VS2005 : Managed Wrapper for Unmanaged code : Ellipsis dilema

  • Thread starter Thread starter Olivier Matrot
  • Start date Start date
O

Olivier Matrot

Hello,
I'm in the process of writing a managed wrapper on legacy C++ code.
I have to deal with methods that are using the ellipsis.

For instance,the legacy method is :
void CLog::Print (LPCSTR szMask, ...)
{

CHAR szText[512];

va_list arglist;

va_start (arglist, szMask);

wvsprintfA (szText, szMask, arglist);

va_end (arglist);

}


The wrapper looks like this :
void Print(System::String ^ Text, array<Object^>^ args )

{

CStringA s((LPCTSTR)UnmanagedString(Text));

m_pLog->Print(s, args);

}

I'm calling the wrapper with this call :

log->Print("Hello, CallCount is [%d]", gcnew
array<Object^>(1){(System::Int32)1});

And the result is "Hello CallCount [3226976]" instead of "Hello CallCount
[1]"

Where is the problem ?

TIA.
 
Hello,
I'm in the process of writing a managed wrapper on legacy C++ code.
I have to deal with methods that are using the ellipsis.

For instance,the legacy method is :
void CLog::Print (LPCSTR szMask, ...)
{

CHAR szText[512];

va_list arglist;

va_start (arglist, szMask);

wvsprintfA (szText, szMask, arglist);

va_end (arglist);

}

The wrapper looks like this :
void Print(System::String ^ Text, array<Object^>^ args )

{

CStringA s((LPCTSTR)UnmanagedString(Text));

m_pLog->Print(s, args);
Click to expand...

Here, you are passing an array<Object^>^ as the 1st ellipsis argument,
while the native part (eg, wvsprintf) expects an int. It not a
surprise it doesn't work..

I do not see any simple solution. IMHO, you should anyway get way of
those ugly, type-unsafe- c-styleish, ellipsis. You could for example
design your wrapper class to work like a stream or a StringBuilder.

Arnaud
 
I agree with you, it is better to do the formating in managed C++.
However, I'm interrested in dealing with this for my personnal skills.
Olivier.

Hello,
I'm in the process of writing a managed wrapper on legacy C++ code.
I have to deal with methods that are using the ellipsis.

For instance,the legacy method is :
void CLog::Print (LPCSTR szMask, ...)
{

CHAR szText[512];

va_list arglist;

va_start (arglist, szMask);

wvsprintfA (szText, szMask, arglist);

va_end (arglist);

}

The wrapper looks like this :
void Print(System::String ^ Text, array<Object^>^ args )

{

CStringA s((LPCTSTR)UnmanagedString(Text));

m_pLog->Print(s, args);
Click to expand...

Here, you are passing an array<Object^>^ as the 1st ellipsis argument,
while the native part (eg, wvsprintf) expects an int. It not a
surprise it doesn't work..

I do not see any simple solution. IMHO, you should anyway get way of
those ugly, type-unsafe- c-styleish, ellipsis. You could for example
design your wrapper class to work like a stream or a StringBuilder.

Arnaud
Click to expand...
 
I've changed the parameter in my code below with your suggestion. It
becomes:
void Print(System::String ^ Text, ... array<Object^>^ args )
The result is the same.

David Anton said:
Why not use a C++/CLI parameter array?
e.g.,
void test(... array<System::Object^> ^ParamArray)
--
http://www.tangiblesoftwaresolutions.com
C++ to C#
C++ to VB
C++ to Java
C++ to C++/CLI
Instant C#: VB to C# converter
Instant VB: C# to VB converter
Instant C++: convert VB or C# to C++/CLI


Olivier Matrot said:
Hello,
I'm in the process of writing a managed wrapper on legacy C++ code.
I have to deal with methods that are using the ellipsis.

For instance,the legacy method is :
void CLog::Print (LPCSTR szMask, ...)
{

CHAR szText[512];

va_list arglist;

va_start (arglist, szMask);

wvsprintfA (szText, szMask, arglist);

va_end (arglist);

}


The wrapper looks like this :
void Print(System::String ^ Text, array<Object^>^ args )

{

CStringA s((LPCTSTR)UnmanagedString(Text));

m_pLog->Print(s, args);

}

I'm calling the wrapper with this call :

log->Print("Hello, CallCount is [%d]", gcnew
array<Object^>(1){(System::Int32)1});

And the result is "Hello CallCount [3226976]" instead of "Hello
CallCount
[1]"

Where is the problem ?

TIA.
Click to expand...
Click to expand...
 
Back
Top