Undefined symbols for architecture x86_64: JNI_CreateJavaVM OS-X Xcode - java-native-interface

I am attempting to compile the following C++ code for creating a Java Virtual Machine, using Xcode on Mac OS-X Mountain Lion:
JNIEnv * createVM(JavaVM **jvm){
JNIEnv * retEnv = NULL;
JavaVMInitArgs vm_args = *new JavaVMInitArgs();
JavaVMOption *options = new JavaVMOption[1];
std::string sJavaInstallPath = "-Djava.class.path=" + findJavaPath();
if(sJavaInstallPath == "-Djava.class.path="){
return NULL;
}
options[0].optionString = const_cast<char*>(sJavaInstallPath.c_str());
vm_args.version = JNI_VERSION_1_6;
vm_args.nOptions = 1;
vm_args.options = options;
vm_args.ignoreUnrecognized = false;
long status = JNI_CreateJavaVM(jvm, (void**)&retEnv, &vm_args);
if(status == JNI_ERR){
std::cout << "Failure: Unable to load JVM \t Exit" << std::endl;
}else if(status == JNI_OK){
std::cout << "CreateVM:\t\tJVM loaded successfully" << std::endl;
}
delete options;
return retEnv;
};
Which gives the following error:
Undefined symbols for architecture x86_64: "_JNI_CreateJavaVM".
From these similar questions:
undefined reference to `JNI_CreateJavaVM' linux
How can I use JNI in C++ to use a Java class?
undefined symbol: JNI_CreateJavaVM in Linux
http://lists.apple.com/archives/java-dev/2005/Apr/msg00067.html
http://lists.apple.com/archives/java-dev/2005/Apr/msg00068.html
I gather that I'm missing a Library reference to libjvm.dylib. The problem is that I'm not sure how to get Xcode to recognize/include the library for linking,
I've tried putting the following directories (all of which contain a libjvm.dylib file) into the Library Search Paths list in Xcode, with no result:
/System/Library//Java/JavaVirtualMachines/1.6.0.jdk/Contents/Libraries/
/Library//Java/JavaVirtualMachines/jdk1.7.0_51.jdk/Contents/Home/jre/lib/server/
/System/Library/Frameworks/JavaVM.framework/Libraries/
I'm still quite new to OS-X and Xcode, and the answers to the other questions all relate to compiling using gcc or g++, which I have even less experience with, so what I need to know is: how do I tell the Xcode compiler where to find libjvm.dylib, and also which libjvm.dylib is the one I should be linking to?

OK, found the solution,
the equivalent of g++'s
-L<library/to/link>
setting in Xcode is the Link Binary with Libraries setting under the Build Phases tab,
not the Library Search Paths list that I mentioned above.
and the library I needed to link to was actually /System/Library/Frameworks/JavaVM.framework/JavaVM,
not libjvm.dylib

Related

Undefined reference to 'getcwd' and 'mkdir' when building Google test project

even if I have been strictly following the tutorials from this post, I can't get my google test demo program to compile.
I'm using Eclipse on Windows 10 x64, and the ARM GCC embedded toolchain to compile my code, since I will eventually need to run unit tests on embedded devices.
My problem is that when I try to Build the project I get those errors :
c:/program files (x86)/gnu tools arm embedded/9 2019-q4-major/bin/../lib/gcc/arm-none-eabi/9.2.1/../../../../arm-none-eabi/bin/ld.exe: ./contrib/gtest/gtest-all.o: in function `testing::internal::FilePath::GetCurrentDir()':
C:\Users\Hugo\eclipse\eclipse-workspace\test_gtest\Debug/../contrib/gtest/gtest-all.cc:9598: undefined reference to `getcwd'
c:/program files (x86)/gnu tools arm embedded/9 2019-q4-major/bin/../lib/gcc/arm-none-eabi/9.2.1/../../../../arm-none-eabi/bin/ld.exe: ./contrib/gtest/gtest-all.o: in function `testing::internal::FilePath::CreateFolder() const':
C:\Users\Hugo\eclipse\eclipse-workspace\test_gtest\Debug/../contrib/gtest/gtest-all.cc:9823: undefined reference to `mkdir'
collect2.exe: error: ld returned 1 exit status
make: *** [makefile:61: test_gtest.elf] Error 1
It comes more precisely from those lines of code in the gtest_all.cc file:
For the undefined reference to 'getcwd'
FilePath FilePath::GetCurrentDir() {
#if GTEST_OS_WINDOWS_MOBILE || GTEST_OS_WINDOWS_PHONE || \
GTEST_OS_WINDOWS_RT || GTEST_OS_ESP8266 || GTEST_OS_ESP32
// These platforms do not have a current directory, so we just return
// something reasonable.
return FilePath(kCurrentDirectoryString);
#elif GTEST_OS_WINDOWS
char cwd[GTEST_PATH_MAX_ + 1] = { '\0' };
return FilePath(_getcwd(cwd, sizeof(cwd)) == nullptr ? "" : cwd);
#else
char cwd[GTEST_PATH_MAX_ + 1] = { '\0' };
char* result = getcwd(cwd, sizeof(cwd));
# if GTEST_OS_NACL
// getcwd will likely fail in NaCl due to the sandbox, so return something
// reasonable. The user may have provided a shim implementation for getcwd,
// however, so fallback only when failure is detected.
return FilePath(result == nullptr ? kCurrentDirectoryString : cwd);
# endif // GTEST_OS_NACL
return FilePath(result == nullptr ? "" : cwd);
#endif // GTEST_OS_WINDOWS_MOBILE
}
For the undefined reference to 'mkdir' :
bool FilePath::CreateFolder() const {
#if GTEST_OS_WINDOWS_MOBILE
FilePath removed_sep(this->RemoveTrailingPathSeparator());
LPCWSTR unicode = String::AnsiToUtf16(removed_sep.c_str());
int result = CreateDirectory(unicode, nullptr) ? 0 : -1;
delete [] unicode;
#elif GTEST_OS_WINDOWS
int result = _mkdir(pathname_.c_str());
#elif GTEST_OS_ESP8266
// do nothing
int result = 0;
#else
int result = mkdir(pathname_.c_str(), 0777);
#endif // GTEST_OS_WINDOWS_MOBILE
if (result == -1) {
return this->DirectoryExists(); // An error is OK if the directory exists.
}
return true; // No error.
}
I have checked that unistd.h was included. I've been searching a lot but can't seem to find any similar error as mine. The closest I could find has been solved by people using CMake to compile it, but I'm not using CMake at all here.
AFAIK, getcwd() and mkdir() is platform dependent.
It seem there have been similar issue with other library:
https://github.com/purduesigbots/pros/issues/176
As above link, you can try to define stub for missing symbols.
In my working platform, getcwd() and mkdir() even get removed from the header.
In such case, you can edit gtest directly, for example:
FilePath FilePath::GetCurrentDir() {
#if GTEST_OS_CUSTOM_PLATFORM
return kCurrentDirectoryString;
...

Xcode gives Mach-O linker error: "_JNI_CreateJavaVM", referenced from: _main in main.o

I get this error from testing JNI:
Undefined symbols for architecture x86_64:
"_JNI_CreateJavaVM", referenced from:
_main in main.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)
Here is c++ code:
#include <jni.h>
#include <iostream>
using namespace std;
int main()
{
int res;
JavaVMInitArgs vm_args;
JavaVMOption options[3];
JavaVM *jvm;
JNIEnv *env;
jmethodID mid;
options[0].optionString = "-Djava.compiler=NONE";
options[1].optionString = "-Djava.class.path = /Users/stephen/course/test/Test";
options[2].optionString = "-verbose:NONE";
vm_args.version = JNI_VERSION_1_8;
vm_args.nOptions = 3;
vm_args.options = options;
vm_args.ignoreUnrecognized = JNI_TRUE;
res = JNI_CreateJavaVM(&jvm,(void**)&env,&vm_args);
if(res == JNI_ERR){
cout << "Error invoking the JVM";
return 1;
}
cout <<"create JVM successfully!"<<endl;
jclass cls = env->FindClass("/Users/stephen/course/Qt-project/test/Test");
if(cls != 0){
cout<<"find class successfully!" << endl;
}
mid = env->GetMethodID(cls,"sayHello","stephen");
if(mid != 0){
cout<<"Invoke method successfully!" << endl;
}
jvm->DestroyJavaVM();
return 0;
}
Here is java code:
public class Test
{
public static void sayHello(String s){
System.out.print("hello I am" + s + "\n");
}
}
I add the include path of " jdk/include; jdk/include/darwin" the project, also I add lib path of " jdk/jre/lib/server" to the project to get the libjvm.dylib. The c++ standard library of my project is libstdc++(gnu c++ standard library.
But I can't solve this problem as expected.
Take a look here for a sample code where JVM library is linked with your project:
https://github.com/mkowsiak/jnicookbook/tree/master/recipes/recipeNo028
Take a look at Makefile. Especially, here:
main: recipeNo028_main.o
ld -o lib/recipeNo028_main -L${JAVA_HOME}/jre/lib/server/ \
-ljvm \
$(MAC_OS_FLAGS) \
lib/recipeNo028_main.o
where jvm lib is linked with the code, and here:
CC=llvm-gcc
MAC_OS_FLAGS=-rpath ${JAVA_HOME}/jre/lib/server -L/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.12.sdk -demangle -dynamic -arch x86_64 -macosx_version_min 10.12.0 -lSystem
where all required libs are added to your code as well. It should work. Try to compile sample code. You can find more samples here: http://jnicookbook.owsiak.org
Update
How to use arbitrary JDK version for compilation.
First, take a look at all installations you have
/usr/libexec/java_home -V
This will produce something like this
/usr/libexec/java_home -V
Matching Java Virtual Machines (4):
9, x86_64: "Java SE 9" /Library/Java/JavaVirtualMachines/jdk-9.jdk/Contents/Home
1.8.0_144, x86_64: "Java SE 8" /Library/Java/JavaVirtualMachines/jdk1.8.0_144.jdk/Contents/Home
1.8.0_111, x86_64: "Java SE 8" /Library/Java/JavaVirtualMachines/jdk1.8.0_111.jdk/Contents/Home
1.7.0_80, x86_64: "Java SE 7" /Library/Java/JavaVirtualMachines/jdk1.7.0_80.jdk/Contents/Home
Then, before running make, simply set JAVA_HOME to whatever you like
export JAVA_HOME=$(/usr/libexec/java_home -v 9)
Now, your code will use version that you have chosen.

Undefined Reference to JNI_CreateJavaVM Eclipse and Linux

I am using Ubuntu 10.04 with Omnet++ 4.0p1 and JSimpleModule (which uses SWIG to make Java wrappers for the C++ methods found in Omnet++). I am trying to create a simulation in Java using the 2 above libraries, and when I am trying to build the project I am getting 'Undefined Reference to JNI_CreateJavaVM' as an error in JUtil.cc (which is code supplied by JSimpleModule). I have been looking around and I am including all of the proper libraries and it still isn't fixing anything. In the Omnet++ IDE (Eclipse) I am including:
/usr/lib/jvm/java-6-openjdk/include
/usr/lib/jvm/java-6-openjdk/include/linux
/usr/lib/jvm/java-6-openjdk/jre/lib/i386/client
and am linking:
/usr/lib/jvm/java-6-openjdk/jre/lib
/usr/lib/jvm/java-6-openjdk/jre/lib/client
-ljvm
I also tried compiling from the terminal using opp_makemake (which creates a Makefile) with the following parameters:
-I/usr/lib/jvm/java-6-openjdk/include
-I/usr/lib/jvm/java-6-openjdk/include/linux
-I/usr/lib/jvm/java-6-openjdk/jre/lib/i386/client
-L/usr/lib/jvm/java-6-openjdk/jre/lib
-L/usr/lib/jvm/java-6-openjdk/jre/lib/client -ljvm
Here is the beginning of JUtil.cc up to the error (note: jni.h is included in JUtil.h):
#include "JUtil.h"
#include "JSimpleModule.h"
//#define DEBUGPRINTF printf
#define DEBUGPRINTF (void)
#ifdef _WIN32
#define PATH_SEP ";"
#else
#define PATH_SEP ":"
#endif
// This will come from the generated SimkernelJNI_registerNatives.cc
void SimkernelJNI_registerNatives(JNIEnv *jenv);
JavaVM *JUtil::vm;
JNIEnv *JUtil::jenv;
void JUtil::initJVM()
{
DEBUGPRINTF("Starting JVM...\n");
JavaVMInitArgs vm_args;
JavaVMOption options[10];
int n = 0;
const char *classpath = getenv("CLASSPATH");
if (!classpath)
opp_error("CLASSPATH environment variable is not set");
// FIXME remove hack once IDE builds the classpath corretcly
const char *classpath2 = getenv("CLASSPATH2");
std::string classpathOption = std::string("-Djava.class.path=")+(classpath2 ? classpath2 : "")+PATH_SEP+(classpath ? classpath : "");
options[n++].optionString = (char *)classpathOption.c_str(); /* user classes */
options[n++].optionString = (char *)"-Djava.library.path=."; /* set native library path */
//options[n++].optionString = "-Djava.compiler=NONE"; /* disable JIT */
//options[n++].optionString = "-verbose:jni"; /* print JNI-related messages */
//options[n++].optionString = "-verbose:class"; /* print class loading messages */
vm_args.version = JNI_VERSION_1_2;
vm_args.options = options;
vm_args.nOptions = n;
vm_args.ignoreUnrecognized = true;
int res = JNI_CreateJavaVM(&vm, (void **)&jenv, &vm_args);
if (res<0)
opp_error("Could not create Java VM: JNI_CreateJavaVM returned %d", res);
DEBUGPRINTF("Registering native methods...\n");
SimkernelJNI_registerNatives(jenv);
DEBUGPRINTF("Done.\n");
}
If someone knows anything on how to fix this, that would be greatly appreciated. Thanks.

"Access violation" Error on displaying string of simple Oracle Query (VS10 Exp C++)

I am struggling with an issue regarding running a SQL statement to an Oracle database through C++, using occi. My code is as follows:
#include <iostream>
#include "occi.h"
namespace oc = oracle::occi;
int main() {
std::cout << "Setting up environment...\n";
oc::Environment * env = oc::Environment::createEnvironment();
std::cout << "Setting up connection...\n";
oc::Connection * conn = env->createConnection("user","pass","server");
std::cout << "Creating statement...\n";
//Very simply query...
oc::Statement * stmt = conn->createStatement("SELECT '1' FROM dual");
std::cout << "Executing query...\n";
oc::ResultSet * rs = stmt->executeQuery();
while(rs->next()) {
std::cout << rs->getString(1) << std::endl; //Error is thrown at this line, but after printing since I can see '1' on the console.
}
stmt->closeResultSet(rs);
conn->terminateStatement(stmt);
env->terminateConnection(conn);
oc::Environment::terminateEnvironment(env);
return 0;
}
The error that is shown is:
Unhandled exception at 0x1048ad7a (msvcp100d.dll) in MyDatabaseApp.exe: 0xC0000005: Access violation reading location 0xccccccd0.
My program stops inside 'xstring' at the following line of code:
#if _ITERATOR_DEBUG_LEVEL == 0
....
#else /* _ITERATOR_DEBUG_LEVEL == 0 */
typedef typename _Alloc::template rebind<_Elem>::other _Alty;
_String_val(_Alty _Al = _Alty())
: _Alval(_Al)
{ // construct allocator from _Al
....
}
~_String_val()
{ // destroy the object
typename _Alloc::template rebind<_Container_proxy>::other
_Alproxy(_Alval);
this->_Orphan_all(); //<----------------------Code stops here
_Dest_val(_Alproxy, this->_Myproxy);
_Alproxy.deallocate(this->_Myproxy, 1);
this->_Myproxy = 0;
}
#endif /* _ITERATOR_DEBUG_LEVEL == 0 */
If I change my query to:
oc::Statement * stmt = conn->createStatement("SELECT 1 FROM dual");
and the loop statement to:
std::cout << rs->getInt(1) << std::endl;
It works fine with no errors. I think this is because getting an integer simply returns a primitive, but when an object is being returned it is blowing up (I think on a destructor, but I'm not sure why...)
I have been playing around with this for hours today, and I am pretty stuck.
Some information about my system:
OS - Windows XP
Oracle Version - 10g
IDE - Microsoft Visual Studio 2010 Express C++
My project properties are as follows:
C/C++ - General - Additional Include Directories = C:\oracle\product\10.2.0\client_1\oci\include;%(AdditionalIncludeDirectories)
C/C++ - Code Generation - Multi-threaded Debug DLL (/MDd)
Linker - General - Additional Library Directories = C:\oracle\product\10.2.0\client_1\oci\lib\msvc\vc8;%(AdditionalLibraryDirectories)
Linked - Input - Additional Dependencies = oraocci10.lib;oraocci10d.lib;%(AdditionalDependencies)
I hope I haven't been confusing with too much info... Any help or insight would be great, Thanks in advance!
EDIT If I rewrite my loop, storing the value in a local variable, the error is thrown at the end of the loop:
while(rs->next()) {
std::string s = rs->getString(1); //s is equal to "1" as expected
std::cout << s << std::endl; //This is executed successfully
} //Error is thrown here
Usually such kind of problems come from differences in build environments (IDE) of end user and provider.
Check this.
Related problems:
Unhandled exception at 0x523d14cf (msvcr100d.dll)?
Why does this program crash: passing of std::string between DLLs
First try to use correct lib and dll. If compiled in debug mode then all libs and dlls must be debug. Use VC++ Modules view to be sure that proper DLL loaded.
I was lucky with my application to have all libs compiled for MSVC2010. So I just check debug and release mode DLLs and got working application.
I revisited this issue about a month ago and I found that the MSVC2010 occi library was built for Oracle 11g. We are running Oracle 10g, so I had to use the MSVC2005 library. So I installed the outdated IDE and loaded the Debug library and it worked (for some reason the release version wouldn't work though).
EDIT
For anyone who is having the same problem I was, if downgrading the IDE from MSVC2010 to MSVC2005 with the appropriate libraries doesn't work, you could try upgrading the Oracle client from 10g to 11g and use the MSVC2010 library, as suggested by harvyS. In retrospect this would've probably been the better solution.

Problem with LZO

So I am trying to use LZO in my application. Here is how I have included it:
#include "lzoconf.h"
#include "lzodefs.h"
#include "lzo1x.h"
/* portability layer */
static const char *progname = NULL;
#define WANT_LZO_MALLOC 1
#define WANT_XMALLOC 1
#include "portab.h"
Then in the application I do:
if (lzo_init() != LZO_E_OK)
{
printf("internal error - lzo_init() failed !!!\n");
printf("(this usually indicates a compiler bug - try recompiling\nwithout optimizations, and enable '-DLZO_DEBUG' for diagnostics)\n");
return 4;
}
It compiles ok. No errors or warnings during compilation.
When I try to run my application though, there are two errors:
/home/richard/client/src/portab.h:145: undefined reference to `__lzo_align_gap'
Which points at this line in portab.h:
if (__lzo_align_gap(p, (lzo_uint) sizeof(lzo_align_t)) != 0)
{
printf("%s: C library problem: malloc() returned mis-aligned pointer!\n", progname);
exit(1);
}
return p;
And in my application:
/home/richard/client/src/main.cc:108: undefined reference to `__lzo_init_v2'
Which points to:
if (lzo_init() != LZO_E_OK)
{
printf("internal error - lzo_init() failed !!!\n");
printf("(this usually indicates a compiler bug - try recompiling\nwithout optimizations, and enable '-DLZO_DEBUG' for diagnostics)\n");
return 4;
}
I have all the header files inside my source directory:
config.h
lzo1x.h
lzoconf.h
lzodefs.h
miniacc.h
portab.h
portab_a.h
What am I doing wrong?
I am compiling my application in Ubuntu 10.10 in Anjuta ide.
Headers is not enough, you need to link to the libraries. Have you read the documentation?