|Re: [jgit-dev] [RFC] Cassandra based storage layer for JGit|
On Fri, Oct 22, 2010 at 3:04 PM, Chris Aniszczyk <caniszczyk@xxxxxxxxx> wrote: > On Fri, Oct 22, 2010 at 4:30 PM, Shawn Pearce <spearce@xxxxxxxxxxx> wrote: >> The implementation sits on top of hector, a Java client library for >> Cassandra. Given the huge dependency tree involved here we might not >> contribute the code to the JGit project at the Eclipse Foundation. I >> don't have the desire to manage this giant stack of non-Eclipse >> libraries in Orbit. > > Yikes, Cassandra is a bit messy in terms of its dependencies. Other > than that, things don't look too bad... many of those dependencies are > already in Orbit. I created a few CQs for some of the main > dependencies to see what the Eclipse.org IP team will say. I'll put > this in my court to track. > - https://dev.eclipse.org/ipzilla/show_bug.cgi?id=4563 (hector) > - https://dev.eclipse.org/ipzilla/show_bug.cgi?id=4564 (cassandra) > - https://dev.eclipse.org/ipzilla/show_bug.cgi?id=4565 (jug) > - https://dev.eclipse.org/ipzilla/show_bug.cgi?id=4566 (perf4j) Is Thrift already in Orbit? Another option is we just code directly against Thrift. There isn't a whole lot we are getting from hector. It does give us connection pooling (which is important in Cassandra usage), but otherwise its mostly giving us an API that looks like the Thirft API would be if we were going raw to Cassandra over Thrift. Given our typical "low dependency" stance in JGit, I might be inclined to find a way to just use Thrift directly and do the rest ourselves in our Cassandra storage code. Similar to how we use JSch, but otherwise do everything else ourselves. -- Shawn.
Back to the top