The Drunken Lizard Pub

We have recently shifted some computers around in the office, which means an upgrade to our Linux workstation.

Right now, we are running Ubuntu 10.04 which is a Long Term Supported version. We mentioned in the past that we would likely use this version for Book III development. However, 10.04 is only supported until spring of 2013, and so we are now decided whether or not to upgrade to the latest LTS version of Ubuntu- 12.04. Or, do we put 10.04 back on the new Linux workstation?

Is there anyone here who has legitimate concerns over us upgrading to the newest LTS Ubuntu?

Whew! I read the thread title and was afraid you were thinking about dropping Linux support entirely.

I would personally suggest against upgrading to 12.04 for your Linux build system. That would result in builds that require the very latest library versions and are incompatible with many people's systems.10.04 will continue to get updates for server packages until 2015. Might that be enough for you to keep running it?

Much as I love Ubuntu, I think it might make sense for you to run CentOS on your build system. CentOS tracks RedHat Enterprise, which uses older technology and has a very long support lifecycle. If you wanted to install CentOS 5, it gets updates until 2017 and will produce binaries compatible with probably 99% of customers' systems.

Of course, you'll still get pretty good coverage even with the latest Ubuntu release, but expect to get some posts on the Linux support board about binary compatibility issues. It really comes down to how much convenience you're willing to give up to support a fraction of a fraction of your customer base.

Thanks for the feedback/suggestions. One of the reasons we selected Ubuntu is that it tends to be a very easy-to-use OS for most people. So, our strategy is that we will make sure it runs nice on a basic Ubuntu install, but beyond that more experienced Linux users will need to figure out how to run it on their personal set up. We just don't have the manpower to give 100% support to Linux.

Furthermore (this is where it turns into bad news) after Book III we are likely to change development platforms for the following game. Unity is our leading option right now, but as of today, they don't support Linux...which means, we may not support Linux either. More on this later.

BasiliskWrangler wrote:Thanks for the feedback/suggestions. One of the reasons we selected Ubuntu is that it tends to be a very easy-to-use OS for most people. So, our strategy is that we will make sure it runs nice on a basic Ubuntu install, but beyond that more experienced Linux users will need to figure out how to run it on their personal set up. We just don't have the manpower to give 100% support to Linux.

That makes sense. You should get very good binary compatibility between CentOS and Ubuntu, but of course you'll get 100% compatibility on the distribution you build on.

BasiliskWrangler wrote:Furthermore (this is where it turns into bad news) after Book III we are likely to change development platforms for the following game. Unity is our leading option right now, but as of today, they don't support Linux...which means, we may not support Linux either. More on this later.

Unity sounds like a great option for you. You lose Linux support, but gain consoles, mobile devices, etc. which is sure to be a net positive. It sounds like there's at least a chance of Unity supporting Linux in the future, and it will definitely support Flash. In the worst case I'd likely be able to play upcoming games on my Xbox or iPhone.

I'd be sad to see Linux support go - it was what originally drew me to Eschalon - but it wouldn't be the end of the world.

I still use Ubuntu 9.10 for my Linux builds.

It shouldn't really matter which version you use to build - you can build against older glibc symbol versions if you pass the flag "-D_GNU_SOURCE=1" to the compiler. This forces it to compile and link using the GNU extensions to the standard POSIX libraries, which are compatible with older distributions.

And don't forget about my Linux development articles - they've got lots of other tips for achieving a "universal" binary that will run on almost any Linux distribution.

I strongly recommend taking advantage of VirtualBox. Since VBox and various Linux distros are free, you can setup an effective testing pipeline. The only thing you cannot perfectly test is the graphics. The VBox display drivers are hit and miss, but you can at least verify that the libs are correct and that the game launches properly. (But for what it's worth, I've had fantastic success with OpenGL games in VBox. They render just fine for me.) Sound is pretty darn reliable in VBox. You can have various distros installing in the background while you work on other stuff.

I still use Ubuntu 11.04. I loathe the new GNOME and Unity. A major impediment for you developers would be going from GTK 2.x to GTK 3.x other than that, I guess it's just a question of taste.

I'm not an Ubuntu user, I use Fedora, but I've never had any trouble running Eschalon games on any version of Fedora so put me in with the "go with what you've been doing" crowd.

On a side note, I also hate Unity (the desktop environment) and GNOMES Hell, but it's a moot point since I've been using XFCE spins for years now.

Edit: As far as the Unity engine goes, Wasteland 2 will be using the Unity engine and have indicated that a Linux port is already in the works and they will be using the Linux port alpha source code for the Linux version of their game, so potentially going with the Unity engine may not mean ditching Linux after all.

Unity 4 will include Linux support:
http://www.youtube.com/watch?v=Bpi4J11Z ... r_embedded
Skip ahead to 4:00 for the Tuxtastic news straight from the horse's mouth.

sirdilznik wrote:Unity 4 will include Linux support:
http://www.youtube.com/watch?v=Bpi4J11Z ... r_embedded
Skip ahead to 4:00 for the Tuxtastic news straight from the horse's mouth.

Nice! Great news.

I'm running Kubuntu 12.04 on my laptop and it runs better than anything I've ever had on this craptop. ^_^

1.7 GHz dual core Intel P4
1 GB of DDR ram. not sure of speed
integrated graphics.




I hate Gnome. Give me KDE anyday.

BW,
my concern would be what happened to us before, where Ubuntu was using libraries not common to most standard Linux distros. eg: glibc version.
I would still prefer you used Debian stable as your base sytem, that way, everybody should have minimum difficulty in running the compilation straight out of the box. It's not the "ease of use of the distro" that you are compiling on that's the main criteria, rather it should be a distro that conforms to a base set of 'safe' stable versions, then every reasonably up to date distro will handle things just fine.

Unity dismays me. It's an horrible interface, and hopefully is already dying the death it deserves.
As above, go for the mainstream of users, not just a subset who just had to have the latest thing, regardless of whether it was an improvement or not. There's no reason why Unity users can't use a game compiled on a stable base Linux system (assuming there will still be any Unity users left by the time it came out).

Lastly, don't forget our cloth map, you promised!

It looks like Ubuntu LTS (current version 12.04) is what Steam bases their client on. So, the current Ubuntu LTS* edition is what we will continue to use in our studio. It will be up to the home user to get the game working on whatever other distro they have.


*LTS stands for "Long Term Support" - Most versions of Ubuntu are only supported for 6 months, LTS versions are supported for 3 years. The current LTS version is 12.04, which came out in April 2012 and will be supported until 2015.

BasiliskWrangler wrote:It looks like Ubuntu LTS (current version 12.04) is what Steam bases their client on. So, the current Ubuntu LTS* edition is what we will continue to use in our studio. It will be up to the home user to get the game working on whatever other distro they have.

I agree that this is a sensible course of action. You should get compatibility with almost everybody's systems for the foreseeable future with a minimum of hassle.