Hi Denis, Thanks for investigating on this; and even more, for sharing conclusions. It is a pity I could not detect this side-effect by myself. It's also a pity I get this feedback so late, I thought introducing such a regression would have been the cause of immediate complaints. IMHO, it shows Jenkins is under-used by most people. About coverage, it's up to people who consume these coverage reports to decide what they need/want. I am afraid all this coverage things are currently not used at all, so no-one would care in case we remove Emma or Jacoco. Currently, we still have to use Emma if we want coverage since we don't have a way to get "continuous" reports easily with Jacoco (BTW Max made some tentative usage of Sonar, and Jonathan Fuerth submitted a GSoC subject for a Jacoco plugin in Jenkins). Then I think the wiser and simpler solution would be to disable Jacoco. We can think of Jacoco as a progress since it allows developers to consume reports directly in their IDE (see a 3-weeks older mail of mine for more details). But, once again, nobody use it, so we can ignore it. In a mid to long term view, we will get rid of Emma to use only Jacoco. Currently, we are stuck to Emma and then it seems like we need to disable Jacoco. I do that immediatly in parent. Regards, -- Mickael Istria Eclipse developer at JBoss, by Red Hat <http://www.jboss.org/tools> My blog <http://mickaelistria.wordpress.com> - My Tweets<http://twitter.com/mickaelistria> _______________________________________________ jbosstools-dev mailing list jbosstools-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/jbosstools-dev

Thanks for mentioning that Denis (M.), and also thanks to you for your investigation and help on this bug ;) On 04/03/2012 01:04 PM, Denis Maliarevich wrote: > FYI, > the related jira is JBIDE-11360. > > 3 апреля 2012 г. 12:08 пользователь Mickael Istria <mistria(a)redhat.com&gt; написал: > Hi Denis, > Thanks for investigating on this; and even more, for sharing conclusions. It is a pity I could not detect this side-effect by myself. It's also a pity I get this feedback so late, I thought introducing such a regression would have been the cause of immediate complaints. IMHO, it shows Jenkins is under-used by most people. > > About coverage, it's up to people who consume these coverage reports to decide what they need/want. I am afraid all this coverage things are currently not used at all, so no-one would care in case we remove Emma or Jacoco. > > Currently, we still have to use Emma if we want coverage since we don't have a way to get "continuous" reports easily with Jacoco (BTW Max made some tentative usage of Sonar, and Jonathan Fuerth submitted a GSoC subject for a Jacoco plugin in Jenkins). > Then I think the wiser and simpler solution would be to disable Jacoco. > > We can think of Jacoco as a progress since it allows developers to consume reports directly in their IDE (see a 3-weeks older mail of mine for more details). But, once again, nobody use it, so we can ignore it. > > In a mid to long term view, we will get rid of Emma to use only Jacoco. > Currently, we are stuck to Emma and then it seems like we need to disable Jacoco. I do that immediatly in parent. > > Regards, > -- > Mickael Istria > Eclipse developer at JBoss, by Red Hat > My blog - My Tweets > > _______________________________________________ > jbosstools-dev mailing list > jbosstools-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/jbosstools-dev > -- Mickael Istria Eclipse developer at JBoss, by Red Hat My blog - My Tweets _______________________________________________ jbosstools-dev mailing list jbosstools-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/jbosstools-dev

On 04/03/2012 04:38 PM, Max Rydahl Andersen wrote: > btw. I've seen this exact error in other cases than emma/jacoco conflicts and discussed it with Marc Hoffman before about possible fixes to this. That's very interesting. I'm trying to make some experiments locally to see if the issue may come from jacoco by itself jacoco+bytecode, or if it is a tricky issue that is consequence of Emma+Jacoco+bytecode.

Mickhael, I was looking at build failures for jsf and vpe components and found out that new jacoco coverage is the reason of tests errors like Tests in error: testAliasBean(org.jboss.tools.jsf.vpe.myfaces.test.MyFacesComponentTest): Invalid length 65532 in LocalVariableTable in class file coverage reports printed out to log are empty. So what happened is we lost old coverage and new one breaks the builds. We need to fix test failures ether by disabling code coverage for now or to make jacoco to work whichever is easier. I disabled coverage for jsf component nightly build from trunk and it helped to get all tests back, it failing now because there is no coverage report at all, which is easy fixable by disabling coverage in build configuration. WDYT? Denis _______________________________________________ jbosstools-dev mailing list jbosstools-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/jbosstools-dev

On 04/04/2012 08:54 AM, Max Rydahl Andersen wrote: > Any chance the stacktrace/context reveals which class is the problem and we could attach the related .class file to https://sourceforge.net/apps/trac/eclemma/ticket/196 as tried in https://sourceforge.net/projects/eclemma/forums/forum/614869/topic/5163999 > ? > In our case, it is clear that the issue seems to come from jacoco/emma conflict. Since I disabled Jacoco, a lot of things are back to work, and the test Denis made on JSF that disabled emma and kept jacoco was fine too. So we don't hit the issue as described on the ticket, and only have trouble with jacoco/emma conflict.