--- /dev/null
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2016, Olivier MATZ <zer0@droids-corp.org>
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include <stdarg.h>
+#include <ctype.h>
+#include <errno.h>
+
+#include <ecoli_malloc.h>
+#include <ecoli_log.h>
+#include <ecoli_string.h>
+#include <ecoli_test.h>
+#include <ecoli_strvec.h>
+#include <ecoli_node.h>
+#include <ecoli_parse.h>
+#include <ecoli_complete.h>
+#include <ecoli_node_seq.h>
+#include <ecoli_node_str.h>
+#include <ecoli_node_option.h>
+#include <ecoli_node_sh_lex.h>
+
+EC_LOG_TYPE_REGISTER(node_sh_lex);
+
+struct ec_node_sh_lex {
+ struct ec_node gen;
+ struct ec_node *child;
+};
+
+static size_t eat_spaces(const char *str)
+{
+ size_t i = 0;
+
+ /* skip spaces */
+ while (isblank(str[i]))
+ i++;
+
+ return i;
+}
+
+/*
+ * Allocate a new string which is a copy of the input string with quotes
+ * removed. If quotes are not closed properly, set missing_quote to the
+ * missing quote char.
+ */
+static char *unquote_str(const char *str, size_t n, int allow_missing_quote,
+ char *missing_quote)
+{
+ unsigned s = 1, d = 0;
+ char quote = str[0];
+ char *dst;
+ int closed = 0;
+
+ dst = ec_malloc(n);
+ if (dst == NULL) {
+ errno = ENOMEM;
+ return NULL;
+ }
+
+ /* copy string and remove quotes */
+ while (s < n && d < n && str[s] != '\0') {
+ if (str[s] == '\\' && str[s+1] == quote) {
+ dst[d++] = quote;
+ s += 2;
+ continue;
+ }
+ if (str[s] == '\\' && str[s+1] == '\\') {
+ dst[d++] = '\\';
+ s += 2;
+ continue;
+ }
+ if (str[s] == quote) {
+ s++;
+ closed = 1;
+ break;
+ }
+ dst[d++] = str[s++];
+ }
+
+ /* not enough room in dst buffer (should not happen) */
+ if (d >= n) {
+ ec_free(dst);
+ errno = EMSGSIZE;
+ return NULL;
+ }
+
+ /* quote not closed */
+ if (closed == 0) {
+ if (missing_quote != NULL)
+ *missing_quote = str[0];
+ if (allow_missing_quote == 0) {
+ ec_free(dst);
+ errno = EBADMSG;
+ return NULL;
+ }
+ }
+ dst[d++] = '\0';
+
+ return dst;
+}
+
+static size_t eat_quoted_str(const char *str)
+{
+ size_t i = 0;
+ char quote = str[0];
+
+ while (str[i] != '\0') {
+ if (str[i] != '\\' && str[i+1] == quote)
+ return i + 2;
+ i++;
+ }
+
+ /* unclosed quote, will be detected later */
+ return i;
+}
+
+static size_t eat_str(const char *str)
+{
+ size_t i = 0;
+
+ /* eat chars until we find a quote, space, or end of string */
+ while (!isblank(str[i]) && str[i] != '\0' &&
+ str[i] != '"' && str[i] != '\'')
+ i++;
+
+ return i;
+}
+
+static struct ec_strvec *tokenize(const char *str, int completion,
+ int allow_missing_quote, char *missing_quote)
+{
+ struct ec_strvec *strvec = NULL;
+ size_t off = 0, len, suboff, sublen;
+ char *word = NULL, *concat = NULL, *tmp;
+ int last_is_space = 1;
+
+ strvec = ec_strvec();
+ if (strvec == NULL)
+ goto fail;
+
+ while (str[off] != '\0') {
+ if (missing_quote != NULL)
+ *missing_quote = '\0';
+ len = eat_spaces(&str[off]);
+ if (len > 0)
+ last_is_space = 1;
+ off += len;
+
+ len = 0;
+ suboff = off;
+ while (str[suboff] != '\0') {
+ if (missing_quote != NULL)
+ *missing_quote = '\0';
+ last_is_space = 0;
+ if (str[suboff] == '"' || str[suboff] == '\'') {
+ sublen = eat_quoted_str(&str[suboff]);
+ word = unquote_str(&str[suboff], sublen,
+ allow_missing_quote, missing_quote);
+ } else {
+ sublen = eat_str(&str[suboff]);
+ if (sublen == 0)
+ break;
+ word = ec_strndup(&str[suboff], sublen);
+ }
+
+ if (word == NULL)
+ goto fail;
+
+ len += sublen;
+ suboff += sublen;
+
+ if (concat == NULL) {
+ concat = word;
+ word = NULL;
+ } else {
+ tmp = ec_realloc(concat, len + 1);
+ if (tmp == NULL)
+ goto fail;
+ concat = tmp;
+ strcat(concat, word);
+ ec_free(word);
+ word = NULL;
+ }
+ }
+
+ if (concat != NULL) {
+ if (ec_strvec_add(strvec, concat) < 0)
+ goto fail;
+ ec_free(concat);
+ concat = NULL;
+ }
+
+ off += len;
+ }
+
+ /* in completion mode, append an empty string in the vector if
+ * the input string ends with space */
+ if (completion && last_is_space) {
+ if (ec_strvec_add(strvec, "") < 0)
+ goto fail;
+ }
+
+ return strvec;
+
+ fail:
+ ec_free(word);
+ ec_free(concat);
+ ec_strvec_free(strvec);
+ return NULL;
+}
+
+static int
+ec_node_sh_lex_parse(const struct ec_node *gen_node,
+ struct ec_parse *state,
+ const struct ec_strvec *strvec)
+{
+ struct ec_node_sh_lex *node = (struct ec_node_sh_lex *)gen_node;
+ struct ec_strvec *new_vec = NULL;
+ struct ec_parse *child_parse;
+ const char *str;
+ int ret;
+
+ if (ec_strvec_len(strvec) == 0) {
+ new_vec = ec_strvec();
+ } else {
+ str = ec_strvec_val(strvec, 0);
+ new_vec = tokenize(str, 0, 0, NULL);
+ }
+ if (new_vec == NULL && errno == EBADMSG) /* quotes not closed */
+ return EC_PARSE_NOMATCH;
+ if (new_vec == NULL)
+ goto fail;
+
+ ret = ec_node_parse_child(node->child, state, new_vec);
+ if (ret < 0)
+ goto fail;
+
+ if ((unsigned)ret == ec_strvec_len(new_vec)) {
+ ret = 1;
+ } else if (ret != EC_PARSE_NOMATCH) {
+ child_parse = ec_parse_get_last_child(state);
+ ec_parse_unlink_child(state, child_parse);
+ ec_parse_free(child_parse);
+ ret = EC_PARSE_NOMATCH;
+ }
+
+ ec_strvec_free(new_vec);
+ new_vec = NULL;
+
+ return ret;
+
+ fail:
+ ec_strvec_free(new_vec);
+ return -1;
+}
+
+static int
+ec_node_sh_lex_complete(const struct ec_node *gen_node,
+ struct ec_comp *comp,
+ const struct ec_strvec *strvec)
+{
+ struct ec_node_sh_lex *node = (struct ec_node_sh_lex *)gen_node;
+ struct ec_comp *tmp_comp = NULL;
+ struct ec_strvec *new_vec = NULL;
+ struct ec_comp_iter *iter = NULL;
+ struct ec_comp_item *item = NULL;
+ char *new_str = NULL;
+ const char *str;
+ char missing_quote = '\0';
+ int ret;
+
+ if (ec_strvec_len(strvec) != 1)
+ return 0;
+
+ str = ec_strvec_val(strvec, 0);
+ new_vec = tokenize(str, 1, 1, &missing_quote);
+ if (new_vec == NULL)
+ goto fail;
+
+ /* we will store the completions in a temporary struct, because
+ * we want to update them (ex: add missing quotes) */
+ tmp_comp = ec_comp(ec_comp_get_state(comp));
+ if (tmp_comp == NULL)
+ goto fail;
+
+ ret = ec_node_complete_child(node->child, tmp_comp, new_vec);
+ if (ret < 0)
+ goto fail;
+
+ /* add missing quote for full completions */
+ if (missing_quote != '\0') {
+ iter = ec_comp_iter(tmp_comp, EC_COMP_FULL);
+ if (iter == NULL)
+ goto fail;
+ while ((item = ec_comp_iter_next(iter)) != NULL) {
+ str = ec_comp_item_get_str(item);
+ if (ec_asprintf(&new_str, "%c%s%c", missing_quote, str,
+ missing_quote) < 0) {
+ new_str = NULL;
+ goto fail;
+ }
+ if (ec_comp_item_set_str(item, new_str) < 0)
+ goto fail;
+ ec_free(new_str);
+ new_str = NULL;
+
+ str = ec_comp_item_get_completion(item);
+ if (ec_asprintf(&new_str, "%s%c", str,
+ missing_quote) < 0) {
+ new_str = NULL;
+ goto fail;
+ }
+ if (ec_comp_item_set_completion(item, new_str) < 0)
+ goto fail;
+ ec_free(new_str);
+ new_str = NULL;
+ }
+ }
+
+ ec_comp_iter_free(iter);
+ ec_strvec_free(new_vec);
+
+ ec_comp_merge(comp, tmp_comp);
+
+ return 0;
+
+ fail:
+ ec_comp_free(tmp_comp);
+ ec_comp_iter_free(iter);
+ ec_strvec_free(new_vec);
+ ec_free(new_str);
+
+ return -1;
+}
+
+static void ec_node_sh_lex_free_priv(struct ec_node *gen_node)
+{
+ struct ec_node_sh_lex *node = (struct ec_node_sh_lex *)gen_node;
+
+ ec_node_free(node->child);
+}
+
+static size_t
+ec_node_sh_lex_get_children_count(const struct ec_node *gen_node)
+{
+ struct ec_node_sh_lex *node = (struct ec_node_sh_lex *)gen_node;
+
+ if (node->child)
+ return 1;
+ return 0;
+}
+
+static int
+ec_node_sh_lex_get_child(const struct ec_node *gen_node, size_t i,
+ struct ec_node **child, unsigned int *refs)
+{
+ struct ec_node_sh_lex *node = (struct ec_node_sh_lex *)gen_node;
+
+ if (i >= 1)
+ return -1;
+
+ *refs = 1;
+ *child = node->child;
+ return 0;
+}
+
+static struct ec_node_type ec_node_sh_lex_type = {
+ .name = "sh_lex",
+ .parse = ec_node_sh_lex_parse,
+ .complete = ec_node_sh_lex_complete,
+ .size = sizeof(struct ec_node_sh_lex),
+ .free_priv = ec_node_sh_lex_free_priv,
+ .get_children_count = ec_node_sh_lex_get_children_count,
+ .get_child = ec_node_sh_lex_get_child,
+};
+
+EC_NODE_TYPE_REGISTER(ec_node_sh_lex_type);
+
+struct ec_node *ec_node_sh_lex(const char *id, struct ec_node *child)
+{
+ struct ec_node_sh_lex *node = NULL;
+
+ if (child == NULL)
+ return NULL;
+
+ node = (struct ec_node_sh_lex *)ec_node_from_type(&ec_node_sh_lex_type, id);
+ if (node == NULL) {
+ ec_node_free(child);
+ return NULL;
+ }
+
+ node->child = child;
+
+ return &node->gen;
+}
+
+/* LCOV_EXCL_START */
+static int ec_node_sh_lex_testcase(void)
+{
+ struct ec_node *node;
+ int testres = 0;
+
+ node = ec_node_sh_lex(EC_NO_ID,
+ EC_NODE_SEQ(EC_NO_ID,
+ ec_node_str(EC_NO_ID, "foo"),
+ ec_node_option(EC_NO_ID,
+ ec_node_str(EC_NO_ID, "toto")
+ ),
+ ec_node_str(EC_NO_ID, "bar")
+ )
+ );
+ if (node == NULL) {
+ EC_LOG(EC_LOG_ERR, "cannot create node\n");
+ return -1;
+ }
+ testres |= EC_TEST_CHECK_PARSE(node, 1, "foo bar");
+ testres |= EC_TEST_CHECK_PARSE(node, 1, " foo bar");
+ testres |= EC_TEST_CHECK_PARSE(node, 1, " 'foo' \"bar\"");
+ testres |= EC_TEST_CHECK_PARSE(node, 1, " 'f'oo 'toto' bar");
+ testres |= EC_TEST_CHECK_PARSE(node, -1, " foo toto bar'");
+ ec_node_free(node);
+
+ /* test completion */
+ node = ec_node_sh_lex(EC_NO_ID,
+ EC_NODE_SEQ(EC_NO_ID,
+ ec_node_str(EC_NO_ID, "foo"),
+ ec_node_option(EC_NO_ID,
+ ec_node_str(EC_NO_ID, "toto")
+ ),
+ ec_node_str(EC_NO_ID, "bar"),
+ ec_node_str(EC_NO_ID, "titi")
+ )
+ );
+ if (node == NULL) {
+ EC_LOG(EC_LOG_ERR, "cannot create node\n");
+ return -1;
+ }
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "", EC_NODE_ENDLIST,
+ "foo", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ " ", EC_NODE_ENDLIST,
+ "foo", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "f", EC_NODE_ENDLIST,
+ "foo", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo", EC_NODE_ENDLIST,
+ "foo", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo ", EC_NODE_ENDLIST,
+ "bar", "toto", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo t", EC_NODE_ENDLIST,
+ "toto", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo b", EC_NODE_ENDLIST,
+ "bar", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo bar", EC_NODE_ENDLIST,
+ "bar", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo bar ", EC_NODE_ENDLIST,
+ "titi", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo toto bar ", EC_NODE_ENDLIST,
+ "titi", EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "x", EC_NODE_ENDLIST,
+ EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo barx", EC_NODE_ENDLIST,
+ EC_NODE_ENDLIST);
+ testres |= EC_TEST_CHECK_COMPLETE(node,
+ "foo 'b", EC_NODE_ENDLIST,
+ "'bar'", EC_NODE_ENDLIST);
+
+ ec_node_free(node);
+ return testres;
+}
+/* LCOV_EXCL_STOP */
+
+static struct ec_test ec_node_sh_lex_test = {
+ .name = "node_sh_lex",
+ .test = ec_node_sh_lex_testcase,
+};
+
+EC_TEST_REGISTER(ec_node_sh_lex_test);