1 /*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
4 This file is part of systemd.
6 Copyright 2012 Kay Sievers <kay@vrfy.org>
8 systemd is free software; you can redistribute it and/or modify it
9 under the terms of the GNU Lesser General Public License as published by
10 the Free Software Foundation; either version 2.1 of the License, or
11 (at your option) any later version.
13 systemd is distributed in the hope that it will be useful, but
14 WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 Lesser General Public License for more details.
18 You should have received a copy of the GNU Lesser General Public License
19 along with systemd; If not, see <http://www.gnu.org/licenses/>.
29 * Strbuf stores given strings in a single continuous allocated memory
30 * area. Identical strings are de-duplicated and return the same offset
31 * as the first string stored. If the tail of a string already exists
32 * in the buffer, the tail is returned.
34 * A trie (http://en.wikipedia.org/wiki/Trie) is used to maintain the
35 * information about the stored strings.
37 * Example of udev rules:
40 * read rules file: /usr/lib/udev/rules.d/99-systemd.rules
41 * rules contain 196608 bytes tokens (16384 * 12 bytes), 39742 bytes strings
42 * 23939 strings (207859 bytes), 20404 de-duplicated (171653 bytes), 3536 trie nodes used
46 struct strbuf *strbuf_new(void) {
49 str = new0(struct strbuf, 1);
53 str->buf = new0(char, 1);
58 str->root = new0(struct strbuf_node, 1);
70 static void strbuf_node_cleanup(struct strbuf_node *node) {
73 for (i = 0; i < node->children_count; i++)
74 strbuf_node_cleanup(node->children[i].child);
79 /* clean up trie data, leave only the string buffer */
80 void strbuf_complete(struct strbuf *str) {
84 strbuf_node_cleanup(str->root);
88 /* clean up everything */
89 void strbuf_cleanup(struct strbuf *str) {
93 strbuf_node_cleanup(str->root);
98 static int strbuf_children_cmp(const struct strbuf_child_entry *n1,
99 const struct strbuf_child_entry *n2) {
100 return n1->c - n2->c;
103 static void bubbleinsert(struct strbuf_node *node,
105 struct strbuf_node *node_child) {
107 struct strbuf_child_entry new = {
111 int left = 0, right = node->children_count;
113 while (right > left) {
114 int middle = (right + left) / 2 ;
115 if (strbuf_children_cmp(&node->children[middle], &new) <= 0)
121 memmove(node->children + left + 1, node->children + left,
122 sizeof(struct strbuf_child_entry) * (node->children_count - left));
123 node->children[left] = new;
125 node->children_count ++;
128 /* add string, return the index/offset into the buffer */
129 ssize_t strbuf_add_string(struct strbuf *str, const char *s, size_t len) {
131 struct strbuf_node *node;
134 struct strbuf_child_entry *child;
135 struct strbuf_node *node_child;
141 /* search string; start from last character to find possibly matching tails */
149 for (depth = 0; depth <= len; depth++) {
150 struct strbuf_child_entry search;
152 /* match against current node */
153 off = node->value_off + node->value_len - len;
154 if (depth == len || (node->value_len >= len && memcmp(str->buf + off, s, len) == 0)) {
155 str->dedup_len += len;
160 /* lookup child node */
161 c = s[len - 1 - depth];
163 child = bsearch(&search, node->children, node->children_count,
164 sizeof(struct strbuf_child_entry),
165 (__compar_fn_t) strbuf_children_cmp);
172 buf_new = realloc(str->buf, str->len + len+1);
177 memcpy(str->buf + off, s, len);
179 str->buf[str->len++] = '\0';
182 node_child = new0(struct strbuf_node, 1);
185 node_child->value_off = off;
186 node_child->value_len = len;
188 /* extend array, add new entry, sort for bisection */
189 child = realloc(node->children, (node->children_count + 1) * sizeof(struct strbuf_child_entry));
197 node->children = child;
198 bubbleinsert(node, c, node_child);