strlen, wcslen and UNICODE
Gerd
I write some source code that should work in both cases - when
compiled with UNICODE enabled and disabled. So I define the strings
using TCHAR. But when I want to work with the strings (calculate the
length, or concatenate two strings), how can I do it independently? To
determine the length there are "strlen" for ascii-strings and "wcslen"
for UNICODE. Is there a way to write on call for both or would I have
to do it like this:
#ifdef _UNICODE
strlen (...);
#else
wcslen (...);
#endif
Many thanks for help.
Gerd
Ulrich Eckhardt
1. The easiest way out is to use _tcslen(), which itself is a macro that
resolves to strlen() or wcslen(), depending on _UNICODE.
2. You could save yourself some hassle if you simply used a TCHAR-based
string class like std::basic_string<TCHAR> or MFC's/ATL's CString. This
requires C++ though, I don't know if you're using that.
3. Just don't. The interesting feature of TCHAR is that you can support a
codebase that compiles in two different ways and thus supports win95 and
win98. Now, first thing here: do you really want to support ten-year old
desktop OSs? Secondly, if you really want to, you could look at the
Unicode-libraries for win9x. Those are helpers to get _UNICODE-compiled
programs to run to a sufficient extent on win9x.
In general, ask yourself what kind of string type is the most appropriate?
If it is wchar_t, use this, if it is char, use that. If you want to support
Unicode, either use wchar_t (95%) or some specialised library like ICU
(99%). For mere logging, use char/ASCII.
= just my 2cc from maintaining a TCHAR program for quite some time =
Uli
--
Sator Laser GmbH
Geschäftsführer: Ronald Boers, Amtsgericht Hamburg HR B62 932
Gerd
Ulrich Eckhardt schrieb:
> 1. The easiest way out is to use _tcslen(), which itself is a macro that
> resolves to strlen() or wcslen(), depending on _UNICODE.
Thanks, I will do it.
> 2. You could save yourself some hassle if you simply used a TCHAR-based
> string class like std::basic_string<TCHAR> or MFC's/ATL's CString. This
> requires C++ though, I don't know if you're using that.
The customer requires a 100% pure C application, so this is not
possible.
> 3. Just don't. The interesting feature of TCHAR is that you can support a
> codebase that compiles in two different ways and thus supports win95 and
> win98. Now, first thing here: do you really want to support ten-year old
> desktop OSs? Secondly, if you really want to, you could look at the
> Unicode-libraries for win9x. Those are helpers to get _UNICODE-compiled
> programs to run to a sufficient extent on win9x.
The app must run W2k and more recent versions (due to hardwaredriver ,
which is not available for older systems)
> In general, ask yourself what kind of string type is the most appropriate?
> If it is wchar_t, use this, if it is char, use that. If you want to support
> Unicode, either use wchar_t (95%) or some specialised library like ICU
> (99%). For mere logging, use char/ASCII.
The app must support greek, swedish, french, german, so UNICODE would
be best.
> = just my 2cc from maintaining a TCHAR program for quite some time =
>
> Uli
>
> --
> Sator Laser GmbH
> Geschäftsführer: Ronald Boers, Amtsgericht Hamburg HR B62 932
Thanks,
Gerd
Klueless
> determine the length there are "strlen" for ascii-strings and "wcslen"
_tcslen and you can find more such functions here:
<http://msdn2.microsoft.com/en-us/library/tsbaswba(VS.80).aspx>