CreateFile, error

[gunnar66]

I try to create a file, the filename i will get from a string. What do I do wrong. It works ok in Visual C++, but in Emededd C++ i get an error.Some one told me that i just have to convert ,char* filnavn1 to a wide char format string Can some one show me how I do that. HANDLE hFile1; char *filename1; char *token; char seps[] = ","; char string[] = "\\windows\\2004-10-T-gr1-tstTag1.csv,\\windows\\2004-10-T-gr1-tstTag2.csv"; token = strtok( string, seps ); filnavn1 = token; hFile1 = CreateFile (filnavn1, // Open .txt GENERIC_READ, // Open for reading 0, // Do not share NULL, // No security OPEN_EXISTING, // Existing file only FILE_ATTRIBUTE_NORMAL, // Normal file NULL); // No template file Get this error: cannot convert parameter 1 from 'char *' to 'const unsigned short *

[Cedric Moonen]

Instead of using char*, try to use TCHAR*. So, for example for string: TCHAR* string = _T("\\windows\\2004-10-T-gr1-tstTag1.csv,\\windows\\2004-10-T-gr1-tstTag2.csv"); And instead of strtok, use wcstok.

[gunnar66]

Still get an error cstok' : cannot convert parameter 2 from 'char [2]' to 'const unsigned short * int ch = ','; TCHAR* token; int i = 0; int ch = ','; TCHAR* filnavn1; TCHAR* string = _T("\\windows\\2004-10-T-gr1-tstTag1.csv,\\windows\\2004-10-T-gr1-tstTag2.csv");

[Cedric Moonen]

Always use TCHAR instead of chars. This error comes because you pass a char array instead of a TCHAR array.

[jan larsen]

As I told you in another thread, there seems to be something fishy about the way your compiler handles the typedefs for TCHAR, PSTR, LPSTR and so on. When you invoke CreateFile, you actually invoke either CreateFileA or CreateFileW based on whether UNICODE is defined or not. Based on the error message, it seems that you are calling CreateFileW, which takes a unsigned short * as the filename parameter. This indicates that UNICODE is defined. I have no idea about what you mean when you mention "Emededd C++", but somewhere UNICODE is defined either explicit or implicit. You have some considerations to do: 1. If you are sure that you only need to build for unicode, then you could start to use the proper functions explicit like this:

HANDLE hFile1;
unsigned short *filename1;
unsigned short *token;
unsigned short seps[] = L",";
unsigned short string[] = L"\\windows\\2004-10-T-gr1-tstTag1.csv,\\windows\\2004-10-T-gr1-tstTag2.csv";
token = wcstok( string, seps );
filnavn1 = token;
hFile1 = CreateFile (filnavn1, // Open .txt
GENERIC_READ, // Open for reading
0, // Do not share
NULL, // No security
OPEN_EXISTING, // Existing file only
FILE_ATTRIBUTE_NORMAL, // Normal file
NULL); // No template file

2. Or you could find out why your compiler insists on using the *W functions, but in the same breath refuses to translate TCHAR to unsigned short. I would go for #2 :-) By the way, for runtime translation of string types: look up MultiByteToWideChar(), but as I said, you main problem is the confusion of string types. "After all it's just text at the end of the day. - Colin Davies "For example, when a VB programmer comes to my house, they may say 'does your pool need cleaning, sir ?' " - Christian Graus

[Antony M Kancidrowski]

jan larsen wrote: I have no idea about what you mean when you mention "Emededd C++" FYI, Microsoft eMbedded C++[^] Is a development IDE for writing code for Windows CE based devices. It instists on UNICODE. Ant. I'm hard, yet soft.
I'm coloured, yet clear.
I'm fruity and sweet.
I'm jelly, what am I? Muse on it further, I shall return! - David Walliams (Little Britain)

[jan larsen]

Ah, "Embedded", I had a vague idea that this was probably the case :-), I didn't know it insisted on UNICODE. "After all it's just text at the end of the day. - Colin Davies "For example, when a VB programmer comes to my house, they may say 'does your pool need cleaning, sir ?' " - Christian Graus

[jan larsen]

Thanks to Antony for clearing up some things here. Embedded C++ insists on UNICODE, this means that if you want to build for both platforms, you should use the proper typedefs when working with strings. Michael Dunn has written a very good article on this: The Complete Guide to C++ Strings[^] So it seems that the compiler confusion springs from: When compiling for the embedded environment, it will allways link to wide character functions, this means that you can't use char, and that, when you compile for the embedded environment, have to define UNICODE. "After all it's just text at the end of the day. - Colin Davies "For example, when a VB programmer comes to my house, they may say 'does your pool need cleaning, sir ?' " - Christian Graus

[gunnar66]

TCHAR* filnavn1 did it. Did not know that it was UNICODE. All works fine now. Thanks to all

[Antony M Kancidrowski]

Sorry, I think it is rather that CE requires UNICODE therefore the API implementations are all UNICODE. Ant. I'm hard, yet soft.
I'm coloured, yet clear.
I'm fruity and sweet.
I'm jelly, what am I? Muse on it further, I shall return! - David Walliams (Little Britain)