Hi,
it should be local with respect to a ruta script/module/block (-> RutaEnvironment) and local to the processed CAS. Recently, a bug was reported that could maybe cause this: https://issues.apache.org/jira/browse/UIMA-5423 The environments are not resetted if the script declaring/initializing the variable is called by another one which is called by the main one (two stacked scripts). The bug is fixed in the current trunk. Is this the case in your application? If not, I'll investigate it further. Best, Peter Am 17.05.2017 um 14:41 schrieb Hugues de Mazancourt: > Hi all, > > I was struggling for several days on a memory leak on a server running a Ruta > parser. > I guess I finally got the cause: I was using a STRINGLIST variable that I > though was local to an analysis (a JCas), but apparently, it is global to the > Ruta interpreter. In my case, since I was adding elements to the list and > never cleared it. Things went wrong... > Can you confirm that behavior ? > If this is the case, what happens when there are several Ruta threads > involved ? (I’m creating the instances with > UIMAFramework.produceAnalysisEngine(…) with multiple simultaneous requests) > > Thanks, > > Hugues de Mazancourt > http://about.me/mazancourt > > > >
