On Thu, 2015-11-26 at 16:42 +0530, Chetan Mehrotra wrote: > Thats strange. The test passes for me on Mongo 2.6.4 > > > OAK-3651 - Remove HierrachialCacheInvalidator > > May be the cache logic was hiding that issue as earlier cache would > only invalidate if the backing document got changed in Mongo which in > this single node test would not happen. So next entry would be > fetched > from cache itself. Now with above change the cache actually gets > invalidated and hence the issue surfaces. So may be a real issue
Filed https://issues.apache.org/jira/browse/OAK-3683 then to track this. I just realised that the test passes on Java 7 but fails on Java 8. Will update the issue. Robert > Chetan Mehrotra > > > On Thu, Nov 26, 2015 at 4:35 PM, Robert Munteanu <romb...@apache.org> > wrote: > > Hi, > > > > I get consistent test failures on MongoDB ( 2.6.9 and 3.0.2 ) > > > > > > Failed tests: testInterestingStrings[MongoFixture: > > MongoDB](org.apache.jackrabbit.oak.plugins.document.BasicDocumentSt > > oreT > > est): failure to round-trip brokensurrogate through MongoDB > > expected:<[?]> but was:<[�]> > > > > I ran a `git bisect` and, surprisingly, the commit that triggers > > this > > seems to be > > > > commit 086bae0f5495a4374d8fb2fb0e68687aef8d25ca > > Author: Chetan Mehrotra <chet...@apache.org> > > Date: Thu Nov 19 05:15:50 2015 +0000 > > > > OAK-3651 - Remove HierrachialCacheInvalidator > > > > git-svn-id: https://svn.apache.org/repos/asf/jackrabbit/oak/tru > > nk@1 > > 715092 13f79535-47bb-0310-9956-ffa450edef68 > > > > :040000 040000 e7b435248e6b621d9ef981380efb621b7160a645 > > 2283b37542dbc9112926a0f2ede77826fbdd9abd M oak-core > > > > Is anyone else seeing this? > > > > Thanks, > > > > Robert