• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 326
  • Last Modified:

Win32 Unicode Compatibility for Win95/98

I'm planning on migrating my Win32 app to Unicode, but need to maintain compatibility with Win95/98 in a single executable.  Although the C Runtime Library Unicode functions (wcslen, wprintf etc.) are all supported in both NT and Win95/98, the Win32 library functions are not - the Windows 95/98 implementation of these functions only accept ANSI characters and no Unicode.

Question is:  Is there an easy way to implement this Win95/98 and NT Unicode support in a single executable?  The only way I can think of is to create my own "wrapper" functions for all Win32 calls, and to conditionally convert the arguments/returns to/from Unicode based on whether I'm running Win95/98.  Has anyone done this and wants to share their code?  Is there a pre-written library that I can buy or download that provides this conversion?

Surely I'm not the first to come across this huge limitation in Win95 and Win98!  I'm happy to give more points if you have a good solution!
0
tphipps
Asked:
tphipps
1 Solution
 
xyuCommented:
Use MultiByteToWideChar & WideCharToMultiByte fuctions...

or You can go to http://www.geocities.com/SiliconValley/1741/
download Natlib template library... and look at nlstring.h/nlstring.cpp for SzToWide, SzToStd, etc. functions to convert between UNICODE/ASCII ....
0
 
tphippsAuthor Commented:
Good try xyu, but not what I'm looking for.  Converting the text from native to Unicode and vice-versa is not the problem.  The problem is the fact that I'd need to write a routine to do this for every one of the hundreds of Win32 function calls in order to get Win95 support.

Surely there's a better way...
0
 
ArkadiyCommented:
Can you please be more specific? What functions are missing? I use Unicode for user-readable output, and TextOutW is indeed supported in 95 as well as in NT.

In case you _really_ need a solution, what about your own string class? it can return different string depending on the context:

class CMyString
{
...
   operator const char*(){return Ansi string}
   operator const WCHAR *(){return Unicode string}
...
}

You can even derive it from MFC strings, if you wish
0
Cloud Class® Course: Certified Penetration Testing

This CPTE Certified Penetration Testing Engineer course covers everything you need to know about becoming a Certified Penetration Testing Engineer. Career Path: Professional roles include Ethical Hackers, Security Consultants, System Administrators, and Chief Security Officers.

 
MirkwoodCommented:
Whether you like it or not. Win95 & 98 do not support UNICODE.
Altough most functions are defined in win95/win98, they are just not implemented and return OK.
Officially win95/win98 were meant for users at home and not in office environments.
Win95/win98 does however support DBCS. Double byte Character system. That means that some character may be displayable on win95/win98.

Our approach is the following and it is more or less what you described. Our middle and bottom layers are completly unicode. We don't use the operating system to do string manipulation since that is not implemented on win95/98.
As soon as we go to the operating system we use defines. In WINNT the define just points to the OS directly. In win95, a conversion is made to the local DBCS system used.
BTW: we do the same with the C++ library functions to allow the software to be compiled with different C compilers. E.g. strncpy puts a 0 byte after the string on some compilers and on other compilers it does not.
0
 
tphippsAuthor Commented:
Thanks Mirkwood - looks like there's no easy solution to this - because you're talking about using #defines, your implementation sounds as if you need to have 2 executable sets - one for NT and another for Win95.

I hear MSFT is using a post-link tool to modify the output executable directly to support Win95 and Unicode.
0
 
MirkwoodCommented:
Thanks for the points.
Yes you are right. We have seperate installation for Win95 and WinNT. We did that since testing the OS on every function would decrease the performance. I now it's only one CMP and one JNE assembly statement but still. Some interfaces are also totally different (shell extensions for example) and on NT we log to the event log and handle security different.


0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

Join & Write a Comment

Featured Post

Keep up with what's happening at Experts Exchange!

Sign up to receive Decoded, a new monthly digest with product updates, feature release info, continuing education opportunities, and more.

Tackle projects and never again get stuck behind a technical roadblock.
Join Now