Slashdot Mirror
XFree 4.1.0 Out
A reader writes "After some release candidates, final XFree86 4.1.0 is out. Check it from XFree86 FTP server. Their website isn't updated yet." Check the README for more information.
← Back to Stories (view on slashdot.org)
Sure, if the fbdev code was accelerated "enough" and supported all the primitives for 3D and whatever else they want, then it would be easier. If or course fbdev was on all the OSes they wanted XFree86 to work on. None of that appears to be true.
Now they could divert that effort in making and extending graphics drivers from X to fbdev itself, but that would limit the OS choice and be a whole lot harder. Writing kernel code is very painful compared to writing user land code. Actually it isn't the writing that is so hard, but the debugging. My relatively few kernel modifications have taken about 4 to 5 times longer then I had guessed to debug, and I had already bumped the number up because I thought I knew how hard it would be...
Plus XFree86 is used by a lot more then just Linux, and I believe many of the XFree86 commiters are not Linux-only users (in fact the only XF86 commiter I personally know is a rabid anti-Linux zealot). So I doubt they would want to abandon support for OSes that didn't have a usable fbdev.
One small point in this arguments favor is that they have a fbdev driver, and have continued work on non-fbdev drivers.
So it seems to be the lesser of N evils to essentially have a user land device driver of inordinate complexity.
You can have multiple X servers running, one in each virtual terminal. I have a set of scripts that automatically launch a new X server with some minimal twm setting, so as to run applications full screen: view DVI/PS/PDF documents, JPG/GIF pictures, HTML documents with embedded pictures, play games, etc. Very handy when you're in a console switching period and your preferred window manager is the linux kernel's virtual terminal facility.
In case you don't want to roll your own from scratch, I can make my scripts available on request: they work very well for me, although it would require quite a lot of work to package them into a RPM, and you'll have to dig the CVS for versions that used to work with older X setups.
Unix sucks so much.
-- Faré @ TUNES.org
-- Faré @ TUNES.org
Reflection & Cybernet
If you have a Matrox G400/450 card then look out for mplayer http://mplayer.sourceforge.net/homepage/ It playes all the formats divx, avi, mpeg .... that have the codec in windows.
Even if you don't have G400/450 you would get a decent performance.
--
Sukanta
Faster computers mean that we can increase the abbstraction layers between the hardware and the application to make it EASIER and FASTER to develop complex applications.
>>>>>>>>>
Yuck. Ugly UNIX thinking at its worst. Faster computers should allow you to do more, not devote more resources to administrative overhead.
Unix thinking? Please. Windows has generally been much worse about hardware-hungry.
And really, the basic theory is sound. As you point out, as computers get faster you can use that speed to make old programs run faster or to make new programs possible at the same perceived speed. Sure, Excel's a pig, but 99 times out of 100 I'd rather use Excel than Visicalc.
But even if you're just going to make the old programs over again with absolutely no feature improvements, it often makes sense to burn processor time to save developer time and make the software cheaper. That's why you see the rise of languages like Perl, Python, C++, and Java and the near disappearance of significant assembly work.
CPU effciency is only one of many criteria you can optimize for. And since more than 90% of computers spend more than 90% of their time idle, other things (like usability and development cost) are often more important.
They are denying Anonymous access now. I guess they wanted to avoid the Slashdot affect as much as possible.
"Imagination is the only weapon in the war against reality." -Jules de Gautier
I don't like reading Freshmeat. Too much crap
WHAT?! You mean you didn't like this week's 4 dozen incarnations of MySQL enabled MP3 playlist maintaining programs? You must be crazy...
---
python -c "x='python -c %sx=%s; print x%%(chr(34),repr(x),chr(34))%s'; print x%(chr(34),repr(x),chr(34))"
I've been following the latest DRI updates (binary) from http://dri.sourceforge.net and currently the only version available is for 4.0.3 (obviously).
:)
So my question is: How new are the DRI drivers included in 4.1.0? Should I assume they're the current CVS from the DRI project? Could someone please verify this before I destroy my current drivers?
Thanks!
ftp://ftp.calderasystems.com/pub/mirrors/xfree866
ftp://carroll.cac.psu.edu/pub/XFree86
ftp://ftp.cs.umn.edu/pub/XFree86
ftp://download.sourceforge.net/pub/mirrors/XFree8
ftp://ftp.freesoftware.com/pub/XFree86
ftp://ftp.infomagic.com/pub/mirrors/XFree86
ftp://mirror.sftw.com/pub/XFree86
ftp://phyppro1.phy.bnl.gov/pub/XFree86
ftp://ftp.rge.com/pub/X/XFree86
ftp://ftp.valinux.com/pub/mirrors/xfree86
Hey, does anyone know if there have been any improvements into getting XFree86 to play nice with the Linux framebuffer console?
Currently, I feel the best way to play a full screen game is in a virtual terminal under fbdev rather than through X, so that I can keep my desktop and the game separated. Ctrl-Alt-F3 takes me to my game, and Alt-F7 takes me back to X. Plus, the high-resolution framebuffer console is very cool just for shell usage too.
Unfortunately for me though, the XFree86 driver for my card clashes with the framebuffer driver from the kernel. So in order to have my super-cool framebuffer console *and* X, I have to use the fbdev version of X. This works and all, but it is unnaccelerated which means I can't do much. Fortunately, most of the games I have are SDL based, so I can just flip to a console to play. But something like OpenGL? Not a chance.
Now can anyone explain to me why there are two video drivers on my system to clash in the first place? It seems logical to me that everyone should use a framebuffer driver (I have a Voodoo3, and it is supported in the kernel) and that X should ride on top. The kernel has every other driver, so why is the video driver so special that X has its own?
Wouldn't it be easier for the XFree folks if they didn't have to worry about making video drivers, and instead it was all taken care of by the kernel? IMO, the only driver they should develop is the fbdev version.
Anyone else have any insights to this? Any reason why DRI couldn't be applied to an fbdev Xserver?
-Justin
570. Add euro locales and some other missing locales to locale.alias and locale.dir (#4662, 4665, 4667, Mike Harris). 569. Fix Romanian XKB map (#4664, Mike Harris). 568. Spell Portuguese correctly in XKB lst files (#4663, Mike Harris).
EEK! Mike Harris is the premier of Ontario, Canada. And also I guy I hate. Now he's working on Free Software?!?! What is the world coming to....
--Volrath50
Recent Changes to the XFree86 4.1 branch
Below is a recent extract from the XFree86 change log for the 4.1 branch. The full change log can be found in the XFree86 source tree (xc/programs/Xserver/hw/xfree86/CHANGELOG).
XFree86 development code can be accessed directly from the CVS repository. Information about this can be found on our CVS page.
XFree86 4.1.0 (2 June 2001)
619. Disable PCI resource conflict checking for Linux/Alpha (Jay Estabrook).
XFree86 4.0.99.902 (1 June 2001)
618. Fix Linux xf86GetPciSizeFromOS() parsing when the kernel is 64 bit
and any base or size is larger than 32 bits in magnitude (#4732,
David S. Miller).
617. Make XDarwin ddx pass up proper right and middle mouse button numbers
and fix mouse button 5 (Christoph Pfisterer and Torrey T. Lyons).
616. Restore backwards compatibility from 4.0.[2,3] to 4.1.0 for
the i810, r128 and radeon DRI drivers (Gareth Hughes).
615. Fix a problem when using patterns of horizontal lines with the mga
video overlay (#A.442, Ewald Snel).
614. Xinstall.sh updates and bug fixes (David Dawes).
613. Remove duplicate XineramaLibrary section in X11.tmpl (#4731,
Mike Harris).
612. Enable building DRI for Linux/ppc, and fix a drm-related bug
for Linux/ppc (#4728, 4730, Michel Dänzer).
611. Document Options for the r128 and fbdev drivers (#4727, 4729,
Michel Dänzer).
610. Add a BuildBindist switch which causes a file containing the XFree86
version number to be installed in ProjectRoot, include this in
the Xbin bindist tarballs, and turn on this switch in the bindist
host.def files. The purpose is to allow the installer script to
easily identify which version the bindist tarballs are (David Dawes).
609. Resync bindist and Xinstall.sh with changes made for 4.0.3 (David Dawes).
608. Fix the Shape extension's XShapeCombineMask to handle cases where
src_mask is None according to the spec. This reportedly fixes an
X server crash (#4715, Huver).
607. Make sure -UXF86DRI is after -DXF86DRI when compiling vfb/miinitext.c
(#4714, Frederic Lepied).
606. Fix ATI Radeon driver on Alpha. Seems as though the BIOS doesn't
like Re-POSTing and memory setup gets confused. (Jay Estabrook, Jeff
Weidemeier)
605. Fix build for Cygwin/XFree86 (#4711,#4713 Harold Hunt).
604. Fix problem with Xinstall.sh on Darwin 1.3.x (#A.431, Stefan Pantos).
603. Update Xinstall.sh and Darwin bindist directories to optionally
install Quartz support and to add an x86 distribution (Torrey T. Lyons).
XFree86 4.0.99.901 (29 May 2001)
602. Add missing return value for miSetPixmapDepths() (#4708,
ISHIKAWA Mutsumi).
601. Fill in the v4l man page template with some useful information (#4707,
Gerd Knorr).
600. Fix FFB OpenGL SwapBuffers (#4705, David S. Miller).
599. Work around a problem building the rstart specs doc with a symlinked
build tree (David Dawes).
598. Remove SPARC-specific byte-swapping code that would not work on older
SPARC CPUs (part of #4653, David S. Miller).
597. NULLify mapVidMem() and remove DEV_MEM #define for Linux/SPARC
(#4651, David S. Miller).
596. Fix Glint 300SX+Delta support. Add faster 500TX text acceleration
based on other code (Alan Hourihane).
595. Fixing MTRR split code (hopefully) (Egbert Eich).
594. Fixing coredump when doing vbeFree() twice: S3 Virge and C&T
(Egbert Eich).
593. Fixing HWCursor for mga driver in fbdev mode (Egbert Eich).
592. Fix xmh's use of XtNewString() with getenv (#4694, Tim Waugh).
591. Xdm/PAM fixes: leave it to PAM to observe whether or not an account
is locked, and reinitialize credentials after calling initgroups(),
because sometimes the credentials pam_setcred() gives are in the
form of group membership (#4693, Mike Harris).
590. Add an encodings file for standard box drawing characters for
VT100-compatible terminals (#4691, Juliusz Chroboczek).
589. Fix warnings when building mieq.c (#4689, Adam Sulmicki).
588. Fix some bugs in the cz and sk entried in XKB's keymap/xfree86 file
(#4692, Ivan Pascal).
587. Add 'hr' entries to XKB's keymap/xfree86 and rules/xfree86.lst files
(#4687, Nerijus Baliunas).
586. Include in shape.h to get Region typedef (#4686,
Adam Sulmicki).
585. Acceleration fixes for GLINT Permedia1 (Alan Hourihane).
584. Ensure glint driver chips don't exceed the specified virtual sizes.
(Alan Hourihane).
583. Remove all VGA'isms from the glint driver, it doesn't need them
(Alan Hourihane).
582. Support the Delta in the glint driver, needed for boards that have
the Delta connected to the rasterizer, as it acts as an arbiter for
the bus. Resolves acceleration troubles. (Alan Hourihane).
581. Add an lv entry to XKB's keymap/xfree86 file (#4685, Nerijus Baliunas).
580. Fix some typos in XKB's xfree86.lst file (#4684, Nerijus Baliunas).
579. Add DDXOSVERRORF ifdefs to the XFree86 ddx code that make use of the
OsVendorVErrorFProc feature (#4678, Michel Dänzer).
578. Convert the r128 driver's "UseBIOSDisplay" option into a more general
"Display" option (#4678, Michel Dänzer).
577. Treat GL_POINT like GL_POINTS and GL_LINE like GL_LINES in the sunffb
DRI driver (#4677, David S. Miller)
576. Fix bsdLib.rules and bsdLib.tmpl problems that show up when
X11ProjectRoot is defined (#4676, Johnny C. Lam).
575. Fix Trident XVideo colorkey at depth 15, 24 (Alan Hourihane).
574. Fix a typo in the lv XKB description, and fix things so that it gets
installed (#4675, 4679, Andris Pavenis).
573. Fix some apm driver bugs, including one that prevented acceleration
from working (#4674, Loïc Grenié).
572. Fix 555 (depth 15) palette handling in the i810 driver (#4673,
Andrew C. Aitchison).
571. [SECURITY] Fix authentication issues with mmap() on drm devices
(Jeff Hartmann).
570. Add euro locales and some other missing locales to locale.alias and
locale.dir (#4662, 4665, 4667, Mike Harris).
569. Fix Romanian XKB map (#4664, Mike Harris).
568. Spell Portuguese correctly in XKB lst files (#4663, Mike Harris).
567. Fix new ioperm calls in lnx_video.c for Alpha that are not needed
(Jay Estabrook).
566. Fix problems with assembler file dependencies when using gccmakedep
with the build (Frederic Lepied).
565. Finish DRI resync, including tdfx driver updates for textured video
support (VA Linux Systems).
564. Fix formatting of max clock reported by DDC (Marc La France).
563. Update Japanese localization of XDarwin help file (Toshimitsu Tanaka).
562. Update XDarwin man pages, help files, and version info. Add option to
build XDarwin.app bundle for deployment (Torrey Lyons).
XFree86 4.0.99.900 (18 May 2001)
561. Add an XKB description for Latvian (lv) keyboards (#A.411, Ilya Ketris).
560. Resync with DRI CVS trunk (VA Linux Systems).
559. Savage driver updates, including compiler warning fixes, document
the "ShadowStatus" option in the man page, and fix an argument
mismatch between ShadowWait and SavageWaitQueue (#4661, Tim Roberts).
558. Update the wacom driver to add a "ScreenNo" option to allow a tablet
to be attached to a screen in a multi-head setup, and to add auto-
detection of USB line and max parameters of USB tablets (#4640,
Frederic Lepied).
557. Add a README file that has information about enabling the extra buttons
on the IBM Rapid Access keyboard (#4639, Dennis Bjorklund).
556. Fix some Slovene/Slovak confusion in locale.dir/locale.alias files
(#4638, Kamil Toman).
555. New XKB keymaps for cz and sk (#4634, 4637, Kamil Toman).
554. Updates for the iso8859-2 Compose file (#4634, Kamil Toman).
553. Check V_CSYNC in the r128 driver, and fix building with R128_DEBUG
enabled (#4631, Michel Dänzer).
552. Mesa 3.4.2 (and later) import.
551. More build & warning fixes (Marc La France).
550. Fix bug that caused hardware cursors to be temporarily moved during mode
switches (Marc La France).
549. Optimise HARDWARE_CURSOR_AND_SOURCE_WITH_MASK case (Marc La France).
548. Move xf86CursorScreenRec definition into xf86CursorPriv.h
(Marc La France).
547. Fix BIOS retrievals in MGA driver (Marc La France).
546. Fix ATIProbe() for newer Rage128 and Radeon chips (Marc La France).
545. Add temporary workaround in ATI driver for interrupts that occur on
PowerPC's upon PCI master-aborts (Marc La France).
544. Update XDarwin to use fb and support Render (Torrey Lyons).
543. Back out sunleo conversion to fb. This driver is too heavily dependent
on cfb32 for a simple fb conversion (Marc La France).
542. Miscellaneous build/warning fixes (Marc La France).
541. More prep work for SunOS (Marc La France).
540. Fix libXft build on SunOS (Marc La France).
539. Another makedepend bug fix (Marc La France).
538. Fix use of xftcache utility during !UseInstalled builds (Marc La France).
feast and enjoy.
-bacchusrx.
Life after capitalism? The participatory economics project
There is a functioning unofficial apt source for XFree 4 debs: http://people.debian.org/~cpbotha/ but 4.1 isn't there yet. Perhaps later