tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[utils] Adding cstring_array_split_ignore_consecutive

Browse Source
This commit is contained in:
Al
2016-05-25 17:07:20 -04:00
parent 6baa7087fe
commit b1816e9b70
2 changed files with 21 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
src/string_utils.c
View File

@@ -829,17 +829,23 @@ inline int64_t cstring_array_token_length(cstring_array *self, uint32_t i) {
}
}
cstring_array *cstring_array_split(char *str, const char *separator, size_t separator_len, size_t *count) {
static cstring_array *cstring_array_split_options(char *str, const char *separator, size_t separator_len, bool ignore_consecutive, size_t *count) {
*count = 0;
char_array *array = char_array_new_size(strlen(str));
bool last_was_separator = false;
while (*str) {
if ((separator_len == 1 && *str == separator[0]) || (memcmp(str, separator, separator_len) == 0)) {
if (!ignore_consecutive || !last_was_separator) {
char_array_push(array, '\0');
}
str += separator_len;
last_was_separator = true;
} else {
char_array_push(array, *str);
str++;
last_was_separator = false;
}
}
char_array_push(array, '\0');
@@ -850,6 +856,17 @@ cstring_array *cstring_array_split(char *str, const char *separator, size_t sepa
return string_array;
}
cstring_array *cstring_array_split(char *str, const char *separator, size_t separator_len, size_t *count) {
return cstring_array_split_options(str, separator, separator_len, true, count);
}
cstring_array *cstring_array_split_ignore_consecutive(char *str, const char *separator, size_t separator_len, size_t *count) {
return cstring_array_split_options(str, separator, separator_len, false, count);
}
cstring_array *cstring_array_split_no_copy(char *str, char separator, size_t *count) {
*count = 0;
char *ptr = str;

2
src/string_utils.h
View File

@@ -180,6 +180,8 @@ char **cstring_array_to_strings(cstring_array *self);
// Split on delimiter
cstring_array *cstring_array_split(char *str, const char *separator, size_t separator_len, size_t *count);
// Split on delimiter, ignore multiple consecutive delimiters
cstring_array *cstring_array_split_ignore_consecutive(char *str, const char *separator, size_t separator_len, size_t *count);
// Split on delimiter by replacing (single character) separator with the NUL byte in the original string
cstring_array *cstring_array_split_no_copy(char *str, char separator, size_t *count);