|
Is it possible to release the memory allocated for an item in a deque when you remove an item from a deque using either pop_back() or erase(). I'm trying to avoid the large delay when deleting the deque object.
typedef struct
{
BYTE b;
WORD w;
DWORD d;
}S;
void func(void)
{
deque<S> *pDeque=new deque<S>;
S s={1,2,3};
for (int i=0;i < 100000;i++)
{
s.d=i;
pDeque->push_back(s)
}
while (pDeque->size())
pDeque->pop_back();
delete pDeque;
}
Systems AXIS Ltd - Software for Business ...
|
|
|
|
|
Steve Thresher wrote:
I'm trying to avoid the large delay when deleting the deque object
simple. don't delete the deque.
Prefer storing STL containers on the stack rather than on the heap, after all that's one of the huge benefits of using STL containers. In this case prefer passing it in as a parameter and recycling it. If you've profiled the method this should give some kind of performance boost.
Passing the container also has the benefit of not allocating space for an element every time you push_back. The push_back is what's making the destructor so slow, because the allocator has to destroy all the elements you've just created, as well as the (comparatively small) empty deque.
If you can keep you head when all about you
Are losing theirs and blaming it on you;
If you can dream - and not make dreams your master;
If you can think - and not make thoughts you aim;
Yours is the Earth and everything that's in it.
Rudyard Kipling
|
|
|
|
|
As far as I know using the stack makes no difference as the deque will still be deleted when it goes out of scope and I will still be left with a pause.
How are you supposed to add elements if you don't use push_back()?
Systems AXIS Ltd - Software for Business ...
|
|
|
|
|
I'm surprised that the memory for the item isn't being deleted in the pop_back(). You've you stepped into the pop_back() code in the debugger to verify this, or looked at the source?
Even if it is though, I don't think it would help in your example. ie. Whether they are deleted in your pop_back() loop or by the delete pDeque would have the same affect.
I think the problem here is the sheer size of the container and the fact that you are doing 100,000 new/delete (malloc/free). There are several fast malloc/free implementations which would help or maybe you shouldn't be using a deque for this.
Neville Franks, Author of ED for Windows. Free Trial at www.getsoft.com
|
|
|
|
|
Hi, i did develop an ATL component, based on a Composite Control. This Component must be used on a VB App or a Delphi APP. I want to set an alignment property(e.g align to bottom, align to top, align to client....)like the one present in some controls. I tried to do this by processing messages sent to my control, but none of them are related to Parent App movement. Any idea how can I accomplish this?
Tnx
|
|
|
|
|
I am writing the addin in ATL. To implement the desired feature i want to access the information about the classes included in the workspace. Anyone knows how to get the information.
Thanx
-qur
|
|
|
|
|
I am trying to create an object derived from IFont... I want to add a few properties...
I have used the wizard to create my IMyFont and replaced IDispatch by IFont in the interface declaration.
Also replaced in the BEGIN_COM_MAP(CMyFont) the COM_INTERFACE_ENTRY(IDispatch) by COM_INTERFACE_ENTRY(IFont) but still I have some compilation errors ... cannot instantiate abstract class.
I suspect I have to change something in the list of derived class from CMyFont... mess about with
public IDispatchImpl<imyfont, &iid_imyfont,="" &libid_vistactrllib="">
but how?
Any idea welcome!
Thanks,
Jerry
|
|
|
|
|
Development of an ATL Control Container(CRetailMapX) to host Mappoint control using Visual Studio .NET(debug build)
Implementation and Problem Description : I did develop a ATL in-process server component and added a ATL composite control named CRetailMapX (which implements all the default interfaces, those comes when this control was added through the studio wizard).
Steps I followed are like this:
(1) Imported Mappoint MappointControl.ocx to get all the GUIDs for the interfaces/object exposed by mappoint control
#import "C:\Program Files\Microsoft MapPoint\MappointControl.ocx" raw_interfaces_only, raw_native_types, no_namespace, named_guids
(2) Declared member variable of type CAxWindow(also tried with CAxWindow2), to manipulate the window that hosts the Mappoint control.
Also declared IMappointCtrlPtr m_pMapCtl and MapPtr m_pMap, will be used later when mappoint control created on map.
(3) On OnInitDialog() added the code like this
// Display size for the control
RECT rect;
SetRect(&rect, 0,0,100,100);
// Get string from the CLSID of mappoint control
OLECHAR szControlName[40];
StringFromGUID2(CLSID_MappointControl, szControlName, 40);
// Create window based on mappoint control CLSID
>>>>> HWND hWnd = m_MapControl.Create(m_hWnd, rect, OLE2CT(szControlName), WS_CHILD|WS_VISIBLE, 0, 5000);
// Get the IMappointCtrl pointer
HRESULT hr = m_MapControl.QueryControl(&pMapCtl);
if(SUCCEEDED(hr))
{
// Hardcoded variant variable to open North America map
VARIANT vtMyInt;
vtMyInt.iVal = 1;
vtMyInt.vt = VT_I2;
// Create new map, this call require to create new NA map
>>>>> pMapCtl->NewMap(vtMyInt, &pMap);
}
Here is the problem. Everything goes fine till NewMap() method of IMappointCtrl interface called, during the call to NewMap(), execution goes through OnInPlaceActivate() function of CAxHostWindow and asserts.
If all the assertions are ignored, mappoint control will be displayed/created, but no user interactions (by mouse or by keyboard) will be supported.
Even I tried CAxWindow::Attachcontrol() instead of CAxWindow::Create(), but the result is the same, assertions!. Reason is that OnInPlaceActivate() of CaxHostWindow in VS6.0 doesn’t have ATLASSERT statements as compared to OnInPlaceActivate() of CAxHostWindow in VS7.0
Going through the OnInPlaceActivate() of CAxHostWindow, tells that this In-place activation event should not be generated second time. As I mentioned, this event occurred when NewMap() function of Mappoint control is called to show/create a map. As per my understanding of this problem, when first time CAxWindow::Create() is called, prompts in-place activation event which calls OnInPlaceActivateEx() where m_bInPlaceActive was set to TRUE and also queried for m_spInPlaceObjectWindowless. After this when NewMap() is called, also prompt the in-place activation event and calling OnInPlaceActivate(), where all the above assertions comes.
I don't know how to avoid second time generation of in-place activation event, but NewMap() method of IMappointCtrl interface has to be called to create/show either European or North America map.
As per my experimentation in this way for hosting other activex controls (Calender and MSFlexGrid but Mappoint control) works fine, only the difference here is that, just calling CAxWindow::Create() for those controls was enough to displays/creates them, those controls doesn't require us to call something like NewMap()( in case of Mappoint control) to create/display them.
I tried the same implementation on VS6.0, got only the last one of the three assertions mentioned above, sinceCAxHostWindow:: OnInPlaceActivate () doesn’t have ATLASSERT statements to assert for in-place not activated yet and m_spInPlaceObjectWindowless is NULL, but CAxHostWindow:: OnInPlaceActivate() in VS7.0 has those ATLASSERT statements.
|
|
|
|
|
i have a vector of structur like that :
struct CharacterCoordinate<br />
{<br />
GLfloat xpos;<br />
GLfloat zpos;<br />
int Xahead,Zahead;<br />
double pasZ;<br />
double pasX;<br />
<br />
<br />
};
and i want to remove just some of them from the vector and i make like follow:
<br />
void Scene::RemoveCharacter(int amount)<br />
{<br />
CharacterCoordinate PtrCreat;<br />
<br />
int numCharacter = CharacterPos.size();<br />
<br />
for(int i=0;i<amount;i++)<br />
{<br />
<br />
PtrCreat = CharacterPos[numCharacter - 1];<br />
CharacterPos.pop_back();<br />
delete PtrCreat;<br />
<br />
<br />
}<br />
}<br />
but it don't work,what's the problem in that??
|
|
|
|
|
There's quite a few problems in that snippet mate.. But if you want to remove an element or range of elements from a vector, then use vector::erase. Calling delete on stack allocated objects is a pretty bad idea too.
Ryan
|
|
|
|
|
Please use the 'Preview' button to check for problems before posting your messages. It'll make it much easier to answer your questions. Thanks 
- Mike
|
|
|
|
|
|
You're missing the '<' in your for header. I had to look at the page source to see what your code for that line was.
I'm assuming that RemoveCharacter() is just supposed to remove amount entries from the end of the CharacterPos vector. There are a number of problems with your code:
First, you're using vector::pop_back() which decreases the size of the vector each time, but you're always getting the item at numCharacter - 1. After one iteration, that index no longer exists. I think you want CharacterPos[CharacterPos.size()-1] instead. Note that you can get the last element with CharacterPos.back(), which is equivalent.
Second, what type is CharacterPos? Is it a vector<CharacterCoordinate> or a vector<CharacterCoordinate*>? We need to know. Please include the error you're getting as well.
As an aside, if CharacterPos is of type vector<CharacterCoordinate>, then you don't need a for-loop. Look up the vector::resize function.
- Mike
|
|
|
|
|
I get the following error when I build my atl project.
"myProductLookup error PRJ0019: A tool returned an error code: "Performing registration""
The catch is I am attempting to use a c api call. I have setup the project properties the same as for creating a regular cpp dll with the api calls. Any ideas how I can incorporate the api's into my project?
Overall goal is to execute a query in my api against a legacy system from a web application. Kind of like cgi. Thoughs?
Thanks.
Courtney
|
|
|
|
|
Well, didn't got it clearly, but the Performing registration stuff happened, when the Visual Studio tries to register your DLL (through calling the from-your-dll-exported function DllRegister or how it is exactly). If there's not such a method or if it failed inside, you'll get this message.
Then, if you have some COM enabled library that should be registered, you have to correct the registration code.
If it is old plain DLL, then you have to switch off the project settings that tries to register the library after building it.
|
|
|
|
|
I would like to write a wrapper around the deque implementation whereby the user cannot access the original deque implementation and I'm wondering if it would be best to derive a new template class from deque eg.
<br />
template <class T><br />
class CMyClass : private deque<T><br />
{<br />
};<br />
or just have a private workarea within my new class eg.
<br />
template <class T><br />
class CMyClass<br />
{<br />
private:<br />
deque d;<br />
<br />
public:<br />
};<br />
Can anybody offer a suggestion as to which is 'more correct' and why. The main problem I am having with deriving my class from deque is how to handle the use of the overloaded [] (subscript) operator. If I do nothing the compiler gives an error due to private inheritance.
Any suggestions will be gratefully received.
Systems AXIS Ltd - Software for Business ...
|
|
|
|
|
Being inaccesible due to private inheritance is tantamount, from the user point of view, to not having this particular member present. If you find the error message annoying nevertheless (I also think it can be misleading), your only alternative is to have a private member. None solution is in principle preferred over the other, though private inheritance simplifies a little the process of publicizing private member functions:
struct A
{
void f();
void g();
};
struct B: private A
{
using A::f;
};
struct BB
{
void f(){a.f();}
private:
A a;
};
Joaquín M López Muñoz
Telefónica, Investigación y Desarrollo
|
|
|
|
|
I finally worked out how to call the overloaded subscript operator for the base class as follows...
<br />
template <class T><br />
class CMyClass : private deque<T><br />
{<br />
<pre><br />
T& operator[](DWORD subscript)<br />
{<br />
return((*(dynamic_cast<deque<T>*>(this)))[dwSubscript]);<br />
}<br />
</pre><br />
}<br />
Its very ugly but it works and thats good enough for me. Thanks for taking the time to help me out. Very much appreciated.
Systems AXIS Ltd - Software for Business ...
|
|
|
|
|
|
Thanks for the info. Very much appreciated.
Systems AXIS Ltd - Software for Business ...
|
|
|
|
|
Hi,
I'm using this libary to convert PNG's to dibs. When i use a function in this lib, however, when i try to compile it, it gives me errors like this.
error LNK2005: _malloc already defined in LIBCMT.lib(malloc.obj)
And it gives them for a lot of functions. I do know what they mean i just don't know what to do about it....
I can compile it when i use Debug Multitread DLL.
I already contacted the auther of this project by mail but he doesn't seem to respond.
Thanks in advance
|
|
|
|
|
This happens if you use one setting for which runtime to use, and the lib was built using another. If it works when you use 'Debug Multithread DLL' but not when you use 'Debug Multithread', guess which one they used  ?
Steve S
[This signature space available for rent]
|
|
|
|
|
Hi,
I need the solution or design where I should write the data strcut like this -
MyStruct{
string1 ;
string2;
float[ ]
}s;
s[ ];
I have written string1 and string2 using vector as these are dynamic arrays but I am not able to write dynamic arrays of float. I am planning to use something like vector here again.
Can somebody help me in this regard?
Regards,
MB
|
|
|
|
|
Use std::vector? It should work.
Best regards,
Alexandru Savescu
P.S. Interested in art? Visit this!
|
|
|
|
|
This is same structure I am looking to store or read
Name1, Name2
Surname1, Surmane2
MiddleName1,MiddleName2 //This structure has been stored in vector successfully
12, 12, 13
15, 14, 12
12, 34, 45 //Marks
// Marks in some test and these tests can grow up to 100 or 200
// and this arrays are not able to store in above vector
Please suggest me code or design for this problem.
Regards,
MB
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
