Look at IVY-387 [1] and the related issues (and vote for it if you want).
[1] https://issues.apache.org/jira/browse/IVY-387 Gilles 2008/2/12, Brown, Carlton <[EMAIL PROTECTED]>: > > I think I may have found a bug, can someone confirm this behavior for > me: > > > > It seems that the <deliver> functionality treats relative paths in an > unexpected way. In the deliverpattern attribute, it appears that paths > are resolved relative to the ${user.dir} system property. This is > different from the usual Ant behavior of resolving them relative to > ${basedir}. I will describe the steps to reproduce below. > > > > In a directory called parent I have an Ant script build-parent.xml > containing this line: > > <ant dir="child-dir" antfile="build-child.xml" target="publish"/> > > > > Of particular interest is the "dir" attribute, which sets the value of > the property ${basedir} in build-child.xml to child-dir. > > > > In the build-child.xml, I have this target: > > <target name="publish"> > > <ivy:deliver > > deliverpattern="output/jar/[artifact].[ext]" > > settingsRef="module-settings"/> > > </target> > > > > The expected behavior is for the ivy file to be delivered to > child-dir/output/jar. This succeeds when I run the publish target > directly from within build-child.xml. However, it misbehaves when it > is called via <ant> from build-parent.xml. The ivy file is delivered > to parent-dir/output/jar instead of child-dir. > > > > A workaround is to use absolute paths in the deliverpattern attribute by > employing the ${basedir} property, like this: > > <target name="publish"> > > <ivy:deliver > > deliverpattern="${basedir}/output/jar/[artifact].[ext]" > > settingsRef="module-settings"/> > > </target> > > > > This makes the task work as I intend for both tests. However, this > shows that Ivy is not treating relative paths as they are normally > treated in Ant, at least in the case of <deliver>. This is also true > of the implicit deliver that happens in <publish>, the same behavior > occurs there. > > > > Can someone confirm whether this is a bug? > > > > Thanks, > > Carlton > > > > > ----------------------------------------- > ==================================================== > This message contains PRIVILEGED and CONFIDENTIAL > information that is intended only for use by the > named recipient. If you are not the named recipient, > any disclosure, dissemination, or action based on > the contents of this message is prohibited. In such > case please notify us and destroy and delete all > copies of this transmission. Thank you. > ==================================================== -- Gilles Scokart
