Hudson moves to Github | Hacker News Reader

Hudson moves to Github

h-online.com

42 pointsgizzlon23 comments

A statement from Sonatype on this:

First, I’d like to address some misinformation. The use of Github itself was never an issue. It was how the original movement of the sources to Github was executed, and why, that created tension. Github is just a tool and it is a better choice for source control, at least in the short term, for several reasons:

...

We proposed using Github on the Hudson list, and in short order agreement was reached and the move was initiated. Winston did the infrastructure work last night to push the sources for Hudson over to Github. It was really that Simple. It’s amazing how smoothly things go when you clearly communicate your intentions to all stakeholders. We hope to keep improving the infrastructure for Hudson so if you’re interested please join the Hudson Dev list!

http://www.sonatype.com/people/2011/02/hudson-moves-to-githu...

This guy seems incapable of making a statement about the Oracle Hudson without taking potshots at it's creators and former leaders.

Making a big decision is always a simple matter when you have only 2 people left to talk to.

FWIW, we have followed along this move at Etsy, where we've been using this CI server for many months now.

We started out with Hudson, but after observing the unethical shenanigans from Oracle (particularly the erasure of the author's contributions from their wiki), we moved to Jenkins.

This is ridiculous. Oracle and Sonatype are just trying to confuse the users into downloading the old and undeveloped hudson by moving to github. From my perspective this is a childish and immature move by both companies.

Undeveloped? I didn't follow every detail of the Hudson-Jenkins story but I thought the split was relatively recent. Have the codebases diverged that much since then? I would expect to still be relatively the same. Or are you referring to future development?

Yes, I am referring to future development. While it's possible that some development will continue on Hudson, the main contributors are sided with Jenkins.

How Ironic indeed.

Oracle insist that the hosting must be on java.net. Sonatype state that it's 'standard practice' to host on a 'blessed' repository.

Short memories indeed.

FWIW, here's the Oracle rep saying that Github wasn't an issue: http://java.net/projects/hudson/lists/dev/archive/2011-02/me...

Is this simply a move to sow seeds of confusion or point to something more fundamentally wrong with java.net?

I visited java.net recently to find out more about the Java Advanced Imaging project (which I discovered has some issues) but after seeing page after page of locked projects it will be a long time before I return, if ever.

The reason Jenkins nee Hudson was moving to GitHub in the first place was that Java.net was fundamentally broken; long outages, SVN was clearly not the right way for a project like Hudson to be run (commits came from all plugin authors, so the repo was touched constantly), fairly useless interface. The week long unnotified outage was just the final straw, the community had been asking for GitHub for a long time.

That they've now done this shows just how bitter, twisted and mercenary they've become. After this whole affair started from a move to GitHub, it's now magically the right thing? I very much doubt it. It's Oracle doing everything they can to try and screw Kohsuke and Jenkins.

Note that this is the same Jason Van Zyl that passive-aggressively deleted Kohsuke's edit history from the Hudson wiki.

> this is the same Jason Van Zyl that passive-aggressively deleted Kohsuke's edit history

I don't think this was passive at all. It was an aggressive-aggressive action, his response to Kohsuke's inquiries on the subject were passive-aggressive.

I'm not sure who has been coming off worse in these post-fork actions - Oracle or Sonatype (Van Zyl's company).

I'd say Oracle: most developers are already primed to hate them, and Sonatype is mostly seen as a side-dish/complicit. They also are pretty much unknown, so I'd expect most people to forget them soon-ish.

Van Zyl himself may be in a different position, personal dislike sticks a lot more.

Oracle already had the lowest reputation possible. Sonatype had no reputation before the incident, and now they have a bad one. So I think Sonatype loses, because the way consultants like Sonatype get hired is quite different than how Oracle products get bought.

Oh well, at least it's fun to be an egomaniac that destroys communities...

Given how long it takes to load each message in their mailing list archive, I'd say java.net has issues.

Wow, the bad press really got to Oracle. I'm impressed. Anyway, they have proven to be unreliable and have shown a very dangerous view of Open Source.

I hope this turns out well for the original contributors, but I'm afraid that months from now "the regular folk" will have forgotten this quarrel and chose the proven name Hudson over Jenkins. I will do my best to prevent this in my wider network.

If there was a way to down-vote a project on Github, I'd down-vote Hudson into oblivion. Just let it go Oracle, you lost. Stop being like the kid that pulls the girl's ponytail because she called you a fatty. You are a fatty.

This could be an interesting feature to have on GitHub for sure. :)

The closest approximation to this is how many people are watching a project. Sometimes when I can't figure out which fork is the main one, I use this.

What a mess. It's really too bad that Oracle seems to be fouling up every aspect of their relationship to the Java community. I am truly worried about the fate of JVM libraries and languages.

Will the SHA-1s match up and will it be easy to pull from one into the other?

If anyone's looking for an alternative (yes, an expensive one), I've been extremely happy with Atlassian's Bamboo CI tool for a couple years now.

and more to the point, Team City is free for up to 20 devs.

Potentially Hudson has a richer range of plugins, but both TC + Bamboo are a lot nicer to use.

Reply on news.ycombinator.com