----- Original Message ----- From: <[EMAIL PROTECTED]> To: "Tomcat Developers List" <[EMAIL PROTECTED]> Sent: Tuesday, April 09, 2002 1:45 PM Subject: Re: cvs commit: jakarta-tomcat-connectors/coyote/src/java/org/apache/coyote/tomcat3 Tomcat3Request.java
> On Tue, 9 Apr 2002, Bill Barker wrote: > > > > Unless I'm missing something. > > > > These fields are recycled per-request, so we lose the value of > > remoteHost/remoteAddr and will have to look it up again. If the buffers are > > identical, then the o.a.c.Request field is recycled as well. At least for > > remoteHost, I'd prefer the copy. > > But how would that work with AJP or other protocols ? Only the protocol > can know if the host/addr are unchanged and optimize for it - I don't > think the container has enough information to do that. > That's a good point. > I'm still missing something I guess :-) > > I can try to preserve the host/addr at the http11 level if you think > it's important ( just to make Remy nervous :-) > It's not extremely important. It's just a nice optimization for people who use remoteHost. If you want to concentrate on JK, I'll volunteer to make Remy nervous. :-) > > I'd be +1 for making get/setAttribute final on the core Request. > > +1 too. Not a big priority as long as you and Lary are reviewing > the commits and prevent me from braking too much :-) > > Costin > > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>