致命信号 11 (SIGSEGV)，在第二次 JNI String 调用 Java 方法之后

Question

I am kinda stuck and lost on this problem: I want to pass a string parameter from a C++ Method (Native C), that I call via ROS to a Java method. In this java method, a robot I am programming should use some Text-to-Speak internal method to say the written thing (the tts method is given by the company of the robot). But whenever I call this Method in C++ from ROS, the first call is properly handled, but on the second call the program crashes with the previous stated error code. Here is my C++ code: (myEnv/myVM/myClass are set before in the init method) (log is just a small own method which does System.out.println() for C++)

void onSpeech(const std_msgs::String::ConstPtr &msg){
    log("RECEIVED SPEECH");
    log(msg->data.c_str());

    myVM->AttachCurrentThread(&myEnv, nullptr);
    jstring stringMsg = myEnv->NewStringUTF(msg->data.c_str());
    log("CALL JAVA METHOD");
    jmethodID mid = myEnv->GetStaticMethodID(myClass, "setSpeech", "(Ljava/lang/String;)V");
    myEnv->CallStaticVoidMethod(myClass, mid, stringMsg);

    log("START WAITING");
    ros::Time begin = ros::Time().now();
    while(begin + ros::Duration(2,0) > ros::Time().now()){ }
    log("WAIT FINISHED");

    log("CALL RELEASE");
    myEnv->ReleaseStringUTFChars(stringMsg, msg->data.c_str());
    myVM->DetachCurrentThread();
}

And here is the Java Method we call. I copy the passed String so often, because I want to copy the data of the word and not the pointer, and the C++ Code can release the stringMsg variable without affecting the java code.

public static void setSpeech(String text) {
    System.out.println("JAVA CALL");
    System.out.println("PASSED TEXT: " + text);
    char[] newChar = text.toCharArray();
    System.out.println("CHAR: " + Arrays.toString(newChar));
    String newString = String.copyValueOf(newChar);
    System.out.println("NEW STRING: " + newString);
    //Below is just the ttsmethod like this: startSpeak(newString)
}

This is the complete Log, I first send a msg with the data "Hello 1" which worked as expected and the robot said it correct. But after that I send another message with the data "Hello 2" and the program crashes before the newChar could be set. Which in my mind doesn't make sense, because we even System.out.println() the text one line before.

I/CPP: Hello 1
I/CPP: CALL JAVA METHOD
I/System.out: JAVA CALL
I/System.out: PASSED TEXT: Hello 1
I/System.out: CHAR: [H, e, l, l, o,  , 1]
I/System.out: NEW STRING: Hello 1
I/CPP: START WAITING
I/System.out: SPEECH STARTED: Hello 1
I/System.out: SPEECH FINISHED
I/CPP: WAIT FINISHED
I/CPP: CALL RELEASE
//Now I call the method again
I/CPP: RECEIVED SPEECH
I/CPP: Hello 2
I/CPP: CALL JAVA METHOD
I/System.out: JAVA CALL
I/System.out: PASSED TEXT: Hello 2
A/libc: Fatal signal 11 (SIGSEGV), code 1, fault addr 0x38206f in tid 8890 (Thread-117)

Any help would be appreciated, I'm really lost why the code works in the "PASSED TEXT" line and then in the newChar line it suddenly crashes. Thanks

btw this is the logcat output for "Hello 9" and "Hello 9"

2022-12-19 13:44:28.785 9455-9499/rob.rosloomo I/ROSCPP_NDK: RECEIVED SPEECH
2022-12-19 13:44:28.786 9455-9499/rob.rosloomo I/ROSCPP_NDK: Hello 9
2022-12-19 13:44:28.787 9455-9499/rob.rosloomo I/ROSCPP_NDK: CALL JAVA METHOD
2022-12-19 13:44:28.788 9455-9499/rob.rosloomo I/System.out: JAVA CALL
2022-12-19 13:44:28.788 9455-9499/rob.rosloomo I/System.out: PASSED TEXT: Hello 9
2022-12-19 13:44:28.788 9455-9499/rob.rosloomo A/libc: Fatal signal 11 (SIGSEGV), code 1, fault addr 0x39206f in tid 9499 (Thread-130)
2022-12-19 13:44:28.893 2750-2750/? I/DEBUG: *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
2022-12-19 13:44:28.893 2750-2750/? I/DEBUG: Build fingerprint: 'i-Buddie/VA50EC_i_Buddie/VA50EC_1:5.1.1/LMY47Z/78:user/test-keys'
2022-12-19 13:44:28.893 2750-2750/? I/DEBUG: Revision: '0'
2022-12-19 13:44:28.893 2750-2750/? I/DEBUG: ABI: 'x86'
2022-12-19 13:44:28.894 2750-2750/? I/DEBUG: pid: 9455, tid: 9499, name: Thread-130  >>> rob.rosloomo <<<
2022-12-19 13:44:28.894 2750-2750/? I/DEBUG: signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x39206f
2022-12-19 13:44:28.920 2750-2750/? I/DEBUG:     eax f3cd0040  ebx f3bfca2c  ecx f3cd0000  edx 00000001
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     esi 0039206f  edi f3ef006c
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     xcs 00000023  xds 0000002b  xes 0000002b  xfs 00000097  xss 0000002b
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     eip f38b4e12  ebp e2a23d78  esp e2a23d40  flags 00210206
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG: backtrace:
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     #00 pc 001a5e12  /system/lib/libart.so (art::gc::allocator::RosAlloc::RefillRun(art::Thread*, unsigned int)+290)
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     #01 pc 001a60fa  /system/lib/libart.so (art::gc::allocator::RosAlloc::AllocFromRun(art::Thread*, unsigned int, unsigned int*, unsigned int*, unsigned int*)+666)
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     #02 pc 00155f8d  /system/lib/libart.so (art::mirror::Object* art::gc::space::RosAllocSpace::AllocCommon<true>(art::Thread*, unsigned int, unsigned int*, unsigned int*, unsigned int*)+109)
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     #03 pc 0040571d  /system/lib/libart.so (_ZN3art6mirror5Array5AllocILb0EEEPS1_PNS_6ThreadEPNS0_5ClassEijNS_2gc13AllocatorTypeEb.constprop.104+989)
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     #04 pc 00406a72  /system/lib/libart.so (artAllocArrayFromCodeResolvedRosAlloc+162)
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     #05 pc 000e5815  /system/lib/libart.so (art_quick_alloc_array_resolved_rosalloc+37)
2022-12-19 13:44:28.921 2750-2750/? I/DEBUG:     #06 pc 00079bfb  /data/dalvik-cache/x86/system@framework@boot.oat
2022-12-19 13:44:29.120 2750-2750/? I/DEBUG: Tombstone written to: /data/tombstones/tombstone_06
2022-12-19 13:44:29.120 3124-9562/? W/ActivityManager:   Force finishing activity 1 rob.rosloomo/.MainActivity
2022-12-19 13:44:29.129 3124-3255/? E/JavaBinder: !!! FAILED BINDER TRANSACTION !!!
2022-12-19 13:44:29.131 3124-9562/? E/JavaBinder: !!! FAILED BINDER TRANSACTION !!!
2022-12-19 13:44:29.131 3124-9562/? W/ActivityManager: Exception thrown during pause

Answer 1

https://stackoverflow.com/users/882003/john had the right tipp with this comment: ReleaseStringUTFChars should only be called on strings allocated with GetStringUTFChars . Been a long time since I've done JNI but remove that line and I think you are OK