Willy, I tested these changes in my environment and it all works. So you may move forward with these changes.
I really appreciate the help. - Ryan Brock On Thu, Sep 11, 2014 at 12:59 AM, Willy Tarreau <w...@1wt.eu> wrote: > Hi Ryan, > > On Wed, Sep 10, 2014 at 04:36:24PM -0500, Ryan Brock wrote: > > Here is the change I made on the very latest 1.6 master dev. It is only > in > > the one spot. There was discussion of changing it in a few other spots > were > > status was checked for less than 200, but I wouldn't know how to test > that > > correctly so I didn't feel it was a good idea to change it. > > > > This change works for the tests I did with just WebSocket connections. > > I've looked at the surrounding code a bit, and am seeing a few other > changes > to be applied for this patch to avoid possible side effects : > 1) we need to allow rspadd to work there. > 2) we want to prevent cache controls from being performed on 101 > 3) we want to avoid Connection header mangling on 101 > > Please could you check if the attached patch still works fine for you ? It > includes your change plus the other ones above. If that's OK, we can merge > it after checking if the doc needs to be updated or not (since we could > already delete/deny headers in 101, I think the doc is already OK). > > Thanks, > Willy > >