You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by Steve Loughran <st...@apache.org> on 2009/03/05 15:01:24 UTC
creating a branch for Hadoop-3628
Following up the discussion we had recently on doing big changes via Git
versus hadoop branches, can we try this out by creating a branch for the
service lifecycle stuff, HADOOP-3628
https://issues.apache.org/jira/browse/HADOOP-3628
I will put my changes in there and bring it up in sync with SVN_HEAD on
a regular basis, while I get the proposal into state that people are
happy with.
-this will let me hook our local hudson server up to it
-it will make it easier for me to work across machines
-I'm talking on this topic at apachecon, and it would be better to say
"play with this branch" rather than "apply these patches to SVN head"
I have all the signoff from HP management to grant the code to apache,
so there are no barriers to putting this stuff into the apache SVN
repository.
If people are happy with the idea, I will work with somebody competent
(Nigel?) to put in a branch and put my changes into it. Right now my
code is in sync with SVN head, unit tests seem to be working, etc, etc,
so it's only SVN that I need to set up
-steve
Re: creating a branch for Hadoop-3628
Posted by Steve Loughran <st...@apache.org>.
Owen O'Malley wrote:
>
> On Mar 5, 2009, at 11:41 AM, Brian Bockelman wrote:
>
>> Dumb comment:
>
> Not really
>
>> Although I know what issue HADOOP-3638 is referring to, it's not a
>> great way for the branch name to be self-descriptive.
>
> Actually, I find it much more natural to label branches with jira
> numbers than names. It might help if there is a HADOOP-*.readme at the
> top level that describes the branch in words.
>
Bone. Branch is in, with covering text
https://svn.apache.org/repos/asf/hadoop/core/branches/HADOOP-3628/HADOOP-3628.txt
Now that I'm sure that the text has appeared in a branch and not trunk,
I will leave the machine to commit all the other changes. That includes
a bit more than just the 3268 patch, as this machine contains most of my
little extra diagnostics tweaks that has crept in; more stack traces and
information on who is refusing to accept connections on
ConnectionRefused exceptions (that happens enough that I should make a
utility method, one that also includes the local hostname/address for
extra diganostics)
Re: creating a branch for Hadoop-3628
Posted by Steve Loughran <st...@apache.org>.
Owen O'Malley wrote:
>
> On Mar 5, 2009, at 11:41 AM, Brian Bockelman wrote:
>
>> Dumb comment:
>
> Not really
>
>> Although I know what issue HADOOP-3638 is referring to, it's not a
>> great way for the branch name to be self-descriptive.
>
> Actually, I find it much more natural to label branches with jira
> numbers than names. It might help if there is a HADOOP-*.readme at the
> top level that describes the branch in words.
good idea. I shall do that, then.
Re: creating a branch for Hadoop-3628
Posted by Owen O'Malley <om...@apache.org>.
On Mar 5, 2009, at 11:41 AM, Brian Bockelman wrote:
> Dumb comment:
Not really
> Although I know what issue HADOOP-3638 is referring to, it's not a
> great way for the branch name to be self-descriptive.
Actually, I find it much more natural to label branches with jira
numbers than names. It might help if there is a HADOOP-*.readme at the
top level that describes the branch in words.
-- Owen
Re: creating a branch for Hadoop-3628
Posted by Brian Bockelman <bb...@cse.unl.edu>.
Dumb comment:
Although I know what issue HADOOP-3638 is referring to, it's not a
great way for the branch name to be self-descriptive. Seems like it'd
be a possible source of confusion (but, on the other hand, it gives
folks an unambiguous way to look up precisely what the branch is; it's
just I can see myself fat-fingering HADOOP-3838 instead of HADOOP-3638
and working on the wrong branch or something like that).
Brian
On Mar 5, 2009, at 1:36 PM, Doug Cutting wrote:
> Owen O'Malley wrote:
>> I think we should keep the branch structure flat, with just
>> hadoop/core/branches/HADOOP-3638
>
> I thought you'd raise that right after I sent that last message. I
> agree: there are tools out of our control that we'd like to use
> (like git-svn & eclipse) that assume the branches directory has a
> flat structure. So I'm +1 for a flat structure.
>
> Doug
Re: creating a branch for Hadoop-3628
Posted by Doug Cutting <cu...@apache.org>.
Owen O'Malley wrote:
> I think we should keep the branch structure flat, with just
>
> hadoop/core/branches/HADOOP-3638
I thought you'd raise that right after I sent that last message. I
agree: there are tools out of our control that we'd like to use (like
git-svn & eclipse) that assume the branches directory has a flat
structure. So I'm +1 for a flat structure.
Doug
Re: creating a branch for Hadoop-3628
Posted by Owen O'Malley <om...@apache.org>.
On Mar 5, 2009, at 10:56 AM, Doug Cutting wrote:
> Steve Loughran wrote:
>> Following up the discussion we had recently on doing big changes
>> via Git versus hadoop branches, can we try this out by creating a
>> branch for the service lifecycle stuff, HADOOP-3628
>> https://issues.apache.org/jira/browse/HADOOP-3628
>
> That sounds fine to me. +1
I'm +1 too.
> We need a naming convention for such feature branches. They should
> incorporate the jira issue id. I wonder if we should put them in a
> branch sub-directory? so yours might be:
>
> http://svn.apache.org/repos/asf/hadoop/core/branches/feature/HADOOP-3638
I think we should keep the branch structure flat, with just
hadoop/core/branches/HADOOP-3638
I can't see that causing any confusion, and I'm worried that if we
have subdirectories in the branches, that the git and eclipse plugins
may have problems with them.
-- Owen
Re: creating a branch for Hadoop-3628
Posted by Doug Cutting <cu...@apache.org>.
Steve Loughran wrote:
> Following up the discussion we had recently on doing big changes via Git
> versus hadoop branches, can we try this out by creating a branch for the
> service lifecycle stuff, HADOOP-3628
> https://issues.apache.org/jira/browse/HADOOP-3628
That sounds fine to me. +1
We need a naming convention for such feature branches. They should
incorporate the jira issue id. I wonder if we should put them in a
branch sub-directory? so yours might be:
http://svn.apache.org/repos/asf/hadoop/core/branches/feature/HADOOP-3638
> If people are happy with the idea, I will work with somebody competent
> (Nigel?) to put in a branch and put my changes into it.
I don't think you should need help to do this. It's simply takes an
'svn cp' to start a branch.
Doug