Menu
▾
▴
#390 Out of Memory in MS Project Guesser
The error occurs as:
Java.lang.OutOfMemoryError: Java heap space
at au.gov.naa.digipres.xena.plugin.project.ProjectCompObj.readByteArray(ProjectCompObj.java:168)
at au.gov.naa.digipres.xena.plugin.project.ProjectCompObj.<init>(ProjectCompObj.java:56)
at au.gov.naa.digipres.xena.plugin.project.MsProjectGuesser.guess(MsProjectGuesser.java:99)
at au.gov.naa.digipres.xena.kernel.guesser.GuesserManager.getBestGuess(GuesserManager.java:374)
at au.gov.naa.digipres.xena.kernel.guesser.GuesserManager.mostLikelyType(GuesserManager.java:258)
at au.gov.naa.digipres.xena.core.Xena.getMostLikelyType(Xena.java:287)
at au.gov.naa.digipres.xena.core.Xena.getMostLikelyType(Xena.java:272)
at au.gov.naa.digipres.dpr.task.step.NormalisationStep.getAuxiliaryDataObjects(NormalisationStep.java:1284)
at au.gov.naa.digipres.dpr.task.step.NormalisationStep.doNormalisation(NormalisationStep.java:525)
at au.gov.naa.digipres.dpr.task.step.NormalisationStep.doProcessing(NormalisationStep.java:264)
at au.gov.naa.digipres.dpr.task.step.Step.process(Step.java:207)
at au.gov.naa.digipres.dpr.ui.task.StepProcessingThread.run(StepProcessingThread.java:106)
The issue turned out to be in the ProjectCompObj class. This class is used by the guessing to see if the file is a Microsoft Project file. The problem with the class is that it assumes that the file is, and at a few points looks for 4 byte fields representing the length of the next field. As the maximum size of a 4 byte integer can be very high, in the case of files that are not Microsoft Project files or corrupt Microsoft Project files then an attempt to get too high an amount of memory can result. To fix this I have set the maximum field size to 1024 bytes. If the field is large than this the object simply returns, guessing that the file is not a Microsoft Project file.
This change is on the testing branch. This change updates the version of the project plugin from 2.3.1 to 2.3.2.