On 2018-04-12, Nicolas Lalevée wrote: > As far as I can see, the classpath used by checkForkedPath is the > proper one. The function which manipulates the classpath to add the > Ant runtime [1] is called before. So I should start looking into the > AntClassLoader which is improperly finding the Ant classes. Maybe we > should « isolate » it.
Sounds reasonable, so we can ensure it really only contains what will be on -classpath as we.. > Or maybe that check for duplicate ant jar is only useful when > includeantruntime is _not_ set to « no ». Since includeantruntime is > true by default, it is nice that Ant is printing a warning when it > also find Ant classes in the provided classpath, it is a common > pitfall. But when includeantruntime is explicitely set to false, then > I would say that the user know what he's doing, thus no need for > special check. I'm not sure. To be honest the check and message are there so we get fewer bug reports by helping people figure out their problem themselves. At one point in time this has been a very common problem, which likely predated the includeAntRuntime attribute. If we manage to isolate the classloader (the infrastructure should be there) then we shouldn't need to disable the check. Stefan --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org For additional commands, e-mail: dev-h...@ant.apache.org
