Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
Pig >> mail # dev >> Fixing a broken dependency // can we include a patched piece of JRuby source code in Pig?


Copy link to this message
-
Re: Fixing a broken dependency // can we include a patched piece of JRuby source code in Pig?
Won't a lot of people already have their version of JRuby and not want a special one?  I'm fine with having a patched version on github and referring it in our release notes.  I'm not wild about including a version of JRuby with Pig, for both licensing reasons and because our tar file is bloated enough as it is.

Alan.

On Mar 23, 2012, at 11:38 AM, Daniel Dai wrote:

> Hi, Jonathan,
> What bug is it? Last time when I try, it seems work well for me. We
> can leave a small hole and describe the limitation clearly in release
> notes/code comments/javadocs, we can also provide a link to the ticket
> tracking the issue. I remember we did something similar for javacc
> before. However, I don't think we shall include a JRuby patch in Pig.
>
> Daniel
>
> On Fri, Mar 23, 2012 at 10:01 AM, Jonathan Coveney <[EMAIL PROTECTED]> wrote:
>> First off: JRuby patch is almost done. It's passing tests, I have some more
>> to add, but I think the definitive version to work off will be out today
>> (assuming we can reconcile what follows :)
>>
>> I hit a bug in JRuby that is pretty impossible to avoid (it's a bug in the
>> way files were found on the classpath). I figured out the bug and let the
>> JRuby devs know and they patched master, but that means that our version is
>> still buggy. I put a patched version of the file in the Pig project pending
>> a new JRuby release, and this works, but there are two issues:
>> 1) Is this how we want this to be structued? It's weird to have this random
>> file in there, but on the other hand, it's a clean and clear fix.
>> 2) Is this legal? JRuby has a kind of odd triple license and I think you
>> can choose 1 for pieces that aren't explicitly GPL (of which there are very
>> few). One of those licenses is the CPL, which Apache says is kosher as long
>> as you're explicit, but I don't know. Is this fine? Should I talk to JRuby
>> or Apache legal?
>>
>> I suppose the alternative would be to publish a patched version of JRuby
>> (we could fork it on Github) and depend on that.
>>
>> I appreciate your comments
>> Jon
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB