/////////////////////////////////////////////////////////////////////////////
-// 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
//-----------------------------------------------------------------------------
{
// 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 )
{
- 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'))
+ // 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:
+ {
+ 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 )
+ {
+ 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)
{
- m_Cache[i].End1 = stpos;
- m_Cache[i].End2 = pos + 1;
+ 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);
}
}
-void wxHtmlTagsCache::QueryTag(int at, int* end1, int* end2)
+wxHtmlTagsCache::~wxHtmlTagsCache()
{
- if (m_Cache == NULL) return;
- if (m_Cache[m_CachePos].Key != at)
+ delete m_Cache;
+}
+
+void wxHtmlTagsCache::QueryTag(const wxString::const_iterator& at,
+ const wxString::const_iterator& inputEnd,
+ wxString::const_iterator *end1,
+ wxString::const_iterator *end2,
+ bool *hasEnding)
+{
+ if (Cache().empty())
{
- int delta = (at < m_Cache[m_CachePos].Key) ? -1 : 1;
- do
- {
- m_CachePos += delta;
+ *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 (m_Cache[m_CachePos].Key != at);
+ while (Cache()[m_CachePos].Key != at);
+ }
+
+ switch ( Cache()[m_CachePos].type )
+ {
+ 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;
}
// wxHtmlTag
//-----------------------------------------------------------------------------
-IMPLEMENT_CLASS(wxHtmlTag,wxObject)
-
-wxHtmlTag::wxHtmlTag(const wxString& source, int pos, int end_pos,
+wxHtmlTag::wxHtmlTag(wxHtmlTag *parent,
+ const wxString *source,
+ const wxString::const_iterator& pos,
+ const wxString::const_iterator& end_pos,
wxHtmlTagsCache *cache,
- wxHtmlEntitiesParser *entParser) : wxObject()
+ wxHtmlEntitiesParser *entParser)
{
- int i;
- wxChar 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:
- 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') &&
+ while ((i < end_pos) &&
+ ((c = *(i++)) != wxT(' ') && c != wxT('\r') &&
c != wxT('\n') && c != wxT('\t') &&
- c != wxT('>')))
+ c != wxT('>') && c != wxT('/')))
{
- if ((c >= wxT('a')) && (c <= wxT('z')))
+ 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('>'))
+ 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
+ enum
{
- ST_BEFORE_NAME = 1,
+ 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++];
+ c = *(i++);
- if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
+ if (c == wxT('>') && !(state == ST_VALUE && quote != 0))
{
if (state == ST_BEFORE_EQ || state == ST_NAME)
{
m_ParamNames.Add(pname);
- m_ParamValues.Add(wxEmptyString);
+ m_ParamValues.Add(wxGetEmptyString());
}
else if (state == ST_VALUE && quote == 0)
{
else if (!IS_WHITE(c))
{
m_ParamNames.Add(pname);
- m_ParamValues.Add(wxEmptyString);
+ m_ParamValues.Add(wxGetEmptyString());
pname = c;
state = ST_NAME;
}
if (!IS_WHITE(c))
{
if (c == wxT('"') || c == wxT('\''))
- quote = c, pvalue = wxEmptyString;
+ quote = c, pvalue = wxGetEmptyString();
else
quote = 0, pvalue = c;
state = ST_VALUE;
break;
}
}
-
+
#undef IS_WHITE
- }
- m_Begin = i;
+ }
+ 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));
+ }
+ }
+}
- 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;
+wxHtmlTag::~wxHtmlTag()
+{
+ wxHtmlTag *t1, *t2;
+ t1 = m_FirstChild;
+ while (t1)
+ {
+ t2 = t1->GetNextSibling();
+ delete t1;
+ t1 = t2;
+ }
}
bool wxHtmlTag::HasParam(const wxString& par) const
{
- return (m_ParamNames.Index(par, FALSE) != wxNOT_FOUND);
+ return (m_ParamNames.Index(par, false) != wxNOT_FOUND);
}
-wxString wxHtmlTag::GetParam(const wxString& par, bool with_commas) const
+wxString wxHtmlTag::GetParam(const wxString& par, bool with_quotes) const
{
- int index = m_ParamNames.Index(par, FALSE);
+ int index = m_ParamNames.Index(par, false);
if (index == wxNOT_FOUND)
- return wxEmptyString;
- if (with_commas)
+ return wxGetEmptyString();
+ if (with_quotes)
{
// VS: backward compatibility, seems to be never used by wxHTML...
wxString s;
return m_ParamValues[index];
}
-int wxHtmlTag::ScanParam(const wxString& par, wxChar *format, void *param) const
+int wxHtmlTag::ScanParam(const wxString& par,
+ const char *format,
+ void *param) const
{
wxString parval = GetParam(par);
return wxSscanf(parval, format, param);
}
-bool wxHtmlTag::GetParamAsColour(const wxString& par, wxColour *clr) const
+int wxHtmlTag::ScanParam(const wxString& par,
+ const wchar_t *format,
+ void *param) const
{
- wxString str = GetParam(par);
-
- if (str.IsEmpty()) return FALSE;
- if (str.GetChar(0) == wxT('#'))
- {
- unsigned long tmp;
- if (ScanParam(par, wxT("#%lX"), &tmp) != 1)
- return FALSE;
- *clr = wxColour((unsigned char)((tmp & 0xFF0000) >> 16),
- (unsigned char)((tmp & 0x00FF00) >> 8),
- (unsigned char)(tmp & 0x0000FF));
- return TRUE;
- }
- else
+ wxString parval = GetParam(par);
+ return wxSscanf(parval, format, param);
+}
+
+/* static */
+bool wxHtmlTag::ParseAsColour(const wxString& str, wxColour *clr)
+{
+ wxCHECK_MSG( clr, false, wxT("invalid colour argument") );
+
+ // handle colours defined in HTML 4.0 first:
+ if (str.length() > 1 && str[0] != wxT('#'))
{
- // Handle colours defined in HTML 4.0:
- #define HTML_COLOUR(name,r,g,b) \
- if (str.IsSameAs(wxT(name), FALSE)) \
- { *clr = wxColour(r,g,b); return TRUE; }
+ #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("teal", 0x00,0x80,0x80)
HTML_COLOUR("aqua", 0x00,0xFF,0xFF)
#undef HTML_COLOUR
- return FALSE;
}
+
+ // 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;
+ if ( !HasParam(par) )
+ return false;
+
long i;
- bool succ = GetParam(par).ToLong(&i);
+ if ( !GetParam(par).ToLong(&i) )
+ return false;
+
*clr = (int)i;
- return succ;
+ return true;
+}
+
+bool
+wxHtmlTag::GetParamAsIntOrPercent(const wxString& par,
+ int* value,
+ bool& isPercent) const
+{
+ 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 compatiblity only,
+ // VS: this function is for backward compatibility only,
// never used by wxHTML
wxString s;
size_t cnt = m_ParamNames.GetCount();
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