Fixes 02dfe83c4b
The previous `provides="mkinifs=3.6.0"` was a requirement because alpine's
kernels had a minimum dependency in `mkinitfs>=3.6.0` to make sure that
support for compressed modules existed. That requirement is no longer
in place[1], since the dependency is now on the `initramfs-generator`
virtual package. Also, since `initramfs-generator` is a virtual package
and not a real package, versioning it might mess up APK. The documentation
explictly states that virtual packages are those with unversioned provides
and a provider_priority. So to avoid having problems in the future, let's
strictly stick to it.
[1] 4fac61c826
Alpine provides a 'initramfs-generator' thing now that the alpine
linux-* depend on explicitly. This sets postmarketos-mkinitfs to be the
preferred provider of that virtual package.
Fixes:
ERROR: unable to select packages:
postmarketos-mkinitfs-1.4.1-r5:
conflicts: mkinitfs-3.6.1-r2[mkinitfs=3.6.0]
satisfies: postmarketos-base-20-r1[postmarketos-mkinitfs]
postmarketos-base-nofde-20-r1[postmarketos-mkinitfs]
.pmbootstrap-20220610.215815[postmarketos-mkinitfs]
.pmbootstrap-20220610.215815[mkinitfs]
mkinitfs-3.6.1-r2:
conflicts: postmarketos-mkinitfs-1.4.1-r5
satisfies: .pmbootstrap-20220610.215815[mkinitfs]
linux-rpi-5.15.45-r1[initramfs-generator]
.pmbootstrap-20220610.215815:
masked in: cache
satisfies: world[.pmbootstrap=20220610.215815]
Also see:
30d7b88324 (aports) main/mkinitfs: provide initramfs-generator
4fac61c826 (aports) main/linux-lts: replace mkinitfs with initramfs-generator
There was lately a discussion upstream due to apk 2.12.10 breaking
some packages in aports[1]. The conclusion in IRC is that provides
were designed to be either versioned or have a provider_priority.
Otherwise, the provides will be treated as a "virtual package".
Which is not what we want here.
[1] https://gitlab.alpinelinux.org/alpine/apk-tools/-/issues/10847
The patch was meant well. But there are problems with yelp in its
current form in postmarketOS:
* It's practically empty. In the default install, there is only one
application that has some sort of help text, which is Chatty. But
when you click on the entry, there is only one line of generic
information about the app. Not useful for users.
* Yelp is not adapted for small screens, the one line of help text is
cut off on the PinePhone.
So let's not ship it by default in edge and v22.06 for now.
This reverts commit 5ef50b12f2.
[ci:skip-vercheck]: modifying _pmb_recommends doesn't need a version
bump.
At the moment, mkinitfs loads some file system modules always (ext4),
some dynamically (brtfs) and some not at all (f2fs, vfat), expecting
them to be built-in.
To support generic kernels (e.g. from Alpine) properly, use the
partition type detection for both boot and root partition and always
load the necessary modules before mounting.
An increasing number of devices relies on the vfat file system in the
initfs (e.g. for UEFI boot). Having to add all the modules required for
it to work to deviceinfo_modules_initfs is annoying and inconsistent,
in particular for the additionally chosen "nls" modules. Since we
already bundle several file systems by default, let's just add vfat
as well to simplify this.
Ideally mkinitfs should detect the file systems chosen at install time
in the future and only include the necessary modules, but until then
this just adds a couple of kilobytes to the initramfs.
closes#1515
This patch HAS NOT been tested to see if it actually fixes#1515. I do
not have my pinephone keyboard with me and will not till v22.06 is out.
Can someone please test this patch before merging?
Note: this is being merged upstream but needs a refactor of the
Makefile. See:
https://lists.sr.ht/~mil/sxmo-devel/patches/32630
Since we cannot get the refactor of the Makefile out before the next
release, this is the next best thing
[ci:skip-build] already built successfully in CI.
This replaces the previous `udevadm test` hack with a combination of
`udevd`, `udevadm trigger` and `udevadm settle` in order to coldplug all
available devices and make them available for use via libinput / unl0kr.
Relates to: #1411
[ci:skip-build] already built successfully in CI
We decided against pvr hacks, quoting craftyguy from issue 1316:
> We discussed this in our team meeting, and ultimately decided that
> supporting pvr in pmaports/pmOS has many of the same problems that
> we saw when trying to support hallium/libhybris, something we
> decided against. For much of the same reasons, we've decided that we
> want to avoid adding the proprietary userspace components required to
> support pvr in pmOS.
No device port is using this package anymore, but we forgot to remove
it. After branching v22.06 I realized that it doesn't build anymore
and as stated in MR 2702, it's unmaintained.
Related: https://builds.sr.ht/~postmarketos/job/767700
Fix the build failure when built with newer gcc:
/home/pmos/build/src/gcc-6.4.0/gcc/config/i386/i386.c: In function 'int function_arg_advance_32(CUMULATIVE_ARGS*, machine_mode, const_tree, long int, long int)':
/home/pmos/build/src/gcc-6.4.0/gcc/config/i386/i386.c:8743:18: error: converting to 'bool' from 'std::nullptr_t' requires direct-initialization [-fpermissive]
8743 | bool error_p = NULL;
| ^~~~
Patch source: https://svnweb.freebsd.org/ports/head/lang/gcc7/files/patch-gcc_config_i386_i386.c?view=markup&pathrev=545751
Fix the build failure when built with newer gcc:
/home/pmos/build/src/gcc-4.9.2/gcc/reload1.c:89:24: error: use of an operand of type 'bool' in 'operator++' is forbidden in C++17
89 | (this_target_reload->x_spill_indirect_levels)
| ~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~
/home/pmos/build/src/gcc-4.9.2/gcc/reload1.c:444:7: note: in expansion of macro 'spill_indirect_levels'
444 | spill_indirect_levels++;
| ^~~~~~~~~~~~~~~~~~~~~
Patch source: https://github.com/ChrisMacGregor/esp-open-sdk/blob/master/1001-fix-reload1-compile-error.patch
This should be reverted when this Alpine issue has been closed:
https://gitlab.alpinelinux.org/alpine/aports/-/issues/13778
sxmo-common already had this for some reason.
[ci:ignore-count]
[ci:skip-build] Already built successfuly on CI in MR
the networkmanager package in alpine was refactored and the wifi and
wwan components were split off into separate subpackages. without these,
wifi and wwan networking are broken.
also see: aports 830af82c6c
[ci:ignore-count]
This code wasn't actually working for me on samsung-m0, so I
investigated and came up with that this fixes it. I'm not sure if you
are supposed to be able to omit the deviceinfo_ prefix? Either way, it
doesn't seem to work, so let's fix it.
Fixes fad7c7ef8e
XWayland is a phosh dependency. Having it in pmb_recommends adds
nothing and can be confusing, since power users would expect that
removing it from world would remove the package. That will never
be the case since phosh depends on it.
[ci:skip-build] already built successfully in CI
Adding bluez to "depends" in order to enable bluetooth service (doesn't
work on packages in "_pmb_recommends").
Adding blueman to "_pmb_recommends" to allow uninstallation by the
user.
Closes https://gitlab.com/postmarketOS/pmaports/-/issues/1465
nmcli and nmtui are still missing in Gnome.
Move NetworkManager dependencies from postmarketos-ui-phosh, and phosh
will get them when depending on postmarketos-ui-gnome.
postmarketos-ui-i3wm|mate.post-deinstall:
Contains removal of file "61-autologin.conf"|"65-autologin.conf".
However, this file is removed by apk automatically unpon
deinstallation.
postmarketos-ui-i3wm|mate|xfce4.pre-deinstall:
Contains disabling service "lightdm". Removing display manager service
upon deinstallation isn't done by other UIs either. It can lead to
unexpected behaviour when installing another UI using lightdm and then
uninstall MATE UI.
Additionally
postmarketos-ui-i3wm.post-upgrade:
Removed postmarketos-ui-i3wm.post-upgrade and replaced it with a link to
postmarketos-ui-i3wm.post-install, like it's done in MATE and Xfce4.
I usually end up installing it on new installs. Why not enable it by
default and have a more polished experience for users? Most users will
use a gtk app and/or use the gtk file chooser
mesa-egl is usually installed by dependencies of ui packages, but when
choosing the "none" ui option in pmbootstrap, it isn't selected, leading
to libEGL.so.1 not being found.
Add it as an explicit dependency.
[ci:skip-build]: already built successfully in CI
Add and rearrange dependencies with new apps, and fix angelfish package
name.
Adds the following packages to extras:
- alligator (rss reader)
- kasts (podcasts)
- tokodon (mastodon client)
Move neochat from a required to an extra dependency, since it's a
matrix client (which many people may not need).
Add plasma-mobile-sounds to main dependency, so that kclock (and dialer
in the future) have sounds to choose from.
Rename plasma-angelfish to angelfish dependency since plasma-angelfish
was renamed in alpine at some point.
Remove the part of the description of the package that says numeric
passwords are required (no longer the case).
When moving the code to it's own repo I changed the license to GPL-2.0,
I forgot to update the APKBUILD.
[ci:skip-build] already built successfully in CI
Shared package for shared gnome and phosh ui configurations
Additionally, added missing dependencies:
Both Phosh and GNOME need elogind to do sleep inhibition and to
provide pam modules. gnome-bluetooth has some small udev rule
that helps marking devices for rfkill. Given that most devices
in pmOS have a bluetooth, this should be useful as a dependency.
These were most likely copied from phosh, but not necessary in
GNOME. They safe space in the interface, but GNOME Shell is really
not that useful in very narrow devices, so adding a configuration
for such cases should be unnecessary.
The dconf configuration and installation file was needed because the
schemas `sm.puri.phoc.application.org-gnome-maps` and
`sm.puri.phoc.application.firefox` are not schemas that can be overriden.
The actual schema is `sm.puri.phoc.application`, which is a relocatable
schema. Therefore, we could use an override to specify the default,
but with a simple configuration file, gsettings cannot relocate the
schema. Therefore, the dconf configuration makes sense for these kind
of schemas. For regular schemas that can be overriden with an override
file, it is simpler to just use it.
Phosh takes care of it in a smarter way[1]. The custom override also has the
consequence of never showing closing buttons in docked mode, which can be
an inconvenience.
[1] https://gitlab.gnome.org/World/Phosh/phosh/-/blob/main/src/docked-manager.c#L292-L295
Or in case the lines change with time:
```C
phosh_docked_manager_set_enabled (PhoshDockedManager *self, gboolean enable)
if (enable)
g_settings_reset (self->wm_settings, WM_KEY_LAYOUT);
else
g_settings_set_string (self->wm_settings, WM_KEY_LAYOUT, "appmenu:");
```
Phosh already installs these schemas by default[1] and makes them specific
to the Phosh desktop. Originally these might have been needed because
either Phosh didn't ship the overrides or/and tinydm wasn't setting the
environment variable XDG_CURRENT_DESKTOP properly[2]. To verify that the
behavior of the Phosh override is working as expected one can do the following
from an ssh session (so that the full DE envvars are not set):
```bash
$ export DISPLAY=:0 # so that gsettings reset works
$ gsettings reset org.gnome.desktop.interface clock-show-date
$ gsettings get org.gnome.desktop.interface clock-show-date # false
$ sudo rm /usr/share/glib-2.0/schemas/000-postmarketos.gschema.override
$ sudo glib-compile-schemas --strict /usr/share/glib-2.0/schemas/
$ gsettings get org.gnome.desktop.interface clock-show-date # true (which is the default)
$ export XDG_CURRENT_DESKTOP=GNOME:Phosh
$ gsettings get org.gnome.desktop.interface clock-show-date # false (from phosh override)
```
[1] https://gitlab.gnome.org/World/Phosh/phosh/-/blob/main/data/00_sm.puri.Phosh.gschema.override
[2] https://gitlab.com/postmarketOS/tinydm/-/merge_requests/12
firefox scaling was set to true before `mobile-config-firefox` existed.
In new installations false is the default, so explicitly stating false is
only necessary for upgrades. Since the default was changed in one and a half
years ago in fef3af0d75 when the
package was still in version 4, removing it should not have any impact. If
anybody had an edge installation not upgrade for more than 1.5 years, it would
definitely be wiser to reinstall than to upgrade.
Glade can be useful if you're building GTK apps on your phone when
making use of convergence, so let's unhide it. The reason it was
hidden was that it gets pulled in when building Phosh, but this
might be solvable in a better way.
Hiding of org.gnome.Extensions can probably be dropped when
https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/31086
has been merged.
Enable screen locking and embed virtual keyboard. For devices with a
hardware keyboard, this is unfavorable. For touch devices, however,
that's needed for unlocking the screen.
Additionally enable single-click behavior in file manager. This applies
also to the desktop icons. Again this is unfavorable for conventional
devices but advantageous for touch devices.
unclutter-xfixes version 1.6 introduced new options --hide-on-touch
and --start-hidden. That seems appropriate enough to cover the use
cases touchscreen usage (hide cursor) and mouse/touchpad usage (show
cursor).
The main change is to move the postmarketOS-specific Xfce4 configs
from /etc/skel to /etc/xdg. This improves the config handling because
the location in /etc/xdg can be defined as the default fallback
configs.
Moving the gtk-3.0 folder to /etc/xdg did not work. Therefore this
one remains in /etc/skel.
Beyond this, there are some minor changes from the xfce4-phone
repository:
- Change icon path of whiskermenu button to new location
- Apply pmOS wallpaper to more monitors
- Several clean-ups in the config files
- Project naming scheme is changed to "lk2nd-msmXXXX"
- lk1st-msm8916: Experimental support for running as the "first" lk
- Secondary core bringup on non-psci firmware is now done by lk2nd
- rproc selection is now handled by lk2nd
- Some display information improvements
- New devices and various improvements for existing ones
[ci:skip-build]: already built successfully in CI
Fixes error: postmarketos-base-18-r0: trying to overwrite etc/fstab owned by alpine-baselayout-data-3.2.0-r20.
caused by 9ecba8a514
[ci:skip-build] already built successfully in CI
This installs/starts udiskie, a udisk2 helper, that facilitates
auto-mounting of removable drives.
Note support for umounting is broken in the current Portfolio release,
so disks have to be umounted manually:
udiskie-umount /run/media/<user>/*
I decided to add this to the `depends` and not `pmb_recommends`, because
I felt like 1) this is a feature that folks would expect to work, 2) by
having it in depends, it'll 'fix' existing installs where mounting
removable disks automatically is not implemented
MMS support (via mmsd-tng) involves sending/receiving network
requests/responses over the wwan interface. If it's ipv4-only and the
device is connected to some other ipv4 network on another iface (like
wifi), this can cause the rp_filter to reject responses on wwan iface
because it incorrectly thinks they are martian packets.
This does theoretically disable some "security" feature in the kernel,
but it's worth noting that:
1) rp_filter isn't implemented at all in the kernel for ipv6
2) other distros (mobian, pureos at least) are also disabling rp_filter
3) this seems to be a relatively common problem with folks using mms on
pmOS, since many carriers' data networks are ipv4-only
also see:
https://gitlab.com/kop316/mmsd/-/merge_requests/55/diffs?commit_id=b22c253fb939ff1eb949ea4e628706e6a28c851a
[ci:skip-build] already built successfully in CI
ondev-boot-mount.sh expects the 'real' losetup (it uses --show), instead
of the busybox version (which doesn't support that parameter).
Alpine broke this in 31372fab76 when they split out losetup from
util-linux
[ci:skip-build]: already built successfully in CI
Allow calling the script with:
SPLASH_CONFIG=$srcdir/splash.ini devicepkg_build $startdir $pkgname
In order to pass a custom splash config, as it will be used with the
PineNote (MR 2910).
The postmarketOS icon was moved to a new location (see issue
pmaports#1418). Therefore the button of whiskermenu in the panel
becomes empty. This commit creates a symlink at the old location
to restore the icon on existing Xfce4 installations.
[ci:skip-build] already built successfully in CI
This copies rules from /etc/udev/rules.d as well which ensures that
device-specific udev rules (e.g. for Nokia N900) are available in the
initfs.
Relates to: #1411
Prevent apk from pulling in pipewire-pulse instead of pulseaudio, as
this leads to various audio issues with Phosh in postmarketOS (volume
buttons not working, no sound in various applications).
Pipewire will be a great replacement for pulseaudio, but let's rather
enable it intentionally for Phosh once we fixed everything that's
currently breaking with it. Without this patch it may get installed when
installing or upgrading other packages (not entirely sure how it gets
triggered yet, but I have seen it myself multiple times and we've gotten
reports about it).
[ci:skip-build] already built successfully in CI
