hadoop-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Scott Carey <sc...@richrelevance.com>
Subject Re: [DISCUSSION] development process of Hadoop
Date Tue, 10 May 2011 18:06:22 GMT

On 5/6/11 7:16 AM, "Marcos Ortiz" <mlortiz@uci.cu> wrote:

>+1 for Git
>We migrated from SVN to Git for our completed infrastructure, for many
>- Git use much less space than SVN, all the changes are in a single .git

FWIW, svn 1.7 will have a single DB file too.  Though that project has
some chaos at the moment too and the release of 1.7 may be soon or a ways
away.  It is still slow over the network compared to git.  It is also
adding 'svn patch'.

>- Git is awesome for branching
>- Another great advantage is that there are many developers that know
>Git, and how the development process can be greatly improved.

There are also many developers who know svn, and many who don't know git.
That is not a clear win.

>PostgreSQL, one of my favorites open source projects that I use on my
>daily work, migrated the development process to Git from CVS.

Almost anything is better than CVS.

I don't feel that the primary cause of hadoop's situation is due to svn.
Git would help with merging patches that have become stale for sure, and
especially help on the client side for developers who need maintain many
concurrent contexts.  But there are many significant process issues at the
heart of the problem that are not due to the tools.

>Marcos Luís Ortíz Valmaseda
>  Software Engineer (Large-Scaled Distributed Systems)
>  University of Information Sciences,
>  La Habana, Cuba
>  Linux User # 418229
>  http://about.me/marcosortiz

View raw message