On Jan 13, 2011, at 10:18 PM, Nigel Daley wrote:
Folks,
As I look more at the impact of the common/MR/HDFS project split on
what and how we release Hadoop, I feel like the split needs an
adjustment. Many folks I've talked to agree that the project split
has caused us a splitting headache. I think 1 relatively small
change could alleviate some of that.
CURRENT SVN REPO:
hadoop / [common, mapreduce, hdfs] / trunk
hadoop / [common, mapreduce, hdfs] / branches
PROPOSAL:
hadoop / trunk / [common, mapreduce, hdfs]
hadoop / branches / [common, mapreduce, hdfs]
Moving the source trees back together is ok, but will cause a fair
amount of churn for those of us that depend on the git versions of the
repository. Using Todd's hack may be able to fix it again at least for
each individual user.
I assume you meant to propose:
hadoop/ {trunk, branches/*, tags/* } / {common, hdfs, mapreduce}
which means that you can make checkouts, branches and tags with a
single command. Your proposal as stated would break all of the tools
that count on standard layouts of subversion repositories, such as the
subversion to git gateways and eclipse.
We currently have other stuff at the top level of hadoop: hive, logos,
nightly, pig, site, and zookeeper. Clearly hive, pig, and zookeeper
should be removed. The others are just versioned and aren't branched.
I'm fine with leaving them at the top level as "extra" bits, but it
should be decided.
-- Owen
