Question

这段代码会产生巨大的内存泄漏。你能帮我找出它发生的地方吗？

此代码执行以下操作：

1）它获得了IHTMLDocuments2接口 2）请求所有标签集合 3）迭代整个集合 4）并将一些标签的数据添加到列表

IDispatch* pDisp;
pDisp = this->GetHtmlDocument();

if (pDisp != NULL ) 
{
    IHTMLDocument2* pHTMLDocument2;
    HRESULT hr;
    hr = pDisp->QueryInterface( IID_IHTMLDocument2,(void**)&pHTMLDocument2 );
    if (hr == S_OK)
    {
                    // I know that I could use IHTMLDocument3 interface to get collection by ID
                    // but it didn't worked and returned NULL on each call.
        IHTMLElementCollection* pColl = NULL;
                    // get all tags
        hr = pHTMLDocument2->get_all( &pColl );
        if (hr == S_OK && pColl != NULL)
        {
            LONG celem;
            hr = pColl->get_length( &celem );
            if ( hr == S_OK )
            {
                                    //iterate through all tags
                                    // if I iterate this block of code in cycle, it 
                                    // uses memory available upto 2GBs and then
                                    // app crashes
                for ( int i=0; i< celem; i++ )
                {                       
                    VARIANT varIndex;
                    varIndex.vt = VT_UINT;
                    varIndex.lVal = i;
                    VARIANT var2;
                    VariantInit( &var2 );
                    IDispatch* pElemDisp = NULL;
                    hr = pColl->item( varIndex, var2, &pElemDisp );
                    if ( hr == S_OK && pElemDisp != NULL)
                    {
                        IHTMLElement* pElem;
                        hr = pElemDisp->QueryInterface(IID_IHTMLElement,(void **)&pElem);
                        if ( hr == S_OK)
                        {                   
                                                            // check INPUT tags only
                            BSTR tagNameStr = L"";
                            pElem->get_tagName(&tagNameStr);
                            CString tagname(tagNameStr);
                            SysFreeString(tagNameStr);
                            tagname.MakeLower();
                            if (tagname != "input")
                            {
                                continue;
                            }
                                                            //get ID attribute
                            BSTR bstr = L"";
                            pElem->get_id(&bstr);
                            CString idStr(bstr);
                            SysFreeString(bstr);

                            if (RequiredTag(pElem)) 
                            {       
                                AddTagToList(pElem);
                            }
                                                            //release all objects
                            pElem->Release();
                        }
                        pElemDisp->Release();
                    }
                }
            }
                            // I looked over this code snippet many times and couldn't find what I'm missing here...
            pColl->Release();
        }
        pHTMLDocument2->Release();
    }
    pDisp->Release();       
}

Answer 1

在你的循环中，对于没有标记名为"input"的每个已检索元素（这将是大多数元素），您在调用pElem->Release()时不会调用continue，所以你正在泄漏他们：

if (tagname != "input") 
{ 
    pElem->Release(); // <-- add this
    continue; 
}

话虽如此，您应该重新编写代码以使用ATL的智能指针类（CComPtr，CComQIPtr，CComBSTR等）为您管理内存，这样您就可以不必再自己手动释放所有东西，例如：

CComPtr<IDispatch> pDisp;
pDisp.Attach(this->GetHtmlDocument());
if (pDisp.p != NULL)    
{   
    CComQIPtr<IHTMLDocument2> pHTMLDocument2(pDisp);   
    if (pHTMLDocument2.p != NULL)   
    {   
        CComPtr<IHTMLElementCollection> pColl;   
        pHTMLDocument2->get_all(&pColl);
        if (pColl.p != NULL)
        {   
            LONG celem;   
            if (SUCCEEDED(pColl->get_length(&celem)))   
            {   
                for (LONG i = 0; i < celem; ++i)   
                {                          
                    VARIANT varIndex;   
                    varIndex.vt = VT_UINT;   
                    varIndex.lVal = i;   
                    VARIANT var2;   
                    VariantInit( &var2 );   
                    CComPtr<IDispatch> pElemDisp;   
                    pColl->item( varIndex, var2, &pElemDisp );   
                    if (pElemDisp.p != NULL)   
                    {   
                        CComQIPtr<IHTMLElement> pElem(pElemDisp);   
                        if (pElem.p != NULL)   
                        {                      
                            CComBSTR tagNameStr;   
                            pElem->get_tagName(&tagNameStr);   

                            if (lstrcmpiW(tagNameStr.m_str, L"input") != 0)
                                continue;   

                            CComBSTR idStr;   
                            pElem->get_id(&idStr);   

                            if (RequiredTag(pElem))    
                                AddTagToList(pElem);   
                        }   
                    }   
                }   
            }   
        }   
    }   
}

Webbrowser（COM）中的内存泄漏

1 个答案: