summaryrefslogtreecommitdiffstats
path: root/src/tree_schema.c
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--src/tree_schema.c2178
1 files changed, 2178 insertions, 0 deletions
diff --git a/src/tree_schema.c b/src/tree_schema.c
new file mode 100644
index 0000000..5c897bf
--- /dev/null
+++ b/src/tree_schema.c
@@ -0,0 +1,2178 @@
+/**
+ * @file tree_schema.c
+ * @author Radek Krejci <rkrejci@cesnet.cz>
+ * @brief Schema tree implementation
+ *
+ * Copyright (c) 2015 - 2018 CESNET, z.s.p.o.
+ *
+ * This source code is licensed under BSD 3-Clause License (the "License").
+ * You may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ * https://opensource.org/licenses/BSD-3-Clause
+ */
+
+#define _GNU_SOURCE /* asprintf, strdup */
+
+#include "tree_schema.h"
+
+#include <assert.h>
+#include <ctype.h>
+#include <dirent.h>
+#include <errno.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <unistd.h>
+
+#include "common.h"
+#include "compat.h"
+#include "context.h"
+#include "dict.h"
+#include "in.h"
+#include "in_internal.h"
+#include "log.h"
+#include "parser_internal.h"
+#include "parser_schema.h"
+#include "path.h"
+#include "plugins_internal.h"
+#include "schema_compile.h"
+#include "schema_compile_amend.h"
+#include "schema_features.h"
+#include "set.h"
+#include "tree.h"
+#include "tree_edit.h"
+#include "tree_schema_free.h"
+#include "tree_schema_internal.h"
+#include "xpath.h"
+
+const char * const ly_devmod_list[] = {
+ [LYS_DEV_NOT_SUPPORTED] = "not-supported",
+ [LYS_DEV_ADD] = "add",
+ [LYS_DEV_DELETE] = "delete",
+ [LYS_DEV_REPLACE] = "replace",
+};
+
+LIBYANG_API_DEF LY_ERR
+lysc_tree_dfs_full(const struct lysc_node *root, lysc_dfs_clb dfs_clb, void *data)
+{
+ struct lysc_node *elem, *elem2;
+ const struct lysc_node_action *action;
+ const struct lysc_node_notif *notif;
+
+ LY_CHECK_ARG_RET(NULL, root, dfs_clb, LY_EINVAL);
+
+ LYSC_TREE_DFS_BEGIN(root, elem) {
+ /* schema node */
+ LY_CHECK_RET(dfs_clb(elem, data, &LYSC_TREE_DFS_continue));
+
+ LY_LIST_FOR(lysc_node_actions(elem), action) {
+ LYSC_TREE_DFS_BEGIN(action, elem2) {
+ /* action subtree */
+ LY_CHECK_RET(dfs_clb(elem2, data, &LYSC_TREE_DFS_continue));
+
+ LYSC_TREE_DFS_END(action, elem2);
+ }
+ }
+
+ LY_LIST_FOR(lysc_node_notifs(elem), notif) {
+ LYSC_TREE_DFS_BEGIN(notif, elem2) {
+ /* notification subtree */
+ LY_CHECK_RET(dfs_clb(elem2, data, &LYSC_TREE_DFS_continue));
+
+ LYSC_TREE_DFS_END(notif, elem2);
+ }
+ }
+
+ LYSC_TREE_DFS_END(root, elem);
+ }
+
+ return LY_SUCCESS;
+}
+
+LIBYANG_API_DEF LY_ERR
+lysc_module_dfs_full(const struct lys_module *mod, lysc_dfs_clb dfs_clb, void *data)
+{
+ const struct lysc_node *root;
+
+ LY_CHECK_ARG_RET(NULL, mod, mod->compiled, dfs_clb, LY_EINVAL);
+
+ /* schema nodes */
+ LY_LIST_FOR(mod->compiled->data, root) {
+ LY_CHECK_RET(lysc_tree_dfs_full(root, dfs_clb, data));
+ }
+
+ /* RPCs */
+ LY_LIST_FOR((const struct lysc_node *)mod->compiled->rpcs, root) {
+ LY_CHECK_RET(lysc_tree_dfs_full(root, dfs_clb, data));
+ }
+
+ /* notifications */
+ LY_LIST_FOR((const struct lysc_node *)mod->compiled->notifs, root) {
+ LY_CHECK_RET(lysc_tree_dfs_full(root, dfs_clb, data));
+ }
+
+ return LY_SUCCESS;
+}
+
+static void
+lys_getnext_into_case(const struct lysc_node_case *first_case, const struct lysc_node **last, const struct lysc_node **next)
+{
+ for ( ; first_case; first_case = (const struct lysc_node_case *)first_case->next) {
+ if (first_case->child) {
+ /* there is something to return */
+ (*next) = first_case->child;
+ return;
+ }
+ }
+
+ /* no children in choice's cases, so go to the choice's sibling instead of into it */
+ (*last) = (*next);
+ (*next) = (*next)->next;
+}
+
+/**
+ * @brief Generic getnext function for ::lys_getnext() and ::lys_getnext_ext().
+ *
+ * Gets next schema tree (sibling) node element that can be instantiated in a data tree. Returned node can
+ * be from an augment. If the @p ext is provided, the function is locked inside the schema tree defined in the
+ * extension instance.
+ *
+ * ::lys_getnext_() is supposed to be called sequentially. In the first call, the @p last parameter is usually NULL
+ * and function starts returning i) the first @p parent's child or ii) the first top level element specified in the
+ * given extension (if provided) or iii) the first top level element of the @p module.
+ * Consequent calls suppose to provide the previously returned node as the @p last parameter and still the same
+ * @p parent and @p module parameters.
+ *
+ * Without options, the function is used to traverse only the schema nodes that can be paired with corresponding
+ * data nodes in a data tree. By setting some @p options the behavior can be modified to the extent that
+ * all the schema nodes are iteratively returned.
+ *
+ * @param[in] last Previously returned schema tree node, or NULL in case of the first call.
+ * @param[in] parent Parent of the subtree where the function starts processing.
+ * @param[in] module In case of iterating on top level elements, the @p parent is NULL and
+ * module must be specified.
+ * @param[in] ext The extension instance to provide a separate schema tree. To consider the top level elements in the tree,
+ * the @p parent must be NULL. Anyway, at least one of @p parent, @p module and @p ext parameters must be specified.
+ * @param[in] options [ORed options](@ref sgetnextflags).
+ * @return Next schema tree node that can be instantiated in a data tree, NULL in case there is no such element.
+ */
+static const struct lysc_node *
+lys_getnext_(const struct lysc_node *last, const struct lysc_node *parent, const struct lysc_module *module,
+ const struct lysc_ext_instance *ext, uint32_t options)
+{
+ const struct lysc_node *next = NULL;
+ ly_bool action_flag = 0, notif_flag = 0;
+
+ LY_CHECK_ARG_RET(NULL, parent || module || ext, NULL);
+
+next:
+ if (!last) {
+ /* first call */
+
+ /* get know where to start */
+ if (parent) {
+ /* schema subtree */
+ next = last = lysc_node_child(parent);
+ } else {
+ /* top level data */
+ if (ext) {
+ lyplg_ext_get_storage(ext, LY_STMT_DATA_NODE_MASK, sizeof last, (const void **)&last);
+ next = last;
+ } else {
+ next = last = module->data;
+ }
+ }
+ if (!next) {
+ /* try to get action or notification */
+ goto repeat;
+ }
+ /* test if the next can be returned */
+ goto check;
+
+ } else if (last->nodetype & (LYS_RPC | LYS_ACTION)) {
+ action_flag = 1;
+ next = last->next;
+ } else if (last->nodetype == LYS_NOTIF) {
+ action_flag = notif_flag = 1;
+ next = last->next;
+ } else {
+ next = last->next;
+ }
+
+repeat:
+ if (!next) {
+ /* possibly go back to parent */
+ if (last && (last->parent != parent)) {
+ last = last->parent;
+ goto next;
+ } else if (!action_flag) {
+ action_flag = 1;
+ if (ext) {
+ lyplg_ext_get_storage(ext, LY_STMT_OP_MASK, sizeof next, (const void **)&next);
+ } else if (parent) {
+ next = (struct lysc_node *)lysc_node_actions(parent);
+ } else {
+ next = (struct lysc_node *)module->rpcs;
+ }
+ } else if (!notif_flag) {
+ notif_flag = 1;
+ if (ext) {
+ lyplg_ext_get_storage(ext, LY_STMT_NOTIFICATION, sizeof next, (const void **)&next);
+ } else if (parent) {
+ next = (struct lysc_node *)lysc_node_notifs(parent);
+ } else {
+ next = (struct lysc_node *)module->notifs;
+ }
+ } else {
+ return NULL;
+ }
+ goto repeat;
+ }
+check:
+ switch (next->nodetype) {
+ case LYS_RPC:
+ case LYS_ACTION:
+ case LYS_NOTIF:
+ case LYS_LEAF:
+ case LYS_ANYXML:
+ case LYS_ANYDATA:
+ case LYS_LIST:
+ case LYS_LEAFLIST:
+ break;
+ case LYS_CASE:
+ if (options & LYS_GETNEXT_WITHCASE) {
+ break;
+ } else {
+ /* go into */
+ lys_getnext_into_case((const struct lysc_node_case *)next, &last, &next);
+ }
+ goto repeat;
+ case LYS_CONTAINER:
+ if (!(next->flags & LYS_PRESENCE) && (options & LYS_GETNEXT_INTONPCONT)) {
+ if (lysc_node_child(next)) {
+ /* go into */
+ next = lysc_node_child(next);
+ } else {
+ last = next;
+ next = next->next;
+ }
+ goto repeat;
+ }
+ break;
+ case LYS_CHOICE:
+ if (options & LYS_GETNEXT_WITHCHOICE) {
+ break;
+ } else if ((options & LYS_GETNEXT_NOCHOICE) || !lysc_node_child(next)) {
+ next = next->next;
+ } else {
+ if (options & LYS_GETNEXT_WITHCASE) {
+ next = lysc_node_child(next);
+ } else {
+ /* go into */
+ lys_getnext_into_case(((struct lysc_node_choice *)next)->cases, &last, &next);
+ }
+ }
+ goto repeat;
+ case LYS_INPUT:
+ if (options & LYS_GETNEXT_OUTPUT) {
+ /* skip */
+ next = next->next;
+ } else {
+ /* go into */
+ next = lysc_node_child(next);
+ }
+ goto repeat;
+ case LYS_OUTPUT:
+ if (!(options & LYS_GETNEXT_OUTPUT)) {
+ /* skip */
+ next = next->next;
+ } else {
+ /* go into */
+ next = lysc_node_child(next);
+ }
+ goto repeat;
+ default:
+ /* we should not be here */
+ LOGINT(module ? module->mod->ctx : parent ? parent->module->ctx : ext->module->ctx);
+ return NULL;
+ }
+
+ return next;
+}
+
+LIBYANG_API_DEF const struct lysc_node *
+lys_getnext(const struct lysc_node *last, const struct lysc_node *parent, const struct lysc_module *module, uint32_t options)
+{
+ return lys_getnext_(last, parent, module, NULL, options);
+}
+
+LIBYANG_API_DEF const struct lysc_node *
+lys_getnext_ext(const struct lysc_node *last, const struct lysc_node *parent, const struct lysc_ext_instance *ext, uint32_t options)
+{
+ return lys_getnext_(last, parent, NULL, ext, options);
+}
+
+const struct lysc_node *
+lysc_ext_find_node(const struct lysc_ext_instance *ext, const struct lys_module *module, const char *name, size_t name_len,
+ uint16_t nodetype, uint32_t options)
+{
+ const struct lysc_node *node = NULL;
+
+ LY_CHECK_ARG_RET(NULL, ext, name, NULL);
+ if (!nodetype) {
+ nodetype = LYS_NODETYPE_MASK;
+ }
+
+ if (module && (module != ext->module)) {
+ return NULL;
+ }
+
+ while ((node = lys_getnext_ext(node, NULL, ext, options))) {
+ if (!(node->nodetype & nodetype)) {
+ continue;
+ }
+
+ if (name_len) {
+ if (!ly_strncmp(node->name, name, name_len)) {
+ return node;
+ }
+ } else {
+ if (!strcmp(node->name, name)) {
+ return node;
+ }
+ }
+ }
+ return NULL;
+}
+
+LIBYANG_API_DEF const struct lysc_node *
+lys_find_child(const struct lysc_node *parent, const struct lys_module *module, const char *name, size_t name_len,
+ uint16_t nodetype, uint32_t options)
+{
+ const struct lysc_node *node = NULL;
+
+ LY_CHECK_ARG_RET(NULL, module, name, NULL);
+ LY_CHECK_CTX_EQUAL_RET(parent ? parent->module->ctx : NULL, module->ctx, NULL);
+ if (!nodetype) {
+ nodetype = LYS_NODETYPE_MASK;
+ }
+
+ while ((node = lys_getnext(node, parent, module->compiled, options))) {
+ if (!(node->nodetype & nodetype)) {
+ continue;
+ }
+ if (node->module != module) {
+ continue;
+ }
+
+ if (name_len) {
+ if (!ly_strncmp(node->name, name, name_len)) {
+ return node;
+ }
+ } else {
+ if (!strcmp(node->name, name)) {
+ return node;
+ }
+ }
+ }
+
+ return NULL;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_find_xpath_atoms(const struct ly_ctx *ctx, const struct lysc_node *ctx_node, const char *xpath, uint32_t options,
+ struct ly_set **set)
+{
+ LY_ERR ret = LY_SUCCESS;
+ struct lyxp_set xp_set;
+ struct lyxp_expr *exp = NULL;
+ uint32_t i;
+
+ LY_CHECK_ARG_RET(NULL, ctx || ctx_node, xpath, set, LY_EINVAL);
+ LY_CHECK_CTX_EQUAL_RET(ctx, ctx_node ? ctx_node->module->ctx : NULL, LY_EINVAL);
+ if (!(options & LYXP_SCNODE_ALL)) {
+ options |= LYXP_SCNODE;
+ }
+ if (!ctx) {
+ ctx = ctx_node->module->ctx;
+ }
+
+ memset(&xp_set, 0, sizeof xp_set);
+
+ /* compile expression */
+ ret = lyxp_expr_parse(ctx, xpath, 0, 1, &exp);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* atomize expression */
+ ret = lyxp_atomize(ctx, exp, NULL, LY_VALUE_JSON, NULL, ctx_node, ctx_node, &xp_set, options);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* allocate return set */
+ ret = ly_set_new(set);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* transform into ly_set */
+ (*set)->objs = malloc(xp_set.used * sizeof *(*set)->objs);
+ LY_CHECK_ERR_GOTO(!(*set)->objs, LOGMEM(ctx); ret = LY_EMEM, cleanup);
+ (*set)->size = xp_set.used;
+
+ for (i = 0; i < xp_set.used; ++i) {
+ if (xp_set.val.scnodes[i].type == LYXP_NODE_ELEM) {
+ ret = ly_set_add(*set, xp_set.val.scnodes[i].scnode, 1, NULL);
+ LY_CHECK_GOTO(ret, cleanup);
+ }
+ }
+
+cleanup:
+ lyxp_set_free_content(&xp_set);
+ lyxp_expr_free(ctx, exp);
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_find_expr_atoms(const struct lysc_node *ctx_node, const struct lys_module *cur_mod, const struct lyxp_expr *expr,
+ const struct lysc_prefix *prefixes, uint32_t options, struct ly_set **set)
+{
+ LY_ERR ret = LY_SUCCESS;
+ struct lyxp_set xp_set = {0};
+ uint32_t i;
+
+ LY_CHECK_ARG_RET(NULL, cur_mod, expr, prefixes, set, LY_EINVAL);
+ LY_CHECK_CTX_EQUAL_RET(ctx_node ? ctx_node->module->ctx : NULL, cur_mod->ctx, LY_EINVAL);
+ if (!(options & LYXP_SCNODE_ALL)) {
+ options = LYXP_SCNODE;
+ }
+
+ /* atomize expression */
+ ret = lyxp_atomize(cur_mod->ctx, expr, cur_mod, LY_VALUE_SCHEMA_RESOLVED, (void *)prefixes, ctx_node, ctx_node,
+ &xp_set, options);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* allocate return set */
+ ret = ly_set_new(set);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* transform into ly_set */
+ (*set)->objs = malloc(xp_set.used * sizeof *(*set)->objs);
+ LY_CHECK_ERR_GOTO(!(*set)->objs, LOGMEM(cur_mod->ctx); ret = LY_EMEM, cleanup);
+ (*set)->size = xp_set.used;
+
+ for (i = 0; i < xp_set.used; ++i) {
+ if ((xp_set.val.scnodes[i].type == LYXP_NODE_ELEM) && (xp_set.val.scnodes[i].in_ctx >= LYXP_SET_SCNODE_ATOM_NODE)) {
+ assert((xp_set.val.scnodes[i].in_ctx == LYXP_SET_SCNODE_ATOM_NODE) ||
+ (xp_set.val.scnodes[i].in_ctx == LYXP_SET_SCNODE_ATOM_VAL) ||
+ (xp_set.val.scnodes[i].in_ctx == LYXP_SET_SCNODE_ATOM_CTX));
+ ret = ly_set_add(*set, xp_set.val.scnodes[i].scnode, 1, NULL);
+ LY_CHECK_GOTO(ret, cleanup);
+ }
+ }
+
+cleanup:
+ lyxp_set_free_content(&xp_set);
+ if (ret) {
+ ly_set_free(*set, NULL);
+ *set = NULL;
+ }
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_find_xpath(const struct ly_ctx *ctx, const struct lysc_node *ctx_node, const char *xpath, uint32_t options,
+ struct ly_set **set)
+{
+ LY_ERR ret = LY_SUCCESS;
+ struct lyxp_set xp_set = {0};
+ struct lyxp_expr *exp = NULL;
+ uint32_t i;
+
+ LY_CHECK_ARG_RET(NULL, ctx || ctx_node, xpath, set, LY_EINVAL);
+ LY_CHECK_CTX_EQUAL_RET(ctx, ctx_node ? ctx_node->module->ctx : NULL, LY_EINVAL);
+ if (!(options & LYXP_SCNODE_ALL)) {
+ options = LYXP_SCNODE;
+ }
+ if (!ctx) {
+ ctx = ctx_node->module->ctx;
+ }
+
+ /* compile expression */
+ ret = lyxp_expr_parse(ctx, xpath, 0, 1, &exp);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* atomize expression */
+ ret = lyxp_atomize(ctx, exp, NULL, LY_VALUE_JSON, NULL, ctx_node, ctx_node, &xp_set, options);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* allocate return set */
+ ret = ly_set_new(set);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* transform into ly_set */
+ (*set)->objs = malloc(xp_set.used * sizeof *(*set)->objs);
+ LY_CHECK_ERR_GOTO(!(*set)->objs, LOGMEM(ctx); ret = LY_EMEM, cleanup);
+ (*set)->size = xp_set.used;
+
+ for (i = 0; i < xp_set.used; ++i) {
+ if ((xp_set.val.scnodes[i].type == LYXP_NODE_ELEM) && (xp_set.val.scnodes[i].in_ctx == LYXP_SET_SCNODE_ATOM_CTX)) {
+ ret = ly_set_add(*set, xp_set.val.scnodes[i].scnode, 1, NULL);
+ LY_CHECK_GOTO(ret, cleanup);
+ }
+ }
+
+cleanup:
+ lyxp_set_free_content(&xp_set);
+ lyxp_expr_free(ctx, exp);
+ if (ret) {
+ ly_set_free(*set, NULL);
+ *set = NULL;
+ }
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_find_lypath_atoms(const struct ly_path *path, struct ly_set **set)
+{
+ LY_ERR ret = LY_SUCCESS;
+ LY_ARRAY_COUNT_TYPE u, v;
+
+ LY_CHECK_ARG_RET(NULL, path, set, LY_EINVAL);
+
+ /* allocate return set */
+ LY_CHECK_RET(ly_set_new(set));
+
+ LY_ARRAY_FOR(path, u) {
+ /* add nodes from the path */
+ LY_CHECK_GOTO(ret = ly_set_add(*set, (void *)path[u].node, 0, NULL), cleanup);
+ if (path[u].pred_type == LY_PATH_PREDTYPE_LIST) {
+ LY_ARRAY_FOR(path[u].predicates, v) {
+ /* add all the keys in a predicate */
+ LY_CHECK_GOTO(ret = ly_set_add(*set, (void *)path[u].predicates[v].key, 0, NULL), cleanup);
+ }
+ }
+ }
+
+cleanup:
+ if (ret) {
+ ly_set_free(*set, NULL);
+ *set = NULL;
+ }
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_find_path_atoms(const struct ly_ctx *ctx, const struct lysc_node *ctx_node, const char *path, ly_bool output,
+ struct ly_set **set)
+{
+ LY_ERR ret = LY_SUCCESS;
+ uint8_t oper;
+ struct lyxp_expr *expr = NULL;
+ struct ly_path *p = NULL;
+
+ LY_CHECK_ARG_RET(ctx, ctx || ctx_node, path, set, LY_EINVAL);
+ LY_CHECK_CTX_EQUAL_RET(ctx, ctx_node ? ctx_node->module->ctx : NULL, LY_EINVAL);
+
+ if (!ctx) {
+ ctx = ctx_node->module->ctx;
+ }
+
+ /* parse */
+ ret = lyxp_expr_parse(ctx, path, strlen(path), 0, &expr);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* compile */
+ oper = output ? LY_PATH_OPER_OUTPUT : LY_PATH_OPER_INPUT;
+ ret = ly_path_compile(ctx, NULL, ctx_node, NULL, expr, oper, LY_PATH_TARGET_MANY, 0, LY_VALUE_JSON, NULL, &p);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* resolve */
+ ret = lys_find_lypath_atoms(p, set);
+
+cleanup:
+ ly_path_free(ctx, p);
+ lyxp_expr_free(ctx, expr);
+ return ret;
+}
+
+LIBYANG_API_DEF const struct lysc_node *
+lys_find_path(const struct ly_ctx *ctx, const struct lysc_node *ctx_node, const char *path, ly_bool output)
+{
+ const struct lysc_node *snode = NULL;
+ struct lyxp_expr *exp = NULL;
+ struct ly_path *p = NULL;
+ LY_ERR ret;
+ uint8_t oper;
+
+ LY_CHECK_ARG_RET(ctx, ctx || ctx_node, NULL);
+ LY_CHECK_CTX_EQUAL_RET(ctx, ctx_node ? ctx_node->module->ctx : NULL, NULL);
+
+ if (!ctx) {
+ ctx = ctx_node->module->ctx;
+ }
+
+ /* parse */
+ ret = lyxp_expr_parse(ctx, path, strlen(path), 0, &exp);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* compile */
+ oper = output ? LY_PATH_OPER_OUTPUT : LY_PATH_OPER_INPUT;
+ ret = ly_path_compile(ctx, NULL, ctx_node, NULL, exp, oper, LY_PATH_TARGET_MANY, 0, LY_VALUE_JSON, NULL, &p);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* get last node */
+ snode = p[LY_ARRAY_COUNT(p) - 1].node;
+
+cleanup:
+ ly_path_free(ctx, p);
+ lyxp_expr_free(ctx, exp);
+ return snode;
+}
+
+char *
+lysc_path_until(const struct lysc_node *node, const struct lysc_node *parent, LYSC_PATH_TYPE pathtype, char *buffer,
+ size_t buflen)
+{
+ const struct lysc_node *iter, *par, *key;
+ char *path = NULL;
+ int len = 0;
+ ly_bool skip_schema;
+
+ if (buffer) {
+ LY_CHECK_ARG_RET(node->module->ctx, buflen > 1, NULL);
+ buffer[0] = '\0';
+ }
+
+ if ((pathtype == LYSC_PATH_DATA) || (pathtype == LYSC_PATH_DATA_PATTERN)) {
+ /* skip schema-only nodes */
+ skip_schema = 1;
+ } else {
+ skip_schema = 0;
+ }
+
+ for (iter = node; iter && (iter != parent) && (len >= 0); iter = iter->parent) {
+ char *s;
+ const char *slash;
+
+ if (skip_schema && (iter->nodetype & (LYS_CHOICE | LYS_CASE | LYS_INPUT | LYS_OUTPUT))) {
+ /* schema-only node */
+ continue;
+ }
+
+ if ((pathtype == LYSC_PATH_DATA_PATTERN) && (iter->nodetype == LYS_LIST)) {
+ key = NULL;
+ while ((key = lys_getnext(key, iter, NULL, 0)) && lysc_is_key(key)) {
+ s = buffer ? strdup(buffer) : path;
+
+ /* print key predicate */
+ if (buffer) {
+ len = snprintf(buffer, buflen, "[%s='%%s']%s", key->name, s ? s : "");
+ } else {
+ len = asprintf(&path, "[%s='%%s']%s", key->name, s ? s : "");
+ }
+ free(s);
+
+ if (buffer && (buflen <= (size_t)len)) {
+ /* not enough space in buffer */
+ break;
+ }
+ }
+ }
+
+ s = buffer ? strdup(buffer) : path;
+ if (parent && (iter->parent == parent)) {
+ slash = "";
+ } else {
+ slash = "/";
+ }
+
+ if (skip_schema) {
+ par = lysc_data_parent(iter);
+ } else {
+ par = iter->parent;
+ }
+
+ if (!par || (par->module != iter->module)) {
+ /* print prefix */
+ if (buffer) {
+ len = snprintf(buffer, buflen, "%s%s:%s%s", slash, iter->module->name, iter->name, s ? s : "");
+ } else {
+ len = asprintf(&path, "%s%s:%s%s", slash, iter->module->name, iter->name, s ? s : "");
+ }
+ } else {
+ /* prefix is the same as in parent */
+ if (buffer) {
+ len = snprintf(buffer, buflen, "%s%s%s", slash, iter->name, s ? s : "");
+ } else {
+ len = asprintf(&path, "%s%s%s", slash, iter->name, s ? s : "");
+ }
+ }
+ free(s);
+
+ if (buffer && (buflen <= (size_t)len)) {
+ /* not enough space in buffer */
+ break;
+ }
+ }
+
+ if (len < 0) {
+ free(path);
+ path = NULL;
+ } else if (len == 0) {
+ if (buffer) {
+ strcpy(buffer, "/");
+ } else {
+ path = strdup("/");
+ }
+ }
+
+ if (buffer) {
+ return buffer;
+ } else {
+ return path;
+ }
+}
+
+LIBYANG_API_DEF char *
+lysc_path(const struct lysc_node *node, LYSC_PATH_TYPE pathtype, char *buffer, size_t buflen)
+{
+ return lysc_path_until(node, NULL, pathtype, buffer, buflen);
+}
+
+LY_ERR
+_lys_set_implemented(struct lys_module *mod, const char **features, struct lys_glob_unres *unres)
+{
+ LY_ERR ret = LY_SUCCESS, r;
+ struct lys_module *mod_iter;
+ const char **imp_f, *all_f[] = {"*", NULL};
+ uint32_t i;
+
+ if (mod->implemented) {
+ /* mod is already implemented, set the features */
+ r = lys_set_features(mod->parsed, features);
+ if (r == LY_EEXIST) {
+ /* no changes */
+ return LY_SUCCESS;
+ } else if (!r) {
+ /* mark the module as changed */
+ mod->to_compile = 1;
+ }
+
+ return r;
+ }
+
+ /* implement, ignore recompilation because it must always take place later */
+ r = lys_implement(mod, features, unres);
+ LY_CHECK_ERR_GOTO(r && (r != LY_ERECOMPILE), ret = r, cleanup);
+
+ if (mod->ctx->flags & LY_CTX_ALL_IMPLEMENTED) {
+ /* implement all the imports as well */
+ for (i = 0; i < unres->creating.count; ++i) {
+ mod = unres->creating.objs[i];
+ if (mod->implemented) {
+ continue;
+ }
+
+ imp_f = (mod->ctx->flags & LY_CTX_ENABLE_IMP_FEATURES) ? all_f : NULL;
+ r = lys_implement(mod, imp_f, unres);
+ LY_CHECK_ERR_GOTO(r && (r != LY_ERECOMPILE), ret = r, cleanup);
+ }
+ }
+
+ /* Try to find module with LYS_MOD_IMPORTED_REV flag. */
+ i = 0;
+ while ((mod_iter = ly_ctx_get_module_iter(mod->ctx, &i))) {
+ if (!strcmp(mod_iter->name, mod->name) && (mod_iter != mod) && (mod_iter->latest_revision & LYS_MOD_IMPORTED_REV)) {
+ LOGVRB("Implemented module \"%s@%s\" was not and will not be imported if the revision-date is missing"
+ " in the import statement. Instead, the revision \"%s\" is imported.", mod->name, mod->revision,
+ mod_iter->revision);
+ break;
+ }
+ }
+
+cleanup:
+ return ret;
+}
+
+/**
+ * @brief Check whether it may be needed to (re)compile a module from a particular dependency set
+ * and if so, add it into its dep set.
+ *
+ * Dependency set includes all modules that need to be (re)compiled in case any of the module(s)
+ * in the dep set are (re)compiled.
+ *
+ * The reason for recompilation is possible disabled nodes and updating
+ * leafref targets to point to the newly compiled modules. Using the import relation, the
+ * dependency is reflexive because of possible foreign augments and deviations, which are compiled
+ * during the target module compilation.
+ *
+ * - every module must belong to exactly one dep set
+ * - implement flag must be ignored because it can be changed during dep set compilation
+ *
+ * @param[in] mod Module to process.
+ * @param[in,out] ctx_set Set with all not-yet-processed modules.
+ * @param[in,out] dep_set Current dependency set to update.
+ * @param[in] aux_set Set of traversed non-compiled modules, should be empty on first call.
+ * @return LY_ERR value.
+ */
+static LY_ERR
+lys_unres_dep_sets_create_mod_r(struct lys_module *mod, struct ly_set *ctx_set, struct ly_set *dep_set,
+ struct ly_set *aux_set)
+{
+ struct lys_module *mod2;
+ struct lysp_import *imports;
+ uint32_t i;
+ LY_ARRAY_COUNT_TYPE u, v;
+ ly_bool found;
+
+ if (LYS_IS_SINGLE_DEP_SET(mod)) {
+ /* is already in a separate dep set */
+ if (!lys_has_dep_mods(mod)) {
+ /* break the dep set here, no modules depend on this one */
+ return LY_SUCCESS;
+ }
+
+ if (ly_set_contains(aux_set, mod, NULL)) {
+ /* it was traversed */
+ return LY_SUCCESS;
+ }
+
+ /* add a new auxiliary module */
+ LY_CHECK_RET(ly_set_add(aux_set, mod, 1, NULL));
+ } else {
+ if (!ly_set_contains(ctx_set, mod, &i)) {
+ /* it was already processed */
+ return LY_SUCCESS;
+ }
+
+ /* remove it from the set, we are processing it now */
+ ly_set_rm_index(ctx_set, i, NULL);
+
+ /* add a new dependent module into the dep set */
+ LY_CHECK_RET(ly_set_add(dep_set, mod, 1, NULL));
+ }
+
+ /* process imports of the module and submodules */
+ imports = mod->parsed->imports;
+ LY_ARRAY_FOR(imports, u) {
+ mod2 = imports[u].module;
+ LY_CHECK_RET(lys_unres_dep_sets_create_mod_r(mod2, ctx_set, dep_set, aux_set));
+ }
+ LY_ARRAY_FOR(mod->parsed->includes, v) {
+ imports = mod->parsed->includes[v].submodule->imports;
+ LY_ARRAY_FOR(imports, u) {
+ mod2 = imports[u].module;
+ if (LYS_IS_SINGLE_DEP_SET(mod2) && !lys_has_dep_mods(mod2)) {
+ /* break the dep set here, no modules depend on this one */
+ continue;
+ }
+
+ LY_CHECK_RET(lys_unres_dep_sets_create_mod_r(imports[u].module, ctx_set, dep_set, aux_set));
+ }
+ }
+
+ /* process modules and submodules importing this module */
+ for (i = 0; i < mod->ctx->list.count; ++i) {
+ mod2 = mod->ctx->list.objs[i];
+ found = 0;
+
+ imports = mod2->parsed->imports;
+ LY_ARRAY_FOR(imports, u) {
+ if (imports[u].module == mod) {
+ found = 1;
+ break;
+ }
+ }
+
+ if (!found) {
+ LY_ARRAY_FOR(mod2->parsed->includes, v) {
+ imports = mod2->parsed->includes[v].submodule->imports;
+ LY_ARRAY_FOR(imports, u) {
+ if (imports[u].module == mod) {
+ found = 1;
+ break;
+ }
+ }
+
+ if (found) {
+ break;
+ }
+ }
+ }
+
+ if (found) {
+ LY_CHECK_RET(lys_unres_dep_sets_create_mod_r(mod2, ctx_set, dep_set, aux_set));
+ }
+ }
+
+ return LY_SUCCESS;
+}
+
+/**
+ * @brief Add all simple modules (that have nothing to (re)compile) into separate dep sets.
+ *
+ * @param[in,out] ctx_set Set with all not-yet-processed modules.
+ * @param[in,out] main_set Set of dependency module sets.
+ * @return LY_ERR value.
+ */
+static LY_ERR
+lys_unres_dep_sets_create_single(struct ly_set *ctx_set, struct ly_set *main_set)
+{
+ LY_ERR ret = LY_SUCCESS;
+ struct lys_module *m;
+ uint32_t i = 0;
+ struct ly_set *dep_set = NULL;
+
+ while (i < ctx_set->count) {
+ m = ctx_set->objs[i];
+ if (LYS_IS_SINGLE_DEP_SET(m)) {
+ /* remove it from the set, we are processing it now */
+ ly_set_rm_index(ctx_set, i, NULL);
+
+ /* this module can be in a separate dep set (but there still may be modules importing this one
+ * that depend on imports of this one in case it defines groupings) */
+ LY_CHECK_GOTO(ret = ly_set_new(&dep_set), cleanup);
+ LY_CHECK_GOTO(ret = ly_set_add(dep_set, m, 1, NULL), cleanup);
+ LY_CHECK_GOTO(ret = ly_set_add(main_set, dep_set, 1, NULL), cleanup);
+ dep_set = NULL;
+ } else {
+ ++i;
+ }
+ }
+
+cleanup:
+ ly_set_free(dep_set, NULL);
+ return ret;
+}
+
+LY_ERR
+lys_unres_dep_sets_create(struct ly_ctx *ctx, struct ly_set *main_set, struct lys_module *mod)
+{
+ LY_ERR ret = LY_SUCCESS;
+ struct lys_module *m;
+ struct ly_set *dep_set = NULL, *ctx_set = NULL, aux_set = {0};
+ uint32_t i;
+ ly_bool found;
+
+ assert(!main_set->count);
+
+ /* start with a duplicate set of modules that we will remove from */
+ LY_CHECK_GOTO(ret = ly_set_dup(&ctx->list, NULL, &ctx_set), cleanup);
+
+ /* first create all dep sets with single modules */
+ LY_CHECK_GOTO(ret = lys_unres_dep_sets_create_single(ctx_set, main_set), cleanup);
+
+ if (mod && !ly_set_contains(ctx_set, mod, NULL)) {
+ /* dep set for this module has already been created, nothing else to do */
+ goto cleanup;
+ }
+
+ while (ctx_set->count) {
+ /* create new dep set */
+ LY_CHECK_GOTO(ret = ly_set_new(&dep_set), cleanup);
+
+ if (mod) {
+ /* use the module create a dep set with the rest of its dependent modules */
+ LY_CHECK_GOTO(ret = lys_unres_dep_sets_create_mod_r(mod, ctx_set, dep_set, &aux_set), cleanup);
+ } else {
+ /* use first ctx mod to create a dep set with the rest of its dependent modules */
+ LY_CHECK_GOTO(ret = lys_unres_dep_sets_create_mod_r(ctx_set->objs[0], ctx_set, dep_set, &aux_set), cleanup);
+ }
+ ly_set_erase(&aux_set, NULL);
+ assert(dep_set->count);
+
+ /* check whether there is any module that will be (re)compiled */
+ found = 0;
+ for (i = 0; i < dep_set->count; ++i) {
+ m = dep_set->objs[i];
+ if (m->to_compile) {
+ found = 1;
+ break;
+ }
+ }
+
+ if (found) {
+ /* if there is, all the implemented modules need to be recompiled */
+ for (i = 0; i < dep_set->count; ++i) {
+ m = dep_set->objs[i];
+ if (m->implemented) {
+ m->to_compile = 1;
+ }
+ }
+ }
+
+ /* add the dep set into main set */
+ LY_CHECK_GOTO(ret = ly_set_add(main_set, dep_set, 1, NULL), cleanup);
+ dep_set = NULL;
+
+ if (mod) {
+ /* we need dep set only for this module */
+ break;
+ }
+ }
+
+#ifndef NDEBUG
+ LOGDBG(LY_LDGDEPSETS, "dep sets created (%" PRIu32 "):", main_set->count);
+ for (i = 0; i < main_set->count; ++i) {
+ struct ly_set *iter_set = main_set->objs[i];
+
+ LOGDBG(LY_LDGDEPSETS, "dep set #%" PRIu32 ":", i);
+ for (uint32_t j = 0; j < iter_set->count; ++j) {
+ m = iter_set->objs[j];
+ LOGDBG(LY_LDGDEPSETS, "\t%s", m->name);
+ }
+ }
+#endif
+
+cleanup:
+ assert(ret || main_set->objs);
+ ly_set_erase(&aux_set, NULL);
+ ly_set_free(dep_set, NULL);
+ ly_set_free(ctx_set, NULL);
+ return ret;
+}
+
+void
+lys_unres_glob_revert(struct ly_ctx *ctx, struct lys_glob_unres *unres)
+{
+ uint32_t i, j, idx, temp_lo = 0;
+ struct lysf_ctx fctx = {.ctx = ctx};
+ struct ly_set *dep_set;
+ LY_ERR ret;
+
+ for (i = 0; i < unres->implementing.count; ++i) {
+ fctx.mod = unres->implementing.objs[i];
+ assert(fctx.mod->implemented);
+
+ /* make the module correctly non-implemented again */
+ fctx.mod->implemented = 0;
+ lys_precompile_augments_deviations_revert(ctx, fctx.mod);
+ lysc_module_free(&fctx, fctx.mod->compiled);
+ fctx.mod->compiled = NULL;
+
+ /* should not be made implemented */
+ fctx.mod->to_compile = 0;
+ }
+
+ for (i = 0; i < unres->creating.count; ++i) {
+ fctx.mod = unres->creating.objs[i];
+
+ /* remove the module from the context */
+ ly_set_rm(&ctx->list, fctx.mod, NULL);
+
+ /* remove it also from dep sets */
+ for (j = 0; j < unres->dep_sets.count; ++j) {
+ dep_set = unres->dep_sets.objs[j];
+ if (ly_set_contains(dep_set, fctx.mod, &idx)) {
+ ly_set_rm_index(dep_set, idx, NULL);
+ break;
+ }
+ }
+
+ /* free the module */
+ lys_module_free(&fctx, fctx.mod, 1);
+ }
+
+ /* remove the extensions as well */
+ lysf_ctx_erase(&fctx);
+
+ if (unres->implementing.count) {
+ /* recompile previous context because some implemented modules are no longer implemented,
+ * we can reuse the current to_compile flags */
+ ly_temp_log_options(&temp_lo);
+ ret = lys_compile_depset_all(ctx, &ctx->unres);
+ ly_temp_log_options(NULL);
+ if (ret) {
+ LOGINT(ctx);
+ }
+ }
+}
+
+void
+lys_unres_glob_erase(struct lys_glob_unres *unres)
+{
+ uint32_t i;
+
+ for (i = 0; i < unres->dep_sets.count; ++i) {
+ ly_set_free(unres->dep_sets.objs[i], NULL);
+ }
+ ly_set_erase(&unres->dep_sets, NULL);
+ ly_set_erase(&unres->implementing, NULL);
+ ly_set_erase(&unres->creating, NULL);
+
+ assert(!unres->ds_unres.whens.count);
+ assert(!unres->ds_unres.musts.count);
+ assert(!unres->ds_unres.leafrefs.count);
+ assert(!unres->ds_unres.disabled_leafrefs.count);
+ assert(!unres->ds_unres.dflts.count);
+ assert(!unres->ds_unres.disabled.count);
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_set_implemented(struct lys_module *mod, const char **features)
+{
+ LY_ERR ret = LY_SUCCESS;
+ struct lys_glob_unres *unres = &mod->ctx->unres;
+
+ LY_CHECK_ARG_RET(NULL, mod, LY_EINVAL);
+
+ /* implement */
+ ret = _lys_set_implemented(mod, features, unres);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ if (!(mod->ctx->flags & LY_CTX_EXPLICIT_COMPILE)) {
+ /* create dep set for the module and mark all the modules that will be (re)compiled */
+ LY_CHECK_GOTO(ret = lys_unres_dep_sets_create(mod->ctx, &unres->dep_sets, mod), cleanup);
+
+ /* (re)compile the whole dep set (other dep sets will have no modules marked for compilation) */
+ LY_CHECK_GOTO(ret = lys_compile_depset_all(mod->ctx, unres), cleanup);
+
+ /* unres resolved */
+ lys_unres_glob_erase(unres);
+ }
+
+cleanup:
+ if (ret) {
+ lys_unres_glob_revert(mod->ctx, unres);
+ lys_unres_glob_erase(unres);
+ }
+ return ret;
+}
+
+/**
+ * @brief Resolve (find) all imported and included modules.
+ *
+ * @param[in] pctx Parser context.
+ * @param[in] pmod Parsed module to resolve.
+ * @param[out] new_mods Set with all the newly loaded modules.
+ * @return LY_ERR value.
+ */
+static LY_ERR
+lysp_resolve_import_include(struct lysp_ctx *pctx, struct lysp_module *pmod, struct ly_set *new_mods)
+{
+ struct lysp_import *imp;
+ LY_ARRAY_COUNT_TYPE u, v;
+
+ pmod->parsing = 1;
+ LY_ARRAY_FOR(pmod->imports, u) {
+ imp = &pmod->imports[u];
+ if (!imp->module) {
+ LY_CHECK_RET(lys_parse_load(PARSER_CTX(pctx), imp->name, imp->rev[0] ? imp->rev : NULL, new_mods, &imp->module));
+
+ if (!imp->rev[0]) {
+ /* This module must be selected for the next similar
+ * import without revision-date to avoid incorrect
+ * derived identities in the ::lys_module.identities.
+ */
+ imp->module->latest_revision |= LYS_MOD_IMPORTED_REV;
+ }
+ }
+ /* check for importing the same module twice */
+ for (v = 0; v < u; ++v) {
+ if (imp->module == pmod->imports[v].module) {
+ LOGWRN(PARSER_CTX(pctx), "Single revision of the module \"%s\" imported twice.", imp->name);
+ }
+ }
+ }
+ LY_CHECK_RET(lysp_load_submodules(pctx, pmod, new_mods));
+
+ pmod->parsing = 0;
+
+ return LY_SUCCESS;
+}
+
+/**
+ * @brief Generate path of the given paresed node.
+ *
+ * @param[in] node Schema path of this node will be generated.
+ * @param[in] parent Build relative path only until this parent is found. If NULL, the full absolute path is printed.
+ * @return NULL in case of memory allocation error, path of the node otherwise.
+ * In case the @p buffer is NULL, the returned string is dynamically allocated and caller is responsible to free it.
+ */
+static char *
+lysp_path_until(const struct lysp_node *node, const struct lysp_node *parent, const struct lysp_module *pmod)
+{
+ const struct lysp_node *iter, *par;
+ char *path = NULL, *s;
+ const char *slash;
+ int len = 0;
+
+ for (iter = node; iter && (iter != parent) && (len >= 0); iter = iter->parent) {
+ if (parent && (iter->parent == parent)) {
+ slash = "";
+ } else {
+ slash = "/";
+ }
+
+ s = path;
+ par = iter->parent;
+ if (!par) {
+ /* print prefix */
+ len = asprintf(&path, "%s%s:%s%s", slash, pmod->mod->name, iter->name, s ? s : "");
+ } else {
+ /* prefix is the same as in parent */
+ len = asprintf(&path, "%s%s%s", slash, iter->name, s ? s : "");
+ }
+ free(s);
+ }
+
+ if (len < 0) {
+ free(path);
+ path = NULL;
+ } else if (len == 0) {
+ path = strdup("/");
+ }
+
+ return path;
+}
+
+/**
+ * @brief Build log path for a parsed extension instance.
+ *
+ * @param[in] pcxt Parse context.
+ * @param[in] ext Parsed extension instance.
+ * @param[out] path Generated path.
+ * @return LY_ERR value.
+ */
+static LY_ERR
+lysp_resolve_ext_instance_log_path(const struct lysp_ctx *pctx, const struct lysp_ext_instance *ext, char **path)
+{
+ char *buf = NULL;
+ uint32_t used = 0, size = 0;
+
+ if (ext->parent_stmt & LY_STMT_NODE_MASK) {
+ /* parsed node path */
+ buf = lysp_path_until(ext->parent, NULL, PARSER_CUR_PMOD(pctx));
+ LY_CHECK_ERR_RET(!buf, LOGMEM(PARSER_CTX(pctx)), LY_EMEM);
+ size = used = strlen(buf);
+
+ /* slash */
+ size += 1;
+ buf = realloc(buf, size + 1);
+ LY_CHECK_ERR_RET(!buf, LOGMEM(PARSER_CTX(pctx)), LY_EMEM);
+ used += sprintf(buf + used, "/");
+ } else {
+ /* module */
+ size += 1 + strlen(PARSER_CUR_PMOD(pctx)->mod->name) + 1;
+ buf = realloc(buf, size + 1);
+ LY_CHECK_ERR_RET(!buf, LOGMEM(PARSER_CTX(pctx)), LY_EMEM);
+ used += sprintf(buf + used, "/%s:", PARSER_CUR_PMOD(pctx)->mod->name);
+ }
+
+ /* extension name */
+ size += 12 + strlen(ext->name) + 2;
+ buf = realloc(buf, size + 1);
+ LY_CHECK_ERR_RET(!buf, LOGMEM(PARSER_CTX(pctx)), LY_EMEM);
+ used += sprintf(buf + used, "{extension='%s'}", ext->name);
+
+ /* extension argument */
+ if (ext->argument) {
+ size += 1 + strlen(ext->argument);
+ buf = realloc(buf, size + 1);
+ LY_CHECK_ERR_RET(!buf, LOGMEM(PARSER_CTX(pctx)), LY_EMEM);
+ used += sprintf(buf + used, "/%s", ext->argument);
+ }
+
+ *path = buf;
+ return LY_SUCCESS;
+}
+
+/**
+ * @brief Resolve (find) all extension instance records and finish their parsing.
+ *
+ * @param[in] pctx Parse context with all the parsed extension instances.
+ * @return LY_ERR value.
+ */
+static LY_ERR
+lysp_resolve_ext_instance_records(struct lysp_ctx *pctx)
+{
+ LY_ERR r;
+ struct lysf_ctx fctx = {.ctx = PARSER_CTX(pctx)};
+ struct lysp_ext_instance *exts, *ext;
+ const struct lys_module *mod;
+ uint32_t i;
+ LY_ARRAY_COUNT_TYPE u;
+ char *path = NULL;
+
+ /* first finish parsing all extension instances ... */
+ for (i = 0; i < pctx->ext_inst.count; ++i) {
+ exts = pctx->ext_inst.objs[i];
+ LY_ARRAY_FOR(exts, u) {
+ ext = &exts[u];
+
+ /* find the extension definition */
+ LY_CHECK_RET(lysp_ext_find_definition(PARSER_CTX(pctx), ext, &mod, &ext->def));
+
+ /* resolve the argument, if needed */
+ LY_CHECK_RET(lysp_ext_instance_resolve_argument(PARSER_CTX(pctx), ext));
+
+ /* find the extension record, if any */
+ ext->record = lyplg_ext_record_find(mod->name, mod->revision, ext->def->name);
+ }
+ }
+
+ /* ... then call the parse callback */
+ for (i = 0; i < pctx->ext_inst.count; ++i) {
+ exts = pctx->ext_inst.objs[i];
+ u = 0;
+ while (u < LY_ARRAY_COUNT(exts)) {
+ ext = &exts[u];
+ if (!ext->record || !ext->record->plugin.parse) {
+ goto next_iter;
+ }
+
+ /* set up log path */
+ if ((r = lysp_resolve_ext_instance_log_path(pctx, ext, &path))) {
+ return r;
+ }
+ LOG_LOCSET(NULL, NULL, path, NULL);
+
+ /* parse */
+ r = ext->record->plugin.parse(pctx, ext);
+
+ LOG_LOCBACK(0, 0, 1, 0);
+ free(path);
+
+ if (r == LY_ENOT) {
+ /* instance should be ignored, remove it */
+ lysp_ext_instance_free(&fctx, ext);
+ LY_ARRAY_DECREMENT(exts);
+ if (u < LY_ARRAY_COUNT(exts)) {
+ /* replace by the last item */
+ *ext = exts[LY_ARRAY_COUNT(exts)];
+ } /* else if there are no more items, leave the empty array, we are not able to free it */
+ continue;
+ } else if (r) {
+ /* error */
+ return r;
+ }
+
+next_iter:
+ ++u;
+ }
+ }
+
+ return LY_SUCCESS;
+}
+
+LY_ERR
+lys_parse_submodule(struct ly_ctx *ctx, struct ly_in *in, LYS_INFORMAT format, struct lysp_ctx *main_ctx,
+ LY_ERR (*custom_check)(const struct ly_ctx *, struct lysp_module *, struct lysp_submodule *, void *),
+ void *check_data, struct ly_set *new_mods, struct lysp_submodule **submodule)
+{
+ LY_ERR ret;
+ struct lysp_submodule *submod = NULL, *latest_sp;
+ struct lysp_yang_ctx *yangctx = NULL;
+ struct lysp_yin_ctx *yinctx = NULL;
+ struct lysp_ctx *pctx;
+ struct lysf_ctx fctx = {.ctx = ctx};
+
+ LY_CHECK_ARG_RET(ctx, ctx, in, LY_EINVAL);
+
+ switch (format) {
+ case LYS_IN_YIN:
+ ret = yin_parse_submodule(&yinctx, ctx, main_ctx, in, &submod);
+ pctx = (struct lysp_ctx *)yinctx;
+ break;
+ case LYS_IN_YANG:
+ ret = yang_parse_submodule(&yangctx, ctx, main_ctx, in, &submod);
+ pctx = (struct lysp_ctx *)yangctx;
+ break;
+ default:
+ LOGERR(ctx, LY_EINVAL, "Invalid schema input format.");
+ ret = LY_EINVAL;
+ break;
+ }
+ LY_CHECK_GOTO(ret, error);
+ assert(submod);
+
+ /* make sure that the newest revision is at position 0 */
+ lysp_sort_revisions(submod->revs);
+
+ /* decide the latest revision */
+ latest_sp = (struct lysp_submodule *)ly_ctx_get_submodule2_latest(submod->mod, submod->name);
+ if (latest_sp) {
+ if (submod->revs) {
+ if (!latest_sp->revs) {
+ /* latest has no revision, so mod is anyway newer */
+ submod->latest_revision = latest_sp->latest_revision;
+ /* the latest_sp is zeroed later when the new module is being inserted into the context */
+ } else if (strcmp(submod->revs[0].date, latest_sp->revs[0].date) > 0) {
+ submod->latest_revision = latest_sp->latest_revision;
+ /* the latest_sp is zeroed later when the new module is being inserted into the context */
+ } else {
+ latest_sp = NULL;
+ }
+ } else {
+ latest_sp = NULL;
+ }
+ } else {
+ submod->latest_revision = 1;
+ }
+
+ if (custom_check) {
+ LY_CHECK_GOTO(ret = custom_check(ctx, NULL, submod, check_data), error);
+ }
+
+ if (latest_sp) {
+ latest_sp->latest_revision = 0;
+ }
+
+ lys_parser_fill_filepath(ctx, in, &submod->filepath);
+
+ /* resolve imports and includes */
+ LY_CHECK_GOTO(ret = lysp_resolve_import_include(pctx, (struct lysp_module *)submod, new_mods), error);
+
+ if (format == LYS_IN_YANG) {
+ lysp_yang_ctx_free(yangctx);
+ } else {
+ lysp_yin_ctx_free(yinctx);
+ }
+ *submodule = submod;
+ return LY_SUCCESS;
+
+error:
+ if (!submod || !submod->name) {
+ LOGERR(ctx, ret, "Parsing submodule failed.");
+ } else {
+ LOGERR(ctx, ret, "Parsing submodule \"%s\" failed.", submod->name);
+ }
+ lysp_module_free(&fctx, (struct lysp_module *)submod);
+ if (format == LYS_IN_YANG) {
+ lysp_yang_ctx_free(yangctx);
+ } else {
+ lysp_yin_ctx_free(yinctx);
+ }
+ return ret;
+}
+
+/**
+ * @brief Add ietf-netconf metadata to the parsed module. Operation, filter, and select are added.
+ *
+ * @param[in] pctx Parse context.
+ * @param[in] mod Parsed module to add to.
+ * @return LY_SUCCESS on success.
+ * @return LY_ERR on error.
+ */
+static LY_ERR
+lysp_add_internal_ietf_netconf(struct lysp_ctx *pctx, struct lysp_module *mod)
+{
+ struct lysp_ext_instance *extp;
+ struct lysp_stmt *stmt;
+ struct lysp_import *imp;
+
+ /*
+ * 1) edit-config's operation
+ */
+ LY_ARRAY_NEW_RET(mod->mod->ctx, mod->exts, extp, LY_EMEM);
+ LY_CHECK_ERR_RET(!extp, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "md_:annotation", 0, &extp->name));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "operation", 0, &extp->argument));
+ extp->format = LY_VALUE_SCHEMA;
+ extp->prefix_data = mod;
+ extp->parent = mod;
+ extp->parent_stmt = LY_STMT_MODULE;
+ extp->flags = LYS_INTERNAL;
+
+ extp->child = stmt = calloc(1, sizeof *extp->child);
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "type", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enumeration", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_TYPE;
+
+ stmt->child = calloc(1, sizeof *stmt->child);
+ stmt = stmt->child;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enum", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "merge", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_ENUM;
+
+ stmt->next = calloc(1, sizeof *stmt->child);
+ stmt = stmt->next;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enum", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "replace", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_ENUM;
+
+ stmt->next = calloc(1, sizeof *stmt->child);
+ stmt = stmt->next;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enum", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "create", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_ENUM;
+
+ stmt->next = calloc(1, sizeof *stmt->child);
+ stmt = stmt->next;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enum", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "delete", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_ENUM;
+
+ stmt->next = calloc(1, sizeof *stmt->child);
+ stmt = stmt->next;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enum", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "remove", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_ENUM;
+
+ /*
+ * 2) filter's type
+ */
+ LY_ARRAY_NEW_RET(mod->mod->ctx, mod->exts, extp, LY_EMEM);
+ LY_CHECK_ERR_RET(!extp, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "md_:annotation", 0, &extp->name));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "type", 0, &extp->argument));
+ extp->format = LY_VALUE_SCHEMA;
+ extp->prefix_data = mod;
+ extp->parent = mod;
+ extp->parent_stmt = LY_STMT_MODULE;
+ extp->flags = LYS_INTERNAL;
+
+ extp->child = stmt = calloc(1, sizeof *extp->child);
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "type", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enumeration", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_TYPE;
+
+ stmt->child = calloc(1, sizeof *stmt->child);
+ stmt = stmt->child;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enum", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "subtree", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_ENUM;
+
+ stmt->next = calloc(1, sizeof *stmt->child);
+ stmt = stmt->next;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "enum", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "xpath", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_ENUM;
+
+ /* if-feature for enum allowed only for YANG 1.1 modules */
+ if (mod->version >= LYS_VERSION_1_1) {
+ stmt->child = calloc(1, sizeof *stmt->child);
+ stmt = stmt->child;
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "if-feature", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "xpath", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_IF_FEATURE;
+ }
+
+ /*
+ * 3) filter's select
+ */
+ LY_ARRAY_NEW_RET(mod->mod->ctx, mod->exts, extp, LY_EMEM);
+ LY_CHECK_ERR_RET(!extp, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "md_:annotation", 0, &extp->name));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "select", 0, &extp->argument));
+ extp->format = LY_VALUE_SCHEMA;
+ extp->prefix_data = mod;
+ extp->parent = mod;
+ extp->parent_stmt = LY_STMT_MODULE;
+ extp->flags = LYS_INTERNAL;
+
+ extp->child = stmt = calloc(1, sizeof *extp->child);
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "type", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "yang_:xpath1.0", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_TYPE;
+
+ if (LY_ARRAY_COUNT(mod->exts) == 3) {
+ /* first extension instances */
+ assert(pctx->main_ctx == pctx);
+ LY_CHECK_RET(ly_set_add(&pctx->ext_inst, mod->exts, 1, NULL));
+ }
+
+ /* create new imports for the used prefixes */
+ LY_ARRAY_NEW_RET(mod->mod->ctx, mod->imports, imp, LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "ietf-yang-metadata", 0, &imp->name));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "md_", 0, &imp->prefix));
+ imp->flags = LYS_INTERNAL;
+
+ LY_ARRAY_NEW_RET(mod->mod->ctx, mod->imports, imp, LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "ietf-yang-types", 0, &imp->name));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "yang_", 0, &imp->prefix));
+ imp->flags = LYS_INTERNAL;
+
+ return LY_SUCCESS;
+}
+
+/**
+ * @brief Add ietf-netconf-with-defaults "default" metadata to the parsed module.
+ *
+ * @param[in] pctx Parse context.
+ * @param[in] mod Parsed module to add to.
+ * @return LY_SUCCESS on success.
+ * @return LY_ERR on error.
+ */
+static LY_ERR
+lysp_add_internal_ietf_netconf_with_defaults(struct lysp_ctx *pctx, struct lysp_module *mod)
+{
+ struct lysp_ext_instance *extp;
+ struct lysp_stmt *stmt;
+ struct lysp_import *imp;
+
+ /* add new extension instance */
+ LY_ARRAY_NEW_RET(mod->mod->ctx, mod->exts, extp, LY_EMEM);
+
+ /* fill in the extension instance fields */
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "md_:annotation", 0, &extp->name));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "default", 0, &extp->argument));
+ extp->format = LY_VALUE_SCHEMA;
+ extp->prefix_data = mod;
+ extp->parent = mod;
+ extp->parent_stmt = LY_STMT_MODULE;
+ extp->flags = LYS_INTERNAL;
+
+ extp->child = stmt = calloc(1, sizeof *extp->child);
+ LY_CHECK_ERR_RET(!stmt, LOGMEM(mod->mod->ctx), LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "type", 0, &stmt->stmt));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "boolean", 0, &stmt->arg));
+ stmt->format = LY_VALUE_SCHEMA;
+ stmt->prefix_data = mod;
+ stmt->kw = LY_STMT_TYPE;
+
+ if (LY_ARRAY_COUNT(mod->exts) == 1) {
+ /* first extension instance */
+ assert(pctx->main_ctx == pctx);
+ LY_CHECK_RET(ly_set_add(&pctx->ext_inst, mod->exts, 1, NULL));
+ }
+
+ /* create new import for the used prefix */
+ LY_ARRAY_NEW_RET(mod->mod->ctx, mod->imports, imp, LY_EMEM);
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "ietf-yang-metadata", 0, &imp->name));
+ LY_CHECK_RET(lydict_insert(mod->mod->ctx, "md_", 0, &imp->prefix));
+ imp->flags = LYS_INTERNAL;
+
+ return LY_SUCCESS;
+}
+
+LY_ERR
+lys_parse_in(struct ly_ctx *ctx, struct ly_in *in, LYS_INFORMAT format,
+ LY_ERR (*custom_check)(const struct ly_ctx *ctx, struct lysp_module *mod, struct lysp_submodule *submod, void *data),
+ void *check_data, struct ly_set *new_mods, struct lys_module **module)
+{
+ struct lys_module *mod = NULL, *latest, *mod_dup = NULL;
+ LY_ERR ret;
+ struct lysp_yang_ctx *yangctx = NULL;
+ struct lysp_yin_ctx *yinctx = NULL;
+ struct lysp_ctx *pctx = NULL;
+ struct lysf_ctx fctx = {.ctx = ctx};
+ char *filename, *rev, *dot;
+ size_t len;
+ ly_bool module_created = 0;
+
+ assert(ctx && in && new_mods);
+
+ if (module) {
+ *module = NULL;
+ }
+
+ mod = calloc(1, sizeof *mod);
+ LY_CHECK_ERR_RET(!mod, LOGMEM(ctx), LY_EMEM);
+ mod->ctx = ctx;
+
+ /* parse */
+ switch (format) {
+ case LYS_IN_YIN:
+ ret = yin_parse_module(&yinctx, in, mod);
+ pctx = (struct lysp_ctx *)yinctx;
+ break;
+ case LYS_IN_YANG:
+ ret = yang_parse_module(&yangctx, in, mod);
+ pctx = (struct lysp_ctx *)yangctx;
+ break;
+ default:
+ LOGERR(ctx, LY_EINVAL, "Invalid schema input format.");
+ ret = LY_EINVAL;
+ break;
+ }
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* make sure that the newest revision is at position 0 */
+ lysp_sort_revisions(mod->parsed->revs);
+ if (mod->parsed->revs) {
+ LY_CHECK_GOTO(ret = lydict_insert(ctx, mod->parsed->revs[0].date, 0, &mod->revision), cleanup);
+ }
+
+ /* decide the latest revision */
+ latest = ly_ctx_get_module_latest(ctx, mod->name);
+ if (latest) {
+ if (mod->revision) {
+ if (!latest->revision) {
+ /* latest has no revision, so mod is anyway newer */
+ mod->latest_revision = latest->latest_revision & (LYS_MOD_LATEST_REV | LYS_MOD_LATEST_SEARCHDIRS);
+ /* the latest is zeroed later when the new module is being inserted into the context */
+ } else if (strcmp(mod->revision, latest->revision) > 0) {
+ mod->latest_revision = latest->latest_revision & (LYS_MOD_LATEST_REV | LYS_MOD_LATEST_SEARCHDIRS);
+ /* the latest is zeroed later when the new module is being inserted into the context */
+ } else {
+ latest = NULL;
+ }
+ } else {
+ latest = NULL;
+ }
+ } else {
+ mod->latest_revision = LYS_MOD_LATEST_REV;
+ }
+
+ if (custom_check) {
+ LY_CHECK_GOTO(ret = custom_check(ctx, mod->parsed, NULL, check_data), cleanup);
+ }
+
+ /* check whether it is not already in the context in the same revision */
+ mod_dup = ly_ctx_get_module(ctx, mod->name, mod->revision);
+ if (mod_dup) {
+ /* nothing to do */
+ LOGVRB("Module \"%s@%s\" is already present in the context.", mod_dup->name,
+ mod_dup->revision ? mod_dup->revision : "<none>");
+ goto cleanup;
+ }
+
+ /* check whether there is not a namespace collision */
+ mod_dup = ly_ctx_get_module_latest_ns(ctx, mod->ns);
+ if (mod_dup && (mod_dup->revision == mod->revision)) {
+ LOGERR(ctx, LY_EINVAL, "Two different modules (\"%s\" and \"%s\") have the same namespace \"%s\".",
+ mod_dup->name, mod->name, mod->ns);
+ ret = LY_EINVAL;
+ goto cleanup;
+ }
+
+ switch (in->type) {
+ case LY_IN_FILEPATH:
+ /* check that name and revision match filename */
+ filename = strrchr(in->method.fpath.filepath, '/');
+ if (!filename) {
+ filename = in->method.fpath.filepath;
+ } else {
+ filename++;
+ }
+ rev = strchr(filename, '@');
+ dot = strrchr(filename, '.');
+
+ /* name */
+ len = strlen(mod->name);
+ if (strncmp(filename, mod->name, len) ||
+ ((rev && (rev != &filename[len])) || (!rev && (dot != &filename[len])))) {
+ LOGWRN(ctx, "File name \"%s\" does not match module name \"%s\".", filename, mod->name);
+ }
+ if (rev) {
+ len = dot - ++rev;
+ if (!mod->parsed->revs || (len != LY_REV_SIZE - 1) || strncmp(mod->parsed->revs[0].date, rev, len)) {
+ LOGWRN(ctx, "File name \"%s\" does not match module revision \"%s\".", filename,
+ mod->parsed->revs ? mod->parsed->revs[0].date : "none");
+ }
+ }
+
+ break;
+ case LY_IN_FD:
+ case LY_IN_FILE:
+ case LY_IN_MEMORY:
+ /* nothing special to do */
+ break;
+ case LY_IN_ERROR:
+ LOGINT(ctx);
+ ret = LY_EINT;
+ goto cleanup;
+ }
+ lys_parser_fill_filepath(ctx, in, &mod->filepath);
+
+ if (latest) {
+ latest->latest_revision &= ~(LYS_MOD_LATEST_REV | LYS_MOD_LATEST_SEARCHDIRS);
+ }
+
+ /* add internal data in case specific modules were parsed */
+ if (!strcmp(mod->name, "ietf-netconf")) {
+ LY_CHECK_GOTO(ret = lysp_add_internal_ietf_netconf(pctx, mod->parsed), cleanup);
+ } else if (!strcmp(mod->name, "ietf-netconf-with-defaults")) {
+ LY_CHECK_GOTO(ret = lysp_add_internal_ietf_netconf_with_defaults(pctx, mod->parsed), cleanup);
+ }
+
+ /* add the module into newly created module set, will also be freed from there on any error */
+ LY_CHECK_GOTO(ret = ly_set_add(new_mods, mod, 1, NULL), cleanup);
+ module_created = 1;
+
+ /* add into context */
+ ret = ly_set_add(&ctx->list, mod, 1, NULL);
+ LY_CHECK_GOTO(ret, cleanup);
+ ctx->change_count++;
+
+ /* resolve includes and all imports */
+ LY_CHECK_GOTO(ret = lysp_resolve_import_include(pctx, mod->parsed, new_mods), cleanup);
+
+ /* resolve extension instance plugin records */
+ LY_CHECK_GOTO(ret = lysp_resolve_ext_instance_records(pctx), cleanup);
+
+ /* check name collisions */
+ LY_CHECK_GOTO(ret = lysp_check_dup_typedefs(pctx, mod->parsed), cleanup);
+ LY_CHECK_GOTO(ret = lysp_check_dup_groupings(pctx, mod->parsed), cleanup);
+ LY_CHECK_GOTO(ret = lysp_check_dup_features(pctx, mod->parsed), cleanup);
+ LY_CHECK_GOTO(ret = lysp_check_dup_identities(pctx, mod->parsed), cleanup);
+
+ /* compile features */
+ LY_CHECK_GOTO(ret = lys_compile_feature_iffeatures(mod->parsed), cleanup);
+
+ /* compile identities */
+ LY_CHECK_GOTO(ret = lys_compile_identities(mod), cleanup);
+
+cleanup:
+ if (ret && (ret != LY_EEXIST)) {
+ if (mod && mod->name) {
+ /* there are cases when path is not available for parsing error, so this additional
+ * message tries to add information about the module where the error occurred */
+ struct ly_err_item *e = ly_err_last(ctx);
+
+ if (e && (!e->path || !strncmp(e->path, "Line ", ly_strlen_const("Line ")))) {
+ LOGERR(ctx, ret, "Parsing module \"%s\" failed.", mod->name);
+ }
+ }
+ }
+ if (!module_created) {
+ fctx.mod = mod;
+ lys_module_free(&fctx, mod, 0);
+ lysf_ctx_erase(&fctx);
+
+ mod = mod_dup;
+ }
+
+ if (format == LYS_IN_YANG) {
+ lysp_yang_ctx_free(yangctx);
+ } else {
+ lysp_yin_ctx_free(yinctx);
+ }
+
+ if (!ret && module) {
+ *module = mod;
+ }
+ return ret;
+}
+
+static LYS_INFORMAT
+lys_parse_get_format(const struct ly_in *in, LYS_INFORMAT format)
+{
+ if (!format && (in->type == LY_IN_FILEPATH)) {
+ /* unknown format - try to detect it from filename's suffix */
+ const char *path = in->method.fpath.filepath;
+ size_t len = strlen(path);
+
+ /* ignore trailing whitespaces */
+ for ( ; len > 0 && isspace(path[len - 1]); len--) {}
+
+ if ((len >= LY_YANG_SUFFIX_LEN + 1) &&
+ !strncmp(&path[len - LY_YANG_SUFFIX_LEN], LY_YANG_SUFFIX, LY_YANG_SUFFIX_LEN)) {
+ format = LYS_IN_YANG;
+ } else if ((len >= LY_YIN_SUFFIX_LEN + 1) &&
+ !strncmp(&path[len - LY_YIN_SUFFIX_LEN], LY_YIN_SUFFIX, LY_YIN_SUFFIX_LEN)) {
+ format = LYS_IN_YIN;
+ } /* else still unknown */
+ }
+
+ return format;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_parse(struct ly_ctx *ctx, struct ly_in *in, LYS_INFORMAT format, const char **features, struct lys_module **module)
+{
+ LY_ERR ret = LY_SUCCESS;
+ struct lys_module *mod;
+
+ if (module) {
+ *module = NULL;
+ }
+ LY_CHECK_ARG_RET(NULL, ctx, in, LY_EINVAL);
+
+ format = lys_parse_get_format(in, format);
+ LY_CHECK_ARG_RET(ctx, format, LY_EINVAL);
+
+ /* remember input position */
+ in->func_start = in->current;
+
+ /* parse */
+ ret = lys_parse_in(ctx, in, format, NULL, NULL, &ctx->unres.creating, &mod);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* implement */
+ ret = _lys_set_implemented(mod, features, &ctx->unres);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ if (!(ctx->flags & LY_CTX_EXPLICIT_COMPILE)) {
+ /* create dep set for the module and mark all the modules that will be (re)compiled */
+ LY_CHECK_GOTO(ret = lys_unres_dep_sets_create(ctx, &ctx->unres.dep_sets, mod), cleanup);
+
+ /* (re)compile the whole dep set (other dep sets will have no modules marked for compilation) */
+ LY_CHECK_GOTO(ret = lys_compile_depset_all(ctx, &ctx->unres), cleanup);
+
+ /* unres resolved */
+ lys_unres_glob_erase(&ctx->unres);
+ }
+
+cleanup:
+ if (ret) {
+ lys_unres_glob_revert(ctx, &ctx->unres);
+ lys_unres_glob_erase(&ctx->unres);
+ } else if (module) {
+ *module = mod;
+ }
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_parse_mem(struct ly_ctx *ctx, const char *data, LYS_INFORMAT format, struct lys_module **module)
+{
+ LY_ERR ret;
+ struct ly_in *in = NULL;
+
+ LY_CHECK_ARG_RET(ctx, data, format != LYS_IN_UNKNOWN, LY_EINVAL);
+
+ LY_CHECK_ERR_RET(ret = ly_in_new_memory(data, &in), LOGERR(ctx, ret, "Unable to create input handler."), ret);
+
+ ret = lys_parse(ctx, in, format, NULL, module);
+ ly_in_free(in, 0);
+
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_parse_fd(struct ly_ctx *ctx, int fd, LYS_INFORMAT format, struct lys_module **module)
+{
+ LY_ERR ret;
+ struct ly_in *in = NULL;
+
+ LY_CHECK_ARG_RET(ctx, fd > -1, format != LYS_IN_UNKNOWN, LY_EINVAL);
+
+ LY_CHECK_ERR_RET(ret = ly_in_new_fd(fd, &in), LOGERR(ctx, ret, "Unable to create input handler."), ret);
+
+ ret = lys_parse(ctx, in, format, NULL, module);
+ ly_in_free(in, 0);
+
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_parse_path(struct ly_ctx *ctx, const char *path, LYS_INFORMAT format, struct lys_module **module)
+{
+ LY_ERR ret;
+ struct ly_in *in = NULL;
+
+ LY_CHECK_ARG_RET(ctx, path, format != LYS_IN_UNKNOWN, LY_EINVAL);
+
+ LY_CHECK_ERR_RET(ret = ly_in_new_filepath(path, 0, &in),
+ LOGERR(ctx, ret, "Unable to create input handler for filepath %s.", path), ret);
+
+ ret = lys_parse(ctx, in, format, NULL, module);
+ ly_in_free(in, 0);
+
+ return ret;
+}
+
+LIBYANG_API_DEF LY_ERR
+lys_search_localfile(const char * const *searchpaths, ly_bool cwd, const char *name, const char *revision,
+ char **localfile, LYS_INFORMAT *format)
+{
+ LY_ERR ret = LY_EMEM;
+ size_t len, flen, match_len = 0, dir_len;
+ ly_bool implicit_cwd = 0;
+ char *wd, *wn = NULL;
+ DIR *dir = NULL;
+ struct dirent *file;
+ char *match_name = NULL;
+ LYS_INFORMAT format_aux, match_format = 0;
+ struct ly_set *dirs;
+ struct stat st;
+
+ LY_CHECK_ARG_RET(NULL, localfile, LY_EINVAL);
+
+ /* start to fill the dir fifo with the context's search path (if set)
+ * and the current working directory */
+ LY_CHECK_RET(ly_set_new(&dirs));
+
+ len = strlen(name);
+ if (cwd) {
+ wd = get_current_dir_name();
+ if (!wd) {
+ LOGMEM(NULL);
+ goto cleanup;
+ } else {
+ /* add implicit current working directory (./) to be searched,
+ * this directory is not searched recursively */
+ ret = ly_set_add(dirs, wd, 0, NULL);
+ LY_CHECK_GOTO(ret, cleanup);
+ implicit_cwd = 1;
+ }
+ }
+ if (searchpaths) {
+ for (uint64_t i = 0; searchpaths[i]; i++) {
+ /* check for duplicities with the implicit current working directory */
+ if (implicit_cwd && !strcmp(dirs->objs[0], searchpaths[i])) {
+ implicit_cwd = 0;
+ continue;
+ }
+ wd = strdup(searchpaths[i]);
+ if (!wd) {
+ LOGMEM(NULL);
+ goto cleanup;
+ } else {
+ ret = ly_set_add(dirs, wd, 0, NULL);
+ LY_CHECK_GOTO(ret, cleanup);
+ }
+ }
+ }
+ wd = NULL;
+
+ /* start searching */
+ while (dirs->count) {
+ free(wd);
+ free(wn); wn = NULL;
+
+ dirs->count--;
+ wd = (char *)dirs->objs[dirs->count];
+ dirs->objs[dirs->count] = NULL;
+ LOGVRB("Searching for \"%s\" in \"%s\".", name, wd);
+
+ if (dir) {
+ closedir(dir);
+ }
+ dir = opendir(wd);
+ dir_len = strlen(wd);
+ if (!dir) {
+ LOGWRN(NULL, "Unable to open directory \"%s\" for searching (sub)modules (%s).", wd, strerror(errno));
+ } else {
+ while ((file = readdir(dir))) {
+ if (!strcmp(".", file->d_name) || !strcmp("..", file->d_name)) {
+ /* skip . and .. */
+ continue;
+ }
+ free(wn);
+ if (asprintf(&wn, "%s/%s", wd, file->d_name) == -1) {
+ LOGMEM(NULL);
+ goto cleanup;
+ }
+ if (stat(wn, &st) == -1) {
+ LOGWRN(NULL, "Unable to get information about \"%s\" file in \"%s\" when searching for (sub)modules (%s)",
+ file->d_name, wd, strerror(errno));
+ continue;
+ }
+ if (S_ISDIR(st.st_mode) && (dirs->count || !implicit_cwd)) {
+ /* we have another subdirectory in searchpath to explore,
+ * subdirectories are not taken into account in current working dir (dirs->set.g[0]) */
+ ret = ly_set_add(dirs, wn, 0, NULL);
+ LY_CHECK_GOTO(ret, cleanup);
+
+ /* continue with the next item in current directory */
+ wn = NULL;
+ continue;
+ } else if (!S_ISREG(st.st_mode)) {
+ /* not a regular file (note that we see the target of symlinks instead of symlinks */
+ continue;
+ }
+
+ /* here we know that the item is a file which can contain a module */
+ if (strncmp(name, file->d_name, len) ||
+ ((file->d_name[len] != '.') && (file->d_name[len] != '@'))) {
+ /* different filename than the module we search for */
+ continue;
+ }
+
+ /* get type according to filename suffix */
+ flen = strlen(file->d_name);
+ if ((flen >= LY_YANG_SUFFIX_LEN + 1) &&
+ !strcmp(&file->d_name[flen - LY_YANG_SUFFIX_LEN], LY_YANG_SUFFIX)) {
+ format_aux = LYS_IN_YANG;
+ } else if ((flen >= LY_YIN_SUFFIX_LEN + 1) &&
+ !strcmp(&file->d_name[flen - LY_YIN_SUFFIX_LEN], LY_YIN_SUFFIX)) {
+ format_aux = LYS_IN_YIN;
+ } else {
+ /* not supportde suffix/file format */
+ continue;
+ }
+
+ if (revision) {
+ /* we look for the specific revision, try to get it from the filename */
+ if (file->d_name[len] == '@') {
+ /* check revision from the filename */
+ if (strncmp(revision, &file->d_name[len + 1], strlen(revision))) {
+ /* another revision */
+ continue;
+ } else {
+ /* exact revision */
+ free(match_name);
+ match_name = wn;
+ wn = NULL;
+ match_len = dir_len + 1 + len;
+ match_format = format_aux;
+ goto success;
+ }
+ } else {
+ /* continue trying to find exact revision match, use this only if not found */
+ free(match_name);
+ match_name = wn;
+ wn = NULL;
+ match_len = dir_len + 1 + len;
+ match_format = format_aux;
+ continue;
+ }
+ } else {
+ /* remember the revision and try to find the newest one */
+ if (match_name) {
+ if ((file->d_name[len] != '@') ||
+ lysp_check_date(NULL, &file->d_name[len + 1],
+ flen - ((format_aux == LYS_IN_YANG) ? LY_YANG_SUFFIX_LEN : LY_YIN_SUFFIX_LEN) - len - 1, NULL)) {
+ continue;
+ } else if ((match_name[match_len] == '@') &&
+ (strncmp(&match_name[match_len + 1], &file->d_name[len + 1], LY_REV_SIZE - 1) >= 0)) {
+ continue;
+ }
+ free(match_name);
+ }
+
+ match_name = wn;
+ wn = NULL;
+ match_len = dir_len + 1 + len;
+ match_format = format_aux;
+ continue;
+ }
+ }
+ }
+ }
+
+success:
+ (*localfile) = match_name;
+ match_name = NULL;
+ if (format) {
+ (*format) = match_format;
+ }
+ ret = LY_SUCCESS;
+
+cleanup:
+ free(wn);
+ free(wd);
+ if (dir) {
+ closedir(dir);
+ }
+ free(match_name);
+ ly_set_free(dirs, free);
+
+ return ret;
+}