[Interest] Qt Android - tgkill - libQt5Purchasing.so

Sat May 11 10:36:02 CEST 2019

Robert,

In the last couple of days I have downgraded the app Qt version to Qt 5.9.8 and the problem remains, the crashes keeps happening. I’m wondering if this actually the same problem you passed me the link to.

In the mean time I have created a bug report on Qt -> https://bugreports.qt.io/browse/QTBUG-75683 <https://bugreports.qt.io/browse/QTBUG-75683> but no replies yet.

My experience in the past about bug reportings in the Qt Purchasing module for Android was simply terrible. I was not given any credit at all and I had to live with it and find a workaround myself…

Today, 4:24 AM on app version 165
Samsung Galaxy S4 (jaltelgt), Android 5.0
Report 1 of 5

*** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
pid: 0, tid: 0 >>> com.imaginando.drc <<<

backtrace:
  #00  pc 000000000003a1c8  /system/lib/libc.so (tgkill+12)
  #01  pc 0000000000017549  /system/lib/libc.so (pthread_kill+52)
  #02  pc 000000000001815b  /system/lib/libc.so (raise+10)
  #03  pc 0000000000014aa5  /system/lib/libc.so (__libc_android_abort+36)
  #04  pc 0000000000012a6c  /system/lib/libc.so (abort+4)
  #05  pc 000000000022eb33  /system/lib/libart.so (_ZN3art7Runtime5AbortEv+170)
  #06  pc 00000000000a9245  /system/lib/libart.so (_ZN3art10LogMessageD1Ev+1360)
  #07  pc 000000000023c86d  /system/lib/libart.so (_ZNK3art6Thread24AssertNoPendingExceptionEv+396)
  #08  pc 00000000000d8ac1  /system/lib/libart.so (_ZN3art11ClassLinker9FindClassEPNS_6ThreadEPKcNS_6HandleINS_6mirror11ClassLoaderEEE+24)
  #09  pc 00000000001de70b  /system/lib/libart.so (_ZN3art3JNI9FindClassEP7_JNIEnvPKc+826)
  #10  pc 0000000000004b3d  /data/app/com.imaginando.drc-1/lib/arm/libQt5Purchasing.so

Best regards,

Nuno

> On 9 May 2019, at 07:35, coroberti . <coroberti at gmail.com> wrote:
> 
> Dear Nuno,
> Yes, JNI in QtPurchasing.
> 
> The issue with the known bug of supposed to be fixed by BogDan recently is that
> Qt-5.9 worked perfect. Qt-5.10 and 5.11 are broken and it's supposed to be fixed very recently somewhere in Qt-5.12,
> but it might be that the fix is not covering all scenario.
> 
> I'd look for all classes in QtPurchasing called by JNI and, if you have a scenario to reproduce the badness,
> make a log prior to every JNI call to see where it breaks.
> 
> Take care.
> 
> Kind regards,
> Robert Iakobashvili
> ............................
> 
> 
> On Thu, May 9, 2019 at 8:25 AM Nuno Santos <nunosantos at imaginando.pt <mailto:nunosantos at imaginando.pt>> wrote:
> Robert,
> 
> Thanks for your input.
> 
> I don’t use JNI for purchasing a long time. I rely solely on the C++ API of QtPurchasing. That’s even more wicked.
> 
> Maybe the problems resides solely inside the QtPurchasing module… I’m the dark!
> 
> Best,
> 
> Nuno
> 
>> On 9 May 2019, at 06:47, coroberti . <coroberti at gmail.com <mailto:coroberti at gmail.com>> wrote:
>> 
>> Dear Nuno,
>> No knowledge, but as a direction, it's kinda JNI of QPurchasing is looking in some scenario
>> for a Java class that is not existing or not installed.
>> 
>> I'd look into all JNI calls that QPurchasing is doing and its Java dependencies.
>> 
>> We had recently a major issue with JNI done not via Qt API that BogDan supposed to get fixed
>> and this issue appeared in Qt-5.10. Therefore, it could be another direction to explore.
>> 
>> Sorry, but jm2c to add.
>> 
>> Kind regards,
>> Robert Iakobashvili
>> ............................
>> 
>> 
>> On Thu, May 9, 2019 at 12:08 AM Nuno Santos <nunosantos at imaginando.pt <mailto:nunosantos at imaginando.pt>> wrote:
>> Thiago,
>> 
>> Thanks for your reply.
>> 
>> Unfortunately no, this is all I can see in Google Play developer console... :(
>> 
>> Unless I’m missing some detail on how to get more information.
>> 
>> All I know it that it is very regular. Before I was using 5.10 and there was no issues. Since I have upgraded to Qt 5.12 this crash started to happen and happens quite often. So often that in some devices the cannot even be open and I don’t have a clue why it happens. 
>> 
>> The biggest problem is that in this cases, people usually slam the app with one star.
>> 
>> Regards,
>> 
>> Nuno
>> 
>> > On 8 May 2019, at 20:47, Thiago Macieira <thiago.macieira at intel.com <mailto:thiago.macieira at intel.com>> wrote:
>> > 
>> > On Wednesday, 8 May 2019 07:21:09 PDT Nuno Santos wrote:
>> >>  #06  pc 00000000000ac2bb  /system/lib/libart.so
>> >> (art::LogMessage::~LogMessage()+1322) #07  pc 0000000000239e37 
>> >> /system/lib/libart.so (art::Thread::AssertNoPendingException() const+358)
>> >> #08  pc 00000000000de40b  /system/lib/libart.so
>> >> (art::ClassLinker::FindClass(art::Thread*, char const*,
>> >> art::Handle<art::mirror::ClassLoader>)+14) #09  pc 00000000001df82b 
>> >> /system/lib/libart.so (art::JNI::FindClass(_JNIEnv*, char const*)+806) #10 
>> >> pc 0000000000005813 
>> >> /data/app/com.imaginando.drc-1/lib/arm/libQt5Purchasing.so
>> > 
>> > All of these are logging a fatal message and then abort()ing the application. 
>> > Can you get that message?
>> > 
>> > -- 
>> > Thiago Macieira - thiago.macieira (AT) intel.com <http://intel.com/>
>> >  Software Architect - Intel System Software Products
>> > 
>> > 
>> > 
>> > _______________________________________________
>> > Interest mailing list
>> > Interest at qt-project.org <mailto:Interest at qt-project.org>
>> > https://lists.qt-project.org/listinfo/interest <https://lists.qt-project.org/listinfo/interest>
>> 
>> _______________________________________________
>> Interest mailing list
>> Interest at qt-project.org <mailto:Interest at qt-project.org>
>> https://lists.qt-project.org/listinfo/interest <https://lists.qt-project.org/listinfo/interest>
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.qt-project.org/pipermail/interest/attachments/20190511/40a46e6d/attachment.html>