Closing device network crashes Jami with error: *** buffer overflow detected ***: jami terminated
Describe your environment
- OS: Linux Debian 10 Buster
- GNOME: 3.30.2, spec: 1.2
- Qt runtime version: 6.2.3
- Jami package "libqt-jami": 6.2.3-2
- Jami repository:
https://dl.jami.net/stable/debian_10 jami/main amd64 jami amd64 <JAME.VERSION>
- Jami version: 20230323.0~dfsg1-1
- What build you are using: dl.jami.net stable
Steps to reproduce
-
Can you reproduce the bug: at will
-
Steps:
-
Using Debian 10 Buster, using GNOME 3.30.2, using Network, open the internet
-
Open Jami 20230323.0~dfsg1-1 It successfully opens and works.
-
Using Debian 10 Buster, using GNOME 3.30.2, Using Network, Close (Éteindre) the internet connection. This first screenshot below shows this.
This second screenshot below shows that the internet is now closed.
-
Leave Jami open. But leave the Network closed. Wait up to 24 hours. The duration you need to wait for depends on various factors. Such as, but not limited to, the resource of your device, CPU, memory, version, etc.
-
Somehow, Jami automatically crashes and exit. This is the first challenge. The needed end result is that Jami does not automatically crash and exit when the device network is closed or not available.
-
The second challenge is that, after Jami exited above, Jami did not notify the user that it exited. Let's called this first user John. His Jami account is a busy supplier reception with heavy traffic. So John presently believes that his Jami is still open. Because he did not close Jami. In other words, John waits for clients to contact him via Jami instant message (IM). Using another device, and another Jami, let's call this second user Jenn, a client, sends IM to John. John won't receive Jenn's IMs. Because his Jami crashed hours ago. But he still does not know this. John wonder why it is so quiet today at the usually busy reception. After a few days, complaints from clients are piling up. As John is not replied to their IM. Using another communication platform than Jami, someone notified John that somehow he is not replying to clients' IM. John investigate. Now realized that for days Jami was not open. Somehow it crashed. John re-open Jami, waves of pooled IMs are coming in. The needed end result is that if somehow Jami exits because of a
*** buffer overflow detected ***: jami terminated
. Before automatically closing itself, Jami somehow notify the user that it is doing so. For example, but not limited to, a popup window. Which communicate to the user the challenge about this*** buffer overflow detected ***: jami terminated
. And asked the user to either click onTry again to detect network
button orExit Jami
button.
-
Additional information
This log shows Jami between its successfully opening to its clashes and exits.
Warning: Ignoring WAYLAND_DISPLAY on Gnome. Use QT_QPA_PLATFORM=wayland to run on Wayland anyway.
Using Qt runtime version: 6.2.3
"notify server name: gnome-shell, vendor: GNOME, version: 3.30.2, spec: 1.2"
qt.webenginecontext:
GLImplementation: desktop
Surface Type: OpenGL
Surface Profile: NoProfile
Surface Version: 3.1
Using Default SG Backend: yes
Using Software Dynamic GL: no
Using Angle: no
Init Parameters:
* application-name Jami
* browser-subprocess-path /usr/lib/libqt-jami/libexec/QtWebEngineProcess
* disable-features DnsOverHttpsUpgrade,ConsolidatedMovementXY,InstalledApp,BackgroundFetch,WebOTP,WebPayments,WebUSB,PictureInPicture
* disable-setuid-sandbox
* disable-speech-api
* disable-web-security
* enable-features NetworkServiceInProcess,TracingServiceInProcess
* enable-main-frame-before-activation
* enable-threaded-compositing
* gpu-preferences <MASKED.FOR.PRIVACY>=
* in-process-gpu
* lang fr-CA
* num-raster-threads 4
* single-process
* use-gl desktop
"Using locale: fr_CA"
09:53:30.620 os_core_unix.c !pjlib 2.12.1 for POSIX initialized
Daemon is running
No migration required
Syncing lrc accounts list with the daemon
Screen saver dbus interface: "org.freedesktop.ScreenSaver"
nmea: No serial ports found
Main window loaded using OpenGLRhi
Qt Quick Layouts: Detected recursive rearrange. Aborting after two iterations.
NetworkManager client initialized, version: 1.14.6 , daemon running: yes , networking enabled: yes
primary network connection: <MASKED.FOR.PRIVACY> default: yes
lookup name is INVALID: "e" "18a397bfea388197"
lookup name is INVALID: "em" "18a397bfea388197"
no primary network connection detected, check network settings
no primary network connection detected, check network settings
*** buffer overflow detected ***: jami terminated
Abandon
In the above log, <MASKED.FOR.PRIVACY> means that this data was manually removed for privacy