--- Niels Provos <[EMAIL PROTECTED]> wrote:
> On 8/6/06, Paul T <[EMAIL PROTECTED]> wrote: > > http support? > > > > Interesting. As a part of some (soon to be > > opensourced) server I have recently implemented a > > 'sufficient' subset of HTTP 1.1 > (get/POST/keepalive). > > There is a very simple HTTP layer implemented in the > current CVS > version of libevent. I wanted a small amount of > code that makes it > easy to quickly instrument http clients and servers. What I see looks nice as a very first step. > Feel free to > have a look and send patches if you like. Unfortunately, patching what is there now makes little sense, because robust keepalive will change a lot. > For patches, I would really appreciate additions to > the regress code > that tests them properly. It makes life so much > easier. I would say that your current regression test approach is also a very first step, because you have your client testing your server == both sides have the same idea about 'what is HTTP protocol'. In reality that is not the case. >From my experience, the only way to really iron out the http 1.1 protocol is to test your server with several clients (http 1.0 and 1.1) written by other people. On another hand, if you limit yourself with HTTP 1.0 only (like it is now) - then you can safely ignore all those tricky issues. May I ask how much of HTTP you are going to support? Rgds.Paul. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com _______________________________________________ Libevent-users mailing list Libevent-users@monkey.org http://monkey.org/mailman/listinfo/libevent-users
