You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pig.apache.org by Olga Natkovich <ol...@yahoo-inc.com> on 2007/11/08 22:09:41 UTC

Porting Pig to hadoop 0.15

Hi,
 
Now that Hadoop 0.15 is out we need to port Pig to be able to work with
it. There are some interface as well as configuration changes which
makes it not backward compatible for Pig. My question is what how we
should do this. The safest way seems to be to branch what we have now
and make Hadoop 0.15 changes on the trunk. This way people have a choice
which version to use for some time with understanding that we would not
be back porting features or bug fixes onto the branch.
 
Comments?
 
Olga

Re: Porting Pig to hadoop 0.15

Posted by Doug Cutting <cu...@apache.org>.
Olga Natkovich wrote:
> I don't necessarily want to prevent changes. I just don't want to signup
> for backporting them from the trunk.

Then maybe just tag now, and, if someone wants to start a backport 
branch then they could branch from that tag?  Or do nothing.  Someone 
can always create a branch or tag later from any revision.

Doug

RE: Porting Pig to hadoop 0.15

Posted by Olga Natkovich <ol...@yahoo-inc.com>.
I don't necessarily want to prevent changes. I just don't want to signup
for backporting them from the trunk.

Olga 

-----Original Message-----
From: Doug Cutting [mailto:cutting@apache.org] 
Sent: Thursday, November 08, 2007 1:18 PM
To: pig-dev@incubator.apache.org
Subject: Re: Porting Pig to hadoop 0.15

Olga Natkovich wrote:
> The safest way seems to be to branch what we have now and make Hadoop 
> 0.15 changes on the trunk. This way people have a choice which version

> to use for some time with understanding that we would not be back 
> porting features or bug fixes onto the branch.

If the intent is not to change it, i.e., you just want a named revision,
then a tag might be more appropriate.

Doug

Re: Porting Pig to hadoop 0.15

Posted by Doug Cutting <cu...@apache.org>.
Olga Natkovich wrote:
> The safest way seems to be to branch what we have now
> and make Hadoop 0.15 changes on the trunk. This way people have a choice
> which version to use for some time with understanding that we would not
> be back porting features or bug fixes onto the branch.

If the intent is not to change it, i.e., you just want a named revision, 
then a tag might be more appropriate.

Doug