 |
|
|
why does URLEncode1 have
if (!isalnum(*pInTmp) && !isalnum(*pInTmp)) ?
is this an error, something supposed to be a double check, in case program is run on computer without ECC ?
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
<pre><code>
CString URLEncode(CString sIn)
{
CString sOut;
const int nLen = sIn.GetLength() + 1;
register LPTSTR pOutTmp = NULL;
LPTSTR pOutBuf = NULL;
register LPTSTR pInTmp = NULL;
LPTSTR pInBuf =(LPTSTR)sIn.GetBuffer(nLen);
//BYTE b = 0;
//alloc out buffer
pOutBuf = (LPTSTR)sOut.GetBuffer(nLen * 3 - 2);//new BYTE [nLen * 3];
if(pOutBuf)
{
pInTmp = pInBuf;
pOutTmp = pOutBuf;
// do encoding
while (*pInTmp)
{
if(isalnum(*pInTmp))
*pOutTmp++ = *pInTmp;
else
if(isspace(*pInTmp) && ((*pInTmp!='\n') && (*pInTmp!='\r')))
*pOutTmp++ = '+';
else
{
*pOutTmp++ = '%';
*pOutTmp++ = toHex(*pInTmp>>4);
*pOutTmp++ = toHex(*pInTmp%16);
}
pInTmp++;
}
*pOutTmp = '\0';
//sOut=pOutBuf;
//delete [] pOutBuf;
sOut.ReleaseBuffer();
}
sIn.ReleaseBuffer();
return sOut;
}
</code></pre>
|
| Sign In·View Thread·PermaLink | 2.00/5 |
|
|
|
|
|
|
CString URLEncode2(CString sIn)
{
CString sOut;
const int nLen = sIn.GetLength();
// do encoding
for (int i = 0; i TCHAR c = sIn.GetAt(i);
if(_istalnum(c))
sOut.AppendChar(c);
else
if(_istspace(c))
sOut.AppendChar(_T('+'));
else
{
int ic = _TINT(c);
sOut.AppendChar(_T('%'));
sOut.AppendChar(_TCHAR(toHex((BYTE)(ic>>4))));
sOut.AppendChar(_TCHAR(toHex((BYTE)(ic%16))));
}
}
sOut.AppendChar(_T('\0'));
return sOut;
}
I think that it works.
Tell me if you have problems.
|
| Sign In·View Thread·PermaLink | 2.00/5 |
|
|
|
|
|
|
I had a bug in the last post.
inline TCHAR toHex(const BYTE &x)
{
return x > 9 ? _T('A') + x-10: _T('0') + x;
}
CString URLEncode2(CString sIn)
{
CString sOut;
const int nLen = sIn.GetLength();
// do encoding
for (int i = 0; i TCHAR c = sIn.GetAt(i);
if(_istalnum(c))
sOut.AppendChar(c);
else
if(_istspace(c))
sOut.AppendChar(_T('+'));
else
{
int ic = _TINT(c);
sOut.AppendChar(_T('%'));
sOut.AppendChar(toHex((BYTE)(ic>>4)));
sOut.AppendChar(toHex((BYTE)(ic%16)));
}
}
sOut.AppendChar(_T('\0'));
return sOut;
}
|
| Sign In·View Thread·PermaLink | 2.00/5 |
|
|
|
|
|
|
|
|
|
great code, thanks 
only thing I was missing is support for linefeeds (CR LF)
As a workaround, I modified:
// do encoding
while (*pInTmp)
{
if(isalnum(*pInTmp))
*pOutTmp++ = *pInTmp;
else
if( isspace(*pInTmp) &&
((*pInTmp!='\n') && // new
(*pInTmp!='\r')) ) // new
*pOutTmp++ = '+';
else
{
*pOutTmp++ = '%';
*pOutTmp++ = toHex(*pInTmp>>4);
*pOutTmp++ = toHex(*pInTmp%16);
}
pInTmp++;
}
that way, both '\r' and '\n' (CR LF) will be converted
properly to %0D%0A.
apart from that, thanks again.
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
in function urlencode1:
if (!isalnum(*pInTmp) && !isalnum(*pInTmp))
it should be:
if (!isalnum(*pInTmp) && !isspace(*pInTmp))
yes ?
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
|
|
|
// guess rafaelcn already found answer, but
// I use something like this
// helper to convert %XX to chars
inline char x2c ( unsigned char* in)
{
register char digit;
digit = ( (*in) >= 'A' ? (((*in) & 0xdf) - 'A')+10 : ((*in) - '0'));
digit *= 16;
in++;
digit += ((*in) >= 'A' ? (((*in) & 0xdf) - 'A')+10 : ((*in) - '0'));
return(digit);
}
//do the work
std::string URLdecode( const std::string& in )
{
int len = in.size();
unsigned char* inBuff = (unsigned char*)(in.c_str());
unsigned char* out = new unsigned char[len+1];
unsigned char* outBuff = out;
while( *inBuff )
{
if ( *inBuff == '%' )
{
inBuff++;
*outBuff = x2c(inBuff);
inBuff++;
}
else
{
*outBuff=*inBuff;
}
outBuff++;
inBuff++;
}
*outBuff =0;
std::string ret ((char*)out);
delete[] out;
return ret;
}
// optimize away use of std::string or
// convert to CString/whatever as needed,
// but:
The real question now is where you have your
URLencoded string from. Could it be encoded in
UTF-8, as URL string nowadays tend to, e.g. if send from browser to server?
Hope your App is UTF-8 aware 
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
AS one of learner,I want to add a dialog to verify whether a user is legal before Windows98's Explore working ,which is like NT's verification dialog.
But I don't know how to do,anyone can help me? Thanks! 
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
|
|
|
I use presented functions in WinSock application. InternetCanonicalizeUrl function is from wininet.dll. I encode XML data and I need fast functions. My function is faster than the standard InternetCanonicalizeUrl.
|
| Sign In·View Thread·PermaLink | 2.00/5 |
|
|
|
|
|
|
|
|
|
I had problems with InternetCanonicalizeUrl if the XML file was massive. Maybe the function is only expecting URLs, not form data?
This URLEncode function works great, however
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
The line:
pOutBuf = new BYTE [nLen + 3 * k];
Should be:
pOutBuf = new BYTE [nLen + 2 * k];
since we are already allocating one byte for the "original" byte and thus only need two additional bytes for the characters needing encoding.
Also, as someone else has noted, you might as well use .GetBuffer on the output string since that avoids a redundant allocation and memcpy.
Marc
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
Hi,
why don't you use
outBuf = (LPBYTE)sOut.GetBuffer(nLen*3)
...
sOut.ReleaseBuffer(reallen);
instead of
outBuf = new BYTE[nLen*3];
...
sOut = outBuf;
delete []outBuf;
this would save memory and a memory copy which is realy slow.
Ciao Matthias
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
Hi,
why don't you use
outBuf = (LPBYTE)sOut.GetBuffer(nLen*3)
...
sOut.ReleaseBuffer(reallen);
instead of
outBuf = new BYTE[nLen*3];
...
sOut = outBuf;
delete []outBuf;
this would save memory and a memory copy which is realy slow.
Ciao Matthias
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
|
|
Heya,
when doing fast conversions like this I usually uses the _alloca() routine. It allocates memory on the stack instead of the heap, which usually makes it faster. And also when the function runs out of scope it automaticallu frees the stack and so the allocated item. 
Sample:
pOutBuf = (LPBYTE)_alloca(nLen * 3); //automatically frees when running out of scope
(It is suitable for smaller allocations, since the stack is limited... )
Cheers
/WW
|
| Sign In·View Thread·PermaLink | |
|
|
|
|
