watch appears twice in this list because we're waiting on upstream bluez to merge the class detection for it
Signed-off-by: Stacy Harper <contact@stacyharper.net>
I kept getting 'Not implemented' in my status bar, output from
sxmo_status.sh, which was caused by sxmo_audio.sh and a alsamicismuted()
outputting "Not implemented" to stdout, effectively obscuring most of
the tiny status bar. I redirected the output to stderr..
Ideally of course we should fix the function, but this will have to do
in the meantime.
Signed-off-by: Peter John Hartman <peterjohnhartman@gmail.com>
I believe these were used in the early days of sxmo to reset the
connection to the modem. eg25-manager now handles this.
Signed-off-by: Peter John Hartman <peterjohnhartman@gmail.com>
It used to be used to update the clock on the desktop widget, but that's
been replaced with sxmo_aligned_sleep.
Signed-off-by: Peter John Hartman <peterjohnhartman@gmail.com>
This scaling value should be specific to devices and screen sizes. I use
sxmo on my destkops and I dont want this value by default from the
swayinitconf that I cant manage.
This restores the key repeat settings that were used with the multikey
script. Without them long-pressing the power button is unreliable.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
This gives us prettier logs. I'm not sure if it breaks anything, since
it will prepend the date and name of shell script to the output now.
Signed-off-by: Anjandev Momi <anjan@momi.ca>
This allows inputhandler to finish before the notification is updated,
which makes using the volume gesture is more responsive.
Signed-off-by: Anjandev Momi <anjan@momi.ca>
People have trouble writing a device profile when the touchscreen
doesnt work. The touchscreen doesnt work so you can't connect to wifi
to ssh in to write the device profile.
To solve this chicken and egg problem, this patch makes sure touchscreen
atleast works by introducing an "unknown" device. I
purposefully went against the linux guidelines for devicenames so that
the name will not conflict with an actual device.
Test procedure:
remove device profile for poco f1 in git repo and from
/usr/share/sxmo/default_hooks /usr/bin/
`doas make install` with this patch
Reboot and the touchscreen atleast still works.
Can someone please test this patch with a onebutton device? It doesnt
matter if the buttons dont work. Lets make sure the touchscreen atleast
works without a device profile.
Signed-off-by: Peter John Hartman <peterjohnhartman@gmail.com>
1. Move the wifi enable/disable to Networks (more convenient)
2. Do not list wifi networks if wifi is disabled.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
This patch is *just* an implementation of the pulse audio menu.
(The bugfix for call mode is now a separate patch, so too the
additions to statusbar.)
1. Implements the sxmo_audio.sh menu for pulse completely using ONLY
pactl (so we don't need pamixer).
2. Cleans up the InCall menu.
A further patch will remove extra alsa functions.
This solution is based on playerctl daemon/client common setups. This is
harmless if the user doesnt have the required binaries or dont have
the daemon running.
Right now all it does is monitor for headphone in/out and update
the statusbar, but this can be useful for other things too.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
Rationale: It gets hard providing support with sxmo being packaged on multiple distros and different
arches using a slightly different stack (superd - pmOS vs systemd - arch/debian, bonsaid -
aarch64 and x86_64 vs no bonsaid - 32bit)
Being able to tell users to run one command when reporting an issue will
help us narrow down whether an issue is a regression, packaging mistake,
or user error.
TODO: when bonsai has version numbering, print it here
Signed-off-by: Peter John Hartman <peterjohnhartman@gmail.com>
This should make some things more intuitive. I'm not happy with having
four entries for Vibrate/Ring options. We should really only have two
toggles: Vibrate, and Ring. But for now this will do.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
With bonsai_tree.json I tried to use a hack to migrate it, using only
the file diff. It was incomplete cause it doest use "interactive" mode
so the .need-migration always remains.
To fix this correctly, first I reverted the defaultconfig function
behavior. Then I adapted checkconfigversion to handle ref files without
explicit versions. For those cases, we also just use diff exit status.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
Signed-off-by: Anjandev Momi <anjan@momi.ca>
Systemd starts a dbus session bus automatically, and runs user services
on it. When we create our own session bus, we can't talk to those
services.
Signed-off-by: Anjandev Momi <anjan@momi.ca>
On sway on ppp the modem got a dedicated device input on sway. It
trigger wake up on some event which break sxmo workflows.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
Signed-off-by: Anjandev Momi <anjan@momi.ca>
This reverts commit df41312cd5.
Now that we freed lisgd and bonsaid process from forked logs, we dont
suffer the issue from superd.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
Signed-off-by: Anjandev Momi <anjan@momi.ca>
Superd fail to manage correctly bonsai and lisgd daemons atm. It doesnt
stop/restart them correctly cause they start async process and their
pipes are used.
This is a superd issue but this break sxmo for a too long time. This
one could be reverted when superd is fixed.
To fix the problems mentionned in the following ticket I worked on new
solution to handle complex key bindings on mobile devices:
https://todo.sr.ht/~mil/sxmo-tickets/524
This solution is structured as client/server state machine in hare lang.
It read a json transition tree and wait for the client to send events or
to change contexts.
Installing bonsai is ofc required :
$ doas apk add bonsai
Dont forget to $ sxmo_migrate.sh sync interactive to setup the json in
your config folder.
I adapted the migration scripts to not break the json. We cannot
versionise it so let's just check if the user one differ from the sxmo
default one.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
Signed-off-by: Anjandev Momi <anjan@momi.ca>
Calling map_to_output while rotating seems to cause a bug. Anyway we
only have to call those one time so let's move this in
sxmo_swayinitconf.
Signed-off-by: Stacy Harper <contact@stacyharper.net>
Signed-off-by: Anjandev Momi <anjan@momi.ca>
Without the check, creation of new wifi-hotspots will silently fail with
keys that are either too short or too long.
Signed-off-by: Maarten van Gompel <proycon@anaproy.nl>
