I suppose you are right, Ant doesn't fail. But, the thing is that according to the Hackystat Build data there was a failure. So, if I run a Hackystat analysis on build data, the analysis will show a failure.
In my development process, I would like to have a Hackystat Build failure only for Compilation and JUnit and exclude Checkstyle. In fact, I guess it would be best to leave out Checkstyle totally from the Ant Build Sensor. thanks, aaron ----- Original Message ----- From: Hongbing Kou <[EMAIL PROTECTED]> Date: Friday, October 7, 2005 10:59 am Subject: Re: [HACKYSTAT-DEV-L] Ant Build Sensor ignore checkstyle errors > Hi, Aaron, > > Please correct me if I get it wrong. I think both Checkstyle and > JUnit failure > will not fail the build. The build will continue and it will > report > checkstyle errors and > test failures at the end. > > Thanks, > Hongbing > > At 10:45 AM 10/7/2005, Aaron Akihisa Kagawa wrote: > >Cedric, > > > >Is there an easy way to ignore a type of error in the Ant Build > Sesnor?>For example, you seem to check three things Compilation, > JUnit, and > >Checkstyle but I want to ignore Checkstyle errors. My reason for > that is > >that we don't fail builds on checkstyle errors, but we do fail > builds on > >the other two. > > > >In addition, if we wanted to add checks to the build sensor, say an > >automated code inspector we would have to hack the > >BuildSensorAntListener class. In the future it would be cool if > it was a > >little more configurable. It seems like the Ant Build Sensor > consists of > >a collection of ant task "sensors". In fact, when I write the > checkstyle>sensor we would have sensors for Junit and Checkstyle, > which makes me > >think that we don't need build entry for every Junit and/or > Checkstyle>Failure. Instead the Snapshot UnitTest, Checkstyle (we > need a better > >more general name for Checkstyle), and Build data could be > connected by > >timestamp or something like that. > > > >thanks, aaron > > >
