Describe the bug
After the last update, Ferdi does not open by itself at the beginning session; then, I try to open it by the terminal and I got that bug:
❯ ferdi
(electron) The default value of app.allowRendererProcessReuse is deprecated, it is currently "false". It will change to be "true" in Electron 9. For more information please check https://github.com/electron/electron/issues/18397
[1] 2135 segmentation fault (core dumped) ferdi
After trying it three times, Ferdi finally opens.
To Reproduce
Steps to reproduce the behavior:
I updated it by yay, I installed it by yay too.
Thanks.
Expected behavior
Screenshots
Desktop:
Debug information: https://debug.getferdi.com/...
Additional context
I also have similar error - core dump, when trying to login to Franz account
Related: #667
Try installing the binary package ferdi-bin, from AUR.
The binary package worked for me for a few hours, but after I rebooted it stopped working again
Installing the ferdi package (not the binary but the 5.5.0-2 package) from the AUR seems to have fixed it for now, let's see if it is consistent
It isn't, after a reboot it started crashing again
@obvionaoe
The binary package worked for me for a few hours, but after I rebooted it stopped working again
Strange, for me the binary is still stable. Rebooted a couple timies to be sure. I think you were on a slightly higher Manjaro kernel, but other that that idk:
Output of inxi -Sx
System: Host: dell-laptop Kernel: 5.4.35-1-MANJARO x86_64 bits: 64 compiler: gcc v: 9.3.0
Desktop: KDE Plasma 5.18.4 Distro: Manjaro Linux
Mine here:
System: Host: joselito-asus-rog Kernel: 5.4.35-1-MANJARO x86_64 bits: 64 compiler: gcc
v: 9.3.0 Desktop: Deepin 15.10.1 Distro: Manjaro Linux
Same experience here, different kernel and DE:
Host: ChanceBox Kernel: 5.5.18-3-MANJARO x86_64 bits: 64 compiler: gcc v: 9.3.0 Desktop: KDE Plasma 5.18.4
Distro: Manjaro Linux
If I remember, I'll try it later with an older kernel and see what happens.
Disabling systray seems to get ferdi not to segfault.
Going to .config/Ferdi/config/settings.json, then changing enableSystemTray to false made Ferdi open at least.
So, it seems like the problem is with the SystemTray.
These electron issues look like they might be related to this: https://github.com/electron/electron/issues/22137, https://github.com/electron/electron/issues/22215, https://github.com/electron/electron/issues/22652
I get the same issue:
$ inxi -Sx
System: Host: laurin-pc Kernel: 5.6.11-1-MANJARO x86_64 bits: 64 compiler: gcc v: 9.3.0 Desktop: KDE Plasma 5.18.5
Distro: Manjaro Linux
$ ferdi
(electron) The default value of app.allowRendererProcessReuse is deprecated, it is currently "false". It will change to be "true" in Electron 9. For more information please check https://github.com/electron/electron/issues/18397
[1] 5198 segmentation fault (core dumped) ferdi
and, like others, it works when disabling systray
For me, the ferdi 5.5.0-4 and ferdi-git AUR packages exit with segmentation fault, but the ferdi-bin 5.5.0-1 package (based on the Fedora RPM) works fine. So maybe the culprit is the electron 8.2.5-2 from Arch vs whatever Electron version is currently used for building the RPM.
Also, this commit, mentioning the issue linked by @dpeukert, looks interesting: https://github.com/zulip/zulip-desktop/pull/909/commits/cbc89a72a202d6155d79d44af2430398d98fdcfa#diff-32b8ac9380f7cf04c01491e8512d188b.
Would it be sensible to revert the upgrade from Electron 8.1.1 to 8.2.3 done via https://github.com/getferdi/ferdi/pull/625?
@kytwb I don't think that's the correct course of action, because using an older version of Electron potentially introduces security vulnerabilities. Even then, the AUR packages are building Ferdi with the Electron installed on the system, which would likely be the latest version.
Probably it's better to work around the issue until it's fixed in upstream Electron. I gather from the Zulip commit that that the crash is caused by creating a Tray icon without setting its context menu immediately, although I am yet to understand the Zulip or the Ferdi codebases well enough to grok why that's the case.
This issue has been automatically marked as stale because it has not had recent activity. Please check if this issue is still relevant and please close it if it's not. This will make sure that our open issues are actually of use and reduce the list of obsolete issues. Thank you for your contributions.
This can probably be closed, right?
My system now runs Electron v9.1.0 and I reinstalled the AUR ferdi (5.5.0-7), I have not experienced any crashes/segfaults.
I rebuilted ferdi modifying the AUR PKGBUILD to go back to native electron (9.1.0) and it seems the segfault problem is indeed solved.
Unrelated but as a piece of information, you might want to save some time and apply the patch by @schinfo for the stuck loading screen bug that still affects electron9, no noticeable problems after that.
This issue has been automatically marked as stale because it has not had recent activity. Please check if this issue is still relevant and please close it if it's not. This will make sure that our open issues are actually of use and reduce the list of obsolete issues. Thank you for your contributions.
Most helpful comment
Disabling systray seems to get ferdi not to segfault.
Going to
.config/Ferdi/config/settings.json, then changingenableSystemTraytofalsemade Ferdi open at least.