annotate c/regexParser/regexParser.cc @ 124:c363a66dc1a7 pairPro

fix
author masa
date Tue, 01 Dec 2015 17:06:26 +0900
parents 188d866227a4
children b061cd8205cc
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
1 #include <stdlib.h>
89
50a146c05192 add NodeNumber in Regex Parser tree
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 87
diff changeset
2 #include <stdio.h>
122
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
3 #include <string.h>
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
4 #include <ctype.h>
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
5 #include "regexParser.h"
115
ca30f8334741 rename createRegexParser.cc to regexParser.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 112
diff changeset
6 #include "error.h"
ca30f8334741 rename createRegexParser.cc to regexParser.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 112
diff changeset
7
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
8 static NodePtr allocateNode();
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
9 static NodePtr createNode(RegexInfoPtr,unsigned char*,NodePtr,NodePtr);
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
10 static NodePtr charClass(RegexInfoPtr);
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
11 static NodePtr group(RegexInfoPtr);
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
12 static void token(RegexInfoPtr);
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
13 static NodePtr regexAtom(RegexInfoPtr);
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
14 NodePtr regex(RegexInfoPtr);
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
15
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
16 /**
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
17 * Create a node of regex parse tree.
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
18 * tokenType
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
19 * regexPosition(state)
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
20 * stateTransitionTable
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
21 */
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
22
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
23 static
116
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
24 NodePtr allocateNode() {
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
25 NodePtr n = (NodePtr)malloc(sizeof(node));
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
26 n->cc = (CharClassPtr)malloc(sizeof(CharClass));
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
27 n->cc->cond = (ConditionList)malloc(sizeof(Condition));
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
28 return n;
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
29 }
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
30
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
31 static
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
32 NodePtr createNode(RegexInfoPtr ri,unsigned char *character, NodePtr left, NodePtr right) {
116
66c633575b53 remove error and warning
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 115
diff changeset
33 NodePtr n = allocateNode();
108
70069d4647a0 implement malloc error checking
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 104
diff changeset
34 if (n == NULL) {
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
35 mallocFailedMessage();
108
70069d4647a0 implement malloc error checking
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 104
diff changeset
36 }
70069d4647a0 implement malloc error checking
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 104
diff changeset
37
89
50a146c05192 add NodeNumber in Regex Parser tree
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 87
diff changeset
38 n->tokenType = ri->tokenType;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
39 n->left = left;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
40 n->right = right;
120
5d29b6a1b50f include Word in Node
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 118
diff changeset
41 n->nodeNumber = ri->nodeNumber;
5d29b6a1b50f include Word in Node
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 118
diff changeset
42 ri->nodeNumber++;
89
50a146c05192 add NodeNumber in Regex Parser tree
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 87
diff changeset
43
50a146c05192 add NodeNumber in Regex Parser tree
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 87
diff changeset
44 if (ri->tokenType == 'a') {
50a146c05192 add NodeNumber in Regex Parser tree
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 87
diff changeset
45 ri->tokenType = 0;
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
46 n->cc->cond->w = getWord(ri->tokenValue);
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
47 } else {
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
48 WordPtr w = (WordPtr)malloc(sizeof(Word));
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
49 w->word = character;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
50 w->length = 1;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
51 n->cc->cond->w = w;
89
50a146c05192 add NodeNumber in Regex Parser tree
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 87
diff changeset
52 }
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
53 return n;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
54 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
55
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
56 // <charClass> ::= '['<literal>'-'<literal>']'
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
57 static
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
58 NodePtr charClass(RegexInfoPtr ri) {
122
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
59 NodePtr n = allocateNode();
108
70069d4647a0 implement malloc error checking
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 104
diff changeset
60 if (n == NULL) {
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
61 mallocFailedMessage();
108
70069d4647a0 implement malloc error checking
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 104
diff changeset
62 }
122
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
63
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
64 n->tokenType = ri->tokenType;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
65 n->nodeNumber = ri->nodeNumber;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
66 ri->nodeNumber++;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
67 n->cc->cond->w = (WordPtr)malloc(sizeof(Word));
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
68
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
69 int i = 0;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
70
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
71 while (ri->ptr[i] != ']') {
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
72 if (ri->ptr[i] == '-') {
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
73 n->cc->begin = ri->ptr[i-1];
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
74 n->cc->end = ri->ptr[i+1];
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
75 }
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
76 i++;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
77 }
124
masa
parents: 122
diff changeset
78 // TODO literal support
122
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
79
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
80 n->cc->cond->w->word = (unsigned char*)malloc(sizeof(unsigned char)*(i+1));
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
81 strncpy((char*)n->cc->cond->w->word, (char*)ri->ptr,i+1);
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
82 n->cc->cond->w->word[i] = '\0';
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
83 ri->ptr += i+1;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
84
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
85 return n;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
86 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
87
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
88 // <literal> ::= [a-z][A-Z][0-9]
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
89 static
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
90 NodePtr literal(RegexInfoPtr ri) {
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
91 NodePtr n = createNode(ri,ri->ptr,0,0);
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
92 return n;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
93 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
94
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
95 // <group> ::= '('<regex>')'
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
96 static
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
97 NodePtr group(RegexInfoPtr ri) {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
98 return regex(ri);
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
99 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
100
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
101 static
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
102 void token(RegexInfoPtr ri) {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
103 while (ri->ptr[0] != '\0') {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
104 if (ri->ptr[0] == '('){
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
105 ri->ptr++;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
106 ri->tokenType = '(';
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
107 ri->tokenValue = NULL;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
108 return;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
109 } else if (ri->ptr[0] == ')') {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
110 ri->ptr++;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
111 ri->tokenType = ')';
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
112 ri->tokenValue = ri->ptr;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
113 return;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
114 } else if (ri->ptr[0] == '[') {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
115 ri->ptr++;
122
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
116 ri->tokenType = 'c';
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
117 ri->tokenValue = ri->ptr;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
118 return;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
119 } else if (ri->ptr[0] == '|'){
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
120 ri->ptr++;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
121 ri->tokenType = '|';
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
122 ri->tokenValue = NULL;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
123 return;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
124 } else if (ri->ptr[0] == '*'){
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
125 ri->ptr++;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
126 ri->tokenType = '*';
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
127 ri->tokenValue = NULL;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
128 return;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
129 } else if (ri->ptr[0] == '\\'){
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
130 // need more proccesing
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
131 /*
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
132 \277
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
133 \0xa5
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
134 \[
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
135 \\
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
136 \utf-8 etc...
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
137 */
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
138 } else {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
139 ri->tokenType = 'a';
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
140 ri->tokenValue = ri->ptr;
122
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
141 while (isalnum(ri->ptr[0])) {
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
142 ri->ptr++;
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
143 }
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
144 return;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
145 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
146 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
147 return;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
148 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
149
124
masa
parents: 122
diff changeset
150 // <regexAtom> ::= <literal>|<charClass>
112
ec485345daf9 some function use static
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 108
diff changeset
151 static
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
152 NodePtr regexAtom(RegexInfoPtr ri) {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
153
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
154 token(ri);
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
155 NodePtr n = NULL;
124
masa
parents: 122
diff changeset
156 if (ri->tokenType == 'c') n = charClass(ri);
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
157
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
158 return n;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
159 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
160
122
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 121
diff changeset
161 // <regex> ::= <regexAtom> | <regexAtom><regex>'*' | <regexAtom>'*' | <regexAtom>'|'<regex> | <regexAtom><regex> | '(' regex ')'
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
162 NodePtr regex(RegexInfoPtr ri) {
124
masa
parents: 122
diff changeset
163 NodePtr n = NULL;
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
164 while (ri->ptr[0]) {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
165 token(ri);
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
166 if (ri->tokenType == '*') {
124
masa
parents: 122
diff changeset
167 // TODO literal support
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
168 unsigned char *syntax = (unsigned char*)malloc(sizeof(unsigned char));
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
169 syntax[0] = '*';
124
masa
parents: 122
diff changeset
170 NodePtr n1 = createNode(ri,syntax,n->right,0);
masa
parents: 122
diff changeset
171
masa
parents: 122
diff changeset
172 unsigned char *syntax1 = (unsigned char*)malloc(sizeof(unsigned char));
masa
parents: 122
diff changeset
173 syntax1[0] = '+';
masa
parents: 122
diff changeset
174
masa
parents: 122
diff changeset
175 n = createNode(ri,syntax1,n->left,n1);
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
176 } else if (ri->tokenType == '|') {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
177 NodePtr n1 = regex(ri);
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
178 unsigned char *syntax = (unsigned char*)malloc(sizeof(unsigned char));
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
179 syntax[0] = '|';
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
180 n = createNode(ri,syntax,n,n1);
124
masa
parents: 122
diff changeset
181 } else if (ri->tokenType == '(') {
masa
parents: 122
diff changeset
182 NodePtr n1 = regex(ri);
masa
parents: 122
diff changeset
183 unsigned char *syntax = (unsigned char*)malloc(sizeof(unsigned char));
masa
parents: 122
diff changeset
184 syntax[0] = '+';
masa
parents: 122
diff changeset
185 n = createNode(ri,syntax,n,n1);
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
186 } else if (ri->tokenType == ')') {
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
187 return n;
124
masa
parents: 122
diff changeset
188 } else if (ri->tokenType == 'a') {
masa
parents: 122
diff changeset
189 NodePtr n1 = literal(ri);
masa
parents: 122
diff changeset
190 unsigned char *syntax = (unsigned char*)malloc(sizeof(unsigned char));
masa
parents: 122
diff changeset
191 syntax[0] = '+';
masa
parents: 122
diff changeset
192 n = createNode(ri,syntax,n,n1);
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
193 } else {
124
masa
parents: 122
diff changeset
194 // return NULL
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
195 NodePtr n1 = regex(ri);
118
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
196 unsigned char *syntax = (unsigned char*)malloc(sizeof(unsigned char));
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
197 syntax[0] = '+';
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents: 116
diff changeset
198 n = createNode(ri,syntax,n,n1);
82
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
199 }
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
200 } return n;
1d9bbf922bb6 add createRegexTree.cc
Masataka Kohagura <kohagura@cr.ie.u-ryukyu.ac.jp>
parents:
diff changeset
201 }