From: Eliot M. <mo...@cs...> - 2012-05-24 12:18:32
|
Dear DaCapo benchmark team -- As some others have noticed under some conditions in the past, I have runs that die when trying to run the eclipse benchmark. Specifically, they throw a Null PointerException in FullSourceWorkspaceTests at line 339, which reads: String jreLibPath = JavaCore.getClasspathVariable("JRE_LIB").toOSString(); This suggests that, under valgrind, getClasspathVariable returns null. I see this with Jikes RVM, HotSpot, and J9, when run under valgrind; for at least J9, the benchmark runs properly when not under valgrind. So, it's a valgrind bug, which I am guessing ultimately has to do with how some syscall is implemented (though it *could* be something else). Does anyone here have insight into what JavaCore.getClasspathVariable does and ho the data are set up that it refers to? It was suggests by Da Feng w.r.t. Jikes RVM that this can have something to do setting up dummyJre, but it seems to me that that would cause failure NOT under valgrind as well. In any case, I thought it reasonable to consult the community before undertaking possibly huge debugging effort to solve this ... Regards -- Eliot |
From: Steve B. <Ste...@an...> - 2012-05-24 12:56:16
|
Hi Eliot, I'm afraid I have not seen this before (not that I can remember, anyway). It is puzzling that it is valgrind-specific... --Steve On 24/05/2012, at 10:18 PM, Eliot Moss wrote: > Dear DaCapo benchmark team -- > > As some others have noticed under some conditions in the past, > I have runs that die when trying to run the eclipse benchmark. > Specifically, they throw a Null PointerException in > FullSourceWorkspaceTests at line 339, which reads: > > String jreLibPath = JavaCore.getClasspathVariable("JRE_LIB").toOSString(); > > This suggests that, under valgrind, getClasspathVariable returns null. > I see this with Jikes RVM, HotSpot, and J9, when run under valgrind; > for at least J9, the benchmark runs properly when not under valgrind. > So, it's a valgrind bug, which I am guessing ultimately has to do with > how some syscall is implemented (though it *could* be something else). > Does anyone here have insight into what JavaCore.getClasspathVariable > does and ho the data are set up that it refers to? > > It was suggests by Da Feng w.r.t. Jikes RVM that this can have something > to do setting up dummyJre, but it seems to me that that would cause > failure NOT under valgrind as well. In any case, I thought it reasonable > to consult the community before undertaking possibly huge debugging > effort to solve this ... > > Regards -- Eliot > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > dacapobench-researchers mailing list > dac...@li... > https://lists.sourceforge.net/lists/listinfo/dacapobench-researchers |