|
Thanx for the help !!!!!!!! 
|
|
|
|
|
Hello everyone,
I find that _stprintf is defined to swprintf. And in the prototype of swprintf, the 1st and the 2nd parameters are both wchar_t. I think the other parameters should also be w_chart type, but in some situations, I always find others use char* type (which seems not unicode character type) as the parameter, like,
--------------------
// buffer is wchar_t* type.
_stprintf (buffer, L"%s", (char*)pointer);
--------------------
Definitions from Microsoft,
--------------------
#define _stprintf swprintf
_CRTIMP int __cdecl swprintf(wchar_t *, const wchar_t *, ...);
--------------------
Can we use char* as the 3rd parameter? If we can, why can it (non-unicode type) be compatible with wchar_t type (unicode type)?
thanks in advance,
George
|
|
|
|
|
actually, _stprinf() is defined to swprintf() only if UNICODE is defined in your project. if not, then it is defined to sprintf().
so, instead of using L"" literals, prefer using _T("")
TOXCCT >>> GEII power
[VisualCalc 3.0 updated ][Flags Beginner's Guide new! ]
|
|
|
|
|
Thank you toxcct,
Why _T is better than L?
My question is whether it is correct to use char* type as the 3rd parameter in my sample. If it is not correct, what potential issues will be?
regards,
George
|
|
|
|
|
L defines unicode literal strings always, when _T defines unicode strings only if the project is compiling in unicode (wchar_t), otherwise, simple char strings.
TOXCCT >>> GEII power
[VisualCalc 3.0 updated ][Flags Beginner's Guide new! ]
|
|
|
|
|
Thanks toxcct!
What header files need to be included if I want to use _L?
By the way, could you come back to the sample I wrote above to comment whether it is correct to use char* (non-unicode type) as the 3rd parameter of _stprintf? Thank a lot! 
regards,
George
|
|
|
|
|
I would suggest you don't use the sprintf family of functions. It has just so many problems it's not funny. Here's some of them:
- It's not type safe.
- It's not extensible.
- It's plain not safe. How big do you make the buffer? Make it too small and you're in for a buffer overrun, which is no fun.
I'd use code like the following:
ostringstream ss;
ss << "The number is " << 42;
cout << ss.str() << endl;
This code assumes the following includes:
<iostream>
<string>
<sstream>
And this:
using namespace std;
With this code none of the listed problems occur.
Steve
|
|
|
|
|
Stephen Hewitt wrote: ss << "The number is " << 42;
Now you'll have to find the universal question 
|
|
|
|
|
Hi Cedric,
What do you mean univeral question?
regards,
George
|
|
|
|
|
No one knows. I have it on good authority that the answer is 42 however.
Steve
|
|
|
|
|
What do you mean 42?
regards,
George
|
|
|
|
|
No, it is just a popular joke. Did you see the movie "The Hitchhiker's Guide to the Galaxy" ?
42 is the answer to the universal question. But it isn't funny at all if you didn't see the movie 
|
|
|
|
|
I see. I have not seen the movie before. I have a lot of souce codes to read. Could you come back to this question and show me an example of mixed unicode type (wchar_t* and _T) and non-unicode type (char*), just as I mentioned at the beginning of this thread, like,
_stprintf (wchar_t*, _T("%S"), char*);
I just can not imagine that we can mix unicode type and non-unicode type in one statement.
regards,
George
|
|
|
|
|
Thank you Steve!
What means "plain not safe"?
Actually, I like your style. But I am reading and learning programs written by others. So, could you come back to the sample I wrote above to comment whether it is correct to use char* (non-unicode type) as the 3rd parameter of _stprintf? Thank a lot!
regards,
George
|
|
|
|
|
George_George wrote: What means "plain not safe"?
This is what I had in mind:
char buffer[1];
sprintf(buffer, "The number is %d", 42);
This code will result in a buffer overrun which at best will crash the program and at worst will open up a security flaw will can be exploited.
Steve
|
|
|
|
|
Thank you Steve!
Maybe I have not made myself understood. My question is not whether the buffer (as the 1st parameter) is large enough (suppose it is large enough in this case, I agree with you about the setbacks of this method), my question is that whether I can input char* (non-unicode type) as the 3rd parameter, while the 1st parameter is wchar_t* type and the 2nd parameter is L (constant string) type.
regards,
George
|
|
|
|
|
Yes. Use %S instead of %s.
Steve
|
|
|
|
|
Thanks Stephen,
I have found that %S is used to specify a wide-character string. Could you show me an example of mixed using of unicode type and non-unicode type, in the following format -- without any errors to print out anything?
I just can not imagine why we can mix the use of unicode type (w_char* and _T) and non-unicode type (char*).
_stprintf (wchar_t*, _T"%s", char*)
regards,
George
|
|
|
|
|
|
Thanks Michael!
Suppose that UNICODE macro is define and we can pass char* type parameter to _stprintf? I can not imagine that -- I think it is not safe. Could you show me an example please to show it works to pass char*?
regards,
George
|
|
|
|
|
Now that I looked more closely at your question  I see where the confusion is coming from. %s means a string of TCHARs, so that will change between ANSI and Unicode builds. The ... in a parameter list means "any number of parameters of any type" - it's totally unrelated to the character set you're building with, but the format string does have to take character sets into account:
TCHAR sz[100];
_stprintf(sz, _T("Hello %hs"), "Bob");Note that I used %hs to mean an ANSI string.
--Mike--
Visual C++ MVP 
LINKS~! Ericahist | PimpFish | CP SearchBar v3.0 | C++ Forum FAQ
|
|
|
|
|
Thank you Mike!
If UNICODE is defined, then in your above example, TCHAR will be wchar_t, _T will be defined as L and "Bob" will still be char* type (non-unicode type). Is that correct?
I am not sure whether your above example will still work if UNICODE is defined since you print a non-unicode type (char*) to a unicode type _T. If your example works, could you show me please?
regards,
George
|
|
|
|
|
|
Hi Mike,
If I correctly understand, you mean debug into sprintf? C runtime library is not an open source provided by Microsoft.
regards,
George
|
|
|
|
|
Hi,
I know the ISO-9660 file format but I'm unble to generate IS0-9660 file system...I don't want to use any CD writing tools or other ISO creation tools for the creation of iso file ie, I have to write the code in vc++ for creating iso file...
Please specify the sites where I'll get the source code or suggest me for the creation of iso file
I'm eagerly waiting for ur valuable suggestions...
Regards,
nag.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
