Posts by heitbaum

It looks like the client has changed the “failure logic”

[MDEV-28634] Client's --ssl-* options (without --ssl-verify-server-cert) are silently ignored if TLS is not possible - Jira
>> The client will only see an error if they also provide the --ssl-verify-server-cert option:<<

I believe the correct fix would be to set sslMode=disabled in the connection from Kodi - as in https://mariadb.com/kb/en/using-tl…java-connector/ (will be the same with c connector I believe)
By default when sslMode is set (not disabled), connector will use "serverSslCert" is set or the default truststore if not. Using default truststore can be disable setting option "fallbackToSystemTrustStore" to false.

Can you confirm it forces ssl or forces ssl verification?

these are the changes

Comparing v3.3.10...v3.4.0 · mariadb-corporation/mariadb-connector-c

MariaDB Connector/C is used to connect applications developed in C/C++ to MariaDB and MySQL databases.The client library is LGPL licensed. - Comparing…

github.com

and this one is the force ssl verification.

make DEFAULT_SSL_VERIFY_SERVER_CERT a cmake option · mariadb-corporation/mariadb-connector-c@82983a3

github.com

Here are the ssl stuff from mariadb and there seems to be some require per user. Without seeing the logs I could not comment further. But is your connection running ssl (and just without the verify?) and it fails with the connector 3.4 (due to the verify) - the comments should be pretty useful in isolating this.

I don’t understand the conversation around the shared library. There has never been a link to /usr/lib for the shared library in LE. The other binaries are correctly linked to the shared library as can be shown by using ldd. I have tried running boblightd via trace and can’t see the library being imported dynamically. Can you please share where you think the dynamic library is being used.

I can see the process running as below (which wouldn’t run if a library is missing.) I can’t run it myself as I don’t have the serial interface that boblightd communicates too.

nuc12:~ # systemctl status service.multimedia.boblightd | cat
● service.multimedia.boblightd.service - boblightd
Loaded: loaded (/storage/.config/system.d/service.multimedia.boblightd.service; enabled; preset: disabled)
Active: active (running) since Mon 2024-05-06 13:37:48 UTC; 22s ago
Main PID: 1958390 (boblightd)
Tasks: 2 (limit: 36590)
Memory: 856.0K (peak: 1.5M)
CPU: 32ms
CGroup: /system.slice/service.multimedia.boblightd.service
└─1958390 boblightd -c /storage/.kodi/userdata/addon_data/service.multimedia.boblightd/boblight.conf

May 06 13:37:48 nuc12 systemd[1]: Started service.multimedia.boblightd.service.
May 06 13:37:48 nuc12 sh[1958414]: cp: can't stat '/storage/.kodi/addons/service.multimedia.boblightd/config/*.sample': No such file or directory

I’m definitely not a bob light expert. I’m assuming that the default /dev/ttyACM0 is the serial port - if that is the same error - then would suggest checking what serial port is on your LE install.

It looks like it doesn’t link to the shared library.

nuc12:~ # ldd .kodi/addons/service.multimedia.boblightd/bin/boblightd
       linux-vdso.so.1 (0x00007fff1bbfd000)
       libusb-1.0.so.0 => /usr/lib/libusb-1.0.so.0 (0x00007ffb71a43000)
       libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x00007ffb717de000)
       libm.so.6 => /usr/lib/libm.so.6 (0x00007ffb71702000)
       libgcc_s.so.1 => /usr/lib/libgcc_s.so.1 (0x00007ffb716d5000)
       libc.so.6 => /usr/lib/libc.so.6 (0x00007ffb71508000)
       libudev.so.1 => /usr/lib/libudev.so.1 (0x00007ffb714c8000)
       /lib64/ld-linux-x86-64.so.2 => /usr/lib64/ld-linux-x86-64.so.2 (0x00007ffb71a5a000)

I have run the boblightd process and “it seems to run”

Quote

# .kodi/addons/service.multimedia.boblightd/bin/boblightd -c /storage/.kodi/userdata/addon_data/servi
ce.multimedia.boblightd/boblight.conf
(InitLog) start of log /storage/.boblight/boblightd.log
(PrintFlags) starting .kodi/addons/service.multimedia.boblightd/bin/boblightd -c /storage/.kodi/userdata/addon_data/service.multimedia.boblightd/boblight.conf
(CConfig::LoadConfigFromFile) opening /storage/.kodi/userdata/addon_data/service.multimedia.boblightd/boblight.conf
(CConfig::CheckConfig) checking config lines
(CConfig::CheckConfig) config lines valid
(CConfig::BuildConfig) building config
(CConfig::BuildConfig) built config successfully
(main) starting devices
(CClientsHandler::Process) opening listening socket on 127.0.0.1:19333
(CClientsHandler::Process) ERROR: bind() 127.0.0.1:19333 Address already in use
(CDevice::Process) device1: starting with output "/dev/ttyACM0"
(CDevice::Process) device1: setting up
(CDeviceRS232::OpenSerialPort) ERROR: device1: /dev/ttyACM0: open() No such file or directory
(CDevice::Process) ERROR: device1: setting up failed, retrying in 10 seconds
^C(SignalHandler) caught SIGINT
(main) signaling devices to stop
(CClientsHandler::Cleanup) disconnecting clients
(CClientsHandler::Cleanup) closing listening socket
(CClientsHandler::Cleanup) clients handler stopped
(main) waiting for devices to stop
(CDevice::Process) device1: closed
(CDevice::Process) device1: stopped
Segmentation fault (core dumped)

Display More

Quote from mglae

Change download source:

Diff

diff --git a/packages/devel/commons-text/package.mk b/packages/devel/commons-text/package.mk
index c0277ca900..654ba2112a 100644
--- a/packages/devel/commons-text/package.mk
+++ b/packages/devel/commons-text/package.mk
@@ -6,7 +6,7 @@ PKG_VERSION="1.11.0"
 PKG_SHA256="4169cb90571fb28fad4c5eea7c1c994c18f1995452f73e8ea7a86087c0e3822e"
 PKG_LICENSE="Apache-2.0"
 PKG_SITE="https://commons.apache.org/proper/commons-text/"
-PKG_URL="https://dlcdn.apache.org/commons/text/binaries/commons-text-${PKG_VERSION}-bin.tar.gz"
+PKG_URL="https://archive.apache.org/dist/commons/text/binaries/commons-text-${PKG_VERSION}-bin.tar.gz"
 PKG_DEPENDS_TARGET="toolchain"
 PKG_LONGDESC="Apache Commons Text is a library focused on algorithms working on strings"
 PKG_TOOLCHAIN="manual"

Display More

You beat me to it https://github.com/LibreELEC/LibreELEC.tv/pull/8825

It’s using the systemd dynamic user. You can see this in /usr/lib/systemd/system/wsdd2.service; and visibility for review in LE13 - https://github.com/LibreELEC/LibreELEC.tv/pull/8705

Quote from jonian

Thank you for looking at this heitbaum! Unfortunately I can't test that image since I have a Raspberry Pi 5.

I’m not sure why I thought your image was Generic. I have compiled a RPi5 test image https://heitbaum.libreelec.tv for you.

ehoitinga this updates bluez to potentially address a DualShock BT device issue. You may want to test this for your scenario

LE12 image with bluez 5.75+. https://heitbaum.libreelec.tv/LibreELEC-Gene…-f46fff4.img.gz includes https://github.com/LibreELEC/LibreELEC.tv/pull/8820

jonian LE12 image with bluez 5.75+ for you to test. https://heitbaum.libreelec.tv/LibreELEC-Gene…-f46fff4.img.gz includes https://github.com/LibreELEC/LibreELEC.tv/pull/8820

Quote from jonian

Regarding my bluetooth issue with a DualShock controller, it has been reported in bluez issue tracker here. LE 12 has the same version (5.73) as the one reported having this problem. Latest bluez version is 7.75 but it is not clear to me if the fix is included. There is a workaround mentioned here, but I'm not sure how to apply it on LE. Maybe the best solution is to downgrade bluez.

See https://github.com/LibreELEC/LibreELEC.tv/pull/8820 - I’m building an image for you to test.

It is supported on Generic-legacy with chrome. No other architectures use X11 anymore which is one of the requirements.

Quote from Antibios

Does anyone know if the halt in builds since 4/6 for all models is development related, or website related, or something else?

Nightlies only get built when there are changes at https://github.com/LibreELEC/LibreELEC.tv - not much changing at the moment as we are in the beta phase of LE12

Patch is now in nightlies.

Fox243 pr raised - will be discussed. https://github.com/LibreELEC/LibreELEC.tv/pull/8761

I’ll draft PR it. Need to discuss with the team on a holistic fix. You are ok to use that image for the time being as it was built using the same build process as our nightly/official images (just with the patch.)

Hi Da Flex - all good. Pretty sure this is the opposite audio=ok and video=bad - I have had it and this is the tracking issue on Generic - https://github.com/LibreELEC/LibreELEC.tv/issues/8488

Can you please test these 2 images.

https://test.libreelec.tv/12.0/Generic/Generic/LibreELEC-Generic.x86_64-12.0-nightly-20240323-353213e.img.gz
- should not work for you

https://heitbaum.libreelec.tv/LibreELEC-Generic.x86_64-12.0-devel-20240323213541-353213e.img.gz
- should work for you (it has the fences PR reverted) I think