All,
In case anyone runs into this one again, we finally found a solution to this
problem.
The solution (sort of) that worked for us was to start Tomcat with the
"server" jvm rather than the "client" jvm. We also tried increasing the
default heap size but that only delayed the appearance of the problem for a
short while. My best guess as to why it works with the server jvm is that we
were running into some memory allocation problems with the client vm. Any
other ideas?
Much thanks goes to Jeffry Guttadauro who was seeing the same problem and
ultimately found this solution!
Troy
-----Original Message-----
From: Troy Landers
Sent: Wednesday, January 10, 2001 10:19 AM
To: '[EMAIL PROTECTED]'
Subject: RE: bizarre JasperException on Solaris box
But if method size was the problem then shouldn't it blow up when I try to
compile it from the command line as well? javac had no problem with the
generated servlet when I compiled it from the command line.
BTW, the java file is roughly 90k and the resulting class file is about 26k.
Thanks,
Troy
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, January 10, 2001 9:48 AM
To: [EMAIL PROTECTED]
Subject: RE: bizarre JasperException on Solaris box
Ask around again ;o) -- the method size limit is actually due to the 16bit
offset inside a class file. It's not javac. JDK 1.3 won't help either
(afaik). I had this problem when using a lot of tags on my page -- they
generate loads of code. You could either using scriptlets to call the
functions you want instead of tags, or (what i did) refactor the page to
minimise the number of actual tags present (I was able to loop a large
section and factor out many of the calls). What is the size of you're
class files / java files? If they're getting on for 64k this is probably
your problem. I'm waiting to move to jsp 1.2 with include/flush=false, to
get away from this problem (ie can have almost a method call of JSP).
Ken.
[EMAIL PROTECTED] on 10/01/2001 10:22:32
Please respond to [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
cc: (bcc: Ken X Horn)
Subject: RE: bizarre JasperException on Solaris box
I've asked around, and nobody here seems to believe that there is a size
limit for java methods. Further, these methods are not that extravagantly
long, namely under 2000 lines... Therefore, I doubt that is the problem.
However, upgrading the JDK to a release version surely wouldn't hurt.
In case we're onto something with this Stronghold thing, is there any
difference in how javac would be called from a page hit by Apache and
forwarded to Tomcat versus a page served by Tomcat itself. Could this be a
bug in the connector?
-----Original Message-----
From: Ralph Einfeldt [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, January 10, 2001 12:46 AM
To: '[EMAIL PROTECTED]'
Subject: AW: bizarre JasperException on Solaris box
If William Brodgen's assumption, that you are hitting a size
limitation in javac for a single method, is right, there are
two things that might help:
- use a more current JDK version
(1.3.0_01 is the newest version I know:
http://java.sun.com/j2se/1.3.0_01/index.html)
- restructure your JSP Source in way that the service
method gets smaller (put parts of your scriptlets in
own methods and just call those methods from your
scriptlets)
> -----Ursprüngliche Nachricht-----
> Von: Troy Landers [mailto:[EMAIL PROTECTED]]
> Gesendet: Dienstag, 9. Januar 2001 18:35
> An: '[EMAIL PROTECTED]'; '[EMAIL PROTECTED]'
> Betreff: RE: bizarre JasperException on Solaris box
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
