AIR/slre
1
0
Fork 0
mirror of https://github.com/cesanta/slre.git synced 2025-03-13 01:05:29 +00:00
Code Issues Packages Projects Releases Wiki Activity

Implemented 0xHH match

Browse Source
This commit is contained in:
Sergey Lyubka 2013-08-02 14:21:48 +01:00
parent 8cfa4336fc
commit 7f8fbe5680
2 changed files with 29 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@ -35,12 +35,12 @@ most.
? Match zero or once ? Match zero or once
x|y Match x or y (alternation operator) x|y Match x or y (alternation operator)
\meta Match one of the meta character: ^$().[]*+?|\ \meta Match one of the meta character: ^$().[]*+?|\
\xHH Match byte with hex value 0xHH
Not yet supported but in progress: Not yet supported but in progress:
[...] Match any character from set. A-Z like ranges supported [...] Match any character from set. A-Z like ranges supported
[^...] Match any character but ones from set [^...] Match any character but ones from set
\xDD Match byte with hex value 0xDD
## API ## API

29
slre.c
View File

@ -74,13 +74,21 @@ struct regex_info {
enum { IGNORE_CASE = 1 }; enum { IGNORE_CASE = 1 };
static int get_op_len(const char *re) { static int get_op_len(const char *re) {
return re[0] == '\\' ? 2 : 1; return re[0] == '\\' && re[1] == 'x' ? 4 :
re[0] == '\\' ? 2 : 1;
} }
static int is_quantifier(const char *re) { static int is_quantifier(const char *re) {
return re[0] == '*' || re[0] == '+' || re[0] == '?'; return re[0] == '*' || re[0] == '+' || re[0] == '?';
} }
static int toi(int x) {
return isdigit(x) ? x - '0' : x - 'W';
}
static int hextoi(const unsigned char *s) {
return (toi(tolower(s[0])) << 4) | toi(tolower(s[1]));
}
static int doh(const char *s, int s_len, static int doh(const char *s, int s_len,
struct slre_cap *caps, struct regex_info *info, int bi); struct slre_cap *caps, struct regex_info *info, int bi);
@ -156,6 +164,18 @@ static int bar(const char *re, int re_len, const char *s, int s_len,
j++; j++;
break; break;
case 'x':
/* Match byte, \xHH where HH is hexadecimal byte representaion */
{
const unsigned char *r = (const unsigned char *) re + i + 2;
FAIL_IF(!(isxdigit(r[0]) && isxdigit(r[1])),
static_error_invalid_metacharacter);
FAIL_IF(hextoi(r) != * (unsigned char *) (s + j),
static_error_no_match);
}
j++;
break;
case '+': case '?': case '*': case '\\': case '(': case ')': case '+': case '?': case '*': case '\\': case '(': case ')':
case '^': case '$': case '.': case '[': case ']': case '^': case '$': case '.': case '[': case ']':
FAIL_IF(re[i + 1] != s[j], static_error_no_match); FAIL_IF(re[i + 1] != s[j], static_error_no_match);
@ -426,6 +446,13 @@ int main(void) {
ASSERT(strcmp(msg, static_error_unexpected_quantifier) == 0); ASSERT(strcmp(msg, static_error_unexpected_quantifier) == 0);
ASSERT(slre_match("()+", "fooklmn", 7, NULL, &msg) == 0); ASSERT(slre_match("()+", "fooklmn", 7, NULL, &msg) == 0);
ASSERT(strcmp(msg, static_error_no_match) == 0); ASSERT(strcmp(msg, static_error_no_match) == 0);
ASSERT(slre_match("\\x", "12", 2, NULL, &msg) == 0);
ASSERT(strcmp(msg, static_error_invalid_metacharacter) == 0);
ASSERT(slre_match("\\xhi", "12", 2, NULL, &msg) == 0);
ASSERT(strcmp(msg, static_error_invalid_metacharacter) == 0);
ASSERT(slre_match("\\x20", "_ J", 3, NULL, &msg) == 2);
ASSERT(slre_match("\\x4A", "_ J", 3, NULL, &msg) == 3);
ASSERT(slre_match("\\d+", "abc123def", 9, NULL, &msg) == 6);
/* Balancing brackets */ /* Balancing brackets */
ASSERT(slre_match("(x))", "fooklmn", 7, NULL, &msg) == 0); ASSERT(slre_match("(x))", "fooklmn", 7, NULL, &msg) == 0);