Submitted by As announced on the Haiku website and reported by IsComputerOn.com, Eric Petit has announced the development of a VMWare graphics driver for Haiku. Based on Be’s driver sample code and code inspired by the X.org driver, Eric’s driver so far implements RECT_BLIT and cursor functions (the latter are disabled as they are still buggy). The driver is already working as can be confirmed from this screenshot. On his initial post to the Haiku mailing list, Eric is asking for feedback, and has made the sources available in this tarball for those brave souls who would like to test the drivers.
This is great news! Hope someone is working on a mouse driver and NIC too!
p
Mouse already works – its not emulating anything but a “normal” mouse. NIC driver is in SVN
Actually, there is mouse driver for vmware in bebits.com that enables the clipboard between the host OS and the VM.
http://www.bebits.com/app/4296
There are two versions, one for zeta and the other for beos; dunno if it works with haiku though.
Edited 2007-01-13 04:11
The more of this the better, since you can bet that most of the world will try Haiku R1 in a virtualized environment before they think of reserving space on a bootable partition.
Now there is a NIC driver for VMware. It’s posted in a SoftwareValet package on BeBits. At one point Haiku used a GCC that could not compile the driver; does anyone know if that’s no longer true? Also, is anyone working on a SoftwareValet for Haiku (kinda important if we’re so interested in having the current repository of BeOS software run on Haiku)?
Edited 2007-01-13 01:28
If the network stack hasn’t been finished, what good is the NIC driver going to do?
I’ll be one that runs in a VM before installation to test the waters.
Without a NIC driver, what good is a netstack?
In any case, you may not have noticed, but the netstack is actually usable now… as in you can ping, irc, ftp, and http now (using wget or net+ since Firefox is having some non-network related issues)
I had read the stack is somewhat operational, but that it took some jumping through hoops to get it to communicate. I also don’t think any web browsing can be done in this state. Hopefully it will be done soon, as TCP/IP is a pretty big component for an OS to be usable today.
I’m looking forward to Haiku getting to beta. That will be a pretty big milestone in and of itself.
“Also, is anyone working on a SoftwareValet for Haiku (kinda important if we’re so interested in having the current repository of BeOS software run on Haiku)?”
One of my several BeOS/Haiku related projects is a simple .pkg file reader/installer. I don’t plan on recreating all the other stuff that SoftwareValet had.
I’ve done a good bit of reverse engineering of the pkg format, and have successfully extracted some of the zipped data inside a pkg file, as well as having a good idea of the various fields within the file.
But there is still a lot of work to do.
Regards,
Ryan Leavengood
Even if the vlance driver didn’t exist for Haiku – you could edit the VMWare VMX file to emulate an e1000 interface which is compatible with the ipro1000 driver that DOES come with Haiku…
I’m continually surprised at how many people don’t know this – as there are pictures of Haiku running Firefox/Net+ in VMWare from well over a year ago now.
I knew that there was an ipro1000 driver and that VMWare can emulate that NIC, but unfortunately it is not the default and most people have no idea. The vlance driver is FAR more important for the simple fact that it’s the default NIC under VMware.
If the VMX file distributed with the image was pre-configured – I’m not sure how important it would be. Ultimately, I gathered that the vlance driver wasn’t nearly as complicated as the ipro1000 anyway, so it’s probably best to use it if it exists… probably easier to debug
It’s nice to see Haiku adding new driver support for virtualization. The majority of people who want to try Haiku will probably do it in a virtual machine first.
It’s nice to see Haiku doing so well, and i think it’s come to a stage where we will see more and more applications for it.
…for those brave souls who would like to test the drivers.
How brave do you need to be to test a driver inside a virtual OS?
Awesome! currently haiku runs fast and smooth on vmware so now with a driver it will be even better
And there goes the little advantage syllable had over haiku,dont wanna seemsarcastic but. BAM!
Last time I tried Syllable is was so buggy. Haiku feels years ahead of Syllable.
Great to hear that the transmission author is back and working on BeOS/Haiku things.
Thanks for you efforts Eric.
Syllable has bugs. So does Haiku. Haiku has more developers, so I’d expect to see them fix more bugs quicker than Syllable. Let’s also not forget that Haiku can draw on existing, debugged, working code that Be released. On Syllable we’re having to write our own desktop and file manager.
What’s Open Source for if you can’t borrow from friends?
Vanders,
Did you ever consider using Open Tracker for Syllable?
Yes, briefly. Syllable isn’t trying to re-create BeOS and OpenTracker obviously depends on a lot of BeOS functionality, so we would have either had to ensure Syllable provided that functionality in the same way BeOS did or re-write large parts of OpenTracker to fit Syllable. Neither of which were really worth it.
Our Desktop & Dock are adequate for what is an Alpha version. They’ll be improved before Syllable 1.0
I think the biggest advantage Haiku has over syllable is that their developers know the code that makes the OS because they had to make it themselves. Syllable got the OS from Kurt, who now longer seems to work on it.
The same is true for Zeta. They inherited code from BeOS and are just tinkering. If palmsource opened the source, that would have made things easier for haiku at first but then you wouldn’t have a team that knows how to put a whole OS together. So I’m glad palmsource didn’t open the source and I’m glad they didn’t wait for a finished kernel either even though it meant more work. On the other hand, they still have a foundation.
It’s sort of how linux started. linus used the closed source minix as a foundation and made linux his own way so he is an expert. Then he was able to get the existing minix community on board. Community is also important for an OS.
You’d be surprised at how much of the codebase has been either written new, rewritten or changed from the original AtheOS code, so it’s not really a case of familiarity. I’m pretty sure that after five years, anyone can be pretty familiar with any codebase. Our biggest problem is manpower; Arno & myself have to know almost the entire codebase, from top to bottom. That’s a pretty tall order, I think you’ll agree.
It’s sort of how linux started. linus used the closed source minix as a foundation and made linux his own way so he is an expert.
You have the details wrong. Minix wasn’t and is not closed source. Linus only used Minix as an early development platform.
Community is also important for an OS.
This however, is true, and probably explains why Haiku has more developers and more people interested in it. Haiku started with a ready-made community of BeOS users to build from. Syllable started with the much smaller AtheOS community.
my point is that there is a difference between tinkering and rearchitecting and if you did much of the latter then you must be in good shape. You can then change the direction of the OS. You know the answer to the why as well as the what.
Minix had source available for buying the book but if I’m not mistaken, you couldn’t really change it (http://en.wikipedia.org/wiki/Linux). Linus came up with a monolithic kernel and that was a very different design than minix.
I think in the case of syllable it would help to have some sense of vision for the direction of the project. What problems are you trying to solve that nobody else is doing. What does “easy to use” mean to you.
Why do we want to change the direction of Syllable? We were quite happy with the direction that AtheOS was taking, which is why Syllable is a continuation of that. There is no need to re-architecture stuff that works, but we have certainly changed the parts that didn’t.
As far as I am aware we answered the question “Why?” nearly five years ago. We’ve has a long term roadmap we’ve been following almost since the creation of Syllable. Perhaps we need to make people more aware of our roadmap, but it’s not that we do not have one.
yeah…not to change the direction but to perhaps clarify more what that direction is