/// <summary>
/// This function will be called when & is found, and it will
/// peek forward to check if its entity, should there be a success
/// indicated by non-zero returned, the pointer will be left at the new byte
/// after entity
/// </summary>
/// <returns>Char (not byte) that corresponds to the entity or 0 if it was not entity</returns>
internal char CheckForEntity(byte[] bHTML, ref int iCurPos, int iDataLength)
{
if (!bDecodeEntities && !bMiniEntities)
{
return((char)0);
}
int iChars = 0;
byte cChar;
//string sEntity="";
// if true it means we are getting hex or decimal value of the byte
bool bCharCode = false;
bool bCharCodeHex = false;
int iEntLen = 0;
int iFrom = iCurPos;
string sEntity;
try
{
/*
* while(!Eof())
* {
* cChar=NextChar();
*/
while (iCurPos < iDataLength)
{
cChar = bHTML[iCurPos++];
// 21/10/05: not necessary
//if(cChar==0)
// break;
if (++iChars <= 2)
{
// the first byte for numbers should be #
if (iChars == 1)
{
if (cChar == '#')
{
iFrom++;
bCharCode = true;
continue;
}
}
else
{
if (bCharCode && cChar == 'x')
{
iFrom++;
iEntLen--;
bCharCodeHex = true;
}
}
}
//Console.WriteLine("Got entity end: {0}",sEntity);
// Break on:
// 1) ; - proper end of entity
// 2) number 10-based entity but current byte is not a number
//if(cChar==';' || (bCharCode && !bCharCodeHex && !char.IsNumber((char)cChar)))
// TODO: browsers appear to be lax about ; requirement for end of entity
// we should really do the same and treat whitespace as termination of entity
if (cChar == ';' || (bCharCode && !bCharCodeHex && !(cChar >= '0' && cChar <= '9')))
{
// lets try speculative quick lookup using just first 2 chars
// this should be successful in almost all cases thus removing need for
// expensive creation of a string
if (!bCharCode && iEntLen > 1)
{
object oChar = oEntities.GetLikelyPresentValue(bHTML[iFrom], bHTML[iFrom + 1]);
if (oChar != null)
{
return((char)((int)oChar));
}
}
// check if its int - this way we can avoid having to create string
if (bCharCode && iEntLen > 0 && !bCharCodeHex)
{
// if mini entities mode is set then we will ignore all numerics
if (bMiniEntities)
{
break;
}
// we have to backdown one char in case when entity did not end with ;
// otherwise we will lose next char in the stream, this correction suggested by Kurt Carlson!
if (cChar != ';')
{
iCurPos--;
}
return((char)ParseUInt(bHTML, iFrom, iEntLen));
}
sEntity = Encoding.Default.GetString(bHTML, iFrom, iEntLen);
if (bCharCode)
{
// NOTE: this may fail due to wrong data format,
// in which case we will return 0, and entity will be
// ignored
if (iEntLen > 0)
{
// if mini entities mode is set then we will ignore all numerics
if (bMiniEntities)
{
break;
}
int iChar;
if (!bCharCodeHex)
{
#if DOTNET20
// we want to avoid exceptions if possible as they are slow
if (!int.TryParse(sEntity, out iChar))
{
if (iChars > 0)
{
if ((iCurPos - iChars) >= 0)
{
iCurPos -= iChars;
}
//PutChars(iChars);
}
return((char)(0));
}
#else
iChar = int.Parse(sEntity);
#endif
}
else
{
#if DOTNET20
// we want to avoid exceptions if possible as they are very slow
if (!int.TryParse(sEntity, System.Globalization.NumberStyles.HexNumber, null, out iChar))
{
if (iChars > 0)
{
if ((iCurPos - iChars) >= 0)
{
iCurPos -= iChars;
}
//PutChars(iChars);
}
return((char)(0));
}
#else
iChar = int.Parse(sEntity, NumberStyles.HexNumber);
#endif
}
return((char)iChar);
}
}
if (iEntLen >= iMinEntityLen && iEntLen <= iMaxEntityLen)
{
object oChar = oEntities.GetLikelyPresentValue(sEntity);
if (oChar != null)
{
return((char)((int)oChar));
}
}
}
//break;
// as soon as entity length exceed max length of entity known to us
// we break up the loop and return nothing found
// NOTE: removed due to entities being generally correct and this code costs 10% of CPU in this function
if (iEntLen > iMaxEntityLen)
{
break;
}
iEntLen++;
}
}
catch //(Exception oEx)
{
//Console.WriteLine("Entity parsing exception: "+oEx.ToString());
}
// if we have not found squat, then we will need to put point back
// to where it was before this function was called
if (iChars > 0)
{
if ((iCurPos - iChars) >= 0)
{
iCurPos -= iChars;
}
//PutChars(iChars);
}
return((char)(0));
}
/// <summary>
/// This function will decode any entities found in a string - not fast!
/// </summary>
/// <returns>Possibly decoded string</returns>
internal static string DecodeEntities(string sData)
{
char cChar;
StringBuilder oSB = new StringBuilder(sData.Length);
string sEntity = "";
try
{
for (int i = 0; i < sData.Length; i++)
{
cChar = sData[i];
if (cChar != '&' || (i + 1 >= sData.Length))
{
oSB.Append(cChar);
}
else
{
// if true it means we are getting hex or decimal value of the byte
bool bCharCode = false;
bool bCharCodeHex = false;
int iEntLen = 0;
int iChars = 0;
int j = i + 1;
int iFrom = i + 1;
for (; j < sData.Length; j++)
{
cChar = sData[j];
if (++iChars <= 2)
{
// the first byte for numbers should be #
if (iChars == 1)
{
if (cChar == '#')
{
iFrom++;
bCharCode = true;
continue;
}
}
else
{
if (bCharCode && cChar == 'x' && !bCharCodeHex)
{
iFrom++;
//iEntLen--;
bCharCodeHex = true;
continue;
}
}
}
//Console.WriteLine("Got entity end: {0}",sEntity);
// Break on:
// 1) ; - proper end of entity
// 2) number 10-based entity but current byte is not a number
//if(cChar==';' || (bCharCode && !bCharCodeHex && !char.IsNumber((char)cChar)))
bool bLastChar = j + 1 >= sData.Length;
if (cChar == ';' || (bCharCode && !bCharCodeHex && !(cChar >= '0' && cChar <= '9')) || (bCharCode && bLastChar))
{
// end of string
if (bLastChar && cChar != ';')
{
iEntLen++;
}
// lets try speculative quick lookup using just first 2 chars
// this should be successful in almost all cases thus removing need for
// expensive creation of a string
if (!bCharCode && iEntLen > 1)
{
// make sure we aint at the end of string
if (i + 2 < sData.Length)
{
object oChar = oAllEntities.GetLikelyPresentValue((byte)sData[i + 1], (byte)sData[i + 2]);
if (oChar != null)
{
oSB.Append((char)((int)oChar));
break;
}
}
}
// check if its int - this way we can avoid having to create string
if (bCharCode && iEntLen > 0 && !bCharCodeHex)
{
sEntity = sData.Substring(iFrom, iEntLen);
int iChar = 0;
bool bSuccess = false;
try
{
iChar = (int)uint.Parse(sEntity);
bSuccess = true;
}
catch
{
}
if (bSuccess)
{
oSB.Append((char)iChar);
// move back once when we got number done without ; at the end
// of it - Firefox and IE do it this way
if (cChar != ';' && !bLastChar)
{
j--;
}
break;
}
else
{
// this will force to add entity as is - probably broken
// or maybe not entity at all
oSB.Append('&');
j = i;
break;
}
}
sEntity = sData.Substring(iFrom, iEntLen);
if (bCharCode)
{
// NOTE: this may fail due to wrong data format,
// in which case we will return 0, and entity will be
// ignored
if (iEntLen > 0)
{
int iChar = 0;
bool bSuccess = false;
#if DOTNET20 && false
if (!bCharCodeHex)
{
bSuccess = int.TryParse(sEntity, out iChar);
}
else
{
bSuccess = int.TryParse(sEntity, System.Globalization.NumberStyles.HexNumber, out iChar);
}
#else
try
{
if (!bCharCodeHex)
{
iChar = int.Parse(sEntity);
}
else
{
iChar = int.Parse(sEntity, NumberStyles.HexNumber);
}
bSuccess = true;
}
catch
{
// some numbers might not be parsed correctly so we will ignore them
}
#endif
if (bSuccess)
{
oSB.Append((char)iChar);
break;
}
else
{
// this will force to add entity as is - probably broken
// or maybe not entity at all
iEntLen = iAllMaxEntityLen + 1;
}
}
}
if (iEntLen >= iAllMinEntityLen && iEntLen <= iAllMaxEntityLen)
{
object oChar = oAllEntities.GetLikelyPresentValue(sEntity);
if (oChar != null)
{
oSB.Append((char)((int)oChar));
break;
}
else
{
// this will force to add entity as is - probably broken
// or maybe not entity at all
iEntLen = iAllMaxEntityLen + 1;
//Utils.Write("");
}
}
}
//break;
// as soon as entity length exceed max length of entity known to us
// we break up the loop and return nothing found
// NOTE: removed due to entities being generally correct and this code costs 10% of CPU in this function
if (iEntLen > iAllMaxEntityLen || bLastChar)
{
// append char that triggered entity thingy in the first place
oSB.Append('&');
j = i;
break;
}
iEntLen++;
}
i = j;
}
}
}
catch (Exception oEx)
{
Console.WriteLine("Entity parsing exception: " + oEx.ToString());
return(sData);
}
return(oSB.ToString());
}
