My memory's a little hazy, but I ran into this specific issue when migrating from my Findbugs plugin to the Gradle Findbugs plugin.
Looking at the command line it was generating, it was listing every individual source or class file (can't remember which it was) as an argument. I was digging through the source for the Findbugs Ant task and noticed that in 2.0.0 those specific arguments were passed in via System.in instead of on the command line. I eventually realized that I hadn't removed the 1.3.9 dependency from my build script, and that was taking precedence in the classloader of the 2.0.0 version that Gradle was trying to provide. Once I removed that dependency, my build worked fine. Andrew Oberstar From: Peter Niederwieser <pnied...@gmail.com> To: user@gradle.codehaus.org Date: 03/07/2012 10:24 AM Subject: [gradle-user] Re: findbugs plugin gets IOException Andrew.Oberstar wrote > > There are improvements between 1.3.9 and 2.0.0 (or > whatever the default Gradle uses is) in passing the arguments to the > findbugs process. > Do you have any details on these improvements? Gradle's FindBugs plugin has defaulted to FindBugs 2.0.0 from the first release (milestone 8). Nevertheless, we get feedback that people on Windows are quickly running into the Windows command line length limit problem, which is a blocker for them. Cheers, Peter -- View this message in context: http://gradle.1045684.n5.nabble.com/findbugs-plugin-gets-IOException-tp5543987p5544634.html Sent from the gradle-user mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe from this list, please visit: http://xircles.codehaus.org/manage_email