char[4] to char[40]

Hey all,

im having this problem here (kind of n00b problem):

i have a struct as follows:

Code:

---

typedef struct
{
        TCHAR strName[40];
        TCHAR strPath[64];
    bool bStarted;
} Program;

---

and im trying to read a file that contains infos to put that into

Program X[50];

later on.

now when im reading the file and trying to put the name specified in the file, like "asd", into the X program im always getting the error that i cant convert char[4] to char[40].

X[i].strName = "asd";

Any help is appreciated.

thx

You need to use strcpy. You also need to properly honour the fact that you're using TCHAR and wrap the string literal with TEXT(), like this: TEXT("asd"). You also should not use a type that's dependent on a compilation flag, such as TCHAR, for data stored in a file - one version of the program would be unable to read files generated by the other!
Instead you should use WCHAR or wchar_t and convert to a different character set manually in ANSI versions of the program - use WideCharToMultiByte() for that.

ok thx for the quick reply gonna try that

another option is TCHAR * strName or String since this is c++

TCHAR* will cause memory troubles (you have to allocate a buffer and remember to deallocate it.)
Both will not work if the Program struct is supposed to be written directly to the file using fwrite/ostream::write.

using WCHAR works now with using strcpy and a reinterpret_cast to convert the WCHAR array to a char *

Code:

---

strcpy(reinterpret_cast<char *>(X[i].strPath), ((*(vIter2)).substr((*(vIter2)).find("=") + 1)).c_str());

---

thx again

What is all this TCHAR bull anyhow? What's wrong with good ol' char* / char[] ? (Ok so I know what's wrong, but meh :p). Let me guess, TCHAR is an MSVC specific re-def type thingy around char buffers? Eww.

well, it's actually from the platform SDK, not the compiler library...

from WTypes.h:

Code:

---

typedef wchar_t WCHAR;

typedef WCHAR TCHAR;

---

So a WCHAR is a TCHAR is a wchar_t, I guess someone just liked his capitals ;)
Of course these tricks save quite some work when moving from a 16 bit to a 32 bit environment, all you now need to do is change the typedef in one header file and recompile, instead of changing potentially hundreds of thousands of declarations.

Quote:

---

Originally Posted by jwenting

well, it's actually from the platform SDK, not the compiler library...

from WTypes.h:

Code:

---

typedef wchar_t WCHAR;

typedef WCHAR TCHAR;

---

So a WCHAR is a TCHAR is a wchar_t.

---

Wrong. If you care to take a closer look, you'll find that if the UNICODE macro is not defined, either WTypes.h isn't even included or TCHAR is a typedef for CHAR, which is a typedef for char.
In addition, you'll find that

Code:

---

const wchar_t *ptr = "hello";

---

doesn't compile, so if WCHAR == TCHAR, then the code wouldn't compile.

And DON'T DON'T DON'T use wchar_t/char reinterpret_casts!!! You'll get nonsensical data.

Yes the whole point of TCHAR is so if you have #define UNICODE, TCHAR will be a typedef for wchar_t (and the TEXT() macro will place the L prefix on strings enclosed in the macro) and if UNICODE is not defined, TCHAR will be a typedef for char (and the TEXT() macro will do nothing).

Code:

---

#ifdef UNICODE

#define __TEXT(q) L##q
#else
#define __TEXT(q) q
#endif

#define TEXT(q) __TEXT(q)

---

That's straight from winnt.h.