软件教程

去除HTML标记的三个实用函数

来源: 本站    类别: 软件教程    日期: 2013/11/1

 

<%

Function RemoveHTML_A(strText)
    Dim nPos1
    Dim nPos2
   
    nPos1 = InStr(strText, "<")
    Do While nPos1>0
        nPos2 = InStr(nPos1+1, strText, ">")
        If nPos2>0 Then
            strText = Left(strText, nPos1 - 1) & Mid(strText, nPos2 + 1)
        Else
            Exit Do
        End If
        nPos1 = InStr(strText, "<")
    Loop
   
    RemoveHTML_A = strText
End Function

Function RemoveHTML_B(strText)
 Dim RegEx

 Set RegEx = New RegExp

 RegEx.Pattern = "<[^>]*>"
 RegEx.Global = True

 RemoveHTML_B = RegEx.Replace(strText, "")
End Function

Function RemoveHTML_C( strText )
    Dim TAGLIST
    TAGLIST = ";!--;!DOCTYPE;A;ACRONYM;ADDRESS;APPLET;AREA;B;BASE;BASEFONT;" &_
              "BGSOUND;BIG;BLOCKQUOTE;BODY;BR;BUTTON;CAPTION;CENTER;CITE;CODE;" &_
              "COL;COLGROUP;COMMENT;DD;DEL;DFN;DIR;DIV;DL;DT;EM;EMBED;FIELDSET;" &_
              "FONT;FORM;FRAME;FRAMESET;HEAD;H1;H2;H3;H4;H5;H6;HR;HTML;I;IFRAME;IMG;" &_
              "INPUT;INS;ISINDEX;KBD;LABEL;LAYER;LAGEND;LI;LINK;LISTING;MAP;MARQUEE;" &_
              "MENU;META;NOBR;NOFRAMES;NOSCRIPT;OBJECT;OL;OPTION;P;PARAM;PLAINTEXT;" &_
              "PRE;Q;S;SAMP;SCRIPT;SELECT;SMALL;SPAN;STRIKE;STRONG;STYLE;SUB;SUP;" &_
              "TABLE;TBODY;TD;TEXTAREA;TFOOT;TH;THEAD;TITLE;TR;TT;U;UL;VAR;WBR;XMP;"

    Const BLOCKTAGLIST = ";APPLET;EMBED;FRAMESET;HEAD;NOFRAMES;NOSCRIPT;OBJECT;SCRIPT;STYLE;"
   
    Dim nPos1
    Dim nPos2
    Dim nPos3
    Dim strResult
    Dim strTagName
    Dim bRemove
    Dim bSearchForBlock
   
    nPos1 = InStr(strText, "<")
    Do While nPos1 > 0
        nPos2 = InStr(nPos1 + 1, strText, ">")
        If nPos2 > 0 Then
            strTagName = Mid(strText, nPos1 + 1, nPos2 - nPos1 - 1)
     strTagName = Replace(Replace(strTagName, vbCr, " "), vbLf, " ")

            nPos3 = InStr(strTagName, " ")
            If nPos3 > 0 Then
                strTagName = Left(strTagName, nPos3 - 1)
            End If
           
            If Left(strTagName, 1) = "/" Then
                strTagName = Mid(strTagName, 2)
                bSearchForBlock = False
            Else
                bSearchForBlock = True
            End If
           
            If InStr(1, TAGLIST, ";" & strTagName & ";", vbTextCompare) > 0 Then
                bRemove = True
                If bSearchForBlock Then
                    If InStr(1, BLOCKTAGLIST, ";" & strTagName & ";", vbTextCompare) > 0 Then
                        nPos2 = Len(strText)
                        nPos3 = InStr(nPos1 + 1, strText, "</" & strTagName, vbTextCompare)
                        If nPos3 > 0 Then
                            nPos3 = InStr(nPos3 + 1, strText, ">")
                        End If
                       
                        If nPos3 > 0 Then
                            nPos2 = nPos3
                        End If
                    End If
                End If
            Else
                bRemove = False
            End If
           
            If bRemove Then
                strResult = strResult & Left(strText, nPos1 - 1)
                strText = Mid(strText, nPos2 + 1)
            Else
                strResult = strResult & Left(strText, nPos1)
                strText = Mid(strText, nPos1 + 1)
            End If
        Else
            strResult = strResult & strText
            strText = ""
        End If
       
        nPos1 = InStr(strText, "<")
    Loop
    strResult = strResult & strText
   
    RemoveHTML_C = strResult
End Function

%>


相关文章


Copyright © 2004 - 2024 CNHWW Inc. All Rights Reserved
石家庄市征红网络科技有限公司版权所有 邮政编码:050051
服务电话:0311-85315152 13931185013 在线客服QQ:81447932 / 81447933 邮箱: cnhww@163.com