Basically, you should create a branch for each fix and make changes there. When done do git rebase against master, and then submit a Pull Request via github. Then we review and merge.
The basic idea is that we can pull in small changes quicker than we can review and merge bigger changes that affect more unrelated stuff. Every set of fixes around a particular test(s) should be on its own branch tracking apache/master I'm currently reviewing the changes you made on your master branch, so far looks good, I might just cherry pick and push HTH -- Itamar Syn-Hershko http://code972.com | @synhershko <https://twitter.com/synhershko> Freelance Developer & Consultant Lucene.NET committer and PMC member On Tue, Feb 3, 2015 at 7:49 PM, Chris Anderson (OSG) <[email protected]> wrote: > I have also been looking into the tests in Lucene.Net.Index and have a few > fixes pushed to master. This is my first time checking in changes with GIT > and I am unclear if I did it correctly. :-) > > Change 1 > /src/Lucene.Net.Tests/core/Index/TestIndexWriterExceptions.cs > > Change 2 > /src/Lucene.Net.Core/Index/IndexWriterConfig.cs > /src/Lucene.Net.Tests/core/Index/TestIndexWriterConfig.cs > > Today I hope to make some progress in > Lucene.Next.Index.TestIndexWriterReader > > Chris > > -----Original Message----- > From: Laimonas Simutis [mailto:[email protected]] > Sent: Tuesday, February 3, 2015 3:54 AM > To: [email protected] > Subject: Re: Core tests > > I am looking currently at the failures in > Lucene.Net.Index.TestBinaryDocValuesUpdates and found the culprits. The > offending class is this one: > > > https://github.com/apache/lucenenet/blob/master/src/Lucene.Net.Core/Index/BufferedUpdates.cs > > First fix is easy, remove the ; that got in there by mistake on line 285. > The second issue is that in Lucene this class is using LinkedHashMap > structure. During the port, a HashMap was used instead which breaks certain > behaviors. Key feature of the LinkedHashMap that Lucene is relying on > appears to be the fact that the order of insertions is reflected in the > iteration of the values. You can see the failure if you run the > TestUpdatesOrder class. > > Has anyone dealt with LinkedHashMap before in .NET? Any implementation > suggestions / pointers? > > Due to the nature of this bug, it might be affecting more than just the > tests in BinaryDocValuesUpdates. > > > On Sun, Feb 1, 2015 at 8:01 AM, Itamar Syn-Hershko <[email protected]> > wrote: > > > I just fixed a couple more > > > > The Index namespace has some concurrency issues, but also many > > failures which relate to the Store ones. > > > > -- > > > > Itamar Syn-Hershko > > http://code972.com | @synhershko <https://twitter.com/synhershko> > > Freelance Developer & Consultant Lucene.NET committer and PMC member > > > > On Sun, Feb 1, 2015 at 12:34 PM, Laimonas Simutis <[email protected]> > > wrote: > > > > > With yesterday's PR which contains more merge related fixes, > > > https://github.com/apache/lucenenet/pull/66, we are reaching the > > milestone > > > of only 100 failing tests in core! Woot woot! (TC is showing 98, > > > locally > > I > > > get ~104). > > > > > > From Codec tests in current core, Lucene40 still has 4 failures, so > > > going after them next and then looking into Lucene.Net.Index test > namespace. > > > Heads up in case anyone else is working on the tests there. > > > > > > > > > Laimonas > > > > > >
