X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/5e8e25e72968ad2d10bbf290291b84de35c7e18b..8a1459665d29fb62f9a89571bd6e0a2d27b05cf0:/src/html/htmltag.cpp diff --git a/src/html/htmltag.cpp b/src/html/htmltag.cpp index 1cfeb5a5b6..6646ac1ed7 100644 --- a/src/html/htmltag.cpp +++ b/src/html/htmltag.cpp @@ -1,35 +1,35 @@ ///////////////////////////////////////////////////////////////////////////// -// 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 // for vsscanf #include - //----------------------------------------------------------------------------- // wxHtmlTagsCache //----------------------------------------------------------------------------- @@ -38,101 +38,236 @@ 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 +{ +}; -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 (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')) - { - dummy[i] = src[pos++]; - if ((dummy[i] >= wxT('a')) && (dummy[i] <= wxT('z'))) dummy[i] -= (wxT('a') - wxT('A')); - i++; + if (*pos == wxT('<')) // tag found: + { + // don't cache comment tags + if ( wxHtmlParser::SkipCommentTag(pos, source.end()) ) + continue; + + size_t tg = Cache().size(); + Cache().push_back(wxHtmlCacheItem()); + + wxString::const_iterator stpos = pos++; + Cache()[tg].Key = stpos; + + int i; + for ( i = 0; + pos < end && i < (int)WXSIZEOF(tagBuffer) - 1 && + *pos != wxT('>') && !wxIsspace(*pos); + ++i, ++pos ) + { + tagBuffer[i] = (wxChar)wxToupper(*pos); } - dummy[i] = 0; - m_Cache[tg].Name = new wxChar[i+1]; - memcpy(m_Cache[tg].Name, dummy, (i+1)*sizeof(wxChar)); + tagBuffer[i] = wxT('\0'); + + Cache()[tg].Name = new wxChar[i+1]; + memcpy(Cache()[tg].Name, tagBuffer, (i+1)*sizeof(wxChar)); - while (pos < lng && src[pos] != wxT('>')) pos++; + while (pos < end && *pos != wxT('>')) + ++pos; - if (src[stpos+1] == wxT('/')) // ending tag: - { - m_Cache[tg].End1 = m_Cache[tg].End2 = -2; + 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--) - if ((m_Cache[i].End1 == -1) && (wxStrcmp(m_Cache[i].Name, dummy+1) == 0)) - { - m_Cache[i].End1 = stpos; - m_Cache[i].End2 = pos + 1; + { + 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; } + } } - else - { - m_Cache[tg].End1 = m_Cache[tg].End2 = -1; + else + { + Cache()[tg].type = wxHtmlCacheItem::Type_NoMatchingEndingTag; + + if (wxIsCDATAElement(tagBuffer)) + { + // 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) + { + 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 (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()) + return; + + if (Cache()[m_CachePos].Key != at) { - int delta = (at < m_Cache[m_CachePos].Key) ? -1 : 1; - do {m_CachePos += delta;} while (m_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 ) + { + 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; } @@ -142,173 +277,349 @@ 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] == '/') { 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++]) != ' ' && c != '\r' && c != '\n' && c != '\t' && - c != '>')) - { - if ((c >= 'a') && (c <= 'z')) c -= ('a' - 'A'); - m_Name += c; + 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 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] != '>') - while ((i < end_pos) && ((c = source[i++]) != '>')) - { - if ((c >= 'a') && (c <= 'z')) c -= ('a' - 'A'); - if (c == '\r' || c == '\n' || c == '\t') c = ' '; // make future parsing a bit simpler - m_Params += c; - if (c == '"') - { - // remove spaces around the '=' character: - if (m_Params.Length() > 1 && - m_Params[m_Params.Length()-2] == ' ') + 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) + { + c = *(i++); + + if (c == wxT('>') && !(state == ST_VALUE && quote != 0)) + { + 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() == ' ') - m_Params.RemoveLast(); - m_Params += '"'; + m_ParamNames.Add(pname); + if (entParser) + m_ParamValues.Add(entParser->Parse(pvalue)); + else + m_ParamValues.Add(pvalue); } - while ((i < end_pos) && (source[i++] == ' ')) {} - if (i < end_pos) i--; - - // ...and copy the value to m_Params: - while ((i < end_pos) && ((c = source[i++]) != '"')) m_Params += c; - m_Params += c; + break; } - else if (c == '\'') - { - while ((i < end_pos) && ((c = source[i++]) != '\'')) m_Params += c; - m_Params += c; + 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(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" }, + }; + + 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; + 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) return TRUE; - if (*st2 == 0) return FALSE; - if (*p2 != *st2) p2 = p; - if (*p2 == *st2) p2++; - if (*st2 == ' ') p2 = p; - else if (*st2 == '=') - { - p2 = p; - while (*st2 != ' ') - { - if (*st2 == '"') - { - st2++; - while (*st2 != '"') 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; - bool comma; - char comma_char; - - if (*st == 0) return ""; - if (*p == 0) return ""; - 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) // found - { - wxString fnd = ""; - st2++; // '=' character - comma = FALSE; - comma_char = '\0'; - if (!with_commas && (*(st2) == '"')) - { - st2++; - comma = TRUE; - comma_char = '"'; - } - else if (!with_commas && (*(st2) == '\'')) - { - st2++; - comma = TRUE; - comma_char = '\''; - } - - while (*st2 != 0) - { - if (comma && *st2 == comma_char) comma = FALSE; - else if ((*st2 == ' ') && (!comma)) break; - fnd += (*(st2++)); - } - if (!with_commas && (*(st2-1) == comma_char)) fnd.RemoveLast(); - return fnd; - } - if (*st2 == 0) return ""; - if (*p2 != *st2) p2 = p; - if (*p2 == *st2) p2++; - if (*st2 == ' ') p2 = p; - else if (*st2 == '=') - { - p2 = p; - while (*st2 != ' ') - { - if (*st2 == '"') - { - st2++; - while (*st2 != '"') st2++; - } - else if (*st2 == '\'') - { - st2++; - while (*st2 != '\'') 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; +} + +wxString wxHtmlTag::GetAllParams() const { - wxString parval = GetParam(par); - return wxSscanf((const wxChar*)parval, format, param); + // 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