2 This file is part of GNUnet
3 (C) 2012 Christian Grothoff (and other contributing authors)
5 GNUnet is free software; you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published
7 by the Free Software Foundation; either version 3, or (at your
8 option) any later version.
10 GNUnet is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with GNUnet; see the file COPYING. If not, write to the
17 Free Software Foundation, Inc., 59 Temple Place - Suite 330,
18 Boston, MA 02111-1307, USA.
21 * @file src/regex/regex_internal.h
22 * @brief common internal definitions for regex library.
23 * @author Maximilian Szengel
25 #ifndef REGEX_INTERNAL_H
26 #define REGEX_INTERNAL_H
28 #include "regex_internal_lib.h"
33 #if 0 /* keep Emacsens' auto-indent happy */
39 * char array of literals that are allowed inside a regex (apart from the
42 #define ALLOWED_LITERALS "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
46 * Transition between two states. Transitions are stored at the states from
47 * which they origin ('from_state'). Each state can have 0-n transitions.
48 * If label is NULL, this is considered to be an epsilon transition.
50 struct REGEX_INTERNAL_Transition
53 * This is a linked list.
55 struct REGEX_INTERNAL_Transition *prev;
58 * This is a linked list.
60 struct REGEX_INTERNAL_Transition *next;
63 * Unique id of this transition.
68 * Label for this transition. This is basically the edge label for the graph.
73 * State to which this transition leads.
75 struct REGEX_INTERNAL_State *to_state;
78 * State from which this transition origins.
80 struct REGEX_INTERNAL_State *from_state;
85 * A state. Can be used in DFA and NFA automatons.
87 struct REGEX_INTERNAL_State;
93 struct REGEX_INTERNAL_StateSet
98 struct REGEX_INTERNAL_State **states;
101 * Number of entries in *use* in the 'states' array.
106 * Length of the 'states' array.
113 * A state. Can be used in DFA and NFA automatons.
115 struct REGEX_INTERNAL_State
118 * This is a linked list to keep states in an automaton.
120 struct REGEX_INTERNAL_State *prev;
123 * This is a linked list to keep states in an automaton.
125 struct REGEX_INTERNAL_State *next;
128 * This is a multi DLL for StateSet_MDLL.
130 struct REGEX_INTERNAL_State *prev_SS;
133 * This is a multi DLL for StateSet_MDLL.
135 struct REGEX_INTERNAL_State *next_SS;
138 * This is a multi DLL for StateSet_MDLL Stack.
140 struct REGEX_INTERNAL_State *prev_ST;
143 * This is a multi DLL for StateSet_MDLL Stack.
145 struct REGEX_INTERNAL_State *next_ST;
153 * Unique state id that is used for traversing the automaton. It is guaranteed
154 * to be > 0 and < state_count.
156 unsigned int traversal_id;
159 * If this is an accepting state or not.
164 * Marking of the state. This is used for marking all visited states when
165 * traversing all states of an automaton and for cases where the state id
166 * cannot be used (dfa minimization).
171 * Marking the state as contained. This is used for checking, if the state is
172 * contained in a set in constant time.
177 * Marking the state as part of an SCC (Strongly Connected Component). All
178 * states with the same scc_id are part of the same SCC. scc_id is 0, if state
179 * is not a part of any SCC.
184 * Used for SCC detection.
189 * Used for SCC detection.
194 * Human readable name of the state. Used for debugging and graph
202 struct GNUNET_HashCode hash;
205 * Linear state ID accquired by depth-first-search. This ID should be used for
206 * storing information about the state in an array, because the 'id' of the
207 * state is not guaranteed to be linear. The 'dfs_id' is guaranteed to be > 0
208 * and < 'state_count'.
213 * Proof for this state.
218 * Number of transitions from this state to other states.
220 unsigned int transition_count;
223 * DLL of transitions.
225 struct REGEX_INTERNAL_Transition *transitions_head;
228 * DLL of transitions.
230 struct REGEX_INTERNAL_Transition *transitions_tail;
233 * Number of incoming transitions. Used for compressing DFA paths.
235 unsigned int incoming_transition_count;
238 * Set of states on which this state is based on. Used when creating a DFA out
239 * of several NFA states.
241 struct REGEX_INTERNAL_StateSet nfa_set;
246 * Type of an automaton.
248 enum REGEX_INTERNAL_AutomatonType
256 * Automaton representation.
258 struct REGEX_INTERNAL_Automaton
261 * Linked list of NFAs used for partial NFA creation.
263 struct REGEX_INTERNAL_Automaton *prev;
266 * Linked list of NFAs used for partial NFA creation.
268 struct REGEX_INTERNAL_Automaton *next;
271 * First state of the automaton. This is mainly used for constructing an NFA,
272 * where each NFA itself consists of one or more NFAs linked together.
274 struct REGEX_INTERNAL_State *start;
277 * End state of the partial NFA. This is undefined for DFAs
279 struct REGEX_INTERNAL_State *end;
282 * Number of states in the automaton.
284 unsigned int state_count;
289 struct REGEX_INTERNAL_State *states_head;
294 struct REGEX_INTERNAL_State *states_tail;
297 * Type of the automaton.
299 enum REGEX_INTERNAL_AutomatonType type;
307 * Canonical regex (result of RX->NFA->DFA->RX)
309 char *canonical_regex;
312 * GNUNET_YES, if multi strides have been added to the Automaton.
319 * Construct an NFA by parsing the regex string of length 'len'.
321 * @param regex regular expression string.
322 * @param len length of the string.
324 * @return NFA, needs to be freed using REGEX_INTERNAL_automaton_destroy.
326 struct REGEX_INTERNAL_Automaton *
327 REGEX_INTERNAL_construct_nfa (const char *regex, const size_t len);
331 * Function that get's passed to automaton traversal and is called before each
332 * next traversal from state 's' using transition 't' to check if traversal
333 * should proceed. Return GNUNET_NO to stop traversal or GNUNET_YES to continue.
335 * @param cls closure for the check.
336 * @param s current state in the traversal.
337 * @param t current transition from state 's' that will be used for the next
340 * @return GNUNET_YES to proceed traversal, GNUNET_NO to stop.
342 typedef int (*REGEX_INTERNAL_traverse_check) (void *cls,
343 struct REGEX_INTERNAL_State * s,
344 struct REGEX_INTERNAL_Transition * t);
348 * Function that is called with each state, when traversing an automaton.
350 * @param cls closure.
351 * @param count current count of the state, from 0 to a->state_count -1.
354 typedef void (*REGEX_INTERNAL_traverse_action) (void *cls,
355 const unsigned int count,
356 struct REGEX_INTERNAL_State * s);
360 * Traverses the given automaton using depth-first-search (DFS) from it's start
361 * state, visiting all reachable states and calling 'action' on each one of
364 * @param a automaton to be traversed.
365 * @param start start state, pass a->start or NULL to traverse the whole automaton.
366 * @param check function that is checked before advancing on each transition
368 * @param check_cls closure for check.
369 * @param action action to be performed on each state.
370 * @param action_cls closure for action
373 REGEX_INTERNAL_automaton_traverse (const struct REGEX_INTERNAL_Automaton *a,
374 struct REGEX_INTERNAL_State *start,
375 REGEX_INTERNAL_traverse_check check,
377 REGEX_INTERNAL_traverse_action action,
381 * Get the canonical regex of the given automaton.
382 * When constructing the automaton a proof is computed for each state,
383 * consisting of the regular expression leading to this state. A complete
384 * regex for the automaton can be computed by combining these proofs.
385 * As of now this function is only useful for testing.
387 * @param a automaton for which the canonical regex should be returned.
389 * @return canonical regex string.
392 REGEX_INTERNAL_get_canonical_regex (struct REGEX_INTERNAL_Automaton *a);
396 * Get the number of transitions that are contained in the given automaton.
398 * @param a automaton for which the number of transitions should be returned.
400 * @return number of transitions in the given automaton.
403 REGEX_INTERNAL_get_transition_count (struct REGEX_INTERNAL_Automaton *a);
407 * Context that contains an id counter for states and transitions as well as a
408 * DLL of automatons used as a stack for NFA construction.
410 struct REGEX_INTERNAL_Context
415 unsigned int state_id;
418 * Unique transition id.
420 unsigned int transition_id;
423 * DLL of REGEX_INTERNAL_Automaton's used as a stack.
425 struct REGEX_INTERNAL_Automaton *stack_head;
428 * DLL of REGEX_INTERNAL_Automaton's used as a stack.
430 struct REGEX_INTERNAL_Automaton *stack_tail;
435 * Adds multi-strided transitions to the given 'dfa'.
437 * @param regex_ctx regex context needed to add transitions to the automaton.
438 * @param dfa DFA to which the multi strided transitions should be added.
439 * @param stride_len length of the strides.
442 REGEX_INTERNAL_dfa_add_multi_strides (struct REGEX_INTERNAL_Context *regex_ctx,
443 struct REGEX_INTERNAL_Automaton *dfa,
444 const unsigned int stride_len);
448 #if 0 /* keep Emacsens' auto-indent happy */