Amebis/wxWidgets
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

1) wxHTML now outputs unrecognized entities verbosely

Browse Source 
2) added string preallocation to the parser


git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@11854 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775
This commit is contained in:
Václav Slavík
2001-10-05 22:03:58 +00:00
parent a350a48841
commit 470252df3c
1 changed files with 16 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
src/html/htmlpars.cpp
View File

@@ -318,7 +318,7 @@ void wxHtmlParser::AddTagHandler(wxHtmlTagHandler *handler)
wxStringTokenizer tokenizer(s, wxT(", ")); wxStringTokenizer tokenizer(s, wxT(", "));
while (tokenizer.HasMoreTokens()) while (tokenizer.HasMoreTokens())
m_HandlersHash.Put(tokenizer.NextToken(), handler); m_HandlersHash.Put(tokenizer.GetNextToken(), handler);
if (m_HandlersList.IndexOf(handler) == wxNOT_FOUND) if (m_HandlersList.IndexOf(handler) == wxNOT_FOUND)
m_HandlersList.Append(handler); m_HandlersList.Append(handler);
@@ -341,7 +341,7 @@ void wxHtmlParser::PushTagHandler(wxHtmlTagHandler *handler, wxString tags)
while (tokenizer.HasMoreTokens()) while (tokenizer.HasMoreTokens())
{ {
key = tokenizer.NextToken(); key = tokenizer.GetNextToken();
m_HandlersHash.Delete(key); m_HandlersHash.Delete(key);
m_HandlersHash.Put(key, handler); m_HandlersHash.Put(key, handler);
} }
@@ -448,6 +448,8 @@ wxString wxHtmlEntitiesParser::Parse(const wxString& input)
const wxChar *c, *last; const wxChar *c, *last;
const wxChar *in_str = input.c_str(); const wxChar *in_str = input.c_str();
wxString output; wxString output;
output.reserve(input.length());
for (c = in_str, last = in_str; *c != wxT('\0'); c++) for (c = in_str, last = in_str; *c != wxT('\0'); c++)
{ {
@@ -456,8 +458,11 @@ wxString wxHtmlEntitiesParser::Parse(const wxString& input)
if (c - last > 0) if (c - last > 0)
output.append(last, c - last); output.append(last, c - last);
if (++c == wxT('\0')) break; if (++c == wxT('\0')) break;
wxString entity; wxString entity;
const wxChar *ent_s = c; const wxChar *ent_s = c;
wxChar entity_char;
for (; (*c >= wxT('a') && *c <= wxT('z')) || for (; (*c >= wxT('a') && *c <= wxT('z')) ||
(*c >= wxT('A') && *c <= wxT('Z')) || (*c >= wxT('A') && *c <= wxT('Z')) ||
(*c >= wxT('0') && *c <= wxT('9')) || (*c >= wxT('0') && *c <= wxT('9')) ||
@@ -465,7 +470,14 @@ wxString wxHtmlEntitiesParser::Parse(const wxString& input)
entity.append(ent_s, c - ent_s); entity.append(ent_s, c - ent_s);
if (*c != wxT(';')) c--; if (*c != wxT(';')) c--;
last = c+1; last = c+1;
output << GetEntityChar(entity); entity_char = GetEntityChar(entity);
if (entity_char)
output << entity_char;
else
{
output.append(ent_s-1, c-ent_s+2);
wxLogDebug(wxT("Unrecognized HTML entity: '%s'"), entity.c_str());
}
} }
} }
if (*last != wxT('\0')) if (*last != wxT('\0'))
@@ -795,7 +807,7 @@ wxChar wxHtmlEntitiesParser::GetEntityChar(const wxString& entity)
} }
if (code == 0) if (code == 0)
return wxT('?'); return 0;
else else
return GetCharForCode(code); return GetCharForCode(code);
} }