Menu

Main Menu
Talk Get Daily Search

Member's Online

    User Name
    Password

    Flatpak on Sailfish OS

    Reply
    Page 10 of 44 | Prev |   8     9   10   11     12   20 | Next | Last
    coderus | # 91 | 2020-02-29, 11:28 | Report

    Just installed flatpak to my XA2
    Code:
    [nemo@Sailfish ~]$ flatpak-runner org.kde.mobile.angelfish
    Starting: org.kde.mobile.angelfish
    [D] unknown:0 - Using Wayland-EGL
    Wayland socket: ../../display/wayland-1
    WAYLAND_DISPLAY="../../display/wayland-1" FLATPAK_MALIIT_CONTAINER_DBUS="unix:abstract=/tmp/dbus-hxDgBBD9wQ,guid=aabaee73b085b4a4461ed6015e5a4a22" flatpak run --env=QT_WAYLAND_FORCE_DPI=289 --filesystem=/system:ro --filesystem=/vendor:ro --filesystem=/odm:ro --device=all --talk-name=org.maliit.server --env=QTWEBENGINE_CHROMIUM_FLAGS=--disable-gpu-compositing --num-raster-threads=1 --enable-viewport --disable-composited-antialiasing --env=QTWEBENGINE_DISABLE_GPU_THREAD=1 --env=QT_QUICK_CONTROLS_MOBILE=1 --env=QT_QUICK_CONTROLS_STYLE=Plasma --env=HYBRIS_EGLPLATFORM_DIR=/usr/lib/arm-linux-gnueabihf/GL/host/lib/libhybris --env=HYBRIS_LINKER_DIR=/usr/lib/arm-linux-gnueabihf/GL/host/lib/libhybris/linker --env=HYBRIS_LD_LIBRARY_PATH=/usr/lib/arm-linux-gnueabihf/GL/host/libexec/droid-hybris/system/lib:/usr/libexec/droid-hybris/system/lib:/vendor/lib:/system/lib --env=LD_LIBRARY_PATH=/usr/lib/arm-linux-gnueabihf/GL/host/lib org.kde.mobile.angelfish
    [D] onCompleted:424 - Warning: specifying an object instance for initialPage is sub-optimal - prefer to use a Component
    [W] unknown:52 - file:///usr/share/flatpak-runner/qml/main.qml:52:26: Unable to assign [undefined] to bool
    i see window with angelfish logo, spinning busyindicator, and nothing happens

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 4 Users Say Thank You to coderus For This Useful Post:
    Amboss, juiceme, mosen, rinigus

     
    rinigus | # 92 | 2020-02-29, 11:50 | Report

    Originally Posted by coderus View Post
    Just installed flatpak to my XA2
    Code:
    [nemo@Sailfish ~]$ flatpak-runner org.kde.mobile.angelfish
    Starting: org.kde.mobile.angelfish
    [D] unknown:0 - Using Wayland-EGL
    Wayland socket: ../../display/wayland-1
    WAYLAND_DISPLAY="../../display/wayland-1" FLATPAK_MALIIT_CONTAINER_DBUS="unix:abstract=/tmp/dbus-hxDgBBD9wQ,guid=aabaee73b085b4a4461ed6015e5a4a22" flatpak run --env=QT_WAYLAND_FORCE_DPI=289 --filesystem=/system:ro --filesystem=/vendor:ro --filesystem=/odm:ro --device=all --talk-name=org.maliit.server --env=QTWEBENGINE_CHROMIUM_FLAGS=--disable-gpu-compositing --num-raster-threads=1 --enable-viewport --disable-composited-antialiasing --env=QTWEBENGINE_DISABLE_GPU_THREAD=1 --env=QT_QUICK_CONTROLS_MOBILE=1 --env=QT_QUICK_CONTROLS_STYLE=Plasma --env=HYBRIS_EGLPLATFORM_DIR=/usr/lib/arm-linux-gnueabihf/GL/host/lib/libhybris --env=HYBRIS_LINKER_DIR=/usr/lib/arm-linux-gnueabihf/GL/host/lib/libhybris/linker --env=HYBRIS_LD_LIBRARY_PATH=/usr/lib/arm-linux-gnueabihf/GL/host/libexec/droid-hybris/system/lib:/usr/libexec/droid-hybris/system/lib:/vendor/lib:/system/lib --env=LD_LIBRARY_PATH=/usr/lib/arm-linux-gnueabihf/GL/host/lib org.kde.mobile.angelfish
    [D] onCompleted:424 - Warning: specifying an object instance for initialPage is sub-optimal - prefer to use a Component
    [W] unknown:52 - file:///usr/share/flatpak-runner/qml/main.qml:52:26: Unable to assign [undefined] to bool
    i see window with angelfish logo, spinning busyindicator, and nothing happens
    It depends on whether you have libhybris version >= 0.0.5.34 on XA2. Unless it's some kind of development device with future SFOS version running on it, you are out luck on XA2. Hopefully, next SFOS release will include that libhybris.

    As for spinning disk, logo and absence of action. Don't forget that we have SFOS app (flatpak-runner) which starts first, then flatpak app is started with Wayland display pointing to the display created by flatpak-runner. Hence, spinning disk is just provided by flatpak-runner itself. I guess in XA2 case, it gets stuck with some hybris trouble and never finishes the flatpak process.

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 4 Users Say Thank You to rinigus For This Useful Post:
    Amboss, juiceme, ljo, mosen

     
    coderus | # 93 | 2020-03-01, 08:18 | Report

    i have 0.0.5.34 on my XA2

    what can be debugged?

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 4 Users Say Thank You to coderus For This Useful Post:
    Amboss, juiceme, ljo, rinigus

     
    rinigus | # 94 | 2020-03-01, 10:59 | Report

    Originally Posted by coderus View Post
    i have 0.0.5.34 on my XA2

    what can be debugged?
    OK, then something is broken. As we seem to have AOSP9 devices (Sony Tama family and Pro1?) working so far, maybe its due to the other base.

    Let's start with the simple test (assuming that all got installed and you did reboot after that):

    Code:
    flatpak run --command=sh org.kde.mobile.angelfish
    That should bring you to the command prompt. As its not depending on hybris, it will show if the flatpak is running at all.

    I wonder if you run flatpak-runner without any arguments (just from launcher) to initialize hybris extension? It is maybe easy to miss in the instructions.

    If flatpak is working (shell in it is accessible), flatpak-runner has generated extension (ls -l ~/.local/share/flatpak/extension/org.freedesktop.Platform.GL.host), then we probably miss somethin in extension.

    Check whether I miss any filesystem that should be there for libhybris in https://github.com/sailfishos-flatpa...runner.cpp#L83 for your device.

    Next, we can run strace on angelfish and see where it breaks. For that, install Sdk:

    Code:
    flatpak install --user org.kde.Sdk//5.14
    Then close flatpak-runner and let's start with the plain flatpak for simplicity (windows minimization will not work, but its not a problem as we don't have any window to start with). Command line is based on what flatpak-runner is composing and showing you in its terminal output as well as addition of -d --command=sh to get into development env

    Code:
    flatpak run --env=QT_WAYLAND_FORCE_DPI=335 --filesystem=/system:ro --filesystem=/vendor:ro --filesystem=/odm:ro --device=all --talk-name=org.maliit.server --env=QTWEBENGINE_CHROMIUM_FLAGS="--disable-gpu-compositing --num-raster-threads=1 --enable-viewport --disable-composited-antialiasing" --env=QTWEBENGINE_DISABLE_GPU_THREAD=1 --env=QT_QUICK_CONTROLS_MOBILE=1 --env=QT_QUICK_CONTROLS_STYLE=Plasma --env=HYBRIS_EGLPLATFORM_DIR=/usr/lib/arm-linux-gnueabihf/GL/host/lib/libhybris --env=HYBRIS_LINKER_DIR=/usr/lib/arm-linux-gnueabihf/GL/host/lib/libhybris/linker --env=HYBRIS_LD_LIBRARY_PATH=/usr/lib/arm-linux-gnueabihf/GL/host/libexec/droid-hybris/system/lib:/usr/libexec/droid-hybris/system/lib:/vendor/lib:/system/lib --env=LD_LIBRARY_PATH=/usr/lib/arm-linux-gnueabihf/GL/host/lib --command=sh -d org.kde.mobile.angelfish
    In Flatpak environment, check if you can get any useful info from strace:

    Code:
    strace /app/bin/angelfish
    Hopefully, strace will get stuck on some hybris call.

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 6 Users Say Thank You to rinigus For This Useful Post:
    Amboss, edp17, juiceme, ljo, mosen, olf

     
    coderus | # 95 | 2020-03-01, 13:15 | Report

    GL was generated after install then rebooted, shell works okay

    not sure what to do with filesystems?

    strace is full of ENOENT: https://gist.github.com/CODeRUS/add4...c89520736ab9e9

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 4 Users Say Thank You to coderus For This Useful Post:
    Amboss, juiceme, mosen, rinigus

     
    rinigus | # 96 | 2020-03-01, 13:55 | Report

    Originally Posted by coderus View Post
    GL was generated after install then rebooted, shell works okay

    not sure what to do with filesystems?

    strace is full of ENOENT: https://gist.github.com/CODeRUS/add4...c89520736ab9e9
    As long as it finds library later, ENOENT is OK.

    Looks like something is off with egl:

    Code:
    writev(6, [{iov_base="\0\7\0w\264[^\327C+!", iov_len=11}, {iov_base="\5", iov_len=1}, {iov_base="libEGL\0", iov_len=7}, {iov_base="eglInitialize(0x3726b78) failed "..., iov_len=48}], 4) = 67
    Anything relevant in dmesg/journal?

    As for filesystems, which ones are mounted on device? (df)

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 4 Users Say Thank You to rinigus For This Useful Post:
    Amboss, juiceme, ljo, mosen

     
    coderus | # 97 | 2020-03-01, 15:31 | Report

    both journal and dmesg are silent.

    Code:
    [nemo@Sailfish ~]$ df
    Filesystem           1K-blocks      Used Available Use% Mounted on
    /dev/sailfish/root     2488624   1366160   1080532  56% /
    devtmpfs               1263452       312   1263140   0% /dev
    tmpfs                  1351608         0   1351608   0% /dev/shm
    tmpfs                  1351608     21356   1330252   2% /run
    tmpfs                  1351608         0   1351608   0% /sys/fs/cgroup
    tmpfs                  1351608         8   1351600   0% /tmp
    tmpfs                  1351608         0   1351608   0% /mnt
    /dev/mmcblk0p71         840320    276924    538772  34% /opt
    /dev/mmcblk0p68         380360    211172    156900  57% /odm
    /dev/mmcblk0p42         112592     97936     14656  87% /firmware
    /dev/mmcblk0p40          65488       496     64992   1% /bt_firmware
    /dev/mmcblk0p2           28144       444     27048   2% /persist
    /dev/mmcblk0p44          12016      7484      4208  64% /dsp
    /dev/mapper/sailfish-home
                          18131468   4322680  13612804  24% /home
    tmpfs                   270324       820    269504   0% /run/user/100000
    /dev/mmcblk1p1        15549952      8128  15541824   0% /run/media/nemo/3334-3431

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 4 Users Say Thank You to coderus For This Useful Post:
    Amboss, juiceme, mosen, rinigus

     
    rinigus | # 98 | 2020-03-01, 18:38 | Report

    Originally Posted by coderus View Post
    both journal and dmesg are silent.

    Code:
    [nemo@Sailfish ~]$ df
    Filesystem           1K-blocks      Used Available Use% Mounted on
    /dev/sailfish/root     2488624   1366160   1080532  56% /
    devtmpfs               1263452       312   1263140   0% /dev
    tmpfs                  1351608         0   1351608   0% /dev/shm
    tmpfs                  1351608     21356   1330252   2% /run
    tmpfs                  1351608         0   1351608   0% /sys/fs/cgroup
    tmpfs                  1351608         8   1351600   0% /tmp
    tmpfs                  1351608         0   1351608   0% /mnt
    /dev/mmcblk0p71         840320    276924    538772  34% /opt
    /dev/mmcblk0p68         380360    211172    156900  57% /odm
    /dev/mmcblk0p42         112592     97936     14656  87% /firmware
    /dev/mmcblk0p40          65488       496     64992   1% /bt_firmware
    /dev/mmcblk0p2           28144       444     27048   2% /persist
    /dev/mmcblk0p44          12016      7484      4208  64% /dsp
    /dev/mapper/sailfish-home
                          18131468   4322680  13612804  24% /home
    tmpfs                   270324       820    269504   0% /run/user/100000
    /dev/mmcblk1p1        15549952      8128  15541824   0% /run/media/nemo/3334-3431
    Doesn't look from your strace that the missing mounts are missed. I am pretty much blank at this point. Maybe we can learn from comparison with some Sailfish app strace when its started.

    I presume that the window of Angelfish did not appear in your test.

    Also, maybe some folder needs to be exported in / . Is there anything obvious there?

    As for no records in dmesg and journal - I wonder where do those writes with egl failed go (https://gist.github.com/CODeRUS/add4...r-strace-L9132)

    Let's get the strace and ask for help in the porters channel. Maybe someone has an idea on what's wrong with XA2 setup.

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 3 Users Say Thank You to rinigus For This Useful Post:
    Amboss, juiceme, mosen

     
    coderus | # 99 | 2020-03-01, 20:50 | Report

    got this from /system/bin/logcat:
    Code:
    03-01 20:50:12.930  7981     5 D vndksupport: Loading /vendor/lib/hw/gralloc.default.so from current namespace instead of sphal namespace.
    03-01 20:50:12.931  7981     5 E cutils-trace: Error opening trace file: No such file or directory (2)
    03-01 20:50:12.931  7981     5 D vndksupport: Loading /vendor/lib/egl/libEGL_adreno.so from current namespace instead of sphal namespace.
    03-01 20:50:12.932  7981     5 D libEGL  : loaded /vendor/lib/egl/libEGL_adreno.so
    03-01 20:50:12.938  7981     5 D vndksupport: Loading /vendor/lib/egl/libGLESv1_CM_adreno.so from current namespace instead of sphal namespace.
    03-01 20:50:12.939  7981     5 D libEGL  : loaded /vendor/lib/egl/libGLESv1_CM_adreno.so
    03-01 20:50:12.952  7981     5 D vndksupport: Loading /vendor/lib/egl/libGLESv2_adreno.so from current namespace instead of sphal namespace.
    03-01 20:50:12.952  7981     5 D libEGL  : loaded /vendor/lib/egl/libGLESv2_adreno.so
    03-01 20:50:12.969  7981     5 I Adreno  : QUALCOMM build                   : 1dcb581, I26dffed9a4
    03-01 20:50:12.969  7981     5 I Adreno  : Build Date                       : 01/10/18
    03-01 20:50:12.969  7981     5 I Adreno  : OpenGL ES Shader Compiler Version: EV031.22.00.01
    03-01 20:50:12.969  7981     5 I Adreno  : Local Branch                     :
    03-01 20:50:12.969  7981     5 I Adreno  : Remote Branch                    : refs/tags/AU_LINUX_ANDROID_LA.UM.6.4.R1.08.00.00.309.049
    03-01 20:50:12.969  7981     5 I Adreno  : Remote Branch                    : NONE
    03-01 20:50:12.969  7981     5 I Adreno  : Reconstruct Branch               : NOTHING
    03-01 20:50:12.969  7981     5 D vndksupport: Loading /vendor/lib/hw/gralloc.default.so from current namespace instead of sphal namespace.
    03-01 20:50:12.969  7981     5 I Adreno  : OpenGrallocModuleAndDevice: Failed to open the Gralloc device, error: -22
    03-01 20:50:12.969  7981     5 W libEGL  : eglInitialize(0xbc110e8) failed (EGL_BAD_ALLOC)

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 6 Users Say Thank You to coderus For This Useful Post:
    Amboss, juiceme, ljo, mosen, olf, rinigus

     
    rinigus | # 100 | 2020-03-02, 06:38 | Report

    @coderus: thank you very much, at least some error message. I should have remembered about logcat...

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 5 Users Say Thank You to rinigus For This Useful Post:
    Amboss, coderus, juiceme, mosen, olf

     
    Page 10 of 44 | Prev |   8     9   10   11     12   20 | Next | Last
vBulletin® Version 3.8.8
Normal Logout