Age | Commit message (Collapse) | Author | Files | Lines |
|
Regression introduced by b1c3dc6ae226db178420e3b5f297b94afc87c94c.
Shutting down the libhal_ctx if the init failed may cause an abort.
This can happen if hald is not yet running at server startup.
X.Org Bug 23213 <http://bugs.freedesktop.org/show_bug.cgi?id=23213>
Tested-by: Stefan Dirsch
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 49046088f10cceaea7da97401d742d3fb59371f5)
|
|
This patch simplifies error handling in the HAL code and fixes a
segfault if libhal_find_device_by_capability() failed.
Fixes http://bugs.gentoo.org/278760
Based on a patch by Martin von Gagern <Martin.vGagern@gmx.net>
Signed-off-by: Rémi Cardona <remi@gentoo.org>
Acked-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit b1c3dc6ae226db178420e3b5f297b94afc87c94c)
|
|
Fixes Solaris bug 6801386 Xorg core dumps on startup if hald not running
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6801386
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
(cherry picked from commit 1e816065e5ec3b9394dc1fa5815457a664e15fd9)
|
|
If HAL isn't available when we try to connect, the registered NameOwnerChanged
signal handler waits until HAL is available. Once we connected to HAL, we
unregister the signal handler again.
This allows HAL to be started in parallel or after the server has started.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 4844bff58f296b2851be4e6b955c3a68d02437a9)
Signed-off-by: Keith Packard <keithp@keithp.com>
|
|
They've promised to fix Solaris printf soon to check for NULL pointers
instead of segfaulting, but that won't help people on existing releases.
|
|
If HAL is restarted, the device list is sent to the server again, leading
first to duplicate devices (and thus duplicate events), and later to a
FatalError "Too many input devices."
dev->config_info contains the UDI for the device. If the UDI of a new devices
is equal to one we already have in the device list, just ignore it.
Signed-off-by: Peter Hutterer <peter.hutterer@redhat.com>
|
|
|
|
|
|
For backwards compatibility with server 1.4.
|
|
X.Org Bug 16874 <http://bugs.freedesktop.org/show_bug.cgi?id=16784>
|
|
Not only was this pretty ugly, but it didn't even work on systems
without strcasestr anyway, due to the define not being in dix-config.h.
Lack of strcasestr is handled transparently with the version from
FreeBSD now anyway, so, huzzah.
|
|
|
|
Removing the device invalidates its ->next pointer. Copy it aside before
destroying the device.
|
|
|
|
The HAL spec says that input.xkb.{rmlv}* can be sent, but if the user
specifies a X-specific {rmlv}, then this is overridden through the use of
input.x11_options.Xkb{RMLV}.
However, the way how the server parses options--by ignoring capitalisation,
underscores and spaces--the HAL and the x11_options would override each other.
So we simply filter the options, letting Xkb{RMLV} override xkb_{rmlv} and
only actually add them to the device after parsing _all_ options.
* rmlv ... rules, model, layout, variant
See Bug 13037 <http://bugs.freedesktop.org/show_bug.cgi?id=13037>
(cherry picked from commit fc35d1e3be201e3821413bb2eeb8d43e1e56ba17)
|
|
It makes my vim look ugly. Put "let c_space_errors=1" into your .vimrc.
(cherry picked from commit 1f54c05cf8a6b82e5fc6362f7f8e8fdc2444b9e8)
|
|
Parse "input.x11_options" and pass every key/name pair to the driver.
Remove check for input.capabilities, because that's part of the fdi files.
Thanks to Dustin Spicuzza <dustin@virtualroadside.com> for the patch.
|
|
Thanks to libdbus' extensive use of assert we won't just get an error, it'll
bring the whole server down for us.
(cherry picked from commit fb07fab2c07e7b0834724541dc47bfba02ba8574)
|
|
(cherry picked from commit 7732898aaa70e076000f6e6aa9420482896ed996)
|
|
This showed up in Xephyr in particular, which denies new device requests.
|
|
|
|
|
|
Actually combine the XKB options into a string, rather than just repeatedly
writing a comma.
|
|
Treat touchpads -- not just mice -- as pointer devices.
|
|
Don't use our DBusError for property getting, because we simply don't care:
this fixes D-Bus error spew to stderr. Thanks Michel Dänzer for debugging
and testing.
|
|
Use an explicit input.xkb.foo namespace, not input.xkb_foo.
|
|
Convert all my license statements to the standard, accepted form:
cf. <20070717142307.GD13478@fooishbar.org>
http://lists.freedesktop.org/archives/xorg/2007-July/026451.html
keithp's license on configure.ac changed with his verbal permission.
|
|
Add support for HAL-based hotplugging, in which we just get the list of
input devices and properties from HAL. Requires an FDI which is not yet
in mainline HAL.
|