new wxHtmlParser core and changes implied by it

git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@11112 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775

src/html/helpdata.cpp

@@ -73,9 +73,10 @@ static int LINKAGEMODE IndexCompareFunc(const void *a, const void *b)
 
 class HP_Parser : public wxHtmlParser
 {
-    public:
+public:
-        void AddText(const char* WXUNUSED(text)) { }
+    wxObject* GetProduct() { return NULL; }
-        wxObject* GetProduct() { return NULL; }
+protected:
+    virtual void AddText(const wxChar* WXUNUSED(txt)) {}
 };
 
 

src/html/htmlpars.cpp

@@ -32,8 +32,32 @@
 #include "wx/fontmap.h"
 #include "wx/html/htmldefs.h"
 #include "wx/html/htmlpars.h"
+#include "wx/dynarray.h"
+#include "wx/arrimpl.cpp"
 
+//-----------------------------------------------------------------------------
+// wxHtmlParser helpers
+//-----------------------------------------------------------------------------
 
+class wxHtmlTextPiece
+{
+public:
+    wxHtmlTextPiece(int pos, int lng) : m_pos(pos), m_lng(lng) {}
+    int m_pos, m_lng;
+};
+
+WX_DECLARE_OBJARRAY(wxHtmlTextPiece, wxHtmlTextPieces);
+WX_DEFINE_OBJARRAY(wxHtmlTextPieces);
+
+struct wxHtmlParserState
+{
+    wxHtmlTag         *m_curTag;
+    wxHtmlTag         *m_tags;
+    wxHtmlTextPieces  *m_textPieces;
+    int                m_curTextPiece;
+    wxString           m_source;
+    wxHtmlParserState *m_nextState;
+};
 
 //-----------------------------------------------------------------------------
 // wxHtmlParser
@@ -42,14 +66,21 @@
 IMPLEMENT_ABSTRACT_CLASS(wxHtmlParser,wxObject)
 
 wxHtmlParser::wxHtmlParser()
-    : wxObject(), m_Cache(NULL), m_HandlersHash(wxKEY_STRING),
+    : wxObject(), m_HandlersHash(wxKEY_STRING),
       m_FS(NULL), m_HandlersStack(NULL)
 {
     m_entitiesParser = new wxHtmlEntitiesParser;
+    m_Tags = NULL;
+    m_CurTag = NULL;
+    m_TextPieces = NULL;
+    m_CurTextPiece = 0;
+    m_SavedStates = NULL;
 }
 
 wxHtmlParser::~wxHtmlParser()
 {
+    while (RestoreState())
+        DestroyDOMTree();
     delete m_HandlersStack;
     m_HandlersHash.Clear();
     m_HandlersList.DeleteContents(TRUE);
@@ -75,62 +106,192 @@ void wxHtmlParser::InitParser(const wxString& source)
 
 void wxHtmlParser::DoneParser()
 {
-    delete m_Cache;
+    DestroyDOMTree();
-    m_Cache = NULL;
 }
 
 void wxHtmlParser::SetSource(const wxString& src)
 {
+    DestroyDOMTree();
     m_Source = src;
-    delete m_Cache;
+    CreateDOMTree();
-    m_Cache = new wxHtmlTagsCache(m_Source);
+    m_CurTag = NULL;
+    m_CurTextPiece = 0;
+}
+
+void wxHtmlParser::CreateDOMTree()
+{
+    wxHtmlTagsCache cache(m_Source);
+    m_TextPieces = new wxHtmlTextPieces;
+    CreateDOMSubTree(NULL, 0, m_Source.Length(), &cache);
+    m_CurTextPiece = 0;
+}
+
+void wxHtmlParser::CreateDOMSubTree(wxHtmlTag *cur,
+                                    int begin_pos, int end_pos,
+                                    wxHtmlTagsCache *cache)
+{
+    if (end_pos <= begin_pos) return;
+
+    wxChar c;
+    int i = begin_pos;
+    int textBeginning = begin_pos;
+    
+    while (i < end_pos)
+    {
+        c = m_Source.GetChar(i);
+
+        if (c == wxT('<'))
+        {
+            // add text to m_TextPieces:
+            if (i - textBeginning > 0)
+                m_TextPieces->Add(
+                    wxHtmlTextPiece(textBeginning, i - textBeginning));
+
+            // if it is a comment, skip it:
+            if (i < end_pos-6 && m_Source.GetChar(i+1) == wxT('!') &&
+                                 m_Source.GetChar(i+2) == wxT('-') &&
+                                 m_Source.GetChar(i+3) == wxT('-'))
+            {
+                // Comments begin with "<!--" and end with "--[ \t\r\n]*>"
+                // according to HTML 4.0
+                int dashes = 0;
+                i += 4;
+                while (i < end_pos)
+                {
+                    c = m_Source.GetChar(i++);
+                    if ((c == wxT(' ') || c == wxT('\n') || 
+                        c == wxT('\r') || c == wxT('\t')) && dashes >= 2) {}
+                    else if (c == wxT('>') && dashes >= 2)
+                    {
+                        textBeginning = i;
+                        break;
+                    }
+                    else if (c == wxT('-'))
+                        dashes++;
+                    else    
+                        dashes = 0;
+                }
+            }
+        
+            // add another tag to the tree:
+            else if (i < end_pos-1 && m_Source.GetChar(i+1) != wxT('/'))
+	        {
+                wxHtmlTag *chd;
+                if (cur) 
+                    chd = new wxHtmlTag(cur, m_Source, 
+                                        i, end_pos, cache, m_entitiesParser);
+                else 
+                {
+                    chd = new wxHtmlTag(NULL, m_Source,
+                                        i, end_pos, cache, m_entitiesParser);
+                    if (!m_Tags) 
+                    {
+                        // if this is the first tag to be created make the root 
+                        // m_Tags point to it:
+                        m_Tags = chd;
+                    }
+                    else
+                    {
+                        // if there is already a root tag add this tag as 
+                        // the last sibling:
+                        chd->m_Prev = m_Tags->GetLastSibling();
+                        chd->m_Prev->m_Next = chd;
+                    }
+                }
+
+                if (chd->HasEnding())
+                {
+                    CreateDOMSubTree(chd,
+                                     chd->GetBeginPos(), chd->GetEndPos1(), 
+                                     cache);
+                    i = chd->GetEndPos2();
+                }
+                else
+                    i = chd->GetBeginPos();
+                textBeginning = i;
+            }
+
+            // ... or skip ending tag:
+            else 
+            {
+                while (i < end_pos && m_Source.GetChar(i) != wxT('>')) i++;
+                textBeginning = i+1;
+            }
+        }
+        else i++;
+    }
+
+    // add remaining text to m_TextPieces:
+    if (end_pos - textBeginning > 0)
+        m_TextPieces->Add(
+            wxHtmlTextPiece(textBeginning, end_pos - textBeginning));
+}
+
+void wxHtmlParser::DestroyDOMTree()
+{
+    wxHtmlTag *t1, *t2;
+    t1 = m_Tags;
+    while (t1)
+    {
+        t2 = t1->GetNextSibling();
+        delete t1;
+        t1 = t2;
+    }
+    m_Tags = m_CurTag = NULL;
+
+    delete m_TextPieces;
+    m_TextPieces = NULL;
+}
+
+void wxHtmlParser::DoParsing() 
+{
+    m_CurTag = m_Tags;
+    m_CurTextPiece = 0;
+    DoParsing(0, m_Source.Length()); 
 }
 
 void wxHtmlParser::DoParsing(int begin_pos, int end_pos)
 {
     if (end_pos <= begin_pos) return;
     
-    char c;
+    wxHtmlTextPieces& pieces = *m_TextPieces;
-    char *temp = new char[end_pos - begin_pos + 1];
+    size_t piecesCnt = pieces.GetCount();
-    int i;
-    int templen;
     
-    templen = 0;
+    while (begin_pos < end_pos)
-    i = begin_pos;
-
-    while (i < end_pos)
     {
-        c = m_Source[(unsigned int) i];
+        while (m_CurTag && m_CurTag->GetBeginPos() < begin_pos)
+            m_CurTag = m_CurTag->GetNextTag();
+        while (m_CurTextPiece < piecesCnt && 
+               pieces[m_CurTextPiece].m_pos < begin_pos)
+            m_CurTextPiece++;
 
-        // continue building word:
+        if (m_CurTextPiece < piecesCnt && 
-        if (c != '<')
+            (!m_CurTag || 
-	    {
+             pieces[m_CurTextPiece].m_pos < m_CurTag->GetBeginPos()))
-            temp[templen++] = c;
+        {
-            i++;
+            // Add text:
+            AddText(m_Source.Mid(pieces[m_CurTextPiece].m_pos,
+                                 pieces[m_CurTextPiece].m_lng));
+            begin_pos = pieces[m_CurTextPiece].m_pos + 
+                        pieces[m_CurTextPiece].m_lng;
+            m_CurTextPiece++;
         }
-
+        else if (m_CurTag)
-        else if (c == '<')
+        {
-	    {
+            // Add tag:
-            wxHtmlTag tag(m_Source, i, end_pos, m_Cache, m_entitiesParser);
+            if (m_CurTag)
-
+            {
-            if (templen)
+                if (m_CurTag->HasEnding())
-	        {
+                    begin_pos = m_CurTag->GetEndPos2();
-                temp[templen] = 0;
+                else 
-                AddText(temp);
+                    begin_pos = m_CurTag->GetBeginPos();
-                templen = 0;
             }
-            AddTag(tag);
+            wxHtmlTag *t = m_CurTag;
-            if (tag.HasEnding()) i = tag.GetEndPos2();
+            m_CurTag = m_CurTag->GetNextTag();
-            else i = tag.GetBeginPos();
+            AddTag(*t);
         }
+        else break;
     }
-
-    if (templen)
-    { // last word of block :-(
-        temp[templen] = 0;
-        AddText(temp);
-    }
-    delete[] temp;
 }
 
 void wxHtmlParser::AddTag(const wxHtmlTag& tag)
@@ -151,7 +312,7 @@ void wxHtmlParser::AddTag(const wxHtmlTag& tag)
 void wxHtmlParser::AddTagHandler(wxHtmlTagHandler *handler)
 {
     wxString s(handler->GetSupportedTags());
-    wxStringTokenizer tokenizer(s, ", ");
+    wxStringTokenizer tokenizer(s, wxT(", "));
 
     while (tokenizer.HasMoreTokens())
         m_HandlersHash.Put(tokenizer.NextToken(), handler);
@@ -164,7 +325,7 @@ void wxHtmlParser::AddTagHandler(wxHtmlTagHandler *handler)
 
 void wxHtmlParser::PushTagHandler(wxHtmlTagHandler *handler, wxString tags)
 {
-    wxStringTokenizer tokenizer(tags, ", ");
+    wxStringTokenizer tokenizer(tags, wxT(", "));
     wxString key;
 
     if (m_HandlersStack == NULL)
@@ -197,6 +358,45 @@ void wxHtmlParser::PopTagHandler()
     m_HandlersStack->DeleteNode(first);
 }
 
+void wxHtmlParser::SetSourceAndSaveState(const wxString& src)
+{
+    wxHtmlParserState *s = new wxHtmlParserState;
+
+    s->m_curTag = m_CurTag;
+    s->m_tags = m_Tags;
+    s->m_textPieces = m_TextPieces;
+    s->m_curTextPiece = m_CurTextPiece;
+    s->m_source = m_Source;
+
+    s->m_nextState = m_SavedStates;
+    m_SavedStates = s;
+
+    m_CurTag = NULL;
+    m_Tags = NULL;
+    m_TextPieces = NULL;
+    m_CurTextPiece = 0;
+    m_Source = wxEmptyString;
+    
+    SetSource(src);
+}
+
+bool wxHtmlParser::RestoreState()
+{
+    if (!m_SavedStates) return FALSE;
+    
+    wxHtmlParserState *s = m_SavedStates;
+    m_SavedStates = s->m_nextState;
+    
+    m_CurTag = s->m_curTag;
+    m_Tags = s->m_tags;
+    m_TextPieces = s->m_textPieces;
+    m_CurTextPiece = s->m_curTextPiece;
+    m_Source = s->m_source;
+    
+    delete s;
+    return TRUE;
+}
+
 //-----------------------------------------------------------------------------
 // wxHtmlTagHandler
 //-----------------------------------------------------------------------------
@@ -256,9 +456,9 @@ wxString wxHtmlEntitiesParser::Parse(const wxString& input)
                    (*c >= wxT('0') && *c <= wxT('9')) ||
                    *c == wxT('_') || *c == wxT('#'); c++) {}
             entity.append(ent_s, c - ent_s);
-            if (*c == wxT(';')) c++;
+            if (*c != wxT(';')) c--;
+            last = c+1;
             output << GetEntityChar(entity);
-            last = c;
         }
     }
     if (*last != wxT('\0'))

src/html/htmltag.cpp

@@ -146,19 +146,35 @@ void wxHtmlTagsCache::QueryTag(int at, int* end1, int* end2)
 
 IMPLEMENT_CLASS(wxHtmlTag,wxObject)
 
-wxHtmlTag::wxHtmlTag(const wxString& source, int pos, int end_pos, 
+wxHtmlTag::wxHtmlTag(wxHtmlTag *parent,
+                     const wxString& source, int pos, int end_pos, 
                      wxHtmlTagsCache *cache,
                      wxHtmlEntitiesParser *entParser) : wxObject()
 {
+    /* 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: */
+    
     int i;
     wxChar c;
 
     // fill-in name, params and begin pos:
     i = pos+1;
-    if (source[i] == wxT('/')) 
-        { m_Ending = TRUE; i++; }
-    else 
-        m_Ending = FALSE;
 
     // find tag's name and convert it to uppercase:
     while ((i < end_pos) && 
@@ -282,6 +298,12 @@ wxHtmlTag::wxHtmlTag(const wxString& source, int pos, int end_pos,
    if (m_End2 > end_pos) m_End2 = end_pos;
 }
 
+wxHtmlTag::~wxHtmlTag()
+{
+    for (wxHtmlTag *t = m_FirstChild; t; t = t->GetNextSibling())
+        delete t;
+}
+
 bool wxHtmlTag::HasParam(const wxString& par) const
 {
     return (m_ParamNames.Index(par, FALSE) != wxNOT_FOUND);
@@ -378,4 +400,41 @@ wxString wxHtmlTag::GetAllParams() const
     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

src/html/m_dflist.cpp

@@ -63,24 +63,19 @@ TAG_HANDLER_BEGIN(DEFLIST, "DL,DT,DD")
         }
         else if (tag.GetName() == wxT("DT"))
         {
-            if (!tag.IsEnding())
+            m_WParser->CloseContainer();
-            {
+            c = m_WParser->OpenContainer();
-                m_WParser->CloseContainer();
+            c->SetAlignHor(wxHTML_ALIGN_LEFT);
-                c = m_WParser->OpenContainer();
+            c->SetMinHeight(m_WParser->GetCharHeight());
-                c->SetAlignHor(wxHTML_ALIGN_LEFT);
-                c->SetMinHeight(m_WParser->GetCharHeight());
-            }
             return FALSE;
         }
-        else if (!tag.IsEnding()) // "DD"
+        else // "DD"
         {
             m_WParser->CloseContainer();
             c = m_WParser->OpenContainer();
             c->SetIndent(5 * m_WParser->GetCharWidth(), wxHTML_INDENT_LEFT);
             return FALSE;
         }
-
-        else return FALSE;
     }
 
 TAG_HANDLER_END(DEFLIST)

src/html/m_list.cpp

@@ -87,34 +87,32 @@ TAG_HANDLER_BEGIN(OLULLI, "OL,UL,LI")
         // List Item:
         if (tag.GetName() == wxT("LI"))
         {
-            if (!tag.IsEnding())
+            m_WParser->GetContainer()->SetIndent(0, wxHTML_INDENT_TOP);
+                // this is to prevent indetation in <li><p> case
+            m_WParser->CloseContainer();
+            m_WParser->CloseContainer();
+
+            c = m_WParser->OpenContainer();
+            c->SetWidthFloat(2 * m_WParser->GetCharWidth(), wxHTML_UNITS_PIXELS);
+            c->SetAlignHor(wxHTML_ALIGN_RIGHT);
+            if (m_Numbering == 0)
+                c->InsertCell(new wxHtmlListmarkCell(m_WParser->GetDC(), m_WParser->GetActualColor()));
+            else
             {
-                m_WParser->GetContainer()->SetIndent(0, wxHTML_INDENT_TOP);
+                wxString mark;
-                    // this is to prevent indetation in <li><p> case
+                mark.Printf(wxT("%i."), m_Numbering);
-                m_WParser->CloseContainer();
+                c->InsertCell(new wxHtmlWordCell(mark, *(m_WParser->GetDC())));
-                m_WParser->CloseContainer();
-
-                c = m_WParser->OpenContainer();
-                c->SetWidthFloat(2 * m_WParser->GetCharWidth(), wxHTML_UNITS_PIXELS);
-                c->SetAlignHor(wxHTML_ALIGN_RIGHT);
-                if (m_Numbering == 0)
-                    c->InsertCell(new wxHtmlListmarkCell(m_WParser->GetDC(), m_WParser->GetActualColor()));
-                else
-                {
-                    wxString mark;
-                    mark.Printf(wxT("%i."), m_Numbering);
-                    c->InsertCell(new wxHtmlWordCell(mark, *(m_WParser->GetDC())));
-                }
-                m_WParser->CloseContainer();
-
-                c = m_WParser->OpenContainer();
-                c->SetIndent(m_WParser->GetCharWidth() / 4, wxHTML_INDENT_LEFT);
-                c->SetWidthFloat(-2 * m_WParser->GetCharWidth(), wxHTML_UNITS_PIXELS);
-
-                m_WParser->OpenContainer();
-
-                if (m_Numbering != 0) m_Numbering++;
             }
+            m_WParser->CloseContainer();
+
+            c = m_WParser->OpenContainer();
+            c->SetIndent(m_WParser->GetCharWidth() / 4, wxHTML_INDENT_LEFT);
+            c->SetWidthFloat(-2 * m_WParser->GetCharWidth(), wxHTML_UNITS_PIXELS);
+
+            m_WParser->OpenContainer();
+
+            if (m_Numbering != 0) m_Numbering++;
+
             return FALSE;
         }
 

src/html/m_pre.cpp

@@ -60,20 +60,18 @@ TAG_HANDLER_BEGIN(PRE, "PRE")
         c = m_WParser->OpenContainer();
         c->SetAlignHor(wxHTML_ALIGN_LEFT);
 
-        wxString src, srcMid;
+        wxString srcMid = 
-
+            m_WParser->GetSource()->Mid(tag.GetBeginPos(),
-        src = *m_WParser->GetSource();
+                                        tag.GetEndPos1() - tag.GetBeginPos());
-        srcMid = src.Mid(tag.GetBeginPos(),
-                         tag.GetEndPos1() - tag.GetBeginPos());
         srcMid.Replace(wxT("\t"), wxT("        "));
         srcMid.Replace(wxT(" "), wxT(" "));
         srcMid.Replace(wxT("\n"), wxT("<br>"));
 
         // It is safe to temporarily change the source being parsed,
         // provided we restore the state back after parsing
-        m_Parser->SetSource(srcMid);
+        m_Parser->SetSourceAndSaveState(srcMid);
         m_Parser->DoParsing();
-        m_Parser->SetSource(src);
+        m_Parser->RestoreState();
 
         m_WParser->CloseContainer();
         c = m_WParser->OpenContainer();

src/html/m_tables.cpp

@@ -496,7 +496,7 @@ TAG_HANDLER_BEGIN(TABLE, "TABLE,TR,TD,TH")
         }
 
 
-        else if (m_Table && !tag.IsEnding())
+        else if (m_Table)
         {
             // new row in table
             if (tag.GetName() == wxT("TR"))

src/html/winpars.cpp

@@ -42,6 +42,8 @@ wxList wxHtmlWinParser::m_Modules;
 
 wxHtmlWinParser::wxHtmlWinParser(wxWindow *wnd) : wxHtmlParser()
 {
+    m_tmpStrBuf = NULL;
+    m_tmpStrBufSize = 0;
     m_Window = wnd;
     m_Container = NULL;
     m_DC = NULL;
@@ -96,7 +98,8 @@ wxHtmlWinParser::~wxHtmlWinParser()
                         if (m_FontsTable[i][j][k][l][m] != NULL) 
                             delete m_FontsTable[i][j][k][l][m];
                     }
-    if (m_EncConv) delete m_EncConv;
+    delete m_EncConv;
+    delete[] m_tmpStrBuf;
 }
 
 
@@ -190,37 +193,43 @@ wxObject* wxHtmlWinParser::GetProduct()
 }
 
 
-
+void wxHtmlWinParser::AddText(const wxChar* txt)
-void wxHtmlWinParser::AddText(const char* txt)
 {
     wxHtmlCell *c;
-    int i = 0, x, lng = strlen(txt);
+    int i = 0, x, lng = wxStrlen(txt);
-    char temp[wxHTML_BUFLEN];
+    register wxChar d;
-    register char d;
     int templen = 0;
 
+    if (lng+1 > m_tmpStrBufSize)
+    {
+        delete[] m_tmpStrBuf;
+        m_tmpStrBuf = new wxChar[lng+1];
+        m_tmpStrBufSize = lng+1;
+    }
+    wxChar *temp = m_tmpStrBuf;
+    
     if (m_tmpLastWasSpace) 
     {
         while ((i < lng) && 
-               ((txt[i] == '\n') || (txt[i] == '\r') || (txt[i] == ' ') || 
+               ((txt[i] == wxT('\n')) || (txt[i] == wxT('\r')) || (txt[i] == wxT(' ')) || 
-                (txt[i] == '\t'))) i++;
+                (txt[i] == wxT('\t')))) i++;
     }
 
     while (i < lng) 
     {
         x = 0;
         d = temp[templen++] = txt[i];
-        if ((d == '\n') || (d == '\r') || (d == ' ') || (d == '\t')) 
+        if ((d == wxT('\n')) || (d == wxT('\r')) || (d == wxT(' ')) || (d == wxT('\t'))) 
 	    {
             i++, x++;
-            while ((i < lng) && ((txt[i] == '\n') || (txt[i] == '\r') || 
+            while ((i < lng) && ((txt[i] == wxT('\n')) || (txt[i] == wxT('\r')) || 
-                                 (txt[i] == ' ') || (txt[i] == '\t'))) i++, x++;
+                                 (txt[i] == wxT(' ')) || (txt[i] == wxT('\t')))) i++, x++;
         }
         else i++;
 
         if (x) 
 	    {
-            temp[templen-1] = ' ';
+            temp[templen-1] = wxT(' ');
             temp[templen] = 0;
             templen = 0;
             if (m_EncConv)