1 |
/* DLGLexerBase.h |
2 |
* |
3 |
* SOFTWARE RIGHTS |
4 |
* |
5 |
* We reserve no LEGAL rights to the Purdue Compiler Construction Tool |
6 |
* Set (PCCTS) -- PCCTS is in the public domain. An individual or |
7 |
* company may do whatever they wish with source code distributed with |
8 |
* PCCTS or the code generated by PCCTS, including the incorporation of |
9 |
* PCCTS, or its output, into commerical software. |
10 |
* |
11 |
* We encourage users to develop software with PCCTS. However, we do ask |
12 |
* that credit is given to us for developing PCCTS. By "credit", |
13 |
* we mean that if you incorporate our source code into one of your |
14 |
* programs (commercial product, research project, or otherwise) that you |
15 |
* acknowledge this fact somewhere in the documentation, research report, |
16 |
* etc... If you like PCCTS and have developed a nice tool with the |
17 |
* output, please mention that you developed it using PCCTS. In |
18 |
* addition, we ask that this header remain intact in our source code. |
19 |
* As long as these guidelines are kept, we expect to continue enhancing |
20 |
* this system and expect to make other tools available as they are |
21 |
* completed. |
22 |
* |
23 |
* ANTLR 1.33 |
24 |
* Terence Parr |
25 |
* Parr Research Corporation |
26 |
* with Purdue University and AHPCRC, University of Minnesota |
27 |
* 1989-2000 |
28 |
*/ |
29 |
|
30 |
#ifndef DLGX_H |
31 |
#define DLGX_H |
32 |
|
33 |
#include "pcctscfg.h" |
34 |
#include "pccts_stdio.h" |
35 |
|
36 |
PCCTS_NAMESPACE_STD |
37 |
|
38 |
#include ATOKEN_H |
39 |
#include ATOKENSTREAM_H |
40 |
|
41 |
class ANTLRParser; // MR1 |
42 |
|
43 |
/* must define what a char looks like; can make this a class too */ |
44 |
typedef char DLGChar; |
45 |
|
46 |
/* Can have it as a class too: (ack this looks weird; is it right?) |
47 |
class DllExportPCCTS DLGChar { |
48 |
private: |
49 |
int c; |
50 |
public: |
51 |
DLGChar(int ch) { c = ch; } |
52 |
int atom() { return c; } |
53 |
}; |
54 |
*/ |
55 |
|
56 |
/* user must subclass this */ |
57 |
class DllExportPCCTS DLGInputStream { |
58 |
public: |
59 |
virtual int nextChar() = 0; |
60 |
}; |
61 |
|
62 |
/* Predefined char stream: Input from FILE */ |
63 |
class DllExportPCCTS DLGFileInput : public DLGInputStream { |
64 |
private: |
65 |
int found_eof; |
66 |
FILE *input; |
67 |
public: |
68 |
DLGFileInput(FILE *f) { input = f; found_eof = 0; } |
69 |
int nextChar() { |
70 |
int c; |
71 |
if ( found_eof ) return EOF; |
72 |
else { |
73 |
c=getc(input); |
74 |
if ( c==EOF ) found_eof = 1; |
75 |
return c; |
76 |
} |
77 |
} |
78 |
void DLGFileReset(FILE *f) {input=f; found_eof = 0; }; // MR11 |
79 |
}; |
80 |
|
81 |
// MR9 Suggested by Bruce Guenter (bruceg@qcc.sk.ca) |
82 |
// MR9 Make DLGStringInput const correct |
83 |
|
84 |
/* Predefined char stream: Input from string */ |
85 |
class DllExportPCCTS DLGStringInput : public DLGInputStream { |
86 |
private: |
87 |
const DLGChar *input; // MR9 |
88 |
const DLGChar *p; // MR9 |
89 |
public: |
90 |
DLGStringInput(const DLGChar *s) { input = s; p = &input[0];} // MR9 |
91 |
int nextChar() |
92 |
{ |
93 |
if (*p) return (int) (unsigned char) *p++; // MR14 |
94 |
else return EOF; |
95 |
} |
96 |
|
97 |
void DLGStringReset(const DLGChar *s) {input=s; p= &input[0]; }; // MR11 // MR16 |
98 |
}; |
99 |
|
100 |
class DllExportPCCTS DLGState { |
101 |
public: |
102 |
DLGInputStream *input; |
103 |
int interactive; |
104 |
int track_columns; |
105 |
int auto_num; |
106 |
int add_erase; |
107 |
int lookc; |
108 |
int char_full; |
109 |
int begcol, endcol; |
110 |
int line; |
111 |
DLGChar *lextext, *begexpr, *endexpr; |
112 |
int bufsize; |
113 |
int bufovf; |
114 |
DLGChar *nextpos; |
115 |
int class_num; |
116 |
int debugLexerFlag; // MR1 |
117 |
ANTLRParser *parser; // MR1 |
118 |
}; |
119 |
|
120 |
/* user must subclass this */ |
121 |
class DllExportPCCTS DLGLexerBase : public ANTLRTokenStream { |
122 |
public: |
123 |
virtual ANTLRTokenType erraction(); |
124 |
|
125 |
protected: |
126 |
DLGInputStream *input; |
127 |
int interactive; |
128 |
int track_columns; |
129 |
DLGChar *_lextext; /* text of most recently matched token */ |
130 |
DLGChar *_begexpr; /* beginning of last reg expr recogn. */ |
131 |
DLGChar *_endexpr; /* beginning of last reg expr recogn. */ |
132 |
int _bufsize; /* number of characters in lextext */ |
133 |
int _begcol; /* column that first character of token is in*/ |
134 |
int _endcol; /* column that last character of token is in */ |
135 |
int _line; /* line current token is on */ |
136 |
int ch; /* character to determine next state */ |
137 |
int bufovf; /* indicates that buffer too small for text */ |
138 |
int charfull; |
139 |
DLGChar *nextpos; /* points to next available position in lextext*/ |
140 |
int cl; |
141 |
int automaton; |
142 |
int add_erase; |
143 |
DLGChar ebuf[70]; |
144 |
_ANTLRTokenPtr token_to_fill; |
145 |
|
146 |
int debugLexerFlag; // MR1 |
147 |
ANTLRParser *parser; // MR1 |
148 |
public: |
149 |
virtual _ANTLRTokenPtr getToken(); // MR12 public |
150 |
virtual void advance(void) = 0; |
151 |
void skip(void); /* erase lextext, look for antoher token */ |
152 |
void more(void); /* keep lextext, look for another token */ |
153 |
void mode(int k); /* switch to automaton 'k' */ |
154 |
void saveState(DLGState *); |
155 |
void restoreState(DLGState *); |
156 |
virtual ANTLRTokenType nextTokenType(void)=0;/* get next token */ |
157 |
void replchar(DLGChar c); /* replace last recognized reg. expr. with |
158 |
a character */ |
159 |
void replstr(const DLGChar *s); /* replace last recognized reg. expr. with |
160 |
a string */ /* MR20 const */ |
161 |
virtual int err_in(); // MR1 |
162 |
virtual void errstd(const char *); // MR1 MR20 const |
163 |
int line() { return _line; } |
164 |
void set_line(int newValue) { _line=newValue; }; // MR1 |
165 |
virtual void newline() { _line++; } |
166 |
DLGChar *lextext() { return _lextext; } |
167 |
|
168 |
int begcol() { return _begcol; } |
169 |
int endcol() { return _endcol; } |
170 |
void set_begcol(int a) { _begcol=a; } |
171 |
void set_endcol(int a) { _endcol=a; } |
172 |
DLGChar *begexpr() { return _begexpr; } |
173 |
DLGChar *endexpr() { return _endexpr; } |
174 |
int bufsize() { return _bufsize; } |
175 |
|
176 |
void setToken(ANTLRAbstractToken *t) { token_to_fill = t; } |
177 |
|
178 |
void setInputStream(DLGInputStream *); |
179 |
DLGLexerBase(DLGInputStream *in, |
180 |
unsigned bufsize=2000, |
181 |
int interactive=0, |
182 |
int track_columns=0); |
183 |
void reset(); // MR19 |
184 |
virtual ~DLGLexerBase() { delete [] _lextext; } |
185 |
virtual void panic(const char *msg); // MR1 MR20 const |
186 |
void trackColumns() { |
187 |
track_columns = 1; |
188 |
this->_begcol = 0; |
189 |
this->_endcol = 0; |
190 |
}; |
191 |
virtual ANTLRParser *setParser(ANTLRParser *p); // MR1 |
192 |
virtual ANTLRParser *getParser(); // MR1 |
193 |
virtual int debugLexer(int value); // MR1 |
194 |
int lexErrCount; // MR12 |
195 |
virtual int printMessage(FILE* pFile, const char* pFormat, ...); // MR23 |
196 |
}; |
197 |
|
198 |
#endif |