I am fine with removing the code in Trunk, after we have reviewed that we don't actually use it!
My only problem at the moment is the "reference size". If somebody has a way to detect compressed OOPS or not, I am happy to remove it! Currently this is a -1 about the removal of the call to get "reference size", which is still relevant because compressed OOPs are a really factor that we need to have (32 bits or 64 reference pointers is something we currently care about!). The JVM guys are about making it possible to retrieve, so in my opinion, we should keep unless Java 9 offers a public way to get them! My problems are more about crushing the JVM with un-mapping byte buffers in MMapDirectory (unfortunately, OpenJDK has no public way to do this). Uwe ----- Uwe Schindler H.-H.-Meier-Allee 63, D-28213 Bremen http://www.thetaphi.de eMail: [email protected] > -----Original Message----- > From: Robert Muir [mailto:[email protected]] > Sent: Wednesday, February 11, 2015 1:01 PM > To: [email protected] > Subject: VOTE: Remove sun.misc.Unsafe calls in RamUsageEstimator > > See issue: LUCENE-6239 > > in my opinion, the objections here are not technical, instead similar to > objections on https://issues.apache.org/jira/browse/LUCENE-6069. > "Its my code, don't try to fix it". > > The problem is, all this sheisty, platform-dependent, unsafe code, well, its > actually totally unnecessary. > > Here is my +1 to remove it. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] For additional > commands, e-mail: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
