Another solution may be to exclude the SymmetricRangeTests from automatic test runs, and to keep this file somewhere for manual tests runs only. What do you think about this ? Can this be implemented in jtreg (for example by "@key hugeMemory" tag) ?
On Thu, Oct 17, 2013 at 2:18 PM, Alan Bateman <alan.bate...@oracle.com>wrote: > On 15/10/2013 01:27, Brian Burkhalter wrote: > >> Ping! >> >> This proposal could use more comments, not to mention review(s). >> >> http://mail.openjdk.java.net/**pipermail/core-libs-dev/2013-** >> September/021264.html<http://mail.openjdk.java.net/pipermail/core-libs-dev/2013-September/021264.html> >> >> Just to follow up on Paul's observation that the test runs with -Xmx8g. > I assume this isn't going to work on 32-bit systems, also I'm sure it will > cause problems in other environments too, esp. when running the tests > concurrently with a pool of VMs running the tests. Paul's suggestion to > have the test just pass if there is insufficient heap is one choice, > another might be to test on a smaller range when there isn't enough memory. > Another potential approach is to have it launch a new VM with -Xmx8g once > you establish that the VM is 64-bit and that there is sufficient memory > available. > > -Alan >