https://bugs.kde.org/show_bug.cgi?id=500339

--- Comment #29 from Ardo Kirsipuu <[email protected]> ---
I'm also having the issue of sometimes not being able to log in (Enter/button
doesn't do anything). Here are my findings after several days of debugging:

None of the following will reproduce it always, it fails rarely (~5%), works
most of the time:
* I can reproduce it when it's initiated by timeout (e.g. 1 minute) - different
timings and other settings like grace period doesn't seem to affect it
* I cannot reproduce when I initiate it myself by via Meta+L (gave it some good
~100 attempts)
* However, I can reproduce it by directly executing from command line
`/usr/libexec/kscreenlocker_greet`, which is strange

Some of the workarounds I found when the problem hits (no restart needed):
* "Switch User" button exits the lock screen (kscreenlocker) and opens the
log-in view (SDDM), where the logging in works
* Clear the input field and wait for 10 seconds until the inputs get hidden and
then try again

This led me to investigate
`/usr/share/plasma/shells/org.kde.plasma.desktop/contents/lockscreen/LockScreenUi.qml`
file and I tried to add some "logging" in there. Being new to Linux, didn't
know where the console.log would be found + writing to file is a bit restricted
in there = so I just created a new UI element to dump my "logs" into.

Initially I thought there might be some race condition in there with how the
states of `uiVisible` and `blockUI` are handled in the event handler functions,
but currently my hypothesis is that the bug is not in that file, because:

* `onPasswordResult` gets called
* when it fails then `onSucceeded` nor `onFailed` ever get called (is there a
third option besides success and fail?)

I don't have any insight what happens or should happen between
`onPasswordResult` and `onSucceeded`/`onFailed`, hopefully somebody knows
better and can make some sense out of this all.

PROBABLY IMPORTANT: If the `authenticator.startAuthenticating()` is also added
to the `onBlockUIChanged` event handler, then I could not reproduce the bug
anymore for some reason. Maybe because the probability goes to 0.05*0.05=0.0025
and my ~100 attempts wasn't enough to reveal it? I guess better understanding
of the backend is needed here.

I have also attached journalctl logs, my custom LockScreenUi.qml file and a
screenshot of it: There are multiple successful attempts (for comparison)
before 02:30 and then the screenshot is about the failure at 02:30:29 (+ 3
additional attempts to press Enter or click the button) followed by the
workaround of clearing the input and waiting for 10 second timeout for the UI
to disappear to try again (and succeed).

Oh, and here's my system info:
Operating System: Fedora Linux 42
KDE Plasma Version: 6.4.0
KDE Frameworks Version: 6.15.0
Qt Version: 6.9.1
Kernel Version: 6.14.11-300.fc42.x86_64 (64-bit)
Graphics Platform: Wayland
Processors: 2 × AMD A6-9210 RADEON R4, 5 COMPUTE CORES 2C+3G
Memory: 16 GiB of RAM (15.1 GiB usable)
Graphics Processor: AMD Radeon R4 Graphics
Manufacturer: LENOVO
Product Name: 80S9
System Version: Lenovo YOGA 510-14AST

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to