Return-Path: Mailing-List: contact tomcat-dev-help@jakarta.apache.org; run by ezmlm Delivered-To: mailing list tomcat-dev@jakarta.apache.org Received: (qmail 99814 invoked from network); 8 Dec 2000 13:01:12 -0000 Received: from merc94.us.sas.com (149.173.6.4) by locus.apache.org with SMTP; 8 Dec 2000 13:01:12 -0000 Received: from merc94.us.sas.com ([127.0.0.1]) by merc94.us.sas.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13) id YHYBAD6J; Fri, 8 Dec 2000 08:00:45 -0500 Received: from 10.19.11.6 by merc94.us.sas.com (InterScan E-Mail VirusWall NT); Fri, 08 Dec 2000 08:00:44 -0500 (Eastern Standard Time) Received: by MERC92.na.sas.com with Internet Mail Service (5.5.2653.19) id ; Fri, 8 Dec 2000 08:00:44 -0500 Message-ID: <764CA2FF49EC054BA086FC8253A52DD743327B@merc09.us.sas.com> From: Larry Isaacs To: "'tomcat-dev@jakarta.apache.org'" Subject: RE: A better proposal for compiling JSP's with debugging info Date: Fri, 8 Dec 2000 08:00:43 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" X-Spam-Rating: locus.apache.org 1.6.2 0/1000/N > My preference would be to set the options on JspInterceptor, using the > server.xml format. > Context and ContextManager shouldn't know/care about what happens in > interceptors unless it's absolutely needed. I think the biggest > priority > right now is to make sure all 9 core objects are "as simple as possible > and not simpler" - i.e. we should review all existing properties and > find > if they are indeed of general interest. I agree. For some reason, I seem to be having trouble actually following the design philosophy in practice. Something must be thicker than I realize. :-) Thanks for fixing the goof in my earlier commit. I'm also not doing a good job of making clear the reasons for this functionality. For the IDE we are developing here at SAS Institute, one of my assignments is to be the "integrator of the web server". The majority of people who will use our IDE to develop servlet's and JSP propabably aren't interested in learning about Tomcat's server.xml. They just want to be able to test what they develop. SAS's IDE is not intended to compete on the open market. Its main value is in the Java library that comes with it. I would hope that the functionality I'm proposing would help Tomcat achieve a better integration with other IDE's as well. I think it is a weakness of Jasper that doesn't make it easy to get debugging information. While making it easier to get debugging information, we might as well try to make the other options easier to control too. Right now, in Tomcat 3.2 I don't see a way to endable debugging info at all in server.xml, and I really don't want to ask users to make Tomcat specific changes to their web.xml's to achieve it. Tomcat 3.3 currently supports enabling it in server.xml, but I would like to bring it to the same level of functionality that I'm trying to reach in Tomcat 3.2. As an illustration of the level of functionality I'm trying to reach, it would be nice if you could have "keepgenerated" and "debugging info" normally off. If a problem turns up in a JSP page, turn them on temporarily, fix the JSP page, then turn them off again. I would like to be able to do that, for contexts I choose to allow it, without modifying the server.xml or needing to keep multiple server.xml's with different options in sync. For these JSP options, would it make sense then, if the JspInterceptor's constructor read a System property to update the built-in defaults using a string like I've proposed. The JspInterceptor's server.xml settings could then override these defaults. > Using notes is always a solution ( since it doesn't create any > dependency), but of course if you think this is important to have them > as > part of the core we can do that. > > On the other side, using properties on JspInterceptor is very easy and > consistent with the rest of tomcat configuration. Even now you can > specify > > , etc > > You can also set the JspInterceptor per context ( if you want different > options for different contexts ). > What do you think ? Adding properties to the JspInterceptor sounds fine. For some reason, I was thinking you would specify something like: which would be a little more complicated than I would prefer. Thanks. Larry