|
Hello everyone,
Confused why the following code can compile when we use :: to invoke non-static function. Especially, in the sample below, is it possible to call non-static member if T is not int? Why compiler allows the following code to compile?
(sorry that I have tried to use the code tag to format the code, but some template statement is not correct after using code tag, so I post the raw code here.)
template <class T>
struct FooTrait {
public:
int foo1();
};
template<>
struct FooTrait <int> {
static int foo1() {cout << "Hello Foo1" << endl; return 0;}
};
template <class T1, class T2=FooTrait<T1>>
struct Foo {
public:
int foo()
{
T2::foo1();
return 0;
}
};
int main()
{
Foo<int> f;
f.foo();
return 0;
}
thanks in advance,
George
modified on Saturday, February 23, 2008 5:09 AM
modified on Saturday, February 23, 2008 5:48 AM
|
|
|
|
|
Hey George, please reformat your code.
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
|
|
|
|
|
Hi CPallini,
I would like to format and I have tried for 10 minutes. But I can not make
template<>
struct FooTrait<int>
to disply when I use <code> tag.
Any advice?
regards,
George
|
|
|
|
|
template <typename T>
use the < > button to set the angle brackets
|
|
|
|
|
Still not working, Rajkumar.
What is your settings for the following choices?
Ignore HTML tags in this message (good for code snippets)
Auto-encode HTML when pasting?
regards,
George
|
|
|
|
|
i mean by < > button is, see the formatting panel when editting the message, u will find the buttons it adds < and > for < and > symbols respectively.
|
|
|
|
|
|
George_George wrote: Any advice?
Yes, use < to show < symbol inside <pre> tags.
If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.
This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke
|
|
|
|
|
|
Hi CPallini,
I have fixed in the original post according to your comments.
regards,
George
|
|
|
|
|
George_George wrote: T2::foo1(); // possible to call non-static member if T is not int?
how do you determined it.
Foo<int> f;
f.foo();
return 0;
change to
Foo<char> f;
f.foo();
return 0;
and try.
|
|
|
|
|
Thanks Rajkumar,
1.
I have tried and when using char, the compile does not pass. So, the conclusion is, the source code is not finalized until we specify a type? And compiler will use the finalized template type to check compile rule?
(for example, in my original code, only int type is used, and only template specialization for int is used and compiled into final code?)
2.
It seems it is allowed to re-declar some method from non-static to static in template specialization class/struct. Is it good code or of any practical usage?
regards,
George
|
|
|
|
|
George_George wrote: So, the conclusion is, the source code is not finalized until we specify a type? And compiler will use the finalized template type to check compile rule?
"A template declaration itself does not generate code; it specifies a family of classes or functions, one or more of which will be generated when referenced by other code." from msdn; may be the answer?
George_George wrote: It seems it is allowed to re-declar some method from non-static to static in template specialization class/struct. Is it good code or of any practical usage?
IMHO why not. template specialization is the additional flexibility. Normally if you are not using it, you will find it sometimes useful.
|
|
|
|
|
Thanks Rajkumar,
Sorry for my English reading.
You mean change the static property of the method has some practical usage and is not bad code?
Rajkumar R wrote: IMHO why not. template specialization is the additional flexibility. Normally if you are not using it, you will find it sometimes useful.
regards,
George
|
|
|
|
|
yes, IMHO. i am not getting a good usage now to give.
George_George wrote: Sorry for my English reading
No. may be i am poor in english
|
|
|
|
|
Thanks Rajkumar,
My question is answered.
regards,
George
|
|
|
|
|
After I add the manifest file, it works.
But still has a problem, when I select an item, the image become invisible, it's fine when I use LVS_EX_BORDERSELECT.
---------------------
#define WINVER 0x0501
#define _WIN32_WINNT 0x0501
m_lstPic.ModifyStyleEx(0,LVS_EX_HIDELABELS);
I think this should be ok, why there is still labels?
----------------------
I used LVS_ICON style.
The problem is that I can't control the size of the label well.
Is there any way?
Thanks
modified on Saturday, February 23, 2008 7:15 AM
|
|
|
|
|
How can I browse filesystem in windows mobile device with SHBrowseForFolder
Hello, everyone.
I plan to develop one application that have to copy file to windows mobile device. Users will specify the directory of windows mobile device. So I have to pop up one dialog to let user to select.I try to use SHBrowseForFolder to select directory in windows mobile device. But it does not work.
There is no 'Mobile device' direcoty under desktop.I can browser this device by window's exporer. I even can access Nokia device with SHBrowseForFolder.The device is Jingpeng S6186. It used windows mobile 5.1.478. ActiveSync's version is 4.5.0.
|
|
|
|
|
Hello everyone,
Some study these days about address of exported function in DLL. It is appreciated if you could review and comment.
(for C++ DLL/EXE only)
1.Inside DLL
For the DLL itself, in its build process, it has to decide the address of each function in the result binary file (.dll), right? I think each exported function in the dll biinary file has a definite absolute address, which is based on the default load base address of the DLL, right?
2. Inside EXE
For the EXE, which utilizes the DLL, since there may be DLL rebase, which changes the base load address of the DLL into EXE process space, the address of the exported function of the DLL is not known until EXE load time and the final rebased address of the exported function of the DLL is put into the IAT of EXE.
In other words, the same function of the DLL, may have different address in different EXEs, which loads the DLL into related process space of EXE.
So, all the places where invokes the dllexported function, will use a function pointer to point to the IAT entry of the related dllexported function to form a indirection level function call. No address of DLL exported function is known and written to the EXE binary until load time address fix-up.
My understanding of (1) and (2) correct? Any comments?
thanks in advance,
George
|
|
|
|
|
George_George wrote: 1.Inside DLL
For the DLL itself, in its build process, it has to decide the address of each function in the result binary file (.dll), right? I think each exported function in the dll biinary file has a definite absolute address, which is based on the default load base address of the DLL, right?
This is correct. The DLL also contains relocation information which contains all the data needed by the loader relocate the DLL if needed.
George_George wrote: 2. Inside EXE
For the EXE, which utilizes the DLL, since there may be DLL rebase, which changes the base load address of the DLL into EXE process space, the address of the exported function of the DLL is not known until EXE load time and the final rebased address of the exported function of the DLL is put into the IAT of EXE.
In other words, the same function of the DLL, may have different address in different EXEs, which loads the DLL into related process space of EXE.
So, all the places where invokes the dllexported function, will use a function pointer to point to the IAT entry of the related dllexported function to form a indirection level function call. No address of DLL exported function is known and written to the EXE binary until load time address fix-up.
This is mostly correct. A minor point:
When you say "DLL rebase" it looks like you mean "relocate". "Relocation" is the when the loader changes the address of a DLL because the address is already taken. "Rebase"ing is similar but is not performed by the loader at runtime but by a tool which alters the actual executable.
Steve
|
|
|
|
|
Thanks for your correction, Steve!
My question is answered.
regards,
George
|
|
|
|
|
I am using gdi plus to draw the string i.e. DrawString() every thing is fine but I could not find any method to change the following.
1.text outline border width and color.
2.text background color.
can any one help me
Trioum
|
|
|
|
|
I think Graphics::DrawString() renders only text,
trioum wrote: 1.text outline border width and color.
use Graphics::DrawRectangle()
trioum wrote: 2.text background color.
use Graphics::FillRectangle()
and order can be background fill, Text, border.
|
|
|
|
|
how can I change the text outline border width and color
Trioum
|
|
|
|
|
trioum wrote: how can I change the text outline border width and color
Did you already done and just want to change 
if not, may be the following is useful. i just learnt how to.
Graphics graphics(hdc);
WCHAR szText[] = L"Sample Text";
Font arialFont(L"Arial", 16);
RectF layoutRect(25.0f, 25.0f, 150.0f, 25.0f);
StringFormat format;
format.SetAlignment(StringAlignmentCenter);
graphics.FillRectangle(&SolidBrush(Color(255, 100, 100, 100)), layoutRect);
graphics.DrawString(szText,11, &arialFont, layoutRect, &format, &SolidBrush(Color(255, 0, 0, 0)));
graphics.DrawRectangle(&Pen(Color::Black, 1), layoutRect);
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
