Re: [XFree86] don't understand how to get started w/ translucent windows
On Thu, 1 Jan 2004, dave giffin wrote: I looked at http://xserver.freedesktop.org/Software/TranslucentWindows and http://xserver.freedesktop.org/Software/Home, where I saw the 'xserver' project. How does this project relate to xfree86? The kdrive/tinyx X server which was previously hosted by xfree86.org and developed within the XFree86 framework has been moved to freedesktop.org where development continues. There is still a kdrive/tinyx X server in XFree86 but (for now) without the more interesting features (such as translucency and shadows). Much of the code is actually shared between kdrive/tinyx/xserver and the traditional XFree86 server -- much of it is actually the sample code from X.org which is also used in many other X servers. If you want to get up to date with xserver at freedesktop.org, have a look at their website, mailing lists, IRC channel. It is really a different project these days. The rest of your questions are probably somewhat off-topic but I'll answer them anyway (but if you want to know more, use fd.o mailing lists/IRC). I don't see anything that says if you want translucent windows, do this Really? It should be there on the fd.o website ;) If I download and compile the 'xserver', does that automatically come with translucent window support? Well, yeah. Hence the demo pictures and the link to the page that describes how to download and build it (on the fd.o site) If it does come with translucent window support, how do I control which windows are translucent and how translucent they are? The translucency is (for now) handled in a compositing manager (xcompmgr) which in principle can decide the translucency itself -- but in practice just composes some windows and adds shadows. Some windows are created with an alpha channel, some are not. The application has a say in the translucency decision if it creates a window with an alpha channel. There are some example programs somewhere in fd.o cvs if you really care. Is there a command line program? Could one be created if there isn't one already? Yes. Are you a developer? Do you like to hack? Then by all means have a go at xserver and xcompmgr. If you are just a user, however, I think it will be better for all involved if you remain patient for a while yet. xserver is very much experimental at this stage and there is a limit to how much hand holding people will do ;) xserver contains a couple of new X extensions, none of them really finalized yet (or used in applications). I expect that if/when they are complete and shown to be useful then XFree86 (and other X servers) will implement them as well. The driver backends in xserver are still *very* rough (or not even implemented for many video cards). If you want fast screen refresh, good driver support and a screen refresh rate that doesn't make your eyes bleed [1], stick with XFree86 :) 1) xserver is currently limited to 60Hz unless you do magical incantations, have the right hardware, and are generally lucky. That is fine for LCDs and awful for relatively modern CRTs with fast phosphors. -Peter But moustaches are a mystery, to evolutionary biologists and to practically everyone else. -- Economist.com ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
RE: [XFree86] Where i send monitor information ??
On Thu, 11 Dec 2003, Marcel Mourguiart wrote: I dont´t have problem with the monitor, i just would like to provide the monitor data to the xfree data base, so they can include it like a supported monitor. try [EMAIL PROTECTED], I think he is collecting them. -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] TinyX/Touchscreen
On Mon, 1 Dec 2003, Sarfraz Nawaz wrote: Is it possible to use elographics touchscreen with TinyX/Kdrive. If yes how to configure it. I know its supported in stock version of XFree86. You'll probably have to do some coding... Send a message to [EMAIL PROTECTED] or drop in at #freedesktop on irc.freenode.net. There are probably others who want the same thing so you can work together on it :) -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Fullscreen DGA really slow
On Sun, 2 Nov 2003, Pieter Hulshoff wrote: I just tried running x with the configuration file from the Morphix cd, but the result is the same: tantalizingly slow. Would it be a SuSE only issue? Perhaps a problem with the driver, or any libraries the driver may link to? Or that /proc/mtrr is different when you run Morphix. Or some AGP stuff is setup differently. -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Fullscreen DGA really slow
On Sun, 2 Nov 2003, Pieter Hulshoff wrote: I've tried editing the /proc/mtrr, but it won't allow me to do so (writing is not allowed, but writing rights of file and directory are set). I was hoping http://lists.debian.org/debian-user/2000/debian-user-200012/msg02130.html Perhaps all the registers were in use already? (or you weren't root?) perhaps the answer could be found by using the Morphix mtrr settings. If that doesn't help, then I should really consider having a good look at my kernel, and try a vanilla kernel or something. Perhaps SuSE played around with the mtrr support in the kernel... Could you write what the morphix mtrr settings are, please? (If you already have done so, could you please do it again with using attachments? I have a procmail filter that discards everything with attachments since I got swamped by the SWEN worm) That way it would be possible to compare the mtrr settings... -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Fullscreen DGA really slow
On Fri, 31 Oct 2003, Pieter Hulshoff wrote: On my SuSE 8.2 system, default xfree86 drivers, which are? I get very slow performance using Fullscreen DGA (using xmame) on my Hercules AIW 9000 card. I've used a Screen updating using DGA is not accellerated. It is just like writing to (slow!) memory. morphix boot cd as well, which use which XFree86 version? and the performance is fine there. Windowed performance (same xmame, same game, same options) is fine as well. Updating Windows is usually accellerated. (Unaccellerated) Write performance to the screen is extremely dependent on telling your CPU that this memory area is to be treated specially: don't blow the CPU cache on trying to accellerate it but combine as many writes into one as possible (to enable burst writing and to not perform useless reads from the screen buffer). This is controlled by the MTRR (Memory Type Range Registers). You can view them and write them as /proc/mtrr under Linux. To see where the screen buffer is located, poke around with lspci -vvv or /proc/pci (or somesuch). -Peter PS: X should enable the right range automatically but sometimes it doesn't work. It probably even tells you that it tries to fiddle with the MTRRs and fails. Take a look at the logfiles. ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Fullscreen DGA really slow
On Sat, 1 Nov 2003, Pieter Hulshoff wrote: XFree86 Version 4.3.0 (Debian 4.3.0-0ds4 20030416150820 [EMAIL PROTECTED]) Release Date: 18 March 2003 X Protocol Version 11, Revision 0, Release 6.6 A pity that both were 4.3.0 :( That rules out the obvious advice of upgrading to a newer X server ;) (but the kernel might be different -- that sometimes matters for MTRR and AGP -- or the X configuration might be different) a driver or an XFree86 configuration issue. Which is why you should check /proc/mtrr and the server logs. -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Fullscreen DGA really slow
On Sat, 1 Nov 2003, Pieter Hulshoff wrote: Here's the /proc/mtrr: reg00: base=0x ( 0MB), size= 512MB: write-back, count=1 reg01: base=0xd000 (3328MB), size= 128MB: write-combining, count=2 reg02: base=0xd800 (3456MB), size= 64MB: write-combining, count=3 That's for the slow DGA setup, right? What about the fast DGA setup (Morphix?) ? If you can get it easily with a quick reboot, that is. -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Fullscreen DGA really slow
On Sat, 1 Nov 2003, Pieter Hulshoff wrote: I've included the config and logfiles in the previous post. Were they attached or included? If they were attached, I didn't see your post at all... SWEN, stupid Windows programs, (what should be) criminally naïve windows users and all that :( -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: Cygwin/XFree86 - Staying or leaving XFree86.org? [Was: Re: Cygwin/XFree86 Bugs?]
On Sun, 26 Oct 2003, David Dawes wrote: When I discussed this with you privately a while ago all I got were disrespectful and insulting responses. Now there is more of the same. Err... No. He was quite reasonable, in my opinion. -Peter ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: Cygwin/XFree86 - Staying or leaving XFree86.org? [Was: Re: Cygwin/XFree86 Bugs?]
On Sun, 26 Oct 2003, Thomas Dickey wrote: On Sun, 26 Oct 2003, Peter Firefly Lund wrote: On Sun, 26 Oct 2003, David Dawes wrote: When I discussed this with you privately a while ago all I got were disrespectful and insulting responses. Now there is more of the same. Err... No. He was quite reasonable, in my opinion. He thought he was, but when I get email from people phrased that way, I don't appreciate it. He really was. You weren't. His reply perhaps wasn't, either, but at least there was a reason for that which was discernible to dispassionate bystanders. -Peter ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: [XFree86] Combining characters
On Sat, 6 Sep 2003, Anuradha Ratnaweera wrote: Now, if I want to render the character 51 of X, in place of the composite character 4001+4010, how should I proceed? Is there a way to map unicode sequences to actual (physical) fonts. Prefarably in the form: I don't think there is a general X way of doing it. You can do it manually in your own programs... or you can take a look at Pango and perhaps modify one of the Indic shapers: http://www.pango.org/ This homepage is not updated anymore since Pango is no longer a blue-sky research project but a standard component of GTK+. http://www.gtk.org/ OpenType probably has a way of handling it - but that would only work for programs that use OpenType... That would be Pango again ;) http://www.microsoft.com/typography/tt/tt.htm http://www.adobe.com/type/opentype/main.html http://www.agfamonotype.com/software/opentype.asp http://www.freetype.org/opentype/ http://www.freetype.org/opentype/ -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] RE: Laptop touchpad and XFree86 4.3.0
On Mon, 4 Aug 2003, Peter Wainwright wrote: More info on XFree hanging on startup when using Synaptics touchpad on some laptops: [...] I now believe that XFree86 is not (solely) responsible for the problems, there is some interaction with the /dev/psaux handling code which is now merged into linux/drivers/char/pc_keyb.c. I put some debugging spew in the XFree86 mouse driver and watched it run; the problem seems to be with the timing of the reset sequence. This as far as I could tell is what happens: Sounds about right. There was a recent thread on this on the linux kernel mailing list (but it looks like they don't have a proper fix yet -- just a decent workaround). You can also google for active ps/2 multiplexing. The first hit is a 27-page well-written spec from Synaptics on how to connect up to four pointing devices to a PS/2 port without confusing legacy drivers too much in a such a way that newer drivers would still be able to distinguish between the different pointer devices if they wanted to. And also such that special features (such as mouse wheels) would work. Interesting stuff. You will also get a couple of older hits from the Linux kernel mailing list (but not the recent thread). -Peter There are only two real problems in computing: computers are too hard to use and too hard to program. -- Brian Kernighan ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: Building XFree 4.3.99.9
On Sat, 26 Jul 2003, Andrew Bevitt wrote: You can try to substitute cc -c by cc -E and look at the produced code. That only prints out a whole load of code and includes to the terminal. Yes :) It runs the preprocessor on the source and nothing else. From what I can see its no different to the code in the files and doesnt really give me any clues. Pipe it into less and search for I2CDevPtr - see if it is defined weirdly as a typedef. It might be a #define instead - you'll be able to see that inside the DDCRead_DDC2() function. You search inside less with the '/' command: Type /I2CDevPtrenter, for example. Use 'n' and 'p' to skip between the matches. -Peter ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: spoofing X-events input?
On Tue, 22 Jul 2003, Roland Pabel wrote: I'd like to know if anyone ever thought of a way to spoof mouse/keyboards events, so you could send special events like clicked() to any application running... No, why would anyone ever want that? ;) If no one knows of an implementation, could you tell me with which of the X11/extensions/ header files to start with? Look for xtrap (and xtest). Some pointers: xc/programs/xtrap/ xc/programs/Xserver/XTrap/ xc/doc/specs/Xext/record.ms (has nice overview of the area) xc/doc/hardcopy/Xext/xtest.PS.gz And then there is XSendEvent, of course... ;) -Peter ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: [XFree86] where are the docs?
On Mon, 21 Jul 2003 [EMAIL PROTECTED] wrote: where is it in the sources? xc/doc/hardcopy/* (gzipped postscript files) xc/doc/specs/* (source files) xc/doc/man/* (man pages) If you don't want to muck around with flags to man you can use pinfo or mc (midnight commander) to view the man files (and some source files, too). -Peter Writing is Nature's way of letting you know how sloppy your thinking is - Guindon ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Xnest
On Sun, 20 Jul 2003, shxu1 wrote: Can you tell me how to use Xnest in the redhat linux 8.0? I have tried many times. It seems Xnest doesn't work properly with redhat linux 8.0. How does it break? Got any error output from it, for example? Can you tell us how you invoke it? What you are trying to do? Etc... -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: Rant (was Re: ATI Drivers.)
On Fri, 18 Jul 2003, Sven Luther wrote: Why do these companies not open source their complete drivers? Because they have intellectual property in their drivers that As if their concurent where not capable of reverse engineering the ^ competition somehow English chose another Latin word than the other Germanic languages -- the weirdness and hodge-podge nature of English strikes again :/ (this is not to spite Sven - merely to rant about English and at the same time increase the chances that an English-only speaker will understand) -Peter ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: [XFree86] A Wiki for this project
On Fri, 18 Jul 2003, Egbert Eich wrote: http://xfree86.linuxwiki.org Brilliant :) -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
RE: LCD Panels Refresh Rates
On Thu, 17 Jul 2003, Chris Edgington wrote: When I run the RH9 display settings applet to change the resolution, it automatically adds HorizSync and VertRefresh values to my XF86Config file (or I should say I believe he's the one adding the values.) Does it work correctly if you add Option NoDDC to your config file? What it comes down to is this I know the current panel can do 1280x1024. For modeset, I'm not doing any timing settings, just telling the BIOS what mode to go into. So, is there a right way to force xfree to allow standard modes 1280x1024 and below? Yes - use HorizSync and VertRefresh to tell it that your monitor can handle them :) -Peter ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: [XFree86] X-Render extension documentation
On Thu, 17 Jul 2003, Damian Dixon wrote: Could anyone point me in the direction of any documentation for the X-Render API please. Keith Packard's papers on keithp.com. Are there any examples of using X-Render and X drawing API together? http://www.eax.com/render/ (some of the links on the page are dead but you should still be able to google it out of the net) the other stuff on eax.com is also quite interesting. Is there an FAQ for the extension? If so where can I find it? Don't think so. Is there a more appropriate maillist for asking questions about X-Render? Sorta. There used to be one on xfree86.org which was killed without prior notice and without telling those who used it (as far as I can tell). Probably an act of spite. Instead, the Xr mailing list on xwin.org has become the place to discuss Render, too. There is also an Xr mailing list on keithp.com but it only contains a handful of spam mails. -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] How to hide the mouse pointer
On Sun, 13 Jul 2003, Pieter Hulshoff wrote: I think I just found an alternative way to achieve my goal. As I'm running 1 program on top of X (an SDL front-end to MAME I wrote this weekend), I can just disable the cursor using the proper SDL commands. Some quick tests seem to indicate this is working as planned. :) An alternative would be to go back about a day or two in either this list or the devel list (can't remember which). -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] command history among xterms
On Sat, 12 Jul 2003, gabe f wrote: not sure if I should ask this here, but here goes: Not really. the command history among my xterms is common. How can I make each xterm No, the command history is shared by your /shells/. xterm has got /nothing/ to do with it. keep its own command history? You are probably using bash as your shell. You can get each copy of bash to use a different history file with the HISTORYFILE environment variable (man bash, then type /HISTORYFILEenter) -Peter The problem isn't really what will we use to replace C?, it's what will we use to replace Fortran and the other languages that were replaced by C?. -- Peter da Silva, comp.arch ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] command history among xterms
On Sat, 12 Jul 2003, Eric Christopherson wrote: I think it's HISTFILE. The manpage, on my system at least, doesn't contain HISTORYFILE. You are absolutely right. I managed to botch the name in the 10 seconds it took to exit less, type fg, and complete the email in pine :( -Peter We also know how cruel the truth often is, and we wonder whether delusion is not more consoling. -- Henri Poincaré ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: FWD: RE: [XFree86] read directly from the video card framebuffer
On Fri, 11 Jul 2003, james amen atayi wrote: the same job. Camtasia for example records on windows the screen. I can allready do it using xwd or other tools to make the screenshots. The problem is that with these tools the mousepointer disappears when Why on Earth didn't you say so at first instead of wasting everybody's time? It seemed you only knew how to make manual screenshots! The request for a screen shot operation that leaves the mouse pointer in the image has come up several times on the XFree86 mailing lists (please google for it). AFAIR the discussions can be summed up as: no, you can't currently do it with a vanilla XFree86, yes, some X servers have an extension to do, yes XFree86 could also conceivably get such an extension, no, nobody has written one yet, yes, there are in principle other solutions but they are hacks too -- and anyway, you have the source code, you can do whatever you like. Please google now. -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Mailing list behaviour and etiquette, in general
On Fri, 11 Jul 2003, Randy Kramer wrote: have responded to a few questions -- I made the (bad) assumption that responding to the list got the answer back to the questioner. No :( Saying that they should subscribe doesn't help, either. Most of them are not able to read/comply with instructions like send us you log file, here is the full path: xxx... Respond to the list and to the questioner (better word?) -- this is an open list and many questioners are not subscribed. (Or something like that -- the shorter, while still being clear to less experienced mail list users, the better.) My preference would be if people got an explanatory letter the first time they wrote to the list telling: 0) that s/h/it is not a customer (in a different wording, of course) 1) that it is a mailing list (and what a mailing list is, perhaps) 2) that there are archives, so the question may already have been answered (with a URL to such an archive) 3) that the answer come to the list but there is no guarantee that it will also be sent to you (i.e. the one sending the email to the list) 4) you can subscribe and unsubscribe easily. Most of these could/should have URLs leading to more comple/wordy explanations. I realize that such a letter would require some fancy scripting to get right (and I am not volunteering) :( Much of the problem can be avoid by other means, for example by having the X server print a version-specific URL instead of an email address when it crashes or is misconfigured. That page should have a FAQ for the most common problems and a link to the mailman interface for the mailing list. It should NOT list the email address directly -- people would just use that without thinking. Another thing is that the footer with the link to the mailman interface really ought to mention the word unsubscribe. -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Xpert, Any software just for 15$ - 40$
On Fri, 11 Jul 2003, David Dawes wrote: When Joe Newuser does this, he gets an automatic reply telling him that he may miss replies to his message if he doesn't subscribe to the list. Disregard (parts of) my last email then. I apologize. -Peter If two men agree on everything, you may be sure that one of them is doing the thinking. -- Lyndon Baines Johnson ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: FWD: RE: [XFree86] read directly from the video card framebuffer
On Fri, 11 Jul 2003, Alan Coopersmith wrote: Actually someone has written one and it's in the XFree86 CVS tree, if you Wonderful! :) look in the attic directories and CVS history files for the remains of Keith Packard's rejected X-Fixes extension. It was merely puzzled for some months why it was rejected. Now I'm bewildered. Thank you, Alan :) -Peter I'm going to explain this as well as I can, given the limits of my space and your attention span. ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Hide mouse pointer under XF86-4.3.0
On Thu, 10 Jul 2003, Michael Heydon wrote: That still has the problem that it only stays as that image while over the root window (iirc). is there any way to force the cursor to stay the same when it is over other windows? Have a program walk the tree of windows regularly and force the cursor for all the windows to be transparent. It's a hack but it sorta works. Once in a while a mouse cursor will be visible briefly -- and some programs may become confused. Perhaps you can live with that? Otherwise you'll have to hack the X server -- or write an X proxy that translates the requests that set the mouse cursor shape :) -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
RE: [XFree86] Hide mouse pointer under XF86-4.3.0
On Thu, 10 Jul 2003, Oli Comber wrote: What about having a very small, 'always on top', background-colour, transparent cursor window in the corner, and (periodically?) set the mouse coords to it..? Even better, I think. -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] read directly from the video card framebuffer
On Thu, 10 Jul 2003, Eric Christopherson wrote: how to capture the video card output. In other way, how can I record the screen as a movie file without making screen shots? thank for your help I'm not sure of anything specifically for making movies/animations out of them, but there are several programs that can capture X protocol events, such as xmon, xscope, and xlab. Also, you could run a VNC server and use rfbproxy to capture its output. Maybe he could use a separate video capture card? Otherwise I'd say that the only way to go would be to take screenshots some number of times per second for as long as necessary but to get a program to do it, of course. -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Hiding mouse cursor
On Sun, 6 Jul 2003, Edoardo Serra wrote: I was thinking, there's a simple way to change the initial position of the cursor ? so i can put it in the lower right corrner at server startup XWarpPointer relative to the root window before any other windows open? -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: twm makes uninitialized memory access after malloc()
On Fri, 4 Jul 2003, Egbert Eich wrote: We have a report in Bugzilla (#464), concerning twm. This test can only be made on NetBSD: Couldn't it be tracked down with valgrind? -Peter ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: [XFree86] Feature Request: Better X11 Compression for client/server
On Sat, 28 Jun 2003 [EMAIL PROTECTED] wrote: Consider SSH. I doubt that any compression (including that which is part of SSH) will make X good over a dial modem, but it does work. I http://www.nomachine.com/index.php -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] BadLength (poly request too large or internal Xliblength error)
On Thu, 26 Jun 2003, Erik Enge wrote: On Fri, Jun 27, 2003 at 12:52:33AM +0200, Peter Firefly Lund wrote: Hmm.. it's getting late in Scandinavia. Can we continue tomorrow? Absolutely! You've been very helpful. I think I'll have to say Monday instead :( The sun just came back to Denmark and I get slow when the weather is hot :/ A diff would be fine; can you include patch command to apply it also? Of course I would :) -Peter Being really good at C++ is like being really good at using rocks to sharpen sticks. -- Thant Tessman ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] BadLength (poly request too large or internal Xliblength error)
On Thu, 26 Jun 2003, Erik Enge wrote: Peter \Firefly\ Lund [EMAIL PROTECTED] writes: Can you send the xdpyinfo output for the boxes? They support the same maximum request size - no surprises there, then. Box 1 supports 24-bit visuals but Box 2 support both 24-bit and 32-bit visuals. I don't know if that might confuse some of your software? (not that I /really/ believe that) My guess is that there is some bug in Render -- or in the software that calls render, namely the combination of the Windows font handling code in Wine and xlib. And also that this is probably provoked by the fonts you have installed. Another set of fonts might work fine. Are you able to compile programs yourself? You might want to get the program xscope from Keith Packard's CVS and run it on both box 1 and box 2 to get a trace of the request streams (warning: they are going to be big). If one of the traces really does contain an enormous request we can rule out the Render code in the X server (and the rest of the X server). The contents of such a request should help in finding out whether to blame xlib, the client-side render code, or Wine. Links: http://keithp.com/cvs.html web interface to the CVS archive + instructions on how to download - ignore what he writes about cvsup and use something like: $ cvs -d :pserver:[EMAIL PROTECTED]:/local/src/CVS login (Logging in to [EMAIL PROTECTED]) CVS password: [hit return here] $ cvs -d :pserver:[EMAIL PROTECTED]:/local/src/CVS co xscope http://keithp.com/~keithp/talks/usenix2003/html/net.html article from Usenix 2003 by Keith Packard and Jim Gettys where they use packet sniffing to gather traces which they then use for performance analyses. There, they use the tool xcap, which is also in Keith's CVS. (as is the actual paper itself, as the module Usenix2003). http://www.doc.ic.ac.uk/~mac/manuals/solaris-manual-pages/solaris/usr/openwin/man/man6/xscope.6.html A man page for the xscope version shipped with a very old version of Solaris (it was still called SunOS, I believe). This is just so you quickly can check whether it sounds useful and doable for you. -Peter Being really good at C++ is like being really good at using rocks to sharpen sticks. -- Thant Tessman ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] dumb proxy
On Thu, 26 Jun 2003, Vladimir wrote: it seems that ssh will not do X forwards on a machine that does not have X installed any way around that? Use -L and -R. Set the DISPLAY manually. Get the MIT cookie on the already running X server and install it manually into ~/.Xauthority on the machine you want the X clients to run on. Summary: Possible but not pleasant. -Peter Being really good at C++ is like being really good at using rocks to sharpen sticks. -- Thant Tessman ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] BadLength (poly request too large or internal Xliblength error)
On Thu, 26 Jun 2003, Erik Enge wrote: Yes, although I can't find any instructions in the xscope CVS module on how to compile it. Are there any instructions for this located somewhere? This seemed to work fine for me with a freshly checked out version of xscope: [EMAIL PROTECTED]:~/keithpcvs/xscope$ xmkmf [EMAIL PROTECTED]:~/keithpcvs/xscope$ make [EMAIL PROTECTED]:~/keithpcvs/xscope$ ./xscope options xmkmf (X Make Makefile) is the general way of generating a Makefile in X. X and related projects use a horrible build system with Imakefiles which are filtered through the C preprocessor by xmkmf in order to generate the real Makefile. Some of the stuff you see in the Imakefile, for example ComplexProgramTarget(xscope), is actually a C preprocessor macro that gets expanded into God knows how much in the real Makefile. See also: http://xwin.org:9673/xwin/BuildSystem If you google a bit, you can also find the original article on Imakefiles and why they were a good idea at the time (they really were!). -Peter Being really good at C++ is like being really good at using rocks to sharpen sticks. -- Thant Tessman ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] BadLength (poly request too large or internal Xliblength error)
On Thu, 26 Jun 2003, Erik Enge wrote: Do they tell us anything more? Don't know yet - I have downloaded them with wget and I'm going to take a look now. -Peter Wars begin when you will, but they do not end when you please. -Niccolo Machiavelli, History of Florence 1521-25 ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] BadLength (poly request too large or internal Xliblength error)
On Thu, 26 Jun 2003, Erik Enge wrote: Do they tell us anything more? Can I get you to rerun the test on box 2 with ./xscope -v2 ? That way, I can get the request number for each X request and see exactly which of the Render requests that fail. It doesn't look like the Render extension in the X server is where the bug is. The code looks very nice and simple (*). I think I have learnt to decipher most of an RenderAddGlyph request now -- except for *) I'm not sure how to calculate the length of a scanline in a glyph pixmap. In fact, I'm not even sure I know which Pixmap type is used for the images of the glyphs that are sent to the X server (it's the first time I try to decipher logs like these). If you don't want to play but just want TXWord to work on test4/box 2 like it does on test3/box 1, I don't blame you. Box 1 uses server-side fonts, which is the old style font technology in X and requires a running font server. Box 2 uses client-side fonts, which is newer, faster, leaner on the band width, and vastly fancier -- and doesn't even require a font server. Reconfigure box 2 to run with a font server instead of using client side fonts and it should all just work. -Peter PS: No, the logs haven't leaked any vital information - just TXWord, the two machine names and the filename of the document involved. ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] BadLength (poly request too large or internal Xliblength error)
On Thu, 26 Jun 2003, Erik Enge wrote: http://home.prium.net:/xscope-box-2 Thanks, I have snarfed it now. I'd really like to figure this one out, so whatever I can do to help you, I'll do it! Hmm.. it's getting late in Scandinavia. Can we continue tomorrow? It looks like all the RenderAddGlyph requests fail - I guess there is a misunderstanding between the client and the server about the pixmap format they should use. I might send you a changed version of xscope that dumps the actual pixmaps used - that way I can at least see what the client thinks it's doing. Would you rather have it as a tarball, a single file (that would be render.c) or a diff (that is, a patch) ? I think I would prefer the diff myself. -Peter Lord, make me chaste--but not yet. ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: XInput: device name in XListInputDevices
On Wed, 25 Jun 2003, Owen Taylor wrote: local-atom = MakeAtom(local-type_name, strlen(local-name), strlen(local-type_name) ? TRUE); -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: [XFree86] BadLength (poly request too large or internal Xliblength error)
On Wed, 25 Jun 2003, Erik Enge wrote: X Error of failed request: BadLength (poly request too large or internal Xlib length error) Can you send the xdpyinfo output for the boxes? Major opcode of failed request: 148 (RENDER) This is fairly new code so it might be a bug in that. Minor opcode of failed request: 20 (RenderAddGlyphs) Serial number of failed request: 5200 Current serial number in output stream: 5323 A normal run of this application usually takes about eight seconds (as timed on box-1). The error shows up after 7 seconds so it is very near the end of the run. The application, for what it's worth, pops up a window, does the conversion and then exits. Does anyone know what this problem might be if it is related to XFree86 in any way and how to solve it so that the application will run on box-2? Here's a vague guess: The application does something with text documents. It might look up some information about some fonts. Wine might in turn use Render to an off-screen pixmap or prepare to use a font or whatever. This then goes wrong. 4.3.0 contains updates to the RENDER extension in the server over 4.2.0. It also has updates in the client side library. You might also have different fonts installed on the two systems. An idea would be to disable the RENDER extension in /etc/XF86Config, restart the X server, and try again. -Peter Give a man a fish, and you'll feed him for a day; Give him a religion, and he'll starve to death while praying for a fish ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] X forwarding; 2nd attempt, can only help me please
On Mon, 23 Jun 2003, Vladimir wrote: display is not set and i dont get any forwarding, any help would be great Read on MIT cookies and X authorization + lots of SSH details. (in other words: use ssh from A (or B) to D directly to forward the X port and copy the cookie (in the .Xauthority file) over manually + set DISPLAY variable. Another idea would be to run a dumb proxy on C -- you could probably also trick SSH into that) What you want should be possible but it won't be easy or simple. And probably very few people have ever tried it and if they did they can't remember the details which would probably end up being very dependent on the versions of SSH involved and on lots of configuration details. -Peter ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] Large Ugly Black Border Surrounding X
On Wed, 18 Jun 2003, Kai Duncan wrote: about 4 inches of black on each side of the screen surrounding the Desktop. I haven't a clue what could be the problem and it just started after I upgraded. That it's a laptop and you asked it for a resolution lower than the maximum the LCD screen can provide and the designers chose to let it show borders instead of trying to resample the screen image badly (stretch it)? See if you can find out what the max resolution for your laptop is and edit the configuration file accordingly. -Peter Some day American bombs will be smart enough to have second thoughts about using wars to solve humanity's problems. -A. Whitney Brown ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [XFree86] [XFree86(TM) Bug Report] VMware and video drivers
On Wed, 5 Mar 2003, ken anderson wrote: I used the vmware_drv.o but i had to download it with Windows XP. it was my understanding that the bin sometimes get messed up this way. It would be nice if you could place a copy on your servers. possibly renamed. Right click and select download, that might work. -Peter Why should George W. Bush care what the American people think? After all they did not vote for him. ___ XFree86 mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xfree86
Re: [Xpert]Xserver problems with /temp/.esd
On Sat, 17 Aug 2002, Sadanand Kota wrote: Hi, When I try to start Xserver by 'startx', the X hangs after some time. When I kill it (Ctrl+Alt+Bakspace), I see these messages. esd: Esound sound daemon already running or stale Unix socket /tmp/.esd/socket This socket already exists indicating esd is already running. Exiting... I even deleted /tmp/.esd and tried to start x. But I am getting the same problem. This is not an X problem as such - it has nothing to do with programming X user interfaces, the internals of the XFree86 X Window server, the xlib library that applications use to communicate with the server, etc... Any idea? It's something to do with some script files that get executed when you start X. For reasons that seemed got at the time there are many potential script being executed behind the scenes. Window managers and the modern Unix/Linux GUIs like KDE and Gnome get started from these. Since you have problems with esd, you are probably running Gnome. If you got into this situation by installing vendorsupplied software, you should contact your Linux distribution vendor for paid support -- or try the free support in the online forums they undoubtedly have. If you got into this situation by installing third party packages (or from source), you should try something like gnome-list. Look for mailing lists at http://gnome.org/ -Peter ___ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
