djencks wrote: > > I think you have this situation, based on the stack trace: > > war with manfest classpath containing foo.jar > foo.jar has jcifs.jar in its manifest classpath. >
I thought of this, which is why I manually checked the manifest files of all the JARs and WARs in the EAR. None of them (seem to) refer to jcifs.jar, nor to any other JARs that *could* refer to jcifs.jar. djencks wrote: > > There are a couple of things we could change here, I'd be interested > in some opinions... > > -- if we run into this error print more information about how we got > there, such as all the jars examined for manifest classpath so far: > "foo.jar" will be on the list fairly near the end. > This would be very useful. It's a bit annoying for G to complain about a missing JAR, but give no information as to why it thinks the JAR is required. djencks wrote: > > -- allow people to have bad manifest classpaths, and just warn loudly > about them rather than refusing to deploy. > Or make it a configurable option - Strict or Relaxed manifest checking. djencks wrote: > > I'm a little worried about (2) because its so easy to miss warnings > no matter how large the type size they are presented in :-). On the > other hand if you include 3rd party jars in your ear they may have > bizarre and incomprehensibly wrong manifest classpaths that you don't > want to fix :-) > This change definitely gets a vote from me. I'd rather have the option to take a risk myself, rather than being rejected out of hand. If G warns me (it thinks) something is missing, and I subsequently get ClassNotFound exceptions, that's my problem. -- View this message in context: http://www.nabble.com/Missing-JAR-when-depolying-EAR-tf4068076s134.html#a11561730 Sent from the Apache Geronimo - Users mailing list archive at Nabble.com.
