[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Which is the best tool /process to migrate VSS (with history) to Subversion

From: Mark Phippard <markphip_at_gmail.com>
Date: Sun, 26 Jun 2016 10:49:52 -0400

On Wed, Jun 22, 2016 at 12:44 AM, Branko Čibej <brane_at_apache.org> wrote:

> On 20.06.2016 15:04, Nico Kadel-Garcia wrote:
> > On Mon, Jun 20, 2016 at 8:45 AM, Stefan Hett <stefan_at_egosoft.com> wrote:
> >> On 6/20/2016 1:07 PM, Yerra Babji wrote:
> >>> Hi,
> >>>
> >>> I am looking for best tool / process to migrate VSS folders (with
> history
> >>> ) to Subversion.
> >>>
> >>> Thanks in advance for your suggestions.
> >> 7 years ago when I migrated VSS to SVN in our company, I used this tool:
> >> http://vssmigrate.codeplex.com/
> >> Had some issues, but it turned out to do the job quite well (check out
> the
> >> reported issues there - think I reported everything I ran into back then
> >> which wasn't reported already).
> >>
> >> 7 years is quite a while ago. So maybe nowadays there are better
> solutions
> >> available. Just google and see if you find a more mature product out
> there.
> >> vss2svn seems to be suggested quite often and there also seems to be an
> >> commercial importer available from Polarion.
> > Unless you have a compelling need for history, I'd suggest you do a
> > simple export o a working copy from VSS to a simple export in
> > Subversiont.
> I have to say I'm mildly put off by your repeated suggestions on this
> list to throw away history to make things "easier" during migrations.
> That "compelling need for history" is very likely to be one of the
> reasons for using a version control system in the first place.

I do not want to get too deeply involved in this thread, but I will just
say that I do agree with Nico that it should at least be put on the table
for discussion whether the history is really needed. In the corporate
repositories I have seen from customers, more of them seem like complete
rubbish than something to be cherished. The commit messages are often
complete crap if not absent entirely, which means the history is largely
crap as well. You can start off on much better footing by being selective
and migrating the snapshots that you do care about. For example, maybe the
commit messages are crap but you still want the history on your release
boundaries. So you can pretty easily export the snapshots of the various
branches and tags and replay those in your new repository so you can at
least see how code changed across releases. And of course you can often
maintain the old repository as a read only archive to be explored when
needed. If you audit this usage over time you can also pretty easily
decide when it is no longer of value.

Anyway, Nico did not say make a blanket statement to not migrate history.
He said "Unless you have a compelling need for history". One obvious
compelling need is if you have good and useful history like we have in the
Subversion project. If that is the case, then I would be in favor of going
the extra mile to preserve it as well. I think anyone would.

But what I see from a lot of people is something like "We want to move from
VSS to SVN. How do we migrate our repository?" It is a valid question to
make sure they have considered how important it is to migrate all history.
I think that is just the default position everyone starts with because if
migrating the history is easy then it is one less decision to have to make.

In the little experience I have with VSS to SVN migrations, which is
largely second hand in providing support for the SVN part, the actual
migration is not terribly difficult. Meaning the tools that exist
generally seem to work properly. The issue is that a lot of VSS
repositories have hidden corruptions that do not surface until you try to
migrate the history. It seems like the corruptions are often in the
history or something, so it is only when you start trying to extract old
data that it surfaces. This is what has historically complicated these VSS
migrations and why some users say it is easy and others have a lot of
trouble. When you have these corruptions the extractions fail and the
migrators cannot run unless you manually fix the corruption.

Mark Phippard
Received on 2016-06-26 16:50:00 CEST

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.