Bug 484541

Summary:	Tons error about missing "cfg_*" properties
Product:	[Plasma] Plasma SDK	Reporter:	orblazer <orblazer.minecraft+kde>
Component:	plasmoidviewer	Assignee:	Plasma Bugs List <plasma-bugs>
Status:	RESOLVED DUPLICATE
Severity:	minor	CC:	bugs.kde.org, cwo.kde, giorgos.tsiapaliokas, orblazer.minecraft+kde
Priority:	NOR
Version First Reported In:	master
Target Milestone:	---
Platform:	Neon
OS:	Linux
Latest Commit:		Version Fixed In:
Sentry Crash Report:

Description orblazer 2024-03-26 18:19:28 UTC

SUMMARY
Since Plama 6, the configuration page emits tons of error about missing `cfg_*` properties during page loading

STEPS TO REPRODUCE
1. run widget with: `plasmoidviewer -a org.kde.plasma.kicker`
2. open settings

OBSERVED RESULT
Errors in terminal :
```
file:///usr/share/plasma/shells/org.kde.plasma.plasmoidviewershell/contents/configuration/ConfigurationAppletPage.qml:54: Error: Cannot assign to non-existent property "cfg_objectName"
file:///usr/share/plasma/shells/org.kde.plasma.plasmoidviewershell/contents/configuration/ConfigurationAppletPage.qml:54: Error: Cannot assign to non-existent property "cfg_iconDefault"
...
file:///usr/share/plasma/shells/org.kde.plasma.plasmoidviewershell/contents/configuration/ConfigurationAppletPage.qml:54: Error: Cannot assign to non-existent property "cfg___33"
```

EXPECTED RESULT
No errors

SOFTWARE/OS VERSIONS
Linux/KDE Plasma: 6.0.2
(available in About System)
KDE Plasma Version: 6.0.2
KDE Frameworks Version: 6.0.0
Qt Version: 6.6.2

ADDITIONAL INFORMATION
This issue could be partially fixed with my MR https://invent.kde.org/plasma/plasma-sdk/-/merge_requests/121 because it's include MR https://invent.kde.org/plasma/plasma-desktop/-/merge_requests/1498.

BUT, this remove only non-config entries, the others errors (config entries) still remaining, like :
```
file:///usr/share/plasma/plasmoids/org.kde.plasma.kicker/contents/ui/ConfigGeneral.qml: Setting initial properties failed: ConfigGeneral does not have a property called cfg_alignResultsToBottomDefault
```

The solution for that should be have some conditions heres https://invent.kde.org/plasma/plasma-sdk/-/blob/c989f5fd99b46f38c07dd2970fa6a2f7046a1677/plasmoidviewer/qmlpackages/shell/contents/configuration/AppletConfiguration.qml#L125-127 (code from my MR) :
- exclude "...Default" (eg. `cfg_alignResultsToBottomDefault`) props
- exclude non-present props (the hardest due to seem impossible without loading twice the page)

Comment 1 ratijas 2024-03-26 18:22:15 UTC

That tends to happen when you build an architecture that relies on magic reflection a bit too much. Also QQmlComponent lacks needed introspection to determine the set of available properties of its loaded type.

Comment 2 orblazer 2024-03-26 22:49:21 UTC

(In reply to ratijas from comment #1)
> That tends to happen when you build an architecture that relies on magic
> reflection a bit too much. Also QQmlComponent lacks needed introspection to
> determine the set of available properties of its loaded type.

Yes and for explain the issue i have simply use an official plasmoid. In the idea the `cfg_*` is really useful but since we need dynamically load the component that become an nightmare to predict.

For me the only 2 ways to deal with that in QML (i think an another way exist with a bit of C++) is :
- load the page before set the props : It's not great because is generate re paint due to changes
- load once in "dummy" mode and filter variable from that (that also generate errors if the config page doesn't support "null" values)

An other way but it's create an big broken is to re change the way to manages settings and store in in single `QtObject` (but i think that could also make some issues with nested object (especially an flat array like `[0, 1, 2]`), if the serializing of `QtObject` is same of `ListModel`)

Comment 3 cwo 2025-02-18 20:08:10 UTC

Thanks for the bug report! This issue has some duplicates, and I'm marking this one a a duplicate of a more recently filed one as that has more current discussion; I've posted a comment there to point people also to the discussion here.

*** This bug has been marked as a duplicate of bug 494417 ***