After reading an excellent answer on Dynamically rethrowing self-defined C++ exceptions as Python exceptions using SWIG, I am trying to throw a custom Exception from C++ and catch it in Python. I have tried to build the simplest example, so that I can see what I'm doing wrong. I would appreciate some help here.
I've started by creating a class called Example which has one method. In this method, I always throw my Exception which extends the Exception base class in C++. This exception is declared in the header.
I'm trying to catch this Exception in Python but so far I am unable to get this to compile. I tried some of the other solutions in the thread mentioned aboved, and while they compiled, I did not seem to be able to catch the Exception without Python aborting.
Here's what I have so far (a very basic project)
example.h
#ifndef EXAMPLE_H
#define EXAMPLE_H
#include <string>
#include <exception>
class Example
{
public:
std::string getName();
};
class ExampleException
{
public:
ExampleException(const std::string &what) { this->eMsg == (std::string(what)); }
const char * what() const throw() { return eMsg.c_str(); }
private:
std::string eMsg;
};
#endif // EXAMPLE_H
example.cpp
#include "example.h"
#include "exception"
using namespace std;
std::string Example::getName()
{
/* Always throw an exception just so we can test this */
throw ExampleException("Couldn't find a name");
return("John");
}
example.i
%module example
%include "example.i"
%include exception.i
%init %{
m_ExampleException = PyErr_NewException("_example.ExampleException", NULL, NULL);
Py_INCREF(m_ExampleException);
PyModule_AddObject(m, "ExampleException", m_ExampleException);
%}
%exception {
try {
$action
} catch (ExampleException &e) {
PyErr_SetString(ExampleException, const_cast<char*>(e.what()));
SWIG_fail;
}
}
%{
#include "example.h"
#define SWIG_FILE_WITH_INIT
static PyObject* m_ExampleException;
%}
test.py
import example
try:
print example.Example_getName()
except ExampleException as e:
print e.what()
And I run the following commands to build the project:
swig -verbose -python -c++ example.i
g++ -fPIC -c example_wrap.cxx -I/usr/include/python2.7 -o example_wrap.o -fPIC
g++ -c example.cpp -I/usr/include/python2.7 -fPIC
g++ -shared example_wrap.o example.o -o _example.so
Unfortunately, I am given the following error message as well as a warning:
example_wrap.cxx: In function ‘void init_example()’:
example_wrap.cxx:3839:82: warning: deprecated conversion from string constant to ‘char*’ [-Wwrite-strings]
m_ExampleException = PyErr_NewException("_example.ExampleException", NULL, NULL);
^
In file included from /usr/include/python2.7/Python.h:80:0,
from example_wrap.cxx:171:
/usr/include/python2.7/object.h:769:20: error: expected primary-expression before ‘)’ token
((PyObject*)(op))->ob_refcnt++)
I feel like I'm missing something obvious and would appreciate an extra pair of eyes. I have not used SWIG and C++ together extensively before.
Thanks for your time
Related
I have some python functions which I run from C++ side and I need them to return std::expected objects. What I did is this:
def do_operation(self):
print("Operation done")
return void_expected()
In Swig code I have something like:
%inline %{
std::expected<void, OperationError> void_expected()
{
return std::expected<void, OperationError>{};
}
%}
%typemap(out) std::expected<void, OperationError> {
if (!$1.has_value()) {
SWIG_exception(SWIG_RuntimeError, "Unexpected error!");
}
Py_RETURN_NONE;
}
But I am getting all these warnings running the python script:
swig/python detected a memory leak of type 'std::expected< void,OperationError > *', no destructor found.
The error message you're seeing hints towards what's going on. Although my older answer on that warning does apply here it shouldn't and that hints as to what's gone wrong. Your expected behaviour is that for the case where a value is returned in C++ None is return in Python. Conversely if there isn't a value in the expected object then a Python exception is raised. In neither case are you expecting SWIG to create a Proxy object on the Python side for the std::expected result.
So the fact that you've seen this warning indicate that something's not as expected in your interface as written. And it turns out to be fairly simple - the typemap you showed comes after the place where it's expected to be used. So it doesn't get applied.
We can confirm this by constructing a small, complete example:
%module test
%include <exception.i>
%{
#include "expected.h" // TODO: include <expected> insteadwhen I really have C++23 impl
%}
%typemap(out) std::expected<void, OperationError> {
if (!$1.has_value()) {
SWIG_exception(SWIG_RuntimeError, "Unexpected error!");
}
Py_RETURN_NONE;
}
%inline %{
class OperationError {};
std::expected<void, OperationError> void_expected() {
return std::expected<void, OperationError>{};
}
%}
(I had to make my own expected.h since I don't have a C++23 compiler to hand!)
// Minimal parts needed to prove the point with this answer
namespace std {
template <typename T, typename E>
struct expected {
bool has_value() const { return true; }
};
}
And some Python to prove it:
import test
test.void_expected()
Which we can then build and run like this:
swig3.0 -c++ -python -py3 test.i
g++ -shared -Wall -Wextra test_wrap.cxx -I/usr/include/python3.7 -m32 -o _test.so
python3 run.py
This runs with no error. Swap the order of the typemap and the %inline with void_expected() declaration/definition around however as in your question and you'll see the exact same error since the typemap isn't getting applied.
Dears,
I use SWIG to generate Python bindings to a C++ API (and it works great!) but I have serious difficulty wrapping a function that takes a vector of enum as argument. I have built a minimal example to simplify the debugging, which I put as an attachment to this issue. It seems to me that the example should work, at least it works well for a vector of integer argument.
The need is pretty simple : we have a C++ method with the following signature:
void Run(const std::vector<double> & in, std::vector<int> & out, std::vector<testing::Status> & status)
where testing::Status is an enum
and we will like to obtain a Python method like:
out, status = Run(in)
Using the attached example, the swig executable does not raise any error, and the Python Run method can be ran, but the output value status cannot be used, an error is raised:
status: (<Swig Object of type 'testing::Status *' at 0x7fa441156450>, <Swig Object of type 'testing::Status *' at 0x7fa441156660>) swig/python detected a memory leak of type 'testing::Status *', no destructor found. swig/python detected a memory leak of type 'testing::Status *', no destructor found.
Here are the different files that can be used to reproduce the error:
mylib.h, the C++ to wrap in Python
#include <vector>
namespace testing
{
typedef enum
{
Ok = 0,
Error = 1,
} Status;
class Algo
{
public:
void Run(const std::vector<double> & in, std::vector<int> & out, std::vector<testing::Status> & status)
{
status.resize(in.size());
out.resize(in.size());
for (int i=0; i<in.size(); ++i) {
out[i] = i;
status[i] = Status::Ok;
}
}
};
}
mymodule.i, the SWIG interface file
%module mymodule
%{
#include "mylib.h"
%}
%include "std_vector.i"
%include "typemaps.i"
%define STD_TEMPLATE(TYPE...)
%template() TYPE;
%apply TYPE& OUTPUT {TYPE&}
%typemap(argout) const TYPE& {
// do nothing for const references
}
%typemap(out) (TYPE&) = (const TYPE&);
%enddef
STD_TEMPLATE (std::vector <int>);
STD_TEMPLATE (std::vector <double>);
STD_TEMPLATE (std::vector < testing::Status >);
%include "mylib.h"
build.sh, the build command line used to compile binaries
${swig_install}/bin/swig \
-I. \
-I${swig_install}/share/swig/${swig_version}/python \
-I${swig_install}/share/swig/${swig_version} \
-c++ -python \
-outdir . \
-o "mymodule.cxx" \
"mymodule.i"
g++ -L${python_install}/lib -lpython3 \
-I${python_install}/include/python \
-I. \
-std=c++11 -shared -fPIC \
mymodule.cxx -o _mymodule.so
run.py, the example in Python that raises the error
import mymodule as mm
algo = mm.Algo()
out, status = algo.Run([1.1, 2.2])
print("out:", out)
print("status:", status)
SWIG doesn't know what to do with the vector of enum outputs. One way is to handle the typemaps yourself:
mylib.i
%module mylib
%{
#include "mylib.h"
%}
%include "std_vector.i"
%include "typemaps.i"
%define STD_TEMPLATE(TYPE...)
%template() TYPE;
%apply TYPE& OUTPUT {TYPE&}
%typemap(argout) const TYPE& {
// do nothing for const references
}
%typemap(out) (TYPE&) = (const TYPE&);
%enddef
STD_TEMPLATE (std::vector <int>);
STD_TEMPLATE (std::vector <double>);
// Don't require an input parameter in Python.
// Create a temporary vector to hold the output result.
%typemap(in,numinputs=0) std::vector<testing::Status>& (std::vector<testing::Status> tmp) %{
$1 = &tmp;
%}
// Create a Python list object the same size as the vector
// and copy and convert the vector contents into it.
%typemap(argout) std::vector<testing::Status>& (PyObject* list) %{
list = PyList_New($1->size());
for(int x = 0; x < $1->size(); ++x)
PyList_SET_ITEM(list, x, PyLong_FromLong($1->at(x)));
$result = SWIG_Python_AppendOutput($result, list);
%}
%include "mylib.h"
Output (same mylib.h and run.py):
out: (0, 1)
status: [0, 0]
I'm trying to embed Python 3.7 in a C application in Windows 10.
#define PY_SSIZE_T_CLEAN
#include <Python.h>
int main () {
Py_Initialize();
PyRun_SimpleString("print('OK')");
}
I use the following command to compile: (MinGW-W64-builds-4.3.4, gcc 7.3.0)
gcc "-IC:/Program Files/Python37_64/include" "-LC:/Program Files/Python37_64/libs" -lpython37 main.c
But it gives the following error:
C:\Users\Paul\AppData\Local\Temp\ccKQF3zu.o:main.c:(.text+0x10): undefined reference to `__imp_Py_Initialize'
C:\Users\Paul\AppData\Local\Temp\ccKQF3zu.o:main.c:(.text+0x25): undefined reference to `__imp_PyRun_SimpleStringFlags'
collect2.exe: error: ld returned 1 exit status
The strange thing is, when I try the same in Go 1.13 (Golang), it does work:
package main
/*
#cgo CFLAGS: "-IC:/Program Files/Python37_64/include"
#cgo LDFLAGS: "-LC:/Program Files/Python37_64/libs" -lpython37
#define PY_SSIZE_T_CLEAN
#include <Python.h>
void run () {
Py_Initialize();
PyRun_SimpleString("print('OK')");
}
*/
import "C"
func main () {
C.run()
}
Compile command:
go build python.go
How to fix this?
I found the solution in this answer.
The argument main.c has to be put somewhere before -lpython37.
So this works:
gcc "-IC:/Program Files/Python37_64/include" "-LC:/Program Files/Python37_64/libs" main.c -lpython37
This program in C runs and compiles well :
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
#include <stdio.h>
#include <string.h>
#include <errno.h>
#include <getopt.h>
#include <atasmart.h>
int main(){
const char *device = "/dev/sda";
int ret;
uint64_t ms;
SkDisk *d;
if ((ret = sk_disk_open(device, &d)) < 0) {
printf("Failed to open disk\n");
return 1;
}
if ((ret = sk_disk_smart_read_data(d)) < 0) {
printf("Failed to read SMART data: \n");
}
if ((ret = sk_disk_smart_get_power_on(d, &ms)) < 0) {
printf("Failed to get power on time:\n");
}
printf("%llu\n", (unsigned long long) ms);
return 0;
}
using:
gcc atatest.c `pkg-config --cflags --libs libatasmart`
However while trying to create python bindings based on that program:
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
#include <stdio.h>
#include <string.h>
#include <errno.h>
#include <getopt.h>
#include <atasmart.h>
#include <Python.h>
static PyObject *pySmart_powerOn(PyObject *self, PyObject *args)
{
const char *device = "/dev/sda";
int ret;
uint64_t ms;
SkDisk *d;
if (!PyArg_ParseTuple(args, "s", &device))
{
return NULL;
}
if ((ret = sk_disk_smart_get_power_on(d, &ms)) < 0) {
return Py_BuildValue("s", "Failed to get power on time");
}
return Py_BuildValue("K", (unsigned long long) ms);
}
static PyMethodDef pySmart_methods[] = {
{ "powerOn", (PyCFunction)pySmart_powerOn, METH_VARARGS, NULL },
{ NULL, NULL, 0, NULL }
};
PyMODINIT_FUNC initpySmart()
{
Py_InitModule3("pySmart", pySmart_methods, "Trial module");
}
I create a shared library using
gcc -shared -I/usr/include/python2.7 `pkg-config --cflags --libs libatasmart` atabind.c -o pySmart.so -fPIC
then I get a warning as follows :, but the file compiles
In file included from /usr/include/python2.7/Python.h:8:0,
from atabind.c:12:
/usr/include/python2.7/pyconfig.h:1158:0: warning: "_POSIX_C_SOURCE" redefined [enabled by default]
/usr/include/features.h:214:0: note: this is the location of the previous definition
when in Python i run
import pySmart
I get
ImportError: ./pySmart.so: undefined symbol: sk_disk_smart_get_power_on
My guess is that the error is caused because I have compiled the pySmart.so shared library with incorrect flags/options.. but I'm unable to figure it out!
You need to specify linker flags (-lfoo) after your source files. That's because of the way how linker works: when you specify a library to it, it checks the library for symbols needed so far. If no symbols needed (as if you didn't get to any source objects yet), it just skips the library.
Try the following commandline:
gcc -shared -I/usr/include/python2.7 \
`pkg-config --cflags libatasmart` \
atabind.c \
`pkg-config --libs libatasmart` \
-o pySmart.so -fPIC
You should include your Python.h first then any std header.
All function, type and macro definitions needed to use the Python/C API are included in your code by
#include "Python.h"
This implies inclusion of the following standard headers:
<stdio.h>, <string.h>, <errno.h>, <limits.h>, <assert.h> and <stdlib.h> (if available).
Since Python may define some pre-processor definitions which affect the standard headers on some systems, you must include Python.h before any standard headers are included.
Alternatively:
just _GNU_SOURCE , and it will be ignored by GNU libc's /usr/include/features.h
I am new to the Python C binding swig and have been trying to solve this problem for a while now. I have an external C library (Example.c) that I would like to call from Python. I read Swig tutorial and able to generate the wrapper in no time. The problem now is that when I invoke the API and I got this:
>>> import Example
>>> dir(Example)
['Example_CreateConnection', 'trimmed to fit the screen']
>>> Example.Example_CreateConnection("")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: in method 'Example_CreateConnection', argument 1 of type 'ExampleChar const *'
It seemed like it cannot find the type ExampleChar. The following is my swig file:
%module Example
%{
#include "ExampleSDK.h"
%}
%include "ExampleTypes.h"
%include "ExampleSDK.h"
ExampleTypes.h looks like this:
#ifndef ExampleTypes_H
#define ExampleTypes_H
typedef wchar_t ExampleChar;
#endif /* ExampleTypes_H */
ExampleSDK.h looks like this:
#ifndef ExampleSDK_H
#define ExampleSDK_H
#include "ExampleTypes.h"
void Example_CreateConnection(const ExampleChar *temp);
#endif /* ExampleSDK_H */
The following are the command lines being invoked to generate the wrapper:
swig -python -I. Example.i
gcc -c Example.c -I/Developer/SDKs/MacOSX10.6.sdk/usr/include/
gcc -c Example_wrap.c -I/usr/include/python2.6 -I.
gcc -bundle -flat_namespace -undefined suppress -o _Example.so Example_wrap.o Example.o -L/usr/lib/python2.6/config/ -lpython2.6
Here is how the Example.c looks like:
#include "runetype.h" // for Mac wchar_t definition
#include "ExampleSDK.h"
void Example_CreateConnection(const ExampleChar *temp)
{
//do nothing
}
I am not sure what is wrong with it. I hope someone will be able to point out the mistake(s) I have done over here. Thank you.
Regards,
Chuan Lim
Last time I used wchat_t with SWIG+Python I ended up needing to add something like:
%include "pywstrings.swg"
%include "pystrings.swg"
%include "std_string.i"
%include "typemaps.i"
%fragment("SWIG_AsVal_wchar_t", "header", fragment="<wchar.h>") {
SWIGINTERN int SWIG_AsVal_wchar_t(PyObject* p, wchar_t* c) {
return SWIG_OK;
}
}
%fragment("SWIG_From_wchar_t", "header", fragment="<wchar.h>") {
SWIGINTERNINLINE PyObject* SWIG_From_wchar_t(wchar_t c) {
return SWIG_Py_Void();
}
}
// Python -> C
%typemap(in) wchar_t const * {
$1 = PyString_to_wchar_t($input);
}
// C -> Python
%typemap(out) wchar_t * {
$result = wchar_t_to_PyObject($1);
}
in my Swig interface file.