1 /* SPDX-License-Identifier: BSD-3-Clause
2 * Copyright 2016, Olivier MATZ <zer0@droids-corp.org>
12 #include <ecoli_malloc.h>
13 #include <ecoli_log.h>
14 #include <ecoli_test.h>
15 #include <ecoli_strvec.h>
16 #include <ecoli_dict.h>
17 #include <ecoli_node.h>
18 #include <ecoli_complete.h>
19 #include <ecoli_parse.h>
20 #include <ecoli_config.h>
21 #include <ecoli_node_many.h>
22 #include <ecoli_node_or.h>
23 #include <ecoli_node_str.h>
24 #include <ecoli_node_int.h>
25 #include <ecoli_node_re_lex.h>
27 EC_LOG_TYPE_REGISTER(node_re_lex);
29 struct regexp_pattern {
36 struct ec_node_re_lex {
37 struct ec_node *child;
38 struct regexp_pattern *table;
42 static struct ec_strvec *
43 tokenize(struct regexp_pattern *table, size_t table_len, const char *str)
45 struct ec_strvec *strvec = NULL;
46 struct ec_dict *attrs = NULL;
64 for (i = 0; i < table_len; i++) {
65 ret = regexec(&table[i].r, &dup[off], 1, &pos, 0);
68 if (pos.rm_so != 0 || pos.rm_eo == 0) {
73 if (table[i].keep == 0)
76 c = dup[pos.rm_eo + off];
77 dup[pos.rm_eo + off] = '\0';
78 EC_LOG(EC_LOG_DEBUG, "re_lex match <%s>\n", &dup[off]);
79 if (ec_strvec_add(strvec, &dup[off]) < 0)
82 if (table[i].attr_name != NULL) {
86 if (ec_dict_set(attrs, table[i].attr_name,
89 if (ec_strvec_set_attrs(strvec,
90 ec_strvec_len(strvec) - 1,
98 dup[pos.rm_eo + off] = c;
113 ec_strvec_free(strvec);
118 ec_node_re_lex_parse(const struct ec_node *node,
119 struct ec_pnode *state,
120 const struct ec_strvec *strvec)
122 struct ec_node_re_lex *priv = ec_node_priv(node);
123 struct ec_strvec *new_vec = NULL;
124 struct ec_pnode *child_parse;
128 if (priv->child == NULL) {
133 if (ec_strvec_len(strvec) == 0) {
134 new_vec = ec_strvec();
136 str = ec_strvec_val(strvec, 0);
137 new_vec = tokenize(priv->table, priv->len, str);
142 ret = ec_parse_child(priv->child, state, new_vec);
146 if ((unsigned)ret == ec_strvec_len(new_vec)) {
148 } else if (ret != EC_PARSE_NOMATCH) {
149 child_parse = ec_pnode_get_last_child(state);
150 ec_pnode_unlink_child(state, child_parse);
151 ec_pnode_free(child_parse);
152 ret = EC_PARSE_NOMATCH;
155 ec_strvec_free(new_vec);
161 ec_strvec_free(new_vec);
165 static void ec_node_re_lex_free_priv(struct ec_node *node)
167 struct ec_node_re_lex *priv = ec_node_priv(node);
170 ec_node_free(priv->child);
171 for (i = 0; i < priv->len; i++) {
172 ec_free(priv->table[i].pattern);
173 ec_free(priv->table[i].attr_name);
174 regfree(&priv->table[i].r);
177 ec_free(priv->table);
181 ec_node_re_lex_get_children_count(const struct ec_node *node)
183 struct ec_node_re_lex *priv = ec_node_priv(node);
191 ec_node_re_lex_get_child(const struct ec_node *node, size_t i,
192 struct ec_node **child, unsigned int *refs)
194 struct ec_node_re_lex *priv = ec_node_priv(node);
199 *child = priv->child;
204 static const struct ec_config_schema ec_node_re_lex_dict[] = {
207 .desc = "The pattern to match.",
208 .type = EC_CONFIG_TYPE_STRING,
212 .desc = "Whether to keep or drop the string matching "
213 "the regular expression.",
214 .type = EC_CONFIG_TYPE_BOOL,
218 .desc = "The optional attribute name to attach.",
219 .type = EC_CONFIG_TYPE_STRING,
222 .type = EC_CONFIG_TYPE_NONE,
226 static const struct ec_config_schema ec_node_re_lex_elt[] = {
228 .desc = "A pattern element.",
229 .type = EC_CONFIG_TYPE_DICT,
230 .subschema = ec_node_re_lex_dict,
233 .type = EC_CONFIG_TYPE_NONE,
237 static const struct ec_config_schema ec_node_re_lex_schema[] = {
240 .desc = "The list of patterns elements.",
241 .type = EC_CONFIG_TYPE_LIST,
242 .subschema = ec_node_re_lex_elt,
246 .desc = "The child node.",
247 .type = EC_CONFIG_TYPE_NODE,
250 .type = EC_CONFIG_TYPE_NONE,
254 static int ec_node_re_lex_set_config(struct ec_node *node,
255 const struct ec_config *config)
257 struct ec_node_re_lex *priv = ec_node_priv(node);
258 struct regexp_pattern *table = NULL;
259 const struct ec_config *patterns, *child, *elt, *pattern, *keep, *attr;
260 char *pattern_str = NULL, *attr_name = NULL;
264 child = ec_config_dict_get(config, "child");
267 if (ec_config_get_type(child) != EC_CONFIG_TYPE_NODE) {
272 patterns = ec_config_dict_get(config, "patterns");
273 if (patterns != NULL) {
274 n = ec_config_count(patterns);
278 table = ec_calloc(n, sizeof(*table));
283 TAILQ_FOREACH(elt, &patterns->list, next) {
284 if (ec_config_get_type(elt) != EC_CONFIG_TYPE_DICT) {
288 pattern = ec_config_dict_get(elt, "pattern");
289 if (pattern == NULL) {
293 if (ec_config_get_type(pattern) != EC_CONFIG_TYPE_STRING) {
297 keep = ec_config_dict_get(elt, "keep");
302 if (ec_config_get_type(keep) != EC_CONFIG_TYPE_BOOL) {
306 attr = ec_config_dict_get(elt, "attr");
307 if (attr != NULL && ec_config_get_type(attr) !=
308 EC_CONFIG_TYPE_STRING) {
312 pattern_str = ec_strdup(pattern->string);
313 if (pattern_str == NULL)
315 if (attr != NULL && attr->string != NULL) {
316 attr_name = ec_strdup(attr->string);
317 if (attr_name == NULL)
321 ret = regcomp(&table[n].r, pattern_str, REG_EXTENDED);
324 "Regular expression <%s> compilation failed: %d\n",
326 if (ret == REG_ESPACE)
332 table[n].pattern = pattern_str;
333 table[n].keep = keep->boolean;
334 table[n].attr_name = attr_name;
342 if (priv->child != NULL)
343 ec_node_free(priv->child);
344 priv->child = ec_node_clone(child->node);
345 for (i = 0; i < (ssize_t)priv->len; i++) {
346 ec_free(priv->table[i].pattern);
347 ec_free(priv->table[i].attr_name);
348 regfree(&priv->table[i].r);
350 ec_free(priv->table);
358 for (i = 0; i < n; i++) {
359 if (table[i].pattern != NULL) {
360 ec_free(table[i].pattern);
361 regfree(&table[i].r);
366 ec_free(pattern_str);
370 static struct ec_node_type ec_node_re_lex_type = {
372 .schema = ec_node_re_lex_schema,
373 .set_config = ec_node_re_lex_set_config,
374 .parse = ec_node_re_lex_parse,
375 .complete = ec_complete_unknown,
376 .size = sizeof(struct ec_node_re_lex),
377 .free_priv = ec_node_re_lex_free_priv,
378 .get_children_count = ec_node_re_lex_get_children_count,
379 .get_child = ec_node_re_lex_get_child,
382 EC_NODE_TYPE_REGISTER(ec_node_re_lex_type);
384 int ec_node_re_lex_add(struct ec_node *node, const char *pattern, int keep,
385 const char *attr_name)
387 const struct ec_config *cur_config = NULL;
388 struct ec_config *config = NULL, *patterns = NULL, *elt = NULL;
391 if (ec_node_check_type(node, &ec_node_re_lex_type) < 0)
394 elt = ec_config_dict();
397 if (ec_config_dict_set(elt, "pattern", ec_config_string(pattern)) < 0)
399 if (ec_config_dict_set(elt, "keep", ec_config_bool(keep)) < 0)
401 if (attr_name != NULL) {
402 if (ec_config_dict_set(elt, "attr",
403 ec_config_string(attr_name)) < 0)
407 cur_config = ec_node_get_config(node);
408 if (cur_config == NULL)
409 config = ec_config_dict();
411 config = ec_config_dup(cur_config);
415 patterns = ec_config_dict_get(config, "patterns");
416 if (patterns == NULL) {
417 patterns = ec_config_list();
418 if (patterns == NULL)
421 if (ec_config_dict_set(config, "patterns", patterns) < 0)
422 goto fail; /* patterns list is freed on error */
425 if (ec_config_list_add(patterns, elt) < 0) {
431 ret = ec_node_set_config(node, config);
432 config = NULL; /* freed */
439 ec_config_free(config);
445 ec_node_re_lex_set_child(struct ec_node *node, struct ec_node *child)
447 const struct ec_config *cur_config = NULL;
448 struct ec_config *config = NULL;
451 if (ec_node_check_type(node, &ec_node_re_lex_type) < 0)
454 cur_config = ec_node_get_config(node);
455 if (cur_config == NULL)
456 config = ec_config_dict();
458 config = ec_config_dup(cur_config);
462 if (ec_config_dict_set(config, "child", ec_config_node(child)) < 0) {
463 child = NULL; /* freed */
466 child = NULL; /* freed */
468 ret = ec_node_set_config(node, config);
469 config = NULL; /* freed */
476 ec_config_free(config);
481 struct ec_node *ec_node_re_lex(const char *id, struct ec_node *child)
483 struct ec_node *node = NULL;
488 node = ec_node_from_type(&ec_node_re_lex_type, id);
492 if (ec_node_re_lex_set_child(node, child) < 0) {
493 child = NULL; /* freed */
505 /* LCOV_EXCL_START */
506 static int ec_node_re_lex_testcase(void)
508 struct ec_node *node;
509 int ret, testres = 0;
511 node = ec_node_re_lex(EC_NO_ID,
512 ec_node_many(EC_NO_ID,
514 ec_node_str(EC_NO_ID, "foo"),
515 ec_node_str(EC_NO_ID, "bar"),
516 ec_node_int(EC_NO_ID, 0, 1000, 0)
521 EC_LOG(EC_LOG_ERR, "cannot create node\n");
525 ret = ec_node_re_lex_add(node, "[a-zA-Z]+", 1, NULL);
526 testres |= EC_TEST_CHECK(ret == 0, "cannot add regexp");
527 ret = ec_node_re_lex_add(node, "[0-9]+", 1, NULL);
528 testres |= EC_TEST_CHECK(ret == 0, "cannot add regexp");
529 ret = ec_node_re_lex_add(node, "=", 1, NULL);
530 testres |= EC_TEST_CHECK(ret == 0, "cannot add regexp");
531 ret = ec_node_re_lex_add(node, "-", 1, NULL);
532 testres |= EC_TEST_CHECK(ret == 0, "cannot add regexp");
533 ret = ec_node_re_lex_add(node, "\\+", 1, NULL);
534 testres |= EC_TEST_CHECK(ret == 0, "cannot add regexp");
535 ret = ec_node_re_lex_add(node, "[ ]+", 0, NULL);
536 testres |= EC_TEST_CHECK(ret == 0, "cannot add regexp");
538 EC_LOG(EC_LOG_ERR, "cannot add regexp to node\n");
543 testres |= EC_TEST_CHECK_PARSE(node, 1, " foo bar 324 bar234");
544 testres |= EC_TEST_CHECK_PARSE(node, 1, "foo bar324");
545 testres |= EC_TEST_CHECK_PARSE(node, 1, "");
546 testres |= EC_TEST_CHECK_PARSE(node, -1, "foobar");
549 testres |= EC_TEST_CHECK_COMPLETE(node,
559 static struct ec_test ec_node_re_lex_test = {
560 .name = "node_re_lex",
561 .test = ec_node_re_lex_testcase,
564 EC_TEST_REGISTER(ec_node_re_lex_test);