[omniORB] omniORB development moved to Subversion

[Michael]

Actually you're missing a lot, git is a lot better for many reasons:
- working offline
- doing a lot of commits to a local repository without polluting the
"master" (if you decide to have one)
- Easy lightweight branching
- Merging works really really well (especially compared to SVN where is
this a total pain)
- Has an object database, so it can easily track files by identity (so
moving files within a repository will be tracked).
- Many more :) (there's so much material online, google for yourself)

It might be worthwhile to watch Linus Torvalds' presentation of git on
youtube (even so he of course is totally biased and has "strong opinions").

On the negative side, organizing your repository structures needs more
thinking upfront and you need external tools to emulate a more SVN like
behaviour (e.g. gitosis which works really great). On the bright side
correcting mistakes is really easy cause branching merging etc. are so
cheap (I assume this is the same for other systems like Hg or bitkeeper
which follow the same basic conecpts, but git seems to be the way to go
right now).

To me it always felt like svn is basically about perfectioning cvs's
misconceptions, so for some of my own prjects I actually moved back from
svn to cvs, for the rest I moved forward to git.

Then again, for ominORB's development model I don't think the versioning
system used really matters (how many people are actively committing or
using unstable branches?!??). And of course, almost everybody knows how
to use sv nowadays.

But in general I would suggest you to get used to rcs systems like git,
because they seem to be they way to go in the future (and if you look
how many oss projects moved to git or mercurial there are definitely
benefits of doing so)

cheers
Michael

Duncan Grisby wrote:
> On Friday 4 September, Stefan Naewe wrote:
> 
>> Why subversion and not git ?
> 
> Because I use Subversion in other things, so I understand it well. I've
> never used git. What am I missing?
> 
> Duncan.
>