Processing os-release in the same buffer that the dmi modalias used caused the
dmi to be recorded as 'dmi: "VERSION_ID=31"'. The cause for that was simply
that the dmi modalias was read but not printed until after the os-release
information was processed.
Fix this two-fold: rearrange that each part now reads and prints in
one go, and rename the buffers so we don't re-use them.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
(cherry picked from commit 850925910f)
Because sometimes it's useful to know what distro a recording was made on, and
the kernel version doesn't always reveal that.
Fixes#428
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
event->time ended up being an uninitialized field. Introduced in 5dc1a7e, the
event here isn't a struct input event but rather our internal event struct.
Fix this and reshuffle the time handling a bit so it's a bit more obvious
here.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
If we want to record multiple events, let's just specify multiple event nodes.
No need for a specific extra argument here.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Positive side-effect - this exposed a bunch of missing #includes that got
pulled in by other headers before.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
The first event we receive is set to a 0ms offset anyway. Setting last_ms to 0
on the first event means the first two events have +0ms offset printed to the
log. Skip it, so the second event has the right offset.
This is human-readable data only, no effect on the recording file itself.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Now that all device quirks are in the quirks subsystem we have to print those
instead of just the udev devices.
Since libinput-record is there to record system devices, the system-installed
quirk list is used (without any commandline overrides right now). This is
useful to capture misconfigurations or missing quirks on the host system.
Fixes https://gitlab.freedesktop.org/libinput/libinput/issues/58
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Leftover from a previous version where printing and handling an event was
identical. Now we may handle events but not actually print them until a bit
later, so other events may have a (wrong) zero timestamp.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Common problem: some touch sequence does something to confuse libinput but it
cannot easily be captureed. The result is a long sequence of touche that need
to be picked apart and isolated.
Print an easy-to-search for message in the evdev output that signals that the
device touch state is now neutral (i.e. no finger down). Same can be achieved
by searching for BTN_TOOL_FINGER but that provides false positives for
switching between one and two fingers.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Collect libinput events together with the evdev events and print them to the
log. This makes it possible to debug the full behavior of a user's machine
rather than having to replay it with potential different race conditions/side
effects.
Example event output:
- evdev:
- [ 2, 314443, 4, 4, 57] # EV_MSC / MSC_SCAN 57
- [ 2, 314443, 1, 57, 1] # EV_KEY / KEY_SPACE 1
- [ 2, 314443, 0, 0, 0] # ------------ SYN_REPORT (0) ---------- +87ms
libinput:
- {time: 2.314443, type: KEYBOARD_KEY, key: 57, state: pressed}
- evdev:
- [ 2, 377203, 4, 4, 57] # EV_MSC / MSC_SCAN 57
- [ 2, 377203, 1, 57, 0] # EV_KEY / KEY_SPACE 0
- [ 2, 377203, 0, 0, 0] # ------------ SYN_REPORT (0) ---------- +63ms
libinput:
- {time: 2.377203, type: KEYBOARD_KEY, key: 57, state: released}
Note that the only way to know that events are within the same frame is to
check the timestamp. libinput keeps those intact which means we can tell that
if we just had an evdev frame with timestamp T and get a pointer motion with
timestamp T, that frame caused the motion event.
So far, only key, pointer and touch events are printed. We also
hardcode-enable tapping where available until we have options to enable this
on the commandline just because that's useful to have.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Only the ones we care about in libinput but for those it's handy to know which
ones are set (especially the LIBINPUT_MODEL ones).
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
This is a tool that does effectively the same job as evemu-record.
evemu has two disadvantages: its API is clunky and hard to extend even for
simple features. And it has a custom file format that requires special
processing but is difficult to extend and hard to write manually. e.g. the
bitmasks require keeping a line number state to know which bit an entry refers
to.
libinput-record records the same data but the output is YAML. That can be
processed easier and extended in the future without breaking the parsing. We
can (in the future) also interleave the evemu output with libinput's debug
output, thus having a single file where the events can be compared and
analysed without the need for replaying. Likewise, we can easily annotate the
file with parsable bits of information without having to shove all that into a
comment (like version numbers of libinput, kernel, etc).
User-visible differences to evemu-record:
* the output file requires an explicit -o or --output-file argument
* no evemu-describe equivalent, if you just want the description simply cancel
before any events are sent
* to see key codes, a --show-keycodes flag must be supplied, otherwise all
'normal' keys end up as KEY_A. This protects against inadvertent information
leakage
* supports a --multiple option to record multiple devices simultaneously. All
recordings have the same time offset, it is thus possible to reproduce bugs
that depend on the interaction of more than one device.
And to answer the question of: why a printf-approach to writing out yaml
instead of a library, it's simply that we want to be able to have real-time
output of the recording.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
