avoid DoS, for now) in the time before > my reply. That sounds pretty optimistic. Maybe it is trivial for you, but from my standpoint it looks like a rather large-ish undertaking to rewrite all the header-parsing code to utilize an AVDictionary. I also just noticed that the user_agent field was already present before I wrote any code in http.c, but I do agree that it should not be treated any different than other header-fields. > > Regards, > > -- > Nicolas George > > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > Regards, Stephan _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
