Has NthArgument changed?

[vefatica]

Code like this, at the beginning of a plugin function

INT WINAPI PLUGGO ( WCHAR *psz )
{
   Printf(L"psz: %s\r\n", psz);
   WCHAR *pArgs = NthArgument(psz, 0, NULL, NULL);
   Printf(L"Args: %s\r\n", pArgs);

suggests that NthArgument behaves differently in v17 from how it behaved in v16. Calling this function with args

e f

gives me

psz:  e f
Args: e f

in v16, and

psz:  e f
Args: e

in v17.

 

[rconn]

Since your example is meaningless, I presume your real issue is something a bit different.

Yes, NthArgument changed in v17. It was removed, and replaced with a renamed function that processes string classes. A stub wrapper named NthArgument was kept, but it doesn't support the "..., NULL, NULL)" format. (Was that ever even documented?) If you are trying to get a pointer to the remainder of the string following argument n (not argument 0 as in your example), use the "nthptr" fourth argument.

 

[vefatica]

Since your example is meaningless, I presume your real issue is something a bit different.

Yes, NthArgument changed in v17. It was removed, and replaced with a renamed function that processes string classes. A stub wrapper named NthArgument was kept, but it doesn't support the "..., NULL, NULL)" format. (Was that ever even documented?) If you are trying to get a pointer to the remainder of the string following argument n (not argument 0 as in your example), use the "nthptr" fourth argument.

I've used that syntax as a quick way to bypass leading whitespace in the function's string argument. How will using ppNthPtr help?
It's no big deal; easy enough to bypass leading whitespace by other means. The hard part might be finding all the places I relied on that behavior.

 

[rconn]

That's a truly horrible perversion of that function. It's far simpler (and about 1000x faster) to skip white space with the SkipWhiteSpace function.

 

[vefatica]

That's a truly horrible perversion of that function. It's far simpler (and about 1000x faster) to skip white space with the SkipWhiteSpace function.

Yeah, well ... that goes way back. I even have my own StripWhiteSpace function. In the case in point, I can even simply increment psz until it' non-white.

 

[vefatica]

Where's SkipWhiteSpace? It's not in my TakeCmd.h (but I haven't updated my SDK).

 

[rconn]

// skip past leading white space and return pointer to first non-space char
DLLExports LPCTSTR SkipWhiteSpace( LPCTSTR pszLine )
{
while (( *pszLine == _TEXT(' ') ) || ( *pszLine == _TEXT('\t') ))
pszLine++;
return pszLine;
}

 

[vefatica]

// skip past leading white space and return pointer to first non-space char
DLLExports LPCTSTR SkipWhiteSpace( LPCTSTR pszLine )
{
while (( *pszLine == _TEXT(' ') ) || ( *pszLine == _TEXT('\t') ))
pszLine++;
return pszLine;
}

I did just that inline. For as long as I can, I'll stay away from exports that differ in various TCC versions (LPWSTR vs. LPCWSTR).