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

[transliterate] no need to strdup transliterator names if they are lowercased, breaking on NUL byte

Browse Source
This commit is contained in:
Al
2015-07-01 14:51:22 -04:00
parent b58877ec6c
commit 46e51ae91e
1 changed files with 14 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
src/transliterate.c
View File

@@ -661,17 +661,23 @@ char *transliterate(char *trans_name, char *str, size_t len) {
log_debug("len = %zu\n", len); log_debug("len = %zu\n", len);
str = strdup(str); str = strdup(str);
trans_name = strdup(trans_name);
// Transliterator names are ASCII strings, so this is fine bool allocated_trans_name = false;
string_lower(trans_name);
if (!string_is_lower(trans_name)) {
trans_name = strdup(trans_name);
// Transliterator names are ASCII strings, so this is fine
string_lower(trans_name);
allocated_trans_name = true;
}
log_debug("lower = %s\n", trans_name); log_debug("lower = %s\n", trans_name);
transliterator_t *transliterator = get_transliterator(trans_name); transliterator_t *transliterator = get_transliterator(trans_name);
if (transliterator == NULL) { if (transliterator == NULL) {
log_warn("transliterator \"%s\" does not exist\n", trans_name); log_warn("transliterator \"%s\" does not exist\n", trans_name);
free(trans_name); if (allocated_trans_name) free(trans_name);
free(str); free(str);
return NULL; return NULL;
} }
@@ -684,7 +690,7 @@ char *transliterate(char *trans_name, char *str, size_t len) {
uint32_t trans_node_id = result.node_id; uint32_t trans_node_id = result.node_id;
free(trans_name); if (allocated_trans_name) free(trans_name);
result = trie_get_prefix_from_index(trans_table->trie, NAMESPACE_SEPARATOR_CHAR, NAMESPACE_SEPARATOR_CHAR_LEN, result.node_id, result.tail_pos); result = trie_get_prefix_from_index(trans_table->trie, NAMESPACE_SEPARATOR_CHAR, NAMESPACE_SEPARATOR_CHAR_LEN, result.node_id, result.tail_pos);
@@ -752,7 +758,7 @@ char *transliterate(char *trans_name, char *str, size_t len) {
ssize_t char_len = 0; ssize_t char_len = 0;
uint8_t *ptr = (uint8_t *)str; uint8_t *ptr = (uint8_t *)str;
uint64_t idx = 0; uint64_t idx = 0;
char *original_str = str; char *original_str = str;
char_array *revisit = NULL; char_array *revisit = NULL;
bool in_revisit = false; bool in_revisit = false;
@@ -776,6 +782,8 @@ char *transliterate(char *trans_name, char *str, size_t len) {
continue; continue;
} }
if (ch == 0) break;
log_debug("Got char '%.*s' at idx=%zu\n", (int)char_len, str + idx, idx); log_debug("Got char '%.*s' at idx=%zu\n", (int)char_len, str + idx, idx);
state = state_transition(trie, str, idx, char_len, prev_state); state = state_transition(trie, str, idx, char_len, prev_state);