Fixed
Status Update
Comments
mi...@gmail.com <mi...@gmail.com> #2
Note: crash happens right after installation. No user interaction is needed, just installing APK.
xa...@google.com <xa...@google.com>
mi...@gmail.com <mi...@gmail.com> #3
Observation: crash happens also with new app built from Android Studio template.
Maybe it's happening only on Cyanogenmod/LineageOS ROMs... I'm not sure.
Maybe it's happening only on Cyanogenmod/LineageOS ROMs... I'm not sure.
je...@google.com <je...@google.com>
ch...@google.com <ch...@google.com> #4
Are all the crashes you see only happening only on Cyanogenmod? I am unable to reproduce it.
mi...@gmail.com <mi...@gmail.com> #5
I see "Android Vitals" and "ANRs and crashes", then screens like in attachments. So Google developer console tells me about crashes, but I don't see details of device ROM or such. But amount of these crashes is huge.
From direct user emails I could setup one configuration where it crashes for me - Cyanogenmod 12.1 on i9250
I can't tell if it's on other OEM ROMs.
Even if it was just on Cyanogenmod/LineageOS, I think it's not for ignoration, userbase of these ROMs is large.
Try to grab Samsung i9250, install CM12.1 on it, and you'll reproduce it.
From direct user emails I could setup one configuration where it crashes for me - Cyanogenmod 12.1 on i9250
I can't tell if it's on other OEM ROMs.
Even if it was just on Cyanogenmod/LineageOS, I think it's not for ignoration, userbase of these ROMs is large.
Try to grab Samsung i9250, install CM12.1 on it, and you'll reproduce it.
ch...@google.com <ch...@google.com> #6
When I try running X-plore_3.95.02-release.apk I get an xml parser exception instead of the native error you got. "Caused by: org.xmlpull.v1.XmlPullParserException: Binary XML file line #1: invalid drawable tag x" Other than the plugin version, is there anything else that was change between X-plore_3.95.01-release.apk and X-plore_3.95.02-release.apk? The assets look quite different.
Would it be possible to attach a toy repro project instead of just the apk?
Would it be possible to attach a toy repro project instead of just the apk?
mi...@gmail.com <mi...@gmail.com> #7
I don't get any XML parser exception. That APK normally works for masses, only it crashed some devices. There are maybe more differences since previous version, but changing Android Studio version is major change. Look at
META-INF/MANIFEST.MF inside APK, it has line
Created-By: Android Gradle 3.0.0-alpha7
that's important clue between non-crashing and crashing version.
And see comment #3 how it's possible to reproduce it with fresh new project. In android studio, click File, New, Project, and make simplest app with AS3. Maybe post your apk here and I'll write you if it crashes my test device.
META-INF/MANIFEST.MF inside APK, it has line
Created-By: Android Gradle 3.0.0-alpha7
that's important clue between non-crashing and crashing version.
And see
ke...@gmail.com <ke...@gmail.com> #8
The same to me
I'm using LineageOS 13 (Android 6.0.1) and I got the bootloop problem
I'm using LineageOS 13 (Android 6.0.1) and I got the bootloop problem
du...@gmail.com <du...@gmail.com> #9
I got an auto-uninstall then an OS crash (but fortuantely reboot without bootloop) on install & each attempt to launch. Cyanogenmod 13.0-20160721 (Android 6.0.1).
du...@gmail.com <du...@gmail.com> #10
PS. Tried it briefly on LineageOS 14.1-20170721 (Android 7.1.2) & no crash.
However, even though Cyanogenmod/LineageOS is more likely to be upgradeable than many Android adaptations, a change of Android major version usually involves some hassle of reinstalling & reconfiguring apps & data so it is likely many will remain on Android 6 for quite a while so this still needs Google to fix it before it hits other apps as they get updated.
However, even though Cyanogenmod/LineageOS is more likely to be upgradeable than many Android adaptations, a change of Android major version usually involves some hassle of reinstalling & reconfiguring apps & data so it is likely many will remain on Android 6 for quite a while so this still needs Google to fix it before it hits other apps as they get updated.
ha...@laposte.net <ha...@laposte.net> #11
I didn't update my device since 3 months. After updating many apps, my device has crashed in loop with this backtrace. I troubleshoot my device for hours until I find out the root cause :[
This issue is highly critical, please fix it !
My device : Kindle Fire HD (tate)
My OS: CM 12.1 (LP 5.1.x)
This issue is highly critical, please fix it !
My device : Kindle Fire HD (tate)
My OS: CM 12.1 (LP 5.1.x)
ra...@gmail.com <ra...@gmail.com> #12
I have two Motorola Photon Q 4G LTE's running Cyanogen Mod 12.1.201510o7-SNAPSHOT-YOG4PA339-xt897.
When Iinstall the update from Google Play, the phones crash and reboot at the end. After a looong bootup, the screen goes blank. I can access the power menu by pressing the power button, but neither Trebuchet nor Apex Pro Launchers will execute.
When Iinstall the update from Google Play, the phones crash and reboot at the end. After a looong bootup, the screen goes blank. I can access the power menu by pressing the power button, but neither Trebuchet nor Apex Pro Launchers will execute.
ra...@gmail.com <ra...@gmail.com> #13
I wanted to follow up and confirm that the app built with Gradle 2.x runs on both of my phones.
ge...@gmail.com <ge...@gmail.com> #14
Any update on this bug? I'll have to revert back to AS 2.3 to avoid this bug.
xa...@google.com <xa...@google.com> #15
We are not prioritizing fixing this for 3.0.
Adam, do you think the new aapt2 steps could have changed something that would make non standard android builds fail?
Adam, do you think the new aapt2 steps could have changed something that would make non standard android builds fail?
me...@googlemail.com <me...@googlemail.com> #17
On my Lineage for Galaxy Grand FINAL[2017/07/20] the following apps (all current versions) cause bootloop:
Greenify
MyJdownloader Remote
Facebook
IFTTT
Greenify
MyJdownloader Remote
IFTTT
ha...@laposte.net <ha...@laposte.net> #18
and now 3C toolbox Pro :(
The list will grow.
The list will grow.
xa...@google.com <xa...@google.com> #19
If you need to build with 3.0 but are running into this bug you can disable aapt2 using -Pandroid.enableAapt2=false on the command line when doing your build.
ha...@laposte.net <ha...@laposte.net> #20
I have tried the workaround. This option has no effect. The device still crashes with the same backtrace.
mi...@gmail.com <mi...@gmail.com> #21
The app list will grow. Even more when Studio 3 will get out of beta.
I believe this bug will be fixed sooner or later. It just needs to get enough weight.
I believe this bug will be fixed sooner or later. It just needs to get enough weight.
dr...@gtempaccount.com <dr...@gtempaccount.com> #22
CyanogenMod has this function getPkgName (https://github.com/CyanogenMod/android_frameworks_base/blob/cm-13.0/libs/androidfw/AssetManager.cpp ). It creates a ResXMLTree on the stack and points it at a buffer from an asset without making a copy. Then it closes the asset before the ResXMLTree is destroyed.
For apps built by aapt, this is benign. However, aapt2 produces UTF-8 string pools, which cause the ResXMLTree's ResStringPool(mStrings)'s mCache to become non-null in ResStringPool::stringAt (https://github.com/CyanogenMod/android_frameworks_base/blob/cm-13.0/libs/androidfw/ResourceTypes.cpp ). Then ResStringPool::uninit dereferences mHeader (which is now dangling), and a crash ensues.
This crash shows up in different ways. On one Cyanogen OS device, the launcher crashes when an app built with aapt2 has been installed, but only if the manifest is large (probably due to how deallocation happens for small vs large blobs). On another device, system_server crashes at boot if an aapt2-built app is installed.
We're attempting to work around this with a custom build of aapt2 that always produces a UTF-16 string pool for the manifest. Results so far are promising.
For apps built by aapt, this is benign. However, aapt2 produces UTF-8 string pools, which cause the ResXMLTree's ResStringPool(mStrings)'s mCache to become non-null in ResStringPool::stringAt (
This crash shows up in different ways. On one Cyanogen OS device, the launcher crashes when an app built with aapt2 has been installed, but only if the manifest is large (probably due to how deallocation happens for small vs large blobs). On another device, system_server crashes at boot if an aapt2-built app is installed.
We're attempting to work around this with a custom build of aapt2 that always produces a UTF-16 string pool for the manifest. Results so far are promising.
ha...@laposte.net <ha...@laposte.net> #23
Great ! Thank you so much for the investigation.
In fact, we had double checked again the workaround (see comment #20 ) :
We’ve cleaned-up the project, all cache and restarted studio in case the option was not considered properly
and finally the workaround works fine, no crash :)
In fact, we had double checked again the workaround (see
We’ve cleaned-up the project, all cache and restarted studio in case the option was not considered properly
and finally the workaround works fine, no crash :)
du...@gmail.com <du...@gmail.com> #24
Another mainstream app now bootloops on Cm Android 6: Yahoo Mail.
At least I now know how to work around without using ADB etc. : Boot into
TWERP recovery, file manager, /data/app, sort by date, delete those apps
updated recently & reboot. Assuming app data is backed up of course.
On 25 Sep 2017 21:56, <buganizer-system@google.com> wrote:
At least I now know how to work around without using ADB etc. : Boot into
TWERP recovery, file manager, /data/app, sort by date, delete those apps
updated recently & reboot. Assuming app data is backed up of course.
On 25 Sep 2017 21:56, <buganizer-system@google.com> wrote:
mi...@gmail.com <mi...@gmail.com> #25
@ comment #19 : where do you put the paramenter in Android Studio? I tried in Run/Debug configurations under Defaults tab to Gradle script parameters, but it produced binary identical APK as without it.
xa...@google.com <xa...@google.com> #26
Put the option under Preferences (or Settings) > "Build, Execution, Deployment" > Compiler
There's a field for command line options
There's a field for command line options
te...@gmail.com <te...@gmail.com> #27
I have a galaxy note 8 with marshmallow rom and cm theming and have xperienced bootloop with xplore file manager ad aitype keyboad latest update. I am concerned that this will happen with more and more, apps if it is not addressed. Please fix this!
te...@gmail.com <te...@gmail.com> #28
Downloaded X-plore_3.95.01-release.apk from first post and it woks on my note 8.
Downloaded X-plore_3.95.02-release.apk and when maneuvering to the download folder where the file is it crashed 2 different file managers. No bootloop but just stopped responding.
I then installed X-plore_3.95.01-release.apk from notifications and bootloop ocurred.
Seems just the apk being called upon in a file manager causes issues. I rebooted to recovery to delete X-plore_3.95.01-release.apk and the data entries and everything works now. Hope this helps in tracking the problem.
Downloaded X-plore_3.95.02-release.apk and when maneuvering to the download folder where the file is it crashed 2 different file managers. No bootloop but just stopped responding.
I then installed X-plore_3.95.01-release.apk from notifications and bootloop ocurred.
Seems just the apk being called upon in a file manager causes issues. I rebooted to recovery to delete X-plore_3.95.01-release.apk and the data entries and everything works now. Hope this helps in tracking the problem.
te...@gmail.com <te...@gmail.com> #29
Note: es file explorer and smart file manager both crash when trying to open downloads with X-plore_3.95.01-release.apk present. But not with solid explorer (able to delete it from there without having to go to recovery) - don't know why but maybe its another clue as to what is happening.
te...@gmail.com <te...@gmail.com> #30
Sorry! I got the files mixed up above. To be clear it is X-plore_3.95.02-release.apk that causes all the Issues mentioned above and not X-plore_3.95.01-release.apk.
ha...@laposte.net <ha...@laposte.net> #31
If you encounter this bug, contact the application author and show him this bug report. He can fix this issue by using the workaround at comment #19 .
di...@gmail.com <di...@gmail.com> #32
Do I understand correctly that Google does not want to fix this? Currently, solid file explorer, Flipboard, L-speed are crashing ANDROID as soon as installed, and nobody cares?
te...@gmail.com <te...@gmail.com> #33
I sent all this info to developer of xplore file manger and he immediately fixed it and sent me an updated apk. I do hope google takes this seriously so no one has to fix anything as far as this bug is concerned.
ad...@google.com <ad...@google.com> #34
Just to be clear, this is a device that introduced a memory corruption bug, which when running an app built with aapt2, happens to be exposed. The device is running with undefined behavior regardless of which tool you use.
We can fix this so as not to trigger the worst case scenario, but there is no guarantee that this won't be broken later by subsequent changes, like legal ordering changes in the resource table.
That being said, I will be writing a CTS test to ensure that OEMs who introduced this "feature" fix the bug.
We can fix this so as not to trigger the worst case scenario, but there is no guarantee that this won't be broken later by subsequent changes, like legal ordering changes in the resource table.
That being said, I will be writing a CTS test to ensure that OEMs who introduced this "feature" fix the bug.
mi...@gmail.com <mi...@gmail.com> #36
Confirmed, no more crashing with build tools 27.0.1
ma...@gmail.com <ma...@gmail.com> #37
"
Just to be clear, this is a device that introduced a memory corruption bug, which when running an app built with aapt2, happens to be exposed. The device is running with undefined behavior regardless of which tool you use.
We can fix this so as not to trigger the worst case scenario, but there is no guarantee that this won't be broken later by subsequent changes, like legal ordering changes in the resource table.
"
I don't understand... Is this a software or a hardware issue? If it's a software issue, why do you mention that "a device introduced a memory corruption bug"?
Can this be fully fixed in software? The second paragraph suggests that no, but it's not clear to me why not...
Just to be clear, this is a device that introduced a memory corruption bug, which when running an app built with aapt2, happens to be exposed. The device is running with undefined behavior regardless of which tool you use.
We can fix this so as not to trigger the worst case scenario, but there is no guarantee that this won't be broken later by subsequent changes, like legal ordering changes in the resource table.
"
I don't understand... Is this a software or a hardware issue? If it's a software issue, why do you mention that "a device introduced a memory corruption bug"?
Can this be fully fixed in software? The second paragraph suggests that no, but it's not clear to me why not...
mi...@gmail.com <mi...@gmail.com> #38
This is software problem, and the software which introduced the bug is Cyanogenmod (now LineageOS), popular alternative Android ROM that users can decide to install on their devices.
The bug was hidden until Android SDK updated its AAPT build tool to a new version, but since Cyanogenmod is installed on many user devices, Google has decided to make some change in their aapt tool so that new Android apps don't trigger the bug in Cyanogenmod.
The bug was hidden until Android SDK updated its AAPT build tool to a new version, but since Cyanogenmod is installed on many user devices, Google has decided to make some change in their aapt tool so that new Android apps don't trigger the bug in Cyanogenmod.
ma...@gmail.com <ma...@gmail.com> #39
Got it. Thank you for the answer!
I've reportedhttps://jira.lineageos.org/plugins/servlet/mobile#issue/BUGBASH-1052 which is very likely caused by this issue. Does anyone on this thread happen to know what the LineageOS bug number for this issue is?
I've reported
te...@gmail.com <te...@gmail.com> #40
I don't know what the number is but I have had 4 apps that have crashed and I reported the problem to the app devs referring to this thread and 3 of the 4 have fixed the issue. Still waiting on the 4th. So if you do the same chances are it will be fixed.
mb...@gmail.com <mb...@gmail.com> #41
I can confirm that build tools 26.0.3 fix it as well.
em...@gmail.com <em...@gmail.com> #42
Is Lineage 14 affected with this bug? Some of my apps are built with build tools 26.0.2, Android Studio 3.1 and nothing happens.
mi...@gmail.com <mi...@gmail.com> #43
Not by my observation. Only older CM versions.
mk...@fejm.pl <mk...@fejm.pl> #44
Is it any method to patch CM 12, 12.1 and 13 to work properly with apps, until app devs do own fixes?
cr...@gmail.com <cr...@gmail.com> #45
There seems to be a flashable zip that solves the problem. You can find it here: https://forum.xda-developers.com/showpost.php?p=74842530&postcount=464 . Could anyone confirm if thispatch is a solution or not? Thanks!
te...@gmail.com <te...@gmail.com> #46
The zip file does not install on my Galaxy Note 8 tablet. This zip is for lollipop but I did get one that was for marshmallow which I have but it did not install either.
wa...@gmail.com <wa...@gmail.com> #47
Is there a fix for CM13 devices?
r1...@gmail.com <r1...@gmail.com> #49
Any fixes for CM11 users? Looked a lot but couldn't find one yet... I have absolutely no intention of upgrading, I like CM11 but not being able to update apps is ridiculous...
mo...@hotmail.com <mo...@hotmail.com> #50
Is this .zip patch for any device with any Lollipop 5.1.1 ROM?
Because I just installed it and my HTC one X with Resurrection Remix ROM (Lollipop 5.1.1) goes black after the HTC logo.
mo...@hotmail.com <mo...@hotmail.com> #51
Anyone?
an...@gmail.com <an...@gmail.com> #52
Is this really fixed? I upgraded to Android Studio 3.2 canary 2, and I just found a report of the same crash with the exact same stack trace in OP since I last published my app built with AS 3.2 !?
I used to have android.enableAapt2=false in gradle.properties, but removed it since it was supposed to be fixed and Android Studio kept complaining about it that it was obsolete and going to be removed by end of 2018.
I used to have android.enableAapt2=false in gradle.properties, but removed it since it was supposed to be fixed and Android Studio kept complaining about it that it was obsolete and going to be removed by end of 2018.
re...@gmail.com <re...@gmail.com> #53
Complain to the maintainers of lineage OS if you want a real perminent fix. All Google can do is try to work around the bug.
[Deleted User] <[Deleted User]> #55
I used to have android.enableAapt2=false in gradle.properties, but removed it since it was supposed to be fixed and Android Studio kept complaining about it that it was obsolete and going to be removed by end of 2018.
Worried that this may cause a same old problem of device reboot and factory reset. Has goole done any workaround so as to avoid this problem ? I am on studio 3.1.4
Worried that this may cause a same old problem of device reboot and factory reset. Has goole done any workaround so as to avoid this problem ? I am on studio 3.1.4
ph...@gmail.com <ph...@gmail.com> #56
Good
ab...@gmail.com <ab...@gmail.com> #58
Privacy violation My privacy has been violated and there are in my account subscriptions why i can see it and there is a false and incorrect user agreement and several activities were opened in my own account I have no knowledge of it and has been informed about this matter the person was not prevented from entering my account
Description
AI-171.4182969, JRE 1.8.0_152-release-884-b01x64 JetBrains s.r.o, OS Windows 7(amd64) v6.1 Service Pack 1, screens 2560x1440
Since building my app with Android Studio 3, the resulting APK causes some Android system crashes at native level on certain devices. It is hardly to debug, crash happens in system code.
backtrace:
native: pc 0000000000029158 /system/lib/libc.so (dlfree+1239)
native: pc 000000000000f3c3 /system/lib/libc.so (free+10)
native: pc 0000000000012ab5 /system/lib/libandroidfw.so (android::ResStringPool::uninit()+38)
native: pc 00000000000137ef /system/lib/libandroidfw.so (android::ResXMLTree::uninit()+12)
native: pc 000000000001380d /system/lib/libandroidfw.so (android::ResXMLTree::~ResXMLTree()+4)
native: pc 0000000000010507 /system/lib/libandroidfw.so (android::AssetManager::getPkgName(char const*)+258)
native: pc 0000000000010569 /system/lib/libandroidfw.so (android::AssetManager::getBasePackageName(unsigned int)+68)
native: pc 0000000000081e9f /system/lib/libandroid_runtime.so
native: pc 000000000027b17b /data/dalvik-cache/arm/system@framework@boot.oat
This backtrace is identical on all affected devices. It may be related to how Gradle tools build resources or manifest. It wakes some sleeping dragon!
Most affected devices (from crash reports provided by Google developer console):
Galaxy S3 Neo Plus (s3ve3g)
Nexus 4 (mako)
Galaxy Grand duos (baffin)
Galaxy S3 (d2att)
Galaxy S2 (GT-I9100)
Galaxy Nexus (maguro)
Nexus 7 (2012) (grouper)
Iconia Tab A211 (picasso_e2)
and many other
I have one device myself where I can reproduce the crash. That is Samsung Galaxy Nexus I9250 (maguro) with Cyanogenmod 12.1 ROM. Both debug and release versions crash the device. The device gets to reboot loop. I can recover device by adb shell and pm uninstall command to uninstall the app.
I confirm that my app previously build using Gradle 2.x plugin doesn't crash system that way. It started exactly with Gradle 3.x plugin.
Attached are 2 versions of my app, one built with Gradle 2.x and not crashing, other build with 3.x and crashing.
You may need to find affected device to reproduce it, but I gave enough hints.
Expect that more developers will produce same affected APKs with Android Studio 3 when it's out of beta, so it's worth fixing build tools.