xref: /openbsd-src/usr.bin/less/decode.c (revision 47911bd667ac77dc523b8a13ef40b012dbffa741)
1 /*	$OpenBSD: decode.c,v 1.3 2001/11/19 19:02:14 mpech Exp $	*/
2 
3 /*
4  * Copyright (c) 1984,1985,1989,1994,1995  Mark Nudelman
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice in the documentation and/or other materials provided with
14  *    the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY
17  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
19  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE
20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
21  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT
22  * OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
23  * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
24  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
25  * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN
26  * IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27  */
28 
29 
30 /*
31  * Routines to decode user commands.
32  *
33  * This is all table driven.
34  * A command table is a sequence of command descriptors.
35  * Each command descriptor is a sequence of bytes with the following format:
36  *	<c1><c2>...<cN><0><action>
37  * The characters c1,c2,...,cN are the command string; that is,
38  * the characters which the user must type.
39  * It is terminated by a null <0> byte.
40  * The byte after the null byte is the action code associated
41  * with the command string.
42  * If an action byte is OR-ed with A_EXTRA, this indicates
43  * that the option byte is followed by an extra string.
44  *
45  * There may be many command tables.
46  * The first (default) table is built-in.
47  * Other tables are read in from "lesskey" files.
48  * All the tables are linked together and are searched in order.
49  */
50 
51 #include "less.h"
52 #include "cmd.h"
53 #include "lesskey.h"
54 
55 extern int erase_char, kill_char;
56 
57 /*
58  * Command table is ordered roughly according to expected
59  * frequency of use, so the common commands are near the beginning.
60  */
61 static unsigned char cmdtable[] =
62 {
63 	'\r',0,				A_F_LINE,
64 	'\n',0,				A_F_LINE,
65 	'e',0,				A_F_LINE,
66 	'j',0,				A_F_LINE,
67 	CONTROL('E'),0,			A_F_LINE,
68 	CONTROL('N'),0,			A_F_LINE,
69 	'k',0,				A_B_LINE,
70 	'y',0,				A_B_LINE,
71 	CONTROL('Y'),0,			A_B_LINE,
72 	CONTROL('K'),0,			A_B_LINE,
73 	CONTROL('P'),0,			A_B_LINE,
74 	'J',0,				A_FF_LINE,
75 	'K',0,				A_BF_LINE,
76 	'Y',0,				A_BF_LINE,
77 	'd',0,				A_F_SCROLL,
78 	CONTROL('D'),0,			A_F_SCROLL,
79 	'u',0,				A_B_SCROLL,
80 	CONTROL('U'),0,			A_B_SCROLL,
81 	' ',0,				A_F_SCREEN,
82 	'f',0,				A_F_SCREEN,
83 	CONTROL('F'),0,			A_F_SCREEN,
84 	CONTROL('V'),0,			A_F_SCREEN,
85 	'b',0,				A_B_SCREEN,
86 	CONTROL('B'),0,			A_B_SCREEN,
87 	ESC,'v',0,			A_B_SCREEN,
88 	'z',0,				A_F_WINDOW,
89 	'w',0,				A_B_WINDOW,
90 	'F',0,				A_F_FOREVER,
91 	'R',0,				A_FREPAINT,
92 	'r',0,				A_REPAINT,
93 	CONTROL('R'),0,			A_REPAINT,
94 	CONTROL('L'),0,			A_REPAINT,
95 	ESC,'u',0,			A_UNDO_SEARCH,
96 	'g',0,				A_GOLINE,
97 	'<',0,				A_GOLINE,
98 	ESC,'<',0,			A_GOLINE,
99 	'p',0,				A_PERCENT,
100 	'%',0,				A_PERCENT,
101 	'{',0,				A_F_BRACKET|A_EXTRA,	'{','}',0,
102 	'}',0,				A_B_BRACKET|A_EXTRA,	'{','}',0,
103 	'(',0,				A_F_BRACKET|A_EXTRA,	'(',')',0,
104 	')',0,				A_B_BRACKET|A_EXTRA,	'(',')',0,
105 	'[',0,				A_F_BRACKET|A_EXTRA,	'[',']',0,
106 	']',0,				A_B_BRACKET|A_EXTRA,	'[',']',0,
107 	ESC,CONTROL('F'),0,		A_F_BRACKET,
108 	ESC,CONTROL('B'),0,		A_B_BRACKET,
109 	'G',0,				A_GOEND,
110 	ESC,'>',0,			A_GOEND,
111 	'>',0,				A_GOEND,
112 	'P',0,				A_GOPOS,
113 
114 	'0',0,				A_DIGIT,
115 	'1',0,				A_DIGIT,
116 	'2',0,				A_DIGIT,
117 	'3',0,				A_DIGIT,
118 	'4',0,				A_DIGIT,
119 	'5',0,				A_DIGIT,
120 	'6',0,				A_DIGIT,
121 	'7',0,				A_DIGIT,
122 	'8',0,				A_DIGIT,
123 	'9',0,				A_DIGIT,
124 
125 	'=',0,				A_STAT,
126 	CONTROL('G'),0,			A_STAT,
127 	':','f',0,			A_STAT,
128 	'/',0,				A_F_SEARCH,
129 	'?',0,				A_B_SEARCH,
130 	ESC,'/',0,			A_F_SEARCH|A_EXTRA,	'*',0,
131 	ESC,'?',0,			A_B_SEARCH|A_EXTRA,	'*',0,
132 	'n',0,				A_AGAIN_SEARCH,
133 	ESC,'n',0,			A_T_AGAIN_SEARCH,
134 	'N',0,				A_REVERSE_SEARCH,
135 	ESC,'N',0,			A_T_REVERSE_SEARCH,
136 	'm',0,				A_SETMARK,
137 	'\'',0,				A_GOMARK,
138 	CONTROL('X'),CONTROL('X'),0,	A_GOMARK,
139 	'E',0,				A_EXAMINE,
140 	':','e',0,			A_EXAMINE,
141 	CONTROL('X'),CONTROL('V'),0,	A_EXAMINE,
142 	':','n',0,			A_NEXT_FILE,
143 	':','p',0,			A_PREV_FILE,
144 	':','x',0,			A_INDEX_FILE,
145 	'-',0,				A_OPT_TOGGLE,
146 	':','t',0,			A_OPT_TOGGLE|A_EXTRA,	't',0,
147 	's',0,				A_OPT_TOGGLE|A_EXTRA,	'o',0,
148 	'_',0,				A_DISP_OPTION,
149 	'|',0,				A_PIPE,
150 	'v',0,				A_VISUAL,
151 	'!',0,				A_SHELL,
152 	'+',0,				A_FIRSTCMD,
153 
154 	'H',0,				A_HELP,
155 	'h',0,				A_HELP,
156 	'V',0,				A_VERSION,
157 	'q',0,				A_QUIT,
158 	':','q',0,			A_QUIT,
159 	':','Q',0,			A_QUIT,
160 	'Z','Z',0,			A_QUIT
161 };
162 
163 static unsigned char edittable[] =
164 {
165 	'\t',0,	    		EC_F_COMPLETE,	/* TAB */
166 	'\17',0,		EC_B_COMPLETE,	/* BACKTAB */
167 	'\14',0,		EC_EXPAND,	/* CTRL-L */
168 	CONTROL('V'),0,		EC_LITERAL,	/* BACKSLASH */
169 	CONTROL('A'),0,		EC_LITERAL,	/* BACKSLASH */
170    	ESC,'l',0,		EC_RIGHT,	/* ESC l */
171 	ESC,'h',0,		EC_LEFT,	/* ESC h */
172 	ESC,'b',0,		EC_W_LEFT,	/* ESC b */
173 	ESC,'w',0,		EC_W_RIGHT,	/* ESC w */
174 	ESC,'i',0,		EC_INSERT,	/* ESC i */
175 	ESC,'x',0,		EC_DELETE,	/* ESC x */
176 	ESC,'X',0,		EC_W_DELETE,	/* ESC X */
177 	ESC,'\b',0,		EC_W_BACKSPACE,	/* ESC BACKSPACE */
178 	ESC,'0',0,		EC_HOME,	/* ESC 0 */
179 	ESC,'$',0,		EC_END,		/* ESC $ */
180 	ESC,'k',0,		EC_UP,		/* ESC k */
181 	ESC,'j',0,		EC_DOWN,	/* ESC j */
182 	ESC,'\t',0,		EC_B_COMPLETE	/* ESC TAB */
183 };
184 
185 /*
186  * Structure to support a list of command tables.
187  */
188 struct tablelist
189 {
190 	struct tablelist *t_next;
191 	char *t_start;
192 	char *t_end;
193 };
194 
195 /*
196  * List of command tables and list of line-edit tables.
197  */
198 static struct tablelist *list_fcmd_tables = NULL;
199 static struct tablelist *list_ecmd_tables = NULL;
200 
201 
202 /*
203  * Initialize the command lists.
204  */
205 	public void
206 init_cmds()
207 {
208 	/*
209 	 * Add the default command tables.
210 	 */
211 	add_fcmd_table((char*)cmdtable, sizeof(cmdtable));
212 	add_ecmd_table((char*)edittable, sizeof(edittable));
213 	get_editkeys();
214 #if USERFILE
215 	/*
216 	 * Try to add the tables in the standard lesskey file "$HOME/.less".
217 	 */
218 	add_hometable();
219 #endif
220 }
221 
222 /*
223  *
224  */
225 	static int
226 add_cmd_table(tlist, buf, len)
227 	struct tablelist **tlist;
228 	char *buf;
229 	int len;
230 {
231 	struct tablelist *t;
232 
233 	if (len == 0)
234 		return (0);
235 	/*
236 	 * Allocate a tablelist structure, initialize it,
237 	 * and link it into the list of tables.
238 	 */
239 	if ((t = (struct tablelist *)
240 			calloc(1, sizeof(struct tablelist))) == NULL)
241 	{
242 		return (-1);
243 	}
244 	t->t_start = buf;
245 	t->t_end = buf + len;
246 	t->t_next = *tlist;
247 	*tlist = t;
248 	return (0);
249 }
250 
251 /*
252  * Add a command table.
253  */
254 	public void
255 add_fcmd_table(buf, len)
256 	char *buf;
257 	int len;
258 {
259 	if (add_cmd_table(&list_fcmd_tables, buf, len) < 0)
260 		error("Warning: some commands disabled", NULL_PARG);
261 }
262 
263 /*
264  * Add an editing command table.
265  */
266 	public void
267 add_ecmd_table(buf, len)
268 	char *buf;
269 	int len;
270 {
271 	if (add_cmd_table(&list_ecmd_tables, buf, len) < 0)
272 		error("Warning: some edit commands disabled", NULL_PARG);
273 }
274 
275 /*
276  * Search a single command table for the command string in cmd.
277  */
278 	public int
279 cmd_search(cmd, table, endtable, sp)
280 	char *cmd;
281 	char *table;
282 	char *endtable;
283 	char **sp;
284 {
285 	char *p;
286 	char *q;
287 	int a;
288 
289 	for (p = table, q = cmd;  p < endtable;  p++, q++)
290 	{
291 		if (*p == *q)
292 		{
293 			/*
294 			 * Current characters match.
295 			 * If we're at the end of the string, we've found it.
296 			 * Return the action code, which is the character
297 			 * after the null at the end of the string
298 			 * in the command table.
299 			 */
300 			if (*p == '\0')
301 			{
302 				a = *++p & 0377;
303 				if (a == A_END_LIST)
304 				{
305 					/*
306 					 * We get here only if the original
307 					 * cmd string passed in was empty ("").
308 					 * I don't think that can happen,
309 					 * but just in case ...
310 					 */
311 					return (A_UINVALID);
312 				}
313 				/*
314 				 * Check for an "extra" string.
315 				 */
316 				if (a & A_EXTRA)
317 				{
318 					*sp = ++p;
319 					a &= ~A_EXTRA;
320 				} else
321 					*sp = NULL;
322 				return (a);
323 			}
324 		} else if (*q == '\0')
325 		{
326 			/*
327 			 * Hit the end of the user's command,
328 			 * but not the end of the string in the command table.
329 			 * The user's command is incomplete.
330 			 */
331 			return (A_PREFIX);
332 		} else
333 		{
334 			/*
335 			 * Not a match.
336 			 * Skip ahead to the next command in the
337 			 * command table, and reset the pointer
338 			 * to the beginning of the user's command.
339 			 */
340 			if (*p == '\0' && p[1] == A_END_LIST)
341 			{
342 				/*
343 				 * A_END_LIST is a special marker that tells
344 				 * us to abort the cmd search.
345 				 */
346 				return (A_UINVALID);
347 			}
348 			while (*p++ != '\0') ;
349 			if (*p & A_EXTRA)
350 				while (*++p != '\0') ;
351 			q = cmd-1;
352 		}
353 	}
354 	/*
355 	 * No match found in the entire command table.
356 	 */
357 	return (A_INVALID);
358 }
359 
360 /*
361  * Decode a command character and return the associated action.
362  * The "extra" string, if any, is returned in sp.
363  */
364 	static int
365 cmd_decode(tlist, cmd, sp)
366 	struct tablelist *tlist;
367 	char *cmd;
368 	char **sp;
369 {
370 	struct tablelist *t;
371 	int action = A_INVALID;
372 
373 	/*
374 	 * Search thru all the command tables.
375 	 * Stop when we find an action which is not A_INVALID.
376 	 */
377 	for (t = tlist;  t != NULL;  t = t->t_next)
378 	{
379 		action = cmd_search(cmd, t->t_start, t->t_end, sp);
380 		if (action != A_INVALID)
381 			break;
382 	}
383 	return (action);
384 }
385 
386 /*
387  * Decode a command from the cmdtables list.
388  */
389 	public int
390 fcmd_decode(cmd, sp)
391 	char *cmd;
392 	char **sp;
393 {
394 	return (cmd_decode(list_fcmd_tables, cmd, sp));
395 }
396 
397 /*
398  * Decode a command from the edittables list.
399  */
400 	public int
401 ecmd_decode(cmd, sp)
402 	char *cmd;
403 	char **sp;
404 {
405 	return (cmd_decode(list_ecmd_tables, cmd, sp));
406 }
407 
408 #if USERFILE
409 	static int
410 gint(sp)
411 	char **sp;
412 {
413 	int n;
414 
415 	n = *(*sp)++;
416 	n += *(*sp)++ * KRADIX;
417 	return (n);
418 }
419 
420 	static int
421 old_lesskey(buf, len)
422 	char *buf;
423 	int len;
424 {
425 	/*
426 	 * Old-style lesskey file.
427 	 * The file must end with either
428 	 *     ...,cmd,0,action
429 	 * or  ...,cmd,0,action|A_EXTRA,string,0
430 	 * So the last byte or the second to last byte must be zero.
431 	 */
432 	if (buf[len-1] != '\0' && buf[len-2] != '\0')
433 		return (-1);
434 	add_fcmd_table(buf, len);
435 	return (0);
436 }
437 
438 	static int
439 new_lesskey(buf, len)
440 	char *buf;
441 	int len;
442 {
443 	char *p;
444 	int c;
445 	int done;
446 	int n;
447 
448 	/*
449 	 * New-style lesskey file.
450 	 * Extract the pieces.
451 	 */
452 	if (buf[len-3] != C0_END_LESSKEY_MAGIC ||
453 	    buf[len-2] != C1_END_LESSKEY_MAGIC ||
454 	    buf[len-1] != C2_END_LESSKEY_MAGIC)
455 		return (-1);
456 	p = buf + 4;
457 	done = 0;
458 	while (!done)
459 	{
460 		c = *p++;
461 		switch (c)
462 		{
463 		case CMD_SECTION:
464 			n = gint(&p);
465 			add_fcmd_table(p, n);
466 			p += n;
467 			break;
468 		case EDIT_SECTION:
469 			n = gint(&p);
470 			add_ecmd_table(p, n);
471 			p += n;
472 			break;
473 		case END_SECTION:
474 			done = 1;
475 			break;
476 		default:
477 			free(buf);
478 			return (-1);
479 		}
480 	}
481 	return (0);
482 }
483 
484 /*
485  * Set up a user command table, based on a "lesskey" file.
486  */
487 	public int
488 lesskey(filename)
489 	char *filename;
490 {
491 	char *buf;
492 	POSITION len;
493 	long n;
494 	int f;
495 
496 	/*
497 	 * Try to open the lesskey file.
498 	 */
499 	f = open(filename, OPEN_READ);
500 	if (f < 0)
501 		return (1);
502 
503 	/*
504 	 * Read the file into a buffer.
505 	 * We first figure out the size of the file and allocate space for it.
506 	 * {{ Minimal error checking is done here.
507 	 *    A garbage .less file will produce strange results.
508 	 *    To avoid a large amount of error checking code here, we
509 	 *    rely on the lesskey program to generate a good .less file. }}
510 	 */
511 	len = filesize(f);
512 	if (len == NULL_POSITION || len < 3)
513 	{
514 		/*
515 		 * Bad file (valid file must have at least 3 chars).
516 		 */
517 		close(f);
518 		return (-1);
519 	}
520 	if ((buf = (char *) calloc((int)len, sizeof(char))) == NULL)
521 	{
522 		close(f);
523 		return (-1);
524 	}
525 	if (lseek(f, (off_t)0, 0) == BAD_LSEEK)
526 	{
527 		free(buf);
528 		close(f);
529 		return (-1);
530 	}
531 	n = read(f, buf, (unsigned int) len);
532 	close(f);
533 	if (n != len)
534 	{
535 		free(buf);
536 		return (-1);
537 	}
538 
539 	/*
540 	 * Figure out if this is an old-style (before version 241)
541 	 * or new-style lesskey file format.
542 	 */
543 	if (buf[0] != C0_LESSKEY_MAGIC || buf[1] != C1_LESSKEY_MAGIC ||
544 	    buf[2] != C2_LESSKEY_MAGIC || buf[3] != C3_LESSKEY_MAGIC)
545 		return (old_lesskey(buf, (int)len));
546 	return (new_lesskey(buf, (int)len));
547 }
548 
549 /*
550  * Add the standard lesskey file "$HOME/.less"
551  */
552 	public void
553 add_hometable()
554 {
555 	char *filename;
556 	PARG parg;
557 
558 	filename = homefile(LESSKEYFILE);
559 	if (filename == NULL)
560 		return;
561 	if (lesskey(filename) < 0)
562 	{
563 		parg.p_string = filename;
564 		error("Cannot use lesskey file \"%s\"", &parg);
565 	}
566 	free(filename);
567 }
568 #endif
569 
570 /*
571  * See if a char is a special line-editing command.
572  */
573 	public int
574 editchar(c, flags)
575 	int c;
576 	int flags;
577 {
578 	int action;
579 	int nch;
580 	char *s;
581 	char usercmd[MAX_CMDLEN+1];
582 
583 	/*
584 	 * An editing character could actually be a sequence of characters;
585 	 * for example, an escape sequence sent by pressing the uparrow key.
586 	 * To match the editing string, we use the command decoder
587 	 * but give it the edit-commands command table
588 	 * This table is constructed to match the user's keyboard.
589 	 */
590 	if (c == erase_char)
591 		return (EC_BACKSPACE);
592 	if (c == kill_char)
593 		return (EC_LINEKILL);
594 
595 	/*
596 	 * Collect characters in a buffer.
597 	 * Start with the one we have, and get more if we need them.
598 	 */
599 	nch = 0;
600 	do {
601 	  	if (nch > 0)
602 			c = getcc();
603 		usercmd[nch] = c;
604 		usercmd[nch+1] = '\0';
605 		nch++;
606 		action = ecmd_decode(usercmd, &s);
607 	} while (action == A_PREFIX);
608 
609 	if (flags & EC_NOHISTORY)
610 	{
611 		/*
612 		 * The caller says there is no history list.
613 		 * Reject any history-manipulation action.
614 		 */
615 		switch (action)
616 		{
617 		case EC_UP:
618 		case EC_DOWN:
619 			action = A_INVALID;
620 			break;
621 		}
622 	}
623 	if (flags & EC_NOCOMPLETE)
624 	{
625 		/*
626 		 * The caller says we don't want any filename completion cmds.
627 		 * Reject them.
628 		 */
629 		switch (action)
630 		{
631 		case EC_F_COMPLETE:
632 		case EC_B_COMPLETE:
633 		case EC_EXPAND:
634 			action = A_INVALID;
635 			break;
636 		}
637 	}
638 	if ((flags & EC_PEEK) || action == A_INVALID)
639 	{
640 		/*
641 		 * We're just peeking, or we didn't understand the command.
642 		 * Unget all the characters we read in the loop above.
643 		 * This does NOT include the original character that was
644 		 * passed in as a parameter.
645 		 */
646 		while (nch > 1) {
647 			ungetcc(usercmd[--nch]);
648 		}
649 	} else
650 	{
651 		if (s != NULL)
652 			ungetsc(s);
653 	}
654 	return action;
655 }
656 
657