There's one Ivy Ant task that had been flying under my radar, the buildnumber <http://ant.apache.org/ivy/history/trunk/use/buildnumber.html>task.
It seems to me that the natural use for the *buildnumber *task is during a build invoked by continuous integration. A CI-invoked Ant target would go into an ivy.xml file and update the /ivy-module/[EMAIL PROTECTED] attribute with the ivy.new.revision property value obtained from *buildnumber*. This could come into play when, under "Dealing with integration versions" on the the Ivy Best Practices page<http://ant.apache.org/ivy/history/trunk/bestpractices.html>, you take the "create automatically a new version for each" approach. I'd be interested in hearing what other uses folks have found for the *buildnumber *task. Also (and actually this has nothing particularly to do with *buildnumber*), I'd be interested in hearing what mechanisms folks are doing for doing an automated update of ivy.xml files. Besides the /ivy-module/[EMAIL PROTECTED] attribute, there's the /ivy-module/dependencies/[EMAIL PROTECTED] attribute, which also can be a good candidate for automated updates in certain situations. I can think of two ways of automatically updating the content of an ivy.xml file: - Use regular expressions. Advantage: everything's right there in the build.xml. Disadvantage: regular expressions. Yuck! - Manipulate the XML programmatically, such as by binding Java objects to the XML schema. This has the disadvantage of requiring Java code (and possibly an additional library not already in the classpath), so you'd have to create a separate Ant task or Java executable. Or is there some other mechanism already out there in some Ant task (Ivy or non-Ivy) that I'm missing, like I'd missed the *buildnumber *task for a while?
