private void Dispose(bool bDisposing)
{
if (!bDisposed)
{
bDisposed = true;
if (oChunk != null)
{
oChunk.Dispose();
oChunk = null;
}
if (sText != null)
{
sText.Dispose();
sText = null;
}
bHTML = null;
if (oE != null)
{
oE.Dispose();
oE = null;
}
if (oTP != null)
{
oTP.Dispose();
oTP = null;
}
}
}
/// <summary>
/// Internally parses tag and returns it from point when open bracket (<) was found
/// </summary>
/// <returns>Chunk</returns>
HTMLchunk GetNextTag()
{
//iCurPos++;
oChunk = oTP.ParseTag(ref iCurPos);
// for backwards compatibility mark closed tags with params as open
if (oChunk.iParams > 0 && bAutoMarkClosedTagsWithParamsAsOpen && oChunk.oType == HTMLchunkType.CloseTag)
{
oChunk.oType = HTMLchunkType.OpenTag;
}
// 012345
// check for start of script
if (oChunk.sTag.Length == 6 && oChunk.sTag[0] == 's' && oChunk.sTag == "script")
{
if (!oChunk.bClosure)
{
oChunk.oType = HTMLchunkType.Script;
oChunk = oTP.ParseScript(ref iCurPos);
return(oChunk);
}
}
oChunk.iChunkLength = iCurPos - oChunk.iChunkOffset;
if (bKeepRawHTML)
{
oChunk.oHTML = oEnc.GetString(bHTML, oChunk.iChunkOffset, oChunk.iChunkLength);
}
return(oChunk);
}
/// <summary>
/// Sets oHTML variable in a chunk to the raw HTML that was parsed for that chunk.
/// </summary>
/// <param name="oChunk">Chunk returned by ParseNext function, it must belong to the same HTMLparser that
/// was initiated with the same HTML data that this chunk belongs to</param>
public void SetRawHTML(HTMLchunk oChunk)
{
// note: this really should have been byte array assigned rather than string
// it would be more correct originality-wise
oChunk.oHTML = oEnc.GetString(bHTML, oChunk.iChunkOffset, oChunk.iChunkLength);
}
/// <summary>
/// Handles META tags that set page encoding
/// </summary>
/// <param name="oP">HTML parser object that is used for parsing</param>
/// <param name="oChunk">Parsed chunk that should contain tag META</param>
/// <param name="bEncodingSet">Your own flag that shows whether encoding was already set or not, if set
/// once then it should not be changed - this is the logic applied by major browsers</param>
/// <returns>True if this was META tag setting Encoding, false otherwise</returns>
public static bool HandleMetaEncoding(HTMLparser oP, HTMLchunk oChunk, ref bool bEncodingSet)
{
if (oChunk.sTag.Length != 4 || oChunk.sTag[0] != 'm' || oChunk.sTag != "meta")
{
return(false);
}
// if we do not use hashmode already then we call conversion explicitly
// this is slow, but METAs are very rare so performance penalty is low
if (!oChunk.bHashMode)
{
oChunk.ConvertParamsToHash();
}
string sKey = oChunk.oParams["http-equiv"] as string;
if (sKey != null)
{
// FIXIT: even though this is happening rare I really don't like lower casing stuff
// that most likely would not need to be - if you feel bored then rewrite this bit
// to make it faster, it is really easy...
switch (sKey.ToLower())
{
case "content-type":
// rare case (appears to work in IE) reported to exist in some pages by Martin B�chtold
case "content-category":
// we might have charset here that may hint at necessity to decode page
// check for possible encoding change
// once encoding is set it should not be changed, but you can be damn
// sure there are web pages out there that do that!!!
if (!bEncodingSet)
{
string sData = oChunk.oParams["content"] as string;
// it is possible we have broken META tag without Content part
if (sData != null)
{
if (oP.SetEncoding(sData))
{
// we may need to re-encode title
if (!bEncodingSet)
{
// here you need to reencode any text that you found so far
// most likely it will be just TITLE, the rest can be ignored anyway
bEncodingSet = true;
}
}
else
{
// failed to set encoding - most likely encoding string
// was incorrect or your machine lacks codepages or something
// else - might be good idea to put warning message here
}
}
}
return(true);
default:
break;
}
}
return(false);
}
