The XFree86 Project recently announced their final RC before the release of XFree86 4.4.
XFree86 4.4-RC3 Released
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
<rant>
Wow, another non-announcement.
When will these guys implement features that we all want? I want transparent windows and an approach to drawing the screen closer to windows so that the graphics card doesnt have to redraw the same dot on the screen 60 times because of the poor way this display server was designed. When will those features be added?
Features we want now, and not the beurocracy committy that is currently here is what we want.
</rant>
Now that no Linux distribution will include XFree86 4.4, this release is useless for everyone. Because of the new license no one will include it in their distribution. By doing this, the XFree86 committed suicide, and also harmed everyone else in the OSS world. Shame on you guys!!!
Just for the record, XFree86 is the only X framework that works to this time, and works good. But the XFree86 project is now history. I don’t want to know how the people, the Linux distributions and the BSDs who contributed feel, but if I would have contributed to the project, I would feel like s**t now. Why can’t you guys just get along and agree on a license?
I can’t think of any Linux or BSD distribution that didn’t give credit to the XFree86 project. No one ever claimed that they wrote the code and no one ever tried to discredit the XFree86 team.
And no, don’t place your bets on XOuvert boys and girls, that project is also dead. If nothing happens, for the next 2 years or so we’ll be stuck with XFree86 4.3.
I know that by now every other project in the OSS world united to solve this, but without the XFree86 team collaborating, it will take a long time to make a new X server for OSS. Imagine how much wasted effort there will be. And with all the problems that XFree86 team made to others, I think that this is the Tip of the Iceberg…
Good luck, I hope you’ll enjoy your work now, ALONE!
Good, the three distributors left that will go through
the hassle of including this package will be happy.
I hope the xserver and xlibs on freedesktop.org will released soon, so I think we can use these to instead of the XFree.
Sounds bad.
What’s the contention. GPL incompatible. How?
I thought I saw somewhere that a solution is a matter of updating the text in a lot of headers or something like that but it sounded doable.
well, let’s don’t forget we have Xouvert…
http://www.xouvert.org
is totally X11R6 compatible… forked from XFree86
Release 2 is due to April…
let’s keep an eye in that…
ah, and Xouvert is 100% open… always a new developer is welcome… AFAIK.
Transparent windows are silly. I tried using *real*[0] transparent windows on OS X, and all they do is slow down your reading speed since the text isn’t as readable due to background variations. I suppose it looks cool for about 10 minuntes though.
I just reread the Xfree license, and I really don’t see the huge issue people are making of it. All it seems to to say is that in the section where it says things like, “This systems uses the Linux Kernel” and “Portions derived from the FreeBSD project” you have to stick “We use Xfree” in. Or, before the Xserver starts up you can’t modify the messages talking about Copyright. It doesn’t seem like a big deal. Really.
[0] I define real transparency as the type where you see stuff behind the top window move, not like the goofy Linux terminal windows that show a shaded version of the background.
I would have loved to have XFree86 4.4 on my Slackware to, but now it won’t be included. It’s useless to compile form source because ATI won’t release any drivers for it since it won’t be included, and the same will happen with NVidia. And no one will adapt their projects to the changes in XFree86 4.4, like the new location of the freetype libraries, inside /usr/X11R6/lib…
Clause number 3 in the license is GPL incompatible, and since every other DE uses XFree86, it’s a problem…And they new it…
<childish>
I hate you, I hate you, I hate you!!! Ugh…I hate you for this, I had a bad day, a bad week, a bad month, and the worst year, and now YOU guys, I hate you!!!!
</childish>
Just my 2 pennies.
So what else does 4.4 have in it besides transparency, which is completely useless anyway. Is it faster, does it use the hardware better? If things are going to get stalled out with regards to X11 development then it just hurts open source in general because of petty licensing squabbles. Stallman strikes again.
“Stallman strikes again.”
This time it’s not Stallman, but Mr. Dawes…That’s life.
XFree86 team didn’t consult anyone on this, they’ve just decided to kick the community in the nutz…Sorry, but this is the cruel reality. Read before you post please.
Keen. We can finally get some support for input devices that aren’t strictly configured in XF86Config. Also, “substantially” better 2D performance for the nv driver.
That said, yeah, they can develop any feature they want that shoulda been there in 1998 at the latest, but no one’s gonna care or use it.
I’m seriously wanting to know; what stands between developers and a group that can release features and updates and bug fixes at at least half the speed of practically any other project? X is a basic piece of software. And it’s not like they’re slow because it’s such a basic thing; look at the Linux and BSD kernels. Updated all the time, and development of other software hinges on them entirely. What’s stopping any old group of developers from just coming along and saying, “Dude, that was easy, I just implemented everything that shoulda been done ten years ago!”?
If you don’t like the licence because it’s not compatible with the so-holy GPL, stop babbling and code.
The Anti-Linux has begun. XFree86 and Sun are teaming up to bring us the no future.
has any distro been released with this ass it’s default X server as opposed to xfree86? when can we hope to swap out our xfree86 w/ xouvert?
The GPL incompatibility isn’t a matter of personal spite or of politics. The Linux userspace is almost completely GPL. This new license is probably not GPL-compatible. Ergo, it is illegal to link these new X libraries to GPL’ed code. The authors of the GPL’ed code chose to put their code under that license, and you have to respect the terms of that license.
It is utterly a practical issue at this point. The XFree86 developers chose to put their code under a license that is (most likely) not GPL-compatible. That’s their prerogative. Distro makers are choosing not to include XFree86 4.4. That’s not only their prerogative, but their legal duty to all the authors of the GPL’ed code that they use in their distro.
Longhorn will utilize 3d hardware for its desktop and gui. If XFree86 is going to have to be forked because of a petty zealot licensing squabble, then linux will fall even further behind on the desktop.
“If you don’t like the licence because it’s not compatible with the so-holy GPL, stop babbling and code.”
Sorry, but if the Linux community expects borad adoption,
it cannot demand end users to “stop babbling and code”.
I’m an end user. I have always downloaded and used free
linux distributions. If XFree.org does not release
a product that can be bundled with a pure-GPL
distribution, then i really couldn’t care less. I’ll just wait
until the makers of such distribution find a replacement,
and use that insted.
If a 3d desktop is in the future of Linux than it has to come from the developer community and be protected by the GPL because it looks like vendors are going to patent it. It might not even be do-able because of the heavily restrictive legalities on software development. It’s illegal to develop a 3d desktop. Doesn’t worry me too much though, I like 2d GNOME just fine.
I think that doing away with Anti-Linux isn’t so bad, and it just goes to show how weak the position on the technology was in the first place if we can’t develop an alternative.
Where have you been? X has already been forked, twice. The most promising for is freedesktop.org, lead by Keith Packard himself. They’ve already got a version on their website that does MacOS-style transparency for X apps, and has a vector graphics API (Cairo). They are planning to base the X server on top of OpenGL and hardware-accelerate all drawing. They’ve got at least two years before Longhorn is released, and they’ve made a lot of progress already. Certainly, they have more tangible results *right now* than Microsoft has shown.
I’m totally aware of these forks, but does this mean there’s going to be another fork. I don’t know what the status of these forks are, eg. how buggy they are, etc. What about drivers from ATI and Nvidia. That’s a huge issue. It only delays things. Are you insinuating that Longhorn won’t come out? Yeah, they’ll just say it was a bad joke and they’ve decided to stop working on anymore OSs. I’m glad to hear that freedesktop.org is going to be based on opengl. That’s the way to go. When can we expect a release of this opengl-based XServer, or is too far off to even speculate at this point?
Hi
If you want to try this new server you can go ahead to fd.org and check this out. there has been no release but the code is available
Longhorn will be available a couple of years from now likely to be late 2006 but fd.org has more concerns than that
regards
Jess
So exactly which part of their new license makes it incompatible with the GPL? Since this is the first I’ve heard of it, I had a look at their site and it looks like they’ve just modified it somewhat to keep people/distro makers from trying to take credit for stuff they themselves did not write (*cough* Lindows *cough*).
I’m totally aware of these forks, but does this mean there’s going to be another fork.
——-
Maybe? But it’ll probably be a temporary one. Picking “one standard X server” really isn’t important since all X servers are compatible. As long as they don’t change the XFree86 driver layer (which would be asinine) it really doesn’t matter.
I don’t know what the status of these forks are, eg. how buggy they are, etc. What about drivers from ATI and Nvidia. That’s a huge issue. It only delays things.
———
Well, kdrive isn’t ready for prime-time, but its quite stable and compatible right now. I wouldn’t worry too much about drivers — NVIDIA makes money supplying workstation 3D cards that work with Linux. Remember, Linux is quite popular in the high-end 3D market these days, with companies like ILM using Linux desktops for their artists. Anyway, initial mutterings suggest that NVIDIA is supportive of the efforts at fd.o.
Are you insinuating that Longhorn won’t come out?
———
No, what I’m saying is that Windows releases are always late, which means that we do not have to worry about Longhorn for two years at a minimum. I also intended to preempt comments about “well, I’ll believe it when I see it” by pointing out that Longhorn’s Direct3D GUI is less tangible to users, right now, than the fd.o X server.
I’m glad to hear that freedesktop.org is going to be based on opengl. That’s the way to go. When can we expect a release of this opengl-based XServer, or is too far off to even speculate at this point?
———
Eventually? Very possible before Longhorn is available. kdrive already exists (and has existed for many years), has some very experienced developers behind, has a clear goal in front of it, and already has some advanced features (compositing desktop, vector library) ready.
Anyone want to start coding a public domain OS? Heh. No licenses, no license headaches.
When you do I’ll make sure to take your code and copyright it and lock you out while I make a bundle of cash sucker !
@Kingston:
Anyone want to start coding a public domain OS? Heh. No licenses, no license headaches.
I don’t know about Public Domain, but Free/Open/NetBSD are pretty close
Public domain is something that cannot be copyrighted, by definition.
I already use FreeBSD and DragonFly. That wasn’t my point
“Public domain is something that cannot be copyrighted, by definition.”
I think a minor clarification is needed. Copies of it can be copyrighted. The public domain code itself can’t be copyrighted. Can’t make too general a statement around here or else some folks’ll get confused.
Hi
“So exactly which part of their new license makes it incompatible with the GPL? Since this is the first I’ve heard of it, I had a look at their site and it looks like they’ve just modified it somewhat to keep people/distro makers from trying to take credit for stuff they themselves did not write (*cough* Lindows *cough*).”
Lindows takes credit for xfree86. no. i havent seen them do that. care to point out a link?
the xfree86 license modification is pretty similar to the old bsd advertising claus. it might look pretty innocent but is has practical problems
1) its considered incompatible with gpl which means that distributions cannot bundle it along with other gpl’ed stuff. qt which is gpl is affected by this. in general its considered advisable to be compatible with gpl when developing free software
http://www.dwheeler.com/essays/gpl-compatible.html
the above points out why. gnu.org has a set of articles on the bsd problems and licenses which they consider to be gpl compatible and they also point out whats their stand on it and why it isnt compatible and which license to use on which circumstances. you can read that to understand the issues.
2) the xfree86 lead guy claims that the new license is pretty close to the original bsd license which begs the question why didnt he just adopt that license instead of bothering to create a new one.
3) the new license doesnt bring any technical advantages ove the old one. every major distributions gives credit to xfree86. it has been pointed out in the openbsd lists( which has rejected the new license due to idealogical reasons) that attaching credit to documentation and software as a requirement in the license is pretty problematic in itself
regards
Jess
I think since this thing has been forked twice in recent history, the license change was intended to prevent people from taking the XFree86 codebase and creating their Xouvert without giving attribution to the original XFree86 team. If you fork now, you have to forever note the contributions of the core team.
The license wasn’t changed to make the distros give credit, it was changed to give the ingrates that forgot where they came from something to think about. IMO, it is going to have the opposite effect of making XFree86 that much more irrelevant faster and giving community support to one of the existing forks. Oh well, such is life when you let the egos rule the day.
How does that make it harder to fork? It certainly adds a cost to distributing derivatives of XFree86, but it adds the same cost to distributing XFree86 itself, so the net result is that its no harder to fork than it was before.
I quite agree with you, that may be the real reason for that licence shift. It may make merges between evolutions of the offical XFree86 and a Fork harder to manage ( although not impossible ).
http://dri.freedesktop.org/~daenzer/debian/dri-trunk-sid/ start building…
wich they could make some better docs on what to do- and maby fix the xserver-inst.sh wich currently points to some stuff that aren`t there, I want it easy, come on freedesktop, now`s your chanche to get the server out there, even if pre-beta
Is that the whole FD.O components for Debian GNU/Linix Sid?
well I could not get it to work, I did however try this:http://www.freedesktop.org/Software/XserverInstallGuide and this script:http://www.freedesktop.org/~sk/xserver-inst.sh some errors in that, and I am not that keen on using like 2 hours fixing stuff, and then I can`t get it to run. I`ll say this much- if they were to put some howto guide up on their page and some bz2 files I`d be happy as a kid on chrismas day.
howto edit config files etc. and it`s somewhat different to start etc: http://freedesktop.org/pipermail/xserver/2004-January/000785.html
I’m using Linux on my servers. No need for XFree86 anymore since I found OS X.
“How does that make it harder to fork?”
Because a fork has to carry the advertising materials and notices that says that it is based on XFree86. This does not make it harder per se, just makes it so that a fork is a less attractive option.
/g
“How does that make it harder to fork?”
It also makes it harder to incorporate changes in the XFree86 codebase into the fork. The forked projects now cannot take patches applied to the XFree86 project and incorporate them without permission of the original author. Starts making it administratively difficult to manage.
IMO, this is all lingering resentment over the Keith Packard issue and Xouvert, as well as the efforts of freedesktop.org.
I wonder… when will there be some full-featured 3dfx Voodoo
drivers? When will there be full-featured SiS 650 drivers?
When will 32bpp be supported for all cards that can handle it?
That sad fact is… They’re trying to “drown the fish”, they
say we’re working on it, and just wait until they can claim
“oh, this hardware is outdated, we don’t have time/no use to
finish it”. And it’s the same crap with many cards. None of
the cards in 5 PCs I have (and they are all very heterogenous)
work fully. I think that’s a reason why Linux will never
make it on the desktop. Even though I like Linux, the video
hardware support isn’t as good as Windows. OK, for newer
cards I don’t complain, but old cards this is unforgivable.
I thought X11 disbanded. I know they are a different group than XFree86, but it was my understanding that Xfree86 just copied and opensourced whatever features were incorporated into the “Official” X.
i think you are misunderstanding ALL the concept…
you must document youself before writing.
* X a protocol, the las version is 11 Revision 6,
also called X11R6 or just X11.
* X.org is an organization.
* XFree86 is a X server, X client and Xlib software
that use the X protocol version 11R6.
* XFree86 core team disbanded, but remain the developers.
* just one dev join X.org and leave XFree86.
* Xouvert and XServer from fd.o are software that also
use X11R6 protocol.
* Qt, GTK+, tcl/tk etc, use X protocol, so, use
Xouvert, or Xserver from fd.o, or XFree86, whatever
you like, always your software will run.
I believe the future is Xouvert (already have the grant of
fd.o)…
i think the best will be RH, MDK, SuSE, Gentoo, GNU must
join forces to Xouvert.
Xouvert code Release one is a fork from XFree86 4.3+cvs_code
(before license change).
ah, and Alan Cox is in the mailing list and he has already
contributed code…. that have to mean something!!!!
Xouvert rulz.
Ok thanks for this info then i won’t try it on Sid either. It tried on Gentoo. And couldn’t get it working either (in ~30 minutes)…
—
XFree 4.4rc2 also still has the old license. Between rc2 and rc2 things changed so either c2 or a CVS version but if everybody does this by theirselves without cooperation it’s gonna be a mess…
“* XFree86 is a X server, X client and Xlib software
that use the X protocol version 11R6.”
Agreed, and it is a “X Window System”
—
/. posted info about who’s gonna boycot the 4.4 release: http://yro.slashdot.org/yro/04/02/18/131223.shtml?tid=104&tid=189
Fedora is gonna do what i dreamed about, supporting multiple X Window Systems: http://www.redhat.com/archives/fedora-devel-list/2004-February/msg0…
Even a BSD won’t be be using it: http://marc.theaimsgroup.com/?l=openbsd-misc&m=107696705911864&w=2 fork is comming.. unfortunately nobody brings up in that thread that it’ll break up several X apps in Ports…
Interesting post which sums the whole problem up including references: http://yro.slashdot.org/comments.pl?sid=97287&cid=8315265
YMMV. That’s what you get for trying pre-alpha CVS code
When I ran the script it just kinda worked, though maybe the CVS tree have gotten a compile bug since then, or I had all the dependencies already.
PS> There is no config file for kdrive. It auto-detects your input devices.
I hear that Rocklyte Systems, designers of the Athene OS, are soon to release their Xserver. My feeling is that it will be leaner and meaner than Xfree86 by a mile, but I’m still waiting to see.
In my opinion, a good, lightweight display technology is a key component that is currently missing in the attempt to make a desktop Linux viable.
<quote>well I could not get it to work, I did however try this:http://www.freedesktop.org/Software/XserverInstallGuide and this script:http://www.freedesktop.org/~sk/xserver-inst.sh s</quote>
well,
I’ve just ‘updated’ & compiled fd.o’s xserver stuff..30 minutes ago,
just added “–enable-dri” to the xserver “autogen.sh”,
(use automake 1.7)
I’m using it right now (xati+xcompmgr+gnome-2.5.4)
and got no problem with it.
did someone make mention of Kdisk…?? I was wondering what that was?
“I’m using Linux on my servers. No need for XFree86 anymore since I found OS X.”
That’s all well and good, but what do you think that Apple’s gonna do for their next release? They just got a port of XFree86 in there, and now they’ll have to either rip it out in favour of another X implementation, or be like the only distributer of XFree86, and I just can’t see them doing that…
I hear that Rocklyte Systems, designers of the Athene OS, are soon to release their Xserver. My feeling is that it will be leaner and meaner than Xfree86 by a mile, but I’m still waiting to see.
Heh – you’re right, this will most likely be the case :-). It’s actually working now and is based on KDrive/TinyX. It compiles to 900k currently, so it’s small. At the moment I’m running it with KDE. It’s not any faster than standard X, but we’ll combine it with a lean window manager to keep the total package small and fast. An attempt will also be made to create a rootless X server for the Athene desktop, as implemented in Mac OS X.
Nothing is being forked here so there’s no controversy, but it is an interesting/different approach to move in a new direction and maintain compatibility with X.
Screenshots will be posted within the next 7 days at rocklytefiles.com.
jhbuild is like any other gnome building system/script, but…
It’ll install FD.o Xserver if u want to…
http://cvs.gnome.org/lxr/source/jhbuild/README