Release Notes for XFree86[TM] 3.3.1
: What's new in 3.3.1?
: Bug fixes
Previous: What's new in 3.3.1?
Next: Known Problems
3.1. Bug fixes
- XFree86 3.3.1 includes The Open Group's public patch 2 for X11R6.3.
- Build problems that showed up on some OSs have been fixed.
- Support for SCO Open Server 5 should now be complete.
- A malloc problem in libXt which showed up on FreeBSD has been fixed.
- Depth-specific DacSpeeds are now implemented for the ET6000.
- Depth-specific DacSpeeds are fixed for the S3 server.
- HW cursor problem with the I128 server has been fixed.
- I128 Series II rev 2 chips are now supported.
- Xterm will now startup on Linux if /etc/termcap is missing.
- Various problems with the S3V server and the SVGA s3v driver have been
fixed.
- A problem with the clock limit for some revisions of the Circus 5434
has been fixed.
- The Mach64 server will now correctly recognise some of the newer
ATI chip revisions, including the Rage II+, Rage Pro and VT3. If
you needed the ChipId/ChipRev workaround when using 3.3, you should
remove those lines from your XF86Config file when upgrading to 3.3.1.
- An initialisation problem in the S3 server that shows up when the
ramdac type is given in the XF86Config file should be fixed.
- The MGA driver now defaults to using the software cursor because
some people have reported problems when using the hardware cursor
with Millennium cards.
- Lockups with the MGA driver that happen on some SVR4 versions have
been fixed. Lockups when the server crashes and dumps core on some
OSs have also been fixed where possible.
- The xterm termcap field for turning off colour has been fixed.
- A server crash that happens when starting some servers on Solaris
has been fixed.
- Some problems with the Trident 9860 and 9685 chips have been fixed.
- A problem with xterm writing an invalid wtmp entry on Linux has been
fixed.
- A PolyPoint bug in the S3V server has been fixed.
- Screen wraparound problems with the S3 server that show up on some
old Number Nine GXE level 10 cards have been fixed.
- A problem with the PCI framebuffer remapping in the S3 server has been
fixed.
- A problem with XF86Setup not showing the correct chipset-specific
README file has been fixed.
- A problem with XF86Setup not setting up the link to the Xserver in
some situations has been fixed.
- Some libXt error/warning messages were partially duplicated, and this
is now fixed.
- Some line drawing problems that show up with the Cirrus 542x chips
have been fixed.
- A problem drawing wide fonts with some Cirrus chips has been fixed.
- Some bugs in some XKB symbols files have been fixed.
- The Chips and Technologies 65555 and 68554 are now detected by
the chips driver.
- `xset dpms' didn't accept some parameters correctly.
- Some raster op bugs in the SVGA server's s3v driver have been fixed.
- Problems with the Mach32 server that show up with some AST motherboards
that have an on-board Mach32 chip have been fixed (see the
README.Mach32 file for details).
- A bug in the ET6000 driver which can cause the server to crash on
non-Linux systems has been fixed.
- An initialisation problem with the ET6000 driver which can result in
a black screen has been fixed.
- The Eraser support in the Wacom driver has been fixed.
- Support for two relative devices has been fixed in the Wacom driver.
- The DPMS state is now correctly reset when switching back to the
Xserver's VT.
- A bug in the 24bpp framebuffer code which caused a server crash
when running StarOffice has been fixed.
- An Xserver bug which could cause a server crash when using lbxproxy
has been fixed.
- A conflict between Imake.rules and Motif.rules has been fixed.
- Some problems that show up with some accelerated servers on SVR4
when using Xqueue have been fixed.
- A problem that can cause a server crash when drawing some arcs has
been fixed.
- The MGA driver now recognises the newer 220MHz Mystique chips.
- The MGA driver has experimental support for the Millennium II.
This really is very new, and is largely untested. There are known
problems (see README.MGA for details).
- The "noaccel" option could cause a lockup with the Cirrus Laguna
chips (546x). This option is now disabled for those chips.
- Some blitter timeouts that show up with Cirrus 7548 and 7555 chips
should now be fixed.
- The handling of the PS/2 mouse protocol is fixed for OpenBSD.
- The SiS driver should now correctly determine the MMIO address.
- Support is added for Number Nine I128 DRAM cards. There are some
text mode restore problems at some depths though.
Release Notes for XFree86[TM] 3.3.1
: What's new in 3.3.1?
: Bug fixes
Previous: What's new in 3.3.1?
Next: Known Problems