Investigate java.lang.OutOfMemory: PermGen space error

[sbesson]

See ​http://hudson.openmicroscopy.org.uk/view/Bio-Formats/job/BIOFORMATS-merge-matlab-stable job.

Since build 67, the Matlab TestBfopen?.testSizeC test has been failing with the following stack trace

Error using bfGetReader (line 79) Java exception occurred: java.lang.OutOfMemoryError: PermGen space 
at java.lang.ClassLoader.defineClass1(Native Method) at java.lang.ClassLoader.defineClass(Unknown Source) 
at java.security.SecureClassLoader.defineClass(Unknown Source) 
at java.net.URLClassLoader.defineClass(Unknown Source) 
at java.net.URLClassLoader.access$000(Unknown Source) 
at java.net.URLClassLoader$1.run(Unknown Source) 
at java.security.AccessController.doPrivileged(Native Method) 
at java.net.URLClassLoader.findClass(Unknown Source) 
at com.mathworks.jmi.CustomURLClassLoader.findClass(ClassLoaderManager.java:760) 
at java.lang.ClassLoader.loadClass(Unknown Source) 
at java.lang.ClassLoader.loadClass(Unknown Source) 
at java.lang.ClassLoader.loadClassInternal(Unknown Source) 
at loci.formats.services.OMEXMLServiceImpl.createOMEXMLMetadata(OMEXMLServiceImpl.java:327) 
at loci.formats.services.OMEXMLServiceImpl.createOMEXMLMetadata(OMEXMLServiceImpl.java:313) 
at loci.formats.services.OMEXMLServiceImpl.createOMEXMLMetadata(OMEXMLServiceImpl.java:307) 
at loci.formats.MetadataTools.createOMEXMLMetadata(MetadataTools.java:489) 
Error in bfopen (line 113) r = bfGetReader(id, stitchFiles); 
Error in TestBfopen/checkFake (line 60) self.data = bfopen(self.filepath); 
Error in TestBfopen/testSizeC (line 91) self.checkFake(['test&sizeC=' num2str(self.sizeC) '.fake']) 
Error in TestCase/run (line 74) f(self); 
Error in TestSuite/run (line 85) this_component_passed = self.TestComponents{k}.run(monitor); 
Error in TestSuite/run (line 85) this_component_passed = self.TestComponents{k}.run(monitor); 
Error in runtests (line 136) did_pass = suite.run(monitor);

Setting -XX:MaxPermSize=256M in a java.opts file is sufficient to make the builds pass again (builds 80, 82 and 83) but -XX:MaxPermSize=128M restores the failing behaviour.
Note this failure is only detected on the stable job.

We may want to identify wether this exception corresponds to a memory leak in the code.

[mlinkert]

Likely worth checking that this is even still an issue, since a lot of things have changed in the Matlab stack over the last few years.

[sbesson]

​https://ci.openmicroscopy.org/view/DEV/job/BIOFORMATS-DEV-merge-matlab/152/ was modified to remove the PermGenSpace? configuration and was successfully passing.