[wxWidgets.git] / tests / benchmarks / htmlparser / htmltag.cpp

/////////////////////////////////////////////////////////////////////////////
// Name:        src/html/htmltag.cpp
// Purpose:     wx28HtmlTag class (represents single tag)
// Author:      Vaclav Slavik
// RCS-ID:      $Id$
// Copyright:   (c) 1999 Vaclav Slavik
// Licence:     wxWindows licence
/////////////////////////////////////////////////////////////////////////////

#include "wx/wxprec.h"

#ifdef __BORLANDC__
    #pragma hdrstop
#endif

#include "htmltag.h"

#ifndef WXPRECOMP
    #include "wx/colour.h"
#endif

#include "htmlpars.h"
#include <stdio.h> // for vsscanf
#include <stdarg.h>


//-----------------------------------------------------------------------------
// wx28HtmlTagsCache
//-----------------------------------------------------------------------------

struct wx28HtmlCacheItem
{
    // this is "pos" value passed to wx28HtmlTag'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(wx28HtmlTagsCache,wxObject)

#define CACHE_INCREMENT  64

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

wx28HtmlTagsCache::wx28HtmlTagsCache(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 = (wx28HtmlCacheItem*) realloc(m_Cache, (m_CacheSize + CACHE_INCREMENT) * sizeof(wx28HtmlCacheItem));
            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] = wxT('\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 wx28HtmlTagsCache::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;
}


//-----------------------------------------------------------------------------
// wx28HtmlTag
//-----------------------------------------------------------------------------

IMPLEMENT_CLASS(wx28HtmlTag,wxObject)

wx28HtmlTag::wx28HtmlTag(wx28HtmlTag *parent,
                     const wxString& source, int pos, int end_pos,
                     wx28HtmlTagsCache *cache,
                     wx28HtmlEntitiesParser *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 '=':
    if (source[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)
        {
            c = source[i++];

            if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
            {
                if (state == ST_BEFORE_EQ || state == ST_NAME)
                {
                    m_ParamNames.Add(pname);
                    m_ParamValues.Add(wxEmptyString);
                }
                else if (state == ST_VALUE && quote == 0)
                {
                    m_ParamNames.Add(pname);
                    if (entParser)
                        m_ParamValues.Add(entParser->Parse(pvalue));
                    else
                        m_ParamValues.Add(pvalue);
                }
                break;
            }
            switch (state)
            {
                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(wxEmptyString);
                        pname = c;
                        state = ST_NAME;
                    }
                    break;
                case ST_BEFORE_VALUE:
                    if (!IS_WHITE(c))
                    {
                        if (c == wxT('"') || c == wxT('\''))
                            quote = c, pvalue = wxEmptyString;
                        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;
            }
        }

        #undef IS_WHITE
    }
    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;
}

wx28HtmlTag::~wx28HtmlTag()
{
    wx28HtmlTag *t1, *t2;
    t1 = m_FirstChild;
    while (t1)
    {
        t2 = t1->GetNextSibling();
        delete t1;
        t1 = t2;
    }
}

bool wx28HtmlTag::HasParam(const wxString& par) const
{
    return (m_ParamNames.Index(par, false) != wxNOT_FOUND);
}

wxString wx28HtmlTag::GetParam(const wxString& par, bool with_commas) const
{
    int index = m_ParamNames.Index(par, false);
    if (index == wxNOT_FOUND)
        return wxEmptyString;
    if (with_commas)
    {
        // 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 wx28HtmlTag::ScanParam(const wxString& par,
                         const wxChar *format,
                         void *param) const
{
    wxString parval = GetParam(par);
    return wxSscanf(parval, format, param);
}

bool wx28HtmlTag::GetParamAsColour(const wxString& par, wxColour *clr) const
{
    wxCHECK_MSG( clr, false, wxT("invalid colour argument") );

    wxString str = GetParam(par);

    // handle colours defined in HTML 4.0 first:
    if (str.length() > 1 && str[0] != wxT('#'))
    {
        #define HTML_COLOUR(name, r, g, b)              \
            if (str.IsSameAs(wxT(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 wx28HtmlTag::GetParamAsInt(const wxString& par, int *clr) const
{
    if ( !HasParam(par) )
        return false;

    long i;
    if ( !GetParam(par).ToLong(&i) )
        return false;

    *clr = (int)i;
    return true;
}

wxString wx28HtmlTag::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;
}

wx28HtmlTag *wx28HtmlTag::GetFirstSibling() const
{
    if (m_Parent)
        return m_Parent->m_FirstChild;
    else
    {
        wx28HtmlTag *cur = (wx28HtmlTag*)this;
        while (cur->m_Prev)
            cur = cur->m_Prev;
        return cur;
    }
}

wx28HtmlTag *wx28HtmlTag::GetLastSibling() const
{
    if (m_Parent)
        return m_Parent->m_LastChild;
    else
    {
        wx28HtmlTag *cur = (wx28HtmlTag*)this;
        while (cur->m_Next)
            cur = cur->m_Next;
        return cur;
    }
}

wx28HtmlTag *wx28HtmlTag::GetNextTag() const
{
    if (m_FirstChild) return m_FirstChild;
    if (m_Next) return m_Next;
    wx28HtmlTag *cur = m_Parent;
    if (!cur) return NULL;
    while (cur->m_Parent && !cur->m_Next)
        cur = cur->m_Parent;
    return cur->m_Next;
}
Commit	Line	Data
c21faa0a VS	1	/////////////////////////////////////////////////////////////////////////////
	2	// Name: src/html/htmltag.cpp
	3	// Purpose: wx28HtmlTag class (represents single tag)
	4	// Author: Vaclav Slavik
	5	// RCS-ID: $Id$
	6	// Copyright: (c) 1999 Vaclav Slavik
	7	// Licence: wxWindows licence
	8	/////////////////////////////////////////////////////////////////////////////
	9
	10	#include "wx/wxprec.h"
	11
	12	#ifdef __BORLANDC__
	13	#pragma hdrstop
	14	#endif
	15
	16	#include "htmltag.h"
	17
	18	#ifndef WXPRECOMP
	19	#include "wx/colour.h"
	20	#endif
	21
	22	#include "htmlpars.h"
	23	#include <stdio.h> // for vsscanf
	24	#include <stdarg.h>
	25
	26
	27	//-----------------------------------------------------------------------------
	28	// wx28HtmlTagsCache
	29	//-----------------------------------------------------------------------------
	30
	31	struct wx28HtmlCacheItem
	32	{
	33	// this is "pos" value passed to wx28HtmlTag's constructor.
	34	// it is position of '<' character of the tag
	35	int Key;
	36
	37	// end positions for the tag:
	38	// end1 is '<' of ending tag,
	39	// end2 is '>' or both are
	40	// -1 if there is no ending tag for this one...
	41	// or -2 if this is ending tag </...>
	42	int End1, End2;
	43
	44	// name of this tag
	45	wxChar *Name;
	46	};
	47
	48
	49	IMPLEMENT_CLASS(wx28HtmlTagsCache,wxObject)
	50
	51	#define CACHE_INCREMENT 64
	52
	53	bool wxIsCDATAElement(const wxChar *tag)
	54	{
9a83f860 VZ	55	return (wxStrcmp(tag, wxT("SCRIPT")) == 0) \|\|
9a83f860 VZ	56	(wxStrcmp(tag, wxT("STYLE")) == 0);
c21faa0a VS	57	}
	58
	59	wx28HtmlTagsCache::wx28HtmlTagsCache(const wxString& source)
	60	{
	61	const wxChar *src = source.c_str();
	62	int lng = source.length();
	63	wxChar tagBuffer[256];
	64
	65	m_Cache = NULL;
	66	m_CacheSize = 0;
	67	m_CachePos = 0;
	68
	69	int pos = 0;
	70	while (pos < lng)
	71	{
	72	if (src[pos] == wxT('<')) // tag found:
	73	{
	74	if (m_CacheSize % CACHE_INCREMENT == 0)
	75	m_Cache = (wx28HtmlCacheItem) realloc(m_Cache, (m_CacheSize + CACHE_INCREMENT) sizeof(wx28HtmlCacheItem));
	76	int tg = m_CacheSize++;
	77	int stpos = pos++;
	78	m_Cache[tg].Key = stpos;
	79
	80	int i;
	81	for ( i = 0;
	82	pos < lng && i < (int)WXSIZEOF(tagBuffer) - 1 &&
	83	src[pos] != wxT('>') && !wxIsspace(src[pos]);
	84	i++, pos++ )
	85	{
	86	tagBuffer[i] = (wxChar)wxToupper(src[pos]);
	87	}
9a83f860	88	tagBuffer[i] = wxT('\0');
c21faa0a VS	89
	90	m_Cache[tg].Name = new wxChar[i+1];
	91	memcpy(m_Cache[tg].Name, tagBuffer, (i+1)*sizeof(wxChar));
	92
	93	while (pos < lng && src[pos] != wxT('>')) pos++;
	94
	95	if (src[stpos+1] == wxT('/')) // ending tag:
	96	{
	97	m_Cache[tg].End1 = m_Cache[tg].End2 = -2;
	98	// find matching begin tag:
	99	for (i = tg; i >= 0; i--)
	100	if ((m_Cache[i].End1 == -1) && (wxStrcmp(m_Cache[i].Name, tagBuffer+1) == 0))
	101	{
	102	m_Cache[i].End1 = stpos;
	103	m_Cache[i].End2 = pos + 1;
	104	break;
	105	}
	106	}
	107	else
	108	{
	109	m_Cache[tg].End1 = m_Cache[tg].End2 = -1;
	110
	111	if (wxIsCDATAElement(tagBuffer))
	112	{
	113	// store the orig pos in case we are missing the closing
	114	// tag (see below)
	115	wxInt32 old_pos = pos;
	116	bool foundCloseTag = false;
	117
	118	// find next matching tag
	119	int tag_len = wxStrlen(tagBuffer);
	120	while (pos < lng)
	121	{
	122	// find the ending tag
	123	while (pos + 1 < lng &&
	124	(src[pos] != '<' \|\| src[pos+1] != '/'))
	125	++pos;
	126	if (src[pos] == '<')
	127	++pos;
	128
	129	// see if it matches
	130	int match_pos = 0;
	131	while (pos < lng && match_pos < tag_len && src[pos] != '>' && src[pos] != '<') {
	132	// cast to wxChar needed to suppress warning in
	133	// Unicode build
	134	if ((wxChar)wxToupper(src[pos]) == tagBuffer[match_pos]) {
	135	++match_pos;
	136	}
	137	else if (src[pos] == wxT(' ') \|\| src[pos] == wxT('\n') \|\|
	138	src[pos] == wxT('\r') \|\| src[pos] == wxT('\t')) {
	139	// need to skip over these
	140	}
	141	else {
	142	match_pos = 0;
	143	}
	144	++pos;
	145	}
	146
	147	// found a match
	148	if (match_pos == tag_len)
	149	{
	150	pos = pos - tag_len - 3;
	151	foundCloseTag = true;
	152	break;
153	}
154	else // keep looking for the closing tag
155	{
156	++pos;
157	}
158	}
159	if (!foundCloseTag)
160	{
161	// we didn't find closing tag; this means the markup
162	// is incorrect and the best thing we can do is to
163	// ignore the unclosed tag and continue parsing as if
164	// it didn't exist:
165	pos = old_pos;
166	}
167	}
168	}
169	}
170
171	pos++;
172	}
173
174	// ok, we're done, now we'll free .Name members of cache - we don't need it anymore:
175	for (int i = 0; i < m_CacheSize; i++)
176	{
177	delete[] m_Cache[i].Name;
178	m_Cache[i].Name = NULL;
179	}
180	}
181
182	void wx28HtmlTagsCache::QueryTag(int at, int* end1, int* end2)
183	{
184	if (m_Cache == NULL) return;
185	if (m_Cache[m_CachePos].Key != at)
186	{
187	int delta = (at < m_Cache[m_CachePos].Key) ? -1 : 1;
188	do
189	{
190	if ( m_CachePos < 0 \|\| m_CachePos == m_CacheSize )
191	{
192	// something is very wrong with HTML, give up by returning an
193	// impossibly large value which is going to be ignored by the
194	// caller
195	*end1 =
196	*end2 = INT_MAX;
197	return;
198	}
199
200	m_CachePos += delta;
201	}
202	while (m_Cache[m_CachePos].Key != at);
203	}
204	*end1 = m_Cache[m_CachePos].End1;
205	*end2 = m_Cache[m_CachePos].End2;
206	}
207
208
209
210
211	//-----------------------------------------------------------------------------
212	// wx28HtmlTag
213	//-----------------------------------------------------------------------------
214
215	IMPLEMENT_CLASS(wx28HtmlTag,wxObject)
216
217	wx28HtmlTag::wx28HtmlTag(wx28HtmlTag *parent,
218	const wxString& source, int pos, int end_pos,
219	wx28HtmlTagsCache *cache,
220	wx28HtmlEntitiesParser *entParser) : wxObject()
221	{
222	/* Setup DOM relations */
223
224	m_Next = NULL;
225	m_FirstChild = m_LastChild = NULL;
226	m_Parent = parent;
227	if (parent)
228	{
229	m_Prev = m_Parent->m_LastChild;
230	if (m_Prev == NULL)
231	m_Parent->m_FirstChild = this;
232	else
233	m_Prev->m_Next = this;
234	m_Parent->m_LastChild = this;
235	}
236	else
237	m_Prev = NULL;
238
239	/* Find parameters and their values: */
240
241	int i;
242	wxChar c;
243
244	// fill-in name, params and begin pos:
245	i = pos+1;
246
247	// find tag's name and convert it to uppercase:
248	while ((i < end_pos) &&
249	((c = source[i++]) != wxT(' ') && c != wxT('\r') &&
250	c != wxT('\n') && c != wxT('\t') &&
251	c != wxT('>')))
252	{
253	if ((c >= wxT('a')) && (c <= wxT('z')))
254	c -= (wxT('a') - wxT('A'));
255	m_Name << c;
256	}
257
258	// if the tag has parameters, read them and "normalize" them,
259	// i.e. convert to uppercase, replace whitespaces by spaces and
260	// remove whitespaces around '=':
261	if (source[i-1] != wxT('>'))
262	{
263	#define IS_WHITE(c) (c == wxT(' ') \|\| c == wxT('\r') \|\| \
264	c == wxT('\n') \|\| c == wxT('\t'))
265	wxString pname, pvalue;
266	wxChar quote;
267	enum
268	{
269	ST_BEFORE_NAME = 1,
270	ST_NAME,
271	ST_BEFORE_EQ,
272	ST_BEFORE_VALUE,
273	ST_VALUE
274	} state;
275
276	quote = 0;
277	state = ST_BEFORE_NAME;
278	while (i < end_pos)
279	{
280	c = source[i++];
281
282	if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
283	{
284	if (state == ST_BEFORE_EQ \|\| state == ST_NAME)
285	{
286	m_ParamNames.Add(pname);
287	m_ParamValues.Add(wxEmptyString);
288	}
289	else if (state == ST_VALUE && quote == 0)
290	{
291	m_ParamNames.Add(pname);
292	if (entParser)
293	m_ParamValues.Add(entParser->Parse(pvalue));
294	else
295	m_ParamValues.Add(pvalue);
296	}
297	break;
298	}
299	switch (state)
300	{
301	case ST_BEFORE_NAME:
302	if (!IS_WHITE(c))
303	{
304	pname = c;
305	state = ST_NAME;
306	}
307	break;
308	case ST_NAME:
309	if (IS_WHITE(c))
310	state = ST_BEFORE_EQ;
311	else if (c == wxT('='))
312	state = ST_BEFORE_VALUE;
313	else
314	pname << c;
315	break;
316	case ST_BEFORE_EQ:
317	if (c == wxT('='))
318	state = ST_BEFORE_VALUE;
319	else if (!IS_WHITE(c))
320	{
321	m_ParamNames.Add(pname);
322	m_ParamValues.Add(wxEmptyString);
323	pname = c;
324	state = ST_NAME;
325	}
326	break;
327	case ST_BEFORE_VALUE:
328	if (!IS_WHITE(c))
329	{
330	if (c == wxT('"') \|\| c == wxT('\''))
331	quote = c, pvalue = wxEmptyString;
332	else
333	quote = 0, pvalue = c;
334	state = ST_VALUE;
335	}
336	break;
337	case ST_VALUE:
338	if ((quote != 0 && c == quote) \|\|
339	(quote == 0 && IS_WHITE(c)))
340	{
341	m_ParamNames.Add(pname);
342	if (quote == 0)
343	{
344	// VS: backward compatibility, no real reason,
345	// but wxHTML code relies on this... :(
346	pvalue.MakeUpper();
347	}
348	if (entParser)
349	m_ParamValues.Add(entParser->Parse(pvalue));
350	else
351	m_ParamValues.Add(pvalue);
352	state = ST_BEFORE_NAME;
353	}
354	else
355	pvalue << c;
356	break;
357	}
358	}
359
360	#undef IS_WHITE
361	}
362	m_Begin = i;
363
364	cache->QueryTag(pos, &m_End1, &m_End2);
365	if (m_End1 > end_pos) m_End1 = end_pos;
366	if (m_End2 > end_pos) m_End2 = end_pos;
367	}
368
369	wx28HtmlTag::~wx28HtmlTag()
370	{
371	wx28HtmlTag t1, t2;
372	t1 = m_FirstChild;
373	while (t1)
374	{
375	t2 = t1->GetNextSibling();
376	delete t1;
377	t1 = t2;
378	}
379	}
380
381	bool wx28HtmlTag::HasParam(const wxString& par) const
382	{
383	return (m_ParamNames.Index(par, false) != wxNOT_FOUND);
384	}
385
386	wxString wx28HtmlTag::GetParam(const wxString& par, bool with_commas) const
387	{
388	int index = m_ParamNames.Index(par, false);
389	if (index == wxNOT_FOUND)
390	return wxEmptyString;
391	if (with_commas)
392	{
393	// VS: backward compatibility, seems to be never used by wxHTML...
394	wxString s;
395	s << wxT('"') << m_ParamValues[index] << wxT('"');
396	return s;
397	}
398	else
399	return m_ParamValues[index];
400	}
401
402	int wx28HtmlTag::ScanParam(const wxString& par,
403	const wxChar *format,
404	void *param) const
405	{
406	wxString parval = GetParam(par);
407	return wxSscanf(parval, format, param);
408	}
409
410	bool wx28HtmlTag::GetParamAsColour(const wxString& par, wxColour *clr) const
411	{
9a83f860	412	wxCHECK_MSG( clr, false, wxT("invalid colour argument") );
c21faa0a VS	413
	414	wxString str = GetParam(par);
	415
	416	// handle colours defined in HTML 4.0 first:
9a83f860	417	if (str.length() > 1 && str[0] != wxT('#'))
c21faa0a VS	418	{
	419	#define HTML_COLOUR(name, r, g, b) \
	420	if (str.IsSameAs(wxT(name), false)) \
	421	{ clr->Set(r, g, b); return true; }
	422	HTML_COLOUR("black", 0x00,0x00,0x00)
	423	HTML_COLOUR("silver", 0xC0,0xC0,0xC0)
	424	HTML_COLOUR("gray", 0x80,0x80,0x80)
	425	HTML_COLOUR("white", 0xFF,0xFF,0xFF)
	426	HTML_COLOUR("maroon", 0x80,0x00,0x00)
	427	HTML_COLOUR("red", 0xFF,0x00,0x00)
	428	HTML_COLOUR("purple", 0x80,0x00,0x80)
	429	HTML_COLOUR("fuchsia", 0xFF,0x00,0xFF)
	430	HTML_COLOUR("green", 0x00,0x80,0x00)
	431	HTML_COLOUR("lime", 0x00,0xFF,0x00)
	432	HTML_COLOUR("olive", 0x80,0x80,0x00)
	433	HTML_COLOUR("yellow", 0xFF,0xFF,0x00)
	434	HTML_COLOUR("navy", 0x00,0x00,0x80)
	435	HTML_COLOUR("blue", 0x00,0x00,0xFF)
	436	HTML_COLOUR("teal", 0x00,0x80,0x80)
	437	HTML_COLOUR("aqua", 0x00,0xFF,0xFF)
	438	#undef HTML_COLOUR
	439	}
	440
	441	// then try to parse #rrggbb representations or set from other well
	442	// known names (note that this doesn't strictly conform to HTML spec,
	443	// but it doesn't do real harm -- but it must be done after the standard
	444	// colors are handled above):
	445	if (clr->Set(str))
	446	return true;
	447
	448	return false;
	449	}
	450
	451	bool wx28HtmlTag::GetParamAsInt(const wxString& par, int *clr) const
	452	{
	453	if ( !HasParam(par) )
	454	return false;
	455
	456	long i;
	457	if ( !GetParam(par).ToLong(&i) )
	458	return false;
	459
	460	*clr = (int)i;
	461	return true;
	462	}
	463
	464	wxString wx28HtmlTag::GetAllParams() const
	465	{
	466	// VS: this function is for backward compatibility only,
	467	// never used by wxHTML
	468	wxString s;
	469	size_t cnt = m_ParamNames.GetCount();
	470	for (size_t i = 0; i < cnt; i++)
	471	{
	472	s << m_ParamNames[i];
	473	s << wxT('=');
	474	if (m_ParamValues[i].Find(wxT('"')) != wxNOT_FOUND)
	475	s << wxT('\'') << m_ParamValues[i] << wxT('\'');
	476	else
	477	s << wxT('"') << m_ParamValues[i] << wxT('"');
	478	}
	479	return s;
	480	}
	481
482	wx28HtmlTag *wx28HtmlTag::GetFirstSibling() const
483	{
484	if (m_Parent)
485	return m_Parent->m_FirstChild;
486	else
487	{
488	wx28HtmlTag cur = (wx28HtmlTag)this;
489	while (cur->m_Prev)
490	cur = cur->m_Prev;
491	return cur;
492	}
493	}
494
495	wx28HtmlTag *wx28HtmlTag::GetLastSibling() const
496	{
497	if (m_Parent)
498	return m_Parent->m_LastChild;
499	else
500	{
501	wx28HtmlTag cur = (wx28HtmlTag)this;
502	while (cur->m_Next)
503	cur = cur->m_Next;
504	return cur;
505	}
506	}
507
508	wx28HtmlTag *wx28HtmlTag::GetNextTag() const
509	{
510	if (m_FirstChild) return m_FirstChild;
511	if (m_Next) return m_Next;
512	wx28HtmlTag *cur = m_Parent;
513	if (!cur) return NULL;
514	while (cur->m_Parent && !cur->m_Next)
515	cur = cur->m_Parent;
516	return cur->m_Next;
517	}