]> git.saurik.com Git - wxWidgets.git/blobdiff - src/html/htmltag.cpp
Remove DoSetSizeHints() call from Create()
[wxWidgets.git] / src / html / htmltag.cpp
index 02f045ce3c8aa8c2062932a490fd11ee89567342..574b01006e9151b64b092a2f7ba92c46b64fbf94 100644 (file)
@@ -1,35 +1,34 @@
 /////////////////////////////////////////////////////////////////////////////
-// Name:        htmltag.cpp
+// Name:        src/html/htmltag.cpp
 // Purpose:     wxHtmlTag class (represents single tag)
 // Author:      Vaclav Slavik
-// RCS-ID:      $Id$
 // Copyright:   (c) 1999 Vaclav Slavik
-// Licence:     wxWindows Licence
+// Licence:     wxWindows licence
 /////////////////////////////////////////////////////////////////////////////
 
+#include "wx/wxprec.h"
 
-#ifdef __GNUG__
-#pragma implementation
+#ifdef __BORLANDC__
+    #pragma hdrstop
 #endif
 
-#include "wx/wxprec.h"
-
-#include "wx/defs.h"
 #if wxUSE_HTML
 
-#ifdef __BORDLANDC__
-#pragma hdrstop
-#endif
+#include "wx/html/htmltag.h"
 
-#ifndef WXPRECOMP
-#include "wx/wx.h"
+#ifndef WX_PRECOMP
+    #include "wx/colour.h"
+    #include "wx/wxcrtvararg.h"
 #endif
 
-#include "wx/html/htmltag.h"
+#include "wx/html/htmlpars.h"
+#include "wx/html/styleparams.h"
+
+#include "wx/vector.h"
+
 #include <stdio.h> // for vsscanf
 #include <stdarg.h>
 
-
 //-----------------------------------------------------------------------------
 // wxHtmlTagsCache
 //-----------------------------------------------------------------------------
@@ -38,101 +37,255 @@ struct wxHtmlCacheItem
 {
     // this is "pos" value passed to wxHtmlTag's constructor.
     // it is position of '<' character of the tag
-    int Key;
+    wxString::const_iterator Key;
+
+    // Tag type
+    enum Type
+    {
+        Type_Normal, // normal tag with a matching ending tag
+        Type_NoMatchingEndingTag, // there's no ending tag for this tag
+        Type_EndingTag // this is ending tag </..>
+    };
+    Type type;
 
     // end positions for the tag:
     // end1 is '<' of ending tag,
     // end2 is '>' or both are
-    // -1 if there is no ending tag for this one...
-    // or -2 if this is ending tag  </...>
-    int End1, End2;
+    wxString::const_iterator End1, End2;
 
     // name of this tag
     wxChar *Name;
 };
 
+// NB: this is an empty class and not typedef because of forward declaration
+class wxHtmlTagsCacheData : public wxVector<wxHtmlCacheItem>
+{
+};
 
-IMPLEMENT_CLASS(wxHtmlTagsCache,wxObject)
+bool wxIsCDATAElement(const wxChar *tag)
+{
+    return (wxStrcmp(tag, wxT("SCRIPT")) == 0) ||
+           (wxStrcmp(tag, wxT("STYLE")) == 0);
+}
 
-#define CACHE_INCREMENT  64
+bool wxIsCDATAElement(const wxString& tag)
+{
+    return (wxStrcmp(tag.wx_str(), wxS("SCRIPT")) == 0) ||
+           (wxStrcmp(tag.wx_str(), wxS("STYLE")) == 0);
+}
 
 wxHtmlTagsCache::wxHtmlTagsCache(const wxString& source)
 {
-    const wxChar *src = source.c_str();
-    int i, tg, pos, stpos;
-    int lng = source.Length();
-    wxChar dummy[256];
-
-    m_Cache = NULL;
-    m_CacheSize = 0;
+    m_Cache = new wxHtmlTagsCacheData;
     m_CachePos = 0;
 
-    pos = 0;
-    while (pos < lng) 
+    wxChar tagBuffer[256];
+
+    const wxString::const_iterator end = source.end();
+    for ( wxString::const_iterator pos = source.begin(); pos < end; ++pos )
     {
-        if (src[pos] == wxT('<'))   // tag found:
+        if (*pos != wxT('<'))
+            continue;
+
+        // possible tag start found:
+
+        // don't cache comment tags
+        if ( wxHtmlParser::SkipCommentTag(pos, end) )
+            continue;
+
+        // Remember the starting tag position.
+        wxString::const_iterator stpos = pos++;
+
+        // And look for the ending one.
+        int i;
+        for ( i = 0;
+              pos < end && i < (int)WXSIZEOF(tagBuffer) - 1 &&
+              *pos != wxT('>') && !wxIsspace(*pos);
+              ++i, ++pos )
+        {
+            tagBuffer[i] = (wxChar)wxToupper(*pos);
+        }
+        tagBuffer[i] = wxT('\0');
+
+        while (pos < end && *pos != wxT('>'))
+            ++pos;
+
+        if ( pos == end )
+        {
+            // We didn't find a closing bracket, this is not a valid tag after
+            // all. Notice that we need to roll back pos to avoid creating an
+            // invalid iterator when "++pos" is done in the loop statement.
+            --pos;
+
+            continue;
+        }
+
+        // We have a valid tag, add it to the cache.
+        size_t tg = Cache().size();
+        Cache().push_back(wxHtmlCacheItem());
+        Cache()[tg].Key = stpos;
+        Cache()[tg].Name = new wxChar[i+1];
+        memcpy(Cache()[tg].Name, tagBuffer, (i+1)*sizeof(wxChar));
+
+        if ((stpos+1) < end && *(stpos+1) == wxT('/')) // ending tag:
         {
-            if (m_CacheSize % CACHE_INCREMENT == 0)
-                m_Cache = (wxHtmlCacheItem*) realloc(m_Cache, (m_CacheSize + CACHE_INCREMENT) * sizeof(wxHtmlCacheItem));
-            tg = m_CacheSize++;
-            m_Cache[tg].Key = stpos = pos++;
-            dummy[0] = 0; i = 0;
-            while (pos < lng && 
-                   src[pos] != wxT('>') &&
-                   src[pos] != wxT(' ') && src[pos] != wxT('\r') && 
-                   src[pos] != wxT('\n') && src[pos] != wxT('\t')) 
+            Cache()[tg].type = wxHtmlCacheItem::Type_EndingTag;
+            // find matching begin tag:
+            for (i = tg; i >= 0; i--)
             {
-                dummy[i] = src[pos++];
-                if ((dummy[i] >= wxT('a')) && (dummy[i] <= wxT('z'))) dummy[i] -= (wxT('a') - wxT('A'));
-                i++;
+                if ((Cache()[i].type == wxHtmlCacheItem::Type_NoMatchingEndingTag) && (wxStrcmp(Cache()[i].Name, tagBuffer+1) == 0))
+                {
+                    Cache()[i].type = wxHtmlCacheItem::Type_Normal;
+                    Cache()[i].End1 = stpos;
+                    Cache()[i].End2 = pos + 1;
+                    break;
+                }
             }
-            dummy[i] = 0;
-            m_Cache[tg].Name = new wxChar[i+1];
-            memcpy(m_Cache[tg].Name, dummy, (i+1)*sizeof(wxChar));
-
-            while (pos < lng && src[pos] != wxT('>')) pos++;
+        }
+        else
+        {
+            Cache()[tg].type = wxHtmlCacheItem::Type_NoMatchingEndingTag;
 
-            if (src[stpos+1] == wxT('/')) // ending tag:
+            if (wxIsCDATAElement(tagBuffer))
             {
-                m_Cache[tg].End1 = m_Cache[tg].End2 = -2;
-                // find matching begin tag:
-                for (i = tg; i >= 0; i--)
-                    if ((m_Cache[i].End1 == -1) && (wxStrcmp(m_Cache[i].Name, dummy+1) == 0)) 
+                // store the orig pos in case we are missing the closing
+                // tag (see below)
+                const wxString::const_iterator old_pos = pos;
+                bool foundCloseTag = false;
+
+                // find next matching tag
+                int tag_len = wxStrlen(tagBuffer);
+                while (pos < end)
+                {
+                    // find the ending tag
+                    while (pos + 1 < end &&
+                           (*pos != '<' || *(pos+1) != '/'))
+                        ++pos;
+                    if (*pos == '<')
+                        ++pos;
+
+                    // see if it matches
+                    int match_pos = 0;
+                    while (pos < end && match_pos < tag_len )
                     {
-                        m_Cache[i].End1 = stpos;
-                        m_Cache[i].End2 = pos + 1;
+                        wxChar c = *pos;
+                        if ( c == '>' || c == '<' )
+                            break;
+
+                        // cast to wxChar needed to suppress warning in
+                        // Unicode build
+                        if ((wxChar)wxToupper(c) == tagBuffer[match_pos])
+                        {
+                            ++match_pos;
+                        }
+                        else if (c == wxT(' ') || c == wxT('\n') ||
+                            c == wxT('\r') || c == wxT('\t'))
+                        {
+                            // need to skip over these
+                        }
+                        else
+                        {
+                            match_pos = 0;
+                        }
+                        ++pos;
+                    }
+
+                    // found a match
+                    if (match_pos == tag_len)
+                    {
+                        pos = pos - tag_len - 3;
+                        foundCloseTag = true;
                         break;
                     }
-            }
-            else 
-            {
-                m_Cache[tg].End1 = m_Cache[tg].End2 = -1;
+                    else // keep looking for the closing tag
+                    {
+                        ++pos;
+                    }
+                }
+                if (!foundCloseTag)
+                {
+                    // we didn't find closing tag; this means the markup
+                    // is incorrect and the best thing we can do is to
+                    // ignore the unclosed tag and continue parsing as if
+                    // it didn't exist:
+                    pos = old_pos;
+                }
             }
         }
-
-        pos++;
     }
 
     // ok, we're done, now we'll free .Name members of cache - we don't need it anymore:
-    for (i = 0; i < m_CacheSize; i++) 
+    for ( wxHtmlTagsCacheData::iterator i = Cache().begin();
+          i != Cache().end(); ++i )
     {
-        delete[] m_Cache[i].Name;
-        m_Cache[i].Name = NULL;
+        wxDELETEA(i->Name);
     }
 }
 
+wxHtmlTagsCache::~wxHtmlTagsCache()
+{
+    delete m_Cache;
+}
 
-
-void wxHtmlTagsCache::QueryTag(int at, int* end1, int* end2)
+void wxHtmlTagsCache::QueryTag(const wxString::const_iterator& at,
+                               const wxString::const_iterator& inputEnd,
+                               wxString::const_iterator *end1,
+                               wxString::const_iterator *end2,
+                               bool *hasEnding)
 {
-    if (m_Cache == NULL) return;
-    if (m_Cache[m_CachePos].Key != at) 
+    if (Cache().empty())
+    {
+        *end1 =
+        *end2 = inputEnd;
+        *hasEnding = true;
+        return;
+    }
+
+    if (Cache()[m_CachePos].Key != at)
+    {
+        int delta = (at < Cache()[m_CachePos].Key) ? -1 : 1;
+        do
+        {
+            m_CachePos += delta;
+
+            if ( m_CachePos < 0 || m_CachePos >= (int)Cache().size() )
+            {
+                if ( m_CachePos < 0 )
+                    m_CachePos = 0;
+                else
+                    m_CachePos = Cache().size() - 1;
+                // something is very wrong with HTML, give up by returning an
+                // impossibly large value which is going to be ignored by the
+                // caller
+                *end1 =
+                *end2 = inputEnd;
+                *hasEnding = true;
+                return;
+            }
+        }
+        while (Cache()[m_CachePos].Key != at);
+    }
+
+    switch ( Cache()[m_CachePos].type )
     {
-        int delta = (at < m_Cache[m_CachePos].Key) ? -1 : 1;
-        do {m_CachePos += delta;} while (m_Cache[m_CachePos].Key != at);
+        case wxHtmlCacheItem::Type_Normal:
+            *end1 = Cache()[m_CachePos].End1;
+            *end2 = Cache()[m_CachePos].End2;
+            *hasEnding = true;
+            break;
+
+        case wxHtmlCacheItem::Type_EndingTag:
+            wxFAIL_MSG("QueryTag called for ending tag - can't be");
+            // but if it does happen, fall through, better than crashing
+
+        case wxHtmlCacheItem::Type_NoMatchingEndingTag:
+            // If input HTML is invalid and there's no closing tag for this
+            // one, pretend that it runs all the way to the end of input
+            *end1 = inputEnd;
+            *end2 = inputEnd;
+            *hasEnding = false;
+            break;
     }
-    *end1 = m_Cache[m_CachePos].End1;
-    *end2 = m_Cache[m_CachePos].End2;
 }
 
 
@@ -142,180 +295,382 @@ void wxHtmlTagsCache::QueryTag(int at, int* end1, int* end2)
 // wxHtmlTag
 //-----------------------------------------------------------------------------
 
-IMPLEMENT_CLASS(wxHtmlTag,wxObject)
-
-wxHtmlTag::wxHtmlTag(const wxString& source, int pos, int end_pos, wxHtmlTagsCache* cache) : wxObject()
+wxHtmlTag::wxHtmlTag(wxHtmlTag *parent,
+                     const wxString *source,
+                     const wxString::const_iterator& pos,
+                     const wxString::const_iterator& end_pos,
+                     wxHtmlTagsCache *cache,
+                     wxHtmlEntitiesParser *entParser)
 {
-    int i;
-    char c;
+    /* Setup DOM relations */
+
+    m_Next = NULL;
+    m_FirstChild = m_LastChild = NULL;
+    m_Parent = parent;
+    if (parent)
+    {
+        m_Prev = m_Parent->m_LastChild;
+        if (m_Prev == NULL)
+            m_Parent->m_FirstChild = this;
+        else
+            m_Prev->m_Next = this;
+        m_Parent->m_LastChild = this;
+    }
+    else
+        m_Prev = NULL;
+
+    /* Find parameters and their values: */
+
+    wxChar c wxDUMMY_INITIALIZE(0);
 
     // fill-in name, params and begin pos:
-    m_Name = m_Params = wxEmptyString;
-    i = pos+1;
-    if (source[i] == wxT('/')) { m_Ending = TRUE; i++; }
-    else m_Ending = FALSE;
+    wxString::const_iterator i(pos+1);
 
     // find tag's name and convert it to uppercase:
-    while ((i < end_pos) && 
-               ((c = source[i++]) != wxT(' ') && c != wxT('\r') && 
-                 c != wxT('\n') && c != wxT('\t') &&
-                 c != wxT('>'))) 
+    while ((i < end_pos) &&
+           ((c = *(i++)) != wxT(' ') && c != wxT('\r') &&
+             c != wxT('\n') && c != wxT('\t') &&
+             c != wxT('>') && c != wxT('/')))
     {
-        if ((c >= wxT('a')) && (c <= wxT('z'))) c -= (wxT('a') - wxT('A'));
-        m_Name += c;
+        if ((c >= wxT('a')) && (c <= wxT('z')))
+            c -= (wxT('a') - wxT('A'));
+        m_Name << c;
     }
 
     // if the tag has parameters, read them and "normalize" them,
-    // i.e. convert to uppercase, replace whitespaces by spaces and 
+    // i.e. convert to uppercase, replace whitespaces by spaces and
     // remove whitespaces around '=':
-    if (source[i-1] != wxT('>'))
-        while ((i < end_pos) && ((c = source[i++]) != wxT('>'))) 
+    if (*(i-1) != wxT('>'))
+    {
+        #define IS_WHITE(c) (c == wxT(' ') || c == wxT('\r') || \
+                             c == wxT('\n') || c == wxT('\t'))
+        wxString pname, pvalue;
+        wxChar quote;
+        enum
+        {
+            ST_BEFORE_NAME = 1,
+            ST_NAME,
+            ST_BEFORE_EQ,
+            ST_BEFORE_VALUE,
+            ST_VALUE
+        } state;
+
+        quote = 0;
+        state = ST_BEFORE_NAME;
+        while (i < end_pos)
         {
-            if ((c >= wxT('a')) && (c <= wxT('z'))) 
-                c -= (wxT('a') - wxT('A'));
-            if (c == wxT('\r') || c == wxT('\n') || c == wxT('\t')) 
-                c = wxT(' '); // make future parsing a bit simpler
-            m_Params += c;
-            if (c == wxT('"')) 
+            c = *(i++);
+
+            if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
             {
-                // remove spaces around the '=' character:
-                if (m_Params.Length() > 1 && 
-                    m_Params[m_Params.Length()-2] == wxT(' '))
+                if (state == ST_BEFORE_EQ || state == ST_NAME)
+                {
+                    m_ParamNames.Add(pname);
+                    m_ParamValues.Add(wxGetEmptyString());
+                }
+                else if (state == ST_VALUE && quote == 0)
                 {
-                    m_Params.RemoveLast();
-                    while (m_Params.Length() > 0 && m_Params.Last() == wxT(' ')) 
-                        m_Params.RemoveLast();
-                    m_Params += wxT('"');
+                    m_ParamNames.Add(pname);
+                    if (entParser)
+                        m_ParamValues.Add(entParser->Parse(pvalue));
+                    else
+                        m_ParamValues.Add(pvalue);
                 }
-                while ((i < end_pos) && (source[i++] == wxT(' '))) {}
-                if (i < end_pos) i--;
-            
-                // ...and copy the value to m_Params:
-                while ((i < end_pos) && ((c = source[i++]) != wxT('"'))) 
-                    m_Params += c;
-                m_Params += c;
+                break;
             }
-            else if (c == wxT('\'')) 
+            switch (state)
             {
-                while ((i < end_pos) && ((c = source[i++]) != wxT('\''))) 
-                    m_Params += c;
-                m_Params += c;
+                case ST_BEFORE_NAME:
+                    if (!IS_WHITE(c))
+                    {
+                        pname = c;
+                        state = ST_NAME;
+                    }
+                    break;
+                case ST_NAME:
+                    if (IS_WHITE(c))
+                        state = ST_BEFORE_EQ;
+                    else if (c == wxT('='))
+                        state = ST_BEFORE_VALUE;
+                    else
+                        pname << c;
+                    break;
+                case ST_BEFORE_EQ:
+                    if (c == wxT('='))
+                        state = ST_BEFORE_VALUE;
+                    else if (!IS_WHITE(c))
+                    {
+                        m_ParamNames.Add(pname);
+                        m_ParamValues.Add(wxGetEmptyString());
+                        pname = c;
+                        state = ST_NAME;
+                    }
+                    break;
+                case ST_BEFORE_VALUE:
+                    if (!IS_WHITE(c))
+                    {
+                        if (c == wxT('"') || c == wxT('\''))
+                            quote = c, pvalue = wxGetEmptyString();
+                        else
+                            quote = 0, pvalue = c;
+                        state = ST_VALUE;
+                    }
+                    break;
+                case ST_VALUE:
+                    if ((quote != 0 && c == quote) ||
+                        (quote == 0 && IS_WHITE(c)))
+                    {
+                        m_ParamNames.Add(pname);
+                        if (quote == 0)
+                        {
+                            // VS: backward compatibility, no real reason,
+                            //     but wxHTML code relies on this... :(
+                            pvalue.MakeUpper();
+                        }
+                        if (entParser)
+                            m_ParamValues.Add(entParser->Parse(pvalue));
+                        else
+                            m_ParamValues.Add(pvalue);
+                        state = ST_BEFORE_NAME;
+                    }
+                    else
+                        pvalue << c;
+                    break;
             }
         }
-   m_Begin = i;
 
-   cache->QueryTag(pos, &m_End1, &m_End2);
-   if (m_End1 > end_pos) m_End1 = end_pos;
-   if (m_End2 > end_pos) m_End2 = end_pos;
+        #undef IS_WHITE
+    }
+    m_Begin = i;
+    cache->QueryTag(pos, source->end(), &m_End1, &m_End2, &m_hasEnding);
+    if (m_End1 > end_pos) m_End1 = end_pos;
+    if (m_End2 > end_pos) m_End2 = end_pos;
+
+#if WXWIN_COMPATIBILITY_2_8
+    m_sourceStart = source->begin();
+#endif
+
+    // Try to parse any style parameters that can be handled simply by
+    // converting them to the equivalent HTML 3 attributes: this is a far cry
+    // from perfect but better than nothing.
+    static const struct EquivAttr
+    {
+        const char *style;
+        const char *attr;
+    } equivAttrs[] =
+    {
+        { "text-align",         "ALIGN"         },
+        { "width",              "WIDTH"         },
+        { "vertical-align",     "VALIGN"        },
+        { "background",         "BGCOLOR"       },
+        { "background-color",   "BGCOLOR"       },
+    };
+
+    wxHtmlStyleParams styleParams(*this);
+    for ( unsigned n = 0; n < WXSIZEOF(equivAttrs); n++ )
+    {
+        const EquivAttr& ea = equivAttrs[n];
+        if ( styleParams.HasParam(ea.style) && !HasParam(ea.attr) )
+        {
+            m_ParamNames.Add(ea.attr);
+            m_ParamValues.Add(styleParams.GetParam(ea.style));
+        }
+    }
 }
 
-    
+wxHtmlTag::~wxHtmlTag()
+{
+    wxHtmlTag *t1, *t2;
+    t1 = m_FirstChild;
+    while (t1)
+    {
+        t2 = t1->GetNextSibling();
+        delete t1;
+        t1 = t2;
+    }
+}
 
 bool wxHtmlTag::HasParam(const wxString& par) const
 {
-    const wxChar *st = m_Params, *p = par;
-    const wxChar *st2, *p2;
-    const wxChar invalid = wxT('\1');
+    return (m_ParamNames.Index(par, false) != wxNOT_FOUND);
+}
 
-    if (*st == 0) return FALSE;
-    if (*p == 0) return FALSE;
-    for (st2 = st, p2 = p; ; st2++) 
+wxString wxHtmlTag::GetParam(const wxString& par, bool with_quotes) const
+{
+    int index = m_ParamNames.Index(par, false);
+    if (index == wxNOT_FOUND)
+        return wxGetEmptyString();
+    if (with_quotes)
     {
-        if (*p2 == 0 && *st2 == wxT('=')) return TRUE;
-        if (*st2 == 0) return FALSE;
-        if (*p2 != *st2) p2 = &invalid;
-        if (*p2 == *st2) p2++;
-        if (*st2 == wxT(' ')) p2 = p;
-        else if (*st2 == wxT('=')) 
-        {
-            p2 = p;
-            while (*st2 != wxT(' ')) 
-            {
-                if (*st2 == wxT('"')) 
-                {
-                    st2++;
-                    while (*st2 != wxT('"')) st2++;
-                }
-                st2++;
-                if (*st2 == 0) return FALSE;
-            }
-        }
+        // VS: backward compatibility, seems to be never used by wxHTML...
+        wxString s;
+        s << wxT('"') << m_ParamValues[index] << wxT('"');
+        return s;
     }
+    else
+        return m_ParamValues[index];
 }
 
+int wxHtmlTag::ScanParam(const wxString& par,
+                         const char *format,
+                         void *param) const
+{
+    wxString parval = GetParam(par);
+    return wxSscanf(parval, format, param);
+}
 
+int wxHtmlTag::ScanParam(const wxString& par,
+                         const wchar_t *format,
+                         void *param) const
+{
+    wxString parval = GetParam(par);
+    return wxSscanf(parval, format, param);
+}
 
-wxString wxHtmlTag::GetParam(const wxString& par, bool with_commas) const
+/* static */
+bool wxHtmlTag::ParseAsColour(const wxString& str, wxColour *clr)
 {
-    const wxChar *st = m_Params, *p = par;
-    const wxChar *st2, *p2;
-    const wxChar invalid = wxT('\1');
-    bool comma;
-    wxChar comma_char;
-
-    if (*st == 0) return wxEmptyString;
-    if (*p == 0) return wxEmptyString;
-    for (st2 = st, p2 = p; ; st2++) 
+    wxCHECK_MSG( clr, false, wxT("invalid colour argument") );
+
+    // handle colours defined in HTML 4.0 first:
+    if (str.length() > 1 && str[0] != wxT('#'))
     {
-        if (*p2 == 0 && *st2 == wxT('='))  // found
-        {
-            wxString fnd = wxEmptyString;
-            st2++; // '=' character
-            comma = FALSE;
-            comma_char = wxT('\0');
-            if (!with_commas && (*(st2) == wxT('"'))) 
-            {
-                st2++;
-                comma = TRUE; 
-                comma_char = wxT('"');
-            }
-            else if (!with_commas && (*(st2) == wxT('\''))) 
-            {
-                st2++; 
-                comma = TRUE;
-                comma_char = wxT('\'');
-            }
-        
-            while (*st2 != 0) 
-            {
-                if (comma && *st2 == comma_char) comma = FALSE;
-                else if ((*st2 == wxT(' ')) && (!comma)) break;
-                fnd += (*(st2++));
-            }
-            if (!with_commas && (*(st2-1) == comma_char)) fnd.RemoveLast();
-            return fnd;
-        }
-        if (*st2 == 0) return wxEmptyString;
-        if (*p2 != *st2) p2 = &invalid;
-        if (*p2 == *st2) p2++;
-        if (*st2 == wxT(' ')) p2 = p;
-        else if (*st2 == wxT('=')) 
-        {
-            p2 = p;
-            while (*st2 != wxT(' ')) 
-            {
-                if (*st2 == wxT('"')) 
-                {
-                    st2++;
-                    while (*st2 != wxT('"')) st2++;
-                }
-                else if (*st2 == wxT('\''))
-                {
-                    st2++;
-                    while (*st2 != wxT('\'')) st2++;
-                }
-                st2++;
-            }
-        }
+        #define HTML_COLOUR(name, r, g, b)              \
+            if (str.IsSameAs(wxS(name), false))         \
+                { clr->Set(r, g, b); return true; }
+        HTML_COLOUR("black",   0x00,0x00,0x00)
+        HTML_COLOUR("silver",  0xC0,0xC0,0xC0)
+        HTML_COLOUR("gray",    0x80,0x80,0x80)
+        HTML_COLOUR("white",   0xFF,0xFF,0xFF)
+        HTML_COLOUR("maroon",  0x80,0x00,0x00)
+        HTML_COLOUR("red",     0xFF,0x00,0x00)
+        HTML_COLOUR("purple",  0x80,0x00,0x80)
+        HTML_COLOUR("fuchsia", 0xFF,0x00,0xFF)
+        HTML_COLOUR("green",   0x00,0x80,0x00)
+        HTML_COLOUR("lime",    0x00,0xFF,0x00)
+        HTML_COLOUR("olive",   0x80,0x80,0x00)
+        HTML_COLOUR("yellow",  0xFF,0xFF,0x00)
+        HTML_COLOUR("navy",    0x00,0x00,0x80)
+        HTML_COLOUR("blue",    0x00,0x00,0xFF)
+        HTML_COLOUR("teal",    0x00,0x80,0x80)
+        HTML_COLOUR("aqua",    0x00,0xFF,0xFF)
+        #undef HTML_COLOUR
     }
+
+    // then try to parse #rrggbb representations or set from other well
+    // known names (note that this doesn't strictly conform to HTML spec,
+    // but it doesn't do real harm -- but it *must* be done after the standard
+    // colors are handled above):
+    if (clr->Set(str))
+        return true;
+
+    return false;
 }
 
+bool wxHtmlTag::GetParamAsColour(const wxString& par, wxColour *clr) const
+{
+    const wxString str = GetParam(par);
+    return !str.empty() && ParseAsColour(str, clr);
+}
 
+bool wxHtmlTag::GetParamAsInt(const wxString& par, int *clr) const
+{
+    if ( !HasParam(par) )
+        return false;
 
-int wxHtmlTag::ScanParam(const wxString& par, wxChar *format, void *param) const
+    long i;
+    if ( !GetParam(par).ToLong(&i) )
+        return false;
+
+    *clr = (int)i;
+    return true;
+}
+
+bool
+wxHtmlTag::GetParamAsIntOrPercent(const wxString& par,
+                                  int* value,
+                                  bool& isPercent) const
 {
-    wxString parval = GetParam(par);
-    return wxSscanf(parval, format, param);
+    const wxString param = GetParam(par);
+    if ( param.empty() )
+        return false;
+
+    wxString num;
+    if ( param.EndsWith("%", &num) )
+    {
+        isPercent = true;
+    }
+    else
+    {
+        isPercent = false;
+        num = param;
+    }
+
+    long lValue;
+    if ( !num.ToLong(&lValue) )
+        return false;
+
+    if ( lValue > INT_MAX || lValue < INT_MIN )
+        return false;
+
+    *value = static_cast<int>(lValue);
+
+    return true;
+}
+
+wxString wxHtmlTag::GetAllParams() const
+{
+    // VS: this function is for backward compatibility only,
+    //     never used by wxHTML
+    wxString s;
+    size_t cnt = m_ParamNames.GetCount();
+    for (size_t i = 0; i < cnt; i++)
+    {
+        s << m_ParamNames[i];
+        s << wxT('=');
+        if (m_ParamValues[i].Find(wxT('"')) != wxNOT_FOUND)
+            s << wxT('\'') << m_ParamValues[i] << wxT('\'');
+        else
+            s << wxT('"') << m_ParamValues[i] << wxT('"');
+    }
+    return s;
+}
+
+wxHtmlTag *wxHtmlTag::GetFirstSibling() const
+{
+    if (m_Parent)
+        return m_Parent->m_FirstChild;
+    else
+    {
+        wxHtmlTag *cur = (wxHtmlTag*)this;
+        while (cur->m_Prev)
+            cur = cur->m_Prev;
+        return cur;
+    }
+}
+
+wxHtmlTag *wxHtmlTag::GetLastSibling() const
+{
+    if (m_Parent)
+        return m_Parent->m_LastChild;
+    else
+    {
+        wxHtmlTag *cur = (wxHtmlTag*)this;
+        while (cur->m_Next)
+            cur = cur->m_Next;
+        return cur;
+    }
+}
+
+wxHtmlTag *wxHtmlTag::GetNextTag() const
+{
+    if (m_FirstChild) return m_FirstChild;
+    if (m_Next) return m_Next;
+    wxHtmlTag *cur = m_Parent;
+    if (!cur) return NULL;
+    while (cur->m_Parent && !cur->m_Next)
+        cur = cur->m_Parent;
+    return cur->m_Next;
 }
 
 #endif