On 29 January 2010 19:22, Ben Smith <[email protected]> wrote:

> To answer your last question first, yes, commit authorization should
> be the same.
>
> Mercurial (aka hg) and git are what I have some experience with, along
> with SVN and CVS - so this is from that context.
>
> Both hg and git are geared around branchy development with multiple
> people doing commits.  As soon as you clone a repository, you've got
> the whole thing there on your disk.  You can hop back and forth in the
> project's history without talking with the server.  You can clone it
> from your disk again, without talking to the server.  Since these are
> local operations, they are very fast.  With subversion, going through
> history and changing branches requires server communication.
>
> Since you've got the repository local, you can make changes and commit
> them locally, then when you're satisfied with them, push them up to
> the server all at once.  This is also handy if you want to hack on
> something but you're going to be offline - suppose you want to do a
> few different things to the source code, fix a bug, add a feature.  In
> the SVN world, you'd end up with several files modified for different
> reasons and have to cherry pick commits to make a coherent history.
> With a local repository, you can commit each change in its context,
> then push them to another repository with a coherent history for free.
>
> One additional useful feature is that you can publish your repository,
> either temporarily hosted from your machine in a trusted environment,
> or set up something more secure for long term public hosting.  This
> way a maintainer could go through your repo and cherry pick the
> patches they'd like to put in the main source archive, and they can do
> this with full project history on both your end and their end.  As
> soon as they clone your repository, they have all the changes they
> need locally and don't need to hit your repository again until they
> want an update.
>
> These tools allow for more flexible development workflows while at the
> same time making it simpler and faster to branch and merge changes
> from many sources.
>
> And I'm not against SVN, it is an effective tool.
>
> -b
>
> On Jan 29, 10:13 am, Adam Bark <[email protected]> wrote:
> > On 24 January 2010 18:35, Ben Smith <[email protected]>
> wrote:
> >
> > > Also, to open up a fun can of worms - now seems like a good time to
> > > move to a dvcs, particularly Mercurial since google code supports it.
> > > Are there compelling reasons to stick with SVN?  I'd used git before
> > > this discussion came up in August, and have been working with both git
> > > and Mercurial since.  A lot of the issue ports and backports would
> > > have been easier with one of these, due to the overhead in tracking
> > > down historical diffs to find out which things changed and when.  I
> > > would personally prefer working with hg over SVN.
> >
> > I still don't really "get it" as far as dvcs's go can you explain it at
> all?
> > Does everybody that had rights on the SVN server get it on Mercurial?
> >
> > Cheers,
> > Adam.
>

Thanks a lot for that explanation it was very helpful, quick question
though; if two people are working on the same thing how does it go about
merging the two different sets of commits into something coherent?

Cheers,
Adam.

-- 
You received this message because you are subscribed to the Google Groups 
"pyglet-users" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/pyglet-users?hl=en.

Reply via email to