Please create an account to participate in the Slashdot moderation system


Forgot your password?

XFree86 News 200

PseudoMan was the first with the news: XFree86 3.3.4 has finally been released (yes, you can actually see the contents of the directory now). Rumour has it that the new release contains support for various Matrox cards, and may be the last release before we see 3.9 show up. Update: 07/20 06:05 by J : It seems that the first public beta of 4.0, 3.9.15, is now available. xinerama, here I come!
This discussion has been archived. No new comments can be posted.

XFree86 3.3.4 Released

Comments Filter:
  • by Anonymous Coward
    I agree that the font configuration in X is utterly archaic.

    However, using xfsft and a bit of tweaking XFree86's fonts look every bit as good as Windows (and certainly much better than Solaris's default font configuration):

    cd /usr/X11/lib/X11/fonts/ttf
    cp /mnt/c/windows/fonts/*.ttf .
    ttmkfdir > fonts.dir
    /etc/rc.d/init.d/xfs restart

  • by Anonymous Coward
    Any monitor that can do 1024x768 at a decent refresh rate (75Hz) can do 1152x900 at an acceptiable one (72Hz). (monitors are analog devices, it doesn't have to be on the box to be possible.. It has to be within bandwidth and frequency constraints)

    If your 17 can't do it then you are probably pushing the refresh too high.. It can do it, but only at a low refresh.. I would never push my refresh below 72Hz.

    Here, save up for a bit and check this out:


    $209 - 17" PX-780 .26 1280x1024x70Hz (there are cheaper ones that can meet 1152x900)
    $297 - 19" KDS VS-195 .26 1600x1200x75Hz
    $824 - 21" CTX EX1300 1800x1440x76Hz

    Sure, you can get cheaper ones.. But with hardware, you usually get what you pay for..

    At home I have a Viewsonic P817, running at 1800x1440x80Hz (soon to be 2048x1536@85Hz when my G400Max comes in).. My wallet may be hurting, (@~$1,400) buy my eyes have never felt better.
  • by Anonymous Coward
    XFree86 is done on a closed development model. Yes, the result is free and the source is freely available, but the development is still quite closed.

    I really think that if they were to change this, it would accelerate the pace of XFree86 development, which I consider to be way behind the curve in how fast it's evolving compared to other projects of the same significance.

    To some extent, though, the choice of a closed development model is to allow them to have greater ability to work with hardware vendors and software contributors that have restrictive requirements. There are clearly two sides to this coin; it lets things happen that wouldn't happen otherwise (more hardware support, more cool-neat-features), but it also lets things happen that wouldn't happen otherwise (being put in wierd positions by vendors such as with the NVidia stuff).

    It may be possible for the XFree86 team to organize their in-development tree into friendly parts and unfriendly parts, where the former is stuff that could be made available by anoncvs and the latter can't. This might be a compromise situation that could make more people happy than the current scenario.
  • by Anonymous Coward
    It's not very closed. All you have to do is send in a request to join. Has anyone ever actually been turned down?

    I can't show you mail that says 'bugger off' but I definitley got that impression.
    When I read about SGI releasing GLX as open source I also read this on the Precsion Insight site:

    Programmers who are interested in working with the DRI are encouraged to join the XFree86 Project.

    As I was interested in working on this project, I went over the the XFree86 site and studied their procedures. They say

    One of the XFree86 Project's scarcest and most valued resources is its developers. We're never short on things that need to be done, just short of people to do them. If you're interested in donating some of your spare time to help advance XFree86, we'd like to hear from you.

    To join The XFree86 Project as a non-voting member, send email to requesting a membership application form, and briefly state the reason why you wish to become a member. It is very rare that we knock back membership requests, but we are looking for members who will be active in developing and/or testing rather than people simply looking for early access to new code.

    So I wrote a short e-mail stating my reasons to join and asked if they have a task that was suited to introduce me to the project.

    The reaction was not a TO DO list, but a mail from XFree86 Prez Dirk Hohndel that told me rather to join some other related project that was run by another SuSE guy, Simon Pogaric. Thus I contacted him and frankly, IMHO he was not looking for any help, he had no TO DO list either.

    This was not what I expected. As I did not want to force my help on people I did not pursue matters further and looked for some other stuff (after all there is enough work).

    I might be paranoid but I have the feeling to have been gotten into some competition between two rivaling groups (Red Hat, PI vs. SuSE).

    The whole matter rather annoyed me because I think such large projects should have enough tasks (documenting, code cleansing, implementing) where good coding skills (in my case 18 years of programming, plus strong scientific background) would help and that would allow one to get accustomed to the code base.

    Other large projects like egcs or FreeBSD work that way and offer a kind apprenticeship system. With XFree86 I have my doubts.

  • XFree86 is done on a closed development model. Yes, the result is free and the source is freely available, but the development is still quite closed.

    It's not very closed. All you have to do is send in a request to join. Has anyone ever actually been turned down? Would you also claim that Linux development is closed because you have to actually get off your proverbial backside to subscribe to the linux-kernel mailing list?

    You can quite happily write new drivers for the current X servers, based on the released source. If you want to do something more involved, then you should join the team and get on the mailing lists. What could be simpler?

  • by Erich ( 151 )
    HMm... I'm still going to go for the DualHead support because (A) one day I might have enough room for a couple of monitors (maybe 1 monitor + 1 LCD screen?) (B) the second output can be TV-out for playing games on a large screen (Go Koules!), and (C) who can argue with a 360mhz ramdac?

    Plus, the 2D performance of the matrox cards is just amazing. If I can get 10% better performance in the 2D world over a TNT2 then even if the 3D performance is half as good I'd go for the Matrox card.

    But, that's just me. I use 2D much more than 3d.

  • by Erich ( 151 )
    I think this is largely Netscape's issue... I think Netscape is taking smaller fonts and making them larger by scaling... 9pt at 18pt or whatnot (but doesn't do a good job of scaling like TeX)
  • Okay, I think I found a way to get around this.

    In ~/.netscape/preferences.js, there is an option called user_pref("intl.font_spec_list". Then next to it is a LARGE list of fonts (it's all one line, by the way.) The final font name is the default font for most things. Change it's -0- to -160- or whatever you want. Then save it (make sure Netscape isn't running), and change to root and chown preferences.js for root, then use chmod to make root the only person that can write to it (but still let users read it.) I used the Midnight Commander (regular version in an xterm or console) to do the chmod'ing and chown'ing. Of course, if you run netscape as root, you'll have to make a script that copies the correct preferences file under a different name to preferences.js.

    God I hope this is fixed soon.
  • Hum. Well, Verdana looks nice in small sizes (compared to seems the text on the side bars on Slashdot always looked like murder on it.) But in bold, large sizes such as the headlines for Slashdot, it looks like crap. Not enough space. Any ideas?
  • Well, I run at 1280x1024, and the fonts in Netscape often look absolutely crappy. They have extra chunks all over the place -- looks like they thought the fonts should be getting antialiased, even though they really aren't..

    At any rate, I wouldn't mind if there were different systems for the simple 1-bit fonts, and another for antialiased font. Just use whichever works better for the situation. If you need speed, use the traditional stuff. If you want it to look pretty, use antialiasing (or whatever other fun technique they come up with).

  • I know that the font system tends to suck sometimes, but you shouldn't have to edit the fonts.dir by hand -- use 'mkfontdir'
  • The problem is LOTS of people (especially in the Linux camp) release broken fonts. Mkfontdir (under FreeBSD and IRIX) needs the entire fontstring embedded in the font file itself in order to work, not just the name of the font. What I've done is created a new driectory of these broken fonts where I can go through and create the fonts.dir by hand. For examples of broken fonts, check out
  • Your font problem is due to scaling: bitmap fonts don't scale at all well, and Type 1 fonts don't really have enough scaling hints to work well at screen resolutions. TrueType fonts are better for this --- ideally, one would use TrueType on screen and Type 1 for printing, but then you have to find fonts with matching metrics....

    Antialiasing has a bigger problem: X fonts are depth-1 bitmaps. Changing this would involve major, incompatible changes to the X protocol and Xlib, breaking every application. Or a whole new API added on top of the existing one, making the new server even more complex (= bloated and buggy, trying to make the two font systems work together properly) and making life hell for the X toolkit and application folks. This might be an idea for X Version 12, but I expect it won't happen in any X11 release.
  • by Brandon S. Allbery ( 500 ) on Tuesday July 20, 1999 @04:28AM (#1794740)
    > XFree could start by opening up its codebase a little.

    Once upon a time, it was open. Then certain Linux distribution maintainers (no longer around) decided it'd be neat to include outdated, buggy pre-alpha X releases in their distributions --- and redirected all the bug reports to the XFree folks. They Were Not Happy, and I don't blame them.

    The upshot here is that *we* screwed up, and the XFree folks got burned badly as a result. If we want to see more open XFree86 development, we're going to have to prove to them that we're not going to pull stunts like that any more.

    (Unfortunately, with Red Hat's fondness for including prerelease stuff in their distributions --- "prepatch" kernels and Perl "m" releases, to name some from the 5.x era --- I'm not sure I'd trust them to keep their mitts off prerelease XFree86 code.)
  • XFree could start by opening up its codebase a little. Last time I checked you had to be a developer to get early access to code. Nor could I find any public archives of the developers mailing lists.

    "Being a developer" implies a commitment that may discourage occasional developers and patch submitters (such as myself).

    Don't get me wrong; I respect and value the work that the XFree developers do.
  • Branden deserves a lot of credit. He's committed to getting X 3.3.4 into Potato before it freezes, and so he's psychologically psyched himself up for this release. I imagine that pre-release (ie: "staging") .debs will be available very soon, with 3.3.4 being rolled into Potato not too long from then.

    Now that I've said all that, Adam, just s/Branden/Adam/ and it's still true. =)

  • by Enry ( 630 )
    I have one of these puppies on order at Goes anyone have one and can comment on how well it works under X, and more importantly, what version of XF86 is needed?
  • This is just what we needed! Soon, XF86-4.0 will be ours! The world is a better place and the birds are singing! Plug it in... plug it in.

    Do I sound excited? YOU BET!
  • The ~4-5 year 17" monitor I am using has done 1600x1280@60Hz before. Also, it is pin sharp at 1280x1024. And monitor technology has improved greatly since then.
  • Can people please post some detils -- my poor little modem can't cope with the download :-(
  • "Broken" fonts? You mean like PostScript and TrueType fonts? There are utilities available for making fonts.scale files for TrueType and PostScript fonts. 'ttmkfdir' for the TrueType side, and something else I can't remember for the PostScript side. (Search Freshmeat.)
  • Well, they DID say the hardware support for the XF4 prerelease would be highly limited...
  • Ummm. I built it easily on a Debian potato box today. Just 'make World'. Or didn't you read the READMEs?
  • by demon ( 1039 )
    Well, yes, XFree86 keeps all the .cf's for all the commercial UNIXen provided in the X Consortium's source distribution.
  • Well, of course, XFree86 isn't even under the BSD license. It's still under the X Consortium license, has been for a long time, and hopefully will stay that way, contrary to some past attempts by the (not-so-)Open Group.
  • How did you get dual-head to work on Linux? I've got cards and monitors for x86 that I can't use at home and my sparc5 has two fb's but only one works with RH6. Even KDE only works on one screen under Solaris7. This is bugging me, but not enough to go mucking around for a solution. A pointer to HOW-TOs would be appreciated.

  • Some highlights... direct access to vidram, multimonitor support, anti-aliased fonts, remote display, 3D acceleration and remote 3D display...

    Where else have the others moved on exactly?

    At the moment Windows is still playing catchup on the remote display issue.
  • Still gives the BitBLT timeout error on GD5446 cards. The GD5446 chipset was awfully popular for it to be dropped from X in 1997.
  • by Matts ( 1628 ) on Tuesday July 20, 1999 @03:02AM (#1794756) Homepage
    I really don't get it. Font handling is a well understood technology, and yet XFree still falls short. Fonts (even true-type fonts) look terrible under XFree - they look _far_ superior under (for example) Solaris' X server. And I'm afraid to say it, fonts just look a lot better under MacOS or Windows. It's a real shame, because I think XFree would be a lot more usable with a decent font engine underneath - and yes, I've tried both TrueType font engines for XFree.

    Anyone know of any progress being made in this area?

    Also font setup is appalling. I can't believe you have to edit font.dir files for each directory - why on earth wouldn't the server do this for you? I was astonished at the amount of work it took to get a few TrueType fonts working before the perl TrueType tools came out to do some of the work for you.

    I guess you could consider this a bug report. :)


    perl -e 'print scalar reverse q(\)-: ,hacker Perl another Just)'
  • The ZX80 didn't have a rampack, AFAIK. That was the ZX81. :)

    Porting the entire of XFree86 4.0 into 256 bytes of memory... Am I allowed to skimp on the fonts?

  • About xFree86, i don't think it ports to solaris (unless it is solaris x86, which may work), but sparc computers use different hardware to the x86 computer, so this is a non-issue.

    I am running right now on Sparc Solaris using XFree86 3.3.3. It doesn't have some of the features I would like like loadable server modules but otherwise seems to work. All of the original Sun drivers from X11R6.3 are still there. Indeed they would have had to still be there for XFree86 to support Sparc Linux, as it does.

  • Nice... You partially answered the reason why I suggest not running a 15" unit above 1024x768... The bandwidth required for good refresh above that typically requires that you go out and pay more money, so you might just as well get a 17" unit for that amount of money. I'm all for smooth fonts, but I believe eye strain to be more dependent on refresh than dot blockiness. And I'm suggesting that you do get more pixels with dual head.

    In my dual-head system,
    I am currently running
    2048x768 : 1.5 Mpixels
    2304x864 : 1.99 Mpix
    2560x1024: 2.6 Mpix

    Normal single screen modes:
    1024x768 : 0.8 Mpix
    1280x1024: 1.3 Mpix
    1600x1200: 1.92 Mpix

    So I get as many or more pixels, more screen surface area, save money, better refresh if I go down a tube size.

    My suggestion wasn't about DPI. On paper, dots are almost everything. On screen, flicker happens. I'm not sure how close your eyes are to your screen, but I'm not anywhere near bothered by minute text blockiness as I am with flicker and high tube prices for marginal size and performance gain.
  • 15" = 1152x900

    You think THIS is a good idea? A 15" unit shouldn't be run above 1024x760 for ergonomic reasons.

    But really, I'm ahead of the game. People blow good cash on a 21" monitor when they should go dual head with 17's and 19's. I'm willing to gamble that the two 19's cost less than a single 21" can give you better than 70% more total pixels at a better refresh rate with more than 70% additional screen surface area. That is from my own analysis. I'd post the numbers, but I lost them. I considered getting a 19" when the costed about 400$, but I found a pair of cheap 17" for about 350$, an extra video card for the remainder savings (Matrox Millennium 8MB - solid units) and come out way ahead. MetroX also supports multiple screens on all Matrox products.
  • That's 1.

    So far, I see no great contradiction to his (admittedly made-up) statistic.
  • 15" = 1152x900
    17" = 1280x1024
    19" = 1600x1200
    21" = 1880x1440

    Great idea. However, I have yet to see any monitor which is even capable of those resolutions at the sizes you have indicated. I don't know of any 15" monitors which can do more than 1024x768, and I can't even get my 17" higher than that (never mind that it says quite plainly on the box that it should be possible). All of the 19" monitors I've found can't do more than 1280, and the 21-inchers can't do more than 1600.
  • Most likely your request came before the DRI stuff was released to XFree86 (at which point I usually deflected people since the stuff > they were looking for simply wasn't there, yet).

    This is correct. I posted it in February, while the DRI code was released in June.

    Normally everyone who sends email to XFree86@XFree86.Org and states "I would like to work on ABC" with "ABC" somewhat more informative than "XFree86" or "drivers" will get an application form within a few days.

    That's what I anticipated (and reading this now explicitly from your side doesn't exactly make me feel better :-)

    As to the generic issue here, yes, I think that XFree86 should open up its development a bit. And guess what, we will.

    Glad to read this. I want to stress here that I don't want you folks to lower your quality criteria, just be more transparent, please. The usual way to do this is having some public forum of discussion. At that time I applied I did not find one.
    I don't want to transmit the message that the XFree86 folks are snobs, but rather (as the title of this message suggests) that the Xfree86 could make it easier for people to find out what is going and eventually participate.

    The release of the 3.9.x snapshots is a first step in that direction, more will follow.

    I would rather be able to follow the developers discussion (and join it on occasion). Without knowing where the project is heading I find it rather useless (for me) to have snapshots, with the exception of analyzing it to find out what is going on. :)

    To illustrate that I am not just ranting let me end this post with a link that might be interesting for the BSD crowd: riva-glx []

  • I don't know if there would really be much of an increase in release speeds. In fact, it would probably end up rather like Mozilla. (imho)

    At the last count that I saw, xfree had ~1.5 million lines of code. That's a whole lot of code to understand before you can start to do any serious development.

    it would a rather hefty time commitment, and xfree strikes me as the kind of project where you'd want to know how the whole thing works before you started mucking around too much.

    who knows about 4.0, though. a lot of things have been rewritten from scratch, if i read correctly. could be a whole different ballpark. but i doubt it.
  • It does. Mine's been working fine for some time now.
  • Definitely ... I'm using Verdana as my browser font (running the xfstt font server) and it definitely looks very nice on a 15" 1024x768 screen, even with Netscape's questionable rendering technbiques. (Can't wait to see how it'll look in Nav 5 though!)

    Plus, there's just something deliciously ironic about taking something free from Microsoft (their web fonts collection) and them not getting any platform lock-in in return.

  • Damn - i was there last night checking for this... But it was not there *I suspect squid* ....

    Today, BOOM, It's there! But flooded...

    Marques Johansson
  • is that $300 us? I just bought a VooDoo3/16mb pci 2000 for $99 at compusa, imagine if i bought it online - $89ish ... Besides that - All the Voodoo cards (AFAIK) are under $100 these days (minus the new Voodoos 3000,4000) ....

    Furthermore - Riva 128/TNT cards with GL support are also $99 at compusa...

    In X i can get 32bit color at all modes my monitor supports, and the framerates are still awesome.

    What home user really needs much beyond that?? (attn smartasses: don't even bother answering that question)

    Marques Johansson
  • And the easy choice is...3Dfx?

    I have a Creative TNT card, and q3test runs in the seconds-per-frame range. Maybe I've just installed the libraries wrong, but as of now the TNT drivers perform very poorly.

    Stripping away religious issues, what 3D card PERFORMS best in the games available _right_now_ for Linux?

  • Well, you do have choices even now for accelerated
    3D under linux

    1) 3dfx

    Binary-only drivers using glide. It's worked under linux for ages.

    2) Matrox G200/G400

    Vendor supplied near, but not complete documentation.

    3) TNT/TNT2

    Vendor supplied/supported 3D acceleration with full source.

    Not a hard choice for me =) (well, ok, G400 is still a possibility, it seems to be slightly cheaper and some of the features are really nice. Not that I have space for two monitors, so the dual head support is useless)

  • Heh. how many hours 'till its in Incoming? :)

  • Sony 200SX/ES (17") will do 1280x1024 (60Hz).

    ViewSonic 815 will do 1880x1440. Good luck getting an off-the-shelf video card that will drive that... (most seem to max out at 1600x1200).

    to be honest, on my system here (ATI RageII Pro driving a GW2K "VX900" 19" monitor at 1600x1200), I see ZERO difference between turning on "sharpen screen fonts" and having it off, in Windows NT.

    Anti-aliasing does less and less at higher resolutions.

  • by drig ( 5119 )
    The wire protocol is the same, so X apps run on another system will still display on XFree 4.0. 4.0 adds GLX, so OpenGL apps will run nicer even then 3.x. This'll definitely help SGI->Linux, I'm not sure about Solaris. Anyway, yes, your Solaris apps will still display on Linux nicely.
  • Check out

    It'll tell you a quick update on the difference.
  • I don't have a Permedia2, but I ran a grep on the sources and there appears to be a lot of support. YMMV.
  • I've never gotten him to compile on my box. It just doesn't work.
  • "Broken" fonts? You mean like PostScript and TrueType fonts?

    No, I think he's talking about fonts like the nexus font (among others). If you've ever used xfontsel, those fonts don't show up because they don't have a true font string embedded in them, I ran into this problem trying to use the gtk fontselect box...fonts under X just suck.

  • Just as a matter of interest, I seem to recall that if you use LibGGI with the DGA target and have fbcon, you can change the screen resolution/mode on-the-fly, which you cannot easily do with plain DGA.

    And, of course, you get the usual LibGGI advantage of extreme display independence. No more releasing separate console and X versions.
  • see the Berlin project. Not that that kind of 3d flashiness is in and of itself a design goal, but the current design allows for it.
  • You cannot change resolution/mode at all with DGA. It's not within the DGA API. You can use the vidmode extension, which is XFree86 specific, to change resolution but not depth (as of 3.3.3).

    This is true, and it is in fact exactly my point ("cannot be done with plain DGA") -- DGA by itself cannot change the resolution or depth. Even with vidmode, however, you still can't change the depth on-the-fly. At least not yet.

    Programs like xmame and snes9x exploit this, with DGA used for speed and vidmode to change res.

    I was aware of this.

    LibGGI's DGA target can use fbcon, if present, to change both the resolution and depth, while getting access to the framebuffer via DGA. That's just kind of neat, and if vidmode or some other extension evolves to do the same, that's fine.

    But you're a Berlin developer, no wonder you know nothing about X or XFree86 extensions.

    Why the hostility? There's no need to resort to these sort of attacks just to correct a simple omission on my part. Honestly, what reason do you have here to feel either angry OR threatened? If GGI was intended to replace X in most situations, LibGGI wouldn't display on X, and they wouldn't have XGGI (an XFree86-derived server that uses LibGGI), would they?

    Chill out, man. Us Berlin folks are just playing around with a heavily CORBA-ized Fresco-esque GUI. It's no threat to your hegemony unless Berlin turns out to be better for users and developers than X is. If it turns out to be a bad idea, you're safe. Let the ideas speak for themselves -- on their own merits and demerits -- and leave the ad hominem attacks out of this.

    On the other hand, how much longer are you going to keep grafting extensions onto X to work around deficiencies in the initial design? It never hurts to experiment with a clean design once and a while under such conditions, and I really don't think the Berlin folks deserve to be vilified for doing so.

  • Netscape is broken. Try this:

    1) Install TrueType fonts. Use the xfs server from Redhat 6.0 or xfstt.

    2) Install the Arial font from Windows according to instructions with the TT font renderer.

    3) In Netscape's preferences Appearance/Fonts, use Arial as the default font, click on the Allow Scaling button.

    4) In the same place, type the number 16 (16 point font) in the textbox next to the "Allow Scaling" button.

    5) Save preferences

    At this point your fonts should be MUCH better on all pages, and comparable to the Windows handling of fonts. This works for my home 15' monitor at 1024x768 and my 21' at work at 1024x1280. This is an OLD problem with Netscape, one that Mozilla doesn't have (thank god).

    Oh, one problem with this setup. Netscape doesn't save the point size of scalable fonts, but rather defaults to 12. You have to enter the '16' into the text box every time you start Netscape...


  • by John Fulmer ( 5840 ) on Tuesday July 20, 1999 @04:46AM (#1794782)
    There are actually two issues here...

    1) The Type1 font engine was donated by IBM many many moons ago. It works, and sometimes well, if you have a good font, but has never been optimized.

    2) Many of the standard XFree fonts were donated and they weren't really high quality.

    Personally, I find that TrueType fonts look very nice in X (with RedHat 6.0's xfs (freetype) or xfstt). I've compared them with the local NT box with a 'real' TT font renderer, and they are at least as good.

    One thing to remmeber is that Netscape is broken as regards scaleable fonts. That's why some pages look really odd with tiny fonts. However, if you do a trick (deals with typing in the font size in preferences), my Netscape fonts look as good as NT's on all pages.

  • Does anyone know if work is being done to support the NeoMagic MagicGraph 256ZX (NM2260) chipset? A friend of mine recently bought a Dell Inspiron 3500 that has this chipset, and if no X servers support it, it's completely useless to her.
  • > Has anyone ever actually been turned down?

    Yep. Well, maybe not turned down, but certainly not replied to. A year or two ago, I sent a message to their main address volunteering to help with the Matrox driver, and never heard a peep back.

  • Hi,

    X 3.3.5 should be released in a week or two.

    Not everything made it in this release...
  • So, my multi-headed system has languished, ever since I went to the 2.2 kernel, which broke Metro-X for reasons unbeknownst. Has anybody yet tried using the new XFree's multihead support? (Xfree 3 got horribly confused by it...couldn't even display on just one card.) My system has two Matrox Millenium II's, which, given what XFree's website says, should work. I hope. And anybody played with Xinerama yet? Any WM's able to handle it?
  • Unfortunately, I'm nowhere near a good enough C progammer (or any other kind of programmer, for that matter) to be able to make code contributions to XFree86. Who's working on the development of that part of the code? I'd be happy to act as a guinea pig, if somebody can kick me in the right direction.
  • The debian maintainer of x, Branden Robinson compiles on my box. I already have both downloaded, and have started applying the debian patches to 3.3.4. 3.9.15 is all his, tho.

    This is NOT an official word from Branden. There is no timeline as to when this will be available from debian.
  • I said no timeline. :|

    Fortunately, the build system(and debian/patches/*) makes this new version easier to work with.
  • well, while you work, I will work. :)
  • Umm, that is why they are putting it into unstable (potato) and not stable (slink).

    Now, if they had a buggy/unstable version of X in potato when potato became stable, that would be a different story.

  • And not only that, you forgot the large number of people with Matrox G200's. Those are accelerated, too -- they're actually quite fast under Linux (and cheap! :). So yeah -- only a few people with accelerated 3D? Maybe a year or so ago, but not now.
  • The fact that no one is coming forward to defend this wild theory is instructive, wouldn't you say?
  • If you antialias 10 point fonts or lower, the of course they'll be blurry then, the only way to get around jaggy fonts at small sizes is to not use them. Besides, you haven't heard the same thing, you brought up the subject in the first place.

    The, shall we say "novel", theory about how antialiasing works, by playing with your eye focus, simply isn't born out by any facts. I eagerly await revelation to the contrary.
  • Has any of the XFree86 mirrors gotten these files yet? I checked the two most local to me ( and and neither had the development snapshot. The main ftp server is /.ing bad:

    /bin/ls -CF: Too many open files in system.
  • You might be able to sync up to that resolution, but will you be able to resolve all the pixels ? What is the dot pitch on the P817 ? 0.22mm horizontal, I believe. You have about 0.75x20in viewable horizontal distance, which leaves you with approx

    25.4*20*0.75=1731 dots per row

    In other words, you might be able to sync up to the higher resolution, but you aren't really getting better than 1800x1440 pixels.

  • Did you make sure that each of the 1024x1280 pixels were really resolved ? it's one thing for the gun to be able to sync up to the resolution , and another thing for your shadow mask to have a fine enough pitch to actually display it.
  • Do you have actual evidence that antialiasing causes eyestrain? Have any studies been done on this?

  • At last, we have support for CID fonts. Now if the Japanese font makers would only get off their collective asses (*cough*Morisawa*cough*) and release fonts under a fucking REASONABLE license, then I'd be happy...

  • Yes, the time is almost upon us. I'd like to see nVidia pick up the ball and run with it now. I have to make a decision sometime soon for a new vid card, and I would love to have a couple choices. Voodoo3, G400, or TNT2/Ultra. Hmmm, choices choices...

    I'd like to see what a DRI driver can do for Q3Test, as this is what was holding back cards like the TNT2 and G400 from performing well.

    With LAN tournaments coming up, I would love to be able to compete with Q3Test/Q2 native on Linux. That would certainly raise eyebrows for the Windows folk :)
  • Personally, I find that TrueType fonts look very nice in X (with RedHat 6.0's xfs (freetype) or xfstt). I've compared them with the local NT box with a 'real' TT font renderer, and they are at least as good.

    Indeed they do look quite nice if you use a good font. However, you're forgotting about antialiasing. That's part of what makes Windows tt rendering so good. Freetype already supports this, so it shouldn't be too hard to hack something in, though the proper way is to use alpha-blended drawables, I expect. Perhaps with the hardware accelerated Imlib extension in XFree 5.0?

    2) Many of the standard XFree fonts were donated and they weren't really high quality.

    This is definitely true. Microsoft, oddly enough, has commissioned some very good truetype screen fonts. Unfortunately, there're not redistributable, but you can download them here [] for free; unzip works on the Windows 'self-extracting' .exe files.

    One thing to remeber is that Netscape is broken as regards scaleable fonts. That's why some pages look really odd with tiny fonts. However, if you do a trick (deals with typing in the font size in preferences), my Netscape fonts look as good as NT's on all pages.

    I think this is more a function of the lack of resolution-independence in the OS. Most operating systems make different (wrong) assumptions about the physical resolution of your monitor, so a webpage that looks reasonable under one OS won't on another. See this tidbits article [] for details.
  • what version of unzip are you using?

    Ah, I spoke too soon. unzip works on the Win 3.1 versions (I think it just skips the executable part and looks for the zip header). The Win32 versions don't work--you're right--and some of them aren't available in the old format.

    I tried wine on them, but the (older--981018) version I had handy didn't work. Other options: borrow a Mac or Windows machine, or try decoding the mac format one. There are some tools for doing that under linux, but I didn't have any luck.
  • I've got a TNT, and the current GLX driver are pretty darned slow, even when compared to the performance I've heard from the G200 driver. The driver doesn't seem to be at all optimized, and since there are no specs (afaik), nobody's working on it. Everytime I 'cvs update', I see all the G200 stuff updated, and usually none of the TNT stuff. NVidia either needs to finish their driver or release specs!
  • you must have an incredibly old monitor...I have a low end 19" (an optiquest v95), and it will happily do 1800x1440 (great for having a whole bunch of files displayed when you're coding).

    Like some other people have mentioned, you probably want to shell out some cash for your monitor. It's one of the parts of your computer that will actually last more than a few years, so go ahead and splurge on it.

    "The value of a man resides in what he gives,
    and not in what he is capable of receiving."

  • by dirty ( 13560 )
    anti-aliasing is a really disgusting hack. Since it slightly blurs the fonts your eyes perceive them as being out of focus, so they try to refocus, this doesn't work, so they refocus yet again, and again, and again, and again. This causes a great deal of strain for your eyes. Running at a higher resolution with larger font sizes looks better and is much nicer to your eyes.
  • Actually I think you can hack around the no-save problem. I did this once with netscape on my computer (no idea if it was for the exact same problem though). Do the above steps, then exit netscape. Go into ~/.netscape and edit whatever the preferences file is (i think it's prefs.js or something like that). Find the font line and everywhere you see a -0- for the font size replace it with -16-. Note: these directions might not be 100%, i have only done this once on one computer and it was a little while ago. If someone has more (better) knowledge please post it. It's a really ugly hack, hopefully mozilla will be useable soon (for every day browsing atleast) and we can finally get some decent looking fonts.
  • sounds like the font server got borked. first try /etc/rc.d/init.d/xfs stop ; /etc/rc.d/init.d/xfs start. If that fails, try rebooting (I know it's a very windows-ish solution, but it works sometimes).
  • by dirty ( 13560 )
    I have heard the same thing, and from personal experience, I would say it's true. My monitor at work (windows, dell 17inch, 1024x768x75hz) causes my eyes to hurt more than my monitor at home (linux, acer 17inch, 1280x1024x60hz). My home monitor runs at a lower refresh rate and a higher resolution which should cause more eyestrain, not less. The only thing I can think of is anti-aliased fonts.
  • Actually any XF86 source can be GPL'd. Even the GNU project uses X11 and XFree86--if it's good enough for RMS, it's good enough for you, too. =)

    X11 doesn't have the advertising clause of the BSD licence, so basically X11 code can ``become'' any other licence. It's truly all things to all people.


  • I got multihead working on my experimental XF86 3.9 setup a number of months ago. The first thing you should know is that few devices other than the Matrox work (the S3 and Trident do not work, at least not now).

    The second thing to know is that 3.9 is highly unstable, especially with multihead. Feel free to fix bugs and submit patches if you do hack around with the 3.9.15 release, though. =)

    (I actually played with multihead on a Microchannel/XGA-2 system, but that's another story).


  • looking over there web site it looks pretty nice. the next X should be really good.. I may need to get new hardware so I can really take advantage of the new features...
  • "Precision Insight has been provided with funding and support from Red Hat and SGI to integrate the GLX extension for 3D rendering in an X11 window. The 3D core rendering component is the Mesa library. SGI has released the sources to the extension framework under an open license, which essentially provides the glue system. Precision Insight has integrated these components into this XFree86 X Server and added a Direct Rendering Infrastructure. Direct Rendering provides a highly optimized path for sending 3D data directly to the graphics hardware. This release demonstrates a sample implementation of direct rendering by providing a single path of 3D hardware accelerated rendering for the GMX2000 graphics card. Future releases will support much broader implementations of hardware accelerated direct rendering on a wide range of 3D capable graphics devices."

    If they support OpenGL for my card (i740) half as well as it is supported under Win98, then I don't have to boot into Windows to play Quake anymore!


    I heard the dude (too buzzed to remeber his name) from XFree86 talk about 4.0 at the LinuxWorldExpo and I have been very excited ever since. True type font support (tho I already got that set up, and they will be using the same thing (xfsft and FreeType)) plus more OpenGL and Multi-head support.

    Then if KDE comes out with a great 2.0 (Especially with high color icons) w/ a better looking widget set, then the future indeed looks bright for Linux on the desktop.
    (I like GNOME, it looks way better than KDE, but it just doesn't seem to work as well at this point (except for it's file manager, which is way faster than kfm))

    alright.. i'll stop with my drunken ramblings.. but DAMMIT, I'M EXCITED!

  • "a fix to get XFree86-3.3.x to compile under glibc-2.1.2 "

    Ack.. I hope by previous releases you dont mean 2.0.x . just finished downloading all the source, and about to try compiling with 2.1.1 on RedHat 6.0
  • Quit buying cheap, crappy monitors. My Dell 17" at work is (right now) running at 1280x1024... for a good 19" try the Hitachi Superscan 750, it can easily do 1600x1200
  • Does anyone know if the Glint (3DLabs) driver supports Permedia2 based cards, or just the higher end chips?

    "Trouble is, just because it's obvious doesn't mean it's true"

  • The question was in reference to the drivers for 3.9.15, I have been running the 3.3.x glint drivers since they were released on SuSE's site. The release notes said the card support for 3.9.15 was limited and though it says glint and 3DLabs I would prefer not to DL and compile it if it inly works on the actual glint chips and not on the permedia2.

    "Trouble is, just because it's obvious doesn't mean it's true"

  • I have a 2 year old ViewSonic (P775) which displays up to 1600x1200. I imagine there are better 17" monitors out there now.

    "Trouble is, just because it's obvious doesn't mean it's true"

  • Close, but not quite, I think. The NVidia people picked up the code written by the "crowd" for the G200 and ported it to the TNT, not the other way around.

    And it seems like it's more a GLX crowd than a G200 crowd now, despite the mailing list name. Stuff like GART and DMA support isn't Matrox specific.
  • And will unsupported in August according to the Register.

    Why can't intel stick with something that it is good at.. like well ummm like.. so okay.. x86 processors..
  • Yeah, right... only 3 or 4 Linux users in the whole world have either a Voodoo {1,Rush,2,Banshee,3} or nVida Riva TNT(2) or Amiga Warp3D!!
    I for sure know more than 4 Linux users who have supported hardware, and besides, if you check Linux 3D [], you'll se that support for other cards is under way as well!!
    Remove the capital letters from the e-mail-address
  • No offence, but that's a really dim comment. I've got a 17" that can't display 1024x768 without getting a bit blurry at the edges, and I think many other people are stuck with even worse monitors. And most of us cannot afford to shell out for a posh Ilyama or something similar...

    Anti-aliasing may not be the 'right' solution in the most anal sense but it makes life better for lots of people. e.g. back in 1991 Acorn introduced a fully scalable, anti-aliased font system on their machines and all I had was a 50Hz telly but it still looked pretty nice. Today Netscape does some 'orrible things to my fonts and I can't afford to buy a posher monitor, so I for one would really really like some more apps to use it.

    Hey ho.
  • Where is the changelog and other assorted documentation thats supposed to be in the release?

    "Due to a few important changes that came after 3.3.4 was finalized , a 3.3.5 release (which will include binaries) will be made in the next couple of weeks."

    Ok, so they're releasing this version, which is known to be somewhat incomplete under a full blown version name. Why? Shouldn't they just call it a pre-release or a beta? It's only a couple weeks until the Real Deal comes out. Why say "Well, We have this new version of Xfree, but its got problems and we'll issue the fixes under the next version." Doesn't this sound like some idiot software company out of Redmond who releases service packs to fix service packs?

    Ok, that was a little too much of a parallel. But do you see my point. If an Xfree86 release addmittively sucks, don't give it the entitlement of a full version number. Just call it 3.3.x-pre or something and let the world know: "For bleeding edge users only." At least they were half-thinking like that . . . they left out the documentation so idiots like me can't see if I need it for my Banshee. . .

  • ok, disclaimer: this is based on my experience only

    BFWIW, I've found that the Voodoo 2 is the "best" solution for the games I want to play under Linux (Quake, Quake 2, Quake 3).

    For the record, I bought a Diamond Viper 770 (TNT2 Ultra) and it plays Q2 ok, just Q3 is too slow (yet). The 3dfx plays all these games very well.

    But, I got the TNT2 card because the Voodoo is running out of gas and XF86 4 with DRI is what I expect to use with the TNT2 to replace it.

    Granted, gaming is not my primary activity in Linux, but I'm getting really sick of dual-booting back to Win98 to play games.
  • by roady ( 30728 )
    Hi, I Know most of you Linux guy are not concerned, but what is the compatibility with X11 ? I run a Solaris box, and there are already so many linux software hardly protable on other Unices.
  • by cbarry ( 70212 ) on Monday July 19, 1999 @09:59PM (#1794862)

    It's kinda sad how short the XFree team is on developers when more or less 99.999% of Linux users use X and 100% of distributions package it. It could really use some more commercial support from RedHat and SUSE, though they have helped a little bit in the past (RHat donated NeoMagic code once...).

    For information on becoming an XFree86 developer, please visit the XFree86 developer [] page.

    Also, you non-programmers that use X can do your part by knowing that RedHat [] and other commercial Linux vendors have ears for their customers and showing concern for the frequency of XFree86 release cycles is a good way to let them know that support for X development is very important to the success of Linux.

  • by DirkHohndel ( 70583 ) on Tuesday July 20, 1999 @05:26AM (#1794864)
    The support load is one of the key problems behind the current
    somewhat closed approach. There are other issues (the devel
    sources often contain drivers that were written under NDA
    and for which we haven't received permission to release,
    yet. Those obviously can't be publicly available).

    The 3.9.15 release is somewhat a test case. If we receive
    tons of support email from people trying to use it and
    asking for help, then we might revert back to the closed
    cycle that we did before. I certainly hope that none
    of the distributions will attempt to include 3.9.15.
    It is definitely not ready for that. SuSE will NOT include
    it on their next distribution, btw...

    Don't get me wrong. Bug reports (and of course, patches)
    are extremely welcome. I saw another comment that we didn't
    respond to bug reports. My answer to that is simple.
    We get so many reports, and there are only so few people
    to respond. Usually none of them go unseen and as long
    as they contain a fix or the fix is obvious, things
    usually get fixed as well.

    Of course, the 800 or so bug reports "my Trio3D card
    doesn't work" didn't really help to fix the problem...

  • by DirkHohndel ( 70583 ) on Tuesday July 20, 1999 @05:17AM (#1794865)
    Sorry if things went wrong that time. I get tons
    of emails a day, so I must admit that I don't
    remember the incident that you are commenting on.

    There is no competition whatsoever between the
    work that PI does and the work that SuSE does
    for 3D. I am sure that Frank LaMonica from PI
    will be happy to comment on his take on the issue.

    Most likely your request came before the
    DRI stuff was released to XFree86 (at which point
    I usually deflected people since the stuff they
    were looking for simply wasn't there, yet).

    Normally everyone who sends email to XFree86@XFree86.Org
    and states "I would like to work on ABC" with "ABC"
    somewhat more informative than "XFree86" or "drivers"
    will get an application form within a few days.
    And those people are always added to the devel

    As to the generic issue here, yes, I think that
    XFree86 should open up its development a bit.
    And guess what, we will. The release of the
    3.9.x snapshots is a first step in that direction,
    more will follow.

    Dirk Hohndel

Things are not as simple as they seems at first. - Edward Thorp