work properly this time on FreeBSD/sparc64 so stop using workaround that
defaults to old 'keyboard' one and remove erroneous patch by me.
Pointed out by: Matthias Muthmann (mmuthmann at gmx.net)
Tested by: Matthias Muthmann (Ultra10/type5c), me (Ultra5/type5)
Obtained from: Xorg CVS
soon-to-be-committed-upstream patch to fix the build after the removal of
xf86drmCompat.c. Also includes the fix for ports/32121, which had been
committed.
[2] Add the PCI ID for Radeon X600, 0x5B62 to a place that was missed.
PR: [1] ports/32121
Submitted by: [1] alane
[2] Vladimir Kushnir <vkushnir@i.kiev.ua>
NO_SUID_XSERVER to disable this. The end result is same as it was before
with x11/wrapper installed as the code from there is included in X
server. It is also the same as xorg-server port does (from which this
changes were taken). This should conclude removal of dependency on
x11/wrapper from meta port for above reasons. [1]
- (patch-bsdResource.c) Correct the ranges of the bus resource windows on
sparc64. This makes PCI video cards work that hang off directly from
the Host-to-PCI bridge in machines like the Blade 100. [2]
- (patch-HALlib_bindings.c patch-mga_driver.c) Revert the MGA HALlib
MGASetDisplayStart interface back to HALSetDisplayStart (taken from
upstream CVS) [3]
Noticed by: (bugsgrief at bugsgrief.net) [1]
Submitted by: marius [2]
Reported and tested by: Steven Friedrich (FreeBSD at InsightBB.com) [3]
Ports changes:
- Drop dependency on x11/wrapper. X server already contains this code
(programs/Xserver/os/utils.c rev. 3.58).
- Install default SecurityPolicy with XFree86-4-libraries, thus avoiding
plists and who-owns-what complications between different fooServer ports.
Copy default file to real one if that doesn't exist and remove both upon
uninstall if they are the same.
- Revert change to Imake.rules rev. 3.132 in devel/imake-4 port, since
couple of ports break with this change
(devel/imake-4/files/patch-Imake.rules)
- Add grehans patches for PPC [1]
- Several of directories that were previously in /etc/mtree/BSD.x11-4.dist
were moved to respective ports pkg-plists
PR: [1] 78345, 78346, 78348, 78352, 78353, 78354, 78355, 78356
This makes PCI video cards work that hang off directly from
the Host-to-PCI bridge in machines like the Blade 100. [1]
- Default to old 'keyboard' driver for sparc64. Patches for 'kbd'
one were not enough so until new driver is fixed, this change
makes Xorg work with sparc64 keyboards. Knob WITH_SPARC_KBD is
provided to ease efforts of fixing 'kbd'.
Submitted by: marius [1]
is a port of the server pieces of the periodic snapshot of X.Org, with the
development DRI drivers enabled. The purpose of this port is to bring
bleeding-edge graphics support to FreeBSD.
With Patch:
(pII 400Mhz)
real 1m12.597s
user 0m14.706s
sys 0m12.061s
(athlon 2500+ 3disk raid 0)
real 0m12.912s
user 0m2.884s
sys 0m3.128s
Without Patch:
(pII 400Mhz)
real 1m43.233s
user 0m14.565s
sys 0m13.951s
(athlon 2500+ 3disk raid 0)
real 0m22.967s
user 0m2.925s
sys 0m3.576s
Submitted by: ahze
Xdmx is proxy X server that provides multi-head support for multiple
displays attached to different machines (each of which is running a
typical X server). When Xinerama is used with Xdmx, the multiple displays
on multiple machines are presented to the user as a single unified screen.
Port based on xorg-nestserver.
Requested and tested by: Henry Miller (6.8.1 version)
Release notes are available at
http://www.x.org/X11R6.8.2/doc/RELNOTES.html
Thanks to kris and krion for running several cluster test builds,
maintainers of GNOME for prompt responses, portmgr for postponing ports
freeze for this update, testers on FreeBSD-X11@ list and others that I
might have mised here.
Also included:
- fix for ATI Mobility on Dell Inspiron 7500 (obtained from Marc Aurele La
France; obtained and tested by julian)
- fix for kbd driver on Sparc64 (tested by Aaron Dudek, Michael G. Jung and
Matthias Muthmann), which still appears to have problems with some
keyboards - so
- fix for kbd driver on PC98 (reported and tested by NAKAJI Hiroyuki; PR
ports/77217)
- fix for i810 on HP D530 (obtained from Egbert Eich; obtained and tested
by Anders Nor Berle; PR ports/74757)
- Install driver manpages with 'x' postfix so the ones that clash with
FreeBSD ones are easier to obtain. [2]
- Fix plist.
PR: ports/25576 [2]
Obtained from: XFree86 CVS [1]
Reported and tested by: julian [1]
Font cache is not build by default, which makes related lines in
default configure file wrong, which in turn causes xfs to error out
on startup with default file. Tweak generation of said file in a way
that would make proper configuration depending on existance of font
caching code, so that this patch can be sent upstream. [2]
Inspired by: Jose M Rodriguez (josemi at freebsd.jazztel.es) [1]
Bits and pieces for [1] stolen from:
src/etc/rc.d/xfs, apache2 and squid ports
Noticed by: Gary Dunn (knowtree at aloha.com) [2]
Add rcNG startup script. [2]
Noticed by: Gary Dunn (knowtree at aloha.com) [1]
Inspired by: Jose M Rodriguez (josemi at freebsd.jazztel.es) [2]
Bits and pieces for [2] stolen from:
src/etc/rc.d/xfs, apache2 and squid ports
[2] Put the manpages from section 4 under 4x. For example, you can do
"man 4x vga" to get the X vga(4x) manpage, still do man 4 cirrus to get
cirrus(4x), and man -a vga to get both vga manpages.
[3] Apply several important-looking fixes to ATI drivers from the stable branch:
- Fix mode setting with Mach64 on some laptops.
- Don't crash with mach64 for some PCI configurations
- Improve hang detection on r128s. (may improve speed in DRI, too)
- Avoid a hang on server regen on r128 with DRI.
- Fix the DynamicClocks support, which could have caused hangs with DRI on
original Radeons.
PR: [2] ports/25576
Submitted by: [2] fanf
Requested by: [1] peter
Obtained from: [3] XORG-6_8-branch
of features supported by the open-source Matrox driver for X.Org and XFree86.
This is a newer version than that discussed in the original PR, and has only
been tested briefly by myself on a G400 with X.Org 6.8.1. Note that though the
readme.txt says the file is redistributable, the only licensing information to
be found is the click-through on the website which says redistribution is not
allowed.
PR: ports/55699
Submitted by: Jens Rehsack, rehsack at liwing dot de
change that occurred.
[2] Install the xorg.conf referred to by xorg.conf(5). Installation was
disabled during the XF86 name avoidance phase, but could still be of use.
[3] Add a section about the Composite extension to xorg.conf.
[4] Add MakeDllModules YES to the WITH_DEBUG flag's options. While it won't
package correctly, you can't actually debug modules (where crashes almost always
happen) without it.
[5] Merge fixes from XORG-6_8-branch for i810 driver crashes in on CloseScreen,
ringbuffer lockups, and various warnings.
[6] Recognize R_ALPHA_SREL32 relocation type in elfloader.
PR: [1] ports/75425
[2] ports/73915
[6] ports/73257
Submitted by: [1] Marc Silver, marcs at draenor dot org
[2] Frank J. Laszlo, laszlof at vonostingroup dot com
[6] Ken Stailey kstailey at yahoo dot com
[6] r1.4 of X.Org CVS.
Approved by: portmgr
I had moved libXThrStub into libX11, because it seemed unnecessary. However,
because libc_r symbols are also weak, the weak stubs in libX11 would interfere,
however they don't if they are in a library linked by libX11. The most common
manifestation of the problem was:
GThread-ERROR **: file gthread-posix.c: line 137 (g_thread_impl_init): error
'Invalid argument' during 'pthread_getschedparam (pthread_self(), &policy,
&sched)'
in glib apps.
PR: ports/75477
Submitted by: KIMURA Yasuhiro, yasu at utahime dot org
Testing by: marcus
DRI drivers are incompatible with the old libGL in XFree86, they have been
repocopied to graphics/xfree86-dri. Also note that with this commit the html
manpages are going away, and with it the runtime dependency on perl for the
imake port.
Release notes at: http://www.x.org/X11R6.8.1/RELNOTES.txt
Props to: kris (multiple cluster runs and sorting through logs)
lesi (fixing all the issues in those logs, and more)
submitted by Satoshi Kimura to both x11@freebsd.org and XFree86. It is
already in XFree86 CVS (Changelog #125) so the diff was taken from there.
Also install two files with Japanese documentation.
Submitted by: Satoshi Kimura
Fix behaviour of keyboard under sparc64, where keycodes were off by one
(i.e., when pressing 'd' on keyboard, server would get code for 'f'). [2]
PR: ports/71937 [1]
Submitted by: Satoshi Kimura [1], marius [2]
Approved by: portmgr (linimon), mentor (anholt)
