On Thu, 7 Nov 2002, James Mitchell wrote:
> Date: Thu, 7 Nov 2002 12:51:22 -0500 > From: James Mitchell <[EMAIL PROTECTED]> > Reply-To: Struts Developers List <[EMAIL PROTECTED]> > To: Struts Developers List <[EMAIL PROTECTED]> > Subject: RE: [Vote] Modify classloading in RequestUtils > > Absolutely, I never intended to commit unless I had buy-in from everyone. > > I don't claim to understand the full impact of this, so your comments as > well as Roberts are always welcome. > > > By reverting the code (my local copy anyway) back to its original: > test.junit <== SUCCESSFUL Total time: 16 seconds > test.tomcat.32 <== NOT SUCCESSFUL (Details below) > test.tomcat.33 <== SUCCESSFUL Total time: 54 seconds > test.tomcat.40 <== SUCCESSFUL Total time: 53 seconds > test.tomcat.41 <== SUCCESSFUL Total time: 59 seconds Personally, I would feel a lot better if we could try the modified version on lots of different servers (in addition to Tomcat), and with struts.jar either inside or outside the webapp (the tests above only put it inside). Can we figure out a way to: * Build a patched distribution of Struts 1.1 nightlies with this being the only change * Design a simple test strategy that can be easily replicated across the different servers and environments * Volunteer to test and report back on one or more environments? and verify that we're not going to destroy the ability to run Struts someplace we haven't thought about yet? Yes, we need to deal with Tomcat 3.2, and this patch does so -- but fixing support on 3.2 is not worth breaking support on something else. Craig -- To unsubscribe, e-mail: <mailto:struts-dev-unsubscribe@;jakarta.apache.org> For additional commands, e-mail: <mailto:struts-dev-help@;jakarta.apache.org>