But this will break many apps. Mine for example.
Couldn't that be detected at jsp->java time? I believe that would not hurt that much. In fact... (not having read the spec) ...what's the whole point of creating the object at that time? IMHO, if it doesn't exist, just null, and everyone should be mostly happy. Or maybe I'm completely wrong? Well, in any case, detecting that at jsp->java time would give a compatible implementation.
If you'd like me to submit a patch explaining more clearly what I mean, please tell me which class is the affected one (the one that changed recently), and I will send that patch to you ASAP.
Yours,
Antonio Fiol
Tim Funk wrote:
The spec says using new() is OK. So tomcat does since it is *much* faster than reflection.
-Tim
Rupert Jones wrote:
Hi there
[Running Tomcat 5.0.19]
I have a servlet which forwards a Boolean object resource (request.setAttribute("someBoolean", someBoolean);) on to a jsp page. As you know there is no default constructor for this object (new Boolean()). In previous versions of Tomcat, well 5.0.16 and less, this didnt become a problem, as long as the Boolean was not null. In 5.0.19 the jsp compile throws an error because it includes code to make a new object were it null.
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
smime.p7s
Description: S/MIME Cryptographic Signature
