Compare commits

..

7 commits
main ... 1.11.1

Author SHA1 Message Date
Peter Hutterer
ff5186651b libinput 1.11.1
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2018-06-19 14:32:00 +10:00
Peter Hutterer
68f124b979 tools: fix grab argument passing for libinput debug-events
The &grab pointer we used to pass as userdata was the address of the function
argument which goes out of scope at the end of the function. This works fine
for devices immediately opened but when a device connects later, the address
may have been re-used since and it's content is undefined. If not NULL, we
end up grabbing the device.

Instead pass the grab option in which is guaranteed to live until the end of
main.

https://gitlab.freedesktop.org/libinput/libinput/issues/26

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 0a13223c39)
2018-06-15 12:37:31 +10:00
Peter Hutterer
e1fe0cd5c5 touchpad: don't pair external touchpads with lid/tablet mode switches
https://gitlab.freedesktop.org/libinput/libinput/issues/29

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 46b64c7363)
2018-06-15 12:37:05 +10:00
Peter Hutterer
ba29b3a96c COPYING: change a cgit to a gitlab URL
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 66d8c79e16)
2018-06-15 12:37:01 +10:00
Daniel Stone
3320c6df90 doc: Fix URLs for GitLab, HTTPS
The GitLab migrations means that bugs should now be reported there
rather than Bugzilla. Though the repository is still available via
anongit, cloning through GitLab allows use of HTTPS.

All freedesktop.org URLs are also preferentially served over HTTPS
rather than unsecured HTTP.

Signed-off-by: Daniel Stone <daniels@collabora.com>
(cherry picked from commit 94dba68f96)
2018-06-15 12:36:58 +10:00
Peter Hutterer
168aad057d filter: cap trackpoint scale factor at 1.0
Otherwise we scale up lower-resolution trackpoints' movements, resulting in a
jumpy cursor.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 483123d490)
2018-06-15 12:36:56 +10:00
Peter Hutterer
156e53dc12 udev: the IBM X41 Tablet must not disable the keyboard in tablet mode
https://bugs.freedesktop.org/show_bug.cgi?id=106799

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 07eefba84b)
2018-06-15 12:36:52 +10:00
543 changed files with 33574 additions and 110422 deletions

View file

@ -1,82 +0,0 @@
AlignAfterOpenBracket: Align
# This option we want but it's frequently broken and causes bad
# misalignment. The canary is wheel_click_count_parser, if that works
# we can actually enable it.
# AlignArrayOfStructures: Left
AlignConsecutiveAssignments: false
AlignConsecutiveDeclarations: false
AlignConsecutiveMacros: true
AlignConsecutiveShortCaseStatements:
Enabled: true
AcrossEmptyLines: false
AcrossComments: true
AlignCaseColons: false
AlignEscapedNewlines: Right
AlignOperands: Align
AlignTrailingComments: true
AllowAllArgumentsOnNextLine: false
AllowAllParametersOfDeclarationOnNextLine: false
AllowShortBlocksOnASingleLine: false
AllowShortCaseLabelsOnASingleLine: false
AllowShortFunctionsOnASingleLine: false
AllowShortIfStatementsOnASingleLine: false
AllowShortLoopsOnASingleLine: false
AlwaysBreakBeforeMultilineStrings: true
BinPackArguments: false
BinPackParameters: OnePerLine
BraceWrapping:
AfterFunction: true
BreakAfterReturnType: All
BreakBeforeBraces: Custom
BreakStringLiterals: false
ColumnLimit: 88
ContinuationIndentWidth: 8
Cpp11BracedListStyle: false
IncludeBlocks: Regroup
IncludeCategories:
- Regex: '^(<|")config\.h(>|")'
Priority: 0
SortPriority: 0
- Regex: '^<.*'
Priority: 1
SortPriority: 0
- Regex: '^"util-.*'
Priority: 2
SortPriority: 0
- Regex: '.*'
Priority: 3
SortPriority: 0
IndentCaseLabels: false
IndentGotoLabels: false
IndentWidth: 8
MaxEmptyLinesToKeep: 1
PointerAlignment: Right
ReflowComments: true
RemoveEmptyLinesInUnwrappedLines: true
RemoveParentheses: MultipleParentheses
RemoveSemicolon: true
SkipMacroDefinitionBody: true
SortIncludes: true
SpaceAfterCStyleCast: false
SpaceAfterLogicalNot: false
SpaceBeforeAssignmentOperators: true
SpaceBeforeCaseColon: false
SpaceBeforeParens: ControlStatementsExceptControlMacros
SpacesInContainerLiterals: true
SpacesInParens: Custom
TabWidth: 8
UseTab: ForContinuationAndIndentation
ForEachMacros:
- ARRAY_FOR_EACH
- list_for_each
- list_for_each_safe
- tp_for_each_touch
- range_for_each
- litest_log_group
- litest_with_logcapture
- litest_with_parameters
- litest_with_event_frame
- udev_list_entry_foreach
# END_TEST is defined as something that enforces a line break
Macros: [ "CASE_RETURN_STRING(s)=case s: return s", "START_TEST(s)=static void s(void)", "END_TEST=enum foo;"]

View file

@ -1 +0,0 @@
include/**/*

View file

@ -1,7 +0,0 @@
# optin.core.unix.Malloc: disabled so we can use __attribute__((cleanup)) without leak complaints
# optin.core.EnumCastOutOfRange: disabled because we use a lot of "wrong" enum values for testing
# and internally and don't want those values leak into the public API
Checks: >
-clang-analyzer-unix.Malloc,
-clang-analyzer-optin.core.EnumCastOutOfRange
WarningsAsErrors: '*'

View file

@ -1 +0,0 @@
include/*

View file

@ -1,25 +0,0 @@
# https://editorconfig.org/
root = true
[*]
charset = utf-8
end_of_line = lf
trim_trailing_whitespace = true
insert_final_newline = true
[*.{c,h}]
indent_size = 8
indent_style = tab
[*.py]
indent_size = 4
indent_style = space
[{meson.build,meson_options.txt}]
indent_size = 8
indent_style = tab
[*.sym]
indent_size = 8
indent_style = tab

View file

@ -1,21 +0,0 @@
# This file contains revisions to be ignored by git blame.
# These revisions are expected to be formatting-only changes.
#
# Calling `git blame --ignore-revs-file .git-blame-ignore-revs` will
# tell git blame to ignore changes made by these revisions when assigning
# assigning blame, as if the change never happened.
#
# You can enable this as a default for your local repository by running
# `git config blame.ignoreRevsFile .git-blame-ignore-revs`
# Important: if you do this, then switch to a branch without this file,
# `git blame` will fail with an error.
#
# Run clang-format over the code
2a1095924b0be60f822bc0ff20d567e209a9db73
# Add trailing commas to prevent clang-format oddities
aebf3cd4915adc576cf055926caa621e098ec32b
# Fix some inconsistent whitespace.
60abf1575560cb8e2220ae48a67e7f74b5dc70dd

File diff suppressed because it is too large Load diff

View file

@ -1,803 +0,0 @@
# vim: set expandtab shiftwidth=2 tabstop=8 textwidth=0 filetype=yaml:
{# You're looking at the template here, so you can ignore the below
warning. This is the right file to edit #}
########################################
# #
# THIS FILE IS GENERATED, DO NOT EDIT #
# #
########################################
# To change the gitlab CI, edit .gitlab-ci/ci.template and/or .gitlab-ci/config.yaml
# and run ci-fairy generate-template. For details, see
# https://freedesktop.pages.freedesktop.org/ci-templates/ci-fairy.html#templating-gitlab-ci-yml
# This is a bit complicated for two reasons:
# - we really want to run dnf/apt/... only once, updating on the test runner for
# each job takes forever. So we create a container image for each distribution
# tested, then run the tests on this container image.
#
# This is handled by the ci-templates, ensuring containers are only rebuilt
# when the TAG changes.
#
# - GitLab only allows one script: set per job but we have a bunch of commands
# we need to re-run for each build (meson && ninja && etc). YAML cannot merge
# arrays so we're screwed.
#
# So instead we use a default_build template and override everything with
# variables. The only two variables that matter:
# MESON_ARGS=-Denable-something=true
# NINJA_ARGS=dist ... to run 'ninja -C builddir dist'
# Note that you cannot use scripts: in any target if you expect default_build
# to work.
#
#
# All jobs must follow the naming scheme of
# <distribution>:<version>@activity:
# e.g. fedora:31@build-default
.templates_sha: &template_sha c6aeb16f86e32525fa630fb99c66c4f3e62fc3cb
include:
- project: 'freedesktop/ci-templates'
ref: *template_sha
file:
- '/templates/ci-fairy.yml'
{% for distro in distributions|sort(attribute="name") %}
# {{ distro.name.capitalize() }} container builder template
- '/templates/{{distro.name}}.yml'
{% endfor %}
workflow:
rules:
# do not duplicate pipelines on merge pipelines
- if: $CI_COMMIT_BRANCH && $CI_OPEN_MERGE_REQUESTS && $CI_PIPELINE_SOURCE == "push"
when: never
# merge pipeline
- if: &is-merge-attempt $GITLAB_USER_LOGIN == "marge-bot" && $CI_PIPELINE_SOURCE == "merge_request_event"
variables:
FDO_RUNNER_JOB_PRIORITY_TAG_X86_64: priority:high
FDO_RUNNER_JOB_PRIORITY_TAG_X86_64_KVM: priority:high-kvm
FDO_RUNNER_JOB_PRIORITY_TAG_AARCH64: priority:high-aarch64
# post-merge pipeline
- if: &is-post-merge $GITLAB_USER_LOGIN == "marge-bot" && $CI_PIPELINE_SOURCE == "push"
variables:
FDO_RUNNER_JOB_PRIORITY_TAG_X86_64: priority:high
FDO_RUNNER_JOB_PRIORITY_TAG_X86_64_KVM: priority:high-kvm
FDO_RUNNER_JOB_PRIORITY_TAG_AARCH64: priority:high-aarch64
# Pre-merge pipeline
- if: &is-pre-merge $CI_PIPELINE_SOURCE == "merge_request_event"
# Push to a branch on a fork
- if: $CI_COMMIT_BRANCH
stages:
- sanity check # CI/commit checks
- prep # prep work like rebuilding the container images if there is a change
- build # for actually building and testing things in a container
- test-suite # for running the test suite in a VM
- test-suite-no-libwacom # for running the test suite in a VM (libwacom disabled)
- valgrind # for running the test suite under valgrind in a VM
- distro # distribs test
- deploy # trigger wayland's website generation
- container_clean # clean up unused container images (scheduled jobs only)
variables:
###############################################################################
# This is the list of packages required to build libinput with the default #
# configuration. #
# #
# Run dnf install/apt-get install/.. with the list of packages for your #
# distribution #
# #
# See the documentation here: #
# https://wayland.freedesktop.org/libinput/doc/latest/building.html #
###############################################################################
{% for distro in distributions %}
{{"%-17s" | format(distro.name.upper() + '_PACKAGES:')}} '{{ distro.packages|join(' ')}}'
{% endfor %}
############################ end of package lists #############################
# these tags should be updated each time the list of packages is updated
# changing these will force rebuilding the associated image
# Note: these tags have no meaning and are not tied to a particular
# libinput version
{% for distro in distributions %}
{{"%-13s"| format(distro.name.upper() + '_TAG:')}}'{{distro.tag}}'
{% endfor %}
FDO_UPSTREAM_REPO: libinput/libinput
MESON_BUILDDIR: "build dir"
NINJA_ARGS: ''
MESON_ARGS: ''
MESON_TEST_ARGS: '--no-suite=hardware'
# udev isn't available/working properly in the containers
UDEV_NOT_AVAILABLE: 1
GIT_DEPTH: 1
# Default priority for non-merge pipelines
FDO_RUNNER_JOB_PRIORITY_TAG_X86_64: "" # Empty tags are ignored by gitlab
FDO_RUNNER_JOB_PRIORITY_TAG_X86_64_KVM: kvm
FDO_RUNNER_JOB_PRIORITY_TAG_AARCH64: aarch64
.policy:
retry:
max: 2
when:
- runner_system_failure
- stuck_or_timeout_failure
# cancel run when a newer version is pushed to the branch
interruptible: true
dependencies: []
.policy-retry-on-failure:
retry:
max: 1
when:
- runner_system_failure
- stuck_or_timeout_failure
# cancel run when a newer version is pushed to the branch
interruptible: true
dependencies: []
.default_artifacts:
artifacts:
name: "meson-logs-$CI_JOB_NAME"
when: always
expire_in: 1 week
paths:
- $MESON_BUILDDIR/meson-logs
- $MESON_BUILDDIR/valgrind.*.log
reports:
junit: $MESON_BUILDDIR/*junit*.xml
.fdo-runner-tags:
tags:
- $FDO_RUNNER_JOB_PRIORITY_TAG_X86_64
#################################################################
# #
# sanity check stage #
# #
#################################################################
fail-if-fork-is-not-public:
extends:
- .fdo-runner-tags
stage: sanity check
script:
- |
if [ $CI_PROJECT_VISIBILITY != "public" ]; then
echo "*************************************************************************************"
echo "Project visibility must be set to 'public'"
echo "Change this in $CI_PROJECT_URL/edit under 'Visibility, project features, permissions'"
echo "*************************************************************************************"
exit 1
fi
except:
- main@libinput/libinput
# Re-generate the CI script and make sure it's the one currently checked in
# If this job fails, re-generate the gitlab-ci.yml script, see
# $SRCDIR/.gitlab-ci/generate-gitlab-ci.py
#
check-ci-script:
extends:
- .fdo.ci-fairy
- .fdo-runner-tags
stage: sanity check
script:
- ci-fairy generate-template --verify && exit 0 || true
- >
printf "%s\n" \
"Committed gitlab-ci.yml differs from generated gitlab-ci.yml. Please verify" \
"https://wayland.freedesktop.org/libinput/doc/latest/contributing.html"
- exit 1
#
# Verify that commit messages are as expected, etc.
#
check-commit:
extends:
- .fdo.ci-fairy
- .fdo-runner-tags
stage: sanity check
script:
- ci-fairy -vv check-commits --junit-xml=results.xml && exit 0 || true
- >
printf "%s\n" \
"Error checking commit format. Please verify" \
"https://wayland.freedesktop.org/libinput/doc/latest/contributing.html"
- exit 1
except:
- main@libinput/libinput
variables:
GIT_DEPTH: 100
artifacts:
reports:
junit: results.xml
#
# Check for trailing whitespaces
#
check-whitespace:
extends:
- .fdo.ci-fairy
- .fdo-runner-tags
stage: sanity check
script:
- .gitlab-ci/whitespace-check.py $(git ls-files)
#
# pre-commit hooks
#
pre-commit-hooks:
extends:
- .fdo.ci-fairy
- .fdo-runner-tags
stage: sanity check
script:
- python3 -m venv venv
- source venv/bin/activate
- pip3 install pre-commit
- pre-commit run --all-files
- git diff --exit-code || (echo "ERROR - Code style errors found, please fix" && false)
#################################################################
# #
# prep stage #
# #
#################################################################
{% for distro in distributions %}
{% for version in distro.versions %}
{{distro.name}}:{{version}}@container-prep:
extends:
{% if distro.qemu_based %}
- .fdo.qemu-build@{{distro.name}}
{% else %}
- .fdo.container-build@{{distro.name}}
{% endif %}
- .policy
- .fdo-runner-tags
{% if distro.qemu_based %}
tags:
- $FDO_RUNNER_JOB_PRIORITY_TAG_X86_64_KVM
{% endif %}
stage: prep
variables:
GIT_STRATEGY: none
FDO_DISTRIBUTION_VERSION: '{{version}}'
FDO_DISTRIBUTION_PACKAGES: ${{distro.name.upper()}}_PACKAGES
FDO_DISTRIBUTION_TAG: ${{distro.name.upper()}}_TAG
{% endfor %}
{% endfor %}
#################################################################
# #
# container clean stage #
# run during the clean stage #
# #
#################################################################
#
# This stage will look for the container images we currently have in
# the registry and will remove any that are not tagged with the provided
# $container_image:$tag
#
.container-clean:
extends:
- .policy
- .fdo.ci-fairy
- .fdo-runner-tags
stage: container_clean
script:
# Go to your Profile, Settings, Access Tokens
# Create a personal token with 'api' scope, copy the value.
# Go to CI/CD, Schedules, schedule a new monthly job (or edit the existing one)
# Define a variable of type File named AUTHFILE. Content is that token
# value.
- ci-fairy -v --authfile $AUTHFILE delete-image
--repository $FDO_DISTRIBUTION_NAME/$FDO_DISTRIBUTION_VERSION
--exclude-tag $FDO_DISTRIBUTION_TAG
allow_failure: true
only:
- schedules
{% for distro in distributions %}
{% for version in distro.versions %}
{{distro.name}}:{{version}}@container-clean:
extends:
- .policy
- .container-clean
variables:
GIT_STRATEGY: none
CURRENT_CONTAINER_IMAGE: $CI_REGISTRY_IMAGE/{{distro.name}}/$FDO_DISTRIBUTION_VERSION:$FDO_DISTRIBUTION_TAG
FDO_DISTRIBUTION_VERSION: '{{version}}'
FDO_DISTRIBUTION_TAG: ${{distro.name.upper()}}_TAG
{% endfor %}
{% endfor %}
#################################################################
# #
# build stage #
# #
#################################################################
.build@template:
extends:
- .policy
- .default_artifacts
- .fdo-runner-tags
stage: build
script:
- .gitlab-ci/meson-build.sh
# Run meson and meson test in the container image through qemu
.build-in-vng@template:
extends:
- .policy
- .default_artifacts
tags:
- $FDO_RUNNER_JOB_PRIORITY_TAG_X86_64_KVM
variables:
MESON_BUILDDIR: build_dir
VNG_KERNEL: {{ vng.kernel }}
script:
# first build in the host container
- .gitlab-ci/meson-build.sh --skip-test
- mkdir -p $MESON_BUILDDIR
- curl -LO $VNG_KERNEL
- export -p > .vngenv
# runs the test suite only
- |
vng --run ./bzImage \
--user root \
--overlay-rwdir=$HOME \
--append HOME=$HOME \
--overlay-rwdir=$(pwd) \
--rwdir=$MESON_BUILDDIR \
--exec "source $PWD/.vngenv; rm $PWD/.vngenv; .gitlab-ci/meson-build.sh --skip-setup --skip-build --run-test"
#
# Fedora
#
.check_tainted: &check_tainted |
# make sure the kernel is not tainted
if [[ "$(ssh localhost -p 5555 cat /proc/sys/kernel/tainted)" -gt 0 ]];
then
echo tainted kernel ;
exit 1 ;
fi
# Run meson and meson test in the qemu image
.build-in-qemu@template:
extends:
- .policy
tags:
- $FDO_RUNNER_JOB_PRIORITY_TAG_X86_64_KVM
variables:
MESON_BUILDDIR: build_dir
script:
# start our vm, no args required
- /app/vmctl start || (echo "Error - Failed to start the VM." && exit 1)
- *check_tainted
- "scp -r $PWD vm:"
- echo "CI_JOB_ID=\"$CI_JOB_ID\"" > sshenv
- echo "CI_JOB_NAME=\"$CI_JOB_NAME\"" >> sshenv
- echo "MESON_ARGS=\"$MESON_ARGS\"" >> sshenv
- echo "MESON_BUILDDIR=\"$MESON_BUILDDIR\"" >> sshenv
- echo "MESON_TEST_ARGS=\"$MESON_TEST_ARGS\"" >> sshenv
- echo "NINJA_ARGS=\"$NINJA_ARGS\"" >> sshenv
- "scp sshenv vm:~/$CI_PROJECT_NAME/.meson_environment"
- /app/vmctl exec "cd $CI_PROJECT_NAME ; .gitlab-ci/meson-build.sh" && touch .success || true
# no matter the results of the tests, we want to fetch the logs
- scp -r vm:$CI_PROJECT_NAME/$MESON_BUILDDIR .
- *check_tainted
- /app/vmctl stop
- if [[ ! -e .success ]] ;
then
exit 1 ;
fi
artifacts:
name: "qemu-meson-logs-$CI_JOB_NAME"
when: always
expire_in: 1 week
paths:
- $MESON_BUILDDIR/meson-logs
- console.out
reports:
junit: $MESON_BUILDDIR/*junit*.xml
# Run in a test suite. Special variables:
# - SUITES: the meson test suites to run, or
# - SUITE_NAMES: all elements will be expanded to libinput-test-suite-$value
# Set one or the other, not both.
.test-suite-vm:
extends:
- .build-in-vng@template
stage: test-suite
variables:
# remove the global --no-suite=hardware
MESON_TEST_ARGS: ''
LITEST_JOBS: 4
before_script:
- if ! [[ -z $SUITE_NAMES ]]; then SUITES=$(echo $SUITE_NAMES | sed 's/\([^ ]*\)/libinput-test-suite-\1/g'); fi
- echo "Testing $SUITES"
- export MESON_TEST_ARGS="$MESON_TEST_ARGS $SUITES"
{# qemu tests are only done for the latest version of any distribution #}
{% for distro in distributions if distro.use_for_qemu_tests %}
{% set version = "{}".format(distro.versions|last()) %}
.{{distro.name}}:{{version}}@test-suite-vm:
extends:
- .fdo.distribution-image@{{distro.name}}
- .test-suite-vm
variables:
FDO_DISTRIBUTION_VERSION: {{version}}
FDO_DISTRIBUTION_TAG: ${{distro.name.upper()}}_TAG
needs:
- "{{distro.name}}:{{version}}@container-prep"
{% for suite in test_suites %}
vm-{{suite.name}}:
extends:
- .{{distro.name}}:{{version}}@test-suite-vm
variables:
SUITE_NAMES: '{{suite.suites|join(' ')}}'
vm-{{suite.name}}-no-libwacom:
extends:
- vm-{{suite.name}}
stage: test-suite-no-libwacom
variables:
MESON_ARGS: '-Dlibwacom=false'
{% endfor %}
{% for suite in test_suites %}
vm-valgrind-{{suite.name}}:
stage: valgrind
extends:
- vm-{{suite.name}}
- .policy-retry-on-failure
variables:
MESON_TEST_ARGS: '--setup=valgrind'
LITEST_JOBS: 0
retry:
max: 2
rules:
- if: $GITLAB_USER_LOGIN != "marge-bot"
{% endfor %}
{% endfor %}{# for if distro.use_for_qemu_tests #}
{% for distro in distributions if distro.use_for_custom_build_tests %}
{% set version = "{}".format(distro.versions|last()) %}
.{{distro.name}}-build@template:
extends:
- .fdo.distribution-image@{{distro.name}}
- .build@template
variables:
FDO_DISTRIBUTION_VERSION: '{{version}}'
FDO_DISTRIBUTION_TAG: ${{distro.name.upper()}}_TAG
needs:
- "{{distro.name}}:{{version}}@container-prep"
default-build-release@{{distro.name}}:{{version}}:
stage: distro
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dbuildtype=release"
CFLAGS: "-Werror"
clang-tidy@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
NINJA_ARGS: ''
MESON_TEST_ARGS: ''
CC: 'clang'
script:
- .gitlab-ci/meson-build.sh
- ninja -C "$MESON_BUILDDIR" clang-tidy
# Below jobs are build option combinations. We only
# run them on one image, they shouldn't fail on one distro
# when they succeed on another.
build-no-libwacom@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dlibwacom=false"
build-no-libwacom-nodeps@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dlibwacom=false"
before_script:
- dnf remove -y libwacom libwacom-devel
build-no-mtdev@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dmtdev=false"
build-no-mtdev-nodeps@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dmtdev=false"
before_script:
- dnf remove -y mtdev mtdev-devel
build-no-lua@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dlua-plugins=disabled"
build-no-lua-nodeps@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dlua-plugins=disabled"
before_script:
- dnf remove -y lua lua-devel
build-docs@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Ddocumentation=true"
build-no-docs-nodeps@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Ddocumentation=false"
before_script:
- dnf remove -y doxygen graphviz
build-no-debuggui@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Ddebug-gui=false"
build-no-debuggui-nodeps@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Ddebug-gui=false"
before_script:
- dnf remove -y gtk3-devel gtk4-devel
build-no-tests@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dtests=false"
build-no-tests-nodeps@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_ARGS: "-Dtests=false"
before_script:
- dnf remove -y check-devel
valgrind@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
variables:
MESON_TEST_ARGS: '--suite=valgrind --no-suite=hardware --setup=valgrind'
# Python checks, only run on Fedora
usr-bin-env-python@{{distro.name}}:{{version}}:
extends:
- .{{distro.name}}-build@template
script:
- |
if git grep -l '^#!/usr/bin/python'; then
echo "Use '/usr/bin/env python3' in the above files";
/bin/false
fi
# A job to check we're actually running all test suites in the CI
check-test-suites:
extends:
- .{{distro.name}}-build@template
script:
- meson setup builddir
- meson introspect builddir --test | jq -r '.[].name' | grep 'libinput-test-suite' | sort > meson-testsuites
- |
cat <<EOF > ci-testsuites ;
{% for suite in test_suites %}
{% for name in suite.suites %}
libinput-test-suite-{{name}}
{% endfor %}
{% endfor %}
EOF
- sort -o ci-testsuites ci-testsuites
- diff -u8 -w ci-testsuites meson-testsuites || (echo "Some test suites are not run in the CI" && false)
only:
changes:
- "meson.build"
- ".gitlab-ci.yml"
{% endfor %}
#
# coverity run
#
# This requires the COVERITY_SCAN_TOKEN. Log into scan.coverity.com and get
# the token from the respective project settings page.
# Schedule a pipeline and set a variable COVERITY_SCAN_TOKEN with the token value.
# https://gitlab.freedesktop.org/$CI_PROJECT_PATH/-/pipeline_schedules
# Email from coverity will be sent to the GITLAB_USER_EMAIL that scheduled the
# job.
#
# Coverity ratelimits submissions and the coverity tools download is about
# 700M, do not run this too often.
#
coverity:
extends:
- .fdo.distribution-image@debian
- .policy
- .fdo-runner-tags
stage: build
variables:
FDO_DISTRIBUTION_VERSION: 'stable'
FDO_DISTRIBUTION_TAG: $DEBIAN_TAG
# so git-describe works, or should work
GIT_DEPTH: 200
only:
variables:
- $COVERITY_SCAN_TOKEN
script:
- curl https://scan.coverity.com/download/linux64
-o /tmp/cov-analysis-linux64.tgz
--form project=$CI_PROJECT_NAME
--form token=$COVERITY_SCAN_TOKEN
- tar xfz /tmp/cov-analysis-linux64.tgz
# coverity has special build options in meson, make sure we enable those
- meson coverity-build -Ddocumentation=false -Dcoverity=true
- cov-analysis-linux64-*/bin/cov-build --dir cov-int ninja -C coverity-build
- tar cfz cov-int.tar.gz cov-int
- curl https://scan.coverity.com/builds?project=$CI_PROJECT_NAME
--form token=$COVERITY_SCAN_TOKEN --form email=$GITLAB_USER_EMAIL
--form file=@cov-int.tar.gz --form version="$(git describe --tags)"
--form description="$(git describe --tags) / $CI_COMMIT_TITLE / $CI_COMMIT_REF_NAME:$CI_PIPELINE_ID"
artifacts:
name: "coverity-submit-data"
when: always
expire_in: 1 week
paths:
- cov-int.tar.gz
needs:
- "debian:stable@container-prep"
#################################################################
# #
# distro stage #
# #
#################################################################
{% for distro in distributions %}
{% if not distro.qemu_based %}
{% for version in distro.versions %}
{{distro.name}}:{{version}}@default-build:
stage: distro
extends:
- .build@template
- .fdo.distribution-image@{{distro.name}}
variables:
FDO_DISTRIBUTION_VERSION: '{{version}}'
FDO_DISTRIBUTION_TAG: ${{distro.name.upper()}}_TAG
{# Where we have extra_variables defined, add them to the list #}
{% if distro.build is defined and distro.build.extra_variables is defined %}
{% for var in distro.build.extra_variables %}
{{var}}
{% endfor %}
{% endif %}
needs:
- "{{distro.name}}:{{version}}@container-prep"
{% endfor %}
{% else %}
{% set version = "{}".format(distro.versions|last()) %}
{{distro.name}}:{{version}}@default-build:
stage: distro
extends:
- .build-in-qemu@template
- .fdo.distribution-image@{{distro.name}}
variables:
FDO_DISTRIBUTION_VERSION: '{{version}}'
FDO_DISTRIBUTION_TAG: ${{distro.name.upper()}}_TAG
{# Where we have extra_variables defined, add them to the list #}
{% if distro.build is defined and distro.build.extra_variables is defined %}
{% for var in distro.build.extra_variables %}
{{var}}
{% endfor %}
{% endif %}
needs:
- "{{distro.name}}:{{version}}@container-prep"
{% endif %}
{% endfor %}
#################################################################
# #
# deploy stage #
# #
#################################################################
{% for distro in distributions if distro.name == "fedora" %}
{% set version = "{}".format(distro.versions|last()) %}
build rpm:
extends:
- .fdo.distribution-image@fedora
- .policy
- .fdo-runner-tags
stage: deploy
variables:
FDO_DISTRIBUTION_VERSION: '{{version}}'
FDO_DISTRIBUTION_TAG: $FEDORA_TAG
needs:
- "fedora:{{version}}@container-prep"
script:
- meson "$MESON_BUILDDIR"
- VERSION=$(meson introspect "$MESON_BUILDDIR" --projectinfo | jq -r .version)
- sed -e "s/@PIPELINEID@/${CI_PIPELINE_ID}/"
-e "s/@GITVERSION@/${CI_COMMIT_SHA}/"
-e "s/@VERSION@/${VERSION}/" .gitlab-ci/libinput.spec.in > libinput.spec
- git config --local user.name 'gitlab CI'
- git config --local user.email 'noreply@nowhere'
- git add libinput.spec && git commit -m 'Add libinput.spec for build testing' libinput.spec
- cd "$MESON_BUILDDIR"
- meson dist --no-test
- rpmbuild -ta meson-dist/libinput*.tar.xz
{% endfor %}
wayland-web:
stage: deploy
trigger: wayland/wayland.freedesktop.org
variables:
MESON_ARGS: '-Ddocumentation=true -Ddebug-gui=false -Dlibwacom=false -Dtests=false'
MESON_BUILDDIR: 'builddir'
rules:
- if: '$CI_PIPELINE_SOURCE == "schedule"'
when: never
- if: '$CI_COMMIT_BRANCH == "main" && $GITLAB_USER_LOGIN != "marge-bot" && $CI_PROJECT_PATH == $FDO_UPSTREAM_REPO'
when: on_success
- when: never

View file

@ -1,217 +0,0 @@
# This file contains the configuration for the gitlab ci.
# See the .gitlab-ci/generate-gitlab-ci.py file for more info
#
# We're happy to rebuild all containers when one changes.
.default_tag: &default_tag '2026-01-09.0'
distributions:
- name: fedora
tag: *default_tag
versions:
- '42'
- '43'
use_for_custom_build_tests: true
use_for_qemu_tests: true
packages:
- git-core
- gcc
- gcc-c++
- pkgconf-pkg-config
- meson
- check-devel
- libudev-devel
- libevdev-devel
- doxygen
- graphviz
- python3-sphinx
- python3-recommonmark
- python3-sphinx_rtd_theme
- python3-pytest-xdist
- libwacom-devel
- cairo-devel
- gtk4-devel
- glib2-devel
- mtdev-devel
- diffutils
- wayland-protocols-devel
- black # for the Python black job, optional
- clang # for the clang-tidy build, optional
- clang-tools-extra # for clang-tidy, optional
- jq # for the test suite check job, optional
- rpmdevtools # for the rpm build job, optional
- valgrind # for the valgrind run, optional
# below packages are for the qemu runs, so optional
- systemd-udev # for the qemu run
- qemu-img
- qemu-system-x86-core
- qemu-system-aarch64-core
- jq
- python3-click
- python3-rich
- virtme-ng
- lua-devel
- name: debian
tag: *default_tag
versions:
- 'stable'
packages:
- git
- gcc
- g++
- pkg-config
- meson
- check
- libudev-dev
- libevdev-dev
- doxygen
- graphviz
- python3-sphinx
- python3-recommonmark
- python3-sphinx-rtd-theme
- python3-pytest-xdist
- libwacom-dev
- libcairo2-dev
- libgtk-3-dev
- libglib2.0-dev
- libmtdev-dev
- curl # for the coverity job
- lua5.4-dev
- name: ubuntu
tag: *default_tag
versions:
- '25.10'
packages:
- git
- gcc
- g++
- pkg-config
- meson
- check
- libudev-dev
- libevdev-dev
- doxygen
- graphviz
- python3-sphinx
- python3-recommonmark
- python3-sphinx-rtd-theme
- python3-pytest-xdist
- libwacom-dev
- libcairo2-dev
- libgtk-3-dev
- libglib2.0-dev
- libmtdev-dev
- lua5.4-dev
- name: arch
tag: *default_tag
versions:
- 'rolling'
packages:
- git
- gcc
- pkgconfig
- meson
- check
- libsystemd
- libevdev
- python-pytest-xdist
- libwacom
- gtk4
- mtdev
- diffutils
- lua
build:
extra_variables:
- "MESON_ARGS: '-Ddocumentation=false'" # python-recommonmark is no longer in the repos
- name: alpine
tag: *default_tag
versions:
- 'latest'
packages:
- git
- gcc build-base
- pkgconfig
- meson
- check-dev
- eudev-dev
- libevdev-dev
- libwacom-dev
- cairo-dev
- gtk4.0-dev
- mtdev-dev
- bash
- lua5.4-dev
build:
extra_variables:
- "MESON_ARGS: '-Ddocumentation=false' # alpine does not have python-recommonmark"
# We don't run the tests on alpine. The litest-selftest fails
# for any tcase_add_exit_test/tcase_add_test_raise_signal
# but someone more invested in musl will have to figure that out.
- "MESON_TEST_ARGS: '' # litest-selftest fails on musl"
test_suites:
- name: touchpad
suites:
- touchpad
- name: touchpad_palm
suites:
- touchpad_palm
- name: touchpad_dwt
suites:
- touchpad_dwt
- name: tap
suites:
- touchpad_tap
- name: tap-drag
suites:
- touchpad_tap_drag
- name: tap-palm
suites:
- touchpad_tap_palm
- name: touchpad-buttons
suites:
- touchpad_buttons
- name: tablet
suites:
- tablet
- name: tablet_left_handed
suites:
- tablet_left_handed
- name: tablet_proximity_tip
suites:
- tablet_proximity
- tablet_tip
- name: tablet_eraser
suites:
- tablet_eraser
- name: gestures
suites:
- gestures
- name: backends
suites:
- path
- udev
- name: misc
suites:
- log
- misc
- quirks
- device
- name: other devices
suites:
- keyboard
- pad
- switch
- trackball
- trackpoint
- totem
- touch
- name: pointer
suites:
- pointer
- name: lua
suites:
- lua
vng:
kernel: https://gitlab.freedesktop.org/api/v4/projects/libevdev%2Fhid-tools/packages/generic/kernel-x86_64/v6.14/bzImage

View file

@ -1,149 +0,0 @@
# This specfile should not be used outside the CI
# Its main purpose is to sound alarm if files disappear or are added that
# weren't intended.
%global udevdir %(pkg-config --variable=udevdir udev)
%global pipelineid @PIPELINEID@
%global gitversion @GITVERSION@
Name: libinput
Version: @VERSION@
Release: %{pipelineid}git%{gitversion}%{?dist}
Summary: Input device library
License: MIT
URL: http://www.freedesktop.org/wiki/Software/libinput/
Source0: %{name}-%{version}.tar.xz
# No BuildRequires, we rely on the container setup to have
# all the requires installed
%description
libinput is a library that handles input devices for display servers and other
applications that need to directly deal with input devices.
It provides device detection, device handling, input device event processing
and abstraction so minimize the amount of custom input code the user of
libinput need to provide the common set of functionality that users expect.
%package devel
Summary: Development files for %{name}
Requires: %{name}%{?_isa} = %{version}-%{release}
%description devel
The %{name}-devel package contains libraries and header files for
developing applications that use %{name}.
%package utils
Summary: Utilities and tools for debugging %{name}
Requires: %{name}%{?_isa} = %{version}-%{release}
Requires: python3-pyudev python3-libevdev
%description utils
The %{name}-utils package contains tools to debug hardware and analyze
%{name}.
%package test
Summary: libinput integration test suite
Requires: %{name}%{?_isa} = %{version}-%{release}
%description test
The %{name}-test package contains the libinput test suite. It is not
intended to be run by users.
%prep
%autosetup -S git -n %{name}-%{version}
%build
%meson -Dtests=true \
-Dinstall-tests=true \
-Dudev-dir=%{udevdir}
%meson_build
%install
%meson_install
%post
%{?ldconfig}
%ldconfig_postun
%files
%doc COPYING
%dir %{_sysconfdir}/libinput
%{_libdir}/libinput.so.*
%{udevdir}/libinput-device-group
%{udevdir}/libinput-fuzz-extract
%{udevdir}/libinput-fuzz-to-zero
%{udevdir}/rules.d/80-libinput-device-groups.rules
%{udevdir}/rules.d/90-libinput-fuzz-override.rules
%{_bindir}/libinput
%dir %{_libexecdir}/libinput/
%{_libexecdir}/libinput/libinput-debug-events
%{_libexecdir}/libinput/libinput-list-devices
%{_mandir}/man1/libinput.1*
%{_datadir}/libinput/*.quirks
%dir %{_datadir}/zsh
%dir %{_datadir}/zsh/site-functions
%{_datadir}/zsh/site-functions/*
%{_mandir}/man1/libinput-list-devices.1*
%{_mandir}/man1/libinput-debug-events.1*
%files devel
%{_includedir}/libinput.h
%{_libdir}/libinput.so
%{_libdir}/pkgconfig/libinput.pc
%files utils
%{_libexecdir}/libinput/libinput-debug-gui
%{_libexecdir}/libinput/libinput-debug-tablet
%{_libexecdir}/libinput/libinput-debug-tablet-pad
%{_libexecdir}/libinput/libinput-list-kernel-devices
%{_libexecdir}/libinput/libinput-measure
%{_libexecdir}/libinput/libinput-measure-fuzz
%{_libexecdir}/libinput/libinput-measure-touchpad-tap
%{_libexecdir}/libinput/libinput-measure-touchpad-pressure
%{_libexecdir}/libinput/libinput-measure-touch-size
%{_libexecdir}/libinput/libinput-measure-touchpad-size
%{_libexecdir}/libinput/libinput-quirks
%{_libexecdir}/libinput/libinput-record
%{_libexecdir}/libinput/libinput-replay
%{_libexecdir}/libinput/libinput-analyze
%{_libexecdir}/libinput/libinput-analyze-buttons
%{_libexecdir}/libinput/libinput-analyze-per-slot-delta
%{_libexecdir}/libinput/libinput-analyze-recording
%{_libexecdir}/libinput/libinput-analyze-touch-down-state
%{_mandir}/man1/libinput-debug-gui.1*
%{_mandir}/man1/libinput-debug-tablet.1*
%{_mandir}/man1/libinput-debug-tablet-pad.1*
%{_mandir}/man1/libinput-list-kernel-devices.1*
%{_mandir}/man1/libinput-measure.1*
%{_mandir}/man1/libinput-measure-fuzz.1*
%{_mandir}/man1/libinput-measure-touchpad-tap.1*
%{_mandir}/man1/libinput-measure-touch-size.1*
%{_mandir}/man1/libinput-measure-touchpad-size.1*
%{_mandir}/man1/libinput-measure-touchpad-pressure.1*
%{_mandir}/man1/libinput-quirks.1*
%{_mandir}/man1/libinput-quirks-list.1*
%{_mandir}/man1/libinput-quirks-validate.1*
%{_mandir}/man1/libinput-record.1*
%{_mandir}/man1/libinput-replay.1*
%{_mandir}/man1/libinput-analyze.1*
%{_mandir}/man1/libinput-analyze-buttons.1*
%{_mandir}/man1/libinput-analyze-per-slot-delta.1*
%{_mandir}/man1/libinput-analyze-recording.1*
%{_mandir}/man1/libinput-analyze-touch-down-state.1*
%files test
%{_libexecdir}/libinput/libinput-test
%{_libexecdir}/libinput/libinput-test-suite
%{_libexecdir}/libinput/libinput-test-utils
%{_mandir}/man1/libinput-test.1*
%{_mandir}/man1/libinput-test-suite.1*
%changelog
* Wed Jul 15 2020 Peter Hutterer <peter.hutterer@redhat.com>
- Add basic spec file for package build testing

View file

@ -1,88 +0,0 @@
#!/usr/bin/env bash
#
# This script is sourced from here:
# https://gitlab.freedesktop.org/whot/meson-helper
#
# SPDX-License-Identifier: MIT
set -x
if [[ -f .meson_environment ]]; then
. .meson_environment
fi
# If test args are set, we assume we want to run the tests
MESON_RUN_TEST="$MESON_TEST_ARGS"
while [[ $# -gt 0 ]]; do
case $1 in
--skip-setup)
shift
MESON_SKIP_SETUP="1"
;;
--skip-build)
shift
MESON_SKIP_BUILD="1"
;;
--skip-test)
shift
MESON_RUN_TEST=""
;;
--run-test)
shift
MESON_RUN_TEST="1"
;;
*)
echo "Unknow commandline argument $1"
exit 1
;;
esac
done
if [[ -z "$MESON_BUILDDIR" ]]; then
echo "\$MESON_BUILDDIR undefined."
exit 1
fi
# emulate a few gitlab variables to make it easier to
# run and debug locally.
if [[ -z "$CI_JOB_ID" ]] || [[ -z "$CI_JOB_NAME" ]]; then
echo "Missing \$CI_JOB_ID or \$CI_JOB_NAME".
CI_PROJECT_NAME=$(basename "$PWD")
CI_JOB_ID=$(date +%s)
CI_JOB_NAME="$CI_PROJECT_NAME-job-local"
echo "Simulating gitlab environment: "
echo " CI_JOB_ID=$CI_JOB_ID"
echo " CI_JOB_NAME=$CI_JOB_NAME"
fi
if [[ -n "$FDO_CI_CONCURRENT" ]]; then
jobcount="-j$FDO_CI_CONCURRENT"
export MESON_TESTTHREADS="$FDO_CI_CONCURRENT"
fi
echo "*************************************************"
echo "builddir: $MESON_BUILDDIR"
echo "meson args: $MESON_ARGS"
echo "ninja args: $NINJA_ARGS"
echo "meson test args: $MESON_TEST_ARGS"
echo "job count: ${jobcount-0}"
echo "*************************************************"
set -e
if [[ -z "$MESON_SKIP_SETUP" ]]; then
rm -rf "$MESON_BUILDDIR"
meson setup "$MESON_BUILDDIR" $MESON_ARGS
fi
meson configure "$MESON_BUILDDIR"
if [[ -z "$MESON_SKIP_BUILD" ]]; then
if [[ -n "$NINJA_ARGS" ]]; then
ninja_args="--ninja-args $NINJA_ARGS"
fi
meson compile -v -C "$MESON_BUILDDIR" $jobcount $ninja_args
fi
if [[ -n "$MESON_RUN_TEST" ]]; then
meson test -C "$MESON_BUILDDIR" $MESON_TEST_ARGS --print-errorlogs
fi

View file

@ -1,2 +0,0 @@
#!/bin/sh
scan-build -v --status-bugs -plist-html "$@"

View file

@ -1,119 +0,0 @@
#!/usr/bin/env python3
# SPDX-License-Identifier: MIT
from pathlib import Path
from dataclasses import dataclass
import argparse
import itertools
import os
import sys
@dataclass
class WhitespaceError:
message: str
lineno: int
nlines: int = 1
column: None | int = None
ncolumns: int = 1
def test_duplicate_empty_lines(lines: list[str]) -> list[WhitespaceError]:
errors = []
for idx, (l1, l2) in enumerate(itertools.pairwise(lines)):
if not l1 and not l2:
errors.append(WhitespaceError("Duplicated empty lines", idx, nlines=2))
return errors
def test_tab_after_space(lines: list[str]) -> list[WhitespaceError]:
errors = []
for idx, l in enumerate(lines):
index = l.find(" \t")
if index > -1:
errors.append(
WhitespaceError(
"Tab after space", idx, nlines=index, column=index, ncolumns=2
)
)
return errors
def test_trailing_whitespace(lines: list[str]) -> list[WhitespaceError]:
errors = []
for idx, l in enumerate(lines):
if l.rstrip() != l:
errors.append(WhitespaceError("Trailing whitespace", idx))
return errors
def test_empty_line_between_braces(lines: list[str]) -> list[WhitespaceError]:
errors = []
for idx in range(len(lines) - 3):
l1 = lines[idx]
l2 = lines[idx + 1]
l3 = lines[idx + 2]
if l1.strip() == "}" and l3.strip() == "}" and l2.strip() == "":
errors.append(WhitespaceError("Empty line between closing braces", idx + 1))
return errors
def main():
parser = argparse.ArgumentParser(description="Whitespace checker script")
parser.add_argument(
"files",
metavar="FILES",
type=Path,
nargs="+",
help="The files to check",
)
args = parser.parse_args()
have_errors: bool = False
if os.isatty(sys.stderr.fileno()):
red = "\x1b[0;31m"
reset = "\x1b[0m"
else:
red = ""
reset = ""
for file in args.files:
lines = [l.rstrip("\n") for l in file.open().readlines()]
errors = []
errors.extend(test_tab_after_space(lines))
errors.extend(test_trailing_whitespace(lines))
if any(file.name.endswith(suffix) for suffix in [".c", ".h"]):
if not file.parts[0] == "include":
errors.extend(test_duplicate_empty_lines(lines))
errors.extend(test_empty_line_between_braces(lines))
for e in errors:
print(f"{red}ERROR: {e.message} in {file}:{reset}", file=sys.stderr)
print(f"{'-' * 72}", file=sys.stderr)
lineno = max(0, e.lineno - 5)
for idx, l in enumerate(lines[lineno : lineno + 10]):
if e.lineno <= lineno + idx < e.lineno + e.nlines:
prefix = "->"
hl = red
nohl = reset
else:
prefix = " "
hl = ""
nohl = ""
print(f"{hl}{lineno + idx:3d}: {prefix} {l.rstrip()}{nohl}")
print(f"{'-' * 72}", file=sys.stderr)
if errors:
have_errors = True
if have_errors:
sys.exit(1)
if __name__ == "__main__":
main()

View file

@ -1,32 +0,0 @@
<!--
Before your file a feature request, please read
https://wayland.freedesktop.org/libinput/doc/latest/what-is-libinput.html
The amount of developer time libinput has available is very small.
Requesting a feature is no guarantee that it will get implemented. Someone
(you!) needs to step up to do the work.
-->
## Summary
<!-- Summarize the requested feature in a few sentences. -->
## Feature details
<!-- A step-by-step list of what the feature should achieve (where applicable) -->
## Affected Hardware
<!-- Which hardware types would be affected by this -->
## Implementation in Other Systems
<!-- Does this feature already exist elsewhere? How does it work there? Try
to provide as many details as possible -->
/label ~enhancement ~"needs triage"

View file

@ -1,35 +0,0 @@
## Summary
<!--
Summarize the bug encountered concisely. See
https://wayland.freedesktop.org/libinput/doc/latest/reporting-bugs.html for
detailed instructions to report bugs
-->
## Steps to reproduce
<!-- How to reproduce the issue on a developer machine - this is very important -->
## Required information
<!-- Note: if your libinput version is older than the current stable version,
please reproduce with a current version instead -->
- libinput version:
- hardware information:
- `libinput record` output: do not paste, **attach** the file
- `libinput debug-events --verbose` output: do not paste, **attach the file**
<!--
Paste any other relevant logs - please use code blocks (```) to format
console output, logs, and code as it's very hard to read otherwise.)
Do not paste logs longer than 10 lines, **attach** those instead.
If your libinput record is longer than 5-10s, we will not be able to process
it.
-->
/label ~"bug" ~"needs triage"

View file

@ -1,31 +0,0 @@
repos:
- repo: https://github.com/pre-commit/pre-commit-hooks
rev: v5.0.0
hooks:
- id: end-of-file-fixer
- id: trailing-whitespace
- id: no-commit-to-branch
args: ['--branch', 'main']
- repo: https://github.com/astral-sh/ruff-pre-commit
rev: v0.12.2
hooks:
- id: ruff-check
args: ['--ignore=E741,E501', '--extend-exclude=subprojects', '.']
- id: ruff-format
args: ['--check', '--diff']
- repo: https://gitlab.freedesktop.org/freedesktop/ci-templates.git
rev: e195d80f35b45cc73668be3767b923fd76c70ed5
hooks:
- id: check-commits
- id: generate-template
- repo: local
hooks:
- id: run-sed-script
name: Check for whitespace errors
entry: ./.gitlab-ci/whitespace-check.py
language: system
- repo: https://github.com/pre-commit/mirrors-clang-format
rev: v20.1.6
hooks:
- id: clang-format
types_or: [c]

View file

@ -1,244 +0,0 @@
# This is a set of bugbot commands for issues and merge requests - setting any of the
# bugbot::foo labels will trigger gitlab-triage to run with this ruleset (well, the
# one we have on the main branch at the time)
#
# Note that for adding labels, the label must first created in the project.
# Re-used in issues and mrs
.close_needinfo: &close_needinfo
name: "Close stale needinfo bugs"
conditions:
labels:
- "bugbot::needinfo-timeout"
actions:
remove_labels:
- "bugbot::needinfo-timeout"
comment: |
I'm closing this bug because some information we requested a while ago was never supplied and
we're not able to continue without this information.
Please feel free to re-open.
status: "close"
.remind_needinfo: &remind_needinfo
name: "Remind users of needinfo bugs"
conditions:
labels:
- "bugbot::needinfo-reminder"
actions:
labels:
- "waiting on reporter"
remove_labels:
- "bugbot::needinfo-reminder"
comment: |
Hi. This is a friendly reminder that the maintainers are waiting on some information by
you (or maybe someone cc'd on this bug). If the information is not provided we may not
be able to proceed with this issue or merge request. Please check the recent comments, thanks.
.help_needed: &help_needed
name: "Remind users help is needed"
conditions:
labels:
- "bugbot::help-needed"
actions:
labels:
- "help needed"
remove_labels:
- "bugbot::help-needed"
comment: |
Hi. This issue or merge request needs help. This simply means that for
the foreseeable future, the maintainers will not have time to work on
this.
If this is a request for a new feature, then the feature is unlikely to
be implemented unless you or another contributor files a merge request.
If a merge request already exists maybe it needs finishing which often
involves adding documentation or tests.
If this is an issue affecting a specific device then it is unlikely to be
fixed. This may be because it requires specific hardware to reproduce or
it may be that the use case is niche enough that the maintainers do not
have time to implement it.
In short, to resolve this issue or get this merge request into libinput
help is needed.
.libinput_record: &libinput_record
name: "Request libinput record output"
conditions:
labels:
- "bugbot::libinput-record"
actions:
remove_labels:
- "bugbot::libinput-record"
comment: |
Looks like we may need some extra information. Please **attach** (do not paste) the full output
of `libinput record` and `libinput debug-events --verbose` (if you haven't yet).
The [documentation](https://wayland.freedesktop.org/libinput/doc/latest/tools.html#libinput-record-and-libinput-replay)
has some information on what we're looking for to be able to triage bugs.
.hid_recorder: &hid_recorder
name: "Request hid-recorder output"
conditions:
labels:
- "bugbot::hid-recorder"
actions:
remove_labels:
- "bugbot::hid-recorder"
comment: |
Looks like we may need some extra information that isn't yet available in this issue.
Please **attach** (do not paste) the output of [`hid-recorder`](https://github.com/hidutils/hid-recorder/)
for this device (run `sudo hid-recorder` without argument and it will let you pick the device).
This should show the data the kernel receives from the device and may provide a hint on what's going on here.
.udev_hid_bpf: &udev_hid_bpf
name: "Punt to udev-hid-bpf"
conditions:
labels:
- "bugbot::udev-hid-bpf"
actions:
remove_labels:
- "bugbot::udev-hid-bpf"
comment: |
This issue looks like it could or should be fixed with [udev-hid-bpf](https://gitlab.freedesktop.org/libevdev/udev-hid-bpf/).
udev-hid-bpf is a collection of BPF programs that modify the HID Report Descriptor and/or HID Reports from the device,
making it possible to change the properties of a device and/or the events to make them compatible with the
expectations the kernel and userspace has of such devices.
Please see the [udev-hid-bpf documentation](https://libevdev.pages.freedesktop.org/udev-hid-bpf/) for details
and/or in particular the [udev-hid-bpf tutorial](https://libevdev.pages.freedesktop.org/udev-hid-bpf/tutorial.html)
if you need to enable a new device.
status: "close"
resource_rules:
issues:
rules:
- name: "Close kernel bugs"
conditions:
labels:
- "bugbot::kernel"
actions:
labels:
- "kernel"
remove_labels:
- "bugbot::kernel"
comment: |
This bug looks like a kernel issue and it cannot be fixed
in libinput directly. I'm closing this bug but do feel free
to continue discussing the issue here.
Kernel bugs are usually best sent to the [`linux-input` list](https://lore.kernel.org/linux-input/).
status: "close"
- name: "Expect a merge request"
conditions:
labels:
- "bugbot::expect-mr"
actions:
remove_labels:
- "bugbot::expect-mr"
comment: |
I'm closing this bug in anticipation of a merge request that fixes this issue.
If you are a new contributor, please see
[the freedesktop.org wiki](https://gitlab.freedesktop.org/freedesktop/freedesktop/-/wikis/home)
on how to get permissions to fork a project and file a merge request.
The [libinput documentation](https://wayland.freedesktop.org/libinput/doc/latest/contributing.html)
also has more details on how to get started.
status: "close"
- name: "Point to 60-evdev.hwdb"
conditions:
labels:
- "bugbot::evdev-hwdb"
actions:
remove_labels:
- "bugbot::evdev-hwdb"
comment: |
Looks like this issue may be solved with a device-specific entry in systemd's hwdb.
You should have a /usr/lib/udev/hwdb.d/60-evdev.hwdb file which includes those quirks. Please see
the top of the file for instructions and follow those. Once the quirk is confirmed working
this issue (or parts thereof) should be fixed and you can submit a pull request to the
[systemd](https://github.com/systemd/systemd/) repository to get that quirk included.
Please link to the systemd issue here, thanks.
I'm closing this issue now, if the hwdb entry does not fix this issue here, please re-open.
status: "close"
- name: "Close bug for reopening"
conditions:
labels:
- "bugbot::close"
actions:
remove_labels:
- "bugbot::close"
comment: |
I'm temporarily closing this bug. The bug may not be fixed yet (see any comments above)
and we **want you to reopen it when/if it becomes actionable again**.
This process may feel unfamiliar but unfortunately closing/re-opening is the only action
all GitLab users are permitted to do. So we close it, you re-open it when whatever
above has been addressed and then we know we need to look at it again.
This issue may be closed more than once in a similar fashion but I only leave this
comment once since now you understand how it works. :smile:
For a detailed explanation on the how and why of this process please see
the [Closed Issues wiki page](https://gitlab.freedesktop.org/libinput/libinput/-/wikis/Closed-Issues).
status: "close"
- name: "Re-close bug for reopening"
conditions:
labels:
- "bugbot::re-close"
actions:
remove_labels:
- "bugbot::re-close"
comment: |
I'm temporarily closing this bug again. This is not a final close, see my comments above for the open/close process.
status: "close"
- *udev_hid_bpf
- *libinput_record
- *hid_recorder
- *close_needinfo
- *remind_needinfo
- *help_needed
merge_requests:
rules:
- name: "Remind contributor of commit rules"
conditions:
labels:
- "bugbot::commit-rules"
actions:
remove_labels:
- "bugbot::commit-rules"
comment: |
Hi. Looks like the pipeline failed because one or more of the commits in this MR do not meet our requirements.
Most commonly this the format of the commit message itself. The "Test summary" above has the details.
Please see [our docs for commit messages](https://wayland.freedesktop.org/libinput/doc/latest/contributing.html#commit-messages)
and [our docs for submitting code](https://wayland.freedesktop.org/libinput/doc/latest/contributing.html#submitting-code)
that explain how to amend and force-push to this repo.
- name: "Remind contributor that info needs to be in commit messages"
conditions:
labels:
- "bugbot::info-in-commit-message"
actions:
remove_labels:
- "bugbot::info-in-commit-message"
comment: |
Hi. Thanks for the merge request. I'm here to request that you add
some documentation about this merge request to the
**commit message** (or messages). You may have already written some
in the merge request description and in many cases it's fine to
copy/paste that into the commit message(s).
The reason is simple: once merged, no-one really looks at this this
page here anymore. The git log on the other hand is what developers
will use to understand the code so the information must be quickly
accessible via git.
Please see [our docs for submitting code](https://wayland.freedesktop.org/libinput/doc/latest/contributing.html#submitting-code)
that explain how to amend and force-push to this repo.
- *udev_hid_bpf
- *libinput_record
- *hid_recorder
- *close_needinfo
- *remind_needinfo
- *help_needed

117
CODING_STYLE Normal file
View file

@ -0,0 +1,117 @@
- Indentation in tabs, 8 characters wide, spaces after the tabs where
vertical alignment is required (see below)
- Max line width 80ch, do not break up printed strings though
- Break up long lines at logical groupings, one line for each logical group
int a = somelongname() +
someotherlongname();
if (a < 0 &&
(b > 20 & d < 10) &&
d != 0.0)
somelongfunctioncall(arg1,
arg2,
arg3);
- Function declarations: return type on separate line, {} on separate line,
arguments broken up as above.
static inline int
foobar(int a, int b)
{
}
void
somenamethatiswaytoolong(int a,
int b,
int c)
{
}
- /* comments only */, no // comments
- variable_name, not VariableName or variableName. same for functions.
- no typedefs of structs, enums, unions
- if it generates a compiler warning, it needs to be fixed
- if it generates a static checker warning, it needs to be fixed or
commented
- declare variables at the top, try to keep them as local as possible.
Exception: if the same variable is re-used in multiple blocks, declare it
at the top.
Exception: basic loop variables, e.g. for (int i = 0; ...)
int a;
int c;
if (foo) {
int b;
c = get_value();
usevalue(c);
}
if (bar) {
c = get_value();
useit(c);
}
- do not mix function invocations and variable definitions.
wrong:
{
int a = foo();
int b = 7;
}
right:
{
int a;
int b = 7;
a = foo();
}
There are exceptions here, e.g. tp_libinput_context(),
litest_current_device()
- if/else: { on the same line, no curly braces if both blocks are a single
statement. If either if or else block are multiple statements, both must
have curly braces.
if (foo) {
blah();
bar();
} else {
a = 10;
}
- public functions MUST be doxygen-commented, use doxygen's @foo rather than
\foo notation
- include "config.h" comes first, followed by system headers, followed by
external library headers, followed by internal headers.
sort alphabetically where it makes sense (specifically system headers)
#include "config.h"
#include <stdio.h>
#include <string.h>
#include <libevdev/libevdev.h>
#include "libinput-private.h"
- goto jumps only to the end of the function, and only for good reasons
(usually cleanup). goto never jumps backwards
- Use stdbool.h's bool for booleans within the library (instead of 'int').
Exception: the public API uses int, not bool.

View file

@ -1,244 +0,0 @@
# Coding style
- Indentation in tabs, 8 characters wide, spaces after the tabs where
vertical alignment is required (see below)
**Note: this file uses spaces due to markdown rendering issues for tabs.
Code must be implemented using tabs.**
- Max line width 80ch, do not break up printed strings though
- Break up long lines at logical groupings, one line for each logical group
```c
int a = somelongname() +
someotherlongname();
if (a < 0 &&
(b > 20 & d < 10) &&
d != 0.0)
somelongfunctioncall(arg1,
arg2,
arg3);
```
- Function declarations: return type on separate line, {} on separate line,
arguments broken up as above.
```c
static inline int
foobar(int a, int b)
{
}
void
somenamethatiswaytoolong(int a,
int b,
int c)
{
}
```
- `/* comments only */`, no `// comments`
- `variable_name`, not `VariableName` or `variableName`. same for functions.
- no typedefs of structs, enums, unions
- if it generates a compiler warning, it needs to be fixed
- if it generates a static checker warning, it needs to be fixed or
commented
- declare variables when they are used first and try to keep them as local as possible.
Exception: basic loop variables, e.g. for (int i = 0; ...) should always be
declared inside the loop even where multiple loops exist
```c
int a;
if (foo) {
int b = 10;
a = get_value();
usevalue(a, b);
}
if (bar) {
a = get_value();
useit(a);
}
int c = a * 100;
useit(c);
```
- avoid uninitialized variables where possible, declare them late instead.
Note that most of libinput predates this style, try to stick with the code
around you if in doubt.
wrong:
```c
int *a;
int b = 7;
... some code ...
a = zalloc(32);
```
right:
```c
int b = 7;
... some code ...
int *a = zalloc(32);
```
- avoid calling non-obvious functions inside declaration blocks for multiple
variables.
bad:
```c
{
int a = 7;
int b = some_complicated_function();
int *c = zalloc(32);
}
```
better:
```c
{
int a = 7;
int *c = zalloc(32);
int b = some_complicated_function();
}
```
There is a bit of gut-feeling involved with this, but the goal is to make
the variable values immediately recognizable.
- Where statements are near-identical and repeated, try to keep them
identical:
bad:
```c
int a = get_some_value(x++);
do_something(a);
a = get_some_value(x++);
do_something(a);
a = get_some_value(x++);
do_something(a);
```
better:
```c
int a;
a = = get_some_value(x++);
do_something(a);
a = get_some_value(x++);
do_something(a);
a = get_some_value(x++);
do_something(a);
```
- if/else: { on the same line, no curly braces if both blocks are a single
statement. If either if or else block are multiple statements, both must
have curly braces.
```c
if (foo) {
blah();
bar();
} else {
a = 10;
}
```
- public functions MUST be doxygen-commented, use doxygen's `@foo` rather than
`\foo` notation
- `#include "config.h"` comes first, followed by system headers, followed by
external library headers, followed by internal headers.
sort alphabetically where it makes sense (specifically system headers)
```c
#include "config.h"
#include <stdio.h>
#include <string.h>
#include <libevdev/libevdev.h>
#include "libinput-private.h"
```
- goto jumps only to the end of the function, and only for good reasons
(usually cleanup). goto never jumps backwards
- Use stdbool.h's bool for booleans within the library (instead of `int`).
Exception: the public API uses int, not bool.
# Git commit message requirements
Our CI will check the commit messages for a few requirements. Below is the
list of what we expect from a git commit.
## Commit message content
A [good commit message](http://who-t.blogspot.com/2009/12/on-commit-messages.html) needs to
answer three questions:
- Why is it necessary? It may fix a bug, it may add a feature, it may
improve performance, reliabilty, stability, or just be a change for the
sake of correctness.
- How does it address the issue? For short obvious patches this part can be
omitted, but it should be a high level description of what the approach
was.
- What effects does the patch have? (In addition to the obvious ones, this
may include benchmarks, side effects, etc.)
These three questions establish the context for the actual code changes, put
reviewers and others into the frame of mind to look at the diff and check if
the approach chosen was correct. A good commit message also helps
maintainers to decide if a given patch is suitable for stable branches or
inclusion in a distribution.
## Commit message format
The canonical git commit message format is:
```
one line as the subject line with a high-level note
full explanation of the patch follows after an empty line. This explanation
can be multiple paragraphs and is largely free-form. Markdown is not
supported.
You can include extra data where required like:
- benchmark one says 10s
- benchmark two says 12s
```
The subject line is the first thing everyone sees about this commit, so make
sure it's on point.
## Commit message technical requirements
- The commit message should use present tense (not past tense). Do write
"change foo to bar", not "changed foo to bar".
- The text width of the commit should be 78 chars or less, especially the
subject line.
- The author must be the name you usually identify as and email address. We do
not accept the default `@users.noreply` gitlab addresses.
```
git config --global user.name Your Name
git config --global user.email your@email
```

View file

@ -1,4 +0,0 @@
Thank you for your interest in contributing to libinput.
Please find more information about how to contribute in
[the documentation](https://wayland.freedesktop.org/libinput/doc/latest/contributing.html).

View file

@ -31,4 +31,4 @@ This copy is provided to provide consistent behavior regardless which kernel
version libinput is compiled against. The header is used during compilation
only, libinput does not link against GPL libraries.
[1] https://gitlab.freedesktop.org/libinput/libinput/blob/main/include/linux/input.h
[1] https://gitlab.freedesktop.org/libinput/libinput/blob/master/include/linux/input.h

View file

@ -5,22 +5,36 @@ libinput is a library that provides a full input stack for display servers
and other applications that need to handle input devices provided by the
kernel.
libinput provides device detection, event handling and abstraction to
minimize the amount of custom input code the user of libinput needs to
libinput provides device detection, event handling and abstraction so
minimize the amount of custom input code the user of libinput need to
provide the common set of functionality that users expect. Input event
processing includes scaling touch coordinates, generating
relative pointer events from touchpads, pointer acceleration, etc.
User documentation
------------------
Architecture
------------
Documentation explaining features available in libinput is available
[here](https://wayland.freedesktop.org/libinput/doc/latest/features.html).
libinput is not used directly by applications. Think of it more as a device
driver than an application library. It is used by the xf86-input-libinput
X.Org driver or Wayland compositors. The typical software stack for a system
running Wayland is:
This includes the [FAQ](https://wayland.freedesktop.org/libinput/doc/latest/faqs.html)
and the instructions on
[reporting bugs](https://wayland.freedesktop.org/libinput/doc/latest/reporting-bugs.html).
@dotfile libinput-stack-wayland.gv
The Wayland compositor may be Weston, mutter, KWin, etc. Note that
Wayland encourages the use of toolkits, so the Wayland client (your
application) does not usually talk directly to the compositor but rather
employs a toolkit (e.g. GTK) to do so. The Wayland client does not know
whether libinput is in use.
The simplified software stack for a system running X.Org is:
@dotfile libinput-stack-xorg.gv
libinput is not employed directly by the X server but by the
xf86-input-libinput driver instead. That driver is loaded by the server
on demand, depending on the xorg.conf.d configuration snippets. The X client
does not know whether libinput is in use.
Source code
-----------
@ -32,7 +46,7 @@ For a list of current and past releases visit:
https://www.freedesktop.org/wiki/Software/libinput/
Build instructions:
https://wayland.freedesktop.org/libinput/doc/latest/building.html
https://wayland.freedesktop.org/libinput/doc/latest/building_libinput.html
Reporting Bugs
--------------
@ -43,25 +57,24 @@ https://gitlab.freedesktop.org/libinput/libinput/issues/
Where possible, please provide the `libinput record` output
of the input device and/or the event sequence in question.
See https://wayland.freedesktop.org/libinput/doc/latest/reporting-bugs.html
for more info.
See @ref reporting_bugs for more info.
Documentation
-------------
- Developer API documentation: https://wayland.freedesktop.org/libinput/doc/latest/development.html
- Developer API documentation: https://wayland.freedesktop.org/libinput/doc/latest/modules.html
- High-level documentation about libinput's features:
https://wayland.freedesktop.org/libinput/doc/latest/features.html
https://wayland.freedesktop.org/libinput/doc/latest/pages.html
- Build instructions:
https://wayland.freedesktop.org/libinput/doc/latest/building.html
https://wayland.freedesktop.org/libinput/doc/latest/building_libinput.html
- Documentation for previous versions of libinput: https://wayland.freedesktop.org/libinput/doc/
Examples of how to use libinput are the debugging tools in the libinput
repository. Developers are encouraged to look at those tools for a
real-world (yet simple) example on how to use libinput.
- A commandline debugging tool: https://gitlab.freedesktop.org/libinput/libinput/tree/main/tools/libinput-debug-events.c
- A GTK application that draws cursor/touch/tablet positions: https://gitlab.freedesktop.org/libinput/libinput/tree/main/tools/libinput-debug-gui.c
- A commandline debugging tool: https://gitlab.freedesktop.org/libinput/libinput/tree/master/tools/libinput-debug-events.c
- A GTK application that draws cursor/touch/tablet positions: https://gitlab.freedesktop.org/libinput/libinput/tree/master/tools/libinput-debug-gui.c
License
-------
@ -75,10 +88,5 @@ libinput is licensed under the MIT license.
> and/or sell copies of the Software, and to permit persons to whom the
> Software is furnished to do so, subject to the following conditions: [...]
See the [COPYING](https://gitlab.freedesktop.org/libinput/libinput/tree/main/COPYING)
See the [COPYING](https://gitlab.freedesktop.org/libinput/libinput/tree/master/COPYING)
file for the full license information.
About
-----
Documentation generated from git commit [__GIT_VERSION__](https://gitlab.freedesktop.org/libinput/libinput/commit/__GIT_VERSION__)

144
circle.yml Normal file
View file

@ -0,0 +1,144 @@
# vim: set expandtab shiftwidth=2 tabstop=8 textwidth=0:
libinput_jobs:
default_settings: &default_settings
working_directory: ~/libinput
# Define a couple of standardized build targets
# that we call from every build container
build_and_test_default: &build_and_test_default
name: Build
command: |
rm -rf build
meson build ${MESON_PARAMS}
meson configure build
ninja -v -C build ${NINJA_ARGS}
environment:
MESON_PARAMS: --prefix=/opt/libinput
build_and_test: &build_and_test
run:
<<: *build_and_test_default
build_no_libwacom: &build_no_libwacom
run:
<<: *build_and_test_default
name: Build - No libwacom
environment:
MESON_PARAMS: -Dlibwacom=false
build_no_debug_gui: &build_no_debug_gui
run:
<<: *build_and_test_default
name: Build - No debug-gui
environment:
MESON_PARAMS: -Ddebug-gui=false
build_no_tests: &build_no_tests
run:
<<: *build_and_test_default
name: Build - No tests
environment:
MESON_PARAMS: -Dtests=false
build_no_docs: &build_no_docs
run:
<<: *build_and_test_default
name: Build - No docs
environment:
MESON_PARAMS: -Ddocumentation=false
build_dist: &build_dist
run:
<<: *build_and_test_default
name: Build - ninja dist
environment:
NINJA_ARGS: dist
ninja_scan_build: &ninja_scan_build
run:
<<: *build_and_test_default
name: Ninja scan-build
environment:
NINJA_ARGS: scan-build
install: &install
run:
name: Installing
command: ninja -v -C build install
export_logs: &export_logs
store_artifacts:
path: ~/libinput/build/meson-logs
fedora_install: &fedora_install
run:
name: Install prerequisites
command: |
dnf upgrade -y libsolv
dnf install -y git gcc gcc-c++ meson check-devel libudev-devel libevdev-devel doxygen graphviz valgrind binutils libwacom-devel cairo-devel gtk3-devel glib2-devel mtdev-devel
fedora_build_all: &fedora_build_all
<<: *default_settings
steps:
- *fedora_install
- checkout
- *build_and_test
- *install
- *export_logs
- *build_no_libwacom
- *build_no_debug_gui
- *build_no_tests
- *build_no_docs
- *build_dist
ubuntu_install: &ubuntu_install
run:
name: Install prerequisites
command: |
apt-get update
apt-get install -y software-properties-common
add-apt-repository universe
apt-get update
apt-get install -y git gcc g++ meson check libudev-dev libevdev-dev doxygen graphviz valgrind binutils libwacom-dev libcairo2-dev libgtk-3-dev libglib2.0-dev libmtdev-dev
ubuntu_build_all: &ubuntu_build_all
<<: *default_settings
steps:
- *ubuntu_install
- checkout
- *build_and_test
- *install
- *export_logs
- *build_no_libwacom
- *build_no_debug_gui
- *build_no_tests
- *build_no_docs
- *build_dist
scan_build_run: &scan_build_run
<<: *default_settings
steps:
- *fedora_install
- run:
name: Install clang and find
command: dnf install -y clang-analyzer findutils
- checkout
- *ninja_scan_build
- *export_logs
- run:
name: Check scan-build results
command: test ! -d ~/libinput/build/meson-logs/scanbuild || test $(find ~/libinput/build/meson-logs/scanbuild -maxdepth 0 ! -empty -exec echo "not empty" \; | wc -l) -eq 0 || (echo "Check scan-build results" && false)
version: 2
jobs:
fedora_latest:
<<: *fedora_build_all
docker:
- image: fedora:latest
ubuntu_17_10:
<<: *ubuntu_build_all
docker:
- image: ubuntu:artful
scan_build:
<<: *scan_build_run
docker:
- image: fedora:latest
workflows:
version: 2
compile:
jobs:
- fedora_latest
- scan_build
- ubuntu_17_10

View file

@ -1,282 +0,0 @@
#compdef libinput
(( $+functions[_libinput_commands] )) || _libinput_commands()
{
local -a commands
commands=(
"list-devices:List all devices recognized by libinput"
"debug-events:Print all events as seen by libinput"
"debug-gui:Show a GUI to visualize libinput's events"
"debug-tablet:Show tablet axis and button values"
"measure:Measure various properties of devices"
"analyze:Analyze device data"
"record:Record the events from a device"
"replay:Replay the events from a device"
)
_describe -t commands 'command' commands
}
__all_seats()
{
# Obviously only works with logind
local -a seats
seats=${(f)"$(loginctl --no-legend --no-pager list-seats 2>/dev/null)"}
if [[ -z $seats ]]; then
# Can always offer seat0, even if we can't enumerate the seats
compadd "$@" - seat0
else
compadd "$@" - $seats
fi
}
(( $+functions[_libinput_list-devices] )) || _libinput_list-devices()
{
_arguments \
'--help[Show help and exit]' \
'--version[show version information and exit]'
}
(( $+functions[_libinput_debug-events] )) || _libinput_debug-events()
{
_arguments \
'--help[Show debug-events help and exit]' \
'--quiet[Only print libinput messages and nothing from this tool]' \
'--verbose[Use verbose output]' \
'--show-keycodes[Make all keycodes visible]' \
'--grab[Exclusively grab all opened devices]' \
'--compress-motion-events[Compress repeated motion events on a TTY]' \
'--device=[Use the given device with the path backend]:device:_files -W /dev/input/ -P /dev/input/' \
'--udev=[Listen for notifications on the given seat]:seat:__all_seats' \
'--apply-to=[Apply configuration options where the device name matches the pattern]:pattern' \
'--disable-sendevents=[Disable send-events option for the devices matching the pattern]:pattern' \
'--set-area=[Set the desired area as "x1/y1 x2/y2" (within \[0.0, 1.0\]) ]' \
'--set-calibration=[Set the first 6 elements of the 3x3 calibration matrix ("1.0 0.0 0.0 0.0 1.0 0.0")]' \
'--set-click-method=[Set the desired click method]:click-method:(none clickfinger buttonareas)' \
'--set-clickfinger-map=[Set button mapping for clickfinger]:tap-map:(( \
lrm\:2-fingers\ right-click\ /\ 3-fingers\ middle-click \
lmr\:2-fingers\ middle-click\ /\ 3-fingers\ right-click \
))' \
'--set-eraser-button-button=[Set button mapping for the eraser button]:eraser-button:(BTN_STYLUS BTN_STYLUS2 BTN_STYLUS3)' \
'--set-eraser-button-mode=[Set the eraser button mode]:eraser-mode:(default button)' \
'--set-pressure-range=[Set the tablet tool pressure range (within range \[0.0, 1.0\])]' \
'--set-profile=[Set pointer acceleration profile]:accel-profile:(adaptive flat custom)' \
'--set-rotation-angle=[Set the rotation angle in degrees]' \
'--set-scroll-button=[Set the button to the given button code]' \
'--set-scroll-method=[Set the desired scroll method]:scroll-method:(none twofinger edge button)' \
'--set-speed=[Set pointer acceleration speed (within range \[-1, 1\])]' \
'--set-tap-map=[Set button mapping for tapping]:tap-map:(( \
lrm\:2-fingers\ right-click\ /\ 3-fingers\ middle-click \
lmr\:2-fingers\ middle-click\ /\ 3-fingers\ right-click \
))' \
+ '(custom pointer acceleration)' \
'--set-custom-points=[Set n points defining a custom acceleration function]' \
'--set-custom-step=[Set the distance along the x axis between the custom points]' \
'--set-custom-type=[Set the type of the acceleration function]:custom-type:(fallback motion scroll)' \
+ '(drag)' \
'--enable-drag[Enable tap-and-drag]' \
'--disable-drag[Disable tap-and-drag]' \
+ '(drag-lock)' \
'--enable-drag-lock[Enable drag-lock]' \
'--disable-drag-lock[Disable drag-lock]' \
+ '(dwt)' \
'--enable-dwt[Enable disable-while-typing]' \
'--disable-dwt[Disable disable-while-typing]' \
+ '(dwtp)' \
'--enable-dwtp[Enable disable-while-trackpointing]' \
'--disable-dwtp[Disable disable-while-trackpointing]' \
+ '(left-handed)' \
'--enable-left-handed[Enable left handed button configuration]' \
'--disable-left-handed[Disable left handed button configuration]' \
+ '(middlebutton)' \
'--enable-middlebutton[Enable middle button emulation]' \
'--disable-middlebutton[Disable middle button emulation]' \
+ '(natural-scrolling)' \
'--enable-natural-scrolling[Enable natural scrolling]' \
'--disable-natural-scrolling[Disable natural scrolling]' \
+ '(plugins)' \
'--enable-plugins[Enable plugins]' \
'--disable-plugins[Disable plugins]' \
+ '(tap-to-click)' \
'--enable-tap[Enable tap-to-click]' \
'--disable-tap[Disable tap-to-click]'
}
(( $+functions[_libinput_debug-gui] )) || _libinput_debug-gui()
{
_arguments \
'--help[Show debug-gui help and exit]' \
'--verbose[Use verbose output]' \
'--grab[Exclusively grab all opened devices]' \
'--device=[Use the given device with the path backend]:device:_files -W /dev/input/ -P /dev/input/' \
'--udev=[Listen for notifications on the given seat]:seat:_libinput_all_seats'
}
(( $+functions[_libinput_debug-tablet] )) || _libinput_debug-tablet()
{
_arguments \
'--help[Show debug-tablet help and exit]' \
'--device=[Use the given device with the path backend]:device:_files -W /dev/input/ -P /dev/input/' \
'--udev=[Use the first tablet device on the given seat]:seat:_libinput_all_seats'
}
(( $+functions[_libinput_measure] )) || _libinput_measure()
{
local curcontext=$curcontext state line ret=1
local features
features=(
"fuzz:Measure touch fuzz to avoid pointer jitter"
"touch-size:Measure touch size and orientation"
"touchpad-tap:Measure tap-to-click time"
"touchpad-pressure:Measure touch pressure"
)
_arguments -C \
'--help[Print help and exit]' \
':feature:->feature' \
'*:: :->option-or-argument'
case $state in
(feature)
_describe -t features 'feature' features
;;
(option-or-argument)
curcontext=${curcontext%:*:*}:libinput-measure-$words[1]:
if ! _call_function ret _libinput_measure_$words[1]; then
_message "unknown feature: $words[1]"
fi
;;
esac
return ret
}
(( $+functions[_libinput_measure_fuzz] )) || _libinput_measure_fuzz()
{
_arguments \
'--help[Show help message and exit]' \
':device:_files -W /dev/input/ -P /dev/input/'
}
(( $+functions[_libinput_measure_touch-size] )) || _libinput_measure_touch-size()
{
_arguments \
'--help[Show help message and exit]' \
'--touch-threshold=[Assume a touch pressure threshold of "down:up"]' \
'--palm-threshold=[Assume a palm threshold of N]' \
':device:_files -W /dev/input/ -P /dev/input/'
}
(( $+functions[_libinput_measure_touchpad-pressure] )) || _libinput_measure_touchpad-pressure()
{
_arguments \
'--help[Show help message and exit]' \
'--touch-threshold=[Assume a touch pressure threshold of "down:up"]' \
'--palm-threshold=[Assume a palm threshold of N]' \
':device:_files -W /dev/input/ -P /dev/input/'
}
(( $+functions[_libinput_measure_touchpad-tap] )) || _libinput_measure_touchpad-tap()
{
_arguments \
'--help[Show help message and exit]' \
'--format=dat[Specify the data format to be printed. The default is "summary"]' \
':device:_files -W /dev/input/ -P /dev/input/'
}
(( $+functions[_libinput_analyze_per-slot-delta] )) || _libinput_analyze_per-slot-delta()
{
_arguments \
'--help[Show help message and exit]' \
':recording:_files'
}
(( $+functions[_libinput_analyze_touch-down-state] )) || _libinput_analyze_touch-down-state()
{
_arguments \
'--help[Show help message and exit]' \
':recording:_files'
}
(( $+functions[_libinput_analyze_recording] )) || _libinput_analyze_recording()
{
_arguments \
'--help[Show help message and exit]' \
':recording:_files'
}
(( $+functions[_libinput_analyze] )) || _libinput_analyze()
{
local curcontext=$curcontext state line ret=1
local features
features=(
"per-slot-delta:analyze relative movement per touch per slot"
"recording:analyze a recording by printing a pretty table"
"touch-down-state:analyze a recording for logical touch down states"
)
_arguments -C \
'--help[Print help and exit]' \
':feature:->feature' \
'*:: :->option-or-argument'
case $state in
(feature)
_describe -t features 'feature' features
;;
(option-or-argument)
curcontext=${curcontext%:*:*}:libinput-analyze-$words[1]:
if ! _call_function ret _libinput_analyze_$words[1]; then
_message "unknown feature: $words[1]"
fi
;;
esac
return ret
}
(( $+functions[_libinput_record] )) || _libinput_record()
{
_arguments \
'--help[Show help message and exit]' \
'--all[Record all /dev/input/event* devices available on the system]' \
'--autorestart=[Terminate the current recording after s seconds of device inactivity]' \
{-o+,--output=}'[Specify the output file to use]:file:_files -g "*.yml"' \
'--multiple[Record multiple devices at once]' \
'--show-keycodes[Show keycodes as-is in the recording]' \
'--with-libinput[Record libinput events alongside device events]' \
'--with-hidraw[Record hidraw events alongside device events]' \
'*::device:_files -W /dev/input/ -P /dev/input/'
}
(( $+functions[_libinput_replay] )) || _libinput_replay()
{
_arguments \
'--help[Show help message and exit]' \
':recording:_files'
}
_libinput()
{
local curcontext=$curcontext state line ret=1
_arguments -C \
{-h,--help}'[Show help message and exit]' \
'--version[Show version information and exit]' \
':command:->command' \
'*:: :->option-or-argument' && return
case $state in
(command)
_libinput_commands && ret=0
;;
(option-or-argument)
curcontext=${curcontext%:*:*}:libinput-$words[1]:
if ! _call_function ret _libinput_$words[1]; then
_message "unknown libinput command: $words[1]"
fi
;;
esac
return ret
}
_libinput

View file

@ -1,12 +0,0 @@
zshcompletiondir = get_option('zshcompletiondir')
if zshcompletiondir == ''
zshcompletiondir = get_option('datadir') / 'zsh' / 'site-functions'
endif
if zshcompletiondir != 'no'
install_data(
'_libinput',
install_dir: zshcompletiondir,
install_mode: 'rw-r--r--',
)
endif

View file

@ -1,8 +1,5 @@
.. _absolute_axes:
==============================================================================
Absolute axes
==============================================================================
/**
@page absolute_axes Absolute axes
Devices with absolute axes are those that send positioning data for an axis in
a device-specific coordinate range, defined by a minimum and a maximum value.
@ -13,7 +10,7 @@ libinput supports three types of devices with absolute axes:
- multi-touch screens
- single-touch screens
- :ref:`graphics tablets <tablet-support>`
- @ref tablet-support "graphics tablets"
Touchpads are technically absolute devices but libinput converts the axis values
to directional motion and posts events as relative events. Touchpads do not count
@ -22,15 +19,11 @@ as absolute devices in libinput.
For all absolute devices in libinput, the default unit for x/y coordinates is
in mm off the top left corner on the device, or more specifically off the
device's sensor. If the device is physically rotated from its natural
position and this rotation was communicated to libinput (e.g. by setting
the device left-handed),
the coordinate origin is the top left corner in the current rotation.
position and this rotation was communicated to libinput (e.g.
@ref libinput_device_config_left_handed_set "by setting the device left-handed"),
the coordinate origin is the top left corner of in the current rotation.
.. _absolute_axes_handling:
------------------------------------------------------------------------------
Handling of absolute coordinates
------------------------------------------------------------------------------
@section absolute_axes_handling Handling of absolute coordinates
In most use-cases, absolute input devices are mapped to a single screen. For
direct input devices such as touchscreens the aspect ratio of the screen and
@ -38,8 +31,8 @@ the device match. Mapping the input device position to the output position is
thus a simple mapping between two coordinates. libinput provides the API for
this with
- **libinput_event_pointer_get_absolute_x_transformed()** for pointer events
- **libinput_event_touch_get_x_transformed()** for touch events
- libinput_event_pointer_get_absolute_x_transformed() for pointer events
- libinput_event_touch_get_x_transformed() for touch events
libinput's API only provides the call to map into a single coordinate range.
If the coordinate range has an offset, the compositor is responsible for
@ -47,98 +40,97 @@ applying that offset after the mapping. For example, if the device is mapped
to the right of two outputs, add the output offset to the transformed
coordinate.
.. _absolute_axes_nores:
------------------------------------------------------------------------------
Devices without x/y resolution
------------------------------------------------------------------------------
@section absolute_axes_nores Devices without x/y resolution
An absolute device that does not provide a valid resolution is considered
buggy and must be fixed in the kernel. Some touchpad devices do not
provide resolution, those devices are correctly handled within libinput
(touchpads are not absolute devices, as mentioned above).
.. _calibration:
------------------------------------------------------------------------------
Calibration of absolute devices
------------------------------------------------------------------------------
@section calibration Calibration of absolute devices
Absolute devices may require calibration to map precisely into the output
range required. This is done by setting a transformation matrix, see
**libinput_device_config_calibration_set_matrix()** which is applied to
libinput_device_config_calibration_set_matrix() which is applied to
each input coordinate.
.. math::
\begin{pmatrix}
cos\theta & -sin\theta & xoff \\
sin\theta & cos\theta & yoff \\
0 & 0 & 1
\end{pmatrix} \begin{pmatrix}
x \\ y \\ 1
\end{pmatrix}
@f[
\begin{pmatrix}
cos\theta & -sin\theta & xoff \\
sin\theta & cos\theta & yoff \\
0 & 0 & 1
\end{pmatrix} \begin{pmatrix}
x \\ y \\ 1
\end{pmatrix}
@f]
:math:`\theta` is the rotation angle. The offsets :math:`xoff` and :math:`yoff` are
specified in device dimensions, i.e. a value of 1 equals one device width or
height. Note that rotation applies to the device's origin, rotation usually
requires an offset to move the coordinates back into the original range.
@f$\theta@f$ is the rotation angle. The offsets @f$xoff@f$ and @f$yoff@f$ are
specified in device dimensions, i.e. a value of 1 equals one device width
or height. Note that rotation applies to the device's origin, rotation
usually requires an offset to move the coordinates back into the original
range.
The most common matrices are:
- 90 degree clockwise:
.. math::
\begin{pmatrix}
0 & -1 & 1 \\
1 & 0 & 0 \\
0 & 0 & 1
\end{pmatrix}
@f$
\begin{pmatrix}
0 & -1 & 1 \\
1 & 0 & 0 \\
0 & 0 & 1
\end{pmatrix}
@f$
- 180 degree clockwise:
.. math::
\begin{pmatrix}
-1 & 0 & 1 \\
0 & -1 & 1 \\
0 & 0 & 1
\end{pmatrix}
@f$
\begin{pmatrix}
-1 & 0 & 1 \\
0 & -1 & 1 \\
0 & 0 & 1
\end{pmatrix}
@f$
- 270 degree clockwise:
.. math::
\begin{pmatrix}
0 & 1 & 0 \\
-1 & 0 & 1 \\
0 & 0 & 1
\end{pmatrix}
@f$
\begin{pmatrix}
0 & 1 & 0 \\
-1 & 0 & 1 \\
0 & 0 & 1
\end{pmatrix}
@f$
- reflection along y axis:
.. math::
\begin{pmatrix}
-1 & 0 & 1 \\
1 & 0 & 0 \\
0 & 0 & 1
\end{pmatrix}
@f$
\begin{pmatrix}
-1 & 0 & 1 \\
1 & 0 & 0 \\
0 & 0 & 1
\end{pmatrix}
@f$
See Wikipedia's
`Transformation Matrix article <http://en.wikipedia.org/wiki/Transformation_matrix>`_
for more information on the matrix maths. See
**libinput_device_config_calibration_get_default_matrix()** for how these
<a href="http://en.wikipedia.org/wiki/Transformation_matrix">Transformation
Matrix article</a> for more information on the matrix maths. See
libinput_device_config_calibration_get_default_matrix() for how these
matrices must be supplied to libinput.
Once applied, any x and y axis value has the calibration applied before it
is made available to the caller. libinput does not provide access to the
raw coordinates before the calibration is applied.
.. _absolute_axes_nonorm:
@section absolute_axes_nonorm Why x/y coordinates are not normalized
------------------------------------------------------------------------------
Why x/y coordinates are not normalized
------------------------------------------------------------------------------
x/y are not given in :ref:`normalized coordinates <motion_normalization>`
x/y are not given in @ref motion_normalization "normalized coordinates"
([0..1]) for one simple reason: the aspect ratio of virtually all current
devices is something other than 1:1. A normalized axes thus is only useful to
determine that the stylus is e.g. at 78% from the left, 34% from the top of
the device. Without knowing the per-axis resolution, these numbers are
meaningless. Worse, calculation based on previous coordinates is simply wrong:
a movement from 0/0 to 50%/50% is not a 45-degree line.
a movement from 0/0 to 50%/50% is not a 45% degree line.
This could be alleviated by providing resolution and information about the
aspect ratio to the caller. Which shifts processing and likely errors into the
caller for little benefit. Providing the x/y axes in mm from the outset
removes these errors.
*/

View file

@ -0,0 +1,117 @@
/**
@page absolute_coordinate_ranges Coordinate ranges for absolute axes
libinput requires that all touchpads provide a correct axis range and
resolution. These are used to enable or disable certain features or adapt
the interaction with the touchpad. For example, the software button area is
narrower on small touchpads to avoid reducing the interactive surface too
much. Likewise, palm detection works differently on small touchpads as palm
interference is less likely to happen.
Touchpads with incorrect axis ranges generate error messages
in the form:
<blockquote>
Axis 0x35 value 4000 is outside expected range [0, 3000]
</blockquote>
This error message indicates that the ABS_MT_POSITION_X axis (i.e. the x
axis) generated an event outside the expected range of 0-3000. In this case
the value was 4000.
This discrepancy between the coordinate range the kernels advertises vs.
what the touchpad sends can be the source of a number of perceived
bugs in libinput.
@section absolute_coordinate_ranges_fix Measuring and fixing touchpad ranges
To fix the touchpad you need to:
-# measure the physical size of your touchpad in mm
-# run touchpad-edge-detector
-# trim the udev match rule to something sensible
-# replace the resolution with the calculated resolution based on physical
settings
-# test locally
-# send a patch to the systemd project
Detailed explanations are below.
[libevdev](http://freedesktop.org/wiki/Software/libevdev/) provides a tool
called **touchpad-edge-detector** that allows measuring the touchpad's input
ranges. Run the tool as root against the device node of your touchpad device
and repeatedly move a finger around the whole outside area of the
touchpad. Then control+c the process and note the output.
An example output is below:
@code
$> sudo touchpad-edge-detector /dev/input/event4
Touchpad SynPS/2 Synaptics TouchPad on /dev/input/event4
Move one finger around the touchpad to detect the actual edges
Kernel says: x [1024..3112], y [2024..4832]
Touchpad sends: x [2445..4252], y [3464..4071]
Touchpad size as listed by the kernel: 49x66mm
Calculate resolution as:
x axis: 2088/<width in mm>
y axis: 2808/<height in mm>
Suggested udev rule:
# <Laptop model description goes here>
evdev:name:SynPS/2 Synaptics TouchPad:dmi:bvnLENOVO:bvrGJET72WW(2.22):bd02/21/2014:svnLENOVO:pn20ARS25701:pvrThinkPadT440s:rvnLENOVO:rn20ARS25701:rvrSDK0E50512STD:cvnLENOVO:ct10:cvrNotAvailable:*
EVDEV_ABS_00=2445:4252:<x resolution>
EVDEV_ABS_01=3464:4071:<y resolution>
EVDEV_ABS_35=2445:4252:<x resolution>
EVDEV_ABS_36=3464:4071:<y resolution>
@endcode
Note the discrepancy between the coordinate range the kernels advertises vs.
what the touchpad sends.
To fix the advertised ranges, the udev rule should be taken and trimmed
before being sent to the [systemd project](https://github.com/systemd/systemd).
An example commit can be found
[here](https://github.com/systemd/systemd/commit/26f667eac1c5e89b689aa0a1daef6a80f473e045).
In most cases the match can and should be trimmed to the system vendor (svn)
and the product version (pvr), with everything else replaced by a wildcard
(*). In this case, a Lenovo T440s, a suitable match string would be: @code
evdev:name:SynPS/2 Synaptics TouchPad:dmi:*svnLENOVO:*pvrThinkPadT440s*
@endcode
@note hwdb match strings only allow for alphanumeric ascii characters. Use a
wildcard (* or ?, whichever appropriate) for special characters.
The actual axis overrides are in the form:
@code
# axis number=min:max:resolution
EVDEV_ABS_00=2445:4252:42
@endcode
or, if the range is correct but the resolution is wrong
@code
# axis number=::resolution
EVDEV_ABS_00=::42
@endcode
Note the leading single space. The axis numbers are in hex and can be found
in *linux/input-event-codes.h*. For touchpads ABS_X, ABS_Y,
ABS_MT_POSITION_X and ABS_MT_POSITION_Y are required.
@note The touchpad's ranges and/or resolution should only be fixed when
there is a significant discrepancy. A few units do not make a difference and
a resolution that is off by 2 or less usually does not matter either.
Once a match and override rule has been found, follow the instructions at
the top of the
[60-evdev.hwdb](https://github.com/systemd/systemd/blob/master/hwdb/60-evdev.hwdb)
file to save it locally and trigger the udev hwdb reload. Rebooting is
always a good idea. If the match string is correct, the new properties will
show up in the
output of
@code
udevadm info /sys/class/input/event4
@endcode
Adjust the command for the event node of your touchpad.
A udev builtin will apply the new axis ranges automatically.
When the axis override is confirmed to work, please submit it as a pull
request to the [systemd project](https://github.com/systemd/systemd).
*/

View file

@ -1,135 +0,0 @@
/**
@mainpage
This is the libinput API reference.
This documentation is aimed at developers of Wayland compositors. User
documentation is available
[here](https://wayland.freedesktop.org/libinput/doc/latest).
@section concepts Concepts
@subsection concepts_initialization Initialization of a libinput context
libinput provides two different backends:
- a @ref libinput_udev_create_context "udev backend" where notifications
about new and removed devices are provided by udev, and
- a @ref libinput_path_create_context "path backend" where
@ref libinput_path_add_device "device addition" and
@ref libinput_path_remove_device "device removal" need to be handled by
the caller.
See section @ref base for information about initializing a libinput context.
@subsection concepts_events Monitoring for events
libinput exposes a single @ref libinput_get_fd "file descriptor" to the
caller. This file descriptor should be monitored by the caller, whenever
data is available the caller **must** immediately call libinput_dispatch().
Failure to do so will result in erroneous behavior.
libinput_dispatch() may result in one or more events being available to the
caller. After libinput_dispatch() a caller **should** call
libinput_get_event() to retrieve and process this event. Whenever
libinput_get_event() returns `NULL`, no further events are available.
See section @ref event for more information about events.
@subsection concepts_seats Device grouping into seats
All devices are grouped into physical and logical seats. Button and key
states are available per-device and per-seat. See @ref seat for more
information.
@subsection concepts_devices Device capabilities
libinput does not use device types. All devices have @ref
libinput_device_has_capability "capabilities" that define which events may
be generated. See @ref device for more information about devices.
Specific event types include:
- @ref event_keyboard
- @ref event_pointer
- @ref event_touch
- @ref event_gesture
- @ref event_tablet
- @ref event_tablet_pad
- @ref event_switch
@subsection concepts_configuration Device configuration
libinput relies on the caller for device configuration. See
@ref config for more information.
@subsection example An example libinput program
The simplest libinput program looks like this:
@code
static int open_restricted(const char *path, int flags, void *user_data)
{
int fd = open(path, flags);
return fd < 0 ? -errno : fd;
}
static void close_restricted(int fd, void *user_data)
{
close(fd);
}
const static struct libinput_interface interface = {
.open_restricted = open_restricted,
.close_restricted = close_restricted,
};
int main(void) {
struct libinput *li;
struct libinput_event *event;
li = libinput_udev_create_context(&interface, NULL, udev);
libinput_udev_assign_seat(li, "seat0");
libinput_dispatch(li);
while ((event = libinput_get_event(li)) != NULL) {
// handle the event here
libinput_event_destroy(event);
libinput_dispatch(li);
}
libinput_unref(li);
return 0;
}
@endcode
@section building_against Building against libinput
libinput provides a
[pkg-config](https://www.freedesktop.org/wiki/Software/pkg-config/) file.
Software that uses libinput should use pkg-config and the
`PKG_CHECK_MODULES` autoconf macro.
Otherwise, the most rudimentary way to compile and link a program against
libinput is:
@verbatim
gcc -o myprogram myprogram.c `pkg-config --cflags --libs libinput`
@endverbatim
For further information on using pkgconfig see the pkg-config documentation.
@section stability Backwards-compatibility
libinput promises backwards-compatibility across all the 1.x.y version. An
application built against libinput 1.x.y will work with any future 1.*.*
release.
@section About
Documentation generated from git commit [__GIT_VERSION__](https://gitlab.freedesktop.org/libinput/libinput/commit/__GIT_VERSION__)
*/

View file

@ -1,52 +0,0 @@
prg_install = find_program('install')
doxygen = find_program('doxygen', required : false)
if not doxygen.found()
error('Program "doxygen" not found or not executable. Try building with -Ddocumentation=false')
endif
dot = find_program('dot', required : false)
if not dot.found()
error('Program "dot" not found or not executable. Try building with -Ddocumentation=false')
endif
mainpage = vcs_tag(command : ['git', 'log', '-1', '--format=%h'],
fallback : 'unknown',
input : 'mainpage.dox',
output : 'mainpage.dox',
replace_string: '__GIT_VERSION__')
src_doxygen = files(
# source files
'../../src/libinput.h',
# style files
'style/header.html',
'style/footer.html',
'style/customdoxygen.css',
'style/bootstrap.css',
'style/libinputdoxygen.css',
)
doxyfiles = []
foreach f : src_doxygen
df = configure_file(input: f,
output: '@PLAINNAME@',
copy : true)
doxyfiles += [ df ]
endforeach
doc_config = configuration_data()
doc_config.set('PACKAGE_NAME', meson.project_name())
doc_config.set('PACKAGE_VERSION', meson.project_version())
doc_config.set('builddir', meson.current_build_dir())
doxyfile = configure_file(input : 'libinput.doxygen.in',
output : 'libinput.doxygen',
configuration : doc_config)
custom_target('doxygen',
input : [ doxyfiles, doxyfile, mainpage ] + src_doxygen,
output : [ 'html' ],
command : [ doxygen, doxyfile ],
install : false,
depends: [ mainpage ],
build_by_default : true)

View file

@ -1,7 +0,0 @@
/**
@page Tablets
- @subpage tablet serial numbers
*/

View file

@ -1,120 +0,0 @@
@import url("https://fonts.googleapis.com/css?family=Roboto+Mono");
dd {
margin-left: 30px;
}
.title {
font-size: 200%;
font-weight: bold;
}
.title .ingroups {
font-size: 50%;
}
h1 {
font-size: 150%;
color: #354C7B;
background: none;
border-bottom: 1px solid #879ECB;
font-size: 150%;
font-weight: normal;
padding-top: 8px;
padding-bottom: 8px;
padding-left: 0px;
width: 100%;
}
h2 {
font-size: 120%;
color: #354C7B;
background: none;
border-bottom: 1px solid #879ECB;
font-size: 150%;
font-weight: normal;
padding-top: 8px;
padding-bottom: 8px;
padding-left: 0px;
width: 100%;
}
.sm-dox li {
float:left;
border-top: 0;
padding-right: 20px;
}
.sm li, .sm a {
position: relative;
}
.sm, .sm ul, .sm li {
list-style: none;
display: block;
line-height: normal;
direction: ltr;
text-align: left;
}
.sm, .sm *, .sm *::before, .sm *::after {
box-sizing: border-box;
}
#main-nav {
padding: 30px;
}
/* Main menu sub-items like file-list, etc */
#main-menu li ul {
display: none;
}
.paramname {
padding-right: 10px;
}
.memtitle {
background-image: none;
background-color: #F0F0F0;
}
.memproto {
background-color: #F0F0F0;
}
.headertitle {
background-image: none;
background-color: #F0F0F0;
}
div.header {
border: none;
}
td.fieldname {
font-family: 'Roboto Mono', monospace;
}
.fieldtable th {
background-image: none;
background-color: #F0F0F0;
}
body {
letter-spacing: 0px;
}
.mdescLeft, .mdescRight, .memItemLeft, .memItemRight, .memTemplItemLeft, .memTemplItemRight, .memTemplParams {
background-color: #F0F0F0;
}
a {
color: #2873b0;
}
.navpath ul {
background-image: none;
background-color: #F0F0F0;
}

282
doc/architecture.dox Normal file
View file

@ -0,0 +1,282 @@
/**
@page architecture libinput's internal architecture
This page provides an outline of libinput's internal architecture. The goal
here is to get the high-level picture across and point out the components
and their interplay to new developers.
The public facing API is in `libinput.c`, this file is thus the entry point
for almost all API calls. General device handling is in `evdev.c` with the
device-type-specific implementations in `evdev-<type>.c`. It is not
necessary to understand all of libinput to contribute a patch.
@ref architecture-contexts is the only user-visible implementation detail,
everything else is purely internal implementation and may change when
required.
@section architecture-contexts The udev and path contexts
The first building block is the "context" which can be one of
two types, "path" and "udev". See libinput_path_create_context() and
libinput_udev_create_context(). The path/udev specific bits are in
`path-seat.c` and `udev-seat.c`. This includes the functions that add new
devices to a context.
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
libudev [label="libudev 'add' event"]
udev [label="libinput_udev_create_context()"];
udev_backend [label="udev-specific backend"];
context [label="libinput context"]
udev -> udev_backend;
libudev -> udev_backend;
udev_backend -> context;
}
@enddot
The udev context provides automatic device hotplugging as udev's "add"
events are handled directly by libinput. The path context requires that the
caller adds devices.
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
path [label="libinput_path_create_context()"];
path_backend [label="path-specific backend"];
xdriver [label="libinput_path_add_device()"]
context [label="libinput context"]
path -> path_backend;
xdriver -> path_backend;
path_backend -> context;
}
@enddot
As a general rule: all Wayland compositors use a udev context, the X.org
stack uses a path context.
Which context was initialized only matters for creating/destroying a context
and adding devices. The device handling itself is the same for both types of
context.
@section architecture-device Device initialization
libinput only supports evdev devices, all the device initialization is done
in `evdev.c`. Much of the libinput public API is also a thin wrapper around
the matching implementation in the evdev device.
There is a 1:1 mapping between libinput devices and `/dev/input/eventX`
device nodes.
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
devnode [label="/dev/input/event0"]
libudev [label="libudev 'add' event"]
xdriver [label="libinput_path_add_device()"]
context [label="libinput context"]
evdev [label="evdev_device_create()"]
devnode -> xdriver;
devnode -> libudev;
xdriver -> context;
libudev -> context;
context->evdev;
}
@enddot
Entry point for all devices is `evdev_device_create()`, this function
decides to create a `struct evdev_device` for the given device node.
Based on the udev tags (e.g. `ID_INPUT_TOUCHPAD`), a @ref
architecture-dispatch is initialized. All event handling is then in this
dispatch.
Rejection of devices and the application of quirks is generally handled in
`evdev.c` as well. Common functionality shared across multiple device types
(like button-scrolling) is also handled here.
@section architecture-dispatch Device-type specific event dispatch
Depending on the device type, `evdev_configure_device` creates the matching
`struct evdev_dispatch`. This dispatch interface contains the function
pointers to handle events. Four such dispatch methods are currently
implemented: touchpad, tablet, tablet pad, and the fallback dispatch which
handles mice, keyboards and touchscreens.
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
evdev [label="evdev_device_create()"]
fallback [label="evdev-fallback.c"]
touchpad [label="evdev-mt-touchpad.c"]
tablet [label="evdev-tablet.c"]
pad [label="evdev-tablet-pad.c"]
evdev -> fallback;
evdev -> touchpad;
evdev -> tablet;
evdev -> pad;
}
@enddot
While `evdev.c` pulls the event out of libevdev, the actual handling of the
events is performed within the dispatch method.
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
evdev [label="evdev_device_dispatch()"]
fallback [label="fallback_interface_process()"];
touchpad [label="tp_interface_process()"]
tablet [label="tablet_process()"]
pad [label="pad_process()"]
evdev -> fallback;
evdev -> touchpad;
evdev -> tablet;
evdev -> pad;
}
@enddot
The dispatch methods then look at the `struct input_event` and proceed to
update the state. Note: the serialized nature of the kernel evdev protocol
requires that the device updates the state with each event but to delay
processing until the `SYN_REPORT` event is received.
@section architecture-configuration Device configuration
All device-specific configuration is handled through `struct
libinput_device_config_FOO` instances. These are set up during device init
and provide the function pointers for the `get`, `set`, `get_default`
triplet of configuration queries (or more, where applicable).
For example, the `struct tablet_dispatch` for tablet devices has a
`struct libinput_device_config_accel`. This struct is set up with the
required function pointers to change the profiles.
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
tablet [label="struct tablet_dispatch"]
config [label="struct libinput_device_config_accel"];
tablet_config [label="tablet_accel_config_set_profile()"];
tablet->config;
config->tablet_config;
}
@enddot
When the matching `libinput_device_config_set_FOO()` is called, this goes
through to the config struct and invokes the function there. Thus, it is
possible to have different configuration functions for a mouse vs a
touchpad, even though the interface is the same.
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
libinput [label="libinput_device_config_accel_set_profile()"];
tablet_config [label="tablet_accel_config_set_profile()"];
libinput->tablet_config;
}
@enddot
@section architecture-filter Pointer acceleration filters
All pointer acceleration is handled in the `filter.c` file and its
associated files.
The `struct motion_filter` is initialized during device init, whenever
deltas are available they are passed to `filter_dispatch()`. This function
returns a set of @ref motion_normalization_customization "normalized coordinates".
All actual acceleration is handled within the filter, the device itself has
no further knowledge. Thus it is possible to have different acceleration
filters for the same device types (e.g. the Lenovo X230 touchpad has a
custom filter).
@dot
digraph context
{
compound=true;
rankdir="LR";
node [
shape="box";
]
fallback [label="fallback deltas"];
touchpad [label="touchpad deltas"];
tablet [label="tablet deltas"];
filter [label="filter_dispatch"];
fallback->filter;
touchpad->filter;
tablet->filter;
flat [label="accelerator_interface_flat()"];
x230 [label="accelerator_filter_x230()"];
pen [label="tablet_accelerator_filter_flat_pen()"];
filter->flat;
filter->x230;
filter->pen;
}
@enddot
Most filters convert the deltas (incl. timestamps) to a motion speed and
then apply a so-called profile function. This function returns a factor that
is then applied to the current delta, converting it into an accelerated
delta. See @ref pointer-acceleration for more details.
the current
*/

241
doc/building.dox Normal file
View file

@ -0,0 +1,241 @@
/**
@page building_libinput libinput build instructions
@tableofcontents
Instructions on how to build libinput and its tools and how to build against
libinput.
The build instruction on this page detail how to overwrite your
system-provided libinput with one from the git repository, see
see @ref reverting_install to revert to the previous state.
@section building Building libinput
libinput uses [meson](https://www.mesonbuild.com) and
[ninja](https://www.ninja-build.org). A build is usually the three-step
process below. A successful build requires the @ref
building_dependencies to be installed before running meson.
@code
$> git clone https://gitlab.freedesktop.org/libinput/libinput
$> cd libinput
$> meson --prefix=/usr builddir/
$> ninja -C builddir/
$> sudo ninja -C builddir/ install
$> sudo udevadm hwdb --update
@endcode
Additional options may also be specified. For example:
@code
$> meson --prefix=/usr -Ddocumentation=false builddir/
@endcode
We recommend that users disable the documentation, it's not usually required
for testing and reduces the number of dependencies needed.
The ```prefix``` or other options can be changed later with the
```mesonconf``` command. For example:
@code
$> mesonconf builddir/ -Dprefix=/some/other/prefix -Ddocumentation=true
$> ninja -C builddir
$> sudo ninja -C builddir/ install
@endcode
Running ``mesonconf builddir/`` with no other arguments lists all
configurable options meson provides.
To rebuild from scratch, simply remove the build directory and run meson
again:
@code
$> rm -r builddir/
$> meson --prefix=....
@endcode
@subsection verifying_install Verifying the install
To verify the install worked correctly, check that libinput.so.x.x.x is in
the library path and that all symlinks point to the new library.
<pre>
$> ls -l /usr/lib64/libinput.*
-rwxr-xr-x 1 root root 946 Apr 28 2015 /usr/lib64/libinput.la
lrwxrwxrwx 1 root root 19 Feb 1 15:12 /usr/lib64/libinput.so -> libinput.so.10.13.0
lrwxrwxrwx 1 root root 19 Feb 1 15:12 /usr/lib64/libinput.so.10 -> libinput.so.10.13.0
-rwxr-xr-x 1 root root 204992 Feb 1 15:12 /usr/lib64/libinput.so.10.13.0
</pre>
@subsection reverting_install Reverting to the system-provided libinput package
The recommended way to revert to the system install is to use the package
manager to reinstall the libinput package. In some cases, this may leave
files in the system (e.g. ```/usr/lib/libinput.la```) but these files are
usually harmless. To definitely remove all files, run the following command
from the libinput source directory:
@code
$> sudo ninja -C builddir/ uninstall
# WARNING: Do not restart the computer/X/the Wayland compositor after
# uninstall, reinstall the system package immediately!
@endcode
The following commands reinstall the current system package for libinput,
overwriting manually installed files.
<ul>
<li><b>Debian/Ubuntu</b> based distributions: ```sudo apt-get install
--reinstall libinput```</li>
<li><b>Fedora 22</b> and later: ```sudo dnf reinstall libinput```</li>
<li><b>RHEL/CentOS/Fedora 21</b> and earlier: ```sudo yum reinstall libinput```</li>
<li><b>openSUSE</b>: ```sudo zypper install --force libinput10```</li>
<li><b>Arch</b>: ```sudo packman -S libinput```</li>
</ul>
@subsection building_selinux SELinux adjustments
On systems with SELinux, overwriting the distribution-provided package with
a manually built libinput may cause SELinux denials. This usually manifests
when gdm does not start because it is denied access to libinput. The journal
shows a log message in the form of:
<pre>
May 25 15:28:42 localhost.localdomain audit[23268]: AVC avc: denied { execute } for pid=23268 comm="gnome-shell" path="/usr/lib64/libinput.so.10.12.2" dev="dm-0" ino=1709093 scontext=system_u:system_r:xdm_t:s0-s0:c0.c1023 tcontext=unconfined_u:object_r:user_home_t:s0 tclass=file permissive=0
May 25 15:28:42 localhost.localdomain org.gnome.Shell.desktop[23270]: /usr/bin/gnome-shell: error while loading shared libraries: libinput.so.10: failed to map segment from shared object
</pre>
The summary of this error message is that gdm's gnome-shell runs in the
```system_u:system_r:xdm_t``` context but libinput is installed with the
context ```unconfined_u:object_r:user_home_t```.
To avoid this issue, restore the SELinux context for any system files.
<pre>
$> sudo restorecon /usr/lib*/libinput.so.*
</pre>
This issue is tracked in https://github.com/mesonbuild/meson/issues/1967.
@subsection building_dependencies Build dependencies
libinput has a few build-time dependencies that must be installed prior to
running configure. In most cases, it is sufficient to install the
dependencies that your distribution uses to build the libinput package.
These can be installed with one of the following commands:
<ul>
<li><b>Debian/Ubuntu</b> based distributions: ```sudo apt-get build-dep
libinput```</li>
<li><b>Fedora 22</b> and later: ```sudo dnf builddep libinput```</li>
<li><b>RHEL/CentOS/Fedora 21</b> and earlier: ```sudo yum-builddep libinput```</li>
<li><b>openSUSE</b>:
<pre>
$> sudo zypper modifyrepo --enable `zypper repos | grep source | awk '{print $5}'`
$> sudo zypper source-install -d libinput10
$> sudo zypper install autoconf automake libtool
$> sudo zypper modifyrepo --disable `zypper repos | grep source | awk '{print $5}'`
</pre>
</li>
<li><b>Arch</b>:
<pre>
$> sudo pacman -S asp
$> cd $(mktemp -d)
$> asp export libinput
$> cd libinput
$> makepkg --syncdeps --nobuild --noextract
</pre>
</li>
</ul>
If dependencies are missing, a message ```No package 'foo' found``` will be
shown during the configure stage. See
<a href="https://who-t.blogspot.com.au/2014/05/configure-fails-with-no-package-foo.html">this blog post here</a>.
for instructions on how to fix it.
@subsection building_libwacom Building without libwacom
libwacom is required by libinput's tablet code to gather additional
information about tablets that is not available from the kernel device
itself. libwacom is required by default but can be skipped when @ref
building.
@code
$> meson --prefix=/usr -Dlibwacom=false builddir
@endcode
It is not recommended to disable libwacom unless libinput is used in an
environment where tablet support is not required. libinput provides tablet
support even without libwacom, but some features may be missing or working
differently.
@subsection building_debug_gui Building without the graphical helper tool
The @ref tools provide commandline features as well as graphical debugging
features. To keep dependencies in check on some builds, the graphical
features of the @ref tools can be disabled. By default, the `debug-gui`
feature of the `libinput` tool is enabled and if the required libraries are
not available, the build will fail. If the feature is not required, use the
``--disable-debug-gui`` argument when @ref building.
@code
$> meson --prefix=/usr -Ddebug-gui=false builddir
@endcode
@subsection building_autotools Building with autotools
<b>libinput no longer supports building with autotools.</b> These
instructions are kept for users for libinput versions up to 1.8.x.
A build with automake is usually the process below. A successful build
requires the @ref building_dependencies to be installed at configure
time.
@code
$> git clone https://gitlab.freedesktop.org/libinput/libinput
$> cd libinput
$> ./autogen.sh --prefix=/usr --libdir=/usr/lib64
$> make
$> sudo make install
$> sudo udevadm hwdb --update
@endcode
@note On Debian-based distributions including Ubuntu and its derivatives skip the
```--libdir=/usr/lib64``` argument.
To uninstall libinput as detailed in section @ref reverting_install, run
@code
$> sudo make uninstall
# WARNING: Do not restart the computer/X/the Wayland compositor after make
# uninstall, reinstall the system package immediately!
@endcode
To disable libwacom as detailed in section @ref building_libwacom, run
@code
$> ./autogen.sh --disable-libwacom --prefix=/usr --libdir=/usr/lib64
@endcode
To disable the graphical helper tool as detailed in section @ref
building_debug_gui, run
@code
$> ./autogen.sh --disable-debug-gui --prefix=/usr --libdir=/usr/lib64
@endcode
@section building_against Building against libinput
libinput provides a
[pkg-config](https://www.freedesktop.org/wiki/Software/pkg-config/) file.
Software that uses libinput should use pkg-config and the
`PKG_CHECK_MODULES` autoconf macro.
Otherwise, the most rudimentary way to compile and link a program against
libinput is:
@code
gcc -o myprogram myprogram.c `pkg-config --cflags --libs libinput`
@endcode
For further information on using pkgconfig see the pkg-config documentation.
*/

File diff suppressed because one or more lines are too long

Before

Width:  |  Height:  |  Size: 97 KiB

After

Width:  |  Height:  |  Size: 78 KiB

View file

@ -1,9 +1,6 @@
/**
.. _button_debouncing:
==============================================================================
Button debouncing
==============================================================================
@page button_debouncing Button debouncing
Physical buttons experience wear-and-tear with usage. On some devices this
can result in an effect called "contact bouncing" or "chatter". This effect
@ -32,8 +29,8 @@ immediately and most release events are delivered immediately. The
does not enable this method unless a faulty event sequence is detected. A
message is printed to the log when spurious deboucing was detected.
libinput's debouncing is supposed to correct hardware damage or
substandard hardware. Debouncing also exists as an accessibility feature
Note that libinput's debouncing intended to correct hardware damage or
substandard hardware. Debouncing is also used as an accessibility feature
but the requirements are different. In the accessibility feature, multiple
physical key presses, usually caused by involuntary muscle movement, must be
filtered to only one key press. This feature must be implemented higher in
@ -45,12 +42,6 @@ extra line is added to show the timeouts used by libinput that
affect the button state handling. The waveform's high and low states
correspond to the buttons 'pressed' and 'released' states, respectively.
.. figure:: button-debouncing-wave-diagram.svg
:align: center
@image html button-debouncing-wave-diagram.svg "Diagram illustrating button debouncing"
Diagram illustrating button debouncing
Some devices send events in bursts, erroneously triggering the button
debouncing detection. Please :ref:`file a bug <reporting_bugs>` if that
occurs for your device.
*/

View file

@ -0,0 +1,103 @@
/**
@page clickpad_softbuttons Clickpad software button behavior
Clickpad is the name given to touchpads without physical buttons below the
touchpad. Instead, the whole touchpad acts as a button and left or right
button clicks are distinguished by the location and/or number of fingers on
the touchpad. <a href="http://www.synaptics.com/en/clickpad.php">"ClickPad" is
a trademark by Synaptics Inc.</a> but for simplicity we refer to any
touchpad with the above feature as Clickpad, regardless of the manufacturer.
A clickpad is always marked with the <a
href="https://www.kernel.org/doc/Documentation/input/event-codes.txt">INPUT_PROP_BUTTONPAD</a> property.
To perform a right-click on a Clickpad, libinput provides @ref
software_buttons and @ref clickfinger.
In the page below, the term "click" shall refer to a physical button press
and/or release of the touchpad, the term "button event" refers to the events
generated by libinput and passed to the caller in response to a click.
@section software_buttons Software button areas
On most clickpads, this is the default behavior. The bottom of the touchpad
is split into three distinct areas generate left, middle or right button
events on click. The height of the button area depends on the hardware but
is usually around 10mm.
Left, right and middle button events can be triggered as follows:
- if a finger is in the main area or the left button area, a click generates
left button events.
- if a finger is in the right area, a click generates right button events.
- if a finger is in the middle area, a click generates middle button events.
The middle button is always centered on the touchpad and smaller in size
than the left or right button. The actual size is device-dependent though as
many touchpads do not have visible markings for the middle button the exact
location of the button is not visibly obvious.
@image html software-buttons.svg "Left, right and middle-button click with software button areas"
@note If middle button emulation is enabled on a clickpad, only left and right
button areas are available. For more details, see
libinput_device_config_middle_emulation_set_enabled().
If fingers are down in the main area in addition to fingers in the
left or right button area, those fingers are are ignored.
A release event always releases the buttons logically down, regardless of
the current finger position
The movement of a finger can alter the button area behavior:
- if a finger starts in the main area and moves into the software button
area, the software buttons do not apply to that finger
- a finger in the software button area does not move the pointer
- if a finger moves out out of the button area it will control the pointer
if it's the first finger in the main area
- once a finger has moved out of the button area, it cannot move back in and
trigger a right or middle button event
On some touchpads, notably the 2015 Lenovo X1 Carbon 3rd series, the very
bottom end of the touchpad is outside of the sensor range but it is possible
to trigger a physical click there. To libinput, the click merely shows up as
a left button click without any positional finger data and it is
impossible to determine whether it is a left or a right click. libinput
ignores such button clicks, this behavior is intentional.
@section clickfinger Clickfinger behavior
This is the default behavior on Apple touchpads.
Here, a left, right, middle button event is generated when one, two, or
three fingers are held down on the touchpad when a physical click is
generated. The location of the fingers does not matter and there are no
software-defined button areas.
@image html clickfinger.svg "One, two and three-finger click with Clickfinger behavior"
On some touchpads, libinput imposes a limit on how the fingers may be placed
on the touchpad. In the most common use-case this allows for a user to
trigger a click with the thumb while leaving the pointer-moving finger on
the touchpad.
@image html clickfinger-distance.svg "Illustration of the distance detection algorithm"
In the illustration above the red area marks the proximity area around the
first finger. Since the thumb is outside of that area libinput considers the
click a single-finger click rather than a two-finger click.
Clickfinger configuration can be enabled through the
libinput_device_config_click_set_method() call. If clickfingers are
enabled on a touchpad with top software buttons, the top area will keep
acting as softbuttons for use with the trackpoint. Clickfingers will be used
everywhere else on the touchpad.
@section special_clickpads Special Clickpads
The Lenovo *40 series laptops have a clickpad that provides two software button sections, one at
the top and one at the bottom. See @ref t440_support "Lenovo *40 series touchpad support"
for details on the top software button.
Some Clickpads, notably some Cypress ones, perform right button detection in
firmware and appear to userspace as if the touchpad had physical buttons.
While physically clickpads, these are not handled by the software and
treated like traditional touchpads.
*/

18
doc/contributing.dox Normal file
View file

@ -0,0 +1,18 @@
/**
@page contributing Contributing to libinput
Contributions to libinput are always welcome. Any patches should be sent to
the
[wayland-devel](https://lists.freedesktop.org/mailman/listinfo/wayland-devel)
email list with a subject prefix of ````[PATCH libinput]````. The easiest
way to achieve that is to run the following command in the libinput
repository:
git config --add format.subjectprefix "PATCH libinput"
Likewise, any new features should also be discussed publicly on the
wayland-devel list.
*/

View file

@ -0,0 +1,303 @@
/**
@page udev_config Static device configuration via udev
libinput supports some static configuration through udev properties.
These properties are read when the device is initially added
to libinput's device list, i.e. before the @ref
LIBINPUT_EVENT_DEVICE_ADDED event is generated.
The following udev properties are supported:
<dl>
<dt>LIBINPUT_CALIBRATION_MATRIX</dt>
<dd>Sets the calibration matrix, see
libinput_device_config_calibration_get_default_matrix(). If unset,
defaults to the identity matrix.
The udev property is parsed as 6 floating point numbers separated by a
single space each (scanf(3) format "%f %f %f %f %f %f").
The 6 values represent the first two rows of the calibration matrix as
described in libinput_device_config_calibration_set_matrix().
Example values are:
@code
ENV{LIBINPUT_CALIBRATION_MATRIX}="1 0 0 0 1 0" # default
ENV{LIBINPUT_CALIBRATION_MATRIX}="0 -1 1 1 0 0" # 90 degree clockwise
ENV{LIBINPUT_CALIBRATION_MATRIX}="-1 0 1 0 -1 1" # 180 degree clockwise
ENV{LIBINPUT_CALIBRATION_MATRIX}="0 1 0 -1 0 1" # 270 degree clockwise
ENV{LIBINPUT_CALIBRATION_MATRIX}="-1 0 1 0 1 0" # reflect along y axis
@endcode
</dd>
<dt>LIBINPUT_DEVICE_GROUP</dt>
<dd>A string identifying the @ref libinput_device_group for this device. Two
devices with the same property value are grouped into the same device group,
the value itself is irrelevant otherwise.
</dd>
<dt>LIBINPUT_IGNORE_DEVICE</dt>
<dd>If set to anything other than "0", the device is ignored by libinput.
See @ref ignoring_devices for more details.
</dd>
<dt>ID_SEAT</dt>
<dd>Assigns the physical @ref seats "seat" for this device. See
libinput_seat_get_physical_name(). Defaults to "seat0".</dd>
<dt>ID_INPUT</dt>
<dd>If this property is set, the device is considered an input device. Any
device with this property missing will be ignored, see @ref
udev_device_type.</dt>
</dd>
<dt>ID_INPUT_KEYBOARD, ID_INPUT_KEY, ID_INPUT_MOUSE, ID_INPUT_TOUCHPAD,
ID_INPUT_TOUCHSCREEN, ID_INPUT_TABLET, ID_INPUT_JOYSTICK,
ID_INPUT_ACCELEROMETER</dt>
<dd>If any of the above is set, libinput initializes the device as the given
type, see @ref udev_device_type. Note that for historical reasons more than
one of these may be set at any time, libinput will select only one of these
to determine the device type. To ensure libinput selects the correct device
type, only set one of them.</dd>
<dt>WL_SEAT</dt>
<dd>Assigns the logical @ref seats "seat" for this device. See
libinput_seat_get_logical_name()
context. Defaults to "default".</dd>
<dt>MOUSE_DPI</dt>
<dd>HW resolution and sampling frequency of a relative pointer device.
See @ref motion_normalization for details.
</dd>
<dt>MOUSE_WHEEL_CLICK_ANGLE</dt>
<dd>The angle in degrees for each click on a mouse wheel. See
libinput_pointer_get_axis_source() for details.
</dd>
<dt>LIBINPUT_MODEL_*</dt>
<dd><b>This prefix is reserved as private API, do not use.</b> See @ref
model_specific_configuration for details.
</dd>
<dt>LIBINPUT_ATTR_*</dt>
<dd><b>This prefix is reserved as private API, do not use.</b> See @ref
model_specific_configuration for details.
</dd>
</dl>
Below is an example udev rule to assign "seat1" to a device from vendor
0x012a with the model ID of 0x034b.
@code
ACTION=="add|change", KERNEL=="event[0-9]*", ENV{ID_VENDOR_ID}=="012a", \
ENV{ID_MODEL_ID}=="034b", ENV{ID_SEAT}="seat1"
@endcode
@section udev_device_type Device type assignment via udev
libinput requires the <b>ID_INPUT</b> property to be set on a device,
otherwise the device will be ignored. In addition, one of <b>
ID_INPUT_KEYBOARD, ID_INPUT_KEY, ID_INPUT_MOUSE, ID_INPUT_TOUCHPAD,
ID_INPUT_TOUCHSCREEN, ID_INPUT_TABLET, ID_INPUT_JOYSTICK,
ID_INPUT_ACCELEROMETER</b> must be set on the device to determine the
device type. The usual error handling applies within libinput and a device
type label does not guarantee that the device is initialized by libinput.
If a device fails to meet the requirements for a device type (e.g. a keyboard
labelled as touchpad) the device will not be available through libinput.
Only one device type should be set per device at a type, though libinput can
handle some combinations for historical reasons.
Below is an example udev rule to remove an <b>ID_INPUT_TOUCHPAD</b> setting
and change it into an <b>ID_INPUT_TABLET</b> setting. This rule would apply
for a device with the vendor/model ID of 012a/034b.
@code
ACTION=="add|change", KERNEL=="event[0-9]*", ENV{ID_VENDOR_ID}=="012a", \
ENV{ID_MODEL_ID}=="034b", ENV{ID_INPUT_TOUCHPAD}="", ENV{ID_INPUT_TABLET}="1"
@endcode
@section ignoring_devices Ignoring specific devices
If a device has the <b>LIBINPUT_IGNORE_DEVICE</b> udev property set to any
value but "0", that device is not initialized by libinput. For a context
created with libinput_udev_create_context(), the device is silently ignored
and never shows up. If the device is added with libinput_path_add_device()
to a context created with libinput_path_create_context(), adding the device
will fail and return NULL (see that function's documentation for more
information).
If the property value is exactly "0", then the property is considered unset
and libinput initializes the device normally.
This property should be used for devices that are correctly detected as
input devices (see @ref udev_device_type) but that should not be used by
libinput. It is recommended that devices that should not be handled as input
devices at all unset the <b>ID_INPUT</b> and related properties instead. The
<b>LIBINPUT_IGNORE_DEVICE</b> property signals that only libinput should
ignore this property but other parts of the stack (if any) should continue
treating this device normally.
@section model_specific_configuration Model-specific configuration
libinput reserves the property prefixes <b>LIBINPUT_MODEL_</b> and
<b>LIBINPUT_ATTR_</b> for model-specific configuration. <b>These prefixes
are reserved as private API, do not use.</b>
The effect of these properties may be to enable or disable certain
features on a specific device or set of devices, to change configuration
defaults or any other reason. The effects of setting these properties, the
format of the property and the value of the property are subject to change
at any time.
@subsection model_specific_configuration_x220fw81 Lenovo x220 with touchpad firmware v8.1
The property <b>LIBINPUT_MODEL_LENOVO_X220_TOUCHPAD_FW81</b> may be set by a
user in a local hwdb file. This property designates the touchpad on a Lenovo
x220 with a touchpad firmware version 8.1. When this firmware version is
installed, the touchpad is imprecise. The touchpad device does not send
continuos x/y axis position updates, a behavior also observed on its
successor model, the Lenovo x230 which has the same firmware version. If the
above property is set, libinput adjusts its behavior to better suit this
particular model.
The touchpad firmware version cannot be detected automatically by libinput,
local configuration is required to set this property. Refer to the libinput
model quirks hwdb for instructions.
This property must not be used for any other purpose, no specific behavior
is guaranteed.
@section hwdb Configuring the hwdb
This section outlines how to add an entry to the <a
href="https://www.freedesktop.org/software/systemd/man/hwdb.html">udev
hwdb</a> and reload properties so they are available to libinput.
The hwdb contains a set of match rules that assign udev properties that are
available to libinput when the device is connected and/or libinput is
initialized. This section only describes the hwdb in relation to libinput,
it is not a full documentation on how the hwdb works.
@note **The use of the hwdb by libinput is not part of the public API. It may
change at any time. Once tested, changes to the hwdb must be submitted
upstream.**
@subsection hwdb_querying Querying the hwdb
libinput only uses device nodes in the form of `/dev/input/eventX` where X
is the number of the specific device. Running `libinput debug-events` lists
all devices currently available to libinput and their event node name:
$> sudo libinput debug-events
-event2 DEVICE_ADDED Power Button seat0 default group1 cap:k
-event5 DEVICE_ADDED Video Bus seat0 default group2 cap:k
-event0 DEVICE_ADDED Lid Switch seat0 default group3 cap:S
...
Note the event node name for your device and translate it into a syspath in
the form of `/sys/class/input/eventX`. This path can be supplied to `udevadm
info`
$> udevadm info
P: /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0D:00/input/input0/event0
N: input/event0
E: DEVNAME=/dev/input/event0
E: DEVPATH=/devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0D:00/input/input0/event0
E: ID_INPUT=1
E: ID_INPUT_SWITCH=1
E: MAJOR=13
E: MINOR=64
E: SUBSYSTEM=input
E: TAGS=:power-switch:
E: USEC_INITIALIZED=7167898
Lines starting with `E:` are udev properties available to libinput. For
example, the above device's `ID_INPUT_SWITCH` property will cause libinput
to treat this device as switch device.
@subsection hwdb_reloading Reloading the hwdb
The actual hwdb is stored in binary file on-disk and must be updated
manually whenever a `.hwdb` file changes. This is required both when a user
manually edits the `.hwdb` file but also when the git tree is updated (and
that update causes a hwdb change).
To update the binary file on-disk, run:
sudo udevadm hwdb --update
Then, to trigger a reload of all properties on your device, run:
sudo udevadm trigger /sys/class/input/eventX
Then check with `udevadm info` whether the properties were updated, see @ref
hwdb_querying. If a new property does not appear on the device, use `udevadm
test` to check for error messages by udev and the hwdb (e.g. syntax errors
in the udev rules files).
sudo udevadm test /sys/class/input/eventX
@subsection hwdb_modifying Modifying the hwdb
This section applies to users that need to add, change, or remove a hwdb
entry for their device. Note that **the hwdb is not part of the public API
and may change at any time**. Once a device has been made to work, the
change must be submitted to the libinput bugzilla, see @ref reporting_bugs.
hwdb entries are only applied if a udev rules calls out to the hwdb with the
right match format. libinput ships with a set of rules to query the hwdb,
the different rules are reflected by their prefix. Again, **this is not part
of the public API**. libinput's matches are
composed of a literal "libinput", then either the device name and dmi
modalias, or the device types and the input modalias. Any part of the hwdb
match line can be a glob by using a literal `*`. For example:
libinput:keyboard:input:b0011v*
libinput:touchpad:input:b0003v05ACp*
libinput:touchpad:input:b0003v05ACp021A*
This type of matching is the preferred one for any removable device (USB,
Bluetooth, ...) that can be uniquely identified by the bustype, vendor and
product ID. The first line matches any keyboard device on the
serial bus (0x0011). The second line matches any touchpad device
with a vendor ID of 0x05AC. The third line matches any touchpad device
with a vendor ID of 0x05AC and a product ID of 0x021A. The `input:b...`
string is available in the device's modalias file
`/sys/class/input/eventX/device/modalias`.
In the case of built-in devices that do not have a unique id, we
need to use the host system's dmi information. For example:
libinput:name:*Lid Switch*:dmi:*svnLENOVO:*pvrThinkPadT440*
This match applies to any device with a name **containing** "Led Switch"
that is on a Lenovo T440 system. The dmi modalias is available in
`/sys/class/dmi/id/modalias`. For example, on the T440 matched above, the
modalias is
dmi:bvnLENOVO:bvrGJET72WW(2.22):bd02/21/2014:svnLENOVO:pn3453453678:pvrThinkPadT440s:rvnLENOVO:rn2012345223:rvrABC123KKK123B:cvnLENOVO:ct10:cvrNotAvailable:
The dmi should always be shortened to the sections that uniquely identify
the type system, in this case the vendor (svn) and the product version (pvr).
Failing to shorten the system may mean that the hwdb match rule only applies
to your specific system, rather than all systems of that type, or that the
rule no longer applies after a firmware update.
The hwdb match string is the first portion of the hwdb entry. The second
portion is the property to set. Each hwdb entry may match on multiple
devices and may apply multiple properties. For example:
libinput:touchpad:input:b0003v05ACp*
THIS_IS_AN_APPLE_DEVICE=1
libinput:touchpad:input:b0003v05ACp*
libinput:name:*Lid Switch*:dmi:*svnLENOVO:*pvrThinkPadT440*
FOO=1
BAR=2
In the example above, any matching touchpad device will have all three
properties applied, the lid switch device only has FOO and BAR.
The hwdb does not allow removing properties. Where a property must be unset,
it should be set to 0.
Any user-specific hwdb entries should be placed in a file
`/etc/udev/hwdb.d/99-somename.hwdb`. See @ref hwdb_reloading for
instructions on how to reload the hwdb once the file is in place.
*/

223
doc/faqs.dox Normal file
View file

@ -0,0 +1,223 @@
/**
@page faq FAQs - Frequently Asked Questions
Frequently asked questions about libinput.
@tableofcontents
@section faq_feature Why doesn't libinput support ...?
First, read @ref what_is_libinput If you have a feature that you think
libinput needs to support, please file a bug report. See @ref reporting_bugs
for more details.
@section faq_fast_mouse My mouse moves too fast, even at the slowest setting
This is a symptom of high-dpi mice (greater than 1000dpi). These devices
need a udev hwdb entry to normalize their motion. See @ref
motion_normalization for a detailed explanation.
@section faq_fast_trackpoint My trackpoint moves too slow or too fast
This is a symptom of an invalid trackpoint range. These devices need a udev
hwdb entry to specify the range available so libinput can adjust the pointer
acceleration accordingly. See @ref trackpoint_range for a detailed explanation.
@section faq_enable_tapping Why isn't touchpad tap-to-click enabled by default
See @ref tapping_default
@section faq_touchpad_pressure Why does my touchpad lose track of touches
The most common cause for this is an incorrect pressure threshold range.
See @ref touchpad_pressure for more info.
@section faq_kinetic_scrolling Kinetic scrolling does not work
The X.Org synaptics driver implemented kinetic scrolling in the driver. It
measures the scroll speed and once the finger leaves the touchpad the driver
keeps sending scroll events for a predetermined time. This effectively
provides for kinetic scrolling without client support but triggers an
unfixable [bug](https://bugs.freedesktop.org/show_bug.cgi?id=38909): the
client cannot know that the events are from a kinetic scroll source. Scroll
events in X are always sent to the current cursor position, a movement of the
cursor after lifting the finger will send the kinetic scroll events to the
new client, something the user does not usually expect. A key event during
the kinetic scroll procedure causes side-effects such as triggering zoom.
libinput does not implement kinetic scrolling for touchpads. Instead it
provides the libinput_event_pointer_get_axis_source() function that enables
callers to implement kinetic scrolling on a per-widget basis, see @ref
scroll_sources.
@section faq_gpl Is libinput GPL-licensed?
No, libinput is MIT licensed. The Linux kernel header file linux/input.h in
libinput's tree is provided to ensure the same behavior regardless of which
kernel version libinput is built on. It does not make libinput GPL-licensed.
@section faq_config_options Where is the configuration stored?
libinput does not store configuration options, it is up to the caller to
manage these and decide which configuration option to apply to each device.
This must be done at startup, after a resume and whenever a new device is
detected.
One commonly used way to configure libinput is to have the Wayland
compositor expose a compositor-specific configuration option. For example,
in a GNOME stack, the gnome-control-center modifies dconf entries. These
changes are read by mutter and applied to libinput. Changing these entries
via the gsettings commandline tool has the same effect.
Another commonly used way to configure libinput is to have xorg.conf.d
snippets. When libinput is used with the xf86-input-libinput driver in an
X.Org stack, these options are read on startup and apply to each device.
Changing properties at runtime with the xinput commandline tool has the same
effect.
In both cases, the selection of available options and how they are exposed
depends on the libinput caller (e.g. mutter or xf86-input-libinput).
@dotfile libinput-stack-gnome.gv
This has an effect on the availability of configuration options: if an
option is not exposed by the intermediary, it cannot be configured by the
client. Also some configuration options that are provided by the
intermediary may not be libinput-specific configuration options.
@section faq_configure_wayland How do I configure my device on Wayland?
See @ref faq_config_options Use the configuration tool provided by your
desktop environment (e.g. gnome-control-center) or direct access to your
desktop environment's configuration storage (e.g. gsettings).
@section faq_configure_xorg How do I configure my device on X?
See @ref faq_config_options If your desktop environment does not provide a
graphical configuration tool you can use an
<a href="https://www.x.org/archive/current/doc/man/man5/xorg.conf.5.xhtml">xorg.conf.d snippet</a>.
Usually, such a snippet looks like this:
<pre>
$> cat /etc/X11/xorg.conf.d/99-libinput-custom-config.conf
Section "InputClass"
Identifier "something to identify this snippet"
MatchDriver "libinput"
MatchProduct "substring of the device name"
Option "some option name" "the option value"
EndSection
</pre>
The identifier is merely a human-readable string that shows up in the log
file. The MatchProduct line should contain the device name or a substring of
the device name that the snippet should apply to. For a full list of option
names and permitted values, see the
<a href="https://www.mankier.com/4/libinput">libinput man page</a>.
xorg.conf.d snippets like the above apply to hotplugged devices but can be
overwritten at runtime by desktop tools. Multiple snippets may be placed
into the same file.
For run-time configuration and testing, the
<a href="https://www.x.org/archive/X11R7.5/doc/man/man1/xinput.1.html">xinput</a>
debugging tool can modify a devices' properties. See the
<a href="https://www.mankier.com/4/libinput">libinput man page</a>
for supported property names and values. Usually, an invocation looks like
this:
<pre>
$> xinput set-prop "the device name" "the property name" value [value2] [value3]
</pre>
@note
Changes performed by xinput do not persist across device hotplugs. xinput is
considered a debugging and testing tool only and should not be used for
permanent configurations.
@section faq_configuration Can you add a configuration option for $FEATURE?
No. At least that's going to be the initial answer. Read <a
href="http://who-t.blogspot.com/2016/04/why-libinput-doesnt-have-lot-of-config.html">Why
libinput doesn't have a lot of configuration options</a> first.
Configuration options for most features are a signal that we are incapable
of handling it correctly. To get to that point, we want to be sure we're
truly incapable of doing so. libinput has several features that
are handled automatically (and correctly) that users wanted to have
configuration options for initially.
So the answer to this question will almost always be 'no'. A configuration
option is, in most cases, a cop-out.
@section faq_synclient Why don't synclient and syndaemon work with libinput?
Synclient and syndaemon rely on X input device properties that are specific
to the xf86-input-synaptics X.Org input driver. Both were written when the
synaptics driver was the only commmon touchpad driver in existence. They
assume that if the properties aren't available, no touchpad is available
either. The xf86-input-libinput X.Org input driver does not export these
driver-specific properties, synclient/syndaemon will thus not detect the
touchpad and refuse to work. Other tools that rely on synclient/syndaemon or
those same properties also do not work with xf86-input-libinput.
Most of syndaemon's functionality is built into libinput, see @ref
disable-while-typing. synclient is merely a configuration tool, see
@ref faq_configure_xorg for similar functionality.
See also the blog posts
<a
href="http://who-t.blogspot.com.au/2017/01/the-definitive-guide-to-synclient.html">The
definitive guide to synclient</a> and
<a href="http://who-t.blogspot.com.au/2016/12/the-future-of-xinput-xmodmap-setxkbmap.html">
The future of xinput, xmodmap, setxkbmap, xsetwacom and other tools under
Wayland</a>
@section faq_tablets Does libinput support non-Wacom tablets?
Yes, though unfortunately many non-Wacom tablets suffer from bad firmware
and don't send the required events. But they should all work nonetheless. If
you have a tablet that does not work with libinput, please @ref
reporting_bugs "file a bug".
@section faq_tablet_capabilities My tablet doesn't work
If you see the message
<pre>
libinput bug: device does not meet tablet criteria. Ignoring this device.
</pre>
or the message
<pre>
missing tablet capabilities [...] Ignoring this device.
</pre>
your tablet device does not have the required capabilities to be treated as
a tablet. This is usually a problem with the device and the kernel driver.
See @ref tablet-capabilities for more details.
@section faq_hwdb_changes How to apply hwdb changes
Sometimes users are asked to test updates to the <a
href="https://www.freedesktop.org/software/systemd/man/hwdb.html">udev
hwdb</a> or patches that include a change to the hwdb. See @ref hwdb for
details on the hwdb and how to modify it locally.
@section faq_timer_offset What causes the "timer offset negative" warning?
libinput relies on the caller to call libinput_dispatch() whenever data is
available on the epoll-fd. Doing so will process the state of all devices
and can trigger some timers to be set (e.g. palm detection, tap-to-click,
disable-while-typing, etc.). Internally, libinput's time offsets are always
based on the event time of the triggering event.
For example, a touch event with time T may trigger a timer for the time T +
180ms. When setting a timer, libinput checks the wall clock time to ensure
that this time T + offset is still in the future. If not, the warning is
logged.
When this warning appears, it simply means that too much time has passed
between the event occurring (and the epoll-fd triggering) and the current
time. In almost all cases this is an indication of the caller being
overloaded and not handling events as speedily as required.
The warning has no immediate effect on libinput's behavior but some of the
functionality that relies on the timer may be impeded (e.g. palms are not
detected as they should be).
*/

129
doc/gestures.dox Normal file
View file

@ -0,0 +1,129 @@
/**
@page gestures Gestures
libinput supports basic gestures on touchpads and other indirect input
devices. Two types of gestures are supported: @ref gestures_pinch and @ref
gestures_swipe. Support for gestures depends on the hardware device, most
touchpads support both gestures and any device that may send gesture events
has the @ref LIBINPUT_DEVICE_CAP_GESTURE capability set.
Note that libinput **does not** support gestures on touchscreens, see
@ref gestures_touchscreens.
@section gestures_lifetime Lifetime of a gesture
A gesture's lifetime has three distinct stages: begin, update and end, each
with their own event types. Begin is sent when the fingers are first set
down or libinput decides that the gesture begins. For @ref gestures_pinch
this sets the initial scale. Any events changing properties of the gesture
are sent as update events. On termination of the gesture, an end event is
sent.
A gesture includes the finger count (see
libinput_event_gesture_get_finger_count()) and that finger count remains the
same for the lifetime of a gesture. Thus, if a user puts down a fourth
finger during a three-finger swipe gesture, libinput will end
the three-finger gesture and, if applicable, start a four-finger swipe
gesture. A caller may decide that those gestures are semantically identical
and continue the two gestures as one single gesture.
@see LIBINPUT_EVENT_GESTURE_PINCH_BEGIN
@see LIBINPUT_EVENT_GESTURE_PINCH_UPDATE
@see LIBINPUT_EVENT_GESTURE_PINCH_END
@see LIBINPUT_EVENT_GESTURE_PINCH_BEGIN
@see LIBINPUT_EVENT_GESTURE_SWIPE_UPDATE
@see LIBINPUT_EVENT_GESTURE_SWIPE_END
@section gestures_pinch Pinch gestures
Pinch gestures are executed when two or more fingers are located on the
touchpad and are either changing the relative distance to each other
(pinching) or are changing the relative angle (rotate). Pinch gestures may
change both rotation and distance at the same time. For such gestures,
libinput calculates a logical center for the gestures and provides the
caller with the delta x/y coordinates of that center, the relative angle of
the fingers compared to the previous event, and the absolute scale compared
to the initial finger position.
@image html pinch-gestures.svg "The pinch and rotate gestures"
The illustration above shows a basic pinch in the left image and a rotate in
the right angle. Not shown is a movement of the logical center if the
fingers move unevenly. Such a movement is supported by libinput, it is
merely left out of the illustration.
Note that while position and angle is relative to the previous event, the
scale is always absolute and a multiplier of the initial finger position's
scale.
@section gestures_swipe Swipe gestures
Swipe gestures are executed when three or more fingers are moved
synchronously in the same direction. libinput provides x and y coordinates
in the gesture and thus allows swipe gestures in any direction, including
the tracing of complex paths. It is up to the caller to interpret the
gesture into an action or limit a gesture to specific directions only.
@image html swipe-gestures.svg "The swipe gestures"
The illustration above shows a vertical three-finger swipe. The coordinates
provided during the gesture are the movements of the logical center.
@section gestures_touchscreens Touchscreen gestures
Touchscreen gestures are **not** interpreted by libinput. Rather, any touch
point is passed to the caller and any interpretation of gestures is up to
the caller or, eventually, the X or Wayland client.
Interpreting gestures on a touchscreen requires context that libinput does
not have, such as the location of windows and other virtual objects on the
screen as well as the context of those virtual objects:
@image html touchscreen-gestures.svg "Context-sensitivity of touchscreen gestures"
In this example, the finger movements are identical but in the left case
both fingers are located within the same window, thus suggesting an attempt
to zoom. In the right case both fingers are located on a window border,
thus suggesting a window movement. libinput only has knowledge of the finger
coordinates (and even then only in device coordinates, not in screen
coordinates) and thus cannot differentiate the two.
@section gestures_softbuttons Gestures with enabled software buttons
If the touchpad device is a @ref touchpads_buttons_clickpads "Clickpad", it
is recommended that a caller switches to @ref clickfinger.
Usually fingers placed in a @ref software_buttons "software button area" is not
considered for gestures, resulting in some gestures to be interpreted as
pointer motion or two-finger scroll events.
@image html pinch-gestures-softbuttons.svg "Interference of software buttons and pinch gestures"
In the example above, the software button area is highlighted in red. The
user executes a three-finger pinch gesture, with the thumb remaining in the
software button area. libinput ignores fingers within the software button
areas, the movement of the remaining fingers is thus interpreted as a
two-finger scroll motion.
@section gestures_twofinger_touchpads Gestures on two-finger touchpads
As of kernel 4.2, many @ref touchpads_touch_partial_mt provide only two
slots. This affects how gestures can be interpreted. Touchpads with only two
slots can identify two touches by position but can usually tell that there
is a third (or fourth) finger down on the touchpad - without providing
positional information for that finger.
Touchpoints are assigned in sequential order and only the first two touch
points are trackable. For libinput this produces an ambiguity where it is
impossible to detect whether a gesture is a pinch gesture or a swipe gesture
whenever a user puts the index and middle finger down first. Since the third
finger does not have positional information, it's location cannot be
determined.
@image html gesture-2fg-ambiguity.svg "Ambiguity of three-finger gestures on two-finger touchpads"
The image above illustrates this ambiguity. The index and middle finger are
set down first, the data stream from both finger positions looks identical.
In this case, libinput assumes the fingers are in a horizontal arrangement
(the right image above) and use a swipe gesture.
*/

View file

@ -7,13 +7,11 @@ OPTIMIZE_OUTPUT_FOR_C = YES
EXTRACT_ALL = YES
EXTRACT_STATIC = YES
MAX_INITIALIZER_LINES = 0
WARNINGS = YES
QUIET = YES
INPUT = "@builddir@"
IMAGE_PATH = "@builddir@"
OUTPUT_DIRECTORY = doc
INPUT = @INPUT@
IMAGE_PATH = "@top_srcdir@/doc/svg" \
"@top_srcdir@/doc/dot"
GENERATE_HTML = YES
HTML_OUTPUT = html
SEARCHENGINE = NO
USE_MATHJAX = YES
MATHJAX_RELPATH = https://cdn.mathjax.org/mathjax/latest
@ -22,13 +20,11 @@ MACRO_EXPANSION = YES
EXPAND_ONLY_PREDEF = YES
PREDEFINED = LIBINPUT_ATTRIBUTE_PRINTF(f, a)= \
LIBINPUT_ATTRIBUTE_DEPRECATED
DOTFILE_DIRS = "@builddir@"
EXAMPLE_PATH = "@builddir@"
SHOW_NAMESPACES = NO
HAVE_DOT = YES
DOTFILE_DIRS = "@top_srcdir@/doc/dot"
HTML_HEADER = "@builddir@/header.html"
HTML_FOOTER = "@builddir@/footer.html"
HTML_EXTRA_STYLESHEET = "@builddir@/bootstrap.css" \
"@builddir@/customdoxygen.css" \
"@builddir@/libinputdoxygen.css"
HTML_HEADER = "@top_srcdir@/doc/style/header.html"
HTML_FOOTER = "@top_srcdir@/doc/style/footer.html"
HTML_EXTRA_STYLESHEET = "@top_srcdir@/doc/style/bootstrap.css" \
"@top_srcdir@/doc/style/customdoxygen.css" \
"@top_srcdir@/doc/style/libinputdoxygen.css"
USE_MDFILE_AS_MAINPAGE = "@top_srcdir@/README.md"

View file

@ -1,8 +1,5 @@
.. _middle_button_emulation:
==============================================================================
Middle button emulation
==============================================================================
/**
@page middle_button_emulation Middle button emulation
Middle button emulation provides users with the ability to generate a middle
click even when the device does not have a physical middle button available.
@ -19,7 +16,7 @@ appropriate to the physical device.
The middle button emulation behavior when combined with other device
buttons, including a physical middle button is device-dependent.
For example, :ref:`clickpad_softbuttons` provides a middle button area when
For example, @ref clickpad_softbuttons provides a middle button area when
middle button emulation is disabled. That middle button area disappears
when middle button emulation is enabled - a middle click can then only be
triggered by a simultaneous left + right click.
@ -29,7 +26,9 @@ enabling/disabling that emulation. Likewise, some devices may allow middle
button emulation but have it disabled by default. This is the case for most
mouse-like devices where a middle button is detected.
libinput provides **libinput_device_config_middle_emulation_set_enabled()** to
enable or disable middle button emulation. See :ref:`faq_configure_wayland`
and :ref:`faq_configure_xorg` for info on how to enable or disable middle
libinput provides libinput_device_config_middle_emulation_set_enabled() to
enable or disable middle button emulation. See @ref faq_configure_wayland
and @ref faq_configure_xorg for info on how to enable or disable middle
button emulation in the Wayland compositor or the X stack.
*/

View file

@ -1,8 +1,5 @@
.. _motion_normalization:
==============================================================================
Normalization of relative motion
==============================================================================
/**
@page motion_normalization Normalization of relative motion
Most relative input devices generate input in so-called "mickeys". A
mickey is in device-specific units that depend on the resolution
@ -27,20 +24,11 @@ libinput applies a dpi-dependent acceleration function. At low speeds, a
movement speed increases, acceleration is applied - at high speeds a low-dpi
device will roughly feel the same as a higher-dpi mouse.
The reason for the normalization is convenience: a caller can assume that a
delta of 1 should result in a movement of 1 pixel on a traditional
(low-dpi) screen. On screens with high resolutions, the caller must scale
according to the UI scale factors.
This normalization only applies to accelerated coordinates, unaccelerated
coordinates are left in device-units. It is up to the caller to interpret
those coordinates correctly.
.. _motion_normalization_touchpad:
------------------------------------------------------------------------------
Normalization of touchpad coordinates
------------------------------------------------------------------------------
@section motion_normalization_touchpad Normalization of touchpad coordinates
Touchpads may have a different resolution for the horizontal and vertical
axis. Interpreting coordinates from the touchpad without taking resolution
@ -48,50 +36,49 @@ into account results in uneven motion.
libinput scales unaccelerated touchpad motion to the resolution of the
touchpad's x axis, i.e. the unaccelerated value for the y axis is:
``y = (x / resolution_x) * resolution_y``.
y = (x / resolution_x) * resolution_y
.. _motion_normalization_tablet:
@section motion_normalization_tablet Normalization of tablet coordinates
------------------------------------------------------------------------------
Normalization of tablet coordinates
------------------------------------------------------------------------------
See @ref tablet-relative-motion
See :ref:`tablet-relative-motion`
.. _motion_normalization_customization:
------------------------------------------------------------------------------
Setting custom DPI settings
------------------------------------------------------------------------------
@section motion_normalization_customization Setting custom DPI settings
Devices usually do not advertise their resolution and libinput relies on
the udev property **MOUSE_DPI** for this information. This property is usually
set via the
`udev hwdb <http://cgit.freedesktop.org/systemd/systemd/tree/hwdb/70-mouse.hwdb>`_.
The ``mouse-dpi-tool`` utility provided by
`libevdev <https://freedesktop.org/wiki/Software/libevdev/>`_ should be
the udev property <b>MOUSE_DPI</b> for this information. This property is usually
set via the <a
href="http://cgit.freedesktop.org/systemd/systemd/tree/hwdb/70-mouse.hwdb">udev hwdb</a>.
The "mouse-dpi-tool" utility provided by <a
href="https://freedesktop.org/wiki/Software/libevdev/">libevdev</a> should be
used to measure a device's resolution.
The format of the property for single-resolution mice is: ::
MOUSE_DPI=resolution@frequency
The format of the property for single-resolution mice is:
@code
MOUSE_DPI=resolution@frequency
@endcode
The resolution is in dots per inch, the frequency in Hz.
The format of the property for multi-resolution mice may list multiple
resolutions and frequencies: ::
MOUSE_DPI=r1@f1 *r2@f2 r3@f3
resolutions and frequencies:
@code
MOUSE_DPI=r1@f1 *r2@f2 r3@f3
@endcode
The default frequency must be pre-fixed with an asterisk.
For example, these two properties are valid: ::
For example, these two properties are valid:
@code
MOUSE_DPI=800@125
MOUSE_DPI=400@125 800@125 *1000@500 5500@500
@endcode
MOUSE_DPI=800@125
MOUSE_DPI=400@125 800@125 *1000@500 5500@500
The behavior for a malformed property is undefined.
The behavior for a malformed property is undefined. If the property is
unset, libinput assumes the resolution is 1000dpi.
If the property is unset, libinput assumes the resolution is 1000dpi.
Note that HW does not usually provide information about run-time
resolution changes, libinput will thus not detect when a resolution
changes to the non-default value.
*/

57
doc/page-hierarchy.dox Normal file
View file

@ -0,0 +1,57 @@
/**
@page misc Users
- @subpage what_is_libinput
- @subpage faq
- @subpage tools
- @subpage reporting_bugs
@page touchpads Touchpads
- @subpage scrolling
- @subpage clickpad_softbuttons
- @subpage tapping
- @subpage gestures
- @subpage touchpad_pressure
- @subpage palm_detection
- @subpage t440_support
- @subpage touchpad_jumping_cursor
- @subpage absolute_coordinate_ranges
- @subpage touchpad_jitter
@page touchscreens Touchscreens
- @subpage absolute_axes
@page pointers Mice, Trackballs, etc.
- @subpage motion_normalization
- @subpage middle_button_emulation
- @subpage button_debouncing
- @subpage trackpoints
@page tablets Graphics Tablets
- @subpage tablet-support
@page general General
- @subpage udev_config
- @subpage seats
- @subpage timestamps
@page developers Developers
Contributions to libinput are always welcome. See the links below for
specific documentation.
- @subpage what_is_libinput
- @subpage contributing
- @subpage architecture
- @subpage building_libinput
- @subpage test-suite
- @subpage tools
- @subpage pointer-acceleration
*/

View file

@ -1,8 +1,5 @@
.. _palm_detection:
==============================================================================
Palm detection
==============================================================================
/**
@page palm_detection Palm detection
Palm detection tries to identify accidental touches while typing, while
using the trackpoint and/or during general use of the touchpad area.
@ -21,34 +18,25 @@ touchpads are less affected by palm touches.
libinput has multiple ways of detecting a palm, each of which depends on
hardware-specific capabilities.
- :ref:`palm_tool`
- :ref:`palm_pressure`
- :ref:`palm_touch_size`
- :ref:`palm_exclusion_zones`
- :ref:`trackpoint-disabling`
- :ref:`disable-while-typing`
- :ref:`disable-while-trackpointing`
- :ref:`stylus-touch-arbitration`
- @ref palm_tool
- @ref palm_pressure
- @ref palm_touch_size
- @ref palm_exclusion_zones
- @ref trackpoint-disabling
- @ref disable-while-typing
- @ref stylus-touch-arbitration
Palm detection is always enabled, with the exception of
disable-while-typing.
.. _palm_tool:
------------------------------------------------------------------------------
Palm detection based on firmware labelling
------------------------------------------------------------------------------
@section palm_tool Palm detection based on firmware labelling
Some devices provide palm detection in the firmware, forwarded by the kernel
as the ``EV_ABS/ABS_MT_TOOL`` axis with a value of ``MT_TOOL_PALM``
as the ```EV_ABS/ABS_MT_TOOL``` axis with a value of ```MT_TOOL_PALM```
(whenever a palm is detected). libinput honors that value and switches that
touch to a palm.
.. _palm_pressure:
------------------------------------------------------------------------------
Palm detection based on pressure
------------------------------------------------------------------------------
@section palm_pressure Palm detection based on pressure
The simplest form of palm detection labels a touch as palm when the pressure
value goes above a certain threshold. This threshold is usually high enough
@ -58,27 +46,19 @@ the threshold again. This ensures that a palm remains a palm even when the
pressure changes as the user is typing.
For some information on how to detect pressure on a touch and debug the
pressure ranges, see :ref:`touchpad_pressure`.
pressure ranges, see @ref touchpad_pressure.
.. _palm_touch_size:
@section palm_touch_size Palm detection based on touch size
------------------------------------------------------------------------------
Palm detection based on touch size
------------------------------------------------------------------------------
On touchpads that support the ``ABS_MT_TOUCH_MAJOR`` axes, libinput can perform
On touchads that support the `ABS_MT_TOUCH_MAJOR` axes, libinput can perform
palm detection based on the size of the touch ellipse. This works similar to
the pressure-based palm detection in that a touch is labelled as palm when
it exceeds the (device-specific) touch size threshold.
For some information on how to detect the size of a touch and debug the
touch size ranges, see :ref:`touchpad_pressure`.
touch size ranges, see @ref touchpad_pressure.
.. _palm_exclusion_zones:
------------------------------------------------------------------------------
Palm exclusion zones
------------------------------------------------------------------------------
@section palm_exclusion_zones Palm exclusion zones
libinput enables palm detection on the left, right and top edges of the
touchpad. Two exclusion zones are defined on the left and right edge of the
@ -92,8 +72,9 @@ Another exclusion zone is defined on the top edge of the touchpad. As with
the edge zones, libinput detects vertical movements out of the edge zone and
avoids palm detection on such touch sequences.
A touch starting in the exclusion zone does not trigger a tap (see
:ref:`tapping`).
Each side edge exclusion zone is divided into a top part and a bottom part.
A touch starting in the top part of the exclusion zone does not trigger a
tap (see @ref tapping).
In the diagram below, the exclusion zones are painted red.
Touch 'A' starts inside the exclusion zone and moves
@ -103,20 +84,17 @@ despite moving out of the exclusion zone.
Touch 'B' starts inside the exclusion zone but moves horizontally out of the
zone. It is considered a valid touch and controls the cursor.
Touch 'C' occurs in the exclusion zone. Despite being a tapping motion, it does
not generate an emulated button event.
Touch 'C' occurs in the top part of the exclusion zone. Despite being a
tapping motion, it does not generate an emulated button event. Touch 'D'
likewise occurs within the exclusion zone but in the bottom half. libinput
will generate a button event for this touch.
.. figure:: palm-detection.svg
:align: center
@image html palm-detection.svg
.. _trackpoint-disabling:
@section trackpoint-disabling Palm detection during trackpoint use
------------------------------------------------------------------------------
Palm detection during trackpoint use
------------------------------------------------------------------------------
If a device provides a
`trackpoint <http://en.wikipedia.org/wiki/Pointing_stick>`_, it is
If a device provides a <a
href="http://en.wikipedia.org/wiki/Pointing_stick">trackpoint</a>, it is
usually located above the touchpad. This increases the likelihood of
accidental touches whenever the trackpoint is used.
@ -125,64 +103,36 @@ certain timeout until after trackpoint activity stops. Touches generated
during this timeout will not move the pointer, and touches started during
this timeout will likewise not move the pointer (allowing for a user to rest
the palm on the touchpad while using the trackstick).
If the touchpad is disabled, the :ref:`top software buttons <t440_support>`
If the touchpad is disabled, the @ref t440_support "top software buttons"
remain enabled.
.. _disable-while-typing:
------------------------------------------------------------------------------
Disable-while-typing
------------------------------------------------------------------------------
@section disable-while-typing Disable-while-typing
libinput automatically disables the touchpad for a timeout after a key
press, a feature traditionally referred to as "disable while typing" and
previously available through the
`syndaemon(1) <http://linux.die.net/man/1/syndaemon>`_ command. libinput does
[syndaemon(1)](http://linux.die.net/man/1/syndaemon) command. libinput does
not require an external command and the feature is currently enabled for all
touchpads but will be reduced in the future to only apply to touchpads where
finger width or pressure data is unreliable.
Notable behaviors of libinput's disable-while-typing feature:
- Two different timeouts are used, after a single key press the timeout is
short to ensure responsiveness. After multiple key events, the timeout is
longer to avoid accidental pointer manipulation while typing.
- Some keys do not trigger the timeout, specifically some modifier keys
- Some keys do not trigger the timeout, specifically some modifier keys
(Ctrl, Alt, Shift, and Fn). Actions such as Ctrl + click thus stay
responsive.
- Touches started while typing do not control the cursor even after typing
has stopped, it is thus possible to rest the palm on the touchpad while
typing.
- Physical buttons work even while the touchpad is disabled. This includes
:ref:`software-emulated buttons <t440_support>`.
- libinput pairs touchpads and keyboards for the disable-while-typing
feature. In the most common case, the internal touchpad is paired only
with the internal keyboard. Typing on an external keyboard will thus not
disable the touchpad. Some devices require a :ref:`quirk <device-quirks>`
to be correctly paired.
@ref t440_support "software-emulated buttons".
Disable-while-typing can be enabled and disabled by calling
**libinput_device_config_dwt_set_enabled()**.
libinput_device_config_dwt_set_enabled().
.. _disable-while-trackpointing:
------------------------------------------------------------------------------
Disable-while-trackpointing
------------------------------------------------------------------------------
libinput automatically disables the touchpad for a timeout after the trackpoint
is moved, a feature referred to as "disable while trackpointing". libinput does
not require an external command and the feature is currently enabled for all
touchpads.
Disable-while-trackpointing can be enabled and disabled by calling
**libinput_device_config_dwtp_set_enabled()**.
.. _stylus-touch-arbitration:
------------------------------------------------------------------------------
Stylus-touch arbitration
------------------------------------------------------------------------------
@section stylus-touch-arbitration Stylus-touch arbitration
A special case of palm detection is touch arbitration on devices that
support styli. When interacting with a stylus on the screen, parts of the
@ -190,38 +140,35 @@ hand may touch the surface and trigger touches. As the user is currently
interacting with the stylus, these touches would interfer with the correct
working of the stylus.
libinput employs a method similar to :ref:`disable-while-typing` to detect
libinput employs a method similar to @ref disable-while-typing to detect
these touches and disables the touchpad accordingly.
.. _thumb-detection:
------------------------------------------------------------------------------
Thumb detection
------------------------------------------------------------------------------
@section thumb-detection Thumb detection
Many users rest their thumb on the touchpad while using the index finger to
move the finger around. For clicks, often the thumb is used rather than the
finger. The thumb should otherwise be ignored as a touch, i.e. it should not
count towards :ref:`clickfinger` and it should not cause a single-finger
movement to trigger :ref:`twofinger_scrolling`.
count towards @ref clickfinger and it should not cause a single-finger
movement to trigger @ref twofinger_scrolling.
libinput uses two triggers for thumb detection: pressure and
location. A touch exceeding a pressure threshold is considered a thumb if it
is within the thumb detection zone.
.. note:: "Pressure" on touchpads is synonymous with "contact area." A large touch
surface area has a higher pressure and thus hints at a thumb or palm
touching the surface.
@note "Pressure" on touchpads is synonymous with "contact area", a large
touch surface area has a higher pressure and thus hints at a thumb or palm
touching the surface.
Pressure readings are unreliable at the far bottom of the touchpad as a
thumb hanging mostly off the touchpad will have a small surface area.
libinput has a definitive thumb zone where any touch is considered a resting
thumb.
.. figure:: thumb-detection.svg
:align: center
@image html thumb-detection.svg
The picture above shows the two detection areas. In the larger (light red)
area, a touch is labelled as thumb when it exceeds a device-specific
pressure threshold. In the lower (dark red) area, a touch is labelled as
thumb if it remains in that area for a time without moving outside.
*/

View file

@ -0,0 +1,156 @@
/**
@page pointer-acceleration Pointer acceleration
libinput uses device-specific pointer acceleration methods, with the default
being the @ref ptraccel-linear. The methods share common properties, such as
@ref ptraccel-velocity.
This page explains the high-level concepts used in the code. It aims to
provide an overview for developers and is not necessarily useful for
users.
@section ptraccel-profiles Pointer acceleration profiles
The profile decides the general method of pointer acceleration.
libinput currently supports two profiles: "adaptive" and "flat". The aptive
profile is the default profile for all devices and takes the current speed
of the device into account when deciding on acceleration. The flat profile
is simply a constant factor applied to all device deltas, regardless of the
speed of motion (see @ref ptraccel-profile-flat). Most of this document
describes the adaptive pointer acceleration.
@section ptraccel-velocity Velocity calculation
The device's speed of movement is measured across multiple input events
through so-called "trackers". Each event prepends a the tracker item, each
subsequent tracker contains the delta of that item to the current position,
the timestamp of the event that created it and the cardinal direction of the
movement at the time. If a device moves into the same direction, the
velocity is calculated across multiple trackers. For example, if a device
moves steadily for 10 events to the left, the velocity is calculated across
all 10 events.
Whenever the movement changes direction or significantly changes speed, the
velocity is calculated from the direction/speed change only. For example, if
a device moves steadily for 8 events to the left and then 2 events to the
right, the velocity is only that of the last 2 events.
An extra time limit prevents events that are too old to factor into the
velocity calculation. For example, if a device moves steadily for 5 events
to the left, then pauses, then moves again for 5 events to the left, only
the last 5 events are used for velocity calculation.
The velocity is then used to calculate the acceleration factor
@section ptraccel-factor Acceleration factor
The acceleration factor is the final outcome of the pointer acceleration
calculations. It is a unitless factor that is applied to the current delta,
a factor of 2 doubles the delta (i.e. speeds up the movement), a factor of
less than 1 reduces the delta (i.e. slows the movement).
Any factor less than 1 requires the user to move the device further to move
the visible pointer. This is called deceleration and enables high precision
target selection through subpixel movements. libinput's current maximum
deceleration factor is 0.3 (i.e. slow down to 30% of the pointer speed).
A factor higher than 1 moves the pointer further than the physical device
moves. This is acceleration and allows a user to cross the screen quickly
but effectively skips pixels. libinput's current maximum acceleration factor
is 3.5.
@section ptraccel-linear Linear pointer acceleration
The linear pointer acceleration method is the default for most pointer
devices. It provides deceleration at very slow movements, a 1:1 mapping for
regular movements and a linear increase to the maximum acceleration factor
for fast movements.
Linear pointer acceleration applies to devices with above 1000dpi resolution
and after @ref motion_normalization is applied.
@image html ptraccel-linear.svg "Linear pointer acceleration"
The image above shows the linear pointer acceleration settings at various
speeds. The line for 0.0 is the default acceleration curve, speed settings
above 0.0 accelerate sooner, faster and to a higher maximum acceleration.
Speed settings below 0 delay when acceleration kicks in, how soon the
maximum acceleration is reached and the maximum acceleration factor.
Extremely low speed settings provide no acceleration and additionally
decelerate all movement by a constant factor.
@section ptraccel-low-dpi Pointer acceleration for low-dpi devices
Low-dpi devices are those with a physical resolution of less than 1000 dots
per inch (dpi). The pointer acceleration is adjusted to provide roughly the
same feel for all devices at normal to high speeds. At slow speeds, the
pointer acceleration works on device-units rather than normalized
coordinates (see @ref motion_normalization).
@image html ptraccel-low-dpi.svg "Pointer acceleration for low-dpi devices"
The image above shows the default pointer acceleration curve for a speed of
0.0 at different DPI settings. A device with low DPI has the acceleration
applied sooner and with a stronger acceleration factor.
@section ptraccel-touchpad Pointer acceleration on touchpads
Touchpad pointer acceleration uses the same approach as the @ref
ptraccel-linear profile, with a constant deceleration factor applied. The
user expectation of how much a pointer should move in response to finger
movement is different to that of a mouse device, hence the constant
deceleration factor.
@image html ptraccel-touchpad.svg "Pointer acceleration curve for touchpads"
The image above shows the touchpad acceleration profile in comparison to the
@ref ptraccel-linear. The shape of the curve is identical but vertically squashed.
@section ptraccel-trackpoint Pointer acceleration on trackpoints
The main difference between trackpoint hardware and mice or touchpads is
that trackpoint speed is a function of pressure rather than moving speed.
But trackpoint hardware is quite varied in how it reacts to user pressure
and unlike other devices it cannot easily be normalized for physical
properties. Measuring pressure objectively across a variety of hardware is
nontrivial.
libinput's pointer acceleration is a function of the total available
pressure range on a device. See @ref trackpoint_range for details.
libinput relies on some system-wide configured properties, specifically the
@ref udev_config. The property that influences trackpoint acceleration is
`LIBINPUT_ATTR_TRACKPOINT_RANGE` which specifies the total delta range for
the trackpoint. See @ref trackpoint_range for details.
Additionally, some trackpoints provide the ability to adjust the sensitivity in
hardware by modifying a sysfs file on the serio node. The udev property
`POINTINGSTICK_SENSITIVITY` indicates the desired value, a udev
builtin is expected to apply this to the device, i.e. libinput does not
handle this property. Once applied, the sensitivity adjusts the deltas
coming out of the hardware. When the sensitivity changes, the trackpoint
range changes and thus the `LIBINPUT_ATTR_TRACKPOINT_RANGE` property
becomes invalid.
As of version 1.9, libinput does not parse the `POINTINGSTICK_CONST_ACCEL` property anymore.
@image html ptraccel-trackpoint.svg "Pointer acceleration curves for trackpoints"
The image above shows the trackpoint acceleration profile for each input
delta.
@section ptraccel-profile-flat The flat pointer acceleration profile
In a flat profile, the acceleration factor is constant regardless of the
velocity of the pointer and each delta (dx, dy) results in an accelerated delta
(dx * factor, dy * factor). This provides 1:1 movement between the device
and the pointer on-screen.
@section ptraccel-tablet Pointer acceleration on tablets
Pointer acceleration for relative motion on tablet devices is a flat
acceleration, with the speed setting slowing down or speeding up the pointer
motion by a constant factor. Tablets do not allow for switchable profiles.
*/

295
doc/reporting-bugs.dox Normal file
View file

@ -0,0 +1,295 @@
/**
@page reporting_bugs Reporting bugs
A new bug can be filed here:
https://gitlab.freedesktop.org/libinput/libinput
When reporting bugs against libinput, please follow the instructions below
and provide the required data. You will need:
- a reliable @ref reporting_bugs_reproducer "reproducer" for the bug
- an @ref evemu "evemu recording" of the device while the bug is reproduced
- device-specific information, see
- @ref reporting_bugs_touchpad
- @ref reporting_bugs_mouse
- @ref reporting_bugs_keyboard
- @ref reporting_bugs_trackpoint
- @ref reporting_bugs_other
- the @ref reporting_bugs_version "libinput version" you are on.
- the @ref reporting_bugs_options "configuration options" you have set
- a bugzilla account
If you don't have all of the above, provide a reason why not. Unless the
reason is justified, the bug will have low priority. Remember, libinput has
a lot of users but very few developers.
Stay technical, on-topic, and keep the description concise.
@section reporting_bugs_version Obtaining the libinput version
If your libinput version is older than the current stable branch, you will
get asked to try the latest version.
If you run a distribution-provided libinput, use the package manager to get
the **full** package name and version of libinput, e.g.
- `rpm -q libinput`
- `dpkg -s libinput10`
If you run a self-compiled version of libinput provide the git commit you
have built or the tarball name.
As a last resort, use `libinput-list-devices --version`
@section reporting_bugs_reproducer Reproducing bugs
Try to identify the bug by reproducing it reliably. Bugs without a
reliable reproducer will have lowest priority. The more specific a bug
description and reproducer is, the easier it is to fix.
Try to replicate the series of events that lead to the bug being triggered.
Narrow it down until you have a reliable sequence that can trigger the bug.
For the vast majority of bugs you should not take longer than 5 seconds or
three interactions (clicks, touches, taps, ...) with the device to
reproduce. If it takes longer than that, you can narrow it down further.
Once you can reproduce it, use the @ref libinput-debug-events helper tool.
The output is textual and can help identify whether the bug is in libinput
at all. Note that any configuration options you have set must be specified
on the commandline, see the @ref libinput-debug-events
"libinput-debug-events" man page. Use the `--verbose` flag to get more
information about how libinput processes events.
If the bug cannot be reproduced with the @ref libinput-debug-events helper,
even with the correct configuration options set, it is likely not a bug in
libinput.
@section reporting_bugs_options libinput configuration settings
libinput has a number of device-specific default configuration settings that
may differ from the ones your desktop environment picks by default. You may
have changed some options in a settings panel or in an the xorg.conf snippet
yourself.
You must provide these options in the bug report, otherwise a developer
reproducing the issue may not be able to do so.
If you are on X11, the current settings can be can be obtained with
`xinput list-props "your device name"`. Use `xinput list` to
obtain the device name.
If you are on Wayland, provide a manual summary of the options you have
changed from the default (e.g. "I enabled tap-to-click").
@section reporting_bugs_touchpad Reporting touchpad bugs
When you file a bug, please attach the following information:
- a virtual description of your input device, see @ref evemu. This is the
most important piece of information, do not forget it!
- the output from udevadm info, see @ref udev_info.
- the vendor model number of your laptop (e.g. "Lenovo Thinkpad T440s")
- and the content of `/sys/class/dmi/id/modalias`.
- run the `touchpad-edge-detectior` tool (provided by libevdev) and verify
that the ranges and sizes it prints match the touchpad (up to 5mm
difference is ok)
If you are reporting a bug related to button event generation:
- does your touchpad have (separate) physical hardware buttons or is the
whole touchpad clickable?
- Are you using software buttons or clickfinger? See @ref
clickpad_softbuttons.
- Do you have @ref tapping enabled?
@section reporting_bugs_mouse Reporting mouse bugs
When you file a bug, please attach the following information:
- a virtual description of your input device, see @ref evemu. This is the
most important piece of information, do not forget it!
- the vendor model number of the device (e.g. "Logitech M325")
- the output from udevadm info, see @ref udev_info.
If the bug is related to the @ref motion_normalization_customization "speed of the mouse":
- the resolution of the mouse as specified by the vendor (in DPI)
- the output of the `mouse-dpi-tool` (provided by libevdev)
@section reporting_bugs_keyboard Reporting keyboard bugs
Is your bug related to a keyboard layout? libinput does not handle keyboard
layouts and merely forwards the physical key events. File the bug with your
desktop environment instead (e.g. GNOME, KDE, ...), that's most likely where
the issue is.
When you file a bug, please attach the following information:
- a virtual description of your input device, see @ref evemu. This is the
most important piece of information, do not forget it!
@section reporting_bugs_trackpoint Reporting trackpoint bugs
When you file a bug, please attach the following information:
- a virtual description of your input device, see @ref evemu. This is the
most important piece of information, do not forget it!
- the vendor model number of the device (e.g. "Logitech M325")
- the output from udevadm info, see @ref udev_info.
- the output of `libinput measure trackpoint-range`
- the sensitivity of the trackpoint (adjust the event node number as needed):
@verbatim
$ cat /sys/class/input/event17/device/device/sensitivity
@endverbatim
@section reporting_bugs_other All other devices
When you file a bug, please attach the following information:
- a virtual description of your input device, see @ref evemu. This is the
most important piece of information, do not forget it!
- the vendor model number of the device (e.g. "Sony Plastation3 controller")
@section udev_info udev information for the device
In many cases, we require the udev properties assigned to the device to
verify whether device-specific quirks were applied. This can be obtained
with ````udevadm info /sys/class/input/eventX```, with the correct event
node for your device. An example output is below:
@code
$ udevadm info /sys/class/input/event4
P: /devices/platform/i8042/serio1/input/input5/event4
N: input/event4
E: DEVNAME=/dev/input/event4
E: DEVPATH=/devices/platform/i8042/serio1/input/input5/event4
E: EVDEV_ABS_00=::41
E: EVDEV_ABS_01=::37
E: EVDEV_ABS_35=::41
E: EVDEV_ABS_36=::37
E: ID_INPUT=1
E: ID_INPUT_HEIGHT_MM=66
E: ID_INPUT_TOUCHPAD=1
E: ID_INPUT_WIDTH_MM=97
E: MAJOR=13
E: MINOR=68
E: SUBSYSTEM=input
E: USEC_INITIALIZED=5463031
@endcode
@section evemu Recording devices with evemu
@note Where available, the @ref libinput-record tools should be used instead
of evemu
<a href="https://www.freedesktop.org/wiki/Evemu/">evemu</a> records the
device capabilities together with the event stream from the kernel. On our
side, this allows us to recreate a virtual device identical to your device
and re-play the event sequence, hopefully triggering the same bug.
evemu-record takes a <tt>/dev/input/eventX</tt> event node, but without arguments
it will simply show the list of devices and let you select:
@code
$ sudo evemu-record > scroll.evemu
Available devices:
/dev/input/event0: Lid Switch
/dev/input/event1: Sleep Button
/dev/input/event2: Power Button
/dev/input/event3: AT Translated Set 2 keyboard
/dev/input/event4: SynPS/2 Synaptics TouchPad
/dev/input/event5: Video Bus
/dev/input/event6: ELAN Touchscreen
/dev/input/event10: ThinkPad Extra Buttons
/dev/input/event11: HDA Intel HDMI HDMI/DP,pcm=3
/dev/input/event12: HDA Intel HDMI HDMI/DP,pcm=7
/dev/input/event13: HDA Intel HDMI HDMI/DP,pcm=8
/dev/input/event14: HDA Intel PCH Dock Mic
/dev/input/event15: HDA Intel PCH Mic
/dev/input/event16: HDA Intel PCH Dock Headphone
/dev/input/event17: HDA Intel PCH Headphone
/dev/input/event18: Integrated Camera
/dev/input/event19: TPPS/2 IBM TrackPoint
Select the device event number [0-19]:
@endcode
Select the device that triggers the issue, then reproduce the bug and Ctrl+C
the process. The resulting recording, ("scroll.evemu" in this example) will
contain the sequence required to reproduce the bug. If the bug fails to
reproduce during recording, simply Ctrl+C and restart evemu-record.
Always start the recording from a neutral state, i.e. without any buttons or
keys down, with the position of the device in the neutral position, without
touching the screen/touchpad.
@note The longer the recording, the harder it is to identify the event
sequence triggering the bug. Please keep the event sequence as short as possible.
To verify that the recording contains the bug, you can replay it on your
device. For example, to replay the sequence recorded in the example above:
@code
$ sudo evemu-play /dev/input/event4 < scroll.evemu
@endcode
If the bug is triggered by replaying on your device, attach the recording to
the bug report.
@note libinput does not affect the evemu recording. libinput and evemu talk
directly to the kernel's device nodes. An evemu recording is not influenced
by the libinput version or whether a libinput context is currently active.
@dotfile evemu.gv
@section fixed_bugs My bug was closed as fixed, what now?
libinput's policy on closing bugs is: once the fix for a given bug is on git
master, the bug is considered fixed and the bugzilla entry will be closed
accordingly.
Of course, unless you actually run git master, the bug will continue to
affect you on your local machine. You are most likely running the
distribution's package and you will need to wait until the distribution has
updated its package accordingly.
<b>Do not re-open a bug just because it hasn't trickled down to your
distribution's package version yet.</b>
Whether the bug fix ends up in your distribution depends on a number of
things. Any given bug fix **may** be cherry-picked into the current stable
branch, depending on its severity, impact, and likelyhood to cause
regressions. Once cherry-picked it will land in the next stable branch
release. These are usually a few weeks apart.
<b>Do not re-open a bug because it wasn't picked into a stable branch
release or because your distribution didn't update to the latest stable
branch release.</b>
Stable branches are usually discontinued when the next release comes out.
Your distribution may pick a patch up immediately and ship the fix
even before the next stable branch update is released. For example, Fedora
does this frequently.
<b>If a bug needs to be fixed urgently, file a bug in your distribution's
bug tracker.</b>
Patches on git master will end up in the next libinput release. Once your
distribution updates to that release, your local libinput version will
contain the fix.
<b>Do not re-open a bug because your distribution didn't update to the
release.</b>
You can always run libinput from git master (see @ref building_libinput).
Even while in development, libinput is very stable so this option isn't as
scary as it may sounds.
@subsection reporting_bugs_reopen When is it ok to re-open a fixed bug?
Any time the bug was considered fixed but it turns out that the fix is
insufficient and/or causes a regression.
However, if the regression is in behavior unrelated to the fix itself it is
usually better to file a new bug to reduce the noise. For example, if a fix
to improve tapping breaks two-finger scrolling behavior, you should file a
new bug but reference the original bug.
*/

106
doc/scrolling.dox Normal file
View file

@ -0,0 +1,106 @@
/**
@page scrolling Scrolling
libinput supports three different types of scrolling methods: @ref
twofinger_scrolling, @ref edge_scrolling and @ref button_scrolling. Some
devices support multiple methods, though only one can be enabled at a time.
As a general overview:
- touchpad devices with physical buttons below the touchpad support edge and
two-finger scrolling
- touchpad devices without physical buttons (@ref clickpad_softbuttons
"clickpads") support two-finger scrolling only
- pointing sticks provide on-button scrolling by default
- mice and other pointing devices support on-button scrolling but it is not
enabled by default
A device may differ from the above based on its capabilities. See
libinput_device_config_scroll_set_method() for documentation on how to
switch methods and libinput_device_config_scroll_get_methods() for
documentation on how to query a device for available scroll methods.
@section horizontal_scrolling Horizontal scrolling
Scroll movements provide vertical and horizontal directions, each
scroll event contains both directions where applicable, see
libinput_event_pointer_get_axis_value(). libinput does not provide separate
toggles to enable or disable horizontal scrolling. Instead, horizontal
scrolling is always enabled. This is intentional, libinput does not have
enough context to know when horizontal scrolling is appropriate for a given
widget. The task of filtering horizontal movements is up to the caller.
@section twofinger_scrolling Two-finger scrolling
The default on two-finger capable touchpads (almost all modern touchpads are
capable of detecting two fingers). Scrolling is triggered by two fingers
being placed on the surface of the touchpad, then moving those fingers
vertically or horizontally.
@image html twofinger-scrolling.svg "Vertical and horizontal two-finger scrolling"
For scrolling to trigger, a built-in distance threshold has to be met but once
engaged any movement will scroll. In other words, to start scrolling a
sufficiently large movement is required, once scrolling tiny amounts of
movements will translate into tiny scroll movements.
Scrolling in both directions at once is possible by meeting the required
distance thresholds to enable each direction separately.
Two-finger scrolling requires the touchpad to track both touch points with
reasonable precision. Unfortunately, some so-called "semi-mt" touchpads can
only track the bounding box of the two fingers rather than the actual
position of each finger. In addition, that bounding box usually suffers from
a low resolution, causing jumpy movement during two-finger scrolling.
libinput does not provide two-finger scrolling on those touchpads.
@section edge_scrolling Edge scrolling
On some touchpads, edge scrolling is available, triggered by moving a single
finger along the right edge (vertical scroll) or bottom edge (horizontal
scroll).
@image html edge-scrolling.svg "Vertical and horizontal edge scrolling"
Due to the layout of the edges, diagonal scrolling is not possible. The
behavior of edge scrolling using both edges at the same time is undefined.
Edge scrolling overlaps with @ref clickpad_softbuttons. A physical click on
a clickpad ends scrolling.
@section button_scrolling On-Button scrolling
On-button scrolling converts the motion of a device into scroll events while
a designated button is held down. For example, Lenovo devices provide a
<a href="http://en.wikipedia.org/wiki/Pointing_stick">pointing stick</a> that emulates
scroll events when the trackstick's middle mouse button is held down.
@note On-button scrolling is enabled by default for pointing sticks. This
prevents middle-button dragging; all motion events while the middle button is
down are converted to scroll events.
@image html button-scrolling.svg "Button scrolling"
The button may be changed with
libinput_device_config_scroll_set_button() but must be on the same device as
the motion events. Cross-device scrolling is not supported but
for one exception: libinput's @ref t440_support enables the use of the middle
button for button scrolling (even when the touchpad is disabled).
@section scroll_sources Scroll sources
libinput provides a pointer axis *source* for each scroll event. The
source can be obtained with the libinput_event_pointer_get_axis_source()
function and is one of **wheel**, **finger**, or **continuous**. The source
information lets a caller decide when to implement kinetic scrolling.
Usually, a caller will process events of source wheel as they come in.
For events of source finger a caller should calculate the velocity of the
scroll motion and upon finger release start a kinetic scrolling motion (i.e.
continue executing a scroll according to some friction factor).
libinput expects the caller to be in charge of widget handling, the source
information is thus enough to provide kinetic scrolling on a per-widget
basis. A caller should cancel kinetic scrolling when the pointer leaves the
current widget or when a key is pressed.
See the libinput_event_pointer_get_axis_source() for details on the
behavior of each scroll source.
See also http://who-t.blogspot.com.au/2015/03/libinput-scroll-sources.html
*/

View file

@ -1,8 +1,5 @@
.. _seats:
==============================================================================
Seats
==============================================================================
/**
@page seats Seats
Each device in libinput is assigned to one seat.
A seat has two identifiers, the physical name and the logical name. The
@ -13,15 +10,11 @@ seats as independent device sets. Alternatively, a compositor may limit
itself to a single logical seat, leaving a second compositor to manage
devices on the other logical seats.
.. _seats_overview:
------------------------------------------------------------------------------
Overview
------------------------------------------------------------------------------
@section Overview
Below is an illustration of how physical seats and logical seats interact:
.. graphviz:: seats-sketch.gv
@dotfile seats-sketch.gv
The devices "Foo", "Bar" and "Spam" share the same physical seat and are
thus available in the same libinput context. Only "Foo" and "Bar" share the
@ -29,19 +22,15 @@ same logical seat. The device "Egg" is not available in the libinput context
associated with the physical seat 0.
The above graph is for illustration purposes only. In libinput, a struct
**libinput_seat** comprises both physical seat and logical seat. From a
@ref libinput_seat comprises both physical seat and logical seat. From a
caller's point-of-view the above device layout is presented as:
.. graphviz:: seats-sketch-libinput.gv
@dotfile seats-sketch-libinput.gv
Thus, devices "Foo" and "Bar" both reference the same struct
**libinput_seat**, all other devices reference their own respective seats.
Thus, devices "Foo" and "Bar" both reference the same struct @ref
libinput_seat, all other devices reference their own respective seats.
.. _seats_and_features:
------------------------------------------------------------------------------
The effect of seat assignment
------------------------------------------------------------------------------
@section seats_and_features The effect of seat assignment
A logical set is interpreted as a group of devices that usually belong to a
single user that interacts with a computer. Thus, the devices are
@ -59,27 +48,26 @@ semantically related. This means for devices within the same logical seat:
two touches down.
libinput provides functions to aid with the above:
**libinput_event_pointer_get_seat_button_count()**,
**libinput_event_keyboard_get_seat_key_count()**, and
**libinput_event_touch_get_seat_slot()**.
libinput_event_pointer_get_seat_button_count(),
libinput_event_keyboard_get_seat_key_count(), and
libinput_event_touch_get_seat_slot().
Internally, libinput counts devices within the same logical seat as related.
Cross-device features only activate if all required devices are in the same
logical seat. For example, libinput will only activate the top software
buttons (see :ref:`t440_support`) if both trackstick and touchpad are assigned
buttons (see @ref t440_support) if both trackstick and touchpad are assigned
to the same logical seat.
.. _changing_seats:
------------------------------------------------------------------------------
Changing seats
------------------------------------------------------------------------------
@section changing_seats Changing seats
A device may change the logical seat it is assigned to at runtime with
**libinput_device_set_seat_logical_name()**. The physical seat is immutable and
libinput_device_set_seat_logical_name(). The physical seat is immutable and
may not be changed.
Changing the logical seat for a device is equivalent to unplugging the
device and plugging it back in with the new logical seat. No device state
carries over across a logical seat change.
*/

View file

@ -252,3 +252,4 @@ blockquote {
margin: 0 24px 0 4px;
padding: 0 12px 0 16px;
}

View file

@ -8,7 +8,7 @@
<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/>
<meta name="generator" content="Doxygen $doxygenversion"/>
<script type="text/javascript" src="https://code.jquery.com/jquery-2.1.1.min.js"></script>
<!--BEGIN PROJECT_NAME--><title>$projectname: $title</title><!--END PROJECT_NAME-->

View file

@ -0,0 +1,47 @@
dd {
margin-left: 30px;
}
.title {
font-size: 400%;
font-weight: 200;
}
h1 {
font-size: 300%;
}
h2 {
font-size: 200%;
}
.sm-dox li {
float:left;
border-top: 0;
padding-right: 20px;
}
.sm li, .sm a {
position: relative;
}
.sm, .sm ul, .sm li {
list-style: none;
display: block;
line-height: normal;
direction: ltr;
text-align: left;
}
.sm, .sm *, .sm *::before, .sm *::after {
box-sizing: border-box;
}
#main-nav {
padding: 30px;
}
/* Main menu sub-items like file-list, etc */
#main-menu li ul {
display: none;
}

File diff suppressed because one or more lines are too long

Before

Width:  |  Height:  |  Size: 105 KiB

After

Width:  |  Height:  |  Size: 105 KiB

View file

Before

Width:  |  Height:  |  Size: 17 KiB

After

Width:  |  Height:  |  Size: 17 KiB

View file

Before

Width:  |  Height:  |  Size: 5.9 KiB

After

Width:  |  Height:  |  Size: 5.9 KiB

View file

Before

Width:  |  Height:  |  Size: 16 KiB

After

Width:  |  Height:  |  Size: 16 KiB

View file

Before

Width:  |  Height:  |  Size: 7 KiB

After

Width:  |  Height:  |  Size: 7 KiB

View file

Before

Width:  |  Height:  |  Size: 28 KiB

After

Width:  |  Height:  |  Size: 28 KiB

View file

@ -138,6 +138,13 @@
style="fill:none;stroke:#000000;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke-miterlimit:4;stroke-dasharray:6, 1;stroke-dashoffset:0;stroke-opacity:1;marker-mid:none;marker-end:url(#Arrow1Lend-2)"
id="path13492"
d="m 38.928571,67.914286 c 0,0 3.508205,24.810617 9.642857,57.857144 6.134651,33.04652 23.277202,79.68584 89.642852,90.35714" />
<rect
style="fill:#000000;fill-opacity:0.3559322;fill-rule:evenodd;stroke:none;stroke-width:3.30510259px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1"
id="rect3490"
width="65.310997"
height="136.12065"
x="7.0411549"
y="7.1355872" />
<text
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:12px;line-height:0%;font-family:Utopia;-inkscape-font-specification:Utopia;text-align:start;writing-mode:lr-tb;text-anchor:start;fill:#000000;fill-opacity:1;stroke:none"
xml:space="preserve"
@ -148,6 +155,13 @@
id="tspan13876"
y="63.628628"
x="33.214291">A</tspan></text>
<rect
style="fill:#000000;fill-opacity:0.3559322;fill-rule:evenodd;stroke:none;stroke-width:3.30527353px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1"
id="rect3490-2"
width="65.272476"
height="136.21509"
x="321.22849"
y="6.8830237" />
<text
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:12px;line-height:0%;font-family:Utopia;-inkscape-font-specification:Utopia;text-align:start;writing-mode:lr-tb;text-anchor:start;display:inline;fill:#000000;fill-opacity:1;stroke:none"
xml:space="preserve"
@ -177,6 +191,16 @@
id="tspan13876-7-9"
y="46.009491"
x="342.27759">C</tspan></text>
<text
style="font-style:normal;font-variant:normal;font-weight:normal;font-stretch:normal;font-size:12px;line-height:0%;font-family:Utopia;-inkscape-font-specification:Utopia;text-align:start;writing-mode:lr-tb;text-anchor:start;display:inline;fill:#000000;fill-opacity:1;stroke:none"
xml:space="preserve"
id="text13874-8-1-4"
y="215.65927"
x="37.970726"><tspan
style="font-size:18px;font-family:Arial;-inkscape-font-specification:Arial"
id="tspan13876-7-9-5"
y="215.65927"
x="37.970726">D</tspan></text>
<circle
style="fill:#000000;fill-opacity:1;stroke:none;stroke-opacity:1"
id="path4401"
@ -184,6 +208,13 @@
cy="24.53549"
r="4.0658817"
transform="scale(-1,1)" />
<circle
style="fill:#000000;fill-opacity:1;stroke:none;stroke-opacity:1"
id="path4401-9"
cx="-36.452721"
cy="194.8819"
r="4.0658817"
transform="scale(-1,1)" />
<rect
width="248.87633"
height="6.8111157"

Before

Width:  |  Height:  |  Size: 7.5 KiB

After

Width:  |  Height:  |  Size: 8.8 KiB

View file

Before

Width:  |  Height:  |  Size: 16 KiB

After

Width:  |  Height:  |  Size: 16 KiB

View file

Before

Width:  |  Height:  |  Size: 34 KiB

After

Width:  |  Height:  |  Size: 34 KiB

View file

@ -1,5 +1,5 @@
<?xml version="1.0" encoding="utf-8" standalone="no"?>
<svg
<svg
width="600" height="480"
viewBox="0 0 600 480"
xmlns="http://www.w3.org/2000/svg"
@ -516,3 +516,4 @@
</g>
</g>
</svg>

Before

Width:  |  Height:  |  Size: 45 KiB

After

Width:  |  Height:  |  Size: 45 KiB

View file

@ -1,7 +1,7 @@
<?xml version="1.0" encoding="utf-8" standalone="no"?>
<!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1//EN"
"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd">
<svg
<svg
width="600" height="480"
viewBox="0 0 600 480"
xmlns="http://www.w3.org/2000/svg"
@ -3745,3 +3745,4 @@
</g>
</g>
</svg>

Before

Width:  |  Height:  |  Size: 377 KiB

After

Width:  |  Height:  |  Size: 377 KiB

View file

@ -1,5 +1,5 @@
<?xml version="1.0" encoding="utf-8" standalone="no"?>
<svg
<svg
width="600" height="480"
viewBox="0 0 600 480"
xmlns="http://www.w3.org/2000/svg"
@ -577,3 +577,4 @@
</g>
</g>
</svg>

Before

Width:  |  Height:  |  Size: 51 KiB

After

Width:  |  Height:  |  Size: 51 KiB

View file

@ -1,5 +1,5 @@
<?xml version="1.0" encoding="utf-8" standalone="no"?>
<svg
<svg
width="600" height="480"
viewBox="0 0 600 480"
xmlns="http://www.w3.org/2000/svg"
@ -77,28 +77,23 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='black' d='M159.2,422.4 L159.2,413.4 M159.2,16.7 L159.2,25.7 '/> <g transform="translate(159.2,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 0.2</tspan></text>
<path stroke='black' d='M185.2,422.4 L185.2,413.4 M185.2,16.7 L185.2,25.7 '/> <g transform="translate(185.2,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 5</tspan></text>
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='black' d='M263.2,422.4 L263.2,413.4 M263.2,16.7 L263.2,25.7 '/> <g transform="translate(263.2,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 0.4</tspan></text>
<path stroke='black' d='M315.2,422.4 L315.2,413.4 M315.2,16.7 L315.2,25.7 '/> <g transform="translate(315.2,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 10</tspan></text>
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='black' d='M367.1,422.4 L367.1,413.4 M367.1,16.7 L367.1,25.7 '/> <g transform="translate(367.1,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 0.6</tspan></text>
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='black' d='M471.1,422.4 L471.1,413.4 M471.1,16.7 L471.1,25.7 '/> <g transform="translate(471.1,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 0.8</tspan></text>
<path stroke='black' d='M445.1,422.4 L445.1,413.4 M445.1,16.7 L445.1,25.7 '/> <g transform="translate(445.1,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 15</tspan></text>
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='black' d='M575.0,422.4 L575.0,413.4 M575.0,16.7 L575.0,25.7 '/> <g transform="translate(575.0,444.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" > 1</tspan></text>
<text><tspan font-family="Arial" > 20</tspan></text>
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -112,7 +107,7 @@
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<g transform="translate(315.1,471.3)" stroke="none" fill="black" font-family="Arial" font-size="12.00" text-anchor="middle">
<text><tspan font-family="Arial" >delta (units/ms)</tspan></text>
<text><tspan font-family="Arial" >delta (units)</tspan></text>
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -124,10 +119,19 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='rgb(148, 0, 211)' d='M516.2,34.7 L558.4,34.7 M55.3,421.2 L75.8,420.3 L96.2,419.2 L116.7,418.0 L137.1,416.9 L157.6,415.8
L178.1,414.7 L198.5,413.7 L219.0,412.8 L239.4,411.8 L259.9,410.9 L280.4,410.1 L300.8,409.3 L321.3,408.5
L341.7,407.8 L362.2,407.1 L382.7,406.4 L403.1,405.8 L423.6,405.2 L444.1,404.6 L464.5,404.1 L485.0,403.5
L505.4,403.0 L525.9,402.6 L546.4,402.1 L566.8,401.7 L575.0,401.5 '/></g>
<path stroke='rgb(148, 0, 211)' d='M516.2,34.7 L558.4,34.7 M55.3,422.4 L60.5,418.5 L65.7,414.6 L70.9,410.7 L76.1,406.8 L81.3,402.9
L86.5,399.0 L91.7,395.1 L96.9,391.2 L102.1,387.3 L107.3,383.5 L112.5,379.6 L117.7,375.7 L122.9,371.8
L128.1,367.9 L133.3,364.0 L138.5,360.1 L143.6,356.2 L148.8,352.3 L154.0,348.4 L159.2,344.5 L164.4,340.6
L169.6,336.7 L174.8,332.8 L180.0,328.9 L185.2,325.0 L190.4,321.1 L195.6,317.2 L200.8,313.3 L206.0,310.0
L211.2,310.0 L216.4,310.0 L221.6,310.0 L226.8,310.0 L232.0,310.0 L237.2,310.0 L242.4,310.0 L247.6,310.0
L252.8,310.0 L258.0,310.0 L263.2,310.0 L268.4,310.0 L273.6,310.0 L278.8,310.0 L284.0,310.0 L289.2,310.0
L294.4,310.0 L299.6,310.0 L304.8,310.0 L310.0,310.0 L315.2,310.0 L320.3,310.0 L325.5,310.0 L330.7,310.0
L335.9,310.0 L341.1,310.0 L346.3,310.0 L351.5,310.0 L356.7,310.0 L361.9,310.0 L367.1,310.0 L372.3,310.0
L377.5,310.0 L382.7,310.0 L387.9,310.0 L393.1,310.0 L398.3,310.0 L403.5,310.0 L408.7,310.0 L413.9,310.0
L419.1,310.0 L424.3,310.0 L429.5,310.0 L434.7,310.0 L439.9,310.0 L445.1,310.0 L450.3,310.0 L455.5,310.0
L460.7,310.0 L465.9,310.0 L471.1,310.0 L476.3,310.0 L481.5,310.0 L486.7,310.0 L491.8,310.0 L497.0,310.0
L502.2,310.0 L507.4,310.0 L512.6,310.0 L517.8,310.0 L523.0,310.0 L528.2,310.0 L533.4,310.0 L538.6,310.0
L543.8,310.0 L549.0,310.0 L554.2,310.0 L559.4,310.0 L564.6,310.0 L569.8,310.0 L575.0,310.0 '/></g>
</g>
<g id="gnuplot_plot_2" ><title>-0.75</title>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -136,10 +140,19 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='rgb( 0, 158, 115)' d='M516.2,52.7 L558.4,52.7 M55.3,420.4 L75.8,418.7 L96.2,416.7 L116.7,414.8 L137.1,412.8 L157.6,410.9
L178.1,409.1 L198.5,407.3 L219.0,405.6 L239.4,404.0 L259.9,402.4 L280.4,401.0 L300.8,399.6 L321.3,398.2
L341.7,397.0 L362.2,395.8 L382.7,394.6 L403.1,393.5 L423.6,392.5 L444.1,391.4 L464.5,390.5 L485.0,389.6
L505.4,388.7 L525.9,387.8 L546.4,387.0 L566.8,386.3 L575.0,386.0 '/></g>
<path stroke='rgb( 0, 158, 115)' d='M516.2,52.7 L558.4,52.7 M55.3,422.4 L60.5,415.3 L65.7,408.1 L70.9,401.0 L76.1,393.8 L81.3,386.7
L86.5,379.6 L91.7,372.4 L96.9,365.3 L102.1,358.1 L107.3,351.0 L112.5,343.9 L117.7,336.7 L122.9,329.6
L128.1,322.4 L133.3,315.3 L138.5,308.2 L143.6,301.0 L148.8,293.9 L154.0,286.7 L159.2,279.6 L164.4,272.5
L169.6,270.7 L174.8,270.7 L180.0,270.7 L185.2,270.7 L190.4,270.7 L195.6,270.7 L200.8,270.7 L206.0,270.7
L211.2,270.7 L216.4,270.7 L221.6,270.7 L226.8,270.7 L232.0,270.7 L237.2,270.7 L242.4,270.7 L247.6,270.7
L252.8,270.7 L258.0,270.7 L263.2,270.7 L268.4,270.7 L273.6,270.7 L278.8,270.7 L284.0,270.7 L289.2,270.7
L294.4,270.7 L299.6,270.7 L304.8,270.7 L310.0,270.7 L315.2,270.7 L320.3,270.7 L325.5,270.7 L330.7,270.7
L335.9,270.7 L341.1,270.7 L346.3,270.7 L351.5,270.7 L356.7,270.7 L361.9,270.7 L367.1,270.7 L372.3,270.7
L377.5,270.7 L382.7,270.7 L387.9,270.7 L393.1,270.7 L398.3,270.7 L403.5,270.7 L408.7,270.7 L413.9,270.7
L419.1,270.7 L424.3,270.7 L429.5,270.7 L434.7,270.7 L439.9,270.7 L445.1,270.7 L450.3,270.7 L455.5,270.7
L460.7,270.7 L465.9,270.7 L471.1,270.7 L476.3,270.7 L481.5,270.7 L486.7,270.7 L491.8,270.7 L497.0,270.7
L502.2,270.7 L507.4,270.7 L512.6,270.7 L517.8,270.7 L523.0,270.7 L528.2,270.7 L533.4,270.7 L538.6,270.7
L543.8,270.7 L549.0,270.7 L554.2,270.7 L559.4,270.7 L564.6,270.7 L569.8,270.7 L575.0,270.7 '/></g>
</g>
<g id="gnuplot_plot_3" ><title>-0.5</title>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -148,10 +161,19 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='rgb( 86, 180, 233)' d='M516.2,70.7 L558.4,70.7 M55.3,416.8 L75.8,412.1 L96.2,406.6 L116.7,401.1 L137.1,395.6 L157.6,390.3
L178.1,385.2 L198.5,380.3 L219.0,375.6 L239.4,371.1 L259.9,366.8 L280.4,362.7 L300.8,358.8 L321.3,355.1
L341.7,351.6 L362.2,348.2 L382.7,345.0 L403.1,341.9 L423.6,338.9 L444.1,336.1 L464.5,333.5 L485.0,330.9
L505.4,328.5 L525.9,326.1 L546.4,323.9 L566.8,321.7 L575.0,320.9 '/></g>
<path stroke='rgb( 86, 180, 233)' d='M516.2,70.7 L558.4,70.7 M55.3,422.4 L60.5,412.0 L65.7,401.6 L70.9,391.2 L76.1,380.9 L81.3,370.5
L86.5,360.1 L91.7,349.7 L96.9,339.3 L102.1,328.9 L107.3,318.5 L112.5,308.2 L117.7,297.8 L122.9,287.4
L128.1,277.0 L133.3,266.6 L138.5,256.2 L143.6,245.8 L148.8,235.5 L154.0,225.1 L159.2,217.6 L164.4,217.6
L169.6,217.6 L174.8,217.6 L180.0,217.6 L185.2,217.6 L190.4,217.6 L195.6,217.6 L200.8,217.6 L206.0,217.6
L211.2,217.6 L216.4,217.6 L221.6,217.6 L226.8,217.6 L232.0,217.6 L237.2,217.6 L242.4,217.6 L247.6,217.6
L252.8,217.6 L258.0,217.6 L263.2,217.6 L268.4,217.6 L273.6,217.6 L278.8,217.6 L284.0,217.6 L289.2,217.6
L294.4,217.6 L299.6,217.6 L304.8,217.6 L310.0,217.6 L315.2,217.6 L320.3,217.6 L325.5,217.6 L330.7,217.6
L335.9,217.6 L341.1,217.6 L346.3,217.6 L351.5,217.6 L356.7,217.6 L361.9,217.6 L367.1,217.6 L372.3,217.6
L377.5,217.6 L382.7,217.6 L387.9,217.6 L393.1,217.6 L398.3,217.6 L403.5,217.6 L408.7,217.6 L413.9,217.6
L419.1,217.6 L424.3,217.6 L429.5,217.6 L434.7,217.6 L439.9,217.6 L445.1,217.6 L450.3,217.6 L455.5,217.6
L460.7,217.6 L465.9,217.6 L471.1,217.6 L476.3,217.6 L481.5,217.6 L486.7,217.6 L491.8,217.6 L497.0,217.6
L502.2,217.6 L507.4,217.6 L512.6,217.6 L517.8,217.6 L523.0,217.6 L528.2,217.6 L533.4,217.6 L538.6,217.6
L543.8,217.6 L549.0,217.6 L554.2,217.6 L559.4,217.6 L564.6,217.6 L569.8,217.6 L575.0,217.6 '/></g>
</g>
<g id="gnuplot_plot_4" ><title>-0.25</title>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -160,10 +182,19 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='rgb(230, 159, 0)' d='M516.2,88.7 L558.4,88.7 M55.3,409.5 L75.8,398.8 L96.2,386.3 L116.7,373.7 L137.1,361.2 L157.6,349.1
L178.1,337.4 L198.5,326.1 L219.0,315.3 L239.4,305.0 L259.9,295.2 L280.4,285.9 L300.8,276.9 L321.3,268.4
L341.7,260.3 L362.2,252.6 L382.7,245.2 L403.1,238.2 L423.6,231.5 L444.1,225.1 L464.5,218.9 L485.0,213.1
L505.4,207.5 L525.9,202.1 L546.4,197.0 L566.8,192.1 L575.0,190.2 '/></g>
<path stroke='rgb(230, 159, 0)' d='M516.2,88.7 L558.4,88.7 M55.3,422.4 L60.5,408.8 L65.7,395.1 L70.9,381.5 L76.1,367.9 L81.3,354.2
L86.5,340.6 L91.7,327.0 L96.9,313.3 L102.1,299.7 L107.3,286.1 L112.5,272.5 L117.7,258.8 L122.9,245.2
L128.1,231.6 L133.3,217.9 L138.5,204.3 L143.6,190.7 L148.8,177.0 L154.0,163.4 L159.2,149.8 L164.4,145.9
L169.6,145.9 L174.8,145.9 L180.0,145.9 L185.2,145.9 L190.4,145.9 L195.6,145.9 L200.8,145.9 L206.0,145.9
L211.2,145.9 L216.4,145.9 L221.6,145.9 L226.8,145.9 L232.0,145.9 L237.2,145.9 L242.4,145.9 L247.6,145.9
L252.8,145.9 L258.0,145.9 L263.2,145.9 L268.4,145.9 L273.6,145.9 L278.8,145.9 L284.0,145.9 L289.2,145.9
L294.4,145.9 L299.6,145.9 L304.8,145.9 L310.0,145.9 L315.2,145.9 L320.3,145.9 L325.5,145.9 L330.7,145.9
L335.9,145.9 L341.1,145.9 L346.3,145.9 L351.5,145.9 L356.7,145.9 L361.9,145.9 L367.1,145.9 L372.3,145.9
L377.5,145.9 L382.7,145.9 L387.9,145.9 L393.1,145.9 L398.3,145.9 L403.5,145.9 L408.7,145.9 L413.9,145.9
L419.1,145.9 L424.3,145.9 L429.5,145.9 L434.7,145.9 L439.9,145.9 L445.1,145.9 L450.3,145.9 L455.5,145.9
L460.7,145.9 L465.9,145.9 L471.1,145.9 L476.3,145.9 L481.5,145.9 L486.7,145.9 L491.8,145.9 L497.0,145.9
L502.2,145.9 L507.4,145.9 L512.6,145.9 L517.8,145.9 L523.0,145.9 L528.2,145.9 L533.4,145.9 L538.6,145.9
L543.8,145.9 L549.0,145.9 L554.2,145.9 L559.4,145.9 L564.6,145.9 L569.8,145.9 L575.0,145.9 '/></g>
</g>
<g id="gnuplot_plot_5" ><title>0</title>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -172,10 +203,19 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='rgb(240, 228, 66)' d='M516.2,106.7 L558.4,106.7 M55.3,398.0 L75.8,377.7 L96.2,354.1 L116.7,330.2 L137.1,306.6 L157.6,283.6
L178.1,261.4 L198.5,240.1 L219.0,219.7 L239.4,200.2 L259.9,181.7 L280.4,163.9 L300.8,147.0 L321.3,130.9
L341.7,115.6 L362.2,101.0 L382.7,87.0 L403.1,73.7 L423.6,61.0 L444.1,48.8 L464.5,37.3 L485.0,26.2
L503.2,16.7 '/></g>
<path stroke='rgb(240, 228, 66)' d='M516.2,106.7 L558.4,106.7 M55.3,422.4 L60.5,405.5 L65.7,388.6 L70.9,371.8 L76.1,354.9 L81.3,338.0
L86.5,321.1 L91.7,304.3 L96.9,287.4 L102.1,270.5 L107.3,253.6 L112.5,236.8 L117.7,219.9 L122.9,203.0
L128.1,186.1 L133.3,169.2 L138.5,152.4 L143.6,135.5 L148.8,118.6 L154.0,101.7 L159.2,84.9 L164.4,68.0
L169.6,51.1 L174.8,49.2 L180.0,49.2 L185.2,49.2 L190.4,49.2 L195.6,49.2 L200.8,49.2 L206.0,49.2
L211.2,49.2 L216.4,49.2 L221.6,49.2 L226.8,49.2 L232.0,49.2 L237.2,49.2 L242.4,49.2 L247.6,49.2
L252.8,49.2 L258.0,49.2 L263.2,49.2 L268.4,49.2 L273.6,49.2 L278.8,49.2 L284.0,49.2 L289.2,49.2
L294.4,49.2 L299.6,49.2 L304.8,49.2 L310.0,49.2 L315.2,49.2 L320.3,49.2 L325.5,49.2 L330.7,49.2
L335.9,49.2 L341.1,49.2 L346.3,49.2 L351.5,49.2 L356.7,49.2 L361.9,49.2 L367.1,49.2 L372.3,49.2
L377.5,49.2 L382.7,49.2 L387.9,49.2 L393.1,49.2 L398.3,49.2 L403.5,49.2 L408.7,49.2 L413.9,49.2
L419.1,49.2 L424.3,49.2 L429.5,49.2 L434.7,49.2 L439.9,49.2 L445.1,49.2 L450.3,49.2 L455.5,49.2
L460.7,49.2 L465.9,49.2 L471.1,49.2 L476.3,49.2 L481.5,49.2 L486.7,49.2 L491.8,49.2 L497.0,49.2
L502.2,49.2 L507.4,49.2 L512.6,49.2 L517.8,49.2 L523.0,49.2 L528.2,49.2 L533.4,49.2 L538.6,49.2
L543.8,49.2 L549.0,49.2 L554.2,49.2 L559.4,49.2 L564.6,49.2 L569.8,49.2 L575.0,49.2 '/></g>
</g>
<g id="gnuplot_plot_6" ><title>0.5</title>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -184,8 +224,9 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='rgb( 0, 114, 178)' d='M516.2,124.7 L558.4,124.7 M55.3,360.4 L75.8,308.6 L96.2,248.7 L116.7,187.8 L137.1,127.6 L157.6,69.2
L176.6,16.7 '/></g>
<path stroke='rgb( 0, 114, 178)' d='M516.2,124.7 L558.4,124.7 M55.3,422.4 L60.5,399.0 L65.7,375.7 L70.9,352.3 L76.1,328.9 L81.3,305.6
L86.5,282.2 L91.7,258.8 L96.9,235.5 L102.1,212.1 L107.3,188.7 L112.5,165.3 L117.7,142.0 L122.9,118.6
L128.1,95.2 L133.3,71.9 L138.5,48.5 L143.6,25.1 L145.5,16.7 '/></g>
</g>
<g id="gnuplot_plot_7" ><title>1</title>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
@ -194,7 +235,9 @@
</g>
</g>
<g fill="none" color="black" stroke="currentColor" stroke-width="1.00" stroke-linecap="butt" stroke-linejoin="miter">
<path stroke='rgb(229, 30, 16)' d='M516.2,142.7 L558.4,142.7 M55.3,300.7 L75.8,199.0 L96.2,81.4 L107.3,16.7 '/></g>
<path stroke='rgb(229, 30, 16)' d='M516.2,142.7 L558.4,142.7 M55.3,422.4 L60.5,392.5 L65.7,362.7 L70.9,332.8 L76.1,303.0 L81.3,273.1
L86.5,243.2 L91.7,213.4 L96.9,183.5 L102.1,153.7 L107.3,123.8 L112.5,93.9 L117.7,64.1 L122.9,34.2
L125.9,16.7 '/></g>
</g>
<g fill="none" color="#FFFFFF" stroke="rgb(229, 30, 16)" stroke-width="2.00" stroke-linecap="butt" stroke-linejoin="miter">
</g>
@ -208,3 +251,4 @@
</g>
</g>
</svg>

Before

Width:  |  Height:  |  Size: 14 KiB

After

Width:  |  Height:  |  Size: 19 KiB

View file

Before

Width:  |  Height:  |  Size: 9.3 KiB

After

Width:  |  Height:  |  Size: 9.3 KiB

View file

Before

Width:  |  Height:  |  Size: 22 KiB

After

Width:  |  Height:  |  Size: 22 KiB

View file

Before

Width:  |  Height:  |  Size: 26 KiB

After

Width:  |  Height:  |  Size: 26 KiB

View file

Before

Width:  |  Height:  |  Size: 21 KiB

After

Width:  |  Height:  |  Size: 21 KiB

View file

Before

Width:  |  Height:  |  Size: 15 KiB

After

Width:  |  Height:  |  Size: 15 KiB

View file

Before

Width:  |  Height:  |  Size: 15 KiB

After

Width:  |  Height:  |  Size: 15 KiB

View file

Before

Width:  |  Height:  |  Size: 22 KiB

After

Width:  |  Height:  |  Size: 22 KiB

View file

Before

Width:  |  Height:  |  Size: 11 KiB

After

Width:  |  Height:  |  Size: 11 KiB

View file

Before

Width:  |  Height:  |  Size: 8.8 KiB

After

Width:  |  Height:  |  Size: 8.8 KiB

View file

Before

Width:  |  Height:  |  Size: 46 KiB

After

Width:  |  Height:  |  Size: 46 KiB

View file

Before

Width:  |  Height:  |  Size: 6.3 KiB

After

Width:  |  Height:  |  Size: 6.3 KiB

View file

Before

Width:  |  Height:  |  Size: 13 KiB

After

Width:  |  Height:  |  Size: 13 KiB

View file

Before

Width:  |  Height:  |  Size: 28 KiB

After

Width:  |  Height:  |  Size: 28 KiB

View file

Before

Width:  |  Height:  |  Size: 11 KiB

After

Width:  |  Height:  |  Size: 11 KiB

50
doc/switches.dox Normal file
View file

@ -0,0 +1,50 @@
/**
@page switches Switches
libinput supports a couple of switches. Unlike button events that come in
press and release pairs, switches are usually toggled once and left at the
setting for an extended period of time.
Only some switches are handled by libinput, see @ref libinput_switch for a
list of supported switches. Switch events are exposed to the caller, but
libinput may handle some switch events internally and enable or disable
specific features based on a switch state.
The order of switch events is guaranteed to be correct, i.e., a switch will
never send consecutive switch on, or switch off, events.
@section switches_lid Lid switch handling
Where available, libinput listens to devices providing a lid switch.
The evdev event code `EV_SW` `SW_LID` is provided as @ref
LIBINPUT_SWITCH_LID. If devices with a lid switch have a touchpad device,
the device is disabled while the lid is logically closed. This is to avoid
ghost touches that can be caused by interference with touchpads and the
closed lid. The touchpad is automatically re-enabled whenever the lid is
openend.
This handling of lid switches is transparent to the user, no notifications
are sent and the device appears as enabled at all times.
On some devices, the device's lid state does not always reflect the physical
state and the lid state may report as closed even when the lid is physicall
open. libinput employs some heuristics to detect user input (specificially
typing) to re-enable the touchpad on those devices.
@section switches_tablet_mode Tablet mode switch handling
Where available, libinput listens to devices providing a tablet mode switch.
This switch is usually triggered on devices that can switch between a normal
laptop layout and a tablet-like layout. One example for such a device is the
Lenovo Yoga.
The event sent by the kernel is `EV_SW` `SW_TABLET_MODE` and is provided as
@ref LIBINPUT_SWITCH_TABLET_MODE. When the device switches to tablet mode,
the touchpad and internal keyboard are disabled. If a trackpoint exists,
it is disabled too. The input devices are automatically re-enabled whenever
tablet mode is disengaged.
This handling of tablet mode switches is transparent to the user, no
notifications are sent and the device appears as enabled at all times.
*/

View file

@ -1,43 +1,29 @@
.. _t440_support:
/**
@page t440_support Lenovo *40 series touchpad support
==============================================================================
Lenovo \*40 series touchpad support
==============================================================================
The Lenovo \*40 series emulates trackstick buttons on the top part of the
The Lenovo *40 series emulates trackstick buttons on the top part of the
touchpads.
.. _t440_support_overview:
@section t440_support_overview Overview
------------------------------------------------------------------------------
Overview
------------------------------------------------------------------------------
The Lenovo \*40 series introduced a new type of touchpad. Previously, all
The Lenovo *40 series introduced a new type of touchpad. Previously, all
laptops had a separate set of physical buttons for the
`trackstick <http://en.wikipedia.org/wiki/Pointing_stick>`_. This
<a href="http://en.wikipedia.org/wiki/Pointing_stick">trackstick</a>. This
series removed these buttons, relying on a software emulation of the top
section of the touchpad. This is visually marked on the trackpad itself,
and clicks can be triggered by pressing the touchpad down with a finger in
the respective area:
.. figure:: top-software-buttons.svg
:align: center
Left, right and middle-button click with top software button areas
@image html top-software-buttons.svg "Left, right and middle-button click with top software button areas"
This page only covers the top software buttons, the bottom button behavior
is covered in :ref:`Clickpad software buttons <clickpad_softbuttons>`.
is covered in @ref clickpad_softbuttons "Clickpad software buttons".
Clickpads with a top button area are marked with the
`INPUT_PROP_TOPBUTTONPAD <https://www.kernel.org/doc/Documentation/input/event-codes.txt>`_
Clickpads with a top button area are marked with the <a
href="https://www.kernel.org/doc/Documentation/input/event-codes.txt">INPUT_PROP_TOPBUTTONPAD</a>
property.
.. _t440_support_btn_size:
------------------------------------------------------------------------------
Size of the buttons
------------------------------------------------------------------------------
@section t440_support_btn_size Size of the buttons
The size of the buttons matches the visual markings on this touchpad.
The width of the left and right buttons is approximately 42% of the
@ -49,23 +35,16 @@ measurements of button presses showed that the size of the buttons needs to
be approximately 10mm high to work reliable (especially when using the
thumb to press the button).
.. _t440_support_btn_behavior:
------------------------------------------------------------------------------
Button behavior
------------------------------------------------------------------------------
@section t440_support_btn_behavior Button behavior
Movement in the top button area does not generate pointer movement. These
buttons are not replacement buttons for the bottom button area but have
their own behavior. Semantically attached to the trackstick device, libinput
re-routes events from these buttons to appear through the trackstick device.
.. graphviz::
digraph top_button_routing
{
@dot
digraph top_button_routing
{
rankdir="LR";
node [shape="box";]
@ -95,8 +74,8 @@ re-routes events from these buttons to appear through the trackstick device.
libinput_tp -> events_tp [arrowhead="none"]
libinput_ts -> events_topbutton [color="red4"]
}
}
@enddot
The top button areas work even if the touchpad is disabled but will be
@ -106,23 +85,20 @@ and must be lifted to generate future buttons. Likewise, movement into the
top button area does not trigger button events, a click has to start inside
this area to take effect.
.. _t440_support_identification:
------------------------------------------------------------------------------
Kernel support
------------------------------------------------------------------------------
@section t440_support_identification Kernel support
The firmware on the first generation of touchpads providing top software
buttons is buggy and announces wrong ranges.
`Kernel patches <https://lkml.org/lkml/2014/3/7/722>`_ are required;
<a href="https://lkml.org/lkml/2014/3/7/722">Kernel patches</a> are required;
these fixes are available in kernels 3.14.1, 3.15 and later but each
touchpad needs a separate fix.
The October 2014 refresh of these laptops do not have this firmware bug
anymore and should work without per-device patches, though
`this kernel commit <http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=02e07492cdfae9c86e3bd21c0beec88dbcc1e9e8>`_
is required.
<a href="http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=02e07492cdfae9c86e3bd21c0beec88dbcc1e9e8">this kernel commit</a> is required.
For a complete list of supported touchpads check
`the kernel source <http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/input/mouse/synaptics.c>`_
(search for "topbuttonpad_pnp_ids").
For a complete list of supported touchpads check <a
href="http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/drivers/input/mouse/synaptics.c">the
kernel source</a> (search for "topbuttonpad_pnp_ids").
*/

View file

@ -1,60 +1,43 @@
.. _tablet-support:
==============================================================================
Tablet support
==============================================================================
/**
@page tablet-support Tablet support
This page provides details about the graphics tablet
support in libinput. Note that the term "tablet" in libinput refers to
graphics tablets only (e.g. Wacom Intuos), not to tablet devices like the
Apple iPad.
.. figure:: tablet.svg
:align: center
@image html tablet.svg "Illustration of a graphics tablet"
Illustration of a graphics tablet
.. _tablet-tools:
------------------------------------------------------------------------------
Pad buttons vs. tablet tools
------------------------------------------------------------------------------
@section tablet-tools Pad buttons vs. tablet tools
Most tablets provide two types of devices. The physical tablet often
provides a number of buttons and a touch ring or strip. Interaction on the
drawing surface of the tablet requires a tool, usually in the shape of a
stylus. The libinput interface exposed by devices with the
**LIBINPUT_DEVICE_CAP_TABLET_TOOL** capability applies only to events generated
stylus. The libinput interface exposed by devices with the @ref
LIBINPUT_DEVICE_CAP_TABLET_TOOL capability applies only to events generated
by tools.
Buttons, rings or strips on the physical tablet hardware (the "pad") are
exposed by devices with the **LIBINPUT_DEVICE_CAP_TABLET_PAD** capability.
exposed by devices with the @ref LIBINPUT_DEVICE_CAP_TABLET_PAD capability.
Pad events do not require a tool to be in proximity. Note that both
capabilities may exist on the same device though usually they are split
across multiple kernel devices.
.. figure:: tablet-interfaces.svg
:align: center
Difference between Pad and Tool buttons
@image html tablet-interfaces.svg "Difference between Pad and Tool buttons"
Touch events on the tablet integrated into a screen itself are exposed
through the **LIBINPUT_DEVICE_CAP_TOUCH** capability. Touch events on a
standalone tablet are exposed through the **LIBINPUT_DEVICE_CAP_POINTER**
through the @ref LIBINPUT_DEVICE_CAP_TOUCH capability. Touch events on a
standalone tablet are exposed through the @ref LIBINPUT_DEVICE_CAP_POINTER
capability. In both cases, the kernel usually provides a separate event
node for the touch device, resulting in a separate libinput device.
See **libinput_device_get_device_group()** for information on how to associate
See libinput_device_get_device_group() for information on how to associate
the touch part with other devices exposed by the same physical hardware.
.. _tablet-tip:
------------------------------------------------------------------------------
Tool tip events vs. tool button events
------------------------------------------------------------------------------
@section tablet-tip Tool tip events vs. tool button events
The primary use of a tablet tool is to draw on the surface of the tablet.
When the tool tip comes into contact with the surface, libinput sends an
event of type **LIBINPUT_EVENT_TABLET_TOOL_TIP**, and again when the tip
event of type @ref LIBINPUT_EVENT_TABLET_TOOL_TIP, and again when the tip
ceases contact with the surface.
Tablet tools may send button events; these are exclusively for extra buttons
@ -68,8 +51,8 @@ tools are capable of detecting 1 gram of pressure.
libinput uses a device-specific pressure threshold to determine when the tip
is considered logically down. As a result, libinput may send a nonzero
pressure value while the tip is logically up. Most application can and
should ignore pressure information until they receive the event of type
**LIBINPUT_EVENT_TABLET_TOOL_TIP**. Applications that require extremely
should ignore pressure information until they receive the event of type @ref
LIBINPUT_EVENT_TABLET_TOOL_TIP. Applications that require extremely
fine-grained pressure sensitivity should use the pressure data instead of
the tip events to determine a logical tip down state and treat the tip
events like axis events otherwise.
@ -78,17 +61,13 @@ Note that the pressure threshold to trigger a logical tip event may be zero
on some devices. On tools without pressure sensitivity, determining when a
tip is down is device-specific.
.. _tablet-relative-motion:
------------------------------------------------------------------------------
Relative motion for tablet tools
------------------------------------------------------------------------------
@section tablet-relative-motion Relative motion for tablet tools
libinput calculates the relative motion vector for each event and converts
it to the same coordinate space that a normal mouse device would use. For
the caller, this means that the delta coordinates returned by
**libinput_event_tablet_tool_get_dx()** and
**libinput_event_tablet_tool_get_dy()** can be used identical to the delta
libinput_event_tablet_tool_get_dx() and
libinput_event_tablet_tool_get_dy() can be used identical to the delta
coordinates from any other pointer event. Any resolution differences between
the x and y axes are accommodated for, a delta of N/N represents a 45 degree
diagonal move on the tablet.
@ -100,25 +79,18 @@ all pen-like tools to absolute mode.
If a tool in relative mode must not use pointer acceleration, callers
should use the absolute coordinates returned by
**libinput_event_tablet_tool_get_x()** and libinput_event_tablet_tool_get_y()
libinput_event_tablet_tool_get_x() and libinput_event_tablet_tool_get_y()
and calculate the delta themselves. Callers that require exact physical
distance should also use these functions to calculate delta movements.
.. _tablet-axes:
------------------------------------------------------------------------------
Special axes on tablet tools
------------------------------------------------------------------------------
@section tablet-axes Special axes on tablet tools
A tablet tool usually provides additional information beyond x/y positional
information and the tip state. A tool may provide the distance to the tablet
surface and the pressure exerted on the tip when in contact. Some tablets
additionally provide tilt information along the x and y axis.
.. figure:: tablet-axes.svg
:align: center
Illustration of the distance, pressure and tilt axes
@image html tablet-axes.svg "Illustration of the distance, pressure and tilt axes"
The granularity and precision of the distance and pressure axes varies
between tablet devices and cannot usually be mapped into a physical unit.
@ -133,13 +105,9 @@ the tablet and the top of the stylus. The angle is measured along the x and
y axis, respectively, a positive tilt angle thus means that the stylus' top
is tilted towards the logical right and/or bottom of the tablet.
.. _tablet-fake-proximity:
@section tablet-fake-proximity Handling of proximity events
------------------------------------------------------------------------------
Handling of proximity events
------------------------------------------------------------------------------
libinput's **LIBINPUT_EVENT_TABLET_TOOL_PROXIMITY** events notify a caller
libinput's @ref LIBINPUT_EVENT_TABLET_TOOL_PROXIMITY events notify a caller
when a tool comes into sensor range or leaves the sensor range. On some
tools this range does not represent the physical range but a reduced
tool-specific logical range. If the range is reduced, this is done
@ -150,16 +118,12 @@ used in relative mode, lying flat on the tablet. Movement typically follows
the interaction normal mouse movements have, i.e. slightly lift the tool and
place it in a separate location. The proximity detection on Wacom
tablets however extends further than the user may lift the mouse, i.e. the
tool may not be lifted out of physical proximity. For such tools, libinput
provides software-emulated proximity.
tool may not be lifted out of physical proximity. For such tools, libinput
provides software-emulated proximity.
Events from the pad do not require proximity, they may be sent any time.
.. _tablet-pressure-offset:
------------------------------------------------------------------------------
Pressure offset on worn-out tools
------------------------------------------------------------------------------
@section tablet-pressure-offset Pressure offset on worn-out tools
When a tool is used for an extended period it can wear down physically. A
worn-down tool may never return a zero pressure value. Even when hovering
@ -175,83 +139,32 @@ than that offset.
Some limitations apply to avoid misdetection of pressure offsets,
specifically:
- pressure offset is only detected on proximity in, and if a device is
capable of detection distances,
- pressure offset is only detected if the distance between the tool and the
tablet is high enough,
- pressure offset is only used if it is 50% or less of the pressure range
available to the tool. A pressure offset higher than 50% indicates either
- pressure offset is only used if it is 20% or less of the pressure range
available to the tool. A pressure offset higher than 20% indicates either
a misdetection or a tool that should be replaced, and
- if a pressure value less than the current pressure offset is seen, the
offset resets to that value.
Pressure offsets are not detected on **LIBINPUT_TABLET_TOOL_TYPE_MOUSE**
and **LIBINPUT_TABLET_TOOL_TYPE_LENS** tools.
Pressure offsets are not detected on @ref LIBINPUT_TABLET_TOOL_TYPE_MOUSE
and @ref LIBINPUT_TABLET_TOOL_TYPE_LENS tools.
.. _tablet-pressure-range:
------------------------------------------------------------------------------
Custom tablet tool pressure ranges
------------------------------------------------------------------------------
On tablets supporting pressure, libinput provides that hardware pressure as
a logical range of ``0.0`` up to ``1.0`` for the maximum supported pressure.
By default, the hardware range thus maps into the following logical range::
hw minimum hw maximum
hw range: |------|-----------------------------------|
logical range: |----|-----------------------------------|
0.0 | 1.0
Tip
Note that libinput always has some built-in thresholds to filter out erroneous
touches with near-zero pressure but otherwise the hardware range maps as-is
into the logical range. The :ref:`tip event <tablet-tip>` threshold is defined
within this range.
For some use-cases the full hardware range is not suitable, it may require either
too light a pressure for the user to interact or it may require too hard a
pressure before the logical maximum is reached. libinput provides
the **libinput_tablet_tool_config_pressure_range_set()** function that allows
reducing the usable range of the tablet::
hw minimum hw maximum
hw range: |----------|-------------------------------|
adjusted range: |------|---------------------|
logical range: |----|---------------------|
0.0 | 1.0
Tip
A reduced range as shown above will result in
- all hw pressure below the new minimum to register as logical pressure ``0.0``
- all hw pressure above the new maximum to register as logical pressure ``1.0``
- the tip event threshold to be relative to the new minimum
In other words, adjusting the pressure range of a tablet tool is equivalent to
reducing the hardware range of said tool. Note that where a custom pressure
range is set, detection of :ref:`tablet-pressure-offset` is disabled.
.. _tablet-serial-numbers:
------------------------------------------------------------------------------
Tracking unique tools
------------------------------------------------------------------------------
@section tablet-serial-numbers Tracking unique tools
Some tools provide hardware information that enables libinput to uniquely
identify the physical device. For example, tools compatible with the Wacom
Intuos 4, Intuos 5, Intuos Pro and Cintiq series are uniquely identifiable
through a serial number. libinput does not specify how a tool can be
identified uniquely, a caller should use **libinput_tablet_tool_is_unique()** to
identified uniquely, a caller should use libinput_tablet_tool_is_unique() to
check if the tool is unique.
libinput creates a struct libinput_tablet_tool on the first proximity in of
this tool. By default, this struct is destroyed on proximity out and
re-initialized on the next proximity in. If a caller keeps a reference to
the tool by using **libinput_tablet_tool_ref()** libinput re-uses this struct
the tool by using libinput_tablet_tool_ref() libinput re-uses this struct
whenever that same physical tool comes into proximity on any tablet
recognized by libinput. It is possible to attach tool-specific virtual state
to the tool. For example, a graphics program such as the GIMP may assign a
@ -263,24 +176,20 @@ If the tool does not have a unique identifier, libinput creates a single
struct libinput_tablet_tool per tool type on each tablet the tool is used
on.
.. _tablet-tool-types:
------------------------------------------------------------------------------
Vendor-specific tablet tool types
------------------------------------------------------------------------------
@section tablet-tool-types Vendor-specific tablet tool types
libinput supports a number of high-level tool types that describe the
general interaction expected with the tool. For example, a user would expect
a tool of type **LIBINPUT_TABLET_TOOL_TYPE_PEN** to interact with a
a tool of type @ref LIBINPUT_TABLET_TOOL_TYPE_PEN to interact with a
graphics application taking pressure and tilt into account. The default
virtual tool assigned should be a drawing tool, e.g. a virtual pen or brush.
A tool of type **LIBINPUT_TABLET_TOOL_TYPE_ERASER** would normally be
mapped to an eraser-like virtual tool. See **libinput_tablet_tool_type**
A tool of type @ref LIBINPUT_TABLET_TOOL_TYPE_ERASER would normally be
mapped to an eraser-like virtual tool. See @ref libinput_tablet_tool_type
for the list of all available tools.
Vendors may provide more fine-grained information about the tool in use by
adding a hardware-specific tool ID. libinput provides this ID to the caller
with **libinput_tablet_tool_get_tool_id()** but makes no promises about the
with libinput_tablet_tool_get_tool_id() but makes no promises about the
content or format of the ID.
libinput currently supports Wacom-style tool IDs as provided on the Wacom
@ -288,11 +197,7 @@ Intuos 3, 4, 5, Wacon Cintiq and Wacom Intuos Pro series. The tool ID can
be used to distinguish between e.g. a Wacom Classic Pen or a Wacom Pro Pen.
It is the caller's responsibility to interpret the tool ID.
.. _tablet-bounds:
------------------------------------------------------------------------------
Out-of-bounds motion events
------------------------------------------------------------------------------
@section tablet-bounds Out-of-bounds motion events
Some tablets integrated into a screen (e.g. Wacom Cintiq 24HD, 27QHD and
13HD series, etc.) have a sensor larger than the display area. libinput uses
@ -301,10 +206,7 @@ quirks are present. Events outside this range will produce coordinates that
may be negative or larger than the tablet's width and/or height. It is up to
the caller to ignore these events.
.. figure:: tablet-out-of-bounds.svg
:align: center
Illustration of the out-of-bounds area on a tablet
@image html tablet-out-of-bounds.svg "Illustration of the out-of-bounds area on a tablet"
In the image above, the display area is shown in black. The red area around
the display illustrates the sensor area that generates input events. Events
@ -312,21 +214,17 @@ within this area will have negative coordinate or coordinates larger than
the width/height of the tablet.
If events outside the logical bounds of the input area are scaled into a
custom range with **libinput_event_tablet_tool_get_x_transformed()** and
**libinput_event_tablet_tool_get_y_transformed()** the resulting value may be
custom range with libinput_event_tablet_tool_get_x_transformed() and
libinput_event_tablet_tool_get_y_transformed() the resulting value may be
less than 0 or larger than the upper range provided. It is up to the caller
to test for this and handle or ignore these events accordingly.
.. _tablet-pad-buttons:
------------------------------------------------------------------------------
Tablet pad button numbers
------------------------------------------------------------------------------
@section tablet-pad-buttons Tablet pad button numbers
Tablet Pad buttons are numbered sequentially, starting with button 0. Thus
button numbers returned by **libinput_event_tablet_pad_get_button_number()**
button numbers returned by libinput_event_tablet_pad_get_button_number()
have no semantic meaning, a notable difference to the button codes returned
by other libinput interfaces (e.g. **libinput_event_tablet_tool_get_button()**).
by other libinput interfaces (e.g. libinput_event_tablet_tool_get_button()).
The Linux kernel requires all input events to have semantic event codes,
but generic buttons like those on a pad cannot easily be assigned semantic
@ -342,20 +240,16 @@ tablet.
Some buttons may have expected default behaviors. For example, on Wacom
Intuos Pro series tablets, the button inside the touch ring is expected to
switch between modes, see :ref:`tablet-pad-modes`. Callers should use
switch between modes, see @ref tablet-pad-modes. Callers should use
external sources like libwacom to identify which buttons have semantic
behaviors.
.. _tablet-left-handed:
------------------------------------------------------------------------------
Tablets in left-handed mode
------------------------------------------------------------------------------
@section tablet-left-handed Tablets in left-handed mode
Left-handed mode on tablet devices usually means rotating the physical
tablet by 180 degrees to move the tablet pad button area to right side of
the tablet. When left-handed mode is enabled on a tablet device (see
**libinput_device_config_left_handed_set()**) the tablet tool and tablet pad
libinput_device_config_left_handed_set()) the tablet tool and tablet pad
behavior changes. In left-handed mode, the tools' axes are adjusted
so that the origin of each axis remains the logical north-east of
the physical tablet. For example, the x and y axes are inverted and the
@ -364,15 +258,11 @@ in its current orientation. On a tablet pad, the ring and strip are
similarly adjusted. The origin of the ring and strips remain the top-most
point.
.. figure:: tablet-left-handed.svg
:align: center
Tablet axes in right- and left-handed mode
@image html tablet-left-handed.svg "Tablet axes in right- and left-handed mode"
Pad buttons are not affected by left-handed mode; the number of each button
remains the same even when the perceived physical location of the button
changes. This is a conscious design decision:
- Tablet pad buttons do not have intrinsic semantic meanings. Re-ordering
the button numbers would not change any functionality.
- Button numbers should not be exposed directly to the user but handled in
@ -385,11 +275,7 @@ symmetric and thus do not support left-handed mode. libinput requires
libwacom to determine if a tablet is capable of being switched to
left-handed mode.
.. _tablet-pad-modes:
------------------------------------------------------------------------------
Tablet pad modes
------------------------------------------------------------------------------
@section tablet-pad-modes Tablet pad modes
Tablet pad modes are virtual groupings of button, ring and strip
functionality. A caller may assign different functionalities depending on
@ -420,42 +306,32 @@ libinput handles the mode groups independently and returns the mode for each
button as appropriate. The mode group is static for the lifetime of the
device.
.. figure:: tablet-intuos-modes.svg
:align: center
Modes on an Intuos Pro-like tablet
@image html tablet-intuos-modes.svg "Modes on an Intuos Pro-like tablet"
In the image above, the Intuos Pro-like tablet provides 4 LEDs to indicate
the currently active modes. The button inside the touch ring cycles through
the modes in a clockwise fashion. The upper-right LED indicates that the
currently active mode is 1, based on 0-indexed mode numbering.
**libinput_event_tablet_pad_get_mode()** would thus return 1 for all button and
libinput_event_tablet_pad_get_mode() would thus return 1 for all button and
ring events on this tablet. When the center button is pressed, the mode
switches to mode 2, the LED changes to the bottom-right and
**libinput_event_tablet_pad_get_mode()** returns 2 for the center button event
libinput_event_tablet_pad_get_mode() returns 2 for the center button event
and all subsequent events.
.. figure:: tablet-cintiq24hd-modes.svg
:align: center
Modes on an Cintiq 24HD-like tablet
@image html tablet-cintiq24hd-modes.svg "Modes on an Cintiq 24HD-like tablet"
In the image above, the Cintiq 24HD-like tablet provides 3 LEDs on each side
of the tablet to indicate the currently active mode for that group of
buttons and the respective ring. The buttons next to the touch ring select
the mode directly. The two LEDs indicate that the mode for the left set of
buttons is currently 0, the mode for the right set of buttons is currently
1, based on 0-indexed mode numbering. **libinput_event_tablet_pad_get_mode()**
1, based on 0-indexed mode numbering. libinput_event_tablet_pad_get_mode()
would thus return 0 for all button and ring events on the left and 1 for all
button and ring events on the right. When one of the three mode toggle
buttons on the right is pressed, the right mode switches to that button's
mode but the left mode remains unchanged.
.. _tablet-touch-arbitration:
------------------------------------------------------------------------------
Tablet touch arbitration
------------------------------------------------------------------------------
@section tablet-touch-arbitration Tablet touch arbitration
"Touch arbitration" is the terminology used when touch events are suppressed
while the pen is in proximity. Since it is almost impossible to use a stylus
@ -464,81 +340,35 @@ touch arbitration serves to reduce the number of accidental inputs.
The wacom kernel driver currently provides touch arbitration but for other
devices arbitration has to be done in userspace.
libinput uses the **libinput_device_group** to decide on touch arbitration
libinput uses the @ref libinput_device_group to decide on touch arbitration
and automatically discards touch events whenever a tool is in proximity.
The exact behavior is device-dependent.
.. _tablet-area:
@section tablet-capabilities Required tablet capabilities
------------------------------------------------------------------------------
Tablet area
------------------------------------------------------------------------------
To handle a tablet correctly, libinput requires a set of capabilities
on the device. When these capabilities are missing, libinput ignores the
device and prints an error to the log. This error messages reads
<pre>
missing tablet capabilities: xy pen btn-stylus resolution. Ignoring this device.
</pre>
or in older versions of libinput simply:
<pre>
libinput bug: device does not meet tablet criteria. Ignoring this device.
</pre>
External tablet devices such as e.g. the Wacom Intuos series can be configured
to reduce the available logical input area. Typically the logical input area
is equivalent to the physical input area but it can be reduced with the
**libinput_device_config_area_set_rectangle()** call. Once reduced, input
events outside the logical input area are ignored and the logical input area
acts as if it represented the extents of the physical tablet.
When a tablet is rejected, it is usually possible to check the issue with
the `evemu-descibe` tool.
.. figure:: tablet-area.svg
:align: center
- **xy** indicates that the tablet is missing the `ABS_X` and/or `ABS_Y`
axis. This indicates that the device is mislabelled and the udev tag
`ID_INPUT_TABLET` is applied to a device that is not a tablet.
- **pen** or **btn-stylus** indicates that the tablet does not have the
`BTN_TOOL_PEN` or `BTN_STYLUS` bit set. libinput requires either or both
of them to be present. This usually indicates a bug in the kernel driver
or the HID descriptors of the device.
- **resolution** indicates that the device does not have a resolution set
for the x and y axes. This can be fixed with a hwdb entry, locate and read
the 60-evdev.hwdb file on your machine to address this.
Tablet area configuration example
In the image above, the area is set to the rectangle 0.25/0.25 to 0.5/0.75.
Even though the tool is roughly at the physical position ``0.5 * width`` and
``0.75 * height``, the return values of
**libinput_event_tablet_tool_get_x_transformed()** and
**libinput_event_tablet_tool_get_y_transformed()** would be close to the
maximum provided in this call.
The size of the tablet reported by **libinput_device_get_size()** always reflects
the physical area, not the logical area.
.. _tablet-eraser-button:
------------------------------------------------------------------------------
Tablet eraser buttons
------------------------------------------------------------------------------
Tablet tools come in a variety of forms but the most common one is a
pen-like tool. Some of these pen-like tools have a virtual eraser at the
tip of the tool - inverting the tool brings the eraser into proximity.
.. figure:: tablet-eraser-invert.svg
:align: center
An pen-like tool used as pen and as eraser by inverting it
Having an eraser as a separate tool is beneficial in many applications as the
eraser tool can be assigned different functionality (colors, paint tools, etc.)
that is easily available.
However, a large proportion of tablet pens have an "eraser button". By
pressing the button the pen switches to be an eraser tool.
On the data level this is not done via a button event, instead the firmware
will pretend the pen tool going out of proximity and the eraser coming
into proximity immediately after - as if the tool was physically inverted.
.. figure:: tablet-eraser-button.svg
:align: center
An pen-like tool used as pen and as eraser by pressing the eraser button
Microsoft mandates this behavior (see
`Windows Pen States <https://learn.microsoft.com/en-us/windows-hardware/design/component-guidelines/windows-pen-states>`_
for details) and thus the overwhelming majority of devices will have
an eraser button that virtually inverts the pen.
Enforcing an eraser button means that users have one button less on the
stylus that they would have otherwise. For some users the eraser button
is in an inconvenient location, others don't want an eraser button at all.
libinput provides an eraser button configuration that allows disabling the
eraser button and turning it into a normal button event. If the eraser button
is disabled, pressing that button will generate a normal tablet tool button
event.
This configuration is only available on pens with an eraser button, not on
with an invert-type eraser.
*/

76
doc/tapping.dox Normal file
View file

@ -0,0 +1,76 @@
/**
@page tapping Tap-to-click behaviour
"Tapping" or "tap-to-click" is the name given to the behavior where a short
finger touch down/up sequence maps into a button click. This is most
commonly used on touchpads, but may be available on other devices.
libinput implements tapping for one, two, and three fingers, where supported
by the hardware, and maps those taps into a left, right, and middle button
click, respectively. Not all devices support three fingers, libinput will
support tapping up to whatever is supported by the hardware. libinput does
not support four-finger taps or any tapping with more than four fingers,
even though some hardware can distinguish between that many fingers.
@section tapping_default Tap-to-click default setting
Tapping is **disabled** by default on most devices, see [this
commit](https://gitlab.freedesktop.org/libinput/libinput/commit/2219c12c3aa45b80f235e761e87c17fb9ec70eae)
because:
- if you don't know that tapping is a thing (or enabled by default), you get
spurious button events that make the desktop feel buggy.
- if you do know what tapping is and you want it, you usually know where to
enable it, or at least you can search for it.
Tapping is **enabled** by default on devices where tapping is the only
method to trigger button clicks. This includes devices without physical
buttons such as touch-capable graphics tablets.
Tapping can be enabled/disabled on a per-device basis. See
libinput_device_config_tap_set_enabled() for details.
@section tapndrag Tap-and-drag
libinput also supports "tap-and-drag" where a tap immediately followed by a
finger down and that finger being held down emulates a button press. Moving
the finger around can thus drag the selected item on the screen.
Tap-and-drag is optional and can be enabled or disabled with
libinput_device_config_tap_set_drag_enabled(). Most devices have
tap-and-drag enabled by default.
Also optional is a feature called "drag lock". With drag lock disabled, lifting
the finger will stop any drag process. When enabled, libinput will ignore a
finger up event during a drag process, provided the finger is set down again
within a implementation-specific timeout. Drag lock can be enabled and
disabled with libinput_device_config_tap_set_drag_lock_enabled().
Note that drag lock only applies if tap-and-drag is be enabled.
@image html tap-n-drag.svg "Tap-and-drag process"
The above diagram explains the process, a tap (a) followed by a finger held
down (b) starts the drag process and logically holds the left mouse button
down. A movement of the finger (c) will drag the selected item until the
finger is released (e). If needed and drag lock is enabled, the finger's
position can be reset by lifting and quickly setting it down again on the
touchpad (d). This will be interpreted as continuing move and is especially
useful on small touchpads or with slow pointer acceleration.
If drag lock is enabled, the release of the mouse buttons after the finger
release (e) is triggered by a timeout. To release the button immediately,
simply tap again (f).
If two fingers are supported by the hardware, a second finger can be used to
drag while the first is held in-place.
@section tap_constraints Constraints while tapping
A couple of constraints apply to the contact to be converted into a press, the most common ones are:
- the touch down and touch up must happen within an implementation-defined timeout
- if a finger moves more than an implementation-defined distance while in contact, it's not a tap
- tapping within @ref clickpad_softbuttons "clickpad software buttons" may not trigger an event
- a tap not meeting required pressure thresholds can be ignored as accidental touch
- a tap exceeding certain pressure thresholds can be ignored (see @ref
palm_detection)
- a tap on the edges of the touchpad can usually be ignored (see @ref
palm_detection)
*/

134
doc/test-suite.dox Normal file
View file

@ -0,0 +1,134 @@
/**
@page test-suite libinput test suite
libinput ships with a number of tests all run automatically on `ninja test`.
The primary test suite is the `libinput-test-suite-runner`. When testing,
the `libinput-test-suite-runner` should always be invoked to check for
behavior changes.
The test suite runner uses
[Check](http://check.sourceforge.net/doc/check_html/) underneath the hood
but most of the functionality is abstracted into *litest* wrappers.
The test suite runner has a make-like job control enabled by the `-j` or
`--jobs` flag and will fork off as many parallel processes as given by this
flag. The default if unspecified is 8. When debugging a specific test case
failure it is recommended to employ test filtures (see @ref test-filtering)
and disable parallel tests. The test suite automatically disables parallel
make when run in gdb.
@section test-config X.Org config to avoid interference
uinput devices created by the test suite are usually recognised by X as
input devices. All events sent through these devices will generate X events
and interfere with your desktop.
Copy the file `$srcdir/test/50-litest.conf` into your `/etc/X11/xorg.conf.d`
and restart X. This will ignore any litest devices and thus not interfere
with your desktop.
@section test-root Permissions required to run tests
Most tests require the creation of uinput devices and access to the
resulting `/dev/input/eventX` nodes. Some tests require temporary udev rules.
<b>This usually requires the tests to be run as root</b>. If not run as
root, the test suite runner will exit with status 77, interpreted as
"skipped" by ninja.
@section test-filtering Selective running of tests
litest's tests are grouped into test groups, test names and devices. A test
group is e.g. "touchpad:tap" and incorporates all tapping-related tests for
touchpads. Each test function is (usually) run with one or more specific
devices. The `--list` commandline argument shows the list of suites and
tests. This is useful when trying to figure out if a specific test is
run for a device.
@code
$ ./test/libinput-test-suite-runner --list
...
pointer:left-handed:
pointer_left_handed_during_click_multiple_buttons:
trackpoint
ms-surface-cover
mouse-wheelclickcount
mouse-wheelclickangle
low-dpi-mouse
mouse-roccat
mouse-wheel-tilt
mouse
logitech-trackball
cyborg-rat
magicmouse
pointer_left_handed_during_click:
trackpoint
ms-surface-cover
mouse-wheelclickcount
mouse-wheelclickangle
low-dpi-mouse
mouse-roccat
mouse-wheel-tilt
mouse
logitech-trackball
cyborg-rat
litest-magicmouse-device
pointer_left_handed:
trackpoint
ms-surface-cover
mouse-wheelclickcount
mouse-wheelclickangle
low-dpi-mouse
mouse-roccat
mouse-wheel-tilt
mouse
...
@endcode
In the above example, the "pointer:left-handed" suite contains multiple
tests, e.g. "pointer_left_handed_during_click" (this is also the function
name of the test, making it easy to grep for). This particular test is run
for various devices including the trackpoint device and the magic mouse
device.
The "no device" entry signals that litest does not instantiate a uinput
device for a specific test (though the test itself may
instantiate one).
The `--filter-test` argument enables selective running of tests through
basic shell-style function name matching. For example:
@code
$ ./test/libinput-test-suite-runner --filter-test="*1fg_tap*"
@endcode
The `--filter-device` argument enables selective running of tests through
basic shell-style device name matching. The device names matched are the
litest-specific shortnames, see the output of `--list`. For example:
@code
$ ./test/libinput-test-suite-runner --filter-device="synaptics*"
@endcode
The `--filter-group` argument enables selective running of test groups
through basic shell-style test group matching. The test groups matched are
litest-specific test groups, see the output of `--list`. For example:
@code
$ ./test/libinput-test-suite-runner --filter-group="touchpad:*hover*"
@endcode
The `--filter-device` and `--filter-group` arguments can be combined with
`--list` to show which groups and devices will be affected.
@section test-verbosity Controlling test output
Each test supports the `--verbose` commandline option to enable debugging
output, see libinput_log_set_priority() for details. The `LITEST_VERBOSE`
environment variable, if set, also enables verbose mode.
@code
$ ./test/libinput-test-suite-runner --verbose
$ LITEST_VERBOSE=1 ninja test
@endcode
*/

36
doc/timestamps.dox Normal file
View file

@ -0,0 +1,36 @@
/**
@page timestamps Timestamps
@section event_timestamps Event timestamps
Most libinput events provide a timestamp in millisecond and/or microsecond
resolution. These timestamp usually increase monotonically, but libinput
does not guarantee that this always the case. In other words, it is possible
to receive an event with a timestamp earlier than the previous event.
For example, if a touchpad has @ref tapping enabled, a button event may have a
lower timestamp than an event from a different device. Tapping requires the
use of timeouts to detect multi-finger taps and/or @ref tapndrag.
Consider the following event sequences from a touchpad and a mouse:
<pre>
Time Touchpad Mouse
---------------------------------
t1 finger down
t2 finger up
t3 movement
t4 tap timeout
</pre>
For this event sequence, the first event to be sent to a caller is in
response to the mouse movement: an event of type @ref
LIBINPUT_EVENT_POINTER_MOTION with the timestamp t3.
Once the timeout expires at t4, libinput generates an event of
@ref LIBINPUT_EVENT_POINTER_BUTTON (press) with a timestamp t1 and an event
@ref LIBINPUT_EVENT_POINTER_BUTTON (release) with a timestamp t2.
Thus, the caller gets events with timestamps in the order t3, t1, t2,
despite t3 > t2 > t1.
*/

237
doc/tools.dox Normal file
View file

@ -0,0 +1,237 @@
/**
@page tools Helper tools
libinput provides a `libinput` tool to query state and events. This tool
takes a subcommand as argument, similar to the **git** command. A full
explanation of the various commands available in the libinput tool is
available in the **libinput(1)** man page.
The most common tools used are:
- `libinput list-devices`: to list locally available devices, see @ref
libinput-list-devices "here"
- `libinput debug-events`: to monitor and debug events, see @ref
libinput-debug-events "here"
- `libinput debug-gui`: to visualize events, see @ref libinput-debug-gui
"here"
- `libinput record`: to record an event sequence for replaying, see @ref
libinput-record "here"
- `libinput measure`: measure properties on a kernel device, see @ref
libinput-measure "here"
Most the tools must be run as root to have access to the kernel's @c
/dev/input/event* device files.
@section libinput-list-devices libinput list-devices
The `libinput list-devices` command shows information about devices
recognized by libinput and can help identifying why a device behaves
different than expected. For example, if a device does not show up in the
output, it is not a supported input device.
@note This tool does **not** show your desktop's configuration, just the
libinput built-in defaults.
@verbatim
$ sudo libinput list-devices
[...]
Device: SynPS/2 Synaptics TouchPad
Kernel: /dev/input/event4
Group: 9
Seat: seat0, default
Size: 97.33x66.86mm
Capabilities: pointer
Tap-to-click: disabled
Tap drag lock: disabled
Left-handed: disabled
Nat.scrolling: disabled
Middle emulation: n/a
Calibration: n/a
Scroll methods: *two-finger
Click methods: *button-areas clickfinger
[...]
@endverbatim
The above listing shows example output for a touchpad. The
`libinput list-devices` command lists general information about the device
(the kernel event node) but also the configuration options. If an option is
"n/a" it does not exist on this device. Otherwise, the tool will show the
default configuration for this device, for options that have more than a
binary state all available options are listed, with the default one prefixed
with an asterisk (*). In the example above, the default click method is
button-areas but clickinger is available.
@note This tool is intended to be human-readable and may change its output
at any time.
@section libinput-debug-events libinput debug-events
The `libinput debug-events` command prints events from devices and can help
to identify why a device behaves different than expected.
@verbatim
$ sudo libinput debug-events --enable-tapping --set-click-method=clickfinger
@endverbatim
All configuration options (enable/disable tapping,
etc.) are available as commandline arguments. To reproduce the event
sequence as your desktop session sees it, ensure that all options are turned
on or off as required. See the **libinput-debug-events(1)** man page or the
`--help` output for information about the available options.
@note When submitting a bug report, always use the `--verbose` flag to get
additional information: `libinput debug-events --verbose <other options>`
An example output from this tool may look like the snippet below.
@verbatim
$ sudo libinput debug-events --enable-tapping --set-click-method=clickfinger
-event2 DEVICE_ADDED Power Button seat0 default group1 cap:k
-event5 DEVICE_ADDED Video Bus seat0 default group2 cap:k
-event0 DEVICE_ADDED Lid Switch seat0 default group3 cap:S
-event1 DEVICE_ADDED Sleep Button seat0 default group4 cap:k
-event4 DEVICE_ADDED HDA Intel HDMI HDMI/DP,pcm=3 seat0 default group5 cap:
-event11 DEVICE_ADDED HDA Intel HDMI HDMI/DP,pcm=7 seat0 default group6 cap:
-event12 DEVICE_ADDED HDA Intel HDMI HDMI/DP,pcm=8 seat0 default group7 cap:
-event13 DEVICE_ADDED HDA Intel HDMI HDMI/DP,pcm=9 seat0 default group8 cap:
-event14 DEVICE_ADDED HDA Intel HDMI HDMI/DP,pcm=10 seat0 default group9 cap:
-event19 DEVICE_ADDED Integrated Camera: Integrated C seat0 default group10 cap:k
-event15 DEVICE_ADDED HDA Intel PCH Dock Mic seat0 default group11 cap:
-event16 DEVICE_ADDED HDA Intel PCH Mic seat0 default group12 cap:
-event17 DEVICE_ADDED HDA Intel PCH Dock Headphone seat0 default group13 cap:
-event18 DEVICE_ADDED HDA Intel PCH Headphone seat0 default group14 cap:
-event6 DEVICE_ADDED ELAN Touchscreen seat0 default group15 cap:t size 305x172mm ntouches 10 calib
-event3 DEVICE_ADDED AT Translated Set 2 keyboard seat0 default group16 cap:k
-event20 DEVICE_ADDED SynPS/2 Synaptics TouchPad seat0 default group17 cap:pg size 100x76mm tap(dl off) left scroll-nat scroll-2fg-edge click-buttonareas-clickfinger dwt-on
-event21 DEVICE_ADDED TPPS/2 IBM TrackPoint seat0 default group18 cap:p left scroll-nat scroll-button
-event7 DEVICE_ADDED ThinkPad Extra Buttons seat0 default group19 cap:k
-event20 POINTER_MOTION +3.62s 2.72/ -0.93
event20 POINTER_MOTION +3.63s 1.80/ -1.42
event20 POINTER_MOTION +3.65s 6.16/ -2.28
event20 POINTER_MOTION +3.66s 6.42/ -1.99
event20 POINTER_MOTION +3.67s 8.99/ -1.42
event20 POINTER_MOTION +3.68s 11.30/ 0.00
event20 POINTER_MOTION +3.69s 21.32/ 1.42
@endverbatim
@section libinput-debug-gui libinput debug-gui
A simple GTK-based graphical tool that shows the behavior and location of
touch events, pointer motion, scroll axes and gestures. Since this tool
gathers data directly from libinput, it is thus suitable for
pointer-acceleration testing.
@note This tool does **not** use your desktop's configuration, just the
libinput built-in defaults.
@verbatim
$ sudo libinput debug-gui --enable-tapping
@endverbatim
As with @ref libinput-debug-events, all options must be specified on the
commandline to emulate the correct behavior.
See the **libinput-debug-gui(1)** man page or the `--help` output for information about
the available options.
@section libinput-record libinput record and libinput replay
The `libinput record` command records the **kernel** events from a specific
device node. The recorded sequence can be replayed with the `libinput
replay` command. This pair of tools is crucial to capturing bugs and
reproducing them on a developer's machine.
@note These tools are shipped with libinput, but the recorded events
are **kernel events** and independent of the libinput context. libinput does not
need to be running, it does not matter whether a user is running X.Org or
Wayland or even what version of libinput is currently running.
The use of the tools is straightforward, just run without arguments, piping
the output into a file:
@verbatim
$ sudo libinput record > touchpad.log
Available devices:
/dev/input/event0: Lid Switch
/dev/input/event1: Sleep Button
/dev/input/event2: Power Button
/dev/input/event3: AT Translated Set 2 keyboard
/dev/input/event4: ThinkPad Extra Buttons
/dev/input/event5: ELAN Touchscreen
/dev/input/event6: Video Bus
/dev/input/event7: HDA Intel HDMI HDMI/DP,pcm=3
/dev/input/event8: HDA Intel HDMI HDMI/DP,pcm=7
/dev/input/event9: HDA Intel HDMI HDMI/DP,pcm=8
/dev/input/event10: HDA Intel HDMI HDMI/DP,pcm=9
/dev/input/event11: HDA Intel HDMI HDMI/DP,pcm=10
/dev/input/event12: HDA Intel PCH Dock Mic
/dev/input/event13: HDA Intel PCH Mic
/dev/input/event14: HDA Intel PCH Dock Headphone
/dev/input/event15: HDA Intel PCH Headphone
/dev/input/event16: Integrated Camera: Integrated C
/dev/input/event17: SynPS/2 Synaptics TouchPad
/dev/input/event18: TPPS/2 IBM TrackPoint
Select the device event number: 17
/dev/input/event17 recording to stdout
@endverbatim
Without arguments, `libinput record` displays the available devices and lets
the user select one. Supply the number (17 in this case for
`/dev/input/event17`) and the tool will print the device information and
events to the file it is redirected to. More arguments are available, see
the **libinput-record(1)** man page.
Reproduce the bug, ctrl+c and attach the output file to a bug report.
For data protection, `libinput record` obscures key codes by default, any
alphanumeric key shows up as letter "a".
@note When reproducing a bug that crashes libinput, run inside `screen` or
`tmux`.
The recording can be replayed with the `libinput replay` command:
@verbatim
$ sudo libinput replay touchpad.log
SynPS/2 Synaptics TouchPad: /dev/input/event19
Hit enter to start replaying
@endverbatim
`libinput replay` creates a new virtual device based on the description in
the log file. Hitting enter replays the event sequence once and the tool
stops once all events have been replayed. Hitting enter again replays the
sequence again, Ctrl+C stops it and removes the virtual device.
Users are advised to always replay a recorded event sequence to ensure they
have captured the bug.
More arguments are available, see the **libinput-record(1)** and
**libinput-replay(1)** man pages.
@subsection libinput-record-multiple Recording multiple devices at once
In some cases, an interaction between multiple devices is the cause for a
specific bug. For example, a touchpad may not work in response to keyboard
events. To accurately reproduce this sequence, the timing between multiple
devices must be correct and we need to record the events in one go.
`libinput record` has a `--multiple` argument to record multiple devices at
once. Unlike the normal invocation, this one requires a number of arguments:
@verbatim
$ sudo libinput record --multiple --output-file=touchpad-bug.log /dev/input/event17 /dev/input/event3
recording to 'touchpad-bug.log'
@endverbatim
As seen above, a user must specify `--multiple` and the `--output-file`.
Finally, all devices to be recorded must be specified on the commandline as
well.
Replaying events is the same as for a single recording:
@verbatim
$ sudo libinput replay touchpad-bug.log
@endverbatim
@subsection libinput-measure Measuring device properties with libinput measure
The `libinput measure` tool is a multiplexer for various sub-tools that can
measure specific properties on the device. These tools generally measure one
thing and one thing only and their usage is highly specific to the tool.
Please see the **libinput-measure(1)** man page for information about what
tools are available and the man page for each respective tool.
*/

File diff suppressed because one or more lines are too long

Before

Width:  |  Height:  |  Size: 147 KiB

90
doc/touchpad-jitter.dox Normal file
View file

@ -0,0 +1,90 @@
/**
@page touchpad_jitter Touchpad jitter
Touchpad jitter describes random movement by a few pixels even when the
user's finger is unmoving.
libinput has a mechanism called a **hysteresis** to avoid that jitter. When
active, movement with in the **hysteresis margin** is discarded. If the
movement delta is larger than the margin, the movement is passed on as
pointer movement. This is a simplified summary, developers should
read the implementation of the hysteresis in `src/evdev.c`.
libinput uses the kernel `fuzz` value to determine the size of the
hysteresis. Users should override this with a udev hwdb entry where the
device itself does not provide the correct value.
@section touchpad_jitter_fuzz_override Overriding the hysteresis margins
libinput provides the debugging tool `libinput measure fuzz` to help edit or
test a fuzz value. This tool is interactive and provides a udev hwdb entry
that matches the device. To check if a fuzz is currently present, simply run
without arguments or with the touchpad's device node:
@verbatim
$ sudo libinput measure fuzz
Using Synaptics TM2668-002: /dev/input/event17
Checking udev property... not set
Checking axes... x=16 y=16
@endverbatim
In the above output, the axis fuzz is set to 16. To set a specific fuzz, run
with the `--fuzz=<value>` argument.
@verbatim
$ sudo libinput measure fuzz --fuzz=8
@endverbatim
The tool will attempt to construct a hwdb file that matches your touchpad
device. Follow the printed prompts.
In the ideal case, the tool will provide you with a file that can be
submitted to the systemd repo for inclusion.
However, hwdb entry creation is difficult to automate and it's likely
that the tools fails in doing so, especially if an existing entry is already
present.
Below is the outline of what a user needs to do to override a device's fuzz
value in case the `libinput measure fuzz` tool fails.
Check with `udevadm info /sys/class/input/eventX` (replace your device node
number) whether an existing hwdb override exists. If the `EVDEV_ABS_`
properties are present, the hwdb overried exists. Find the file that
contains that entry, most likely in `/etc/udev/hwdb.d` or
`/usr/lib/udev/hwdb.d`.
The content of the property is a set of values in the format
`EVDEV_ABS_00=min:max:resolution:fuzz`. You need to set the `fuzz` part,
leaving the remainder of the property as-is. Values may be empty, e.g. a
property that only sets resolution and fuzz reads as `EVDEV_ABS_00=::32:8`.
If no properties exist, your hwdb.entry should look approximately like this:
@verbatim
evdev:name:Synaptics TM2668-002:dmi:*:svnLENOVO*:pvrThinkPadT440s*:
EVDEV_ABS_00=:::8
EVDEV_ABS_01=:::8
EVDEV_ABS_35=:::8
EVDEV_ABS_36=:::8
@endverbatim
Substitute the `name` field with the device name (see the output of
`libinput measure fuzz` and the DMI match content with your hardware. See
@ref hwdb_modifying for details.
Once the hwdb entry has been modified, added, or created, @ref
hwdb_reloading "reload the hwdb". Once reloaded, @ref libinput-record
"libinput record" should show the new fuzz value for the axes.
Restart the host and libinput should pick up the revised fuzz values.
@section kernel_fuzz Kernel fuzz
A fuzz set on an absolute axis in the kernel causes the kernel to apply
hysteresis-like behavior to the axis. Unfortunately, this behavior leads to
inconsistent deltas. To avoid this, libinput sets the kernel fuzz on the
device to 0 to disable this kernel behavior but remembers what the fuzz was
on startup. The fuzz is stored in the `LIBINPUT_FUZZ_XX` udev property, on
startup libinput will check that property as well as the axis itself.
*/

View file

@ -0,0 +1,54 @@
/**
@page touchpad_jumping_cursor Touchpad jumping cursor bugs
A common bug encountered on touchpads is a cursor jump when alternating
between fingers on a multi-touch-capable touchpad. For example, after moving
the cursor a user may use a second finger in the software button area to
physically click the touchpad. Upon setting the finger down, the cursor
exhibits a jump towards the bottom left or right, depending on the finger
position.
When libinput detects a cursor jump it prints a bug warning to the log with
the text <b>"Touch jump detected and discarded."</b> and a link to this page.
In most cases, this is a bug in the kernel driver and to libinput it appears
that the touch point moves from its previous position. The pointer jump can
usually be seen in the evemu-record output for the device:
<pre>
E: 249.206319 0000 0000 0000 # ------------ SYN_REPORT (0) ----------
E: 249.218008 0003 0035 3764 # EV_ABS / ABS_MT_POSITION_X 3764
E: 249.218008 0003 0036 2221 # EV_ABS / ABS_MT_POSITION_Y 2221
E: 249.218008 0003 003a 0065 # EV_ABS / ABS_MT_PRESSURE 65
E: 249.218008 0003 0000 3764 # EV_ABS / ABS_X 3764
E: 249.218008 0003 0001 2216 # EV_ABS / ABS_Y 2216
E: 249.218008 0003 0018 0065 # EV_ABS / ABS_PRESSURE 65
E: 249.218008 0000 0000 0000 # ------------ SYN_REPORT (0) ----------
E: 249.230881 0003 0035 3752 # EV_ABS / ABS_MT_POSITION_X 3752
E: 249.230881 0003 003a 0046 # EV_ABS / ABS_MT_PRESSURE 46
E: 249.230881 0003 0000 3758 # EV_ABS / ABS_X 3758
E: 249.230881 0003 0018 0046 # EV_ABS / ABS_PRESSURE 46
E: 249.230881 0000 0000 0000 # ------------ SYN_REPORT (0) ----------
E: 249.242648 0003 0035 1640 # EV_ABS / ABS_MT_POSITION_X 1640
E: 249.242648 0003 0036 4681 # EV_ABS / ABS_MT_POSITION_Y 4681
E: 249.242648 0003 003a 0025 # EV_ABS / ABS_MT_PRESSURE 25
E: 249.242648 0003 0000 1640 # EV_ABS / ABS_X 1640
E: 249.242648 0003 0001 4681 # EV_ABS / ABS_Y 4681
E: 249.242648 0003 0018 0025 # EV_ABS / ABS_PRESSURE 25
E: 249.242648 0000 0000 0000 # ------------ SYN_REPORT (0) ----------
E: 249.254568 0003 0035 1648 # EV_ABS / ABS_MT_POSITION_X 1648
E: 249.254568 0003 003a 0027 # EV_ABS / ABS_MT_PRESSURE 27
E: 249.254568 0003 0000 1644 # EV_ABS / ABS_X 1644
E: 249.254568 0003 0018 0027 # EV_ABS / ABS_PRESSURE 27
</pre>
In this recording, the pointer jumps from its position 3752/2216 to
1640/4681 within a single frame. On this particular touchpad, this would
represent a physical move of almost 50mm. libinput detects some of these
jumps and discards the movement but otherwise continues as usual. However,
the bug should be fixed at the kernel level.
When you encounter the warning in the log, please generate an evemu
recording of your touchpad and file a bug. See @ref reporting_bugs for more
details.
*/

217
doc/touchpad-pressure.dox Normal file
View file

@ -0,0 +1,217 @@
/**
@page touchpad_pressure Touchpad pressure-based touch detection
libinput uses the touchpad pressure values and/or touch size values to
detect wether a finger has been placed on the touchpad. This is @ref
kernel_pressure_information and combines with a libinput-specific hardware
database to adjust the thresholds on a per-device basis. libinput uses
these thresholds primarily to filter out accidental light touches but
the information is also used for some @ref palm_detection.
Pressure and touch size thresholds are **not** directly configurable by the
user. Instead, libinput provides these thresholds for each device where
necessary. See @ref touchpad_pressure_hwdb for instructions on how to adjust
the pressure ranges and @ref touchpad_touch_size_hwdb for instructions on
how to adjust the touch size ranges.
@section kernel_pressure_information Information provided by the kernel
The kernel sends multiple values to inform userspace about a finger touching
the touchpad. The most basic is the ```EV_KEY/BTN_TOUCH``` boolean event
that simply announces physical contact with the touchpad. The decision when
this event is sent is usually made by the kernel driver and may depend on
device-specific thresholds. These thresholds are transparent to userspace
and cannot be modified. On touchpads where pressure or touch size is not
available, libinput uses ```BTN_TOUCH``` to determine when a finger is
logically down.
Many contemporary touchpad devices provide an absolute pressure axis in
addition to ```BTN_TOUCH```. This pressure generally increases as the pressure
increases, however few touchpads are capable of detecting true pressure. The
pressure value is usually related to the covered area - as the pressure
increases a finger flattens and thus covers a larger area. The range
provided by the kernel is not mapped to a specific physical range and
often requires adjustment. Pressure is sent by the ```ABS_PRESSURE``` axis
for single-touch touchpads or ```ABS_MT_PRESSURE``` on multi-touch capable
touchpads. Some devices can detect multiple fingers but only provide
```ABS_PRESSURE```.
Some devices provide additional touch size information through
the ```ABS_MT_TOUCH_MAJOR/ABS_MT_TOUCH_MINOR``` axes and/or
the ```ABS_MT_WIDTH_MAJOR/ABS_MT_WIDTH_MINOR``` axes. These axes specifcy
the size of the touch ellipse. While the kernel documentation specifies how
these axes are supposed to be mapped, few devices forward reliable
information. libinput uses these values together with a device-specific hwdb
entry. In other words, touch size detection does not work unless a hwdb
entry is present for the device.
@section touchpad_pressure_hwdb Debugging touchpad pressure ranges
This section describes how to determine the touchpad pressure ranges
required for a touchpad device and how to add the required hwdb entry
locally. Note that the hwdb entry is **not public API** and **may change at
any time**. Users are advised to @ref reporting_bugs "report a bug" with the
updated pressure ranges when testing has completed.
Use the ```libinput measure touchpad-pressure``` tool provided by libinput.
This tool will search for your touchpad device and print some pressure
statistics, including whether a touch is/was considered logically down.
@note This tool will only work on touchpads with pressure.
Example output of the tool is below:
<pre>
$ sudo libinput measure touchpad-pressure
Ready for recording data.
Pressure range used: 8:10
Palm pressure range used: 65535
Place a single finger on the touchpad to measure pressure values.
Ctrl+C to exit
&nbsp;
Sequence 1190 pressure: min: 39 max: 48 avg: 43 median: 44 tags: down
Sequence 1191 pressure: min: 49 max: 65 avg: 62 median: 64 tags: down
Sequence 1192 pressure: min: 40 max: 78 avg: 64 median: 66 tags: down
Sequence 1193 pressure: min: 36 max: 83 avg: 70 median: 73 tags: down
Sequence 1194 pressure: min: 43 max: 76 avg: 72 median: 74 tags: down
Touchpad pressure: 47 min: 47 max: 86 tags: down
</pre>
The example output shows five completed touch sequences and one ongoing one.
For each, the respective minimum and maximum pressure values are printed as
well as some statistics. The ```tags``` show that sequence was considered
logically down at some point. This is an interactive tool and its output may
change frequently. Refer to the <i>libinput-measure-touchpad-pressure(1)</i> man
page for more details.
By default, this tool uses the udev hwdb entries for the pressure range. To
narrow down on the best values for your device, specify the 'logically down'
and 'logically up' pressure thresholds with the ```--touch-thresholds``
argument:
<pre>
$ sudo libinput measure touchpad-pressure --touch-thresholds=10:8 --palm-threshold=20
</pre>
Interact with the touchpad and check if the output of this tool matches your
expectations.
@note **This is an interactive process. You will need to re-run the
tool with varying thresholds until you find the right range for your
touchpad. Attaching output logs to a bug will not help, only you with access
to the hardware can figure out the correct ranges.**
Once the thresholds are decided on (e.g. 10 and 8), they can be enabled with
the following hwdb file:
<pre>
$> cat /etc/udev/hwdb.d/99-touchpad-pressure.hwdb
libinput:name:*SynPS/2 Synaptics TouchPad:dmi:*svnHewlett-Packard:*pnHPCompaq6910p*
LIBINPUT_ATTR_PRESSURE_RANGE=10:8
</pre>
The first line is the match line and should be adjusted for the device name
(see evemu-record's output) and for the local system, based on the
information in ```/sys/class/dmi/id/modalias```. The modalias should be
shortened to the specific system's information, usually system vendor (svn)
and product name (pn).
Once in place, you need to run the following to commands, adjusted for your
device's event node (see @ref faq_hwdb_changes):
<pre>
sudo udevadm hwdb --update
sudo udevadm test /sys/class/input/eventX
</pre>
If the pressure range property shows up correctly, restart X or the
Wayland compositor and libinput should now use the correct pressure
thresholds. The @ref tools can be used to verify the correct
functionality first without the need for a restart.
Once the pressure ranges are deemed correct,
@ref reporting_bugs "report a bug" to get the pressure ranges into the
repository.
@section touchpad_touch_size_hwdb Debugging touch size ranges
This section describes how to determine the touchpad touch size ranges
required for a touchpad device and how to add the required hwdb entry
locally. Note that the hwdb entry is **not public API** and **may change at
any time**. Users are advised to @ref reporting_bugs "report a bug" with the
updated pressure ranges when testing has completed.
Use the ```libinput measure touch-size``` tool provided by libinput.
This tool will search for your touchpad device and print some touch size
statistics, including whether a touch is/was considered logically down.
@note This tool will only work on touchpads with the ```ABS_MT_MAJOR``` axis.
Example output of the tool is below:
<pre>
$ sudo libinput measure touch-size --touch-thresholds 10:8 --palm-threshold 14
Using ELAN Touchscreen: /dev/input/event5
&nbsp;
Ready for recording data.
Touch sizes used: 10:8
Palm size used: 14
Place a single finger on the device to measure touch size.
Ctrl+C to exit
&nbsp;
Sequence: major: [ 9.. 11] minor: [ 7.. 9]
Sequence: major: [ 9.. 10] minor: [ 7.. 7]
Sequence: major: [ 9.. 14] minor: [ 6.. 9] down
Sequence: major: [ 11.. 11] minor: [ 9.. 9] down
Sequence: major: [ 4.. 33] minor: [ 1.. 5] down palm
</pre>
The example output shows five completed touch sequences. For each, the
respective minimum and maximum pressure values are printed as well as some
statistics. The ```down``` and ```palm``` tags show that sequence was considered
logically down or a palm at some point. This is an interactive tool and its
output may change frequently. Refer to the <i>libinput-measure-touch-size(1)</i> man
page for more details.
By default, this tool uses the udev hwdb entries for the touch size range. To
narrow down on the best values for your device, specify the 'logically down'
and 'logically up' pressure thresholds with the ```--touch-thresholds``
arguments as in the example above.
Interact with the touchpad and check if the output of this tool matches your
expectations.
@note **This is an interactive process. You will need to re-run the
tool with varying thresholds until you find the right range for your
touchpad. Attaching output logs to a bug will not help, only you with access
to the hardware can figure out the correct ranges.**
Once the thresholds are decided on (e.g. 10 and 8), they can be enabled with
the following hwdb file:
<pre>
$> cat /etc/udev/hwdb.d/99-touchpad-pressure.hwdb
libinput:name:*SynPS/2 Synaptics TouchPad:dmi:*svnHewlett-Packard:*pnHPCompaq6910p*
LIBINPUT_ATTR_TOUCH_SIZE_RANGE=10:8
</pre>
The first line is the match line and should be adjusted for the device name
(see evemu-record's output) and for the local system, based on the
information in ```/sys/class/dmi/id/modalias```. The modalias should be
shortened to the specific system's information, usually system vendor (svn)
and product name (pn).
Once in place, you need to run the following to commands, adjusted for your
device's event node (see @ref faq_hwdb_changes):
<pre>
sudo udevadm hwdb --update
sudo udevadm test /sys/class/input/eventX
</pre>
If the touch size range property shows up correctly, restart X or the
Wayland compositor and libinput should now use the correct thresholds.
The @ref tools can be used to verify the correct functionality first without
the need for a restart.
Once the touch size ranges are deemed correct, @ref reporting_bugs "report a
bug" to get the thresholds into the repository.
*/

File diff suppressed because one or more lines are too long

Before

Width:  |  Height:  |  Size: 181 KiB

After

Width:  |  Height:  |  Size: 166 KiB

Some files were not shown because too many files have changed in this diff Show more