12:02:39 #startmeeting dependency tracking 12:02:39 Meeting started Mon Jul 20 12:02:39 2015 UTC. The chair is colindixon. Information about MeetBot at http://ci.openstack.org/meetbot.html. 12:02:39 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 12:02:39 The meeting name has been set to 'dependency_tracking' 12:02:44 #topic what happened last week 12:02:53 #chair zxiiro abhi_ 12:02:53 Current chairs: abhi_ colindixon zxiiro 12:03:37 #link http://54.148.146.38/dnvtools/sidemenu/ 12:05:07 #info if you go to version skew tab, you see a list of modules with more than one version 12:06:00 #info right now, it’s the groupId and artifactId, the project which the dependencies were found in, and the versions 12:06:24 #info it would be easier if we had pom file instead of project 12:08:08 #info e.g., it would be nice to see the left column stay the same: the groupId and artifactId 12:08:38 #info the second column should be the version # and should list all of the unique version #s for that artifactId and groudId 12:09:00 #info the third colum would be a list of all the pom files that reference the given triple artifactId, groupId and version 12:09:21 #info ideally, across all projects, so the pom file will tell you what project you’re in 12:09:53 #info if you have to, you would prepent the project name e.g., pom.xml would become controller/pom.xml 12:13:08 #info darkdragon says that some of the pom files aren’t being generate properly with an outofmemory issue or something like that 12:13:48 #action darkdragon to send out the error he gets when generating effective pom flles 12:14:39 #link https://wiki.opendaylight.org/view/OpenDaylight_Controller:Installation#1._Out_of_memory_error_-_java.lang.OutOfMemoryError:_PermGen_space usualy setting -Xmx to 1G or 2G and setting MaxPermSize to 512m or 1G, will help 12:15:46 zxiiro: http://54.148.146.38/dnvtools/sidemenu/#dependency 12:17:56 #topic next steps 12:18:39 #Info colindixon notes that the current dependency graph looks wrong, e.g., packetcable, nic, topoprocessing, reservatin, ttp, vpnservice, lispflowmapping and tsdr, all have no dependenices 12:19:29 https://jenkins.opendaylight.org/sandbox/job/dependency-tracking-lithium/4/console 12:20:12 #info he was able to get a jenkins job, but it was throwing an error that it couldn’t find maven 12:20:20 #link https://jenkins.opendaylight.org/sandbox/job/dependency-tracking-lithium/4/console the error is here 12:20:32 #Info zxiiro says that you have to declare a “top-level maven target” 12:21:06 #info zxiiro says the simplest thing to do is to decare a maven action with the goal of clean, to set the maven version 12:23:13 #action darkdragon will try to get the jenkins script so that before the “execute shell” build action, it does a “invoke top-level Maven targets” with a goal fo clean and Maven version of Maven 3.3.3 12:44:10 #topic fixing the jenkins job 12:44:58 #info one idea is to make this a maven job instead of a freestyle job, which might give us a valid maven for the whole time 12:45:30 #info another idea is to actually do the maven help:effective-pom as part of the mavne top-level project, but we need it to run on every project 12:47:57 #info the last option would be to instead use the maven exec to run things, but that’s actually not likely to work either 12:48:12 #info for now, let’s just use this path: /opt/jenkins/tools/hudson.tasks.Maven_MavenInstallation/Maven_3.3.3/bin/mvn 12:48:34 #info that will change whenever the version of maven changes, but for now, it’s probably fine 12:51:20 #action abhi wil change line 39 of his parsar.py script to use that path instead 12:52:56 should cleanup parser.py to conform to pep8 12:53:23 #topic actions for this week 12:53:53 #action abhi to modifiy the jjb file he has to first call maven clean with a maven goal and use maven 3.3.3 so that we get a maven executable to use 12:54:51 #action abhi to modify parsar.py and the web scripts to give the new table structure, i.e., (groupId, artifactId) => list ( versions => list of pom files) 12:55:17 #action abhi to work with zxiiro to figure what’s going on with help:effective-pom failing sometimes 12:55:31 #info after we can do those things, we can figure out if we belive the results 12:57:02 #info for what it’s worth, hard coding the maven executable path seems to be working 12:57:22 #link http://pastebin.com/xRckS4K9 the yaml file he’s using to generate the jenkins job 12:58:51 #link https://github.com/opendaylight/releng-builder/blob/master/jjb/ttp/ttp.yaml#L101 if you copy that maven-target: section into the yaml file you have just before the shell builder, it shoudl work 12:59:51 #info obviously change the goals to just clean 13:00:23 #info you should be able to ignore the settings and global settings 13:00:48 #info we’ll meet on Thursday as usual 13:03:27 #endmeeting