-
Notifications
You must be signed in to change notification settings - Fork 40
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Crashing on the first launch of a patched App #7
Comments
Cannot reproduce on my phone. |
I removed the debug Lspatch manager (b409cd3), deleted all modules, deleted the patched application, cleared the entire cache on the phone, rebooted, downloaded and installed the latest action (b409cd3) manager-v0.6-418-debug.apk, downloaded the telegram apk from the official site, patched it through Lspatch, and the crash repeats on first launch. I would like to mention that if you use the original Lspatch (not a fork) for the patch, the patched application will not launch at all, it will just crash) |
Now I checked what @harshitshah4 did from #6, I downloaded b8c4664, and there is no more crash on startup 👀 |
@JingMatrix, I hope you will make a fix in the next build, since the last build b409cd3 is currently broken, the previous one b8c4664 works correctly, in any case, thank you very much for your work, the fact that you support Lspatch/Lsposed, all the hopes of the community are on you at the moment <3 |
One question, this "b409cd3" work in this moment for me but the "debug" version, the "release" version don't work for the app netease cloud music, why? i can open the app but wont load the module in the app using the "release" version This "b8c4664" also the "release" version don't work for me. android 14 and google play system release "August" Thanks |
@mario6714 Thank you for the information, I will investigate the Release version. If possible, please upload some related logs. |
Could not reproduce your issue with the Release version. |
Issue not fixed, b409cd3 not working at release and debug versions, please note
|
It is possible that libart.so was updated in the August Google Play system update, this is part of the Android Runtime (ART), and perhaps that is why the crash occurs |
After Lspatch tries to load modules for the first time (whether modules are installed or not), the patched application crashes
Here is full debug from logcat
The second launch is successful and the application starts successfully and initializes the modules
|
@skxev Do you mean that LSPatch worked at expected after the second launch ? |
An application patched through LSpatch (without modules) crashes on the first launch, all subsequent times it is launched without crashing, if you mean the LSpatch application itself, then it does not crash |
@skxev According to your description, the first crash of the patched app is indeed normal. |
@JingMatrix I understood your explanation, I closed the issue, but I didn’t understand why there are no crashes in build b8c4664 when first launched, but in the latest build b409cd3 there are and you say that this is normal |
@skxev That part surprises me a bit as well. I will inverstigate it further to see if there is a reasonable explanation. |
How i can do that? because in the debug seem work but in the release not |
In my case the main application loaded fine after the first crash, but is not operational due to a service (possibly in a separate process?) failing.
Is this related? The class Also
Though I have no idea why this is happening, does not seem to be related to LSPatch on the surface. |
@kmod-midori If you suspect that it is the problem of renaming the classname, you may try the debug version to test your hyperthesis. |
The exception is gone in debug version, but the problem is still there. The same application + module combination worked fine in upstream LSPatch, maybe I should investigate further. |
@kmod-midori by worked fine, you meant no crash for the first lauch ? Or you meant the latest build doesn't work for you at all? |
To reliably reproduce this issue: Clear the app data and launch it. |
Both? Upstream LSPatch did not crash at all (in my memory). I came here because the recent Google update basically destroyed upstream LSPosed and LSPatch (all the patched apps do not launch). However, the latest build here have a totally different problem: if I patch QQ with the latest build (integrated mode, without any module), the application launches, but a background process ( I can confirm that the exact same base app (QQ 9.0.60 6478) worked flawlessly with or without any module on upstream LSPatch. Here is log file after clearing app data, include the first (crash) and second (runs but not very well) launch. |
Mmm, step by step for create the lspatch.log? i already clean the data and open the app but cant find the lsptach.log in my phone, where is the file? |
One question, how i can update that file in the lspatch? for test |
@mario6714 You don't need provide the log if you don't know how to. Currently, a first crash is normal behavior, so you can safely ignore this issue. |
@kmod-midori For your problem of QQ, open a new issue. It is not related to here.
|
is ok use the debug version? because the release is not working for me The problem is, release version wont open the app just crash, no matter how much i try to open the app and the debug i think work fine |
@mario6714 Yes, you can surely keep using the debug version. |
Funny error, in the "release" last version, the app open fine but is like the module does not exists jaja |
Information from the Official Lsposed telegram channel, here translate: |
as I wrote the problem is due to the Google system update, I hope for a fix from @JingMatrix |
Based on commits in this repo, this fork already contains necessary changes to work around that ART change, but along the way introduced some compatibility problems. |
oh, that's true, i forgot that @JingMatrix made important changes in recent builds, due to which LSpatch now works on new system updates, but everything is still not perfect, there are bugs like crash on first launch |
This will cause LSPatch to crash at first launch, see #7
However, HookFunction couldn't be replaced by HookArtFunction, which should be investigated deeper. Temporarily revert commit 2a18d73 of JingMatrix/LSPlant to fix #7.
However, HookFunction couldn't be replaced by HookArtFunction (#7). LSPlt doesn't work well with LSPatch, and we should keep using Dobby.
However, we cannot use PLT .inline_hooker for LSPatch (#7).
To avoid a crash in the fisrt launch (#7), we cannot use PLT .inline_hooker to DisableBackgroundVerification.
The reason for a first crash is found. For second launch, there is no more remaining verification function in the memory, so PLT hook won't cause crash. |
This comment was marked as outdated.
This comment was marked as outdated.
To avoid a crash in the fisrt launch (#7), we cannot use PLT .inline_hooker to DisableBackgroundVerification. Remove ccache to avoid difference between CI builds and local builds.
Installed 2681111 and everything working without any crashes or error, thanks @JingMatrix for work |
To avoid a crash in the fisrt launch (#7), we cannot use PLT .inline_hooker to DisableBackgroundVerification. Remove ccache to avoid difference between CI builds and local builds.
Steps to reproduce/复现步骤
Using latest debug build b409cd3, still crashing on first launch
Expected behaviour/预期行为
After the August Google Play system update, the application patched via Lspatch (b409cd3) crashes on first launch.
Actual behaviour/实际行为
JNI DETECTED ERROR IN APPLICATION: JNI NewGlobalRef called with pending exception java.lang.NoClassDefFoundError: Class not found using the boot class loader; no stack trace available
Xposed Module List/Xposed 模块列表
LSPatch version/LSPatch 版本
418
Android version/Android 版本
14
Shizuku version/Shizuku 版本
13.5
Version requirement/版本要求
Apk file/Apk 文件
No response
Logs/日志
logcat-sAndroidRuntimeLSPatchLSPosedLSPatch-MetaLoaderLSPatch-SigBypassLSPosed-BridgeLSPlantLSPosedContextzygisk64LSPltDobbyLSPosedService.__F._20240905105521.txt
The text was updated successfully, but these errors were encountered: