1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
|
#include "prism/util/pm_strpbrk.h"
/**
* This is the slow path that does care about the encoding.
*/
static inline const uint8_t *
pm_strpbrk_multi_byte(const pm_parser_t *parser, const uint8_t *source, const uint8_t *charset, size_t maximum) {
size_t index = 0;
while (index < maximum) {
if (strchr((const char *) charset, source[index]) != NULL) {
return source + index;
}
size_t width = parser->encoding->char_width(source + index, (ptrdiff_t) (maximum - index));
if (width == 0) {
return NULL;
}
index += width;
}
return NULL;
}
/**
* This is the fast path that does not care about the encoding.
*/
static inline const uint8_t *
pm_strpbrk_single_byte(const uint8_t *source, const uint8_t *charset, size_t maximum) {
size_t index = 0;
while (index < maximum) {
if (strchr((const char *) charset, source[index]) != NULL) {
return source + index;
}
index++;
}
return NULL;
}
/**
* Here we have rolled our own version of strpbrk. The standard library strpbrk
* has undefined behavior when the source string is not null-terminated. We want
* to support strings that are not null-terminated because pm_parse does not
* have the contract that the string is null-terminated. (This is desirable
* because it means the extension can call pm_parse with the result of a call to
* mmap).
*
* The standard library strpbrk also does not support passing a maximum length
* to search. We want to support this for the reason mentioned above, but we
* also don't want it to stop on null bytes. Ruby actually allows null bytes
* within strings, comments, regular expressions, etc. So we need to be able to
* skip past them.
*
* Finally, we want to support encodings wherein the charset could contain
* characters that are trailing bytes of multi-byte characters. For example, in
* Shift-JIS, the backslash character can be a trailing byte. In that case we
* need to take a slower path and iterate one multi-byte character at a time.
*/
const uint8_t *
pm_strpbrk(const pm_parser_t *parser, const uint8_t *source, const uint8_t *charset, ptrdiff_t length) {
if (length <= 0) {
return NULL;
} else if (parser->encoding_changed && parser->encoding->multibyte) {
return pm_strpbrk_multi_byte(parser, source, charset, (size_t) length);
} else {
return pm_strpbrk_single_byte(source, charset, (size_t) length);
}
}
|
