bug: Force close when long pressing patch source text field

[J4CKED]

Bug description

When I add a source patch (long press in the “Source URL” field), the application closes.
The link I add: https://raw.githubusercontent.com/crimera/piko/refs/heads/dev/patches-bundle.json

Double-clicking in this field also closes the application.

Version of Morphe (Manager) and the app you are patching

v1.11.0-dev.1 (2026-03-03)

Morphe logs

No log, just fc.

Acknowledgements

• I have checked all open and closed bug reports and this is not a duplicate.
• I have chosen an appropriate title.
• All requested information has been provided properly.
• The bug only affects using Morphe (Manager)

[MarcaDian]

I can't reproduce it.

Please provide the logs after the crash via adb.
What phone and OS version do you have?

[MarcaDian]

By the way, you can add this link:

https://github.com/crimera/piko/tree/dev

[LisoUseInAIKyrios]

I can reproduce:

Process: app.morphe.manager, PID: 11087 java.lang.NullPointerException: Attempt to invoke virtual method 'void android.widget.TextView.setEllipsize(android.text.TextUtils$TruncateAt)' on a null object reference

Stack trace doesn't have any app code.

[MarcaDian]

@LisoUseInAIKyrios Strangely, the debug version doesn't have a crash, but the release version does. Proguard again?

[LisoUseInAIKyrios]

Yes, might be

[LisoUseInAIKyrios]

Maybe another local context error from it optimizing

[MarcaDian]

We can try adding a rule -dontoptimize that removes aggressive optimization rules. In theory, this should not have a significant impact on the size of the finished APK.

[MarcaDian]

We can try adding a rule -dontoptimize that removes aggressive optimization rules

I locally built the release app and it seems like it doesn't crash anymore. If I compare it with the previous size of the installed manager, the difference is 590 kb, which may be due to recent changes. I don't know how proguard works, so I can't judge whether this change is appropriate.

[LisoUseInAIKyrios]

We can try turning off only some proguard optimizations to find which are the problem.

It may be an inlining local variables optimization, which is breaking a local context or some other compose behavior.

Or we can just turn off all optimizations. The last time I checked they don't actually speed apps to any appreciable amount because the JVM already optimizes hotspots. I remember the author of proguard publicly mentioning somewhere that benchmarking with and without optimizations showed less than 1% performance improvement (or maybe it was 0.1%, I cant remember). The only tangible benefit of optimization is more obfuscation, but that doesn't mater when the source code is public.

Edit: List of optimizations. I think we should turn them all off and avoid all proguard optimization issues
https://www.guardsquare.com/manual/configuration/optimizations

[MarcaDian]

@J4CKED Please test this artifact

https://github.com/MorpheApp/morphe-manager/actions/runs/22638264847/artifacts/5747071534

[J4CKED]

@J4CKED Please test this artifact

https://github.com/MorpheApp/morphe-manager/actions/runs/22638264847/artifacts/5747071534

It's ok.