Ian Lynch wrote:
I've checked out all the other ones. None seems to be suitable for OOo,
leaving subversion as the only reasonable alternative to CVS.
So will subversion do everything CVS does and more?
yes.
What are the benefits of using subversion?
It's an incremental improvement on CVS, not a radical change. Benefits
include:
* Directory versioning
CVS only keeps tracks of files, not changesets. This is significant
because adding a feature or fixing a bug often requires changing several
files. CVS gives you no simple way to express the fact that those
changes go together. Subversion is an improvement in this area.
* Renames
With CVS you can't move a file. All you can do is delete a file and make
a new one with the same contents. This means losing all the history of
that file. Also, you can't replace a file by a brand new one with the
same name. Here's an example from INGOTs:
Yesterday I renamed the file pupil.tpl to pupil_list.tpl and created a
new, totally different file called pupils.tpl. So, the history should
look like this:
pupil.tpl --> pupil_list.tpl --> ...
create pupil.tpl --> ...
This simple operation would wreck your CVS history. Not so with Darcs or
Subversion.
* Atomic commits
Say you edit 3 files to add a new feature. If there is a problem at the
server, you want either all changes committed or none. Subversion gives
you that and CVS does not.
* Branching and merging
You want to add a new, complex feature that requires a few weeks or
months to complete. You should create a branch of the project, work on
that feature, and when you're done, merge it with the main project.
Merging with CVS is slow and difficult. Time required is proportional to
the code size. For OOo (8 million lines of code), this is a problem.
With Subversion, it's constant time.
http://svnbook.red-bean.com/en/1.1/ch01s03.html
Will it enable easier participation?
That will depend on how it's used in practice. But it will reduce some
barriers (by making branches easier). The most inmediate benefit will be
just having a system that is easier to work with.
If it got one person full-time commitment, a cost of 3 man months seems a
fairly low price to pay. If on the other hand several CVS die-hards
leave, it could be counter-productive.
I've never met a CVS die-hard, I don't think they exist :-)
I think that getting one person full-time commitment shouldn't be hard.
The branching-merging alone should do that.
It will also make the current developers more efficient. At OOoCon, more
than one OOo developer told me how bad CVS is. I think the time savings
and convenience for 80 current developers will be worth 3 man months.
Look at it this way. If SVN makes the developers 10% more efficient, the
3 man months will be paid for after 1 month.
Cheers,
Daniel.
--
/\/`) http://oooauthors.org
/\/_/ http://opendocumentfellowship.org
/\/_/ No trees were harmed in the creation of this email.
\/_/ However, a significant number of electrons were
/ were severely inconvenienced.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
