Mercurial > hg > Applications > Grep

comparison c/regexParser/main.cc @ 76:d98a036441e2

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
add createNode comment
author masa
date Fri, 28 Aug 2015 20:36:16 +0900
parents 6541eae41a73
children 7f53a587bf97
comparison
equal deleted inserted replaced
75:6541eae41a73 76:d98a036441e2
7 */ 7 */
8 8
9 #include <stdio.h> 9 #include <stdio.h>
10 #include <stdlib.h> 10 #include <stdlib.h>
11 #include <string.h> 11 #include <string.h>
12
13 typedef struct charClass {
14 unsigned char table[256];
15 struct utf8Range {
16 unsigned char *begin;
17 unsigned char *end;
18 struct utf8Range *next;
19 } *rangeList;
20 } CharClass, *CharClassPtr;
21 12
22 typedef struct node { 13 typedef struct node {
23 unsigned char type; 14 unsigned char type;
24 union value { 15 union value {
25 charClass *cc; 16 charClass *cc;
29 struct node *parent; 20 struct node *parent;
30 struct node *left; 21 struct node *left;
31 struct node *right; 22 struct node *right;
32 } Node, *NodePtr; 23 } Node, *NodePtr;
33 24
25 typedef struct charClass {
26 NodePtr nextState[256];
27 struct utf8Range {
28 unsigned char *begin;
29 unsigned char *end;
30 struct utf8Range *next;
31 NodePtr nextState;
32 } *rangeList;
33 } CharClass, *CharClassPtr;
34
34 unsigned char *ptr; 35 unsigned char *ptr;
35 unsigned char tokenType; 36 unsigned char tokenType;
36 int tokenValue; 37 int tokenValue;
37 NodePtr regexHeadNode; 38 NodePtr regexHeadNode;
38 39
39 NodePtr charClass(); 40 NodePtr charClass();
40 NodePtr group(); 41 NodePtr group();
41 NodePtr orexp();
42 NodePtr asterisk();
43 NodePtr regex(); 42 NodePtr regex();
44 NodePtr createNode(unsigned char,NodePtr,NodePtr); 43 NodePtr createNode(unsigned char,NodePtr,NodePtr);
45 extern void token(); 44 void token();
46 extern NodePtr regexAtom(); 45 NodePtr regexAtom();
47 46
48 47
49 bool isLiteral(char c) { 48 bool isLiteral(char c) {
50 if (*ptr > 0x7f) return true; 49 if (*ptr > 0x7f) return true;
51 else if (*ptr == '(') return false; 50 else if (*ptr == '(') return false;
53 else if (*ptr == '|') return false; 52 else if (*ptr == '|') return false;
54 else if (*ptr == '*') return false; 53 else if (*ptr == '*') return false;
55 return true; 54 return true;
56 } 55 }
57 56
57 /**
58 * Create a node of regex parse tree.
59 * tokenType
60 * regexPosition(state)
61 * stateTransitionTable
62 */
58 NodePtr createNode(unsigned char character, NodePtr left, NodePtr right) { 63 NodePtr createNode(unsigned char character, NodePtr left, NodePtr right) {
59 NodePtr n = (NodePtr)malloc(sizeof(Node)); 64 NodePtr n = (NodePtr)malloc(sizeof(Node));
60 n->self = n; 65 n->self = n;
61 n->Value.character = character; 66 n->Value.character = character;
62 n->left = left; 67 n->left = left;