Christopher Schultz schrieb am 01.12.2008 um 16:16:49 (-0500): > Michael Ludwig wrote: > > The very fact that object leak from one web app into another is a > > possibility seems to constitute a strong argument in favour of > > strict monitoring, be it only to shield oneself from errors that > > might be difficult to track down. > > It's rare that an object would "leak" from one app to another. When > Chuck says "more global", he means in the classloader-sense. That > typically means that the object would be passed from the webapp to the > container, which is usually a safe operation.
And the object couldn't dive down into another web app because there would be a ClassCastException. > > May I rephrase this into there being a duty for the Tomcat admin to > > enforce none of the web apps loaded into his container supply a > > library that the container itself already supplies? > > That's a reasonable request, but is often easier said than done. Since > webapps can come with any variety of JAR files (or even re-packaged > JAR files) it might not be possible to enforce this rule. But, if your > webapp breaks the rules, something usually goes wrong (like a > ClassCastException) so the webapp suffers, anyway ;) Thanks for these clarifications. Michael Ludwig --------------------------------------------------------------------- To start a new topic, e-mail: users@tomcat.apache.org To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]