Hi Mark,
This problem lets me remember LUCENE-5355: This one affected the “nice” javadocs created after release. I backported the patch, so the one running the “nice looking javadocs” on java 7 with java 6 bootclasspath can be sure, it was really using the correct bootclasspath. Also the documentation howto on the wiki applies only to builds with this patch. Uwe ----- Uwe Schindler H.-H.-Meier-Allee 63, D-28213 Bremen <http://www.thetaphi.de/> http://www.thetaphi.de eMail: [email protected] From: Mark Miller [mailto:[email protected]] Sent: Monday, January 20, 2014 12:05 AM To: [email protected] Subject: Re: [VOTE] Release Lucene/Solr 4.6.1 RC1 If you are building the artifacts, you have to make sure it's using Java 6 rather than 7. IMO on linux, ant picks up what is set by the JAVA_HOME env variable. - Mark On Sun, Jan 19, 2014 at 5:57 PM, Shawn Heisey <[email protected]> wrote: On 1/19/2014 12:59 PM, Mark Miller wrote: > Unfortunately, I think SOLR-5645 is probably worth a re spin. I think > otherwise, collection reloads will be unstable vs 4.6. Collection reloads were unstable once before, SOLR-4805. A lot of people were hit by that one. I was very glad when SOLR-4805 was fixed and we could tell people that relief was coming. I admit that I don't know why registering in ZK again is bad, but if you think the issue warrants a re-spin, I'm inclined to put my vote in that basket. -1, please respin. In other news, when I tried Friday to run nightly-smoke on the SVN checkout used to build the RC, it failed, said that a jarfile was missing something like "Created by 1.6" ... which seemed more like a problem with my setup than the code. I can obtain the exact error message by running it again, but does anyone know what my problem might be? I have Java 6u26 and 7u25 installed on a Debian system. Thanks, Shawn --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected] -- - Mark
