htmltag.cpp

来自「A*算法 A*算法 A*算法 A*算法A*算法A*算法」· C++ 代码 · 共 534 行 · 第 1/2 页

/////////////////////////////////////////////////////////////////////////////
// Name:        htmltag.cpp
// Purpose:     wxHtmlTag class (represents single tag)
// Author:      Vaclav Slavik
// RCS-ID:      $Id: htmltag.cpp,v 1.49 2005/07/22 17:08:42 ABX Exp $
// Copyright:   (c) 1999 Vaclav Slavik
// Licence:     wxWindows licence
/////////////////////////////////////////////////////////////////////////////


#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA)
#pragma implementation "htmltag.h"
#endif

#include "wx/wxprec.h"

#include "wx/defs.h"
#if wxUSE_HTML

#ifdef __BORLANDC__
#pragma hdrstop
#endif

#ifndef WXPRECOMP
#endif

#include "wx/html/htmltag.h"
#include "wx/html/htmlpars.h"
#include "wx/colour.h"
#include <stdio.h> // for vsscanf
#include <stdarg.h>


//-----------------------------------------------------------------------------
// wxHtmlTagsCache
//-----------------------------------------------------------------------------

struct wxHtmlCacheItem
{
    // this is "pos" value passed to wxHtmlTag's constructor.
    // it is position of '<' character of the tag
    int Key;

    // 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;

    // name of this tag
    wxChar *Name;
};


IMPLEMENT_CLASS(wxHtmlTagsCache,wxObject)

#define CACHE_INCREMENT  64

bool wxIsCDATAElement(const wxChar *tag)
{
    return (wxStrcmp(tag, _T("SCRIPT")) == 0) ||
           (wxStrcmp(tag, _T("STYLE")) == 0);
}

wxHtmlTagsCache::wxHtmlTagsCache(const wxString& source)
{
    const wxChar *src = source.c_str();
    int lng = source.Length();
    wxChar tagBuffer[256];

    m_Cache = NULL;
    m_CacheSize = 0;
    m_CachePos = 0;

    int pos = 0;
    while (pos < lng)
    {
        if (src[pos] == wxT('<'))   // tag found:
        {
            if (m_CacheSize % CACHE_INCREMENT == 0)
                m_Cache = (wxHtmlCacheItem*) realloc(m_Cache, (m_CacheSize + CACHE_INCREMENT) * sizeof(wxHtmlCacheItem));
            int tg = m_CacheSize++;
            int stpos = pos++;
            m_Cache[tg].Key = stpos;

            int i;
            for ( i = 0;
                  pos < lng && i < (int)WXSIZEOF(tagBuffer) - 1 &&
                  src[pos] != wxT('>') && !wxIsspace(src[pos]);
                  i++, pos++ )
            {
                tagBuffer[i] = (wxChar)wxToupper(src[pos]);
            }
            tagBuffer[i] = _T('\0');

            m_Cache[tg].Name = new wxChar[i+1];
            memcpy(m_Cache[tg].Name, tagBuffer, (i+1)*sizeof(wxChar));

            while (pos < lng && src[pos] != wxT('>')) pos++;

            if (src[stpos+1] == wxT('/')) // ending tag:
            {
                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, tagBuffer+1) == 0))
                    {
                        m_Cache[i].End1 = stpos;
                        m_Cache[i].End2 = pos + 1;
                        break;
                    }
            }
            else
            {
                m_Cache[tg].End1 = m_Cache[tg].End2 = -1;

                if (wxIsCDATAElement(tagBuffer))
                {
                    // store the orig pos in case we are missing the closing
                    // tag (see below)
                    wxInt32 old_pos = pos;
                    bool foundCloseTag = false;

                    // find next matching tag
                    int tag_len = wxStrlen(tagBuffer);
                    while (pos < lng)
                    {
                        // find the ending tag
                        while (pos + 1 < lng &&
                               (src[pos] != '<' || src[pos+1] != '/'))
                            ++pos;
                        if (src[pos] == '<')
                            ++pos;

                        // see if it matches
                        int match_pos = 0;
                        while (pos < lng && match_pos < tag_len && src[pos] != '>' && src[pos] != '<') {
                            // cast to wxChar needed to suppress warning in
                            // Unicode build
                            if ((wxChar)wxToupper(src[pos]) == tagBuffer[match_pos]) {
                                ++match_pos;
                            }
                            else if (src[pos] == wxT(' ') || src[pos] == wxT('\n') ||
                                src[pos] == wxT('\r') || src[pos] == 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 // 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 (int i = 0; i < m_CacheSize; i++)
    {
        delete[] m_Cache[i].Name;
        m_Cache[i].Name = NULL;
    }
}

void wxHtmlTagsCache::QueryTag(int at, int* end1, int* end2)
{
    if (m_Cache == NULL) return;
    if (m_Cache[m_CachePos].Key != at)
    {
        int delta = (at < m_Cache[m_CachePos].Key) ? -1 : 1;
        do
        {
            if ( m_CachePos < 0 || m_CachePos == m_CacheSize )
            {
                // 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 = INT_MAX;
                return;
            }

            m_CachePos += delta;
        }
        while (m_Cache[m_CachePos].Key != at);
    }
    *end1 = m_Cache[m_CachePos].End1;
    *end2 = m_Cache[m_CachePos].End2;
}




//-----------------------------------------------------------------------------
// wxHtmlTag
//-----------------------------------------------------------------------------

IMPLEMENT_CLASS(wxHtmlTag,wxObject)

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;

    // 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('>')))
    {
        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
    // remove whitespaces around '=':