Re: [Boost-users] Git vs SVN

Subject: Re: [Boost-users] Git vs SVN
From: Christopher Jefferson (chris_at_[hidden])
Date: 2011-01-28 09:03:07

On 28 Jan 2011, at 13:42, Edward Diener wrote:

> On 1/28/2011 2:12 AM, Anthony Foiani wrote:
>> 
>> Edward --
>> 
>> Edward Diener<eldiener_at_[hidden]>  writes:
>>> I hope such a discussion entails a very strong justification of why
>>> Git is better than Subversion. I still do not buy it, and only find
>>> Git more complicated and harder to use than Subversion with little
>>> advantage.  [...], but no one is bothering to explain why this
>>> latest thing has any value to Boost.
>> 
>> For my own development efforts, I've found Git to be an improvement
>> over Subversion in the following ways:
>> 
>> 1. Detached development.
>> 
>>    The ability to do incremental check-ins without requiring a network
>>    connection is a huge win for me.
> 
> Why do you mean by "incremental checkins" ? If I use SVN I can make as many changes locally as I want.

With 'git' you can commit those incremental checkins to your local repository. You can then decide later to either push them all up to the boost repository, merge them into a single commit, or abandon them.
>> 
>> 3. Experimentation.
>> 
>>    In my experience, branching is cheaper and much lighter-weight in
>>    Git than in SVN.
> 
> Please explain "cheaper and lighter weight" ?
> 
> It is all this rhetoric that really bothers me from developers on the Git bandwagon. I would love to see real technical proof.

I'm not sure what you mean by "technical proof", however we switched from svn to git at work. It is very easy to say "apply the commits X,Y and Z from branch A to branch B", whereas or to keep multiple branches in sync.

We found this basically impossible to do in svn and it is necessary to manually keep track of patches which need applying. boost appears to have a similar problem, requiring frequent manual diffs between head and release to find patches which have not been applied yet.

Chris