basic/extract-word: add EXTRACT_UNESCAPE_SEPARATORS mode

This allows separators to be escaped, for example to allow
"a\:b:c", to be treated as "a:b", "c" with ":" as the separator.
This commit is contained in:
Zbigniew Jędrzejewski-Szmek
2020-08-05 21:29:13 +01:00
committed by Luca Boccassi
parent 866fdcceb4
commit 1e198efcdb
3 changed files with 59 additions and 13 deletions
+15 -10
View File
@@ -86,25 +86,30 @@ int extract_first_word(const char **p, char **ret, const char *separators, Extra
return -EINVAL;
}
if (flags & EXTRACT_CUNESCAPE) {
if (flags & (EXTRACT_CUNESCAPE|EXTRACT_UNESCAPE_SEPARATORS)) {
bool eight_bit = false;
char32_t u;
r = cunescape_one(*p, (size_t) -1, &u, &eight_bit, false);
if (r < 0) {
if (flags & EXTRACT_CUNESCAPE_RELAX) {
s[sz++] = '\\';
s[sz++] = c;
} else
return -EINVAL;
} else {
if ((flags & EXTRACT_CUNESCAPE) &&
(r = cunescape_one(*p, (size_t) -1, &u, &eight_bit, false)) >= 0) {
/* A valid escaped sequence */
assert(r >= 1);
(*p) += r - 1;
if (eight_bit)
s[sz++] = u;
else
sz += utf8_encode_unichar(s + sz, u);
}
} else if ((flags & EXTRACT_UNESCAPE_SEPARATORS) &&
strchr(separators, **p))
/* An escaped separator char */
s[sz++] = c;
else if (flags & EXTRACT_CUNESCAPE_RELAX) {
s[sz++] = '\\';
s[sz++] = c;
} else
return -EINVAL;
} else
s[sz++] = c;
+4 -3
View File
@@ -7,9 +7,10 @@ typedef enum ExtractFlags {
EXTRACT_RELAX = 1 << 0,
EXTRACT_CUNESCAPE = 1 << 1,
EXTRACT_CUNESCAPE_RELAX = 1 << 2,
EXTRACT_UNQUOTE = 1 << 3,
EXTRACT_DONT_COALESCE_SEPARATORS = 1 << 4,
EXTRACT_RETAIN_ESCAPE = 1 << 5,
EXTRACT_UNESCAPE_SEPARATORS = 1 << 3,
EXTRACT_UNQUOTE = 1 << 4,
EXTRACT_DONT_COALESCE_SEPARATORS = 1 << 5,
EXTRACT_RETAIN_ESCAPE = 1 << 6,
} ExtractFlags;
int extract_first_word(const char **p, char **ret, const char *separators, ExtractFlags flags);