Tomcat 3.2 was a bit more forgiving (which it shouldn't have been) about mal-formed JSP tags. One example I can think of in my own past, was that I was forgetting to put the /> to end a singleton tag, but 3.2 didn't care.
Once I moved to 4.x it took forever and a day to figure it out.
Is the tiny bit of JSP in that included page behaving ? are you sure?
Another possibility is whether it is resulting in static or dynamic includes. Sometimes an include statement results in a C style #include, where the JSP code is added to the same source file and there is a single Class output. Dynamic includes create a separate Class for each JSP, execute them separately and then concatenate the output. I'm not going to guess on the behaviour in your case, it's relatively vendor specific and just a useful thing to keep in mind. There are pro's and con's to both but it is a long discussion and may have been discussed before. I recommend a search first. Hope this helps you out though... Dave
Sometimes an include statement results in a C style #include, where the JSP code is added to the same source file and there is a single Class output. Dynamic includes create a separate Class for each JSP, execute them separately and then concatenate the output. I'm not going to guess on the behaviour in your case, it's relatively vendor specific and just a useful thing to keep in mind.
Actually, this is not vendor specific. The jsp:include action includes the output of the included page. The included page gets invoked at request time. The main page and the included page are two separate servlets. The @include directive includes the source of the included file. The included file gets merged into the main file at page-translation time. The main page and the included file become one big servlet. Cheers- - Marty [ February 13, 2003: Message edited by: Marty Hall ]
Originally posted by Marty Hall: Actually, this is not vendor specific.
You're right, but it still relyies on the vendor implementing it correctly. I still tend to refer to it as vendor specific so that people don't just accept the behaviour without trying it out on their container first. We're talking a few years now, but the behaviour of this changed dramatically in the BEA WebLogic server (4.51 and 5.1, I think).
Originally posted by Robert Paris: sure, but why does it result in buffer overflow as @include but not jsp:include?
I'm not Marty but have you tried debugs to trace the flow? I can't think of anything off the top of my head to casue this (but then I'm still grogy from surgery yesterday), it would be pretty specific to your code. Did you want to drop some sample code for us? Dave
Joined: Jul 28, 2002