1.3.2:
Jeremy Huddleston (5):
darwin: Buildfix for Tiger which doesn't have posix_spawn
startx: Choose an unused $DISPLAY by default on all platforms
launchd: Add an option (--with-launchagent-xserver) to set the default X11 server
Use default clientargs, serverargs, and display if none were set
configure.ac: 1.3.2
Julien Cristau (1):
Include <sys/time.h> and <sys/resource.h> unconditionally
1.3.1:
For most users, the main changes in xinit-1.3.1 focus on build system cleanup,
but darwin has seen a fair number of other changes for the LaunchAgent and
LaunchDaemon, specifically with respect to logging to ASL. These changes
mirror similar changes coming in xorg-server-1.11.x (currently git master).
1.3.0:
This release adds support for an xinitrc.d directory to contain startup
script snippets on all systems. It also features a number of modernizations
and cleanups of the code base.
This is mainly just a bunch of updates for OSX Tiger compatability and
SnowLeopard fratures.
Alan Coopersmith (1):
Fix build of startx & xinitrc with Solaris make
Jeremy Huddleston (17):
Apple: privleged_startx: Increase the timeout so slower
machines don't get a stray process warning in their system
log.
startx.cpp: Made some things a little more friendly for
people who have spaces in pathnames.
Apple: Using eval to protect people who have spaces in their
home path.
Support the instant-off API in SnowLeopard.
Apple: sleep() to avoid innocent error message spew
Apple: Make VPATH builds work on darwin
Apple: Use AvailabilityMacros.h for Tiger compatability
OSX: Create privleged_startx on Tiger
OSX Tiger: More compilation fixes...
OSX Tiger: Hopefully the last fix for Tiger users...
Apple: Honor X11_PREFS_DOMAIN if it's set (1.4.2-apple26
and later)
OSX Tiger: Fix discovery of mach port
OSX: Fixed bad prefix
OSX: More prefix fixes
OSX: Don't use login/bash -c hack anymore for org.x.startx
LaunchAgent
Apple: Allow defaults option that maps to a -dpi server
command line in startx
xinit 1.1.1
Adam Jackson (1):
xinit 1.1.0
Jeremy Huddleston (2):
Properly handle using xserverrc
Apple: Don't use launchd on Tiger since it doesn't support
the features we need.
Adam Jackson (2):
Unset session environment variables so startx works within a running session.
xinit 1.0.9
Alan Coopersmith (3):
Use od on /dev/random to generate cookie on Solaris
Check for proper cookie creation on all platforms
Use /bin/ksh instead of /bin/sh on Solaris
Jeremy C. Reed (1):
Use /dev/urandom for cookie it it exists
Jeremy Huddleston (14):
Apple: Fixed issue where display would not be passed correctly to the server if nolisten tcp was enabled.
Apple: Execute /usr/X11/bin/privileged_startx if it's present
Apple: Renamed LaunchAgent to org.x.startx to distinguish it from the org.x.X11 application
Correctly handle users with spaces in $HOME
Apple: privileged_startx is in XINITDIR
Apple: We renamed font_cache.sh to font_cache for better consistency... Changing startx to work with this
Apple: Wait for privileged_startx to finish, so we don't race to create /tmp/.X11-unix
Apple: Added privileged_startx
Apple: Fixed make dist to include a missing header and not include generated sources
Cleaned up handling of defaultserver{,args}
Apple: Use CFProcessPath instead of argv[0] trick.
Apple: Added some OSX version protection, so this will still compile on Tiger and below.
Apple: No need for these bits being here in xinit as of xserver-1.4.0-apple9
Apple: Don't need to spew unneccessary 'directory exists' messages.
Matthieu Herrb (2):
OpenBSD: use openssl rand -hex 16 to generate the cookie.
Replace sprintf() and strcpy() with snprintf().
if no filename given, but this isn't true with the native HP compiler, so
teach them to use '-' explicitly.
XXX using cpp(1) to postprocess manpages is ugly, most of the other X.org
packages seem to have been converted to use sed(1).
Adam Jackson (1):
xinit 1.0.8
David Nusinow (1):
Unbreak installs on non-launchd systems
James Cloos (1):
Add missing PHONY line for automatic ChangeLog generation
Jeremy Huddleston (19):
Added some support for Xquartz on OS-X.
Added Apple launchd support
Changed x11-exec to actually start X11.app. X11.app will be responsible for calling xinit.
OS-X: Added argv[0] hack for finding Xquartz UI and icon as well as claiming its dock icon
Added support for OS-X preferences to disable tcp/ip connections and xauth
OS-X: Made defaults test more robust
OS-X: Need to use bash --login to setup the environment from launchd.
XQuartz doesn't need the -launchd command line argument anymore.
OS-X: Properly set enable_xauth if the user has no defaults set and choose a valid $display for fast-user-switching.
OS-X: Add bindir to $PATH in startx if it's not there
OS-X: Fixed typo
OS-X: Set fontpath in xinitrc to include system / user fonts. Additionally fixed font scaling issues (trac #52)
Apple: use -nocpp arg to xrdb if cpp is not installed
Apple: Cleaned up font path setting and added font caching to xinitrc
Apple: Support spaces in $HOME for font path, and remove duplicate entry
Added org.x.X11.plist.cpp to EXTRA_DIST
Apple: font caching and defaults
Apple: Adding support for an xinitrc.d directory on OSX, so fink doesn't need to clobber files
Apple: Use -c in the launchd plist, so broken .profile/.bashrc don't much up startx
Matthieu Herrb (1):
Use /dev/arandom instead of /dev/urandom to generate the cookie
root (1):
Define APPLE for the rest of us
including:
-- Bug #12448: Don't try to trap 'ERR'.
startx is advertised as a POSIX sh script. These shells don't
necessarily support trapping 'ERR'. This makes startx work again with
dash (and probably others).
-- Fix broken test for /dev/random cookie generation failure.
-- Use /dev/random to generate the cookie if mcookie is unavailable.
-- Setting XAUTHORITY without having actually generated a cookie and created
.Xauthority led to issues if somebody like ssh later came around
and made the .Xauthority file for their own setup. So, simply make
it so that we never fail to create one.
-- Remove .serverauth.$$ file on signals or errors
pkgsrc changes:
Remove cookie maker environment variable from CONFIGURE_ENV, as the
script now does the "right thing".
tested by me, seems to work fine, ok'd by agc@
