CString::Format issue

Hi,

I've recently been forced to change my project from using a MBCS to Unicode. This has thrown up a number of problems with my code one of which is below.

text.Format ("Row %d,Col %d",ln,col);

Thats is the line that I had in when using MBCS, I know how to format strings etc under unicode, and the problem isnt the change, it is what would that string appears as under MBCS, I cant seem to find much on what the %d thing would do, so if some one could tell me how the string text would appear after that code was successful, that would be great

Thanks

Will

Answer this question

CString::Format issue

Normality2000
_wtof is a function that only takes wchar_t chars and this is true if UNICODE is defined or not! Note that "string" is just an MBCS string and L"string" is a Unicode string, also this does not depend on the compiler settings for UNICODE.
Nearly all functions exists in a Unicode and a MBCS version! You have to take care about this. If you don't take care you get thousands of errors.
Note also that nearly all Windows API functions exist in two versions a MBCS and a Unicode version (SetWindowTextA and SetWindowTextW).

So every string constant not defined with _T used in a function that maps to Unicode (like SetWindowText) will cause an error!

Just search the MSDN for tchar, you find this link:
http://msdn.microsoft.com/library/default.asp url=/library/en-us/vccore/html/_core_generic.2d.text_mappings_in_tchar..h.asp
KBaghaei
NO! I am using a lot of modules either in MBCS or Unicode, but I am using the tchar notation and I have no problems at all!
Michael Tsai
Yes, I understand what your saying what am saying is this.

There is one file called martin.cpp

and two projects, project 1 uses unicode and project 2 uses MBCS. But both projects use the file martin.cpp. This would cause the problem because to get the file to compile under project 1 it would have to use the unicode versions of the API but to get it to compile under project 2 it would have to use the MBCS versions of the API. So u see thats why i get the error, becuase the file martin.cpp can be one not the other. I would have to have one file for MBCS and one for unicode. Under these circumstances of sharing the file i would get the error correct

Will
Mr. Jerzy
So this is the reason for my thousands of errors, that i am using unicode versions of methods in the unicode project that is referenced from an MBCS project

correct

Will
Kerberos Mansour
NO! If I use strictly the _w functions and the Unicode function in the API and L"string" for all constants, than you would get again NO errors!
And again if I use just the MBCS versions and the correct string constants I would get no errors neither in an MBCS project nor in an Unicode project!

You get problems because you never cared about the different MBCS and Unicode versions of the API!
Emphyrio
Ok, thanks, just so i can check, this is a ligitimate reason why i am getting loads of errors, for example things like "_wtof" wouldnt compile under MBCS, and now do compile under Unicode. The problem was the one project that was changed to using Unicode, was being referenced in another project (or rather they were sharing the same files). but becuase the other project was still in MBCS, i got thousands of errors come up. This is likely to be the reason correct

Incidently is there a really good article you know of on this I can find some, but nothing great

Will
larkahn
Thats what I just said, i'm saying the reason why it failed is because the two are mixed, its not a solution its the reason why the error is there in first place. I agree
MickT
ok, but if you werent using the tchar notation, thats why you would get all these errors
Rui Silva18093
NO!
If martin.cpp just contains this code and function:
This function will work always:

void FooA()
{
   MessageBoxA(NULL, "MyFoo", "MyFoo", MB_OK);
}

The next too:

void FooW()
{
   MessageBoxW(NULL, L"MyFoo", L"MyFoo", MB_OK);
}

And this also

void FooT()
{
   MessageBox(NULL, _T("MyFoo"), _T("MyFoo"), MB_OK);
}

BUT THIS WILL FAIL IN A UNICODE COMPILE

void FooFail()
{
   MessageBox(NULL, "MyFoo", "MyFoo", MB_OK);
}

Because MessageBox will map to MessageBoxW!

So what can you do:
1. So use either always the Unicode versions
2. Use always the MBCS versions
3. Use the tchar versions.

Never mix!
vtcoder
Use ether:
text.Format (L"Row %d,Col %d",ln,col);

or
text.Format (_T("Row %d,Col %d"),ln,col);

I would prefer the later one, because it allows a project to be dynamically changed from MBCS to UNICODE and vice versa!
CString::Format issue

Answer this question