From our own experience the biggest problem with svn was merging of branches.

We often had features that took several months of development time. We didn't want to commit incomplete or buggy code to our master, so we created feature branches for them and called them "child workspaces" (cws). When a cws was finished, it had to be synchronized with the current master before it could be integrated. Developers also might want to merge in changes from master into their branch in between so that code change conflicts can be be reduced or solved early.

This merging of branches took hours with svn and was a PITA. With Mercurial it usually took only a few minutes and it was done completely on a local computer. No problems with slow or interrupted internet connections.

Beside that, if you ever worked with a DSCM, you never want to go back to a centralized SCM. It's as easy as that. The OOo repo is huge, but with Mercurial it took only a few seconds or minutes to create a clone or 10 minutes to get the source from it. Checking out from svn over http took hours. But I'm sure the pros and cons of DSCM are known.

*If* we had some options, I would like to think about the following ones:

- start with svn and decide later (at least in the next few months we won't need feature branches a lot)

- have svn for the master and do all work on a git repo elsewhere that at times is pushed to the svn master

- start with a git repo from the beginning

Regards,
Mathias

On 14.06.2011 16:48, Christian Grobmeier wrote:
ASF Infra provides GIT for readonly: http://git.apache.org/
This might already help for some people.

As far as I know the GIT writing is in the makings but finished. I
heard it will take another good while to become truth.

I don't want to start a flame war - but it has been said ASF SVN can
handle the volume of the code. And Git reading is available.

Do you see any serious problems if the code will go into svn in the
first step (which will last for a while)?

Otherwise I am afraid there are no options - of course maybe Joe or
somebody else from infra can comment this too.

Cheers,
Christian

On Tue, Jun 14, 2011 at 4:38 PM, Simon Phipps<[email protected]>  wrote:
One factor we need to consider is making it easy to collaborate with
peer/downstream projects, per the podling proposal. LibreOffice has imported
all the code into Git, and I'd assume that downstream projects like
RedOffice are using either Mercurial or Git as well. While SVN is obviously
the Apache standard, it would probably be better to use Git for this project
given both its size and the choices made elsewhere in the OOo community.

S.


----- Original Message ----
From: Joe Schaefer<[email protected]>
Date: 13 June 2011 21:05:52 GMT+01:00
To: [email protected]
Subject: Re: Proposed short term goals


1) volunteers are working on git hosting, but it's taking longer than
we expected.  Additional volunteers are welcome- start by subscribiing
to [email protected] (a public list).

2) svn has improved its merging capabilities since you last ran ooo on it.

3) as I mentioned our svn infrastructure performs well. we even have a
transparent
mirror in europe to cut down on cross-atlantic latency.



----- Original Message ----
From: Stephan Bergmann<[email protected]>
To: [email protected]
Sent: Mon, June 13, 2011 4:00:25 PM
Subject: Re: Proposed short term goals

What are the reasons?  Is it "just" that there is no  appropriate
infrastructure in place, or is there some general rule that  Apache projects
must run on svn instead of something else?

I'm asking  because currently OOo runs on hg, and the short time we did run
on svn (after  cvs, before hg), things were not working really ideally  for
us.

-Stephan

On Mon, Jun 13, 2011 at 9:39 PM, Joe Schaefer<[email protected]
wrote:

Not at this time, no.  FWIW our Subversion server is very  responsive,
even for largish codebases (yay  SSDs).



----- Original Message ----
From: Damjan Jovanovic<[email protected]>
To: [email protected]
Sent: Mon, June 13, 2011 3:37:53 PM
Subject: Re: Proposed  short term goals

OO.o is a very large project, it  might be overkill for Subversion. Is
Git an  option?

Damjan





Reply via email to