Apache Subversion to Migrate to Git | Hacker News Reader

Apache Subversion to Migrate to Git

infoq.com

134 pointsbigsassy35 comments

Now definitively an April Fool's joke.

Greg Stein's latest, and presumably last, comment on the JIRA ticket that started this furore: https://issues.apache.org/jira/browse/INFRA-7524?focusedComm...

(ETA: they've also added "April Fools" to the title of the ticket itself.)

Quoting:

  ----------------------------------------------------

Resolving as "Not a problem". We sure as hell don't want to do this. :-)

My major thanks to [~jfarrell] for the concept.

And plenty of thanks to all of my co-conspirators on this issue. Yes.. even my antagonist [~jimjag] was in on the ruse :-) … the Infra team handled this with perfect aplomb. And the Directors and Exec Officers came in with a perfect level of wrath. Our Subversion teammates showed a great sense of community and circling the wagons… Thank you all for making this work!!!

Last but not least…

THANK YOU to all of you who actually BELIEVED.

At first I thought this was a serious thing even though I knew today is April 1st.

While I was believing the news, the second sentence caught my attention:

This will allow the development of the Subversion repository be developed with Git

O_o

But that's already been possible for a good long while. They wouldn't be the first people to find git-svn to be a better front end to Subversion than the one that comes with Subversion itself :-).

Point is: if your mature project itself is a VCS, it sounds contradicting to use someone else's VCS to host your VCS :)

Give it a year and the Apache Software Foundation's sites will be switching to nginx.

Full disclosure: I'm an ASF member and one of the httpd developers and PMC members.

We're actually planning an experimental migration to nginx as part of Apachecon this coming weekend.

This has to be an April Fool's joke. Unless it's for legitimate competitive intelligence :-)

I'm STILL not convinced that this is a real story, and not an ongoing April Fool's prank. We'll see later in the week.

However, even if it DOES turn out to be legit... how dumb would you have to be to make any major announcement over the Internet on 4/1? Much less raise a question, conduct only a few hours of discussion and polling, and then make your final decision within a few hours... all on a day where most of the people commenting on that Jira ticket probably thought they were participating in a joke.

This is either the most well-played April Fool's joke since "OMG Ponies!!!", or else the absolute dumbest thing I have ever seen Apache do.

Real announcements do happen on April 1st. See for example gmail [1], perhaps one of the most important product announcements in the last decade.

Though I admit, I sometimes expect them to announce that it was a joke all along and shut it down.

[1] http://googlepress.blogspot.ca/2004/04/google-gets-message-l...

I think tons of companies battle with the limitations of "we should eat our own dog food." It tends to be a general rule, even when you as an organization have only marginal overlap with your target market.

I applaud Greg Stein's response (Stein is the founder and VP who was against the move):

> The short of it is: the Apache Subversion project chose this. We want to get our stuff coded and released. For our backend, we don't need the super-huge repositories that Subversion supports. Our project stores some binaries, but we can make Git work for us. We have no need for Subversion's fine-grained authorization ... shoot. We allow ALL ASF committers access to our repository. There are no barriers to the migration here, and some of the stuff that Infra has done for integration with GitHub? Pretty cool. Positives, and only little negatives.

I love seeing this commitment to the community and developers.

Edit though for completeness on the whole "dogfood" issue, this is another comment Stein made:

>So I'll be able to use the svn client to check out from the ASF git repo? Awesome. (...) We can dogfood svn against our own repo.

Unless the author actually has access to the private mailing list where the vote supposedly took place, I still call BS. I believe this was a prank. I'm familiar with many of the commenters on the JIRA and they were very much out of their normal character.

I really thought it was an April fools joke...

Tony Stevenson, on the Apache infrastructure team, seems to think it is:

https://issues.apache.org/jira/browse/INFRA-7524?page=com.at...

Although the discussion is still ongoing. Either it's real or they've really put some serious effort into this one and I will be quite impressed!

I read their Jira thread, and I was convinced it was a joke, too. The project manager mentioned some poll that no one else had heard of, and then other people kept coming up with reasons as to why it made sense to have had a short poll for such a major issue. I just assumed they were stringing everyone else along. :P

Honestly, if this isn't a joke, I think it doesn't really make SVN look bad (or dead). SVN's remaining proponents keep arguing that it works well in a corporate environment, while they agree that git is better for large, open-source projects. Well, SVN is a large, open-source project.

I'm still betting it's a joke, though.

I don't think it's a joke really. The ticket and a couple of comments would look like one. But multiple project leaders commenting more than once and (acting as if) getting quite annoyed with each other would be a massive waste of time for the whole org on a joke that doesn't need to be this long.

At this point, I think it's real...

Given the competition between SVN and Git, this would be a very poor choice for an April fool's joke. You don't put out false information that makes you look like you're ceding to your chief competitor, when you're in the weaker position.

Yeah... that would be the most pointless April fool's joke ever started.

Yeah, it was a joke. [0]

A pretty long one at that.

[0] https://issues.apache.org/jira/browse/INFRA-7524?focusedComm...

I'm kind of confused at the people who thought this was an April Fool joke. Really? If the announcement was on a news site or blog, maybe, but a 'Task' issue on the INFRA JIRA is hardly the appropriate venue. Although, it seems that @jimjag (the ASF director) was pretty pissed at the Subversion PMC and lead [1] for making the decision without consulting the community, just taking a vote from the committers.

Note that the interesting part of this is the move to use Git as the back end storage engine for Subversion eventually, with 'svn' being just a client API. The code is available now in the 'ra-git' branch. [2] Also, see @gstein's comment to the committer, noting that using a Git repo for Subversion code will simplify testing the new libsvn_ra_git client when it's available:

> So I'll be able to use the svn client to check out from the ASF git repo? Awesome. [...] We can dogfood svn against our own repo. Thanks, Stefan

[1] https://twitter.com/jimjag/status/450999640583053312 [2] http://svn.apache.org/viewvc?view=revision&revision=r1583639

SVN was designed as an API that could support multiple backends (they've already moved from one based on BerkeleyDB to the current ad-hoc one). Git was designed as a backend for a version control system that was supposed to get a cleaner frontend API but never did. Set aside the trollish April 1 announcement - a Git-based SVN could actually work.

Well this is embarrassing.

Isn't it? But, well played to the Subversion guys, et al. Was absolutely taken in... ;)

Some interesting discussion going on in this ticket https://issues.apache.org/jira/browse/INFRA-7524

This ticket is linked from the second word of the article. What does this comment add to the discussion?

I just think the PMC just figured out that they cannot compete with git and it's services (github, anyone?) so they are just doing the most natural thing to do using git. This is complemented with the ra_git upcoming release that will enable better svn/git integration. Having a svn that is able to talk with a git backend natively enables users to use tools designed for git also with svn. Many everyday git users i know complain about the verbose workflow that git imposes. Some of them even made scripts that mime svn update like behavior using git stash/pull --rebase and other svn like commands. So svn could really have a place being a git frontend with a simplified workflow

In short: may look silly, but in the long run i don't think it's a bad idea.

I'm really surprised people thought this was serious. That's like RedHat announcing that they're switching their servers to IIS.

Does this mean they won this year's April Fool's? https://issues.apache.org/jira/browse/INFRA-7524?focusedComm...

I just put a doc together on how to migrate from svn to git on my blog at http://jmoses.co/2014/03/21/moving-from-svn-to-git.html for all who is interested.

Well it was just closed as Not a Problem and "April Fools" was added to the title.

A joke or not, this makes SVN look bad.

As a SVN user I don't appreciate the message here. Every piece of technology is a platform that needs mindshare to survive. If it has no mindshare, it might as well be an excellent piece of engineering, but it'll still disappear.

In a perfectly rational world, with perfect communication, and universal understanding of everything, it wouldn't be a big deal - SVN and Git have different best use cases.

But in our real world, this action gives excellent bullets for the anti-SVN people to use now in any discussion bringing Git vs. SVN.

I can't see this working out poorly for the pro-SVN people one way or the other. If you'd probably be better off with git, you should use git - but svn will certainly prove workable enough in the same situation. On the other hand, if you'd be better off with svn, most likely due to its locking functionality, by comparison git is very likely to prove painful enough to warrant a switch.

Ideally this will cement SVN's position in its niche. Most git users seem to deny that this niche even exists, but I think it is there. I see SVN's more natural competitor as Perforce (which currently beats SVN as handily as git does, though from the opposite direction), but its maintainers seem not to be bothering to take it on, presumably due to their perception of SVN's position in the marketplace. Were they to admit that it's lost to git on what has become git's home turf, which it seems that they're doing, perhaps this would help focus attention elsewhere...

April Fools joke!!!

But so good because so very, very plausible.

If it is an April Joke, that would be ironic.

They think people would not take seriously, but as everybody knows, SVN is increasly being replaced with Git and that doesn't seem illogic even for them to use Git.

Reply on news.ycombinator.com