Phil Thompson wrote:
> On Fri, 19 Feb 2010 10:42:51 +0100, Sébastien Petitdemange
> <[email protected]> wrote:
>> Hi Phil,
>>
>> We use the latest sip (4.10) to wrap your C++ library and I think I
>> found a bug in the wrapping code generated by sip.
>>
>> Sip Code:
>>
>> class HwSyncCtrlObj
>> {
>> %TypeHeaderCode
>> #include <HwSyncCtrlObj.h>
>> using namespace lima;
>> %End
>> public:
>>      struct ValidRangesType
>>      {
>>        double min_exp_time;
>>        double max_exp_time;
>>        double min_lat_time;
>>        double max_lat_time;
>>        const char* __repr__();
>> %MethodCode
>> std::ostringstream str;
>> str << *sipCpp;
>> sipRes = str.str().c_str();
>> %End
>>      };
>>
>>      virtual void getValidRanges(HwSyncCtrlObj::ValidRangesType& range
>> /Out/) = 0;
>> };
>>
>> C++ Code result:
>>
>> void sipVH_lima_8(sip_gilstate_t sipGILState,PyObject
>> *sipMethod,HwSyncCtrlObj::ValidRangesType& a0)
>> {
>>     PyObject *resObj = sipCallMethod(0,sipMethod,"");
>>
>>     if (!resObj ||
>>
> sipParseResult(0,sipMethod,resObj,"D4",sipType_HwSyncCtrlObj_ValidRangesType,&a0)
>> < 0)
>>         PyErr_Print();
>>
>>     Py_XDECREF(resObj);
>>     Py_DECREF(sipMethod);
>>
>>     SIP_RELEASE_GIL(sipGILState)
>> }
>>
>> We try to get the pointer of a0 instead of the reference!
>>
>> Is it a bug or miss I something?
> 
> If you are referring to the use of &a0 in the call to sipParseResult() then
> that is correct.
> 
> Phil
> 

Hi Phil,

It's not correct because you want to fill the structure instead of
retrieve the pointer of that structure.

I made a small patch:

void sipVH_lima_8(sip_gilstate_t sipGILState,PyObject
*sipMethod,HwSyncCtrlObj::ValidRangesType& a0)
{
    PyObject *resObj = sipCallMethod(0,sipMethod,"");

    HwSyncCtrlObj::ValidRangesType *tmpValue = NULL
    if (!resObj ||
sipParseResult(0,sipMethod,resObj,"D4",sipType_HwSyncCtrlObj_ValidRangesType,tmpValue)
< 0)
        PyErr_Print();
    else
      a0 = *tmpValue;

    Py_XDECREF(resObj);
    Py_DECREF(sipMethod);

    SIP_RELEASE_GIL(sipGILState)
}


Sorry I wasn't really clear In my first mail...

Regards,

SEB

<<attachment: sebastien_petitdemange.vcf>>

_______________________________________________
PyQt mailing list    [email protected]
http://www.riverbankcomputing.com/mailman/listinfo/pyqt

Reply via email to