| 1 |
cycrow |
1 |
/*
|
|
|
2 |
defines token_stream - class that acts as high level wrapper around
|
|
|
3 |
bod_text_parser. It mimics behaviour of stream.
|
|
|
4 |
|
|
|
5 |
There is potential hazard when loading strings because only finite amount of
|
|
|
6 |
tokens is loaded into memory. If number of tokens forming the string is greater
|
|
|
7 |
than the token unload threshold, seeking back to beginning of string will fail
|
|
|
8 |
|
|
|
9 |
see bod_bob_parser::loadString
|
|
|
10 |
|
|
|
11 |
*/
|
|
|
12 |
|
|
|
13 |
#ifndef TOKEN_STREAM_INCLUDED
|
|
|
14 |
#define TOKEN_STREAM_INCLUDED
|
|
|
15 |
|
|
|
16 |
#include "bob_dom_stream.h"
|
|
|
17 |
#include "bod_text_parser.h"
|
|
|
18 |
|
|
|
19 |
#include "../common/ext_list.h"
|
|
|
20 |
|
|
|
21 |
class token_stream : public ext::stream_base
|
|
|
22 |
{
|
|
|
23 |
public:
|
|
|
24 |
typedef bod_text_parser::token token;
|
|
|
25 |
typedef bod_text_parser::TokenList::size_type size_type;
|
|
|
26 |
typedef int offset_type;
|
|
|
27 |
|
|
|
28 |
private:
|
|
|
29 |
typedef bod_text_parser::TokenList::iterator iterator;
|
|
|
30 |
|
|
|
31 |
bod_text_parser p;
|
|
|
32 |
iterator m_end;
|
|
|
33 |
iterator m_pos;
|
|
|
34 |
size_type m_left;
|
|
|
35 |
size_type m_right;
|
|
|
36 |
bool m_bIgnoreRemarks;
|
|
|
37 |
|
|
|
38 |
void advance(iterator &it, offset_type off)
|
|
|
39 |
{
|
|
|
40 |
if(off > 0){
|
|
|
41 |
for(offset_type i=0; i < off; i++)
|
|
|
42 |
++it;
|
|
|
43 |
}
|
|
|
44 |
else{
|
|
|
45 |
for(offset_type i=off; i < 0; i++)
|
|
|
46 |
--it;
|
|
|
47 |
}
|
|
|
48 |
}
|
|
|
49 |
|
|
|
50 |
void parse(size_t limit)
|
|
|
51 |
{
|
|
|
52 |
m_right+=p.parseBuffer(limit);
|
|
|
53 |
}
|
|
|
54 |
|
|
|
55 |
void unload(size_t count)
|
|
|
56 |
{
|
|
|
57 |
if(count > m_left) count=m_left;
|
|
|
58 |
m_left-=count;
|
|
|
59 |
while(count--){
|
|
|
60 |
delete *(p.tokens.begin());
|
|
|
61 |
p.tokens.erase(p.tokens.begin());
|
|
|
62 |
}
|
|
|
63 |
}
|
|
|
64 |
|
|
|
65 |
public:
|
|
|
66 |
static const int tokenTreshold = 200; // how much tokens to unload
|
|
|
67 |
static const int tokenUnloadTreshold = 2 * 200; // when to start unloading tokens
|
|
|
68 |
|
|
|
69 |
token_stream() { m_bIgnoreRemarks=false; }
|
|
|
70 |
|
|
|
71 |
void rdbuffer(char *pszBuffer, size_t size)
|
|
|
72 |
{
|
|
|
73 |
p.ignoreRemarks(ignoreRemarks());
|
|
|
74 |
p.preParseBuffer(pszBuffer, size);
|
|
|
75 |
m_left=0;
|
|
|
76 |
m_right=p.parseBuffer(1);
|
|
|
77 |
m_pos=p.tokens.begin();
|
|
|
78 |
m_end=p.tokens.end();
|
|
|
79 |
clear(p.tokens.size() ? goodbit : badbit);
|
|
|
80 |
}
|
|
|
81 |
|
|
|
82 |
token* tok() { return good() ? *m_pos : 0; }
|
|
|
83 |
|
|
|
84 |
size_type tell() const { return m_left; }
|
|
|
85 |
size_type avail() const { return m_right; }
|
|
|
86 |
size_type size() const { return p.tokens.size(); }
|
|
|
87 |
|
|
|
88 |
token_stream& operator++() { advance(); return *this; }
|
|
|
89 |
token_stream& operator--() { advance(-1); return *this; }
|
|
|
90 |
|
|
|
91 |
bool ignoreRemarks() const { return m_bIgnoreRemarks; }
|
|
|
92 |
void ignoreRemarks(bool ignore) { m_bIgnoreRemarks=ignore; }
|
|
|
93 |
|
|
|
94 |
bool advance(offset_type offset=1)
|
|
|
95 |
{
|
|
|
96 |
if(offset > (offset_type)m_right)
|
|
|
97 |
parse(offset < tokenTreshold ? tokenTreshold : offset);
|
|
|
98 |
|
|
|
99 |
if(offset > (offset_type)m_right || (offset_type)m_left + offset < 0)
|
|
|
100 |
setstate(failbit);
|
|
|
101 |
else{
|
|
|
102 |
m_left+=offset; m_right-=offset;
|
|
|
103 |
if(m_right==0)
|
|
|
104 |
parse(tokenTreshold);
|
|
|
105 |
|
|
|
106 |
advance(m_pos, offset);
|
|
|
107 |
|
|
|
108 |
if(m_pos==m_end)
|
|
|
109 |
setstate(eofbit);
|
|
|
110 |
else
|
|
|
111 |
clear((state)(rdstate() & ~eofbit));
|
|
|
112 |
}
|
|
|
113 |
if(m_left > tokenUnloadTreshold)
|
|
|
114 |
unload(tokenTreshold);
|
|
|
115 |
return good();
|
|
|
116 |
}
|
|
|
117 |
|
|
|
118 |
token* previous()
|
|
|
119 |
{
|
|
|
120 |
if(m_left==0)
|
|
|
121 |
return 0;
|
|
|
122 |
else
|
|
|
123 |
return *(m_pos-1);
|
|
|
124 |
}
|
|
|
125 |
};
|
|
|
126 |
|
|
|
127 |
#endif // !defined(TOKEN_STREAM_INCLUDED)
|