similar to: [ANNOUNCE] xf86-input-libinput 0.15.0

A couple of cleanups, no big features added. Most of this was in the property handling code. The most visible fix is that we now handle addition of initially disabled devices correctly, i.e. when a device is plugged in while the server is VT-switched away. In the non-logind case this previously caused a segfault. Our tablet support is good enough that we now match against tablets too. The

One significant change that affects packagers: the config snippet has been renamed from 90-libinput.conf to 60-libinput.conf. This should not show any immediate effects on most user's machines but there is a slight potential of user configurations to change if their custom snippets now sort after libinput instead of before. This change is part of the two-step solution to fix the handling of

xf86-input-libinput 0.24.0 is now available. A couple of new features were added this cycle: * tablets now support a configurable pressure curve * tablets now have a 'area ratio' setting to be able to match the tablet aspect ratio with that of the screen * the mouse wheel's click angle now (correctly) affects how fast the scroll wheel triggers. In particular, if you have a wheel

A few patches that have been sitting on master for a while, of which two are documentation patches. The most interesting patch is for tablets: previously we didn't send a separate motion event before the button event which caused some clients to trigger the button event on the previous stylus position. Martin Kepplinger (1): Fix config comment description to match the config Niklas

This release adds support for tablet tool devices. Now, this support should be taken with a grain of salt, it works differently to the xf86-input-wacom driver and thus some capability differences can be observed: First, pad support is still missing, i.e. you cannot use the buttons on the tablet bit itself. There are no specific configuration options just yet, and the way we are aiming for

xf86-input-libinput 1.4.0 is now available. Notable changes are: - tablets now have BTN_STYLUS3 mapped to X button 8 (i.e. the first usable one after the LMR buttons) - tablet now have the left-handed property available so we can actually change them to left-handed. Peter Hutterer (6): Change the capabilities to an enum Initialize the left-handed property for tablet tools

xf86-input-libinput 1.3.0 is now available. The main feature in this version is support for the new 'custom' pointer acceleration profile in libinput 1.23.0. This acceleration profile is quite flexible, so it is exposed via several properties: - "libinput Accel Custom Fallback Points" and "libinput Accel Custom Fallback Step" - "libinput Accel Custom Motion

A couple of cosmetic changes but one addition of a new set of properties: properties named "libinput something Default" show the libinput default for a device. For example: libinput Click Methods Available (284): 1, 1 libinput Click Method Enabled (285): 0, 1 libinput Click Method Enabled Default (286): 1, 0 This device has software buttons and clickfingers

xf86-input-libinput 0.29 is now available. This is the X.Org X server driver to use libinput for input devices. Only one noticeable change: the scroll button lock configuration option available in recent libinput versions is now exposed as the usual set of properties by this driver. Peter Hutterer (4): Implement support for scroll button locks gitlab CI: bump to Fedora 32

The libinput driver splits libinput devices into multiple X devices (e.g. a mouse and a keyboard device) and then routes the events accordingly. In the case where there is a user-specific on one device, a VT switch or suspend/resume cycle sometimes overwrote the configuration with the devices in-order, resulting in the loss of that configuration. e.g. if the keyboard device resumes after the mouse

xf86-input-libinput 0.29 is now available. This is the X.Org X server driver to use libinput for input devices. Only one real fix: we now check for the tool type as well as the serial when we create subdevices for tablet tools. Previously there were some cases where the eraser device wasn't created correctly. David Rosca (1): Also use type to match tablet tool with device Peter

An updated version of the xorg libinput driver is now available. The driver now supports server-side fds and the libinput 0.8 API. Peter Hutterer (3): Support server-side fds Fix for new libinput APIs xf86-input-libinput 0.5.0 git tag: xf86-input-libinput-0.5.0 http://xorg.freedesktop.org/archive/individual/driver/xf86-input-libinput-0.5.0.tar.bz2 MD5:

Biggest change is that the property #defines have moved to a header and we now supply a pkg-config file to get access to that header. This removes the need of string duplication in those portions that want access to the properties. Olivier Fourdan (3): libinput-drv: Add autogen.sh libinput-drv: Move properties to a separate header Move properties to a separate header Peter

xf86-input-libinput 0.16 is now available. This release fixes a long-standing issue with mixed devices that expose a pointer and a keyboard interface at the same time. X requires a strict separation between pointers and keyboards, these devices were usually initialized as pointers and exhibited some keyboard functionality issues (e.g. XKB layouts didn't get applied correctly). With this

This version adds a feature to expose the touch count for touch devices through the X device. It makes it possible for clients to adjust their UI based on the number of supported touches by this device. Also, the draglock code had a bug and would call memset with an invalid size, causing memory corruption. You're advised to update, especially if you expect someone to use the draglock options.

This release contains a bugfix that will likely trigger in future releases of libinput. The driver assumed wrongly that any wheel event has a nonzero discrete event and used the discrete as a divisor. Which is obviously a bad idea, mathematically speaking, because you never know what the future will bring. Hint: it will bring wheel events with a discrete of zero. Alan Coopersmith (2):

Only two commits, only one that really matters: previously the driver used the per-device slot number to fetch the tracking IDs from a driver-internal array. Because devices re-use slots, this could cause a tracking ID mismatch when two touchscreens had the finger down at the same time. This is now fixed by switching to libinput's seat-wide slot number instead. Peter Hutterer (3):

Two new properties, one to enable drag lock, one to disable horizontal scrolling. These properties are notable because they don't have a libinput configuration option. Drag lock is the ability to hold a button logically down when the physical button has been released. This is mainly an accessibility feature and should be handled by the compositor so that the appropriate notifications can be

xf86-input-libinput 0.8.0 is now available. Note that this version requires libinput 0.11 or later. An off-by-one error that caused the driver to crash (especially with -fstack-protector-strong) has been fixed. The excessive scroll speed on touchpads has been fixed too. A new config option/property to change the click method was added. Boyan Ding (1): Initialize variable 'path' to

Just one patch that's been sitting on master for a while now - after a proximity event we immediately post a motion event. This avoids the lines between the last known point before proximity out and the new point. Peter Hutterer (2): Post a motion event after proximity events xf86-input-libinput 0.25.1 git tag: xf86-input-libinput-0.25.1