|
a/src/internfile/myhtmlparse.h |
|
b/src/internfile/myhtmlparse.h |
1 |
=======
|
|
|
2 |
/* myhtmlparse.h: subclass of HtmlParser for extracting text
|
1 |
/* myhtmlparse.h: subclass of HtmlParser for extracting text
|
3 |
*
|
2 |
*
|
4 |
* ----START-LICENCE----
|
3 |
* ----START-LICENCE----
|
5 |
* Copyright 1999,2000,2001 BrightStation PLC
|
4 |
* Copyright 1999,2000,2001 BrightStation PLC
|
6 |
* Copyright 2002,2003,2004 Olly Betts
|
5 |
* Copyright 2002,2003,2004 Olly Betts
|
|
... |
|
... |
33 |
public:
|
32 |
public:
|
34 |
bool in_script_tag;
|
33 |
bool in_script_tag;
|
35 |
bool in_style_tag;
|
34 |
bool in_style_tag;
|
36 |
bool pending_space;
|
35 |
bool pending_space;
|
37 |
string title, sample, keywords, dump;
|
36 |
string title, sample, keywords, dump;
|
|
|
37 |
string ocharset; // This is the charset our user thinks the doc was
|
|
|
38 |
string charset; // This is the charset it was supposedly converted to
|
|
|
39 |
string doccharset; // Set this to value of charset parameter in header
|
38 |
bool indexing_allowed;
|
40 |
bool indexing_allowed;
|
39 |
void process_text(const string &text);
|
41 |
void process_text(const string &text);
|
40 |
void opening_tag(const string &tag, const map<string,string> &p);
|
42 |
void opening_tag(const string &tag, const map<string,string> &p);
|
41 |
void closing_tag(const string &tag);
|
43 |
void closing_tag(const string &tag);
|
42 |
MyHtmlParser() :
|
44 |
MyHtmlParser() :
|